€•ÔŒsphinx.addnodes”Œdocument”“”)”}”(Œ rawsource”Œ”Œchildren”]”(Œtranslations”Œ LanguagesNode”“”)”}”(hhh]”(hŒpending_xref”“”)”}”(hhh]”Œdocutils.nodes”ŒText”“”ŒChinese (Simplified)”…””}”Œparent”hsbaŒ attributes”}”(Œids”]”Œclasses”]”Œnames”]”Œdupnames”]”Œbackrefs”]”Œ refdomain”Œstd”Œreftype”Œdoc”Œ reftarget”Œ)/translations/zh_CN/admin-guide/cgroup-v2”Œmodname”NŒ classname”NŒrefexplicit”ˆuŒtagname”hhhubh)”}”(hhh]”hŒChinese (Traditional)”…””}”hh2sbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ)/translations/zh_TW/admin-guide/cgroup-v2”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒItalian”…””}”hhFsbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ)/translations/it_IT/admin-guide/cgroup-v2”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒJapanese”…””}”hhZsbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ)/translations/ja_JP/admin-guide/cgroup-v2”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒKorean”…””}”hhnsbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ)/translations/ko_KR/admin-guide/cgroup-v2”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒSpanish”…””}”hh‚sbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ)/translations/sp_SP/admin-guide/cgroup-v2”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubeh}”(h]”h ]”h"]”h$]”h&]”Œcurrent_language”ŒEnglish”uh1h hhŒ _document”hŒsource”NŒline”NubhŒtarget”“”)”}”(hŒ.. _cgroup-v2:”h]”h}”(h]”h ]”h"]”h$]”h&]”Œrefid”Œ cgroup-v2”uh1h¡h KhhhžhhŸŒC/var/lib/git/docbuild/linux/Documentation/admin-guide/cgroup-v2.rst”ubhŒsection”“”)”}”(hhh]”(hŒtitle”“”)”}”(hŒControl Group v2”h]”hŒControl Group v2”…””}”(hh·hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhh²hžhhŸh¯h KubhŒ field_list”“”)”}”(hhh]”(hŒfield”“”)”}”(hhh]”(hŒ field_name”“”)”}”(hŒDate”h]”hŒDate”…””}”(hhÑhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÏhhÌhŸh¯h KubhŒ field_body”“”)”}”(hŒ October, 2015”h]”hŒ paragraph”“”)”}”(hhãh]”hŒ October, 2015”…””}”(hhçhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Khháubah}”(h]”h ]”h"]”h$]”h&]”uh1hßhhÌubeh}”(h]”h ]”h"]”h$]”h&]”uh1hÊhŸh¯h KhhÇhžhubhË)”}”(hhh]”(hÐ)”}”(hŒAuthor”h]”hŒAuthor”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÏhjhŸh¯h Kubhà)”}”(hŒTejun Heo ”h]”hæ)”}”(hŒTejun Heo ”h]”(hŒTejun Heo <”…””}”(hjhžhhŸNh NubhŒ reference”“”)”}”(hŒ tj@kernel.org”h]”hŒ tj@kernel.org”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”Œrefuri”Œmailto:tj@kernel.org”uh1jhjubhŒ>”…””}”(hjhžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Khjubah}”(h]”h ]”h"]”h$]”h&]”uh1hßhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1hÊhŸh¯h KhhÇhžhubeh}”(h]”h ]”h"]”h$]”h&]”uh1hÅhh²hžhhŸh¯h Kubhæ)”}”(hXhThis is the authoritative documentation on the design, interface and conventions of cgroup v2. It describes all userland-visible aspects of cgroup including core and specific controller behaviors. All future changes must be reflected in this document. Documentation for v1 is available under :ref:`Documentation/admin-guide/cgroup-v1/index.rst `.”h]”(hX'This is the authoritative documentation on the design, interface and conventions of cgroup v2. It describes all userland-visible aspects of cgroup including core and specific controller behaviors. All future changes must be reflected in this document. Documentation for v1 is available under ”…””}”(hjKhžhhŸNh Nubh)”}”(hŒ@:ref:`Documentation/admin-guide/cgroup-v1/index.rst `”h]”hŒinline”“”)”}”(hjUh]”hŒ-Documentation/admin-guide/cgroup-v1/index.rst”…””}”(hjYhžhhŸNh Nubah}”(h]”h ]”(Œxref”Œstd”Œstd-ref”eh"]”h$]”h&]”uh1jWhjSubah}”(h]”h ]”h"]”h$]”h&]”Œrefdoc”Œadmin-guide/cgroup-v2”Œ refdomain”jdŒreftype”Œref”Œrefexplicit”ˆŒrefwarn”ˆŒ reftarget”Œ cgroup-v1”uh1hhŸh¯h K hjKubhŒ.”…””}”(hjKhžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h K hh²hžhubhŒcomment”“”)”}”(hXÜCONTENTS 1. Introduction 1-1. Terminology 1-2. What is cgroup? 2. Basic Operations 2-1. Mounting 2-2. Organizing Processes and Threads 2-2-1. Processes 2-2-2. Threads 2-3. [Un]populated Notification 2-4. Controlling Controllers 2-4-1. Enabling and Disabling 2-4-2. Top-down Constraint 2-4-3. No Internal Process Constraint 2-5. Delegation 2-5-1. Model of Delegation 2-5-2. Delegation Containment 2-6. Guidelines 2-6-1. Organize Once and Control 2-6-2. Avoid Name Collisions 3. Resource Distribution Models 3-1. Weights 3-2. Limits 3-3. Protections 3-4. Allocations 4. Interface Files 4-1. Format 4-2. Conventions 4-3. Core Interface Files 5. Controllers 5-1. CPU 5-1-1. CPU Interface Files 5-2. Memory 5-2-1. Memory Interface Files 5-2-2. Usage Guidelines 5-2-3. Memory Ownership 5-3. IO 5-3-1. IO Interface Files 5-3-2. Writeback 5-3-3. IO Latency 5-3-3-1. How IO Latency Throttling Works 5-3-3-2. IO Latency Interface Files 5-3-4. IO Priority 5-4. PID 5-4-1. PID Interface Files 5-5. Cpuset 5.5-1. Cpuset Interface Files 5-6. Device 5-7. RDMA 5-7-1. RDMA Interface Files 5-8. DMEM 5-9. HugeTLB 5.9-1. HugeTLB Interface Files 5-10. Misc 5.10-1 Miscellaneous cgroup Interface Files 5.10-2 Migration and Ownership 5-11. Others 5-11-1. perf_event 5-N. Non-normative information 5-N-1. CPU controller root cgroup process behaviour 5-N-2. IO controller root cgroup process behaviour 6. Namespace 6-1. Basics 6-2. The Root and Views 6-3. Migration and setns(2) 6-4. Interaction with Other Namespaces P. Information on Kernel Programming P-1. Filesystem Support for Writeback D. Deprecated v1 Core Features R. Issues with v1 and Rationales for v2 R-1. Multiple Hierarchies R-2. Thread Granularity R-3. Competition Between Inner Nodes and Threads R-4. Other Interface Issues R-5. Controller Issues and Remedies R-5-1. Memory”h]”hXÜCONTENTS 1. Introduction 1-1. Terminology 1-2. What is cgroup? 2. Basic Operations 2-1. Mounting 2-2. Organizing Processes and Threads 2-2-1. Processes 2-2-2. Threads 2-3. [Un]populated Notification 2-4. Controlling Controllers 2-4-1. Enabling and Disabling 2-4-2. Top-down Constraint 2-4-3. No Internal Process Constraint 2-5. Delegation 2-5-1. Model of Delegation 2-5-2. Delegation Containment 2-6. Guidelines 2-6-1. Organize Once and Control 2-6-2. Avoid Name Collisions 3. Resource Distribution Models 3-1. Weights 3-2. Limits 3-3. Protections 3-4. Allocations 4. Interface Files 4-1. Format 4-2. Conventions 4-3. Core Interface Files 5. Controllers 5-1. CPU 5-1-1. CPU Interface Files 5-2. Memory 5-2-1. Memory Interface Files 5-2-2. Usage Guidelines 5-2-3. Memory Ownership 5-3. IO 5-3-1. IO Interface Files 5-3-2. Writeback 5-3-3. IO Latency 5-3-3-1. How IO Latency Throttling Works 5-3-3-2. IO Latency Interface Files 5-3-4. IO Priority 5-4. PID 5-4-1. PID Interface Files 5-5. Cpuset 5.5-1. Cpuset Interface Files 5-6. Device 5-7. RDMA 5-7-1. RDMA Interface Files 5-8. DMEM 5-9. HugeTLB 5.9-1. HugeTLB Interface Files 5-10. Misc 5.10-1 Miscellaneous cgroup Interface Files 5.10-2 Migration and Ownership 5-11. Others 5-11-1. perf_event 5-N. Non-normative information 5-N-1. CPU controller root cgroup process behaviour 5-N-2. IO controller root cgroup process behaviour 6. Namespace 6-1. Basics 6-2. The Root and Views 6-3. Migration and setns(2) 6-4. Interaction with Other Namespaces P. Information on Kernel Programming P-1. Filesystem Support for Writeback D. Deprecated v1 Core Features R. Issues with v1 and Rationales for v2 R-1. Multiple Hierarchies R-2. Thread Granularity R-3. Competition Between Inner Nodes and Threads R-4. Other Interface Issues R-5. Controller Issues and Remedies R-5-1. Memory”…””}”hj„sbah}”(h]”h ]”h"]”h$]”h&]”Œ xml:space”Œpreserve”uh1j‚hh²hžhhŸh¯h K^ubh±)”}”(hhh]”(h¶)”}”(hŒIntroduction”h]”hŒIntroduction”…””}”(hj—hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj”hžhhŸh¯h K`ubh±)”}”(hhh]”(h¶)”}”(hŒTerminology”h]”hŒTerminology”…””}”(hj¨hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj¥hžhhŸh¯h Kcubhæ)”}”(hX"cgroup" stands for "control group" and is never capitalized. The singular form is used to designate the whole feature and also as a qualifier as in "cgroup controllers". When explicitly referring to multiple individual control groups, the plural form "cgroups" is used.”h]”hX â€œcgroupâ€ stands for â€œcontrol groupâ€ and is never capitalized. The singular form is used to designate the whole feature and also as a qualifier as in â€œcgroup controllersâ€. When explicitly referring to multiple individual control groups, the plural form â€œcgroupsâ€ is used.”…””}”(hj¶hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Kehj¥hžhubeh}”(h]”Œterminology”ah ]”h"]”Œterminology”ah$]”h&]”uh1h°hj”hžhhŸh¯h Kcubh±)”}”(hhh]”(h¶)”}”(hŒWhat is cgroup?”h]”hŒWhat is cgroup?”…””}”(hjÏhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjÌhžhhŸh¯h Klubhæ)”}”(hŒ—cgroup is a mechanism to organize processes hierarchically and distribute system resources along the hierarchy in a controlled and configurable manner.”h]”hŒ—cgroup is a mechanism to organize processes hierarchically and distribute system resources along the hierarchy in a controlled and configurable manner.”…””}”(hjÝhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h KnhjÌhžhubhæ)”}”(hXccgroup is largely composed of two parts - the core and controllers. cgroup core is primarily responsible for hierarchically organizing processes. A cgroup controller is usually responsible for distributing a specific type of system resource along the hierarchy although there are utility controllers which serve purposes other than resource distribution.”h]”hXccgroup is largely composed of two parts - the core and controllers. cgroup core is primarily responsible for hierarchically organizing processes. A cgroup controller is usually responsible for distributing a specific type of system resource along the hierarchy although there are utility controllers which serve purposes other than resource distribution.”…””}”(hjëhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h KrhjÌhžhubhæ)”}”(hXtcgroups form a tree structure and every process in the system belongs to one and only one cgroup. All threads of a process belong to the same cgroup. On creation, all processes are put in the cgroup that the parent process belongs to at the time. A process can be migrated to another cgroup. Migration of a process doesn't affect already existing descendant processes.”h]”hXvcgroups form a tree structure and every process in the system belongs to one and only one cgroup. All threads of a process belong to the same cgroup. On creation, all processes are put in the cgroup that the parent process belongs to at the time. A process can be migrated to another cgroup. Migration of a process doesnâ€™t affect already existing descendant processes.”…””}”(hjùhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h KyhjÌhžhubhæ)”}”(hXõFollowing certain structural constraints, controllers may be enabled or disabled selectively on a cgroup. All controller behaviors are hierarchical - if a controller is enabled on a cgroup, it affects all processes which belong to the cgroups consisting the inclusive sub-hierarchy of the cgroup. When a controller is enabled on a nested cgroup, it always restricts the resource distribution further. The restrictions set closer to the root in the hierarchy can not be overridden from further away.”h]”hXõFollowing certain structural constraints, controllers may be enabled or disabled selectively on a cgroup. All controller behaviors are hierarchical - if a controller is enabled on a cgroup, it affects all processes which belong to the cgroups consisting the inclusive sub-hierarchy of the cgroup. When a controller is enabled on a nested cgroup, it always restricts the resource distribution further. The restrictions set closer to the root in the hierarchy can not be overridden from further away.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h K€hjÌhžhubeh}”(h]”Œwhat-is-cgroup”ah ]”h"]”Œwhat is cgroup?”ah$]”h&]”uh1h°hj”hžhhŸh¯h Klubeh}”(h]”Œintroduction”ah ]”h"]”Œintroduction”ah$]”h&]”uh1h°hh²hžhhŸh¯h K`ubh±)”}”(hhh]”(h¶)”}”(hŒBasic Operations”h]”hŒBasic Operations”…””}”(hj(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj%hžhhŸh¯h K‹ubh±)”}”(hhh]”(h¶)”}”(hŒMounting”h]”hŒMounting”…””}”(hj9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj6hžhhŸh¯h KŽubhæ)”}”(hŒzUnlike v1, cgroup v2 has only single hierarchy. The cgroup v2 hierarchy can be mounted with the following mount command::”h]”hŒyUnlike v1, cgroup v2 has only single hierarchy. The cgroup v2 hierarchy can be mounted with the following mount command:”…””}”(hjGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Khj6hžhubhŒ literal_block”“”)”}”(hŒ$# mount -t cgroup2 none $MOUNT_POINT”h]”hŒ$# mount -t cgroup2 none $MOUNT_POINT”…””}”hjWsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h K“hj6hžhubhæ)”}”(hX“cgroup2 filesystem has the magic number 0x63677270 ("cgrp"). All controllers which support v2 and are not bound to a v1 hierarchy are automatically bound to the v2 hierarchy and show up at the root. Controllers which are not in active use in the v2 hierarchy can be bound to other hierarchies. This allows mixing v2 hierarchy with the legacy v1 multiple hierarchies in a fully backward compatible way.”h]”hX—cgroup2 filesystem has the magic number 0x63677270 (â€œcgrpâ€). All controllers which support v2 and are not bound to a v1 hierarchy are automatically bound to the v2 hierarchy and show up at the root. Controllers which are not in active use in the v2 hierarchy can be bound to other hierarchies. This allows mixing v2 hierarchy with the legacy v1 multiple hierarchies in a fully backward compatible way.”…””}”(hjehžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h K•hj6hžhubhæ)”}”(hXvA controller can be moved across hierarchies only after the controller is no longer referenced in its current hierarchy. Because per-cgroup controller states are destroyed asynchronously and controllers may have lingering references, a controller may not show up immediately on the v2 hierarchy after the final umount of the previous hierarchy. Similarly, a controller should be fully disabled to be moved out of the unified hierarchy and it may take some time for the disabled controller to become available for other hierarchies; furthermore, due to inter-controller dependencies, other controllers may need to be disabled too.”h]”hXvA controller can be moved across hierarchies only after the controller is no longer referenced in its current hierarchy. Because per-cgroup controller states are destroyed asynchronously and controllers may have lingering references, a controller may not show up immediately on the v2 hierarchy after the final umount of the previous hierarchy. Similarly, a controller should be fully disabled to be moved out of the unified hierarchy and it may take some time for the disabled controller to become available for other hierarchies; furthermore, due to inter-controller dependencies, other controllers may need to be disabled too.”…””}”(hjshžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Kœhj6hžhubhæ)”}”(hX)While useful for development and manual configurations, moving controllers dynamically between the v2 and other hierarchies is strongly discouraged for production use. It is recommended to decide the hierarchies and controller associations before starting using the controllers after system boot.”h]”hX)While useful for development and manual configurations, moving controllers dynamically between the v2 and other hierarchies is strongly discouraged for production use. It is recommended to decide the hierarchies and controller associations before starting using the controllers after system boot.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h K§hj6hžhubhæ)”}”(hXKDuring transition to v2, system management software might still automount the v1 cgroup filesystem and so hijack all controllers during boot, before manual intervention is possible. To make testing and experimenting easier, the kernel parameter cgroup_no_v1= allows disabling controllers in v1 and make them always available in v2.”h]”hXKDuring transition to v2, system management software might still automount the v1 cgroup filesystem and so hijack all controllers during boot, before manual intervention is possible. To make testing and experimenting easier, the kernel parameter cgroup_no_v1= allows disabling controllers in v1 and make them always available in v2.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Khj6hžhubhæ)”}”(hŒ9cgroup v2 currently supports the following mount options.”h]”hŒ9cgroup v2 currently supports the following mount options.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h K³hj6hžhubhŒblock_quote”“”)”}”(hX¡ nsdelegate Consider cgroup namespaces as delegation boundaries. This option is system wide and can only be set on mount or modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts. Please refer to the Delegation section for details. favordynmods Reduce the latencies of dynamic cgroup modifications such as task migrations and controller on/offs at the cost of making hot path operations such as forks and exits more expensive. The static usage pattern of creating a cgroup, enabling controllers, and then seeding it with CLONE_INTO_CGROUP is not affected by this option. memory_localevents Only populate memory.events with data for the current cgroup, and not any subtrees. This is legacy behaviour, the default behaviour without this option is to include subtree counts. This option is system wide and can only be set on mount or modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts. memory_recursiveprot Recursively apply memory.min and memory.low protection to entire subtrees, without requiring explicit downward propagation into leaf cgroups. This allows protecting entire subtrees from one another, while retaining free competition within those subtrees. This should have been the default behavior but is a mount-option to avoid regressing setups relying on the original semantics (e.g. specifying bogusly high 'bypass' protection values at higher tree levels). memory_hugetlb_accounting Count HugeTLB memory usage towards the cgroup's overall memory usage for the memory controller (for the purpose of statistics reporting and memory protetion). This is a new behavior that could regress existing setups, so it must be explicitly opted in with this mount option. A few caveats to keep in mind: * There is no HugeTLB pool management involved in the memory controller. The pre-allocated pool does not belong to anyone. Specifically, when a new HugeTLB folio is allocated to the pool, it is not accounted for from the perspective of the memory controller. It is only charged to a cgroup when it is actually used (for e.g at page fault time). Host memory overcommit management has to consider this when configuring hard limits. In general, HugeTLB pool management should be done via other mechanisms (such as the HugeTLB controller). * Failure to charge a HugeTLB folio to the memory controller results in SIGBUS. This could happen even if the HugeTLB pool still has pages available (but the cgroup limit is hit and reclaim attempt fails). * Charging HugeTLB memory towards the memory controller affects memory protection and reclaim dynamics. Any userspace tuning (of low, min limits for e.g) needs to take this into account. * HugeTLB pages utilized while this option is not selected will not be tracked by the memory controller (even if cgroup v2 is remounted later on). pids_localevents The option restores v1-like behavior of pids.events:max, that is only local (inside cgroup proper) fork failures are counted. Without this option pids.events.max represents any pids.max enforcemnt across cgroup's subtree. ”h]”hŒdefinition_list”“”)”}”(hhh]”(hŒdefinition_list_item”“”)”}”(hXnsdelegate Consider cgroup namespaces as delegation boundaries. This option is system wide and can only be set on mount or modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts. Please refer to the Delegation section for details. ”h]”(hŒterm”“”)”}”(hŒ nsdelegate”h]”hŒ nsdelegate”…””}”(hj¾hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Kºhj¸ubhŒ definition”“”)”}”(hhh]”hæ)”}”(hXConsider cgroup namespaces as delegation boundaries. This option is system wide and can only be set on mount or modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts. Please refer to the Delegation section for details.”h]”hXConsider cgroup namespaces as delegation boundaries. This option is system wide and can only be set on mount or modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts. Please refer to the Delegation section for details.”…””}”(hjÑhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h K¶hjÎubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj¸ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Kºhj³ubj·)”}”(hXSfavordynmods Reduce the latencies of dynamic cgroup modifications such as task migrations and controller on/offs at the cost of making hot path operations such as forks and exits more expensive. The static usage pattern of creating a cgroup, enabling controllers, and then seeding it with CLONE_INTO_CGROUP is not affected by this option. ”h]”(j½)”}”(hŒfavordynmods”h]”hŒfavordynmods”…””}”(hjïhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h KÂhjëubjÍ)”}”(hhh]”hæ)”}”(hXEReduce the latencies of dynamic cgroup modifications such as task migrations and controller on/offs at the cost of making hot path operations such as forks and exits more expensive. The static usage pattern of creating a cgroup, enabling controllers, and then seeding it with CLONE_INTO_CGROUP is not affected by this option.”h]”hXEReduce the latencies of dynamic cgroup modifications such as task migrations and controller on/offs at the cost of making hot path operations such as forks and exits more expensive. The static usage pattern of creating a cgroup, enabling controllers, and then seeding it with CLONE_INTO_CGROUP is not affected by this option.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h K½hjýubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjëubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h KÂhj³ubj·)”}”(hXpmemory_localevents Only populate memory.events with data for the current cgroup, and not any subtrees. This is legacy behaviour, the default behaviour without this option is to include subtree counts. This option is system wide and can only be set on mount or modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts. ”h]”(j½)”}”(hŒmemory_localevents”h]”hŒmemory_localevents”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h KÊhjubjÍ)”}”(hhh]”hæ)”}”(hX\Only populate memory.events with data for the current cgroup, and not any subtrees. This is legacy behaviour, the default behaviour without this option is to include subtree counts. This option is system wide and can only be set on mount or modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts.”h]”hX\Only populate memory.events with data for the current cgroup, and not any subtrees. This is legacy behaviour, the default behaviour without this option is to include subtree counts. This option is system wide and can only be set on mount or modified through remount from the init namespace. The mount option is ignored on non-init namespace mounts.”…””}”(hj/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h KÅhj,ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h KÊhj³ubj·)”}”(hXåmemory_recursiveprot Recursively apply memory.min and memory.low protection to entire subtrees, without requiring explicit downward propagation into leaf cgroups. This allows protecting entire subtrees from one another, while retaining free competition within those subtrees. This should have been the default behavior but is a mount-option to avoid regressing setups relying on the original semantics (e.g. specifying bogusly high 'bypass' protection values at higher tree levels). ”h]”(j½)”}”(hŒmemory_recursiveprot”h]”hŒmemory_recursiveprot”…””}”(hjMhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h KÔhjIubjÍ)”}”(hhh]”hæ)”}”(hXÏRecursively apply memory.min and memory.low protection to entire subtrees, without requiring explicit downward propagation into leaf cgroups. This allows protecting entire subtrees from one another, while retaining free competition within those subtrees. This should have been the default behavior but is a mount-option to avoid regressing setups relying on the original semantics (e.g. specifying bogusly high 'bypass' protection values at higher tree levels).”h]”hXÓRecursively apply memory.min and memory.low protection to entire subtrees, without requiring explicit downward propagation into leaf cgroups. This allows protecting entire subtrees from one another, while retaining free competition within those subtrees. This should have been the default behavior but is a mount-option to avoid regressing setups relying on the original semantics (e.g. specifying bogusly high â€˜bypassâ€™ protection values at higher tree levels).”…””}”(hj^hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h KÍhj[ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjIubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h KÔhj³ubj·)”}”(hX¡memory_hugetlb_accounting Count HugeTLB memory usage towards the cgroup's overall memory usage for the memory controller (for the purpose of statistics reporting and memory protetion). This is a new behavior that could regress existing setups, so it must be explicitly opted in with this mount option. A few caveats to keep in mind: * There is no HugeTLB pool management involved in the memory controller. The pre-allocated pool does not belong to anyone. Specifically, when a new HugeTLB folio is allocated to the pool, it is not accounted for from the perspective of the memory controller. It is only charged to a cgroup when it is actually used (for e.g at page fault time). Host memory overcommit management has to consider this when configuring hard limits. In general, HugeTLB pool management should be done via other mechanisms (such as the HugeTLB controller). * Failure to charge a HugeTLB folio to the memory controller results in SIGBUS. This could happen even if the HugeTLB pool still has pages available (but the cgroup limit is hit and reclaim attempt fails). * Charging HugeTLB memory towards the memory controller affects memory protection and reclaim dynamics. Any userspace tuning (of low, min limits for e.g) needs to take this into account. * HugeTLB pages utilized while this option is not selected will not be tracked by the memory controller (even if cgroup v2 is remounted later on). ”h]”(j½)”}”(hŒmemory_hugetlb_accounting”h]”hŒmemory_hugetlb_accounting”…””}”(hj|hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h KñhjxubjÍ)”}”(hhh]”(hæ)”}”(hXCount HugeTLB memory usage towards the cgroup's overall memory usage for the memory controller (for the purpose of statistics reporting and memory protetion). This is a new behavior that could regress existing setups, so it must be explicitly opted in with this mount option.”h]”hXCount HugeTLB memory usage towards the cgroupâ€™s overall memory usage for the memory controller (for the purpose of statistics reporting and memory protetion). This is a new behavior that could regress existing setups, so it must be explicitly opted in with this mount option.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h K×hjŠubhæ)”}”(hŒA few caveats to keep in mind:”h]”hŒA few caveats to keep in mind:”…””}”(hj›hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h KÝhjŠubhŒbullet_list”“”)”}”(hhh]”(hŒ list_item”“”)”}”(hXThere is no HugeTLB pool management involved in the memory controller. The pre-allocated pool does not belong to anyone. Specifically, when a new HugeTLB folio is allocated to the pool, it is not accounted for from the perspective of the memory controller. It is only charged to a cgroup when it is actually used (for e.g at page fault time). Host memory overcommit management has to consider this when configuring hard limits. In general, HugeTLB pool management should be done via other mechanisms (such as the HugeTLB controller).”h]”hæ)”}”(hXThere is no HugeTLB pool management involved in the memory controller. The pre-allocated pool does not belong to anyone. Specifically, when a new HugeTLB folio is allocated to the pool, it is not accounted for from the perspective of the memory controller. It is only charged to a cgroup when it is actually used (for e.g at page fault time). Host memory overcommit management has to consider this when configuring hard limits. In general, HugeTLB pool management should be done via other mechanisms (such as the HugeTLB controller).”h]”hXThere is no HugeTLB pool management involved in the memory controller. The pre-allocated pool does not belong to anyone. Specifically, when a new HugeTLB folio is allocated to the pool, it is not accounted for from the perspective of the memory controller. It is only charged to a cgroup when it is actually used (for e.g at page fault time). Host memory overcommit management has to consider this when configuring hard limits. In general, HugeTLB pool management should be done via other mechanisms (such as the HugeTLB controller).”…””}”(hj´hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Kßhj°ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj«ubj¯)”}”(hŒËFailure to charge a HugeTLB folio to the memory controller results in SIGBUS. This could happen even if the HugeTLB pool still has pages available (but the cgroup limit is hit and reclaim attempt fails).”h]”hæ)”}”(hŒËFailure to charge a HugeTLB folio to the memory controller results in SIGBUS. This could happen even if the HugeTLB pool still has pages available (but the cgroup limit is hit and reclaim attempt fails).”h]”hŒËFailure to charge a HugeTLB folio to the memory controller results in SIGBUS. This could happen even if the HugeTLB pool still has pages available (but the cgroup limit is hit and reclaim attempt fails).”…””}”(hjÌhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h KèhjÈubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj«ubj¯)”}”(hŒ¸Charging HugeTLB memory towards the memory controller affects memory protection and reclaim dynamics. Any userspace tuning (of low, min limits for e.g) needs to take this into account.”h]”hæ)”}”(hŒ¸Charging HugeTLB memory towards the memory controller affects memory protection and reclaim dynamics. Any userspace tuning (of low, min limits for e.g) needs to take this into account.”h]”hŒ¸Charging HugeTLB memory towards the memory controller affects memory protection and reclaim dynamics. Any userspace tuning (of low, min limits for e.g) needs to take this into account.”…””}”(hjähžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Kìhjàubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj«ubj¯)”}”(hŒ‘HugeTLB pages utilized while this option is not selected will not be tracked by the memory controller (even if cgroup v2 is remounted later on). ”h]”hæ)”}”(hŒHugeTLB pages utilized while this option is not selected will not be tracked by the memory controller (even if cgroup v2 is remounted later on).”h]”hŒHugeTLB pages utilized while this option is not selected will not be tracked by the memory controller (even if cgroup v2 is remounted later on).”…””}”(hjühžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Kïhjøubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj«ubeh}”(h]”h ]”h"]”h$]”h&]”Œbullet”Œ*”uh1j©hŸh¯h KßhjŠubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjxubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Kñhj³ubj·)”}”(hŒñpids_localevents The option restores v1-like behavior of pids.events:max, that is only local (inside cgroup proper) fork failures are counted. Without this option pids.events.max represents any pids.max enforcemnt across cgroup's subtree. ”h]”(j½)”}”(hŒpids_localevents”h]”hŒpids_localevents”…””}”(hj(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Kùhj$ubjÍ)”}”(hhh]”hæ)”}”(hŒÝThe option restores v1-like behavior of pids.events:max, that is only local (inside cgroup proper) fork failures are counted. Without this option pids.events.max represents any pids.max enforcemnt across cgroup's subtree.”h]”hŒßThe option restores v1-like behavior of pids.events:max, that is only local (inside cgroup proper) fork failures are counted. Without this option pids.events.max represents any pids.max enforcemnt across cgroupâ€™s subtree.”…””}”(hj9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Kôhj6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj$ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Kùhj³ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Kµhj6hžhubeh}”(h]”Œmounting”ah ]”h"]”Œmounting”ah$]”h&]”uh1h°hj%hžhhŸh¯h KŽubh±)”}”(hhh]”(h¶)”}”(hŒ Organizing Processes and Threads”h]”hŒ Organizing Processes and Threads”…””}”(hjjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjghžhhŸh¯h Küubh±)”}”(hhh]”(h¶)”}”(hŒ Processes”h]”hŒ Processes”…””}”(hj{hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjxhžhhŸh¯h Kÿubhæ)”}”(hŒInitially, only the root cgroup exists to which all processes belong. A child cgroup can be created by creating a sub-directory::”h]”hŒ€Initially, only the root cgroup exists to which all processes belong. A child cgroup can be created by creating a sub-directory:”…””}”(hj‰hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjxhžhubjV)”}”(hŒ# mkdir $CGROUP_NAME”h]”hŒ# mkdir $CGROUP_NAME”…””}”hj—sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mhjxhžhubhæ)”}”(hX†A given cgroup may have multiple child cgroups forming a tree structure. Each cgroup has a read-writable interface file "cgroup.procs". When read, it lists the PIDs of all processes which belong to the cgroup one-per-line. The PIDs are not ordered and the same PID may show up more than once if the process got moved to another cgroup and then back or the PID got recycled while reading.”h]”hXŠA given cgroup may have multiple child cgroups forming a tree structure. Each cgroup has a read-writable interface file â€œcgroup.procsâ€. When read, it lists the PIDs of all processes which belong to the cgroup one-per-line. The PIDs are not ordered and the same PID may show up more than once if the process got moved to another cgroup and then back or the PID got recycled while reading.”…””}”(hj¥hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjxhžhubhæ)”}”(hXA process can be migrated into a cgroup by writing its PID to the target cgroup's "cgroup.procs" file. Only one process can be migrated on a single write(2) call. If a process is composed of multiple threads, writing the PID of any thread migrates all threads of the process.”h]”hXA process can be migrated into a cgroup by writing its PID to the target cgroupâ€™s â€œcgroup.procsâ€ file. Only one process can be migrated on a single write(2) call. If a process is composed of multiple threads, writing the PID of any thread migrates all threads of the process.”…””}”(hj³hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjxhžhubhæ)”}”(hXiWhen a process forks a child process, the new process is born into the cgroup that the forking process belongs to at the time of the operation. After exit, a process stays associated with the cgroup that it belonged to at the time of exit until it's reaped; however, a zombie process does not appear in "cgroup.procs" and thus can't be moved to another cgroup.”h]”hXqWhen a process forks a child process, the new process is born into the cgroup that the forking process belongs to at the time of the operation. After exit, a process stays associated with the cgroup that it belonged to at the time of exit until itâ€™s reaped; however, a zombie process does not appear in â€œcgroup.procsâ€ and thus canâ€™t be moved to another cgroup.”…””}”(hjÁhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjxhžhubhæ)”}”(hŒðA cgroup which doesn't have any children or live processes can be destroyed by removing the directory. Note that a cgroup which doesn't have any children and is associated only with zombie processes is considered empty and can be removed::”h]”hŒóA cgroup which doesnâ€™t have any children or live processes can be destroyed by removing the directory. Note that a cgroup which doesnâ€™t have any children and is associated only with zombie processes is considered empty and can be removed:”…””}”(hjÏhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjxhžhubjV)”}”(hŒ# rmdir $CGROUP_NAME”h]”hŒ# rmdir $CGROUP_NAME”…””}”hjÝsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mhjxhžhubhæ)”}”(hŒß"/proc/$PID/cgroup" lists a process's cgroup membership. If legacy cgroup is in use in the system, this file may contain multiple lines, one for each hierarchy. The entry for cgroup v2 is always in the format "0::$PATH"::”h]”hŒèâ€œ/proc/$PID/cgroupâ€ lists a processâ€™s cgroup membership. If legacy cgroup is in use in the system, this file may contain multiple lines, one for each hierarchy. The entry for cgroup v2 is always in the format â€œ0::$PATHâ€:”…””}”(hjëhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M!hjxhžhubjV)”}”(hŒ=# cat /proc/842/cgroup ... 0::/test-cgroup/test-cgroup-nested”h]”hŒ=# cat /proc/842/cgroup ... 0::/test-cgroup/test-cgroup-nested”…””}”hjùsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M&hjxhžhubhæ)”}”(hŒ…If the process becomes a zombie and the cgroup it was associated with is removed subsequently, " (deleted)" is appended to the path::”h]”hŒˆIf the process becomes a zombie and the cgroup it was associated with is removed subsequently, â€œ (deleted)â€ is appended to the path:”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M*hjxhžhubjV)”}”(hŒG# cat /proc/842/cgroup ... 0::/test-cgroup/test-cgroup-nested (deleted)”h]”hŒG# cat /proc/842/cgroup ... 0::/test-cgroup/test-cgroup-nested (deleted)”…””}”hjsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M-hjxhžhubeh}”(h]”Œ processes”ah ]”h"]”Œ processes”ah$]”h&]”uh1h°hjghžhhŸh¯h Kÿubh±)”}”(hhh]”(h¶)”}”(hŒThreads”h]”hŒThreads”…””}”(hj.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj+hžhhŸh¯h M3ubhæ)”}”(hXácgroup v2 supports thread granularity for a subset of controllers to support use cases requiring hierarchical resource distribution across the threads of a group of processes. By default, all threads of a process belong to the same cgroup, which also serves as the resource domain to host resource consumptions which are not specific to a process or thread. The thread mode allows threads to be spread across a subtree while still maintaining the common resource domain for them.”h]”hXácgroup v2 supports thread granularity for a subset of controllers to support use cases requiring hierarchical resource distribution across the threads of a group of processes. By default, all threads of a process belong to the same cgroup, which also serves as the resource domain to host resource consumptions which are not specific to a process or thread. The thread mode allows threads to be spread across a subtree while still maintaining the common resource domain for them.”…””}”(hj<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M5hj+hžhubhæ)”}”(hŒzControllers which support thread mode are called threaded controllers. The ones which don't are called domain controllers.”h]”hŒ|Controllers which support thread mode are called threaded controllers. The ones which donâ€™t are called domain controllers.”…””}”(hjJhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M=hj+hžhubhæ)”}”(hX‰Marking a cgroup threaded makes it join the resource domain of its parent as a threaded cgroup. The parent may be another threaded cgroup whose resource domain is further up in the hierarchy. The root of a threaded subtree, that is, the nearest ancestor which is not threaded, is called threaded domain or thread root interchangeably and serves as the resource domain for the entire subtree.”h]”hX‰Marking a cgroup threaded makes it join the resource domain of its parent as a threaded cgroup. The parent may be another threaded cgroup whose resource domain is further up in the hierarchy. The root of a threaded subtree, that is, the nearest ancestor which is not threaded, is called threaded domain or thread root interchangeably and serves as the resource domain for the entire subtree.”…””}”(hjXhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M@hj+hžhubhæ)”}”(hŒíInside a threaded subtree, threads of a process can be put in different cgroups and are not subject to the no internal process constraint - threaded controllers can be enabled on non-leaf cgroups whether they have threads in them or not.”h]”hŒíInside a threaded subtree, threads of a process can be put in different cgroups and are not subject to the no internal process constraint - threaded controllers can be enabled on non-leaf cgroups whether they have threads in them or not.”…””}”(hjfhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MGhj+hžhubhæ)”}”(hX‰As the threaded domain cgroup hosts all the domain resource consumptions of the subtree, it is considered to have internal resource consumptions whether there are processes in it or not and can't have populated child cgroups which aren't threaded. Because the root cgroup is not subject to no internal process constraint, it can serve both as a threaded domain and a parent to domain cgroups.”h]”hXAs the threaded domain cgroup hosts all the domain resource consumptions of the subtree, it is considered to have internal resource consumptions whether there are processes in it or not and canâ€™t have populated child cgroups which arenâ€™t threaded. Because the root cgroup is not subject to no internal process constraint, it can serve both as a threaded domain and a parent to domain cgroups.”…””}”(hjthžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MLhj+hžhubhæ)”}”(hŒßThe current operation mode or type of the cgroup is shown in the "cgroup.type" file which indicates whether the cgroup is a normal domain, a domain which is serving as the domain of a threaded subtree, or a threaded cgroup.”h]”hŒãThe current operation mode or type of the cgroup is shown in the â€œcgroup.typeâ€ file which indicates whether the cgroup is a normal domain, a domain which is serving as the domain of a threaded subtree, or a threaded cgroup.”…””}”(hj‚hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MShj+hžhubhæ)”}”(hŒžOn creation, a cgroup is always a domain cgroup and can be made threaded by writing "threaded" to the "cgroup.type" file. The operation is single direction::”h]”hŒ¥On creation, a cgroup is always a domain cgroup and can be made threaded by writing â€œthreadedâ€ to the â€œcgroup.typeâ€ file. The operation is single direction:”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MXhj+hžhubjV)”}”(hŒ# echo threaded > cgroup.type”h]”hŒ# echo threaded > cgroup.type”…””}”hjžsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M\hj+hžhubhæ)”}”(hŒyOnce threaded, the cgroup can't be made a domain again. To enable the thread mode, the following conditions must be met.”h]”hŒ{Once threaded, the cgroup canâ€™t be made a domain again. To enable the thread mode, the following conditions must be met.”…””}”(hj¬hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M^hj+hžhubjª)”}”(hhh]”(j¯)”}”(hŒAs the cgroup will join the parent's resource domain. The parent must either be a valid (threaded) domain or a threaded cgroup. ”h]”hæ)”}”(hŒ€As the cgroup will join the parent's resource domain. The parent must either be a valid (threaded) domain or a threaded cgroup.”h]”hŒ‚As the cgroup will join the parentâ€™s resource domain. The parent must either be a valid (threaded) domain or a threaded cgroup.”…””}”(hjÁhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mahj½ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjºhžhhŸh¯h Nubj¯)”}”(hŒ¢When the parent is an unthreaded domain, it must not have any domain controllers enabled or populated domain children. The root is exempt from this requirement. ”h]”hæ)”}”(hŒ¡When the parent is an unthreaded domain, it must not have any domain controllers enabled or populated domain children. The root is exempt from this requirement.”h]”hŒ¡When the parent is an unthreaded domain, it must not have any domain controllers enabled or populated domain children. The root is exempt from this requirement.”…””}”(hjÙhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MdhjÕubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjºhžhhŸh¯h Nubeh}”(h]”h ]”h"]”h$]”h&]”jŒ-”uh1j©hŸh¯h Mahj+hžhubhæ)”}”(hŒ]Topology-wise, a cgroup can be in an invalid state. Please consider the following topology::”h]”hŒ\Topology-wise, a cgroup can be in an invalid state. Please consider the following topology:”…””}”(hjôhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhhj+hžhubjV)”}”(hŒ=A (threaded domain) - B (threaded) - C (domain, just created)”h]”hŒ=A (threaded domain) - B (threaded) - C (domain, just created)”…””}”hjsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mkhj+hžhubhæ)”}”(hX!C is created as a domain but isn't connected to a parent which can host child domains. C can't be used until it is turned into a threaded cgroup. "cgroup.type" file will report "domain (invalid)" in these cases. Operations which fail due to invalid topology use EOPNOTSUPP as the errno.”h]”hX-C is created as a domain but isnâ€™t connected to a parent which can host child domains. C canâ€™t be used until it is turned into a threaded cgroup. â€œcgroup.typeâ€ file will report â€œdomain (invalid)â€ in these cases. Operations which fail due to invalid topology use EOPNOTSUPP as the errno.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mmhj+hžhubhæ)”}”(hXA domain cgroup is turned into a threaded domain when one of its child cgroup becomes threaded or threaded controllers are enabled in the "cgroup.subtree_control" file while there are processes in the cgroup. A threaded domain reverts to a normal domain when the conditions clear.”h]”hXA domain cgroup is turned into a threaded domain when one of its child cgroup becomes threaded or threaded controllers are enabled in the â€œcgroup.subtree_controlâ€ file while there are processes in the cgroup. A threaded domain reverts to a normal domain when the conditions clear.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mshj+hžhubhæ)”}”(hX¡When read, "cgroup.threads" contains the list of the thread IDs of all threads in the cgroup. Except that the operations are per-thread instead of per-process, "cgroup.threads" has the same format and behaves the same way as "cgroup.procs". While "cgroup.threads" can be written to in any cgroup, as it can only move threads inside the same threaded domain, its operations are confined inside each threaded subtree.”h]”hX±When read, â€œcgroup.threadsâ€ contains the list of the thread IDs of all threads in the cgroup. Except that the operations are per-thread instead of per-process, â€œcgroup.threadsâ€ has the same format and behaves the same way as â€œcgroup.procsâ€. While â€œcgroup.threadsâ€ can be written to in any cgroup, as it can only move threads inside the same threaded domain, its operations are confined inside each threaded subtree.”…””}”(hj,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Myhj+hžhubhæ)”}”(hXÞThe threaded domain cgroup serves as the resource domain for the whole subtree, and, while the threads can be scattered across the subtree, all the processes are considered to be in the threaded domain cgroup. "cgroup.procs" in a threaded domain cgroup contains the PIDs of all processes in the subtree and is not readable in the subtree proper. However, "cgroup.procs" can be written to from anywhere in the subtree to migrate all threads of the matching process to the cgroup.”h]”hXæThe threaded domain cgroup serves as the resource domain for the whole subtree, and, while the threads can be scattered across the subtree, all the processes are considered to be in the threaded domain cgroup. â€œcgroup.procsâ€ in a threaded domain cgroup contains the PIDs of all processes in the subtree and is not readable in the subtree proper. However, â€œcgroup.procsâ€ can be written to from anywhere in the subtree to migrate all threads of the matching process to the cgroup.”…””}”(hj:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj+hžhubhæ)”}”(hXWOnly threaded controllers can be enabled in a threaded subtree. When a threaded controller is enabled inside a threaded subtree, it only accounts for and controls resource consumptions associated with the threads in the cgroup and its descendants. All consumptions which aren't tied to a specific thread belong to the threaded domain cgroup.”h]”hXYOnly threaded controllers can be enabled in a threaded subtree. When a threaded controller is enabled inside a threaded subtree, it only accounts for and controls resource consumptions associated with the threads in the cgroup and its descendants. All consumptions which arenâ€™t tied to a specific thread belong to the threaded domain cgroup.”…””}”(hjHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‰hj+hžhubhæ)”}”(hXBecause a threaded subtree is exempt from no internal process constraint, a threaded controller must be able to handle competition between threads in a non-leaf cgroup and its child cgroups. Each threaded controller defines how such competitions are handled.”h]”hXBecause a threaded subtree is exempt from no internal process constraint, a threaded controller must be able to handle competition between threads in a non-leaf cgroup and its child cgroups. Each threaded controller defines how such competitions are handled.”…””}”(hjVhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj+hžhubhæ)”}”(hŒ[Currently, the following controllers are threaded and can be enabled in a threaded cgroup::”h]”hŒZCurrently, the following controllers are threaded and can be enabled in a threaded cgroup:”…””}”(hjdhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M”hj+hžhubjV)”}”(hŒ"- cpu - cpuset - perf_event - pids”h]”hŒ"- cpu - cpuset - perf_event - pids”…””}”hjrsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M—hj+hžhubeh}”(h]”Œthreads”ah ]”h"]”Œthreads”ah$]”h&]”uh1h°hjghžhhŸh¯h M3ubeh}”(h]”Œ organizing-processes-and-threads”ah ]”h"]”Œ organizing processes and threads”ah$]”h&]”uh1h°hj%hžhhŸh¯h Küubh±)”}”(hhh]”(h¶)”}”(hŒ[Un]populated Notification”h]”hŒ[Un]populated Notification”…””}”(hj“hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjhžhhŸh¯h Mubhæ)”}”(hXlEach non-root cgroup has a "cgroup.events" file which contains "populated" field indicating whether the cgroup's sub-hierarchy has live processes in it. Its value is 0 if there is no live process in the cgroup and its descendants; otherwise, 1. poll and [id]notify events are triggered when the value changes. This can be used, for example, to start a clean-up operation after all processes of a given sub-hierarchy have exited. The populated state updates and notifications are recursive. Consider the following sub-hierarchy where the numbers in the parentheses represent the numbers of processes in each cgroup::”h]”hXuEach non-root cgroup has a â€œcgroup.eventsâ€ file which contains â€œpopulatedâ€ field indicating whether the cgroupâ€™s sub-hierarchy has live processes in it. Its value is 0 if there is no live process in the cgroup and its descendants; otherwise, 1. poll and [id]notify events are triggered when the value changes. This can be used, for example, to start a clean-up operation after all processes of a given sub-hierarchy have exited. The populated state updates and notifications are recursive. Consider the following sub-hierarchy where the numbers in the parentheses represent the numbers of processes in each cgroup:”…””}”(hj¡hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŸhjhžhubjV)”}”(hŒ%A(4) - B(0) - C(1) \ D(0)”h]”hŒ%A(4) - B(0) - C(1) \ D(0)”…””}”hj¯sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mªhjhžhubhæ)”}”(hŒâA, B and C's "populated" fields would be 1 while D's 0. After the one process in C exits, B and C's "populated" fields would flip to "0" and file modified events will be generated on the "cgroup.events" files of both cgroups.”h]”hŒøA, B and Câ€™s â€œpopulatedâ€ fields would be 1 while Dâ€™s 0. After the one process in C exits, B and Câ€™s â€œpopulatedâ€ fields would flip to â€œ0â€ and file modified events will be generated on the â€œcgroup.eventsâ€ files of both cgroups.”…””}”(hj½hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjhžhubeh}”(h]”Œun-populated-notification”ah ]”h"]”Œ[un]populated notification”ah$]”h&]”uh1h°hj%hžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒControlling Controllers”h]”hŒControlling Controllers”…””}”(hjÖhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjÓhžhhŸh¯h M´ubh±)”}”(hhh]”(h¶)”}”(hŒEnabling and Disabling”h]”hŒEnabling and Disabling”…””}”(hjçhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjähžhhŸh¯h M·ubhæ)”}”(hŒlEach cgroup has a "cgroup.controllers" file which lists all controllers available for the cgroup to enable::”h]”hŒoEach cgroup has a â€œcgroup.controllersâ€ file which lists all controllers available for the cgroup to enable:”…””}”(hjõhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹hjähžhubjV)”}”(hŒ&# cat cgroup.controllers cpu io memory”h]”hŒ&# cat cgroup.controllers cpu io memory”…””}”hjsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M¼hjähžhubhæ)”}”(hŒNo controller is enabled by default. Controllers can be enabled and disabled by writing to the "cgroup.subtree_control" file::”h]”hŒ‚No controller is enabled by default. Controllers can be enabled and disabled by writing to the â€œcgroup.subtree_controlâ€ file:”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¿hjähžhubjV)”}”(hŒ2# echo "+cpu +memory -io" > cgroup.subtree_control”h]”hŒ2# echo "+cpu +memory -io" > cgroup.subtree_control”…””}”hjsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MÂhjähžhubhæ)”}”(hŒõOnly controllers which are listed in "cgroup.controllers" can be enabled. When multiple operations are specified as above, either they all succeed or fail. If multiple operations on the same controller are specified, the last one is effective.”h]”hŒùOnly controllers which are listed in â€œcgroup.controllersâ€ can be enabled. When multiple operations are specified as above, either they all succeed or fail. If multiple operations on the same controller are specified, the last one is effective.”…””}”(hj-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÄhjähžhubhæ)”}”(hŒåEnabling a controller in a cgroup indicates that the distribution of the target resource across its immediate children will be controlled. Consider the following sub-hierarchy. The enabled controllers are listed in parentheses::”h]”hŒäEnabling a controller in a cgroup indicates that the distribution of the target resource across its immediate children will be controlled. Consider the following sub-hierarchy. The enabled controllers are listed in parentheses:”…””}”(hj;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÉhjähžhubjV)”}”(hŒ?A(cpu,memory) - B(memory) - C() \ D()”h]”hŒ?A(cpu,memory) - B(memory) - C() \ D()”…””}”hjIsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MÎhjähžhubhæ)”}”(hXAs A has "cpu" and "memory" enabled, A will control the distribution of CPU cycles and memory to its children, in this case, B. As B has "memory" enabled but not "CPU", C and D will compete freely on CPU cycles but their division of memory available to B will be controlled.”h]”hX#As A has â€œcpuâ€ and â€œmemoryâ€ enabled, A will control the distribution of CPU cycles and memory to its children, in this case, B. As B has â€œmemoryâ€ enabled but not â€œCPUâ€, C and D will compete freely on CPU cycles but their division of memory available to B will be controlled.”…””}”(hjWhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÑhjähžhubhæ)”}”(hX As a controller regulates the distribution of the target resource to the cgroup's children, enabling it creates the controller's interface files in the child cgroups. In the above example, enabling "cpu" on B would create the "cpu." prefixed controller interface files in C and D. Likewise, disabling "memory" from B would remove the "memory." prefixed controller interface files from C and D. This means that the controller interface files - anything which doesn't start with "cgroup." are owned by the parent rather than the cgroup itself.”h]”hX:As a controller regulates the distribution of the target resource to the cgroupâ€™s children, enabling it creates the controllerâ€™s interface files in the child cgroups. In the above example, enabling â€œcpuâ€ on B would create the â€œcpu.â€ prefixed controller interface files in C and D. Likewise, disabling â€œmemoryâ€ from B would remove the â€œmemory.â€ prefixed controller interface files from C and D. This means that the controller interface files - anything which doesnâ€™t start with â€œcgroup.â€ are owned by the parent rather than the cgroup itself.”…””}”(hjehžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÖhjähžhubeh}”(h]”Œenabling-and-disabling”ah ]”h"]”Œenabling and disabling”ah$]”h&]”uh1h°hjÓhžhhŸh¯h M·ubh±)”}”(hhh]”(h¶)”}”(hŒTop-down Constraint”h]”hŒTop-down Constraint”…””}”(hj~hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj{hžhhŸh¯h Máubhæ)”}”(hXÄResources are distributed top-down and a cgroup can further distribute a resource only if the resource has been distributed to it from the parent. This means that all non-root "cgroup.subtree_control" files can only contain controllers which are enabled in the parent's "cgroup.subtree_control" file. A controller can be enabled only if the parent has the controller enabled and a controller can't be disabled if one or more children have it enabled.”h]”hXÐResources are distributed top-down and a cgroup can further distribute a resource only if the resource has been distributed to it from the parent. This means that all non-root â€œcgroup.subtree_controlâ€ files can only contain controllers which are enabled in the parentâ€™s â€œcgroup.subtree_controlâ€ file. A controller can be enabled only if the parent has the controller enabled and a controller canâ€™t be disabled if one or more children have it enabled.”…””}”(hjŒhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mãhj{hžhubeh}”(h]”Œtop-down-constraint”ah ]”h"]”Œtop-down constraint”ah$]”h&]”uh1h°hjÓhžhhŸh¯h Máubh±)”}”(hhh]”(h¶)”}”(hŒNo Internal Process Constraint”h]”hŒNo Internal Process Constraint”…””}”(hj¥hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj¢hžhhŸh¯h Míubhæ)”}”(hXNon-root cgroups can distribute domain resources to their children only when they don't have any processes of their own. In other words, only domain cgroups which don't contain any processes can have domain controllers enabled in their "cgroup.subtree_control" files.”h]”hXNon-root cgroups can distribute domain resources to their children only when they donâ€™t have any processes of their own. In other words, only domain cgroups which donâ€™t contain any processes can have domain controllers enabled in their â€œcgroup.subtree_controlâ€ files.”…””}”(hj³hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mïhj¢hžhubhæ)”}”(hŒöThis guarantees that, when a domain controller is looking at the part of the hierarchy which has it enabled, processes are always only on the leaves. This rules out situations where child cgroups compete against internal processes of the parent.”h]”hŒöThis guarantees that, when a domain controller is looking at the part of the hierarchy which has it enabled, processes are always only on the leaves. This rules out situations where child cgroups compete against internal processes of the parent.”…””}”(hjÁhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Môhj¢hžhubhæ)”}”(hXœThe root cgroup is exempt from this restriction. Root contains processes and anonymous resource consumption which can't be associated with any other cgroups and requires special treatment from most controllers. How resource consumption in the root cgroup is governed is up to each controller (for more information on this topic please refer to the Non-normative information section in the Controllers chapter).”h]”hXžThe root cgroup is exempt from this restriction. Root contains processes and anonymous resource consumption which canâ€™t be associated with any other cgroups and requires special treatment from most controllers. How resource consumption in the root cgroup is governed is up to each controller (for more information on this topic please refer to the Non-normative information section in the Controllers chapter).”…””}”(hjÏhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mùhj¢hžhubhæ)”}”(hXžNote that the restriction doesn't get in the way if there is no enabled controller in the cgroup's "cgroup.subtree_control". This is important as otherwise it wouldn't be possible to create children of a populated cgroup. To control resource distribution of a cgroup, the cgroup must create children and transfer all its processes to the children before enabling controllers in its "cgroup.subtree_control" file.”h]”hX¬Note that the restriction doesnâ€™t get in the way if there is no enabled controller in the cgroupâ€™s â€œcgroup.subtree_controlâ€. This is important as otherwise it wouldnâ€™t be possible to create children of a populated cgroup. To control resource distribution of a cgroup, the cgroup must create children and transfer all its processes to the children before enabling controllers in its â€œcgroup.subtree_controlâ€ file.”…””}”(hjÝhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj¢hžhubeh}”(h]”Œno-internal-process-constraint”ah ]”h"]”Œno internal process constraint”ah$]”h&]”uh1h°hjÓhžhhŸh¯h Míubeh}”(h]”Œcontrolling-controllers”ah ]”h"]”Œcontrolling controllers”ah$]”h&]”uh1h°hj%hžhhŸh¯h M´ubh±)”}”(hhh]”(h¶)”}”(hŒ Delegation”h]”hŒ Delegation”…””}”(hjþhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjûhžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒModel of Delegation”h]”hŒModel of Delegation”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjhžhhŸh¯h Mubhæ)”}”(hX7A cgroup can be delegated in two ways. First, to a less privileged user by granting write access of the directory and its "cgroup.procs", "cgroup.threads" and "cgroup.subtree_control" files to the user. Second, if the "nsdelegate" mount option is set, automatically to a cgroup namespace on namespace creation.”h]”hXGA cgroup can be delegated in two ways. First, to a less privileged user by granting write access of the directory and its â€œcgroup.procsâ€, â€œcgroup.threadsâ€ and â€œcgroup.subtree_controlâ€ files to the user. Second, if the â€œnsdelegateâ€ mount option is set, automatically to a cgroup namespace on namespace creation.”…””•}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjhžhubhæ)”}”(hXdBecause the resource control interface files in a given directory control the distribution of the parent's resources, the delegatee shouldn't be allowed to write to them. For the first method, this is achieved by not granting access to these files. For the second, files outside the namespace should be hidden from the delegatee by the means of at least mount namespacing, and the kernel rejects writes to all files on a namespace root from inside the cgroup namespace, except for those files listed in "/sys/kernel/cgroup/delegate" (including "cgroup.procs", "cgroup.threads", "cgroup.subtree_control", etc.).”h]”hXxBecause the resource control interface files in a given directory control the distribution of the parentâ€™s resources, the delegatee shouldnâ€™t be allowed to write to them. For the first method, this is achieved by not granting access to these files. For the second, files outside the namespace should be hidden from the delegatee by the means of at least mount namespacing, and the kernel rejects writes to all files on a namespace root from inside the cgroup namespace, except for those files listed in â€œ/sys/kernel/cgroup/delegateâ€ (including â€œcgroup.procsâ€, â€œcgroup.threadsâ€, â€œcgroup.subtree_controlâ€, etc.).”…””}”(hj+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjhžhubhæ)”}”(hX½The end results are equivalent for both delegation types. Once delegated, the user can build sub-hierarchy under the directory, organize processes inside it as it sees fit and further distribute the resources it received from the parent. The limits and other settings of all resource controllers are hierarchical and regardless of what happens in the delegated sub-hierarchy, nothing can escape the resource restrictions imposed by the parent.”h]”hX½The end results are equivalent for both delegation types. Once delegated, the user can build sub-hierarchy under the directory, organize processes inside it as it sees fit and further distribute the resources it received from the parent. The limits and other settings of all resource controllers are hierarchical and regardless of what happens in the delegated sub-hierarchy, nothing can escape the resource restrictions imposed by the parent.”…””}”(hj9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjhžhubhæ)”}”(hŒ³Currently, cgroup doesn't impose any restrictions on the number of cgroups in or nesting depth of a delegated sub-hierarchy; however, this may be limited explicitly in the future.”h]”hŒµCurrently, cgroup doesnâ€™t impose any restrictions on the number of cgroups in or nesting depth of a delegated sub-hierarchy; however, this may be limited explicitly in the future.”…””}”(hjGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M(hjhžhubeh}”(h]”Œmodel-of-delegation”ah ]”h"]”Œmodel of delegation”ah$]”h&]”uh1h°hjûhžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒDelegation Containment”h]”hŒDelegation Containment”…””}”(hj`hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj]hžhhŸh¯h M.ubhæ)”}”(hŒ„A delegated sub-hierarchy is contained in the sense that processes can't be moved into or out of the sub-hierarchy by the delegatee.”h]”hŒ†A delegated sub-hierarchy is contained in the sense that processes canâ€™t be moved into or out of the sub-hierarchy by the delegatee.”…””}”(hjnhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M0hj]hžhubhæ)”}”(hŒÝFor delegations to a less privileged user, this is achieved by requiring the following conditions for a process with a non-root euid to migrate a target process into a cgroup by writing its PID to the "cgroup.procs" file.”h]”hŒáFor delegations to a less privileged user, this is achieved by requiring the following conditions for a process with a non-root euid to migrate a target process into a cgroup by writing its PID to the â€œcgroup.procsâ€ file.”…””}”(hj|hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M3hj]hžhubjª)”}”(hhh]”(j¯)”}”(hŒ>The writer must have write access to the "cgroup.procs" file. ”h]”hæ)”}”(hŒ=The writer must have write access to the "cgroup.procs" file.”h]”hŒAThe writer must have write access to the â€œcgroup.procsâ€ file.”…””}”(hj‘hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M8hjubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjŠhžhhŸh¯h Nubj¯)”}”(hŒ{The writer must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups. ”h]”hæ)”}”(hŒzThe writer must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups.”h]”hŒ~The writer must have write access to the â€œcgroup.procsâ€ file of the common ancestor of the source and destination cgroups.”…””}”(hj©hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M:hj¥ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjŠhžhhŸh¯h Nubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h M8hj]hžhubhæ)”}”(hŒºThe above two constraints ensure that while a delegatee may migrate processes around freely in the delegated sub-hierarchy it can't pull in from or push out to outside the sub-hierarchy.”h]”hŒ¼The above two constraints ensure that while a delegatee may migrate processes around freely in the delegated sub-hierarchy it canâ€™t pull in from or push out to outside the sub-hierarchy.”…””}”(hjÃhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M=hj]hžhubhæ)”}”(hŒ¸For an example, let's assume cgroups C0 and C1 have been delegated to user U0 who created C00, C01 under C0 and C10 under C1 as follows and all processes under C0 and C1 belong to U0::”h]”hŒ¹For an example, letâ€™s assume cgroups C0 and C1 have been delegated to user U0 who created C00, C01 under C0 and C10 under C1 as follows and all processes under C0 and C1 belong to U0:”…””}”(hjÑhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MAhj]hžhubjV)”}”(hŒX~~~~~~~~~~~~~ - C0 - C00 ~ cgroup ~ \ C01 ~ hierarchy ~ ~~~~~~~~~~~~~ - C1 - C10”h]”hŒX~~~~~~~~~~~~~ - C0 - C00 ~ cgroup ~ \ C01 ~ hierarchy ~ ~~~~~~~~~~~~~ - C1 - C10”…””}”hjßsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MEhj]hžhubhæ)”}”(hXpLet's also say U0 wants to write the PID of a process which is currently in C10 into "C00/cgroup.procs". U0 has write access to the file; however, the common ancestor of the source cgroup C10 and the destination cgroup C00 is above the points of delegation and U0 would not have write access to its "cgroup.procs" files and thus the write will be denied with -EACCES.”h]”hXzLetâ€™s also say U0 wants to write the PID of a process which is currently in C10 into â€œC00/cgroup.procsâ€. U0 has write access to the file; however, the common ancestor of the source cgroup C10 and the destination cgroup C00 is above the points of delegation and U0 would not have write access to its â€œcgroup.procsâ€ files and thus the write will be denied with -EACCES.”…””}”(hjíhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MJhj]hžhubhæ)”}”(hXFor delegations to namespaces, containment is achieved by requiring that both the source and destination cgroups are reachable from the namespace of the process which is attempting the migration. If either is not reachable, the migration is rejected with -ENOENT.”h]”hXFor delegations to namespaces, containment is achieved by requiring that both the source and destination cgroups are reachable from the namespace of the process which is attempting the migration. If either is not reachable, the migration is rejected with -ENOENT.”…””}”(hjûhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MQhj]hžhubeh}”(h]”Œdelegation-containment”ah ]”h"]”Œdelegation containment”ah$]”h&]”uh1h°hjûhžhhŸh¯h M.ubeh}”(h]”Œ delegation”ah ]”h"]”Œ delegation”ah$]”h&]”uh1h°hj%hžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒ Guidelines”h]”hŒ Guidelines”…””}”(hj hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj hžhhŸh¯h MXubh±)”}”(hhh]”(h¶)”}”(hŒOrganize Once and Control”h]”hŒOrganize Once and Control”…””}”(hj- hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj* hžhhŸh¯h M[ubhæ)”}”(hX-Migrating a process across cgroups is a relatively expensive operation and stateful resources such as memory are not moved together with the process. This is an explicit design decision as there often exist inherent trade-offs between migration and various hot paths in terms of synchronization cost.”h]”hX-Migrating a process across cgroups is a relatively expensive operation and stateful resources such as memory are not moved together with the process. This is an explicit design decision as there often exist inherent trade-offs between migration and various hot paths in terms of synchronization cost.”…””}”(hj; hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M]hj* hžhubhæ)”}”(hXlAs such, migrating processes across cgroups frequently as a means to apply different resource restrictions is discouraged. A workload should be assigned to a cgroup according to the system's logical and resource structure once on start-up. Dynamic adjustments to resource distribution can be made by changing controller configuration through the interface files.”h]”hXnAs such, migrating processes across cgroups frequently as a means to apply different resource restrictions is discouraged. A workload should be assigned to a cgroup according to the systemâ€™s logical and resource structure once on start-up. Dynamic adjustments to resource distribution can be made by changing controller configuration through the interface files.”…””}”(hjI hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mchj* hžhubeh}”(h]”Œorganize-once-and-control”ah ]”h"]”Œorganize once and control”ah$]”h&]”uh1h°hj hžhhŸh¯h M[ubh±)”}”(hhh]”(h¶)”}”(hŒAvoid Name Collisions”h]”hŒAvoid Name Collisions”…””}”(hjb hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj_ hžhhŸh¯h Mlubhæ)”}”(hŒ¡Interface files for a cgroup and its children cgroups occupy the same directory and it is possible to create children cgroups which collide with interface files.”h]”hŒ¡Interface files for a cgroup and its children cgroups occupy the same directory and it is possible to create children cgroups which collide with interface files.”…””}”(hjp hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mnhj_ hžhubhæ)”}”(hXÌAll cgroup core interface files are prefixed with "cgroup." and each controller's interface files are prefixed with the controller name and a dot. A controller's name is composed of lower case alphabets and '_'s but never begins with an '_' so it can be used as the prefix character for collision avoidance. Also, interface file names won't start or end with terms which are often used in categorizing workloads such as job, service, slice, unit or workload.”h]”hXÞAll cgroup core interface files are prefixed with â€œcgroup.â€ and each controllerâ€™s interface files are prefixed with the controller name and a dot. A controllerâ€™s name is composed of lower case alphabets and â€˜_â€™s but never begins with an â€˜_â€™ so it can be used as the prefix character for collision avoidance. Also, interface file names wonâ€™t start or end with terms which are often used in categorizing workloads such as job, service, slice, unit or workload.”…””}”(hj~ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mrhj_ hžhubhæ)”}”(hŒgcgroup doesn't do anything to prevent name collisions and it's the user's responsibility to avoid them.”h]”hŒmcgroup doesnâ€™t do anything to prevent name collisions and itâ€™s the userâ€™s responsibility to avoid them.”…””}”(hjŒ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mzhj_ hžhubeh}”(h]”Œavoid-name-collisions”ah ]”h"]”Œavoid name collisions”ah$]”h&]”uh1h°hj hžhhŸh¯h Mlubeh}”(h]”Œ guidelines”ah ]”h"]”Œ guidelines”ah$]”h&]”uh1h°hj%hžhhŸh¯h MXubeh}”(h]”Œbasic-operations”ah ]”h"]”Œbasic operations”ah$]”h&]”uh1h°hh²hžhhŸh¯h K‹ubh±)”}”(hhh]”(h¶)”}”(hŒResource Distribution Models”h]”hŒResource Distribution Models”…””}”(hjµ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj² hžhhŸh¯h Mubhæ)”}”(hŒËcgroup controllers implement several resource distribution schemes depending on the resource type and expected use cases. This section describes major schemes in use along with their expected behaviors.”h]”hŒËcgroup controllers implement several resource distribution schemes depending on the resource type and expected use cases. This section describes major schemes in use along with their expected behaviors.”…””}”(hjÃ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj² hžhubh±)”}”(hhh]”(h¶)”}”(hŒWeights”h]”hŒWeights”…””}”(hjÔ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjÑ hžhhŸh¯h M‡ubhæ)”}”(hXnA parent's resource is distributed by adding up the weights of all active children and giving each the fraction matching the ratio of its weight against the sum. As only children which can make use of the resource at the moment participate in the distribution, this is work-conserving. Due to the dynamic nature, this model is usually used for stateless resources.”h]”hXpA parentâ€™s resource is distributed by adding up the weights of all active children and giving each the fraction matching the ratio of its weight against the sum. As only children which can make use of the resource at the moment participate in the distribution, this is work-conserving. Due to the dynamic nature, this model is usually used for stateless resources.”…””}”(hjâ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‰hjÑ hžhubhæ)”}”(hŒÁAll weights are in the range [1, 10000] with the default at 100. This allows symmetric multiplicative biases in both directions at fine enough granularity while staying in the intuitive range.”h]”hŒÁAll weights are in the range [1, 10000] with the default at 100. This allows symmetric multiplicative biases in both directions at fine enough granularity while staying in the intuitive range.”…””}”(hjð hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÑ hžhubhæ)”}”(hŒ™As long as the weight is in range, all configuration combinations are valid and there is no reason to reject configuration changes or process migrations.”h]”hŒ™As long as the weight is in range, all configuration combinations are valid and there is no reason to reject configuration changes or process migrations.”…””}”(hjþ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M”hjÑ hžhubhæ)”}”(hŒe"cpu.weight" proportionally distributes CPU cycles to active children and is an example of this type.”h]”hŒiâ€œcpu.weightâ€ proportionally distributes CPU cycles to active children and is an example of this type.”…””}”(hj hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M˜hjÑ hžhubh¢)”}”(hŒ .. _cgroupv2-limits-distributor:”h]”h}”(h]”h ]”h"]”h$]”h&]”hŒcgroupv2-limits-distributor”uh1h¡h MœhjÑ hžhhŸh¯ubeh}”(h]”Œweights”ah ]”h"]”Œweights”ah$]”h&]”uh1h°hj² hžhhŸh¯h M‡ubh±)”}”(hhh]”(h¶)”}”(hŒLimits”h]”hŒLimits”…””}”(hj0 hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj- hžhhŸh¯h MŸubhæ)”}”(hŒÁA child can only consume up to the configured amount of the resource. Limits can be over-committed - the sum of the limits of children can exceed the amount of resource available to the parent.”h]”hŒÁA child can only consume up to the configured amount of the resource. Limits can be over-committed - the sum of the limits of children can exceed the amount of resource available to the parent.”…””}”(hj> hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¡hj- hžhubhæ)”}”(hŒFLimits are in the range [0, max] and defaults to "max", which is noop.”h]”hŒJLimits are in the range [0, max] and defaults to â€œmaxâ€, which is noop.”…””}”(hjL hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¥hj- hžhubhæ)”}”(hŒ—As limits can be over-committed, all configuration combinations are valid and there is no reason to reject configuration changes or process migrations.”h]”hŒ—As limits can be over-committed, all configuration combinations are valid and there is no reason to reject configuration changes or process migrations.”…””}”(hjZ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M§hj- hžhubhæ)”}”(hŒu"io.max" limits the maximum BPS and/or IOPS that a cgroup can consume on an IO device and is an example of this type.”h]”hŒyâ€œio.maxâ€ limits the maximum BPS and/or IOPS that a cgroup can consume on an IO device and is an example of this type.”…””}”(hjh hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M«hj- hžhubh¢)”}”(hŒ%.. _cgroupv2-protections-distributor:”h]”h}”(h]”h ]”h"]”h$]”h&]”hŒ cgroupv2-protections-distributor”uh1h¡h M®hj- hžhhŸh¯ubeh}”(h]”(Œlimits”j$ eh ]”h"]”(Œlimits”Œcgroupv2-limits-distributor”eh$]”h&]”uh1h°hj² hžhhŸh¯h MŸŒexpect_referenced_by_name”}”j‡ j sŒexpect_referenced_by_id”}”j$ j subh±)”}”(hhh]”(h¶)”}”(hŒProtections”h]”hŒProtections”…””}”(hj‘ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjŽ hžhhŸh¯h M±ubhæ)”}”(hXSA cgroup is protected up to the configured amount of the resource as long as the usages of all its ancestors are under their protected levels. Protections can be hard guarantees or best effort soft boundaries. Protections can also be over-committed in which case only up to the amount available to the parent is protected among children.”h]”hXSA cgroup is protected up to the configured amount of the resource as long as the usages of all its ancestors are under their protected levels. Protections can be hard guarantees or best effort soft boundaries. Protections can also be over-committed in which case only up to the amount available to the parent is protected among children.”…””}”(hjŸ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M³hjŽ hžhubhæ)”}”(hŒGProtections are in the range [0, max] and defaults to 0, which is noop.”h]”hŒGProtections are in the range [0, max] and defaults to 0, which is noop.”…””}”(hj hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MºhjŽ hžhubhæ)”}”(hŒœAs protections can be over-committed, all configuration combinations are valid and there is no reason to reject configuration changes or process migrations.”h]”hŒœAs protections can be over-committed, all configuration combinations are valid and there is no reason to reject configuration changes or process migrations.”…””}”(hj» hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M½hjŽ hžhubhæ)”}”(hŒU"memory.low" implements best-effort memory protection and is an example of this type.”h]”hŒYâ€œmemory.lowâ€ implements best-effort memory protection and is an example of this type.”…””}”(hjÉ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÁhjŽ hžhubeh}”(h]”(Œprotections”j€ eh ]”h"]”(Œprotections”Œ cgroupv2-protections-distributor”eh$]”h&]”uh1h°hj² hžhhŸh¯h M±jŠ }”jÝ jv sjŒ }”j€ jv subh±)”}”(hhh]”(h¶)”}”(hŒAllocations”h]”hŒAllocations”…””}”(hjå hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjâ hžhhŸh¯h MÆubhæ)”}”(hŒÕA cgroup is exclusively allocated a certain amount of a finite resource. Allocations can't be over-committed - the sum of the allocations of children can not exceed the amount of resource available to the parent.”h]”hŒ×A cgroup is exclusively allocated a certain amount of a finite resource. Allocations canâ€™t be over-committed - the sum of the allocations of children can not exceed the amount of resource available to the parent.”…””}”(hjó hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÈhjâ hžhubhæ)”}”(hŒNAllocations are in the range [0, max] and defaults to 0, which is no resource.”h]”hŒNAllocations are in the range [0, max] and defaults to 0, which is no resource.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÍhjâ hžhubhæ)”}”(hŒÏAs allocations can't be over-committed, some configuration combinations are invalid and should be rejected. Also, if the resource is mandatory for execution of processes, process migrations may be rejected.”h]”hŒÑAs allocations canâ€™t be over-committed, some configuration combinations are invalid and should be rejected. Also, if the resource is mandatory for execution of processes, process migrations may be rejected.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÐhjâ hžhubhæ)”}”(hŒK"cpu.rt.max" hard-allocates realtime slices and is an example of this type.”h]”hŒOâ€œcpu.rt.maxâ€ hard-allocates realtime slices and is an example of this type.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÕhjâ hžhubeh}”(h]”Œallocations”ah ]”h"]”Œallocations”ah$]”h&]”uh1h°hj² hžhhŸh¯h MÆubeh}”(h]”Œresource-distribution-models”ah ]”h"]”Œresource distribution models”ah$]”h&]”uh1h°hh²hžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒInterface Files”h]”hŒInterface Files”…””}”(hj>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj;hžhhŸh¯h MÚubh±)”}”(hhh]”(h¶)”}”(hŒFormat”h]”hŒFormat”…””}”(hjOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjLhžhhŸh¯h MÝubhæ)”}”(hŒQAll interface files should be in one of the following formats whenever possible::”h]”hŒPAll interface files should be in one of the following formats whenever possible:”…””}”(hj]hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MßhjLhžhubjV)”}”(hXNew-line separated values (when only one value can be written at once) VAL0\n VAL1\n ... Space separated values (when read-only or multiple values can be written at once) VAL0 VAL1 ...\n Flat keyed KEY0 VAL0\n KEY1 VAL1\n ... Nested keyed KEY0 SUB_KEY0=VAL00 SUB_KEY1=VAL01... KEY1 SUB_KEY0=VAL10 SUB_KEY1=VAL11... ...”h]”hXNew-line separated values (when only one value can be written at once) VAL0\n VAL1\n ... Space separated values (when read-only or multiple values can be written at once) VAL0 VAL1 ...\n Flat keyed KEY0 VAL0\n KEY1 VAL1\n ... Nested keyed KEY0 SUB_KEY0=VAL00 SUB_KEY1=VAL01... KEY1 SUB_KEY0=VAL10 SUB_KEY1=VAL11... ...”…””}”hjksbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MâhjLhžhubhæ)”}”(hŒ½For a writable file, the format for writing should generally match reading; however, controllers may allow omitting later fields or implement restricted shortcuts for most common use cases.”h]”hŒ½For a writable file, the format for writing should generally match reading; however, controllers may allow omitting later fields or implement restricted shortcuts for most common use cases.”…””}”(hjyhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MúhjLhžhubhæ)”}”(hŒÑFor both flat and nested keyed files, only the values for a single key can be written at a time. For nested keyed files, the sub key pairs may be specified in any order and not all pairs have to be specified.”h]”hŒÑFor both flat and nested keyed files, only the values for a single key can be written at a time. For nested keyed files, the sub key pairs may be specified in any order and not all pairs have to be specified.”…””}”(hj‡hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MþhjLhžhubeh}”(h]”Œformat”ah ]”h"]”Œformat”ah$]”h&]”uh1h°hj;hžhhŸh¯h MÝubh±)”}”(hhh]”(h¶)”}”(hŒConventions”h]”hŒConventions”…””}”(hj hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjhžhhŸh¯h Mubjª)”}”(hhh]”(j¯)”}”(hŒDSettings for a single feature should be contained in a single file. ”h]”hæ)”}”(hŒCSettings for a single feature should be contained in a single file.”h]”hŒCSettings for a single feature should be contained in a single file.”…””}”(hjµhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj±ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj®hžhhŸh¯h Nubj¯)”}”(hŒqThe root cgroup should be exempt from resource control and thus shouldn't have resource control interface files. ”h]”hæ)”}”(hŒpThe root cgroup should be exempt from resource control and thus shouldn't have resource control interface files.”h]”hŒrThe root cgroup should be exempt from resource control and thus shouldnâ€™t have resource control interface files.”…””}”(hjÍhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÉubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj®hžhhŸh¯h Nubj¯)”}”(hŒsThe default time unit is microseconds. If a different unit is ever used, an explicit unit suffix must be present. ”h]”hæ)”}”(hŒrThe default time unit is microseconds. If a different unit is ever used, an explicit unit suffix must be present.”h]”hŒrThe default time unit is microseconds. If a different unit is ever used, an explicit unit suffix must be present.”…””}”(hjåhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjáubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj®hžhhŸh¯h Nubj¯)”}”(hŒkA parts-per quantity should use a percentage decimal with at least two digit fractional part - e.g. 13.40. ”h]”hæ)”}”(hŒjA parts-per quantity should use a percentage decimal with at least two digit fractional part - e.g. 13.40.”h]”hŒjA parts-per quantity should use a percentage decimal with at least two digit fractional part - e.g. 13.40.”…””}”(hjýhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjùubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj®hžhhŸh¯h Nubj¯)”}”(hX!If a controller implements weight based resource distribution, its interface file should be named "weight" and have the range [1, 10000] with 100 as the default. The values are chosen to allow enough and symmetric bias in both directions while keeping it intuitive (the default is 100%). ”h]”hæ)”}”(hX If a controller implements weight based resource distribution, its interface file should be named "weight" and have the range [1, 10000] with 100 as the default. The values are chosen to allow enough and symmetric bias in both directions while keeping it intuitive (the default is 100%).”h]”hX$If a controller implements weight based resource distribution, its interface file should be named â€œweightâ€ and have the range [1, 10000] with 100 as the default. The values are chosen to allow enough and symmetric bias in both directions while keeping it intuitive (the default is 100%).”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj®hžhhŸh¯h Nubj¯)”}”(hXšIf a controller implements an absolute resource guarantee and/or limit, the interface files should be named "min" and "max" respectively. If a controller implements best effort resource guarantee and/or limit, the interface files should be named "low" and "high" respectively. In the above four control files, the special token "max" should be used to represent upward infinity for both reading and writing. ”h]”(hæ)”}”(hXIf a controller implements an absolute resource guarantee and/or limit, the interface files should be named "min" and "max" respectively. If a controller implements best effort resource guarantee and/or limit, the interface files should be named "low" and "high" respectively.”h]”hX%If a controller implements an absolute resource guarantee and/or limit, the interface files should be named â€œminâ€ and â€œmaxâ€ respectively. If a controller implements best effort resource guarantee and/or limit, the interface files should be named â€œlowâ€ and â€œhighâ€ respectively.”…””}”(hj-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj)ubhæ)”}”(hŒ‚In the above four control files, the special token "max" should be used to represent upward infinity for both reading and writing.”h]”hŒ†In the above four control files, the special token â€œmaxâ€ should be used to represent upward infinity for both reading and writing.”…””}”(hj;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj)ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j®hj®hžhhŸh¯h Nubj¯)”}”(hXØIf a setting has a configurable default value and keyed specific overrides, the default entry should be keyed with "default" and appear as the first entry in the file. The default value can be updated by writing either "default $VAL" or "$VAL". When writing to update a specific override, "default" can be used as the value to indicate removal of the override. Override entries with "default" as the value must not appear when read. For example, a setting which is keyed by major:minor device numbers with integer values may look like the following:: # cat cgroup-example-interface-file default 150 8:0 300 The default value can be updated by:: # echo 125 > cgroup-example-interface-file or:: # echo "default 125" > cgroup-example-interface-file An override can be set by:: # echo "8:16 170" > cgroup-example-interface-file and cleared by:: # echo "8:0 default" > cgroup-example-interface-file # cat cgroup-example-interface-file default 125 8:16 170 ”h]”(hæ)”}”(hŒ§If a setting has a configurable default value and keyed specific overrides, the default entry should be keyed with "default" and appear as the first entry in the file.”h]”hŒ«If a setting has a configurable default value and keyed specific overrides, the default entry should be keyed with â€œdefaultâ€ and appear as the first entry in the file.”…””}”(hjShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjOubhæ)”}”(hŒLThe default value can be updated by writing either "default $VAL" or "$VAL".”h]”hŒTThe default value can be updated by writing either â€œdefault $VALâ€ or â€œ$VALâ€.”…””}”(hjahžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M$hjOubhæ)”}”(hŒ¼When writing to update a specific override, "default" can be used as the value to indicate removal of the override. Override entries with "default" as the value must not appear when read.”h]”hŒÄWhen writing to update a specific override, â€œdefaultâ€ can be used as the value to indicate removal of the override. Override entries with â€œdefaultâ€ as the value must not appear when read.”…””}”(hjohžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M'hjOubhæ)”}”(hŒuFor example, a setting which is keyed by major:minor device numbers with integer values may look like the following::”h]”hŒtFor example, a setting which is keyed by major:minor device numbers with integer values may look like the following:”…””}”(hj}hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M+hjOubjV)”}”(hŒ7# cat cgroup-example-interface-file default 150 8:0 300”h]”hŒ7# cat cgroup-example-interface-file default 150 8:0 300”…””}”hj‹sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M.hjOubhæ)”}”(hŒ%The default value can be updated by::”h]”hŒ$The default value can be updated by:”…””}”(hj™hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M2hjOubjV)”}”(hŒ*# echo 125 > cgroup-example-interface-file”h]”hŒ*# echo 125 > cgroup-example-interface-file”…””}”hj§sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M4hjOubhæ)”}”(hŒor::”h]”hŒor:”…””}”(hjµhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M6hjOubjV)”}”(hŒ4# echo "default 125" > cgroup-example-interface-file”h]”hŒ4# echo "default 125" > cgroup-example-interface-file”…””}”hjÃsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M8hjOubhæ)”}”(hŒAn override can be set by::”h]”hŒAn override can be set by:”…””}”(hjÑhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M:hjOubjV)”}”(hŒ1# echo "8:16 170" > cgroup-example-interface-file”h]”hŒ1# echo "8:16 170" > cgroup-example-interface-file”…””}”hjßsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MhjOubjV)”}”(hŒm# echo "8:0 default" > cgroup-example-interface-file # cat cgroup-example-interface-file default 125 8:16 170”h]”hŒm# echo "8:0 default" > cgroup-example-interface-file # cat cgroup-example-interface-file default 125 8:16 170”…””}”hjûsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M@hjOubeh}”(h]”h ]”h"]”h$]”h&]”uh1j®hj®hžhhŸh¯h Nubj¯)”}”(hŒÙFor events which are not very high frequency, an interface file "events" should be created which lists event key value pairs. Whenever a notifiable event happens, file modified event should be generated on the file. ”h]”hæ)”}”(hŒ×For events which are not very high frequency, an interface file "events" should be created which lists event key value pairs. Whenever a notifiable event happens, file modified event should be generated on the file.”h]”hŒÛFor events which are not very high frequency, an interface file â€œeventsâ€ should be created which lists event key value pairs. Whenever a notifiable event happens, file modified event should be generated on the file.”…””}”(hj hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MEhj ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hj®hžhhŸh¯h Nubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h Mhjhžhubeh}”(h]”Œconventions”ah ]”h"]”Œconventions”ah$]”h&]”uh1h°hj;hžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒCore Interface Files”h]”hŒCore Interface Files”…””}”(hj8 hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj5 hžhhŸh¯h MLubhæ)”}”(hŒ1All cgroup core files are prefixed with "cgroup."”h]”hŒ5All cgroup core files are prefixed with â€œcgroup.â€”…””}”(hjF hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MNhj5 hžhubj¬)”}”(hX±!cgroup.type A read-write single value file which exists on non-root cgroups. When read, it indicates the current type of the cgroup, which can be one of the following values. - "domain" : A normal valid domain cgroup. - "domain threaded" : A threaded domain cgroup which is serving as the root of a threaded subtree. - "domain invalid" : A cgroup which is in an invalid state. It can't be populated or have controllers enabled. It may be allowed to become a threaded cgroup. - "threaded" : A threaded cgroup which is a member of a threaded subtree. A cgroup can be turned into a threaded cgroup by writing "threaded" to this file. cgroup.procs A read-write new-line separated values file which exists on all cgroups. When read, it lists the PIDs of all processes which belong to the cgroup one-per-line. The PIDs are not ordered and the same PID may show up more than once if the process got moved to another cgroup and then back or the PID got recycled while reading. A PID can be written to migrate the process associated with the PID to the cgroup. The writer should match all of the following conditions. - It must have write access to the "cgroup.procs" file. - It must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups. When delegating a sub-hierarchy, write access to this file should be granted along with the containing directory. In a threaded cgroup, reading this file fails with EOPNOTSUPP as all the processes belong to the thread root. Writing is supported and moves every thread of the process to the cgroup. cgroup.threads A read-write new-line separated values file which exists on all cgroups. When read, it lists the TIDs of all threads which belong to the cgroup one-per-line. The TIDs are not ordered and the same TID may show up more than once if the thread got moved to another cgroup and then back or the TID got recycled while reading. A TID can be written to migrate the thread associated with the TID to the cgroup. The writer should match all of the following conditions. - It must have write access to the "cgroup.threads" file. - The cgroup that the thread is currently in must be in the same resource domain as the destination cgroup. - It must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups. When delegating a sub-hierarchy, write access to this file should be granted along with the containing directory. cgroup.controllers A read-only space separated values file which exists on all cgroups. It shows space separated list of all controllers available to the cgroup. The controllers are not ordered. cgroup.subtree_control A read-write space separated values file which exists on all cgroups. Starts out empty. When read, it shows space separated list of the controllers which are enabled to control resource distribution from the cgroup to its children. Space separated list of controllers prefixed with '+' or '-' can be written to enable or disable controllers. A controller name prefixed with '+' enables the controller and '-' disables. If a controller appears more than once on the list, the last one is effective. When multiple enable and disable operations are specified, either all succeed or all fail. cgroup.events A read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. populated 1 if the cgroup or its descendants contains any live processes; otherwise, 0. frozen 1 if the cgroup is frozen; otherwise, 0. cgroup.max.descendants A read-write single value files. The default is "max". Maximum allowed number of descent cgroups. If the actual number of descendants is equal or larger, an attempt to create a new cgroup in the hierarchy will fail. cgroup.max.depth A read-write single value files. The default is "max". Maximum allowed descent depth below the current cgroup. If the actual descent depth is equal or larger, an attempt to create a new child cgroup will fail. cgroup.stat A read-only flat-keyed file with the following entries: nr_descendants Total number of visible descendant cgroups. nr_dying_descendants Total number of dying descendant cgroups. A cgroup becomes dying after being deleted by a user. The cgroup will remain in dying state for some time undefined time (which can depend on system load) before being completely destroyed. A process can't enter a dying cgroup under any circumstances, a dying cgroup can't revive. A dying cgroup can consume system resources not exceeding limits, which were active at the moment of cgroup deletion. nr_subsys_ Total number of live cgroup subsystems (e.g memory cgroup) at and beneath the current cgroup. nr_dying_subsys_ Total number of dying cgroup subsystems (e.g. memory cgroup) at and beneath the current cgroup. cgroup.freeze A read-write single value file which exists on non-root cgroups. Allowed values are "0" and "1". The default is "0". Writing "1" to the file causes freezing of the cgroup and all descendant cgroups. This means that all belonging processes will be stopped and will not run until the cgroup will be explicitly unfrozen. Freezing of the cgroup may take some time; when this action is completed, the "frozen" value in the cgroup.events control file will be updated to "1" and the corresponding notification will be issued. A cgroup can be frozen either by its own settings, or by settings of any ancestor cgroups. If any of ancestor cgroups is frozen, the cgroup will remain frozen. Processes in the frozen cgroup can be killed by a fatal signal. They also can enter and leave a frozen cgroup: either by an explicit move by a user, or if freezing of the cgroup races with fork(). If a process is moved to a frozen cgroup, it stops. If a process is moved out of a frozen cgroup, it becomes running. Frozen status of a cgroup doesn't affect any cgroup tree operations: it's possible to delete a frozen (and empty) cgroup, as well as create new sub-cgroups. cgroup.kill A write-only single value file which exists in non-root cgroups. The only allowed value is "1". Writing "1" to the file causes the cgroup and all descendant cgroups to be killed. This means that all processes located in the affected cgroup tree will be killed via SIGKILL. Killing a cgroup tree will deal with concurrent forks appropriately and is protected against migrations. In a threaded cgroup, writing this file fails with EOPNOTSUPP as killing cgroups is a process directed operation, i.e. it affects the whole thread-group. cgroup.pressure A read-write single value file that allowed values are "0" and "1". The default is "1". Writing "0" to the file will disable the cgroup PSI accounting. Writing "1" to the file will re-enable the cgroup PSI accounting. This control attribute is not hierarchical, so disable or enable PSI accounting in a cgroup does not affect PSI accounting in descendants and doesn't need pass enablement via ancestors from root. The reason this control attribute exists is that PSI accounts stalls for each cgroup separately and aggregates it at each level of the hierarchy. This may cause non-negligible overhead for some workloads when under deep level of the hierarchy, in which case this control attribute can be used to disable PSI accounting in the non-leaf cgroups. irq.pressure A read-write nested-keyed file. Shows pressure stall information for IRQ/SOFTIRQ. See :ref:`Documentation/accounting/psi.rst ` for details. ”h]”j²)”}”(hhh]”(j·)”}”(hX†cgroup.type A read-write single value file which exists on non-root cgroups. When read, it indicates the current type of the cgroup, which can be one of the following values. - "domain" : A normal valid domain cgroup. - "domain threaded" : A threaded domain cgroup which is serving as the root of a threaded subtree. - "domain invalid" : A cgroup which is in an invalid state. It can't be populated or have controllers enabled. It may be allowed to become a threaded cgroup. - "threaded" : A threaded cgroup which is a member of a threaded subtree. A cgroup can be turned into a threaded cgroup by writing "threaded" to this file. ”h]”(j½)”}”(hŒcgroup.type”h]”hŒcgroup.type”…””}”(hj_ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mdhj[ ubjÍ)”}”(hhh]”(hæ)”}”(hŒ@A read-write single value file which exists on non-root cgroups.”h]”hŒ@A read-write single value file which exists on non-root cgroups.”…””}”(hjp hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MQhjm ubhæ)”}”(hŒaWhen read, it indicates the current type of the cgroup, which can be one of the following values.”h]”hŒaWhen read, it indicates the current type of the cgroup, which can be one of the following values.”…””}”(hj~ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MThjm ubjª)”}”(hhh]”(j¯)”}”(hŒ)"domain" : A normal valid domain cgroup. ”h]”hæ)”}”(hŒ("domain" : A normal valid domain cgroup.”h]”hŒ,â€œdomainâ€ : A normal valid domain cgroup.”…””}”(hj“ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MWhj ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjŒ ubj¯)”}”(hŒa"domain threaded" : A threaded domain cgroup which is serving as the root of a threaded subtree. ”h]”hæ)”}”(hŒ`"domain threaded" : A threaded domain cgroup which is serving as the root of a threaded subtree.”h]”hŒdâ€œdomain threadedâ€ : A threaded domain cgroup which is serving as the root of a threaded subtree.”…””}”(hj« hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MYhj§ ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjŒ ubj¯)”}”(hŒ"domain invalid" : A cgroup which is in an invalid state. It can't be populated or have controllers enabled. It may be allowed to become a threaded cgroup. ”h]”hæ)”}”(hŒœ"domain invalid" : A cgroup which is in an invalid state. It can't be populated or have controllers enabled. It may be allowed to become a threaded cgroup.”h]”hŒ¢â€œdomain invalidâ€ : A cgroup which is in an invalid state. It canâ€™t be populated or have controllers enabled. It may be allowed to become a threaded cgroup.”…””}”(hjÃ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M\hj¿ ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjŒ ubj¯)”}”(hŒH"threaded" : A threaded cgroup which is a member of a threaded subtree. ”h]”hæ)”}”(hŒG"threaded" : A threaded cgroup which is a member of a threaded subtree.”h]”hŒKâ€œthreadedâ€ : A threaded cgroup which is a member of a threaded subtree.”…””}”(hjÛ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M`hj× ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjŒ ubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h MWhjm ubhæ)”}”(hŒQA cgroup can be turned into a threaded cgroup by writing "threaded" to this file.”h]”hŒUA cgroup can be turned into a threaded cgroup by writing â€œthreadedâ€ to this file.”…””}”(hjõ hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mchjm ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj[ ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MdhjX ubj·)”}”(hXÀcgroup.procs A read-write new-line separated values file which exists on all cgroups. When read, it lists the PIDs of all processes which belong to the cgroup one-per-line. The PIDs are not ordered and the same PID may show up more than once if the process got moved to another cgroup and then back or the PID got recycled while reading. A PID can be written to migrate the process associated with the PID to the cgroup. The writer should match all of the following conditions. - It must have write access to the "cgroup.procs" file. - It must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups. When delegating a sub-hierarchy, write access to this file should be granted along with the containing directory. In a threaded cgroup, reading this file fails with EOPNOTSUPP as all the processes belong to the thread root. Writing is supported and moves every thread of the process to the cgroup. ”h]”(j½)”}”(hŒcgroup.procs”h]”hŒcgroup.procs”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M~hjubjÍ)”}”(hhh]”(hæ)”}”(hŒHA read-write new-line separated values file which exists on all cgroups.”h]”hŒHA read-write new-line separated values file which exists on all cgroups.”…””}”(hj$hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mghj!ubhæ)”}”(hŒüWhen read, it lists the PIDs of all processes which belong to the cgroup one-per-line. The PIDs are not ordered and the same PID may show up more than once if the process got moved to another cgroup and then back or the PID got recycled while reading.”h]”hŒüWhen read, it lists the PIDs of all processes which belong to the cgroup one-per-line. The PIDs are not ordered and the same PID may show up more than once if the process got moved to another cgroup and then back or the PID got recycled while reading.”…””}”(hj2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mjhj!ubhæ)”}”(hŒŒA PID can be written to migrate the process associated with the PID to the cgroup. The writer should match all of the following conditions.”h]”hŒŒA PID can be written to migrate the process associated with the PID to the cgroup. The writer should match all of the following conditions.”…””}”(hj@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mphj!ubjª)”}”(hhh]”(j¯)”}”(hŒ6It must have write access to the "cgroup.procs" file. ”h]”hæ)”}”(hŒ5It must have write access to the "cgroup.procs" file.”h]”hŒ9It must have write access to the â€œcgroup.procsâ€ file.”…””}”(hjUhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MthjQubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjNubj¯)”}”(hŒsIt must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups. ”h]”hæ)”}”(hŒrIt must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups.”h]”hŒvIt must have write access to the â€œcgroup.procsâ€ file of the common ancestor of the source and destination cgroups.”…””}”(hjmhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mvhjiubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjNubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h Mthj!ubhæ)”}”(hŒqWhen delegating a sub-hierarchy, write access to this file should be granted along with the containing directory.”h]”hŒqWhen delegating a sub-hierarchy, write access to this file should be granted along with the containing directory.”…””}”(hj‡hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Myhj!ubhæ)”}”(hŒ¸In a threaded cgroup, reading this file fails with EOPNOTSUPP as all the processes belong to the thread root. Writing is supported and moves every thread of the process to the cgroup.”h]”hŒ¸In a threaded cgroup, reading this file fails with EOPNOTSUPP as all the processes belong to the thread root. Writing is supported and moves every thread of the process to the cgroup.”…””}”(hj•hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M|hj!ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M~hjX ubj·)”}”(hXucgroup.threads A read-write new-line separated values file which exists on all cgroups. When read, it lists the TIDs of all threads which belong to the cgroup one-per-line. The TIDs are not ordered and the same TID may show up more than once if the thread got moved to another cgroup and then back or the TID got recycled while reading. A TID can be written to migrate the thread associated with the TID to the cgroup. The writer should match all of the following conditions. - It must have write access to the "cgroup.threads" file. - The cgroup that the thread is currently in must be in the same resource domain as the destination cgroup. - It must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups. When delegating a sub-hierarchy, write access to this file should be granted along with the containing directory. ”h]”(j½)”}”(hŒcgroup.threads”h]”hŒcgroup.threads”…””}”(hj³hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M—hj¯ubjÍ)”}”(hhh]”(hæ)”}”(hŒHA read-write new-line separated values file which exists on all cgroups.”h]”hŒHA read-write new-line separated values file which exists on all cgroups.”…””}”(hjÄhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÁubhæ)”}”(hŒùWhen read, it lists the TIDs of all threads which belong to the cgroup one-per-line. The TIDs are not ordered and the same TID may show up more than once if the thread got moved to another cgroup and then back or the TID got recycled while reading.”h]”hŒùWhen read, it lists the TIDs of all threads which belong to the cgroup one-per-line. The TIDs are not ordered and the same TID may show up more than once if the thread got moved to another cgroup and then back or the TID got recycled while reading.”…””}”(hjÒhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M„hjÁubhæ)”}”(hŒ‹A TID can be written to migrate the thread associated with the TID to the cgroup. The writer should match all of the following conditions.”h]”hŒ‹A TID can be written to migrate the thread associated with the TID to the cgroup. The writer should match all of the following conditions.”…””}”(hjàhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŠhjÁubjª)”}”(hhh]”(j¯)”}”(hŒ8It must have write access to the "cgroup.threads" file. ”h]”hæ)”}”(hŒ7It must have write access to the "cgroup.threads" file.”h]”hŒ;It must have write access to the â€œcgroup.threadsâ€ file.”…””}”(hjõhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŽhjñubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjîubj¯)”}”(hŒjThe cgroup that the thread is currently in must be in the same resource domain as the destination cgroup. ”h]”hæ)”}”(hŒiThe cgroup that the thread is currently in must be in the same resource domain as the destination cgroup.”h]”hŒiThe cgroup that the thread is currently in must be in the same resource domain as the destination cgroup.”…””}”(hj hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjîubj¯)”}”(hŒsIt must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups. ”h]”hæ)”}”(hŒrIt must have write access to the "cgroup.procs" file of the common ancestor of the source and destination cgroups.”h]”hŒvIt must have write access to the â€œcgroup.procsâ€ file of the common ancestor of the source and destination cgroups.”…””}”(hj%hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M“hj!ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjîubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h MŽhjÁubhæ)”}”(hŒqWhen delegating a sub-hierarchy, write access to this file should be granted along with the containing directory.”h]”hŒqWhen delegating a sub-hierarchy, write access to this file should be granted along with the containing directory.”…””}”(hj?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M–hjÁubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj¯ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M—hjX ubj·)”}”(hŒÅcgroup.controllers A read-only space separated values file which exists on all cgroups. It shows space separated list of all controllers available to the cgroup. The controllers are not ordered. ”h]”(j½)”}”(hŒcgroup.controllers”h]”hŒcgroup.controllers”…””}”(hj]hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MžhjYubjÍ)”}”(hhh]”(hæ)”}”(hŒDA read-only space separated values file which exists on all cgroups.”h]”hŒDA read-only space separated values file which exists on all cgroups.”…””}”(hjnhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mšhjkubhæ)”}”(hŒkIt shows space separated list of all controllers available to the cgroup. The controllers are not ordered.”h]”hŒkIt shows space separated list of all controllers available to the cgroup. The controllers are not ordered.”…””}”(hj|hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjkubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjYubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MžhjX ubj·)”}”(hXjcgroup.subtree_control A read-write space separated values file which exists on all cgroups. Starts out empty. When read, it shows space separated list of the controllers which are enabled to control resource distribution from the cgroup to its children. Space separated list of controllers prefixed with '+' or '-' can be written to enable or disable controllers. A controller name prefixed with '+' enables the controller and '-' disables. If a controller appears more than once on the list, the last one is effective. When multiple enable and disable operations are specified, either all succeed or all fail. ”h]”(j½)”}”(hŒcgroup.subtree_control”h]”hŒcgroup.subtree_control”…””}”(hjšhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhj–ubjÍ)”}”(hhh]”(hæ)”}”(hŒXA read-write space separated values file which exists on all cgroups. Starts out empty.”h]”hŒXA read-write space separated values file which exists on all cgroups. Starts out empty.”…””}”(hj«hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¡hj¨ubhæ)”}”(hŒWhen read, it shows space separated list of the controllers which are enabled to control resource distribution from the cgroup to its children.”h]”hŒWhen read, it shows space separated list of the controllers which are enabled to control resource distribution from the cgroup to its children.”…””}”(hj¹hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¤hj¨ubhæ)”}”(hXgSpace separated list of controllers prefixed with '+' or '-' can be written to enable or disable controllers. A controller name prefixed with '+' enables the controller and '-' disables. If a controller appears more than once on the list, the last one is effective. When multiple enable and disable operations are specified, either all succeed or all fail.”h]”hXwSpace separated list of controllers prefixed with â€˜+â€™ or â€˜-â€™ can be written to enable or disable controllers. A controller name prefixed with â€˜+â€™ enables the controller and â€˜-â€™ disables. If a controller appears more than once on the list, the last one is effective. When multiple enable and disable operations are specified, either all succeed or all fail.”…””}”(hjÇhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¨hj¨ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj–ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MhjX ubj·)”}”(hXncgroup.events A read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. populated 1 if the cgroup or its descendants contains any live processes; otherwise, 0. frozen 1 if the cgroup is frozen; otherwise, 0. ”h]”(j½)”}”(hŒ cgroup.events”h]”hŒ cgroup.events”…””}”(hjåhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¹hjáubjÍ)”}”(hhh]”(hæ)”}”(hŒºA read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event.”h]”hŒºA read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event.”…””}”(hjöhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M°hjóubj¬)”}”(hŒšpopulated 1 if the cgroup or its descendants contains any live processes; otherwise, 0. frozen 1 if the cgroup is frozen; otherwise, 0. ”h]”j²)”}”(hhh]”(j·)”}”(hŒWpopulated 1 if the cgroup or its descendants contains any live processes; otherwise, 0.”h]”(j½)”}”(hŒ populated”h]”hŒ populated”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¶hjubjÍ)”}”(hhh]”hæ)”}”(hŒM1 if the cgroup or its descendants contains any live processes; otherwise, 0.”h]”hŒM1 if the cgroup or its descendants contains any live processes; otherwise, 0.”…””}”(hj hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¶hjubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¶hjubj·)”}”(hŒ0frozen 1 if the cgroup is frozen; otherwise, 0. ”h]”(j½)”}”(hŒfrozen”h]”hŒfrozen”…””}”(hj>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¹hj:ubjÍ)”}”(hhh]”hæ)”}”(hŒ(1 if the cgroup is frozen; otherwise, 0.”h]”hŒ(1 if the cgroup is frozen; otherwise, 0.”…””}”(hjOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹hjLubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj:ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¹hjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mµhjóubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjáubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¹hjX ubj·)”}”(hŒñcgroup.max.descendants A read-write single value files. The default is "max". Maximum allowed number of descent cgroups. If the actual number of descendants is equal or larger, an attempt to create a new cgroup in the hierarchy will fail. ”h]”(j½)”}”(hŒcgroup.max.descendants”h]”hŒcgroup.max.descendants”…””}”(hj…hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÀhjubjÍ)”}”(hhh]”(hæ)”}”(hŒ7A read-write single value files. The default is "max".”h]”hŒ;A read-write single value files. The default is â€œmaxâ€.”…””}”(hj–hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¼hj“ubhæ)”}”(hŒ Maximum allowed number of descent cgroups. If the actual number of descendants is equal or larger, an attempt to create a new cgroup in the hierarchy will fail.”h]”hŒ Maximum allowed number of descent cgroups. If the actual number of descendants is equal or larger, an attempt to create a new cgroup in the hierarchy will fail.”…””}”(hj¤hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¾hj“ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÀhjX ubj·)”}”(hŒåcgroup.max.depth A read-write single value files. The default is "max". Maximum allowed descent depth below the current cgroup. If the actual descent depth is equal or larger, an attempt to create a new child cgroup will fail. ”h]”(j½)”}”(hŒcgroup.max.depth”h]”hŒcgroup.max.depth”…””}”(hjÂhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÇhj¾ubjÍ)”}”(hhh]”(hæ)”}”(hŒ7A read-write single value files. The default is "max".”h]”hŒ;A read-write single value files. The default is â€œmaxâ€.”…””}”(hjÓhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÃhjÐubhæ)”}”(hŒšMaximum allowed descent depth below the current cgroup. If the actual descent depth is equal or larger, an attempt to create a new child cgroup will fail.”h]”hŒšMaximum allowed descent depth below the current cgroup. If the actual descent depth is equal or larger, an attempt to create a new child cgroup will fail.”…””}”(hjáhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÅhjÐubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj¾ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÇhjX ubj·)”}”(hX»cgroup.stat A read-only flat-keyed file with the following entries: nr_descendants Total number of visible descendant cgroups. nr_dying_descendants Total number of dying descendant cgroups. A cgroup becomes dying after being deleted by a user. The cgroup will remain in dying state for some time undefined time (which can depend on system load) before being completely destroyed. A process can't enter a dying cgroup under any circumstances, a dying cgroup can't revive. A dying cgroup can consume system resources not exceeding limits, which were active at the moment of cgroup deletion. nr_subsys_ Total number of live cgroup subsystems (e.g memory cgroup) at and beneath the current cgroup. nr_dying_subsys_ Total number of dying cgroup subsystems (e.g. memory cgroup) at and beneath the current cgroup. ”h]”(j½)”}”(hŒcgroup.stat”h]”hŒcgroup.stat”…””}”(hjÿhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MáhjûubjÍ)”}”(hhh]”(hæ)”}”(hŒ7A read-only flat-keyed file with the following entries:”h]”hŒ7A read-only flat-keyed file with the following entries:”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÊhj ubj¬)”}”(hXTnr_descendants Total number of visible descendant cgroups. nr_dying_descendants Total number of dying descendant cgroups. A cgroup becomes dying after being deleted by a user. The cgroup will remain in dying state for some time undefined time (which can depend on system load) before being completely destroyed. A process can't enter a dying cgroup under any circumstances, a dying cgroup can't revive. A dying cgroup can consume system resources not exceeding limits, which were active at the moment of cgroup deletion. nr_subsys_ Total number of live cgroup subsystems (e.g memory cgroup) at and beneath the current cgroup. nr_dying_subsys_ Total number of dying cgroup subsystems (e.g. memory cgroup) at and beneath the current cgroup. ”h]”j²)”}”(hhh]”(j·)”}”(hŒ;nr_descendants Total number of visible descendant cgroups. ”h]”(j½)”}”(hŒnr_descendants”h]”hŒnr_descendants”…””}”(hj)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÍhj%ubjÍ)”}”(hhh]”hæ)”}”(hŒ+Total number of visible descendant cgroups.”h]”hŒ+Total number of visible descendant cgroups.”…””}”(hj:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÍhj7ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj%ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÍhj"ubj·)”}”(hXÐnr_dying_descendants Total number of dying descendant cgroups. A cgroup becomes dying after being deleted by a user. The cgroup will remain in dying state for some time undefined time (which can depend on system load) before being completely destroyed. A process can't enter a dying cgroup under any circumstances, a dying cgroup can't revive. A dying cgroup can consume system resources not exceeding limits, which were active at the moment of cgroup deletion. ”h]”(j½)”}”(hŒnr_dying_descendants”•¨5h]”hŒnr_dying_descendants”…””}”(hjXhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÙhjTubjÍ)”}”(hhh]”(hæ)”}”(hŒçTotal number of dying descendant cgroups. A cgroup becomes dying after being deleted by a user. The cgroup will remain in dying state for some time undefined time (which can depend on system load) before being completely destroyed.”h]”hŒçTotal number of dying descendant cgroups. A cgroup becomes dying after being deleted by a user. The cgroup will remain in dying state for some time undefined time (which can depend on system load) before being completely destroyed.”…””}”(hjihžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÐhjfubhæ)”}”(hŒZA process can't enter a dying cgroup under any circumstances, a dying cgroup can't revive.”h]”hŒ^A process canâ€™t enter a dying cgroup under any circumstances, a dying cgroup canâ€™t revive.”…””}”(hjwhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÕhjfubhæ)”}”(hŒuA dying cgroup can consume system resources not exceeding limits, which were active at the moment of cgroup deletion.”h]”hŒuA dying cgroup can consume system resources not exceeding limits, which were active at the moment of cgroup deletion.”…””}”(hj…hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MØhjfubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjTubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÙhj"ubj·)”}”(hŒxnr_subsys_ Total number of live cgroup subsystems (e.g memory cgroup) at and beneath the current cgroup. ”h]”(j½)”}”(hŒnr_subsys_”h]”hŒnr_subsys_”…””}”(hj£hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÝhjŸubjÍ)”}”(hhh]”hæ)”}”(hŒ]Total number of live cgroup subsystems (e.g memory cgroup) at and beneath the current cgroup.”h]”hŒ]Total number of live cgroup subsystems (e.g memory cgroup) at and beneath the current cgroup.”…””}”(hj´hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÜhj±ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjŸubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÝhj"ubj·)”}”(hŒ€nr_dying_subsys_ Total number of dying cgroup subsystems (e.g. memory cgroup) at and beneath the current cgroup. ”h]”(j½)”}”(hŒnr_dying_subsys_”h]”hŒnr_dying_subsys_”…””}”(hjÒhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MáhjÎubjÍ)”}”(hhh]”hæ)”}”(hŒ_Total number of dying cgroup subsystems (e.g. memory cgroup) at and beneath the current cgroup.”h]”hŒ_Total number of dying cgroup subsystems (e.g. memory cgroup) at and beneath the current cgroup.”…””}”(hjãhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Màhjàubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÎubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Máhj"ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MÌhj ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjûubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MáhjX ubj·)”}”(hX‘cgroup.freeze A read-write single value file which exists on non-root cgroups. Allowed values are "0" and "1". The default is "0". Writing "1" to the file causes freezing of the cgroup and all descendant cgroups. This means that all belonging processes will be stopped and will not run until the cgroup will be explicitly unfrozen. Freezing of the cgroup may take some time; when this action is completed, the "frozen" value in the cgroup.events control file will be updated to "1" and the corresponding notification will be issued. A cgroup can be frozen either by its own settings, or by settings of any ancestor cgroups. If any of ancestor cgroups is frozen, the cgroup will remain frozen. Processes in the frozen cgroup can be killed by a fatal signal. They also can enter and leave a frozen cgroup: either by an explicit move by a user, or if freezing of the cgroup races with fork(). If a process is moved to a frozen cgroup, it stops. If a process is moved out of a frozen cgroup, it becomes running. Frozen status of a cgroup doesn't affect any cgroup tree operations: it's possible to delete a frozen (and empty) cgroup, as well as create new sub-cgroups. ”h]”(j½)”}”(hŒ cgroup.freeze”h]”hŒ cgroup.freeze”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MûhjubjÍ)”}”(hhh]”(hæ)”}”(hŒtA read-write single value file which exists on non-root cgroups. Allowed values are "0" and "1". The default is "0".”h]”hŒ€A read-write single value file which exists on non-root cgroups. Allowed values are â€œ0â€ and â€œ1â€. The default is â€œ0â€.”…””}”(hj*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mähj'ubhæ)”}”(hX‘Writing "1" to the file causes freezing of the cgroup and all descendant cgroups. This means that all belonging processes will be stopped and will not run until the cgroup will be explicitly unfrozen. Freezing of the cgroup may take some time; when this action is completed, the "frozen" value in the cgroup.events control file will be updated to "1" and the corresponding notification will be issued.”h]”hXWriting â€œ1â€ to the file causes freezing of the cgroup and all descendant cgroups. This means that all belonging processes will be stopped and will not run until the cgroup will be explicitly unfrozen. Freezing of the cgroup may take some time; when this action is completed, the â€œfrozenâ€ value in the cgroup.events control file will be updated to â€œ1â€ and the corresponding notification will be issued.”…””}”(hj8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mçhj'ubhæ)”}”(hŒŸA cgroup can be frozen either by its own settings, or by settings of any ancestor cgroups. If any of ancestor cgroups is frozen, the cgroup will remain frozen.”h]”hŒŸA cgroup can be frozen either by its own settings, or by settings of any ancestor cgroups. If any of ancestor cgroups is frozen, the cgroup will remain frozen.”…””}”(hjFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mïhj'ubhæ)”}”(hX:Processes in the frozen cgroup can be killed by a fatal signal. They also can enter and leave a frozen cgroup: either by an explicit move by a user, or if freezing of the cgroup races with fork(). If a process is moved to a frozen cgroup, it stops. If a process is moved out of a frozen cgroup, it becomes running.”h]”hX:Processes in the frozen cgroup can be killed by a fatal signal. They also can enter and leave a frozen cgroup: either by an explicit move by a user, or if freezing of the cgroup races with fork(). If a process is moved to a frozen cgroup, it stops. If a process is moved out of a frozen cgroup, it becomes running.”…””}”(hjThžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Móhj'ubhæ)”}”(hŒœFrozen status of a cgroup doesn't affect any cgroup tree operations: it's possible to delete a frozen (and empty) cgroup, as well as create new sub-cgroups.”h]”hŒ Frozen status of a cgroup doesnâ€™t affect any cgroup tree operations: itâ€™s possible to delete a frozen (and empty) cgroup, as well as create new sub-cgroups.”…””}”(hjbhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mùhj'ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MûhjX ubj·)”}”(hX#cgroup.kill A write-only single value file which exists in non-root cgroups. The only allowed value is "1". Writing "1" to the file causes the cgroup and all descendant cgroups to be killed. This means that all processes located in the affected cgroup tree will be killed via SIGKILL. Killing a cgroup tree will deal with concurrent forks appropriately and is protected against migrations. In a threaded cgroup, writing this file fails with EOPNOTSUPP as killing cgroups is a process directed operation, i.e. it affects the whole thread-group. ”h]”(j½)”}”(hŒcgroup.kill”h]”hŒcgroup.kill”…””}”(hj€hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M hj|ubjÍ)”}”(hhh]”(hæ)”}”(hŒ_A write-only single value file which exists in non-root cgroups. The only allowed value is "1".”h]”hŒcA write-only single value file which exists in non-root cgroups. The only allowed value is â€œ1â€.”…””}”(hj‘hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MþhjŽubhæ)”}”(hŒ°Writing "1" to the file causes the cgroup and all descendant cgroups to be killed. This means that all processes located in the affected cgroup tree will be killed via SIGKILL.”h]”hŒ´Writing â€œ1â€ to the file causes the cgroup and all descendant cgroups to be killed. This means that all processes located in the affected cgroup tree will be killed via SIGKILL.”…””}”(hjŸhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjŽubhæ)”}”(hŒhKilling a cgroup tree will deal with concurrent forks appropriately and is protected against migrations.”h]”hŒhKilling a cgroup tree will deal with concurrent forks appropriately and is protected against migrations.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjŽubhæ)”}”(hŒ™In a threaded cgroup, writing this file fails with EOPNOTSUPP as killing cgroups is a process directed operation, i.e. it affects the whole thread-group.”h]”hŒ™In a threaded cgroup, writing this file fails with EOPNOTSUPP as killing cgroups is a process directed operation, i.e. it affects the whole thread-group.”…””}”(hj»hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjŽubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj|ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M hjX ubj·)”}”(hX cgroup.pressure A read-write single value file that allowed values are "0" and "1". The default is "1". Writing "0" to the file will disable the cgroup PSI accounting. Writing "1" to the file will re-enable the cgroup PSI accounting. This control attribute is not hierarchical, so disable or enable PSI accounting in a cgroup does not affect PSI accounting in descendants and doesn't need pass enablement via ancestors from root. The reason this control attribute exists is that PSI accounts stalls for each cgroup separately and aggregates it at each level of the hierarchy. This may cause non-negligible overhead for some workloads when under deep level of the hierarchy, in which case this control attribute can be used to disable PSI accounting in the non-leaf cgroups. ”h]”(j½)”}”(hŒcgroup.pressure”h]”hŒcgroup.pressure”…””}”(hjÙhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MhjÕubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-write single value file that allowed values are "0" and "1". The default is "1".”h]”hŒcA read-write single value file that allowed values are â€œ0â€ and â€œ1â€. The default is â€œ1â€.”…””}”(hjêhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjçubhæ)”}”(hŒWriting "0" to the file will disable the cgroup PSI accounting. Writing "1" to the file will re-enable the cgroup PSI accounting.”h]”hŒ‰Writing â€œ0â€ to the file will disable the cgroup PSI accounting. Writing â€œ1â€ to the file will re-enable the cgroup PSI accounting.”…””}”(hjøhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjçubhæ)”}”(hŒÃThis control attribute is not hierarchical, so disable or enable PSI accounting in a cgroup does not affect PSI accounting in descendants and doesn't need pass enablement via ancestors from root.”h]”hŒÅThis control attribute is not hierarchical, so disable or enable PSI accounting in a cgroup does not affect PSI accounting in descendants and doesnâ€™t need pass enablement via ancestors from root.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjçubhæ)”}”(hXWThe reason this control attribute exists is that PSI accounts stalls for each cgroup separately and aggregates it at each level of the hierarchy. This may cause non-negligible overhead for some workloads when under deep level of the hierarchy, in which case this control attribute can be used to disable PSI accounting in the non-leaf cgroups.”h]”hXWThe reason this control attribute exists is that PSI accounts stalls for each cgroup separately and aggregates it at each level of the hierarchy. This may cause non-negligible overhead for some workloads when under deep level of the hierarchy, in which case this control attribute can be used to disable PSI accounting in the non-leaf cgroups.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjçubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÕubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MhjX ubj·)”}”(hŒŸirq.pressure A read-write nested-keyed file. Shows pressure stall information for IRQ/SOFTIRQ. See :ref:`Documentation/accounting/psi.rst ` for details. ”h]”(j½)”}”(hŒirq.pressure”h]”hŒirq.pressure”…””}”(hj2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M!hj.ubjÍ)”}”(hhh]”(hæ)”}”(hŒA read-write nested-keyed file.”h]”hŒA read-write nested-keyed file.”…””}”(hjChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj@ubhæ)”}”(hŒpShows pressure stall information for IRQ/SOFTIRQ. See :ref:`Documentation/accounting/psi.rst ` for details.”h]”(hŒ6Shows pressure stall information for IRQ/SOFTIRQ. See ”…””}”(hjQhžhhŸNh Nubh)”}”(hŒ-:ref:`Documentation/accounting/psi.rst `”h]”jX)”}”(hj[h]”hŒ Documentation/accounting/psi.rst”…””}”(hj]hžhhŸNh Nubah}”(h]”h ]”(jcŒstd”Œstd-ref”eh"]”h$]”h&]”uh1jWhjYubah}”(h]”h ]”h"]”h$]”h&]”Œrefdoc”jpŒ refdomain”jgŒreftype”Œref”Œrefexplicit”ˆŒrefwarn”ˆjvŒpsi”uh1hhŸh¯h M hjQubhŒ for details.”…””}”(hjQhžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hj@ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj.ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M!hjX ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjT ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MPhj5 hžhubeh}”(h]”Œcore-interface-files”ah ]”h"]”Œcore interface files”ah$]”h&]”uh1h°hj;hžhhŸh¯h MLubeh}”(h]”Œinterface-files”ah ]”h"]”Œinterface files”ah$]”h&]”uh1h°hh²hžhhŸh¯h MÚubh±)”}”(hhh]”(h¶)”}”(hŒControllers”h]”hŒControllers”…””}”(hj®hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj«hžhhŸh¯h M$ubh¢)”}”(hŒ.. _cgroup-v2-cpu:”h]”h}”(h]”h ]”h"]”h$]”h&]”hŒ cgroup-v2-cpu”uh1h¡h M&hj«hžhhŸh¯ubh±)”}”(hhh]”(h¶)”}”(hŒCPU”h]”hŒCPU”…””}”(hjÊhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjÇhžhhŸh¯h M)ubhæ)”}”(hŒçThe "cpu" controllers regulates distribution of CPU cycles. This controller implements weight and absolute bandwidth limit models for normal scheduling policy and absolute bandwidth allocation model for realtime scheduling policy.”h]”hŒëThe â€œcpuâ€ controllers regulates distribution of CPU cycles. This controller implements weight and absolute bandwidth limit models for normal scheduling policy and absolute bandwidth allocation model for realtime scheduling policy.”…””}”(hjØhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M+hjÇhžhubhæ)”}”(hXIn all the above models, cycles distribution is defined only on a temporal base and it does not account for the frequency at which tasks are executed. The (optional) utilization clamping support allows to hint the schedutil cpufreq governor about the minimum desired frequency which should always be provided by a CPU, as well as the maximum desired frequency, which should not be exceeded by a CPU.”h]”hXIn all the above models, cycles distribution is defined only on a temporal base and it does not account for the frequency at which tasks are executed. The (optional) utilization clamping support allows to hint the schedutil cpufreq governor about the minimum desired frequency which should always be provided by a CPU, as well as the maximum desired frequency, which should not be exceeded by a CPU.”…””}”(hjæhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M0hjÇhžhubhæ)”}”(hX4WARNING: cgroup2 cpu controller doesn't yet fully support the control of realtime processes. For a kernel built with the CONFIG_RT_GROUP_SCHED option enabled for group scheduling of realtime processes, the cpu controller can only be enabled when all RT processes are in the root cgroup. Be aware that system management software may already have placed RT processes into non-root cgroups during the system boot process, and these processes may need to be moved to the root cgroup before the cpu controller can be enabled with a CONFIG_RT_GROUP_SCHED enabled kernel.”h]”hX6WARNING: cgroup2 cpu controller doesnâ€™t yet fully support the control of realtime processes. For a kernel built with the CONFIG_RT_GROUP_SCHED option enabled for group scheduling of realtime processes, the cpu controller can only be enabled when all RT processes are in the root cgroup. Be aware that system management software may already have placed RT processes into non-root cgroups during the system boot process, and these processes may need to be moved to the root cgroup before the cpu controller can be enabled with a CONFIG_RT_GROUP_SCHED enabled kernel.”…””}”(hjôhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M7hjÇhžhubhæ)”}”(hXrWith CONFIG_RT_GROUP_SCHED disabled, this limitation does not apply and some of the interface files either affect realtime processes or account for them. See the following section for details. Only the cpu controller is affected by CONFIG_RT_GROUP_SCHED. Other controllers can be used for the resource control of realtime processes irrespective of CONFIG_RT_GROUP_SCHED.”h]”hXrWith CONFIG_RT_GROUP_SCHED disabled, this limitation does not apply and some of the interface files either affect realtime processes or account for them. See the following section for details. Only the cpu controller is affected by CONFIG_RT_GROUP_SCHED. Other controllers can be used for the resource control of realtime processes irrespective of CONFIG_RT_GROUP_SCHED.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M@hjÇhžhubh±)”}”(hhh]”(h¶)”}”(hŒCPU Interface Files”h]”hŒCPU Interface Files”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjhžhhŸh¯h MHubhæ)”}”(hŒ'All time durations are in microseconds.”h]”hŒ'All time durations are in microseconds.”…””}”(hj!hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MJhjhžhubj¬)”}”(hXy cpu.stat A read-only flat-keyed file. This file exists whether the controller is enabled or not. It always reports the following three stats: - usage_usec - user_usec - system_usec and the following five when the controller is enabled: - nr_periods - nr_throttled - throttled_usec - nr_bursts - burst_usec cpu.weight A read-write single value file which exists on non-root cgroups. The default is "100". For non idle groups (cpu.idle = 0), the weight is in the range [1, 10000]. If the cgroup has been configured to be SCHED_IDLE (cpu.idle = 1), then the weight will show as a 0. cpu.weight.nice A read-write single value file which exists on non-root cgroups. The default is "0". The nice value is in the range [-20, 19]. This interface file is an alternative interface for "cpu.weight" and allows reading and setting weight using the same values used by nice(2). Because the range is smaller and granularity is coarser for the nice values, the read value is the closest approximation of the current weight. cpu.max A read-write two value file which exists on non-root cgroups. The default is "max 100000". The maximum bandwidth limit. It's in the following format:: $MAX $PERIOD which indicates that the group may consume up to $MAX in each $PERIOD duration. "max" for $MAX indicates no limit. If only one number is written, $MAX is updated. cpu.max.burst A read-write single value file which exists on non-root cgroups. The default is "0". The burst in the range [0, $MAX]. cpu.pressure A read-write nested-keyed file. Shows pressure stall information for CPU. See :ref:`Documentation/accounting/psi.rst ` for details. cpu.uclamp.min A read-write single value file which exists on non-root cgroups. The default is "0", i.e. no utilization boosting. The requested minimum utilization (protection) as a percentage rational number, e.g. 12.34 for 12.34%. This interface allows reading and setting minimum utilization clamp values similar to the sched_setattr(2). This minimum utilization value is used to clamp the task specific minimum utilization clamp. The requested minimum utilization (protection) is always capped by the current value for the maximum utilization (limit), i.e. `cpu.uclamp.max`. cpu.uclamp.max A read-write single value file which exists on non-root cgroups. The default is "max". i.e. no utilization capping The requested maximum utilization (limit) as a percentage rational number, e.g. 98.76 for 98.76%. This interface allows reading and setting maximum utilization clamp values similar to the sched_setattr(2). This maximum utilization value is used to clamp the task specific maximum utilization clamp. cpu.idle A read-write single value file which exists on non-root cgroups. The default is 0. This is the cgroup analog of the per-task SCHED_IDLE sched policy. Setting this value to a 1 will make the scheduling policy of the cgroup SCHED_IDLE. The threads inside the cgroup will retain their own relative priorities, but the cgroup itself will be treated as very low priority relative to its peers. ”h]”j²)”}”(hhh]”(j·)”}”(hX6cpu.stat A read-only flat-keyed file. This file exists whether the controller is enabled or not. It always reports the following three stats: - usage_usec - user_usec - system_usec and the following five when the controller is enabled: - nr_periods - nr_throttled - throttled_usec - nr_bursts - burst_usec ”h]”(j½)”}”(hŒcpu.stat”h]”hŒcpu.stat”…””}”(hj:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M\hj6ubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-only flat-keyed file. This file exists whether the controller is enabled or not.”h]”hŒWA read-only flat-keyed file. This file exists whether the controller is enabled or not.”…””}”(hjKhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MMhjHubhæ)”}”(hŒ,It always reports the following three stats:”h]”hŒ,It always reports the following three stats:”…””}”(hjYhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MPhjHubjª)”}”(hhh]”(j¯)”}”(hŒ usage_usec”h]”hæ)”}”(hjlh]”hŒ usage_usec”…””}”(hjnhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MRhjjubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjgubj¯)”}”(hŒ user_usec”h]”hæ)”}”(hjƒh]”hŒ user_usec”…””}”(hj…hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MShjubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjgubj¯)”}”(hŒsystem_usec ”h]”hæ)”}”(hŒsystem_usec”h]”hŒsystem_usec”…””}”(hjœhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MThj˜ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjgubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h MRhjHubhæ)”}”(hŒ6and the following five when the controller is enabled:”h]”hŒ6and the following five when the controller is enabled:”…””}”(hj¶hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MVhjHubjª)”}”(hhh]”(j¯)”}”(hŒ nr_periods”h]”hæ)”}”(hjÉh]”hŒ nr_periods”…””}”(hjËhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MXhjÇubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjÄubj¯)”}”(hŒnr_throttled”h]”hæ)”}”(hjàh]”hŒnr_throttled”…””}”(hjâhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MYhjÞubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjÄubj¯)”}”(hŒthrottled_usec”h]”hæ)”}”(hj÷h]”hŒthrottled_usec”…””}”(hjùhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MZhjõubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjÄubj¯)”}”(hŒ nr_bursts”h]”hæ)”}”(hjh]”hŒ nr_bursts”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M[hjubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjÄubj¯)”}”(hŒburst_usec ”h]”hæ)”}”(hŒ burst_usec”h]”hŒ burst_usec”…””}”(hj'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M\hj#ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjÄubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h MXhjHubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj6ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M\hj3ubj·)”}”(hXcpu.weight A read-write single value file which exists on non-root cgroups. The default is "100". For non idle groups (cpu.idle = 0), the weight is in the range [1, 10000]. If the cgroup has been configured to be SCHED_IDLE (cpu.idle = 1), then the weight will show as a 0. ”h]”(j½)”}”(hŒ cpu.weight”h]”hŒ cpu.weight”…””}”(hjQhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MfhjMubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-write single value file which exists on non-root cgroups. The default is "100".”h]”hŒ[A read-write single value file which exists on non-root cgroups. The default is â€œ100â€.”…””}”(hjbhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M_hj_ubhæ)”}”(hŒJFor non idle groups (cpu.idle = 0), the weight is in the range [1, 10000].”h]”hŒJFor non idle groups (cpu.idle = 0), the weight is in the range [1, 10000].”…””}”(hjphžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mbhj_ubhæ)”}”(hŒdIf the cgroup has been configured to be SCHED_IDLE (cpu.idle = 1), then the weight will show as a 0.”h]”hŒdIf the cgroup has been configured to be SCHED_IDLE (cpu.idle = 1), then the weight will show as a 0.”…””}”(hj~hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mehj_ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjMubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mfhj3ubj·)”}”(hX±cpu.weight.nice A read-write single value file which exists on non-root cgroups. The default is "0". The nice value is in the range [-20, 19]. This interface file is an alternative interface for "cpu.weight" and allows reading and setting weight using the same values used by nice(2). Because the range is smaller and granularity is coarser for the nice values, the read value is the closest approximation of the current weight. ”h]”(j½)”}”(hŒcpu.weight.nice”h]”hŒcpu.weight.nice”…””}”(hjœhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mrhj˜ubjÍ)”}”(hhh]”(hæ)”}”(hŒUA read-write single value file which exists on non-root cgroups. The default is "0".”h]”hŒYA read-write single value file which exists on non-root cgroups. The default is â€œ0â€.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mihjªubhæ)”}”(hŒ)The nice value is in the range [-20, 19].”h]”hŒ)The nice value is in the range [-20, 19].”…””}”(hj»hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mlhjªubhæ)”}”(hXThis interface file is an alternative interface for "cpu.weight" and allows reading and setting weight using the same values used by nice(2). Because the range is smaller and granularity is coarser for the nice values, the read value is the closest approximation of the current weight.”h]”hX"This interface file is an alternative interface for â€œcpu.weightâ€ and allows reading and setting weight using the same values used by nice(2). Because the range is smaller and granularity is coarser for the nice values, the read value is the closest approximation of the current weight.”…””}”(hjÉhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mnhjªubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj˜ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mrhj3ubj·)”}”(hXWcpu.max A read-write two value file which exists on non-root cgroups. The default is "max 100000". The maximum bandwidth limit. It's in the following format:: $MAX $PERIOD which indicates that the group may consume up to $MAX in each $PERIOD duration. "max" for $MAX indicates no limit. If only one number is written, $MAX is updated. ”h]”(j½)”}”(hŒcpu.max”h]”hŒcpu.max”…””}”(hjçhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M~hjãubjÍ)”}”(hhh]”(hæ)”}”(hŒZA read-write two value file which exists on non-root cgroups. The default is "max 100000".”h]”hŒ^A read-write two value file which exists on non-root cgroups. The default is â€œmax 100000â€.”…””}”(hjøhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Muhjõubhæ)”}”(hŒ` for details. ”h]”(j½)”}”(hŒcpu.pressure”h]”hŒcpu.pressure”…””}”(hj}hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MŠhjyubjÍ)”}”(hhh]”(hæ)”}”(hŒA read-write nested-keyed file.”h]”hŒA read-write nested-keyed file.”…””}”(hjŽhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‡hj‹ubhæ)”}”(hŒhShows pressure stall information for CPU. See :ref:`Documentation/accounting/psi.rst ` for details.”h]”(hŒ.Shows pressure stall information for CPU. See ”…””}”(hjœhžhhŸNh Nubh)”}”(hŒ-:ref:`Documentation/accounting/psi.rst `”h]”jX)”}”(hj¦h]”hŒ Documentation/accounting/psi.rst”…””}”(hj¨hžhhŸNh Nubah}”(h]”h ]”(jcŒstd”Œstd-ref”eh"]”h$]”h&]”uh1jWhj¤ubah}”(h]”h ]”h"]”h$]”h&]”Œrefdoc”jpŒ refdomain”j²Œreftype”Œref”Œrefexplicit”ˆŒrefwarn”ˆjvŒpsi”uh1hhŸh¯h M‰hjœubhŒ for details.”…””}”(hjœhžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‰hj‹ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjyubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MŠhj3ubj·)”}”(hXFcpu.uclamp.min A read-write single value file which exists on non-root cgroups. The default is "0", i.e. no utilization boosting. The requested minimum utilization (protection) as a percentage rational number, e.g. 12.34 for 12.34%. This interface allows reading and setting minimum utilization clamp values similar to the sched_setattr(2). This minimum utilization value is used to clamp the task specific minimum utilization clamp. The requested minimum utilization (protection) is always capped by the current value for the maximum utilization (limit), i.e. `cpu.uclamp.max`. ”h]”(j½)”}”(hŒcpu.uclamp.min”h]”hŒcpu.uclamp.min”…””}”(hjÞhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M™hjÚubjÍ)”}”(hhh]”(hæ)”}”(hŒrA read-write single value file which exists on non-root cgroups. The default is "0", i.e. no utilization boosting.”h]”hŒvA read-write single value file which exists on non-root cgroups. The default is â€œ0â€, i.e. no utilization boosting.”…””}”(hjïhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjìubhæ)”}”(hŒfThe requested minimum utilization (protection) as a percentage rational number, e.g. 12.34 for 12.34%.”h]”hŒfThe requested minimum utilization (protection) as a percentage rational number, e.g. 12.34 for 12.34%.”…””}”(hjýhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjìubhæ)”}”(hŒÈThis interface allows reading and setting minimum utilization clamp values similar to the sched_setattr(2). This minimum utilization value is used to clamp the task specific minimum utilization clamp.”h]”hŒÈThis interface allows reading and setting minimum utilization clamp values similar to the sched_setattr(2). This minimum utilization value is used to clamp the task specific minimum utilization clamp.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M“hjìubhæ)”}”(hŒThe requested minimum utilization (protection) is always capped by the current value for the maximum utilization (limit), i.e. `cpu.uclamp.max`.”h]”(hŒThe requested minimum utilization (protection) is always capped by the current value for the maximum utilization (limit), i.e. ”…””}”(hjhžhhŸNh NubhŒtitle_reference”“”)”}”(hŒ`cpu.uclamp.max`”h]”hŒcpu.uclamp.max”…””}”(hj#hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j!hjubhŒ.”…””}”(hjhžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M—hjìubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÚubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M™hj3ubj·)”}”(hX¯cpu.uclamp.max A read-write single value file which exists on non-root cgroups. The default is "max". i.e. no utilization capping The requested maximum utilization (limit) as a percentage rational number, e.g. 98.76 for 98.76%. This interface allows reading and setting maximum utilization clamp values similar to the sched_setattr(2). This maximum utilization value is used to clamp the task specific maximum utilization clamp. ”h]”(j½)”}”(hŒcpu.uclamp.max”h]”hŒcpu.uclamp.max”…””}”(hjKhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¤hjGubjÍ)”}”(hhh]”(hæ)”}”(hŒrA read-write single value file which exists on non-root cgroups. The default is "max". i.e. no utilization capping”h]”hŒvA read-write single value file which exists on non-root cgroups. The default is â€œmaxâ€. i.e. no utilization capping”…””}”(hj\hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MœhjYubhæ)”}”(hŒaThe requested maximum utilization (limit) as a percentage rational number, e.g. 98.76 for 98.76%.”h]”hŒaThe requested maximum utilization (limit) as a percentage rational number, e.g. 98.76 for 98.76%.”…””}”(hjjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŸhjYubhæ)”}”(hŒÈThis interface allows reading and setting maximum utilization clamp values similar to the sched_setattr(2). This maximum utilization value is used to clamp the task specific maximum utilization clamp.”h]”hŒÈThis interface allows reading and setting maximum utilization clamp values similar to the sched_setattr(2). This maximum utilization value is used to clamp the task specific maximum utilization clamp.”…””}”(hjxhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¢hjYubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjGubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¤hj3ubj·)”}”(hX‘cpu.idle A read-write single value file which exists on non-root cgroups. The default is 0. This is the cgroup analog of the per-task SCHED_IDLE sched policy. Setting this value to a 1 will make the scheduling policy of the cgroup SCHED_IDLE. The threads inside the cgroup will retain their own relative priorities, but the cgroup itself will be treated as very low priority relative to its peers. ”h]”(j½)”}”(hŒcpu.idle”h]”hŒcpu.idle”…””}”(hj–hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M°hj’ubjÍ)”}”(hhh]”(hæ)”}”(hŒRA read-write single value file which exists on non-root cgroups. The default is 0.”h]”hŒRA read-write single value file which exists on non-root cgroups. The default is 0.”…””}”(hj§hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M§hj¤ubhæ)”}”(hX1This is the cgroup analog of the per-task SCHED_IDLE sched policy. Setting this value to a 1 will make the scheduling policy of the cgroup SCHED_IDLE. The threads inside the cgroup will retain their own relative priorities, but the cgroup itself will be treated as very low priority relative to its peers.”h]”hX1This is the cgroup analog of the per-task SCHED_IDLE sched policy. Setting this value to a 1 will make the scheduling policy of the cgroup SCHED_IDLE. The threads inside the cgroup will retain their own relative priorities, but the cgroup itself will be treated as very low priority relative to its peers.”…””}”(hjµhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mªhj¤ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj’ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M°hj3ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj/ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MLhjhžhubeh}”(h]”Œcpu-interface-files”ah ]”h"]”Œcpu interface files”ah$]”h&]”uh1h°hjÇhžhhŸh¯h MHubeh}”(h]”(Œcpu”jÆeh ]”h"]”(Œcpu”Œ cgroup-v2-cpu”eh$]”h&]”uh1h°hj«hžhhŸh¯h M)jŠ }”jéj¼sjŒ }”jÆj¼subh±)”}”(hhh]”(h¶)”}”(hŒMemory”h]”hŒMemory”…””}”(hjñhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjîhžhhŸh¯h M³ubhæ)”}”(hXThe "memory" controller regulates distribution of memory. Memory is stateful and implements both limit and protection models. Due to the intertwining between memory usage and reclaim pressure and the stateful nature of memory, the distribution model is relatively complex.”h]”hXThe â€œmemoryâ€ controller regulates distribution of memory. Memory is stateful and implements both limit and protection models. Due to the intertwining between memory usage and reclaim pressure and the stateful nature of memory, the distribution model is relatively complex.”…””}”(hjÿhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mµhjîhžhubhæ)”}”(hŒòWhile not completely water-tight, all major memory usages by a given cgroup are tracked so that the total memory consumption can be accounted and controlled to a reasonable extent. Currently, the following types of memory usages are tracked.”h]”hŒòWhile not completely water-tight, all major memory usages by a given cgroup are tracked so that the total memory consumption can be accounted and controlled to a reasonable extent. Currently, the following types of memory usages are tracked.”…””}”(hj hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M»hjîhžhubjª)”}”(hhh]”(j¯)”}”(hŒ3Userland memory - page cache and anonymous memory. ”h]”hæ)”}”(hŒ2Userland memory - page cache and anonymous memory.”h]”hŒ2Userland memory - page cache and anonymous memory.”…””}”(hj"hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÀhjubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjhžhhŸh¯h Nubj¯)”}”(hŒ4Kernel data structures such as dentries and inodes. ”h]”hæ)”}”(hŒ3Kernel data structures such as dentries and inodes.”h]”hŒ3Kernel data structures such as dentries and inodes.”…””}”(hj:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÂhj6ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjhžhhŸh¯h Nubj¯)”}”(hŒTCP socket buffers. ”h]”hæ)”}”(hŒTCP socket buffers.”h]”hŒTCP socket buffers.”…””}”(hjRhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÄhjNubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjhžhhŸh¯h Nubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h MÀhjîhžhubhæ)”}”(hŒ memory.reclaim Please note that the kernel can over or under reclaim from the target cgroup. If less bytes are reclaimed than the specified amount, -EAGAIN is returned. Please note that the proactive reclaim (triggered by this interface) is not meant to indicate memory pressure on the memory cgroup. Therefore socket memory balancing triggered by the memory reclaim normally is not exercised in this case. This means that the networking layer will not adapt based on reclaim induced by memory.reclaim. ”h]”j²)”}”(hhh]”(j·)”}”(hŒ£memory.current A read-only single value file which exists on non-root cgroups. The total amount of memory currently being used by the cgroup and its descendants. ”h]”(j½)”}”(hŒmemory.current”h]”hŒmemory.current”…””}”(hj¤hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÕhj ubjÍ)”}”(hhh]”(hæ)”}”(hŒ?A read-only single value file which exists on non-root cgroups.”h]”hŒ?A read-only single value file which exists on non-root cgroups.”…””}”(hjµhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÑhj²ubhæ)”}”(hŒRThe total amount of memory currently being used by the cgroup and its descendants.”h]”hŒRThe total amount of memory currently being used by the cgroup and its descendants.”…””}”(hjÃhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÔhj²ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÕhjubj·)”}”(hXmemory.min A read-write single value file which exists on non-root cgroups. The default is "0". Hard memory protection. If the memory usage of a cgroup is within its effective min boundary, the cgroup's memory won't be reclaimed under any conditions. If there is no unprotected reclaimable memory available, OOM killer is invoked. Above the effective min boundary (or effective low boundary if it is higher), pages are reclaimed proportionally to the overage, reducing reclaim pressure for smaller overages. Effective min boundary is limited by memory.min values of all ancestor cgroups. If there is memory.min overcommitment (child cgroup or cgroups are requiring more protected memory than parent will allow), then each child cgroup will get the part of parent's protection proportional to its actual memory usage below memory.min. Putting more memory than generally available under this protection is discouraged and may lead to constant OOMs. If a memory cgroup is not populated with processes, its memory.min is ignored. ”h]”(j½)”}”(hŒ memory.min”h]”hŒ memory.min”…””}”(hjáhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MïhjÝubjÍ)”}”(hhh]”(hæ)”}”(hŒUA read-write single value file which exists on non-root cgroups. The default is "0".”h]”hŒYA read-write single value file which exists on non-root cgroups. The default is â€œ0â€.”…””}”(hjòhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MØhjïubhæ)”}”(hXœHard memory protection. If the memory usage of a cgroup is within its effective min boundary, the cgroup's memory won't be reclaimed under any conditions. If there is no unprotected reclaimable memory available, OOM killer is invoked. Above the effective min boundary (or effective low boundary if it is higher), pages are reclaimed proportionally to the overage, reducing reclaim pressure for smaller overages.”h]”hX Hard memory protection. If the memory usage of a cgroup is within its effective min boundary, the cgroupâ€™s memory wonâ€™t be reclaimed under any conditions. If there is no unprotected reclaimable memory available, OOM killer is invoked. Above the effective min boundary (or effective low boundary if it is higher), pages are reclaimed proportionally to the overage, reducing reclaim pressure for smaller overages.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÛhjïubhæ)”}”(hXEEffective min boundary is limited by memory.min values of all ancestor cgroups. If there is memory.min overcommitment (child cgroup or cgroups are requiring more protected memory than parent will allow), then each child cgroup will get the part of parent's protection proportional to its actual memory usage below memory.min.”h]”hXGEffective min boundary is limited by memory.min values of all ancestor cgroups. If there is memory.min overcommitment (child cgroup or cgroups are requiring more protected memory than parent will allow), then each child cgroup will get the part of parentâ€™s protection proportional to its actual memory usage below memory.min.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mähjïubhæ)”}”(hŒpPutting more memory than generally available under this protection is discouraged and may lead to constant OOMs.”h]”hŒpPutting more memory than generally available under this protection is discouraged and may lead to constant OOMs.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mëhjïubhæ)”}”(hŒNIf a memory cgroup is not populated with processes, its memory.min is ignored.”h]”hŒNIf a memory cgroup is not populated with processes, its memory.min is ignored.”…””}”(hj*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mîhjïubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÝubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mïhjubj·)”}”(hXƒmemory.low A read-write single value file which exists on non-root cgroups. The default is "0". Best-effort memory protection. If the memory usage of a cgroup is within its effective low boundary, the cgroup's memory won't be reclaimed unless there is no reclaimable memory available in unprotected cgroups. Above the effective low boundary (or effective min boundary if it is higher), pages are reclaimed proportionally to the overage, reducing reclaim pressure for smaller overages. Effective low boundary is limited by memory.low values of all ancestor cgroups. If there is memory.low overcommitment (child cgroup or cgroups are requiring more protected memory than parent will allow), then each child cgroup will get the part of parent's protection proportional to its actual memory usage below memory.low. Putting more memory than generally available under this protection is discouraged. ”h]”(j½)”}”(hŒ memory.low”h]”hŒ memory.low”…””}”(hjHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MhjDubjÍ)”}”(hhh]”(hæ)”}”(hŒUA read-write single value file which exists on non-root cgroups. The default is "0".”h]”hŒYA read-write single value file which exists on non-root cgroups. The default is â€œ0â€.”…””}”(hjYhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MòhjVubhæ)”}”(hX…Best-effort memory protection. If the memory usage of a cgroup is within its effective low boundary, the cgroup's memory won't be reclaimed unless there is no reclaimable memory available in unprotected cgroups. Above the effective low boundary (or effective min boundary if it is higher), pages are reclaimed proportionally to the overage, reducing reclaim pressure for smaller overages.”h]”hX‰Best-effort memory protection. If the memory usage of a cgroup is within its effective low boundary, the cgroupâ€™s memory wonâ€™t be reclaimed unless there is no reclaimable memory available in unprotected cgroups. Above the effective low boundary (or effective min boundary if it is higher), pages are reclaimed proportionally to the overage, reducing reclaim pressure for smaller overages.”…””}”(hjghžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MõhjVubhæ)”}”(hXEEffective low boundary is limited by memory.low values of all ancestor cgroups. If there is memory.low overcommitment (child cgroup or cgroups are requiring more protected memory than parent will allow), then each child cgroup will get the part of parent's protection proportional to its actual memory usage below memory.low.”h]”hXGEffective low boundary is limited by memory.low values of all ancestor cgroups. If there is memory.low overcommitment (child cgroup or cgroups are requiring more protected memory than parent will allow), then each child cgroup will get the part of parentâ€™s protection proportional to its actual memory usage below memory.low.”…””}”(hjuhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MþhjVubhæ)”}”(hŒRPutting more memory than generally available under this protection is discouraged.”h]”hŒRPutting more memory than generally available under this protection is discouraged.”…””}”(hjƒhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjVubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjDubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjubj·)”}”(hXúmemory.high A read-write single value file which exists on non-root cgroups. The default is "max". Memory usage throttle limit. If a cgroup's usage goes over the high boundary, the processes of the cgroup are throttled and put under heavy reclaim pressure. Going over the high limit never invokes the OOM killer and under extreme conditions the limit may be breached. The high limit should be used in scenarios where an external process monitors the limited cgroup to alleviate heavy reclaim pressure. ”h]”(j½)”}”(hŒmemory.high”h]”hŒmemory.high”…””}”(hj¡hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MhjubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-write single value file which exists on non-root cgroups. The default is "max".”h]”hŒ[A read-write single value file which exists on non-root cgroups. The default is â€œmaxâ€.”…””}”(hj²hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hj¯ubhæ)”}”(hŒžMemory usage throttle limit. If a cgroup's usage goes over the high boundary, the processes of the cgroup are throttled and put under heavy reclaim pressure.”h]”hŒ Memory usage throttle limit. If a cgroupâ€™s usage goes over the high boundary, the processes of the cgroup are throttled and put under heavy reclaim pressure.”…””}”(hjÀhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj¯ubhæ)”}”(hŒôGoing over the high limit never invokes the OOM killer and under extreme conditions the limit may be breached. The high limit should be used in scenarios where an external process monitors the limited cgroup to alleviate heavy reclaim pressure.”h]”hŒôGoing over the high limit never invokes the OOM killer and under extreme conditions the limit may be breached. The high limit should be used in scenarios where an external process monitors the limited cgroup to alleviate heavy reclaim pressure.”…””}”(hjÎhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj¯ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjubj·)”}”(hXœmemory.max A read-write single value file which exists on non-root cgroups. The default is "max". Memory usage hard limit. This is the main mechanism to limit memory usage of a cgroup. If a cgroup's memory usage reaches this limit and can't be reduced, the OOM killer is invoked in the cgroup. Under certain circumstances, the usage may go over the limit temporarily. In default configuration regular 0-order allocations always succeed unless OOM killer chooses current task as a victim. Some kinds of allocations don't invoke the OOM killer. Caller could retry them differently, return into userspace as -ENOMEM or silently ignore in cases like disk readahead. ”h]”(j½)”}”(hŒ memory.max”h]”hŒ memory.max”…””}”(hjìhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M%hjèubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-write single value file which exists on non-root cgroups. The default is "max".”h]”hŒ[A read-write single value file which exists on non-root cgroups. The default is â€œmaxâ€.”…””}”(hjýhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjúubhæ)”}”(hXMemory usage hard limit. This is the main mechanism to limit memory usage of a cgroup. If a cgroup's memory usage reaches this limit and can't be reduced, the OOM killer is invoked in the cgroup. Under certain circumstances, the usage may go over the limit temporarily.”h]”hXMemory usage hard limit. This is the main mechanism to limit memory usage of a cgroup. If a cgroupâ€™s memory usage reaches this limit and canâ€™t be reduced, the OOM killer is invoked in the cgroup. Under certain circumstances, the usage may go over the limit temporarily.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjúubhæ)”}”(hŒwIn default configuration regular 0-order allocations always succeed unless OOM killer chooses current task as a victim.”h]”hŒwIn default configuration regular 0-order allocations always succeed unless OOM killer chooses current task as a victim.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjúubhæ)”}”(hŒSome kinds of allocations don't invoke the OOM killer. Caller could retry them differently, return into userspace as -ENOMEM or silently ignore in cases like disk readahead.”h]”hŒ¯Some kinds of allocations donâ€™t invoke the OOM killer. Caller could retry them differently, return into userspace as -ENOMEM or silently ignore in cases like disk readahead.”…””}”(hj'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M#hjúubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjèubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M%hjubj·)”}”(hX«memory.reclaim A write-only nested-keyed file which exists for all cgroups. This is a simple interface to trigger memory reclaim in the target cgroup. Example:: echo "1G" > memory.reclaim Please note that the kernel can over or under reclaim from the target cgroup. If less bytes are reclaimed than the specified amount, -EAGAIN is returned. Please note that the proactive reclaim (triggered by this interface) is not meant to indicate memory pressure on the memory cgroup. Therefore socket memory balancing triggered by the memory reclaim normally is not exercised in this case. This means that the networking layer will not adapt based on reclaim induced by memory.reclaim. ”h]”(j½)”}”(hŒmemory.reclaim”h]”hŒmemory.reclaim”…””}”(hjEhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M:hjAubjÍ)”}”(hhh]”(hæ)”}”(hŒ memory.reclaim”h]”hŒecho "1G" > memory.reclaim”…””}”hj€sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M/hjSubhæ)”}”(hŒ™Please note that the kernel can over or under reclaim from the target cgroup. If less bytes are reclaimed than the specified amount, -EAGAIN is returned.”h]”hŒ™Please note that the kernel can over or under reclaim from the target cgroup. If less bytes are reclaimed than the specified amount, -EAGAIN is returned.”…””}”(hjŽhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M1hjSubhæ)”}”(hXMPlease note that the proactive reclaim (triggered by this interface) is not meant to indicate memory pressure on the memory cgroup. Therefore socket memory balancing triggered by the memory reclaim normally is not exercised in this case. This means that the networking layer will not adapt based on reclaim induced by memory.reclaim.”h]”hXMPlease note that the proactive reclaim (triggered by this interface) is not meant to indicate memory pressure on the memory cgroup. Therefore socket memory balancing triggered by the memory reclaim normally is not exercised in this case. This means that the networking layer will not adapt based on reclaim induced by memory.reclaim.”…””}”(hjœhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M5hjSubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjAubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M:hjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj™ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MÐhjzhžhubhæ)”}”(hŒ&The following nested keys are defined.”h]”hŒ&The following nested keys are defined.”…””}”(hjÂhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M<hjzhžhubj¬)”}”(hXF ========== ================================ swappiness Swappiness value to reclaim with ========== ================================ Specifying a swappiness value instructs the kernel to perform the reclaim with that swappiness value. Note that this has the same semantics as vm.swappiness applied to memcg reclaim with all the existing limitations and potential future extensions. memory.peak A read-write single value file which exists on non-root cgroups. The max memory usage recorded for the cgroup and its descendants since either the creation of the cgroup or the most recent reset for that FD. A write of any non-empty string to this file resets it to the current memory usage for subsequent reads through the same file descriptor. memory.oom.group A read-write single value file which exists on non-root cgroups. The default value is "0". Determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. This can be used to avoid partial kills to guarantee workload integrity. Tasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed. If the OOM killer is invoked in a cgroup, it's not going to kill any tasks outside of this cgroup, regardless memory.oom.group values of ancestor cgroups. memory.events A read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. Note that all fields in this file are hierarchical and the file modified event can be generated due to an event down the hierarchy. For the local events at the cgroup level see memory.events.local. low The number of times the cgroup is reclaimed due to high memory pressure even though its usage is under the low boundary. This usually indicates that the low boundary is over-committed. high The number of times processes of the cgroup are throttled and routed to perform direct memory reclaim because the high memory boundary was exceeded. For a cgroup whose memory usage is capped by the high limit rather than global memory pressure, this event's occurrences are expected. max The number of times the cgroup's memory usage was about to go over the max boundary. If direct reclaim fails to bring it down, the cgroup goes to OOM state. oom The number of time the cgroup's memory usage was reached the limit and allocation was about to fail. This event is not raised if the OOM killer is not considered as an option, e.g. for failed high-order allocations or if caller asked to not retry attempts. oom_kill The number of processes belonging to this cgroup killed by any kind of OOM killer. oom_group_kill The number of times a group OOM has occurred. memory.events.local Similar to memory.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events. memory.stat A read-only flat-keyed file which exists on non-root cgroups. This breaks down the cgroup's memory footprint into different types of memory, type-specific details, and other information on the state and past events of the memory management system. All memory amounts are in bytes. The entries are ordered to be human readable, and new entries can show up in the middle. Don't rely on items remaining in a fixed position; use the keys to look up specific values! If the entry has no per-node counter (or not show in the memory.numa_stat). We use 'npn' (non-per-node) as the tag to indicate that it will not show in the memory.numa_stat. anon Amount of memory used in anonymous mappings such as brk(), sbrk(), and mmap(MAP_ANONYMOUS). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not all the memory of such an allocation is mapped anymore. file Amount of memory used to cache filesystem data, including tmpfs and shared memory. kernel (npn) Amount of total kernel memory, including (kernel_stack, pagetables, percpu, vmalloc, slab) in addition to other kernel memory use cases. kernel_stack Amount of memory allocated to kernel stacks. pagetables Amount of memory allocated for page tables. sec_pagetables Amount of memory allocated for secondary page tables, this currently includes KVM mmu allocations on x86 and arm64 and IOMMU page tables. percpu (npn) Amount of memory used for storing per-cpu kernel data structures. sock (npn) Amount of memory used in network transmission buffers vmalloc (npn) Amount of memory used for vmap backed memory. shmem Amount of cached filesystem data that is swap-backed, such as tmpfs, shm segments, shared anonymous mmap()s zswap Amount of memory consumed by the zswap compression backend. zswapped Amount of application memory swapped out to zswap. file_mapped Amount of cached filesystem data mapped with mmap(). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not not all the memory of such an allocation is mapped. file_dirty Amount of cached filesystem data that was modified but not yet written back to disk file_writeback Amount of cached filesystem data that was modified and is currently being written back to disk swapcached Amount of swap cached in memory. The swapcache is accounted against both memory and swap usage. anon_thp Amount of memory used in anonymous mappings backed by transparent hugepages file_thp Amount of cached filesystem data backed by transparent hugepages shmem_thp Amount of shm, tmpfs, shared anonymous mmap()s backed by transparent hugepages inactive_anon, active_anon, inactive_file, active_file, unevictable Amount of memory, swap-backed and filesystem-backed, on the internal memory management lists used by the page reclaim algorithm. As these represent internal list state (eg. shmem pages are on anon memory management lists), inactive_foo + active_foo may not be equal to the value for the foo counter, since the foo counter is type-based, not list-based. slab_reclaimable Part of "slab" that might be reclaimed, such as dentries and inodes. slab_unreclaimable Part of "slab" that cannot be reclaimed on memory pressure. slab (npn) Amount of memory used for storing in-kernel data structures. workingset_refault_anon Number of refaults of previously evicted anonymous pages. workingset_refault_file Number of refaults of previously evicted file pages. workingset_activate_anon Number of refaulted anonymous pages that were immediately activated. workingset_activate_file Number of refaulted file pages that were immediately activated. workingset_restore_anon Number of restored anonymous pages which have been detected as an active workingset before they got reclaimed. workingset_restore_file Number of restored file pages which have been detected as an active workingset before they got reclaimed. workingset_nodereclaim Number of times a shadow node has been reclaimed pswpin (npn) Number of pages swapped into memory pswpout (npn) Number of pages swapped out of memory pgscan (npn) Amount of scanned pages (in an inactive LRU list) pgsteal (npn) Amount of reclaimed pages pgscan_kswapd (npn) Amount of scanned pages by kswapd (in an inactive LRU list) pgscan_direct (npn) Amount of scanned pages directly (in an inactive LRU list) pgscan_khugepaged (npn) Amount of scanned pages by khugepaged (in an inactive LRU list) pgscan_proactive (npn) Amount of scanned pages proactively (in an inactive LRU list) pgsteal_kswapd (npn) Amount of reclaimed pages by kswapd pgsteal_direct (npn) Amount of reclaimed pages directly pgsteal_khugepaged (npn) Amount of reclaimed pages by khugepaged pgsteal_proactive (npn) Amount of reclaimed pages proactively pgfault (npn) Total number of page faults incurred pgmajfault (npn) Number of major page faults incurred pgrefill (npn) Amount of scanned pages (in an active LRU list) pgactivate (npn) Amount of pages moved to the active LRU list pgdeactivate (npn) Amount of pages moved to the inactive LRU list pglazyfree (npn) Amount of pages postponed to be freed under memory pressure pglazyfreed (npn) Amount of reclaimed lazyfree pages swpin_zero Number of pages swapped into memory and filled with zero, where I/O was optimized out because the page content was detected to be zero during swapout. swpout_zero Number of zero-filled pages swapped out with I/O skipped due to the content being detected as zero. zswpin Number of pages moved in to memory from zswap. zswpout Number of pages moved out of memory to zswap. zswpwb Number of pages written from zswap to swap. thp_fault_alloc (npn) Number of transparent hugepages which were allocated to satisfy a page fault. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set. thp_collapse_alloc (npn) Number of transparent hugepages which were allocated to allow collapsing an existing range of pages. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set. thp_swpout (npn) Number of transparent hugepages which are swapout in one piece without splitting. thp_swpout_fallback (npn) Number of transparent hugepages which were split before swapout. Usually because failed to allocate some continuous swap space for the huge page. numa_pages_migrated (npn) Number of pages migrated by NUMA balancing. numa_pte_updates (npn) Number of pages whose page table entries are modified by NUMA balancing to produce NUMA hinting faults on access. numa_hint_faults (npn) Number of NUMA hinting faults. pgdemote_kswapd Number of pages demoted by kswapd. pgdemote_direct Number of pages demoted directly. pgdemote_khugepaged Number of pages demoted by khugepaged. pgdemote_proactive Number of pages demoted by proactively. hugetlb Amount of memory used by hugetlb pages. This metric only shows up if hugetlb usage is accounted for in memory.current (i.e. cgroup is mounted with the memory_hugetlb_accounting option). memory.numa_stat A read-only nested-keyed file which exists on non-root cgroups. This breaks down the cgroup's memory footprint into different types of memory, type-specific details, and other information per node on the state of the memory management system. This is useful for providing visibility into the NUMA locality information within an memcg since the pages are allowed to be allocated from any physical node. One of the use case is evaluating application performance by combining this information with the application's CPU allocation. All memory amounts are in bytes. The output format of memory.numa_stat is:: type N0= N1= ... The entries are ordered to be human readable, and new entries can show up in the middle. Don't rely on items remaining in a fixed position; use the keys to look up specific values! The entries can refer to the memory.stat. memory.swap.current A read-only single value file which exists on non-root cgroups. The total amount of swap currently being used by the cgroup and its descendants. memory.swap.high A read-write single value file which exists on non-root cgroups. The default is "max". Swap usage throttle limit. If a cgroup's swap usage exceeds this limit, all its further allocations will be throttled to allow userspace to implement custom out-of-memory procedures. This limit marks a point of no return for the cgroup. It is NOT designed to manage the amount of swapping a workload does during regular operation. Compare to memory.swap.max, which prohibits swapping past a set amount, but lets the cgroup continue unimpeded as long as other memory can be reclaimed. Healthy workloads are not expected to reach this limit. memory.swap.peak A read-write single value file which exists on non-root cgroups. The max swap usage recorded for the cgroup and its descendants since the creation of the cgroup or the most recent reset for that FD. A write of any non-empty string to this file resets it to the current memory usage for subsequent reads through the same file descriptor. memory.swap.max A read-write single value file which exists on non-root cgroups. The default is "max". Swap usage hard limit. If a cgroup's swap usage reaches this limit, anonymous memory of the cgroup will not be swapped out. memory.swap.events A read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. high The number of times the cgroup's swap usage was over the high threshold. max The number of times the cgroup's swap usage was about to go over the max boundary and swap allocation failed. fail The number of times swap allocation failed either because of running out of swap system-wide or max limit. When reduced under the current usage, the existing swap entries are reclaimed gradually and the swap usage may stay higher than the limit for an extended period of time. This reduces the impact on the workload and memory management. memory.zswap.current A read-only single value file which exists on non-root cgroups. The total amount of memory consumed by the zswap compression backend. memory.zswap.max A read-write single value file which exists on non-root cgroups. The default is "max". Zswap usage hard limit. If a cgroup's zswap pool reaches this limit, it will refuse to take any more stores before existing entries fault back in or are written out to disk. memory.zswap.writeback A read-write single value file. The default value is "1". Note that this setting is hierarchical, i.e. the writeback would be implicitly disabled for child cgroups if the upper hierarchy does so. When this is set to 0, all swapping attempts to swapping devices are disabled. This included both zswap writebacks, and swapping due to zswap store failures. If the zswap store failures are recurring (for e.g if the pages are incompressible), users can observe reclaim inefficiency after disabling writeback (because the same pages might be rejected again and again). Note that this is subtly different from setting memory.swap.max to 0, as it still allows for pages to be written to the zswap pool. This setting has no effect if zswap is disabled, and swapping is allowed unless memory.swap.max is set to 0. memory.pressure A read-only nested-keyed file. Shows pressure stall information for memory. See :ref:`Documentation/accounting/psi.rst ` for details. ”•h]”(j¬)”}”(hX¥ ========== ================================ swappiness Swappiness value to reclaim with ========== ================================ Specifying a swappiness value instructs the kernel to perform the reclaim with that swappiness value. Note that this has the same semantics as vm.swappiness applied to memcg reclaim with all the existing limitations and potential future extensions. ”h]”(j¬)”}”(hŒ¥========== ================================ swappiness Swappiness value to reclaim with ========== ================================ ”h]”hŒtable”“”)”}”(hhh]”hŒtgroup”“”)”}”(hhh]”(hŒcolspec”“”)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K uh1jæhjãubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K uh1jæhjãubhŒtbody”“”)”}”(hhh]”hŒrow”“”)”}”(hhh]”(hŒentry”“”)”}”(hhh]”hæ)”}”(hŒ swappiness”h]”hŒ swappiness”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M?hjubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjubj)”}”(hhh]”hæ)”}”(hŒ Swappiness value to reclaim with”h]”hŒ Swappiness value to reclaim with”…””}”(hj"hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M?hjubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjþubah}”(h]”h ]”h"]”h$]”h&]”uh1jühjãubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhjÞubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhjØubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M>hjÔubhæ)”}”(hŒøSpecifying a swappiness value instructs the kernel to perform the reclaim with that swappiness value. Note that this has the same semantics as vm.swappiness applied to memcg reclaim with all the existing limitations and potential future extensions.”h]”hŒøSpecifying a swappiness value instructs the kernel to perform the reclaim with that swappiness value. Note that this has the same semantics as vm.swappiness applied to memcg reclaim with all the existing limitations and potential future extensions.”…””}”(hjUhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MBhjÔubeh}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M>hjÐubj²)”}”(hhh]”(j·)”}”(hXhmemory.peak A read-write single value file which exists on non-root cgroups. The max memory usage recorded for the cgroup and its descendants since either the creation of the cgroup or the most recent reset for that FD. A write of any non-empty string to this file resets it to the current memory usage for subsequent reads through the same file descriptor. ”h]”(j½)”}”(hŒmemory.peak”h]”hŒmemory.peak”…””}”(hjphžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MOhjlubjÍ)”}”(hhh]”(hæ)”}”(hŒ@A read-write single value file which exists on non-root cgroups.”h]”hŒ@A read-write single value file which exists on non-root cgroups.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MHhj~ubhæ)”}”(hŒŽThe max memory usage recorded for the cgroup and its descendants since either the creation of the cgroup or the most recent reset for that FD.”h]”hŒŽThe max memory usage recorded for the cgroup and its descendants since either the creation of the cgroup or the most recent reset for that FD.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MJhj~ubhæ)”}”(hŒ‰A write of any non-empty string to this file resets it to the current memory usage for subsequent reads through the same file descriptor.”h]”hŒ‰A write of any non-empty string to this file resets it to the current memory usage for subsequent reads through the same file descriptor.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MMhj~ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjlubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MOhjiubj·)”}”(hX®memory.oom.group A read-write single value file which exists on non-root cgroups. The default value is "0". Determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. This can be used to avoid partial kills to guarantee workload integrity. Tasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed. If the OOM killer is invoked in a cgroup, it's not going to kill any tasks outside of this cgroup, regardless memory.oom.group values of ancestor cgroups. ”h]”(j½)”}”(hŒmemory.oom.group”h]”hŒmemory.oom.group”…””}”(hj»hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mahj·ubjÍ)”}”(hhh]”(hæ)”}”(hŒ[A read-write single value file which exists on non-root cgroups. The default value is "0".”h]”hŒ_A read-write single value file which exists on non-root cgroups. The default value is â€œ0â€.”…””}”(hjÌhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MRhjÉubhæ)”}”(hX5Determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. This can be used to avoid partial kills to guarantee workload integrity.”h]”hX5Determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. This can be used to avoid partial kills to guarantee workload integrity.”…””}”(hjÚhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MUhjÉubhæ)”}”(hŒlTasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed.”h]”hŒlTasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed.”…””}”(hjèhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M\hjÉubhæ)”}”(hŒšIf the OOM killer is invoked in a cgroup, it's not going to kill any tasks outside of this cgroup, regardless memory.oom.group values of ancestor cgroups.”h]”hŒœIf the OOM killer is invoked in a cgroup, itâ€™s not going to kill any tasks outside of this cgroup, regardless memory.oom.group values of ancestor cgroups.”…””}”(hjöhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M_hjÉubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj·ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mahjiubj·)”}”(hXkmemory.events A read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. Note that all fields in this file are hierarchical and the file modified event can be generated due to an event down the hierarchy. For the local events at the cgroup level see memory.events.local. low The number of times the cgroup is reclaimed due to high memory pressure even though its usage is under the low boundary. This usually indicates that the low boundary is over-committed. high The number of times processes of the cgroup are throttled and routed to perform direct memory reclaim because the high memory boundary was exceeded. For a cgroup whose memory usage is capped by the high limit rather than global memory pressure, this event's occurrences are expected. max The number of times the cgroup's memory usage was about to go over the max boundary. If direct reclaim fails to bring it down, the cgroup goes to OOM state. oom The number of time the cgroup's memory usage was reached the limit and allocation was about to fail. This event is not raised if the OOM killer is not considered as an option, e.g. for failed high-order allocations or if caller asked to not retry attempts. oom_kill The number of processes belonging to this cgroup killed by any kind of OOM killer. oom_group_kill The number of times a group OOM has occurred. ”h]”(j½)”}”(hŒ memory.events”h]”hŒ memory.events”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MŽhjubjÍ)”}”(hhh]”(hæ)”}”(hŒºA read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event.”h]”hŒºA read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event.”…””}”(hj%hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mdhj"ubhæ)”}”(hŒÅNote that all fields in this file are hierarchical and the file modified event can be generated due to an event down the hierarchy. For the local events at the cgroup level see memory.events.local.”h]”hŒÅNote that all fields in this file are hierarchical and the file modified event can be generated due to an event down the hierarchy. For the local events at the cgroup level see memory.events.local.”…””}”(hj3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mihj"ubj¬)”}”(hX¤low The number of times the cgroup is reclaimed due to high memory pressure even though its usage is under the low boundary. This usually indicates that the low boundary is over-committed. high The number of times processes of the cgroup are throttled and routed to perform direct memory reclaim because the high memory boundary was exceeded. For a cgroup whose memory usage is capped by the high limit rather than global memory pressure, this event's occurrences are expected. max The number of times the cgroup's memory usage was about to go over the max boundary. If direct reclaim fails to bring it down, the cgroup goes to OOM state. oom The number of time the cgroup's memory usage was reached the limit and allocation was about to fail. This event is not raised if the OOM killer is not considered as an option, e.g. for failed high-order allocations or if caller asked to not retry attempts. oom_kill The number of processes belonging to this cgroup killed by any kind of OOM killer. oom_group_kill The number of times a group OOM has occurred. ”h]”j²)”}”(hhh]”(j·)”}”(hŒ¾low The number of times the cgroup is reclaimed due to high memory pressure even though its usage is under the low boundary. This usually indicates that the low boundary is over-committed. ”h]”(j½)”}”(hŒlow”h]”hŒlow”…””}”(hjLhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MrhjHubjÍ)”}”(hhh]”hæ)”}”(hŒ¹The number of times the cgroup is reclaimed due to high memory pressure even though its usage is under the low boundary. This usually indicates that the low boundary is over-committed.”h]”hŒ¹The number of times the cgroup is reclaimed due to high memory pressure even though its usage is under the low boundary. This usually indicates that the low boundary is over-committed.”…””}”(hj]hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MohjZubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjHubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MrhjEubj·)”}”(hX"high The number of times processes of the cgroup are throttled and routed to perform direct memory reclaim because the high memory boundary was exceeded. For a cgroup whose memory usage is capped by the high limit rather than global memory pressure, this event's occurrences are expected. ”h]”(j½)”}”(hŒhigh”h]”hŒhigh”…””}”(hj{hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MzhjwubjÍ)”}”(hhh]”hæ)”}”(hXThe number of times processes of the cgroup are throttled and routed to perform direct memory reclaim because the high memory boundary was exceeded. For a cgroup whose memory usage is capped by the high limit rather than global memory pressure, this event's occurrences are expected.”h]”hXThe number of times processes of the cgroup are throttled and routed to perform direct memory reclaim because the high memory boundary was exceeded. For a cgroup whose memory usage is capped by the high limit rather than global memory pressure, this eventâ€™s occurrences are expected.”…””}”(hjŒhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Muhj‰ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjwubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MzhjEubj·)”}”(hŒ¢max The number of times the cgroup's memory usage was about to go over the max boundary. If direct reclaim fails to bring it down, the cgroup goes to OOM state. ”h]”(j½)”}”(hŒmax”h]”hŒmax”…””}”(hjªhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhj¦ubjÍ)”}”(hhh]”hæ)”}”(hŒThe number of times the cgroup's memory usage was about to go over the max boundary. If direct reclaim fails to bring it down, the cgroup goes to OOM state.”h]”hŒŸThe number of times the cgroupâ€™s memory usage was about to go over the max boundary. If direct reclaim fails to bring it down, the cgroup goes to OOM state.”…””}”(hj»hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M}hj¸ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj¦ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MhjEubj·)”}”(hXoom The number of time the cgroup's memory usage was reached the limit and allocation was about to fail. This event is not raised if the OOM killer is not considered as an option, e.g. for failed high-order allocations or if caller asked to not retry attempts. ”h]”(j½)”}”(hŒoom”h]”hŒoom”…””}”(hjÙhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M‡hjÕubjÍ)”}”(hhh]”(hæ)”}”(hŒdThe number of time the cgroup's memory usage was reached the limit and allocation was about to fail.”h]”hŒfThe number of time the cgroupâ€™s memory usage was reached the limit and allocation was about to fail.”…””}”(hjêhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‚hjçubhæ)”}”(hŒ›This event is not raised if the OOM killer is not considered as an option, e.g. for failed high-order allocations or if caller asked to not retry attempts.”h]”hŒ›This event is not raised if the OOM killer is not considered as an option, e.g. for failed high-order allocations or if caller asked to not retry attempts.”…””}”(hjøhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M…hjçubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÕubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M‡hjEubj·)”}”(hŒ\oom_kill The number of processes belonging to this cgroup killed by any kind of OOM killer. ”h]”(j½)”}”(hŒoom_kill”h]”hŒoom_kill”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M‹hjubjÍ)”}”(hhh]”hæ)”}”(hŒRThe number of processes belonging to this cgroup killed by any kind of OOM killer.”h]”hŒRThe number of processes belonging to this cgroup killed by any kind of OOM killer.”…””}”(hj'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŠhj$ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M‹hjEubj·)”}”(hŒ=oom_group_kill The number of times a group OOM has occurred. ”h]”(j½)”}”(hŒoom_group_kill”h]”hŒoom_group_kill”…””}”(hjEhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MŽhjAubjÍ)”}”(hhh]”hæ)”}”(hŒ-The number of times a group OOM has occurred.”h]”hŒ-The number of times a group OOM has occurred.”…””}”(hjVhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŽhjSubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjAubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MŽhjEubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjAubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mnhj"ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MŽhjiubj·)”}”(hŒÆmemory.events.local Similar to memory.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events. ”h]”(j½)”}”(hŒmemory.events.local”h]”hŒmemory.events.local”…””}”(hjŒhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M“hjˆubjÍ)”}”(hhh]”hæ)”}”(hŒ±Similar to memory.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events.”h]”hŒ±Similar to memory.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‘hjšubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjˆubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M“hjiubj·)”}”(hXbmemory.stat A read-only flat-keyed file which exists on non-root cgroups. This breaks down the cgroup's memory footprint into different types of memory, type-specific details, and other information on the state and past events of the memory management system. All memory amounts are in bytes. The entries are ordered to be human readable, and new entries can show up in the middle. Don't rely on items remaining in a fixed position; use the keys to look up specific values! If the entry has no per-node counter (or not show in the memory.numa_stat). We use 'npn' (non-per-node) as the tag to indicate that it will not show in the memory.numa_stat. anon Amount of memory used in anonymous mappings such as brk(), sbrk(), and mmap(MAP_ANONYMOUS). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not all the memory of such an allocation is mapped anymore. file Amount of memory used to cache filesystem data, including tmpfs and shared memory. kernel (npn) Amount of total kernel memory, including (kernel_stack, pagetables, percpu, vmalloc, slab) in addition to other kernel memory use cases. kernel_stack Amount of memory allocated to kernel stacks. pagetables Amount of memory allocated for page tables. sec_pagetables Amount of memory allocated for secondary page tables, this currently includes KVM mmu allocations on x86 and arm64 and IOMMU page tables. percpu (npn) Amount of memory used for storing per-cpu kernel data structures. sock (npn) Amount of memory used in network transmission buffers vmalloc (npn) Amount of memory used for vmap backed memory. shmem Amount of cached filesystem data that is swap-backed, such as tmpfs, shm segments, shared anonymous mmap()s zswap Amount of memory consumed by the zswap compression backend. zswapped Amount of application memory swapped out to zswap. file_mapped Amount of cached filesystem data mapped with mmap(). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not not all the memory of such an allocation is mapped. file_dirty Amount of cached filesystem data that was modified but not yet written back to disk file_writeback Amount of cached filesystem data that was modified and is currently being written back to disk swapcached Amount of swap cached in memory. The swapcache is accounted against both memory and swap usage. anon_thp Amount of memory used in anonymous mappings backed by transparent hugepages file_thp Amount of cached filesystem data backed by transparent hugepages shmem_thp Amount of shm, tmpfs, shared anonymous mmap()s backed by transparent hugepages inactive_anon, active_anon, inactive_file, active_file, unevictable Amount of memory, swap-backed and filesystem-backed, on the internal memory management lists used by the page reclaim algorithm. As these represent internal list state (eg. shmem pages are on anon memory management lists), inactive_foo + active_foo may not be equal to the value for the foo counter, since the foo counter is type-based, not list-based. slab_reclaimable Part of "slab" that might be reclaimed, such as dentries and inodes. slab_unreclaimable Part of "slab" that cannot be reclaimed on memory pressure. slab (npn) Amount of memory used for storing in-kernel data structures. workingset_refault_anon Number of refaults of previously evicted anonymous pages. workingset_refault_file Number of refaults of previously evicted file pages. workingset_activate_anon Number of refaulted anonymous pages that were immediately activated. workingset_activate_file Number of refaulted file pages that were immediately activated. workingset_restore_anon Number of restored anonymous pages which have been detected as an active workingset before they got reclaimed. workingset_restore_file Number of restored file pages which have been detected as an active workingset before they got reclaimed. workingset_nodereclaim Number of times a shadow node has been reclaimed pswpin (npn) Number of pages swapped into memory pswpout (npn) Number of pages swapped out of memory pgscan (npn) Amount of scanned pages (in an inactive LRU list) pgsteal (npn) Amount of reclaimed pages pgscan_kswapd (npn) Amount of scanned pages by kswapd (in an inactive LRU list) pgscan_direct (npn) Amount of scanned pages directly (in an inactive LRU list) pgscan_khugepaged (npn) Amount of scanned pages by khugepaged (in an inactive LRU list) pgscan_proactive (npn) Amount of scanned pages proactively (in an inactive LRU list) pgsteal_kswapd (npn) Amount of reclaimed pages by kswapd pgsteal_direct (npn) Amount of reclaimed pages directly pgsteal_khugepaged (npn) Amount of reclaimed pages by khugepaged pgsteal_proactive (npn) Amount of reclaimed pages proactively pgfault (npn) Total number of page faults incurred pgmajfault (npn) Number of major page faults incurred pgrefill (npn) Amount of scanned pages (in an active LRU list) pgactivate (npn) Amount of pages moved to the active LRU list pgdeactivate (npn) Amount of pages moved to the inactive LRU list pglazyfree (npn) Amount of pages postponed to be freed under memory pressure pglazyfreed (npn) Amount of reclaimed lazyfree pages swpin_zero Number of pages swapped into memory and filled with zero, where I/O was optimized out because the page content was detected to be zero during swapout. swpout_zero Number of zero-filled pages swapped out with I/O skipped due to the content being detected as zero. zswpin Number of pages moved in to memory from zswap. zswpout Number of pages moved out of memory to zswap. zswpwb Number of pages written from zswap to swap. thp_fault_alloc (npn) Number of transparent hugepages which were allocated to satisfy a page fault. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set. thp_collapse_alloc (npn) Number of transparent hugepages which were allocated to allow collapsing an existing range of pages. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set. thp_swpout (npn) Number of transparent hugepages which are swapout in one piece without splitting. thp_swpout_fallback (npn) Number of transparent hugepages which were split before swapout. Usually because failed to allocate some continuous swap space for the huge page. numa_pages_migrated (npn) Number of pages migrated by NUMA balancing. numa_pte_updates (npn) Number of pages whose page table entries are modified by NUMA balancing to produce NUMA hinting faults on access. numa_hint_faults (npn) Number of NUMA hinting faults. pgdemote_kswapd Number of pages demoted by kswapd. pgdemote_direct Number of pages demoted directly. pgdemote_khugepaged Number of pages demoted by khugepaged. pgdemote_proactive Number of pages demoted by proactively. hugetlb Amount of memory used by hugetlb pages. This metric only shows up if hugetlb usage is accounted for in memory.current (i.e. cgroup is mounted with the memory_hugetlb_accounting option). ”h]”(j½)”}”(hŒmemory.stat”h]”hŒmemory.stat”…””}”(hj»hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M˜hj·ubjÍ)”}”(hhh]”(hæ)”}”(hŒ=A read-only flat-keyed file which exists on non-root cgroups.”h]”hŒ=A read-only flat-keyed file which exists on non-root cgroups.”…””}”(hjÌhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M–hjÉubhæ)”}”(hŒ¹This breaks down the cgroup's memory footprint into different types of memory, type-specific details, and other information on the state and past events of the memory management system.”h]”hŒ»This breaks down the cgroupâ€™s memory footprint into different types of memory, type-specific details, and other information on the state and past events of the memory management system.”…””}”(hjÚhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M˜hjÉubhæ)”}”(hŒ All memory amounts are in bytes.”h]”hŒ All memory amounts are in bytes.”…””}”(hjèhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MœhjÉubhæ)”}”(hŒ´The entries are ordered to be human readable, and new entries can show up in the middle. Don't rely on items remaining in a fixed position; use the keys to look up specific values!”h]”hŒ¶The entries are ordered to be human readable, and new entries can show up in the middle. Donâ€™t rely on items remaining in a fixed position; use the keys to look up specific values!”…””}”(hjöhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MžhjÉubhæ)”}”(hŒIf the entry has no per-node counter (or not show in the memory.numa_stat). We use 'npn' (non-per-node) as the tag to indicate that it will not show in the memory.numa_stat.”h]”hŒ±If the entry has no per-node counter (or not show in the memory.numa_stat). We use â€˜npnâ€™ (non-per-node) as the tag to indicate that it will not show in the memory.numa_stat.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¢hjÉubj¬)”}”(hXsanon Amount of memory used in anonymous mappings such as brk(), sbrk(), and mmap(MAP_ANONYMOUS). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not all the memory of such an allocation is mapped anymore. file Amount of memory used to cache filesystem data, including tmpfs and shared memory. kernel (npn) Amount of total kernel memory, including (kernel_stack, pagetables, percpu, vmalloc, slab) in addition to other kernel memory use cases. kernel_stack Amount of memory allocated to kernel stacks. pagetables Amount of memory allocated for page tables. sec_pagetables Amount of memory allocated for secondary page tables, this currently includes KVM mmu allocations on x86 and arm64 and IOMMU page tables. percpu (npn) Amount of memory used for storing per-cpu kernel data structures. sock (npn) Amount of memory used in network transmission buffers vmalloc (npn) Amount of memory used for vmap backed memory. shmem Amount of cached filesystem data that is swap-backed, such as tmpfs, shm segments, shared anonymous mmap()s zswap Amount of memory consumed by the zswap compression backend. zswapped Amount of application memory swapped out to zswap. file_mapped Amount of cached filesystem data mapped with mmap(). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not not all the memory of such an allocation is mapped. file_dirty Amount of cached filesystem data that was modified but not yet written back to disk file_writeback Amount of cached filesystem data that was modified and is currently being written back to disk swapcached Amount of swap cached in memory. The swapcache is accounted against both memory and swap usage. anon_thp Amount of memory used in anonymous mappings backed by transparent hugepages file_thp Amount of cached filesystem data backed by transparent hugepages shmem_thp Amount of shm, tmpfs, shared anonymous mmap()s backed by transparent hugepages inactive_anon, active_anon, inactive_file, active_file, unevictable Amount of memory, swap-backed and filesystem-backed, on the internal memory management lists used by the page reclaim algorithm. As these represent internal list state (eg. shmem pages are on anon memory management lists), inactive_foo + active_foo may not be equal to the value for the foo counter, since the foo counter is type-based, not list-based. slab_reclaimable Part of "slab" that might be reclaimed, such as dentries and inodes. slab_unreclaimable Part of "slab" that cannot be reclaimed on memory pressure. slab (npn) Amount of memory used for storing in-kernel data structures. workingset_refault_anon Number of refaults of previously evicted anonymous pages. workingset_refault_file Number of refaults of previously evicted file pages. workingset_activate_anon Number of refaulted anonymous pages that were immediately activated. workingset_activate_file Number of refaulted file pages that were immediately activated. workingset_restore_anon Number of restored anonymous pages which have been detected as an active workingset before they got reclaimed. workingset_restore_file Number of restored file pages which have been detected as an active workingset before they got reclaimed. workingset_nodereclaim Number of times a shadow node has been reclaimed pswpin (npn) Number of pages swapped into memory pswpout (npn) Number of pages swapped out of memory pgscan (npn) Amount of scanned pages (in an inactive LRU list) pgsteal (npn) Amount of reclaimed pages pgscan_kswapd (npn) Amount of scanned pages by kswapd (in an inactive LRU list) pgscan_direct (npn) Amount of scanned pages directly (in an inactive LRU list) pgscan_khugepaged (npn) Amount of scanned pages by khugepaged (in an inactive LRU list) pgscan_proactive (npn) Amount of scanned pages proactively (in an inactive LRU list) pgsteal_kswapd (npn) Amount of reclaimed pages by kswapd pgsteal_direct (npn) Amount of reclaimed pages directly pgsteal_khugepaged (npn) Amount of reclaimed pages by khugepaged pgsteal_proactive (npn) Amount of reclaimed pages proactively pgfault (npn) Total number of page faults incurred pgmajfault (npn) Number of major page faults incurred pgrefill (npn) Amount of scanned pages (in an active LRU list) pgactivate (npn) Amount of pages moved to the active LRU list pgdeactivate (npn) Amount of pages moved to the inactive LRU list pglazyfree (npn) Amount of pages postponed to be freed under memory pressure pglazyfreed (npn) Amount of reclaimed lazyfree pages swpin_zero Number of pages swapped into memory and filled with zero, where I/O was optimized out because the page content was detected to be zero during swapout. swpout_zero Number of zero-filled pages swapped out with I/O skipped due to the content being detected as zero. zswpin Number of pages moved in to memory from zswap. zswpout Number of pages moved out of memory to zswap. zswpwb Number of pages written from zswap to swap. thp_fault_alloc (npn) Number of transparent hugepages which were allocated to satisfy a page fault. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set. thp_collapse_alloc (npn) Number of transparent hugepages which were allocated to allow collapsing an existing range of pages. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set. thp_swpout (npn) Number of transparent hugepages which are swapout in one piece without splitting. thp_swpout_fallback (npn) Number of transparent hugepages which were split before swapout. Usually because failed to allocate some continuous swap space for the huge page. numa_pages_migrated (npn) Number of pages migrated by NUMA balancing. numa_pte_updates (npn) Number of pages whose page table entries are modified by NUMA balancing to produce NUMA hinting faults on access. numa_hint_faults (npn) Number of NUMA hinting faults. pgdemote_kswapd Number of pages demoted by kswapd. pgdemote_direct Number of pages demoted directly. pgdemote_khugepaged Number of pages demoted by khugepaged. pgdemote_proactive Number of pages demoted by proactively. hugetlb Amount of memory used by hugetlb pages. This metric only shows up if hugetlb usage is accounted for in memory.current (i.e. cgroup is mounted with the memory_hugetlb_accounting option). ”h]”j²)”}”(hhh]”(j·)”}”(hX anon Amount of memory used in anonymous mappings such as brk(), sbrk(), and mmap(MAP_ANONYMOUS). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not all the memory of such an allocation is mapped anymore. ”h]”(j½)”}”(hŒanon”h]”hŒanon”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M«hjubjÍ)”}”(hhh]”hæ)”}”(hXAmount of memory used in anonymous mappings such as brk(), sbrk(), and mmap(MAP_ANONYMOUS). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not all the memory of such an allocation is mapped anymore.”h]”hXAmount of memory used in anonymous mappings such as brk(), sbrk(), and mmap(MAP_ANONYMOUS). Note that some kernel configurations might account complete larger allocations (e.g., THP) if only some, but not all the memory of such an allocation is mapped anymore.”…””}”(hj.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M§hj+ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M«hjubj·)”}”(hŒXfile Amount of memory used to cache filesystem data, including tmpfs and shared memory. ”h]”(j½)”}”(hŒfile”h]”hŒfile”…””}”(hjLhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¯hjHubjÍ)”}”(hhh]”hæ)”}”(hŒRAmount of memory used to cache filesystem data, including tmpfs and shared memory.”h]”hŒRAmount of memory used to cache filesystem data, including tmpfs and shared memory.”…””}”(hj]hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M®hjZubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjHubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¯hjubj·)”}”(hŒ–kernel (npn) Amount of total kernel memory, including (kernel_stack, pagetables, percpu, vmalloc, slab) in addition to other kernel memory use cases. ”h]”(j½)”}”(hŒkernel (npn)”h]”hŒkernel (npn)”…””}”(hj{hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M´hjwubjÍ)”}”(hhh]”hæ)”}”(hŒˆAmount of total kernel memory, including (kernel_stack, pagetables, percpu, vmalloc, slab) in addition to other kernel memory use cases.”h]”hŒˆAmount of total kernel memory, including (kernel_stack, pagetables, percpu, vmalloc, slab) in addition to other kernel memory use cases.”…””}”(hjŒhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M²hj‰ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjwubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M´hjubj·)”}”(hŒ:kernel_stack Amount of memory allocated to kernel stacks. ”h]”(j½)”}”(hŒkernel_stack”h]”hŒkernel_stack”…””}”(hjªhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M·hj¦ubjÍ)”}”(hhh]”hæ)”}”(hŒ,Amount of memory allocated to kernel stacks.”h]”hŒ,Amount of memory allocated to kernel stacks.”…””}”(hj»hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M·hj¸ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj¦ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M·hjubj·)”}”(hŒ7pagetables Amount of memory allocated for page tables. ”h]”(j½)”}”(hŒ pagetables”h]”hŒ pagetables”…””}”(hjÙhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MºhjÕubjÍ)”}”(hhh]”hæ)”}”(hŒ+Amount of memory allocated for page tables.”h]”hŒ+Amount of memory allocated for page tables.”…””}”(hjêhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mºhjçubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÕubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mºhjubj·)”}”(hŒ™sec_pagetables Amount of memory allocated for secondary page tables, this currently includes KVM mmu allocations on x86 and arm64 and IOMMU page tables. ”h]”(j½)”}”(hŒsec_pagetables”h]”hŒsec_pagetables”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¿hjubjÍ)”}”(hhh]”hæ)”}”(hŒ‰Amount of memory allocated for secondary page tables, this currently includes KVM mmu allocations on x86 and arm64 and IOMMU page tables.”h]”hŒ‰Amount of memory allocated for secondary page tables, this currently includes KVM mmu allocations on x86 and arm64 and IOMMU page tables.”…””}”(hjhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M½hjubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¿hjubj·)”}”(hŒOpercpu (npn) Amount of memory used for storing per-cpu kernel data structures. ”h]”(j½)”}”(hŒpercpu (npn)”h]”hŒpercpu (npn)”…””}”(hj7hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÃhj3ubjÍ)”}”(hhh]”hæ)”}”(hŒAAmount of memory used for storing per-cpu kernel data structures.”h]”hŒAAmount of memory used for storing per-cpu kernel data structures.”…””}”(hjHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÂhjEubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj3ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÃhjubj·)”}”(hŒAsock (npn) Amount of memory used in network transmission buffers ”h]”(j½)”}”(hŒ sock (npn)”h]”hŒ sock (npn)”…””}”(hjfhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÆhjbubjÍ)”}”(hhh]”hæ)”}”(hŒ5Amount of memory used in network transmission buffers”h]”hŒ5Amount of memory used in network transmission buffers”…””}”(hjwhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÆhjtubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjbubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÆhjubj·)”}”(hŒpgsteal_proactive (npn) Amount of reclaimed pages proactively ”h]”(j½)”}”(hŒpgsteal_proactive (npn)”h]”hŒpgsteal_proactive (npn)”…””}”(hj²%hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MChj®%ubjÍ)”}”(hhh]”hæ)”}”(hŒ%Amount of reclaimed pages proactively”h]”hŒ%Amount of reclaimed pages proactively”…””}”(hjÃ%hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MChjÀ%ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj®%ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MChjubj·)”}”(hŒ3pgfault (npn) Total number of page faults incurred ”h]”(j½)”}”(hŒ pgfault (npn)”h]”hŒ pgfault (npn)”…””}”(hjá%hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MFhjÝ%ubjÍ)”}”(hhh]”hæ)”}”(hŒ$Total number of page faults incurred”h]”hŒ$Total number of page faults incurred”…””}”(hjò%hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MFhjï%ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÝ%ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MFhjubj·)”}”(hŒ6pgmajfault (npn) Number of major page faults incurred ”h]”(j½)”}”(hŒpgmajfault (npn)”h]”hŒpgmajfault (npn)”…””}”(hj&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MIhj&ubjÍ)”}”(hhh]”hæ)”}”(hŒ$Number of major page faults incurred”h]”hŒ$Number of major page faults incurred”…””}”(hj!&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MIhj&ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj&ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MIhjubj·)”}”(hŒ?pgrefill (npn) Amount of scanned pages (in an active LRU list) ”h]”(j½)”}”(hŒpgrefill (npn)”h]”hŒpgrefill (npn)”…””}”(hj?&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MLhj;&ubjÍ)”}”(hhh]”hæ)”}”(hŒ/Amount of scanned pages (in an active LRU list)”h]”hŒ/Amount of scanned pages (in an active LRU list)”…””}”(hjP&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MLhjM&ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj;&ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MLhjubj·)”}”(hŒ>pgactivate (npn) Amount of pages moved to the active LRU list ”h]”(j½)”}”(hŒpgactivate (npn)”h]”hŒpgactivate (npn)”…””}”(hjn&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MOhjj&ubjÍ)”}”(hhh]”hæ)”}”(hŒ,Amount of pages moved to the active LRU list”h]”hŒ,Amount of pages moved to the active LRU list”…””}”(hj&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MOhj|&ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjj&ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MOhjubj·)”}”(hŒBpgdeactivate (npn) Amount of pages moved to the inactive LRU list ”h]”(j½)”}”(hŒpgdeactivate (npn)”h]”hŒpgdeactivate (npn)”…””}”(hj&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MRhj™&ubjÍ)”}”(hhh]”hæ)”}”(hŒ.Amount of pages moved to the inactive LRU list”h]”hŒ.Amount of pages moved to the inactive LRU list”…””}”(hj®&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MRhj«&ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj™&ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MRhjubj·)”}”(hŒMpglazyfree (npn) Amount of pages postponed to be freed under memory pressure ”h]”(j½)”}”(hŒpglazyfree (npn)”h]”hŒpglazyfree (npn)”…””}”(hjÌ&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MUhjÈ&ubjÍ)”}”(hhh]”hæ)”}”(hŒ;Amount of pages postponed to be freed under memory pressure”h]”hŒ;Amount of pages postponed to be freed under memory pressure”…””}”(hjÝ&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MUhjÚ&ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÈ&ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MUhjubj·)”}”(hŒ5pglazyfreed (npn) Amount of reclaimed lazyfree pages ”h]”(j½)”}”(hŒpglazyfreed (npn)”h]”hŒpglazyfreed (npn)”…””}”(hjû&hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MXhj÷&ubjÍ)”}”(hhh]”hæ)”}”(hŒ"Amount of reclaimed lazyfree pages”h]”hŒ"Amount of reclaimed lazyfree pages”…””}”(hj'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MXhj 'ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj÷&ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MXhjubj·)”}”(hŒ¢swpin_zero Number of pages swapped into memory and filled with zero, where I/O was optimized out because the page content was detected to be zero during swapout. ”h]”(j½)”}”(hŒ swpin_zero”h]”hŒ swpin_zero”…””}”(hj*'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M]hj&'ubjÍ)”}”(hhh]”hæ)”}”(hŒ–Number of pages swapped into memory and filled with zero, where I/O was optimized out because the page content was detected to be zero during swapout.”h]”hŒ–Number of pages swapped into memory and filled with zero, where I/O was optimized out because the page content was detected to be zero during swapout.”…””}”(hj;'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M[hj8'ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj&'ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M]hjubj·)”}”(hŒpswpout_zero Number of zero-filled pages swapped out with I/O skipped due to the content being detected as zero. ”h]”(j½)”}”(hŒswpout_zero”h]”hŒswpout_zero”…””}”(hjY'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MahjU'ubjÍ)”}”(hhh]”hæ)”}”(hŒcNumber of zero-filled pages swapped out with I/O skipped due to the content being detected as zero.”h]”hŒcNumber of zero-filled pages swapped out with I/O skipped due to the content being detected as zero.”…””}”(hjj'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M`hjg'ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjU'ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mahjubj·)”}”(hŒ6zswpin Number of pages moved in to memory from zswap. ”h]”(j½)”}”(hŒzswpin”h]”hŒzswpin”…””}”(hjˆ'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸ•£h¯h Mdhj„'ubjÍ)”}”(hhh]”hæ)”}”(hŒ.Number of pages moved in to memory from zswap.”h]”hŒ.Number of pages moved in to memory from zswap.”…””}”(hj™'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mdhj–'ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj„'ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mdhjubj·)”}”(hŒ6zswpout Number of pages moved out of memory to zswap. ”h]”(j½)”}”(hŒzswpout”h]”hŒzswpout”…””}”(hj·'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mghj³'ubjÍ)”}”(hhh]”hæ)”}”(hŒ-Number of pages moved out of memory to zswap.”h]”hŒ-Number of pages moved out of memory to zswap.”…””}”(hjÈ'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MghjÅ'ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj³'ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mghjubj·)”}”(hŒ3zswpwb Number of pages written from zswap to swap. ”h]”(j½)”}”(hŒzswpwb”h]”hŒzswpwb”…””}”(hjæ'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mjhjâ'ubjÍ)”}”(hhh]”hæ)”}”(hŒ+Number of pages written from zswap to swap.”h]”hŒ+Number of pages written from zswap to swap.”…””}”(hj÷'hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mjhjô'ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjâ'ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mjhjubj·)”}”(hŒthp_fault_alloc (npn) Number of transparent hugepages which were allocated to satisfy a page fault. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set. ”h]”(j½)”}”(hŒthp_fault_alloc (npn)”h]”hŒthp_fault_alloc (npn)”…””}”(hj(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mohj(ubjÍ)”}”(hhh]”hæ)”}”(hŒ–Number of transparent hugepages which were allocated to satisfy a page fault. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set.”h]”hŒ–Number of transparent hugepages which were allocated to satisfy a page fault. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set.”…””}”(hj&(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mmhj#(ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj(ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mohjubj·)”}”(hŒÇthp_collapse_alloc (npn) Number of transparent hugepages which were allocated to allow collapsing an existing range of pages. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set. ”h]”(j½)”}”(hŒthp_collapse_alloc (npn)”h]”hŒthp_collapse_alloc (npn)”…””}”(hjD(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mthj@(ubjÍ)”}”(hhh]”hæ)”}”(hŒNumber of transparent hugepages which were allocated to allow collapsing an existing range of pages. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set.”h]”hŒNumber of transparent hugepages which were allocated to allow collapsing an existing range of pages. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE is not set.”…””}”(hjU(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MrhjR(ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj@(ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mthjubj·)”}”(hŒcthp_swpout (npn) Number of transparent hugepages which are swapout in one piece without splitting. ”h]”(j½)”}”(hŒthp_swpout (npn)”h]”hŒthp_swpout (npn)”…””}”(hjs(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mxhjo(ubjÍ)”}”(hhh]”hæ)”}”(hŒQNumber of transparent hugepages which are swapout in one piece without splitting.”h]”hŒQNumber of transparent hugepages which are swapout in one piece without splitting.”…””}”(hj„(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mwhj(ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjo(ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mxhjubj·)”}”(hŒ¬thp_swpout_fallback (npn) Number of transparent hugepages which were split before swapout. Usually because failed to allocate some continuous swap space for the huge page. ”h]”(j½)”}”(hŒthp_swpout_fallback (npn)”h]”hŒthp_swpout_fallback (npn)”…””}”(hj¢(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M}hjž(ubjÍ)”}”(hhh]”hæ)”}”(hŒ‘Number of transparent hugepages which were split before swapout. Usually because failed to allocate some continuous swap space for the huge page.”h]”hŒ‘Number of transparent hugepages which were split before swapout. Usually because failed to allocate some continuous swap space for the huge page.”…””}”(hj³(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M{hj°(ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjž(ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M}hjubj·)”}”(hŒFnuma_pages_migrated (npn) Number of pages migrated by NUMA balancing. ”h]”(j½)”}”(hŒnuma_pages_migrated (npn)”h]”hŒnuma_pages_migrated (npn)”…””}”(hjÑ(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M€hjÍ(ubjÍ)”}”(hhh]”hæ)”}”(hŒ+Number of pages migrated by NUMA balancing.”h]”hŒ+Number of pages migrated by NUMA balancing.”…””}”(hjâ(hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M€hjß(ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÍ(ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M€hjubj·)”}”(hŒ‰numa_pte_updates (npn) Number of pages whose page table entries are modified by NUMA balancing to produce NUMA hinting faults on access. ”h]”(j½)”}”(hŒnuma_pte_updates (npn)”h]”hŒnuma_pte_updates (npn)”…””}”(hj)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M„hjü(ubjÍ)”}”(hhh]”hæ)”}”(hŒqNumber of pages whose page table entries are modified by NUMA balancing to produce NUMA hinting faults on access.”h]”hŒqNumber of pages whose page table entries are modified by NUMA balancing to produce NUMA hinting faults on access.”…””}”(hj)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mƒhj)ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjü(ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M„hjubj·)”}”(hŒ6numa_hint_faults (npn) Number of NUMA hinting faults. ”h]”(j½)”}”(hŒnuma_hint_faults (npn)”h]”hŒnuma_hint_faults (npn)”…””}”(hj/)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M‡hj+)ubjÍ)”}”(hhh]”hæ)”}”(hŒNumber of NUMA hinting faults.”h]”hŒNumber of NUMA hinting faults.”…””}”(hj@)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‡hj=)ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj+)ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M‡hjubj·)”}”(hŒ3pgdemote_kswapd Number of pages demoted by kswapd. ”h]”(j½)”}”(hŒpgdemote_kswapd”h]”hŒpgdemote_kswapd”…””}”(hj^)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MŠhjZ)ubjÍ)”}”(hhh]”hæ)”}”(hŒ"Number of pages demoted by kswapd.”h]”hŒ"Number of pages demoted by kswapd.”…””}”(hjo)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŠhjl)ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjZ)ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MŠhjubj·)”}”(hŒ2pgdemote_direct Number of pages demoted directly. ”h]”(j½)”}”(hŒpgdemote_direct”h]”hŒpgdemote_direct”…””}”(hj)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhj‰)ubjÍ)”}”(hhh]”hæ)”}”(hŒ!Number of pages demoted directly.”h]”hŒ!Number of pages demoted directly.”…””}”(hjž)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj›)ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj‰)ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjubj·)”}”(hŒ;pgdemote_khugepaged Number of pages demoted by khugepaged. ”h]”(j½)”}”(hŒpgdemote_khugepaged”h]”hŒpgdemote_khugepaged”…””}”(hj¼)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhj¸)ubjÍ)”}”(hhh]”hæ)”}”(hŒ&Number of pages demoted by khugepaged.”h]”hŒ&Number of pages demoted by khugepaged.”…””}”(hjÍ)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÊ)ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj¸)ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjubj·)”}”(hŒ;pgdemote_proactive Number of pages demoted by proactively. ”h]”(j½)”}”(hŒpgdemote_proactive”h]”hŒpgdemote_proactive”…””}”(hjë)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M“hjç)ubjÍ)”}”(hhh]”hæ)”}”(hŒ'Number of pages demoted by proactively.”h]”hŒ'Number of pages demoted by proactively.”…””}”(hjü)hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M“hjù)ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjç)ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M“hjubj·)”}”(hŒÂhugetlb Amount of memory used by hugetlb pages. This metric only shows up if hugetlb usage is accounted for in memory.current (i.e. cgroup is mounted with the memory_hugetlb_accounting option). ”h]”(j½)”}”(hŒhugetlb”h]”hŒhugetlb”…””}”(hj*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M˜hj*ubjÍ)”}”(hhh]”hæ)”}”(hŒ¹Amount of memory used by hugetlb pages. This metric only shows up if hugetlb usage is accounted for in memory.current (i.e. cgroup is mounted with the memory_hugetlb_accounting option).”h]”hŒ¹Amount of memory used by hugetlb pages. This metric only shows up if hugetlb usage is accounted for in memory.current (i.e. cgroup is mounted with the memory_hugetlb_accounting option).”…””}”(hj+*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M–hj(*ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj*ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M˜hjubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M¦hjÉubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj·ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M˜hjiubj·)”}”(hX‰memory.numa_stat A read-only nested-keyed file which exists on non-root cgroups. This breaks down the cgroup's memory footprint into different types of memory, type-specific details, and other information per node on the state of the memory management system. This is useful for providing visibility into the NUMA locality information within an memcg since the pages are allowed to be allocated from any physical node. One of the use case is evaluating application performance by combining this information with the application's CPU allocation. All memory amounts are in bytes. The output format of memory.numa_stat is:: type N0= N1= ... The entries are ordered to be human readable, and new entries can show up in the middle. Don't rely on items remaining in a fixed position; use the keys to look up specific values! The entries can refer to the memory.stat. ”h]”(j½)”}”(hŒmemory.numa_stat”h]”hŒmemory.numa_stat”…””}”(hja*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M±hj]*ubjÍ)”}”(hhh]”(hæ)”}”(hŒ?A read-only nested-keyed file which exists on non-root cgroups.”h]”hŒ?A read-only nested-keyed file which exists on non-root cgroups.”…””}”(hjr*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M›hjo*ubhæ)”}”(hŒ²This breaks down the cgroup's memory footprint into different types of memory, type-specific details, and other information per node on the state of the memory management system.”h]”hŒ´This breaks down the cgroupâ€™s memory footprint into different types of memory, type-specific details, and other information per node on the state of the memory management system.”…””}”(hj€*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjo*ubhæ)”}”(hXThis is useful for providing visibility into the NUMA locality information within an memcg since the pages are allowed to be allocated from any physical node. One of the use case is evaluating application performance by combining this information with the application's CPU allocation.”h]”hXThis is useful for providing visibility into the NUMA locality information within an memcg since the pages are allowed to be allocated from any physical node. One of the use case is evaluating application performance by combining this information with the applicationâ€™s CPU allocation.”…””}”(hjŽ*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¡hjo*ubhæ)”}”(hŒ All memory amounts are in bytes.”h]”hŒ All memory amounts are in bytes.”…””}”(hjœ*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M§hjo*ubhæ)”}”(hŒ*The output format of memory.numa_stat is::”h]”hŒ)The output format of memory.numa_stat is:”…””}”(hjª*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M©hjo*ubjV)”}”(hŒ2type N0= N1= ...”h]”hŒ2type N0= N1= ...”…””}”hj¸*sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M«hjo*ubhæ)”}”(hŒ´The entries are ordered to be human readable, and new entries can show up in the middle. Don't rely on items remaining in a fixed position; use the keys to look up specific values!”h]”hŒ¶The entries are ordered to be human readable, and new entries can show up in the middle. Donâ€™t rely on items remaining in a fixed position; use the keys to look up specific values!”…””}”(hjÆ*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjo*ubhæ)”}”(hŒ)The entries can refer to the memory.stat.”h]”hŒ)The entries can refer to the memory.stat.”…””}”(hjÔ*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M±hjo*ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj]*ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M±hjiubj·)”}”(hŒ¦memory.swap.current A read-only single value file which exists on non-root cgroups. The total amount of swap currently being used by the cgroup and its descendants. ”h]”(j½)”}”(hŒmemory.swap.current”h]”hŒmemory.swap.current”…””}”(hjò*hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¸hjî*ubjÍ)”}”(hhh]”(hæ)”}”(hŒ?A read-only single value file which exists on non-root cgroups.”h]”hŒ?A read-only single value file which exists on non-root cgroups.”…””}”(hj+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M´hj+ubhæ)”}”(hŒPThe total amount of swap currently being used by the cgroup and its descendants.”h]”hŒPThe total amount of swap currently being used by the cgroup and its descendants.”…””}”(hj+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M·hj+ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjî*ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¸hjiubj·)”}”(hX‰memory.swap.high A read-write single value file which exists on non-root cgroups. The default is "max". Swap usage throttle limit. If a cgroup's swap usage exceeds this limit, all its further allocations will be throttled to allow userspace to implement custom out-of-memory procedures. This limit marks a point of no return for the cgroup. It is NOT designed to manage the amount of swapping a workload does during regular operation. Compare to memory.swap.max, which prohibits swapping past a set amount, but lets the cgroup continue unimpeded as long as other memory can be reclaimed. Healthy workloads are not expected to reach this limit. ”h]”(j½)”}”(hŒmemory.swap.high”h]”hŒmemory.swap.high”…””}”(hj/+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÈhj++ubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-write single value file which exists on non-root cgroups. The default is "max".”h]”hŒ[A read-write single value file which exists on non-root cgroups. The default is â€œmaxâ€.”…””}”(hj@+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M»hj=+ubhæ)”}”(hŒ·Swap usage throttle limit. If a cgroup's swap usage exceeds this limit, all its further allocations will be throttled to allow userspace to implement custom out-of-memory procedures.”h]”hŒ¹Swap usage throttle limit. If a cgroupâ€™s swap usage exceeds this limit, all its further allocations will be throttled to allow userspace to implement custom out-of-memory procedures.”…””}”(hjN+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¾hj=+ubhæ)”}”(hX,This limit marks a point of no return for the cgroup. It is NOT designed to manage the amount of swapping a workload does during regular operation. Compare to memory.swap.max, which prohibits swapping past a set amount, but lets the cgroup continue unimpeded as long as other memory can be reclaimed.”h]”hX,This limit marks a point of no return for the cgroup. It is NOT designed to manage the amount of swapping a workload does during regular operation. Compare to memory.swap.max, which prohibits swapping past a set amount, but lets the cgroup continue unimpeded as long as other memory can be reclaimed.”…””}”(hj\+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÂhj=+ubhæ)”}”(hŒ7Healthy workloads are not expected to reach this limit.”h]”hŒ7Healthy workloads are not expected to reach this limit.”…””}”(hjj+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÈhj=+ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj++ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÈhjiubj·)”}”(hXdmemory.swap.peak A read-write single value file which exists on non-root cgroups. The max swap usage recorded for the cgroup and its descendants since the creation of the cgroup or the most recent reset for that FD. A write of any non-empty string to this file resets it to the current memory usage for subsequent reads through the same file descriptor. ”h]”(j½)”}”(hŒmemory.swap.peak”h]”hŒmemory.swap.peak”…””}”(hjˆ+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÒhj„+ubjÍ)”}”(hhh]”(hæ)”}”(hŒ@A read-write single value file which exists on non-root cgroups.”h]”hŒ@A read-write single value file which exists on non-root cgroups.”…””}”(hj™+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MËhj–+ubhæ)”}”(hŒ…The max swap usage recorded for the cgroup and its descendants since the creation of the cgroup or the most recent reset for that FD.”h]”hŒ…The max swap usage recorded for the cgroup and its descendants since the creation of the cgroup or the most recent reset for that FD.”…””}”(hj§+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÍhj–+ubhæ)”}”(hŒ‰A write of any non-empty string to this file resets it to the current memory usage for subsequent reads through the same file descriptor.”h]”hŒ‰A write of any non-empty string to this file resets it to the current memory usage for subsequent reads through the same file descriptor.”…””}”(hjµ+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÐhj–+ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj„+ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÒhjiubj·)”}”(hŒæmemory.swap.max A read-write single value file which exists on non-root cgroups. The default is "max". Swap usage hard limit. If a cgroup's swap usage reaches this limit, anonymous memory of the cgroup will not be swapped out. ”h]”(j½)”}”(hŒmemory.swap.max”h]”hŒmemory.swap.max”…””}”(hjÓ+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÙhjÏ+ubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-write single value file which exists on non-root cgroups. The default is "max".”h]”hŒ[A read-write single value file which exists on non-root cgroups. The default is â€œmaxâ€.”…””}”(hjä+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÕhjá+ubhæ)”}”(hŒ|Swap usage hard limit. If a cgroup's swap usage reaches this limit, anonymous memory of the cgroup will not be swapped out.”h]”hŒ~Swap usage hard limit. If a cgroupâ€™s swap usage reaches this limit, anonymous memory of the cgroup will not be swapped out.”…””}”(hjò+hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MØhjá+ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÏ+ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÙhjiubj·)”}”(hX2memory.swap.events A read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. high The number of times the cgroup's swap usage was over the high threshold. max The number of times the cgroup's swap usage was about to go over the max boundary and swap allocation failed. fail The number of times swap allocation failed either because of running out of swap system-wide or max limit. When reduced under the current usage, the existing swap entries are reclaimed gradually and the swap usage may stay higher than the limit for an extended period of time. This reduces the impact on the workload and memory management. ”h]”(j½)”}”(hŒmemory.swap.events”h]”hŒmemory.swap.events”…””}”(hj,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mòhj,ubjÍ)”}”(hhh]”(hæ)”}”(hŒºA read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event.”h]”hŒºA read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event.”…””}”(hj!,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÜhj,ubj¬)”}”(hXbhigh The number of times the cgroup's swap usage was over the high threshold. max The number of times the cgroup's swap usage was about to go over the max boundary and swap allocation failed. fail The number of times swap allocation failed either because of running out of swap system-wide or max limit. ”h]”j²)”}”(hhh]”(j·)”}”(hŒNhigh The number of times the cgroup's swap usage was over the high threshold. ”h]”(j½)”}”(hŒhigh”h]”hŒhigh”…””}”(hj:,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mãhj6,ubjÍ)”}”(hhh]”hæ)”}”(hŒHThe number of times the cgroup's swap usage was over the high threshold.”h]”hŒJThe number of times the cgroupâ€™s swap usage was over the high threshold.”…””}”(hjK,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MâhjH,ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj6,ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mãhj3,ubj·)”}”(hŒrmax The number of times the cgroup's swap usage was about to go over the max boundary and swap allocation failed. ”h]”(j½)”}”(hŒmax”h]”hŒmax”…””}”(hji,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mèhje,ubjÍ)”}”(hhh]”hæ)”}”(hŒmThe number of times the cgroup's swap usage was about to go over the max boundary and swap allocation failed.”h]”hŒoThe number of times the cgroupâ€™s swap usage was about to go over the max boundary and swap allocation failed.”…””}”(hjz,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mæhjw,ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhje,ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mèhj3,ubj·)”}”(hŒpfail The number of times swap allocation failed either because of running out of swap system-wide or max limit. ”h]”(j½)”}”(hŒfail”h]”hŒfail”…””}”(hj˜,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Míhj”,ubjÍ)”}”(hhh]”hæ)”}”(hŒjThe number of times swap allocation failed either because of running out of swap system-wide or max limit.”h]”hŒjThe number of times swap allocation failed either because of running out of swap system-wide or max limit.”…””}”(hj©,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mëhj¦,ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj”,ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Míhj3,ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj/,ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Máhj,ubhæ)”}”(hŒéWhen reduced under the current usage, the existing swap entries are reclaimed gradually and the swap usage may stay higher than the limit for an extended period of time. This reduces the impact on the workload and memory management.”h]”hŒéWhen reduced under the current usage, the existing swap entries are reclaimed gradually and the swap usage may stay higher than the limit for an extended period of time. This reduces the impact on the workload and memory management.”…””}”(hjÏ,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mïhj,ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj,ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mòhjiubj·)”}”(hŒœmemory.zswap.current A read-only single value file which exists on non-root cgroups. The total amount of memory consumed by the zswap compression backend. ”h]”(j½)”}”(hŒmemory.zswap.current”h]”hŒmemory.zswap.current”…””}”(hjí,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mùhjé,ubjÍ)”}”(hhh]”(hæ)”}”(hŒ?A read-only single value file which exists on non-root cgroups.”h]”hŒ?A read-only single value file which exists on non-root cgroups.”…””}”(hjþ,hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mõhjû,ubhæ)”}”(hŒEThe total amount of memory consumed by the zswap compression backend.”h]”hŒEThe total amount of memory consumed by the zswap compression backend.”…””}”(hj-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Møhjû,ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjé,ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mùhjiubj·)”}”(hXmemory.zswap.max A read-write single value file which exists on non-root cgroups. The default is "max". Zswap usage hard limit. If a cgroup's zswap pool reaches this limit, it will refuse to take any more stores before existing entries fault back in or are written out to disk. ”h]”(j½)”}”(hŒmemory.zswap.max”h]”hŒmemory.zswap.max”…””}”(hj*-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhj&-ubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-write single value file which exists on non-root cgroups. The default is "max".”h]”hŒ[A read-write single value file which exists on non-root cgroups. The default is â€œmaxâ€.”…””}”(hj;-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mühj8-ubhæ)”}”(hŒZswap usage hard limit. If a cgroup's zswap pool reaches this limit, it will refuse to take any more stores before existing entries fault back in or are written out to disk.”h]”hŒ¯Zswap usage hard limit. If a cgroupâ€™s zswap pool reaches this limit, it will refuse to take any more stores before existing entries fault back in or are written out to disk.”…””}”(hjI-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mÿhj8-ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj&-ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjiubj·)”}”(hX>memory.zswap.writeback A read-write single value file. The default value is "1". Note that this setting is hierarchical, i.e. the writeback would be implicitly disabled for child cgroups if the upper hierarchy does so. When this is set to 0, all swapping attempts to swapping devices are disabled. This included both zswap writebacks, and swapping due to zswap store failures. If the zswap store failures are recurring (for e.g if the pages are incompressible), users can observe reclaim inefficiency after disabling writeback (because the same pages might be rejected again and again). Note that this is subtly different from setting memory.swap.max to 0, as it still allows for pages to be written to the zswap pool. This setting has no effect if zswap is disabled, and swapping is allowed unless memory.swap.max is set to 0. ”h]”(j½)”}”(hŒmemory.zswap.writeback”h]”hŒmemory.zswap.writeback”…””}”(hjg-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhjc-ubjÍ)”}”(hhh]”(hæ)”}”(hŒÃA read-write single value file. The default value is "1". Note that this setting is hierarchical, i.e. the writeback would be implicitly disabled for child cgroups if the upper hierarchy does so.”h]”hŒÇA read-write single value file. The default value is â€œ1â€. Note that this setting is hierarchical, i.e. the writeback would be implicitly disabled for child cgroups if the upper hierarchy does so.”…””}”(hjx-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhju-ubhæ)”}”(hXoWhen this is set to 0, all swapping attempts to swapping devices are disabled. This included both zswap writebacks, and swapping due to zswap store failures. If the zswap store failures are recurring (for e.g if the pages are incompressible), users can observe reclaim inefficiency after disabling writeback (because the same pages might be rejected again and again).”h]”hXoWhen this is set to 0, all swapping attempts to swapping devices are disabled. This included both zswap writebacks, and swapping due to zswap store failures. If the zswap store failures are recurring (for e.g if the pages are incompressible), users can observe reclaim inefficiency after disabling writeback (because the same pages might be rejected again and again).”…””}”(hj†-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hju-ubhæ)”}”(hŒðNote that this is subtly different from setting memory.swap.max to 0, as it still allows for pages to be written to the zswap pool. This setting has no effect if zswap is disabled, and swapping is allowed unless memory.swap.max is set to 0.”h]”hŒðNote that this is subtly different from setting memory.swap.max to 0, as it still allows for pages to be written to the zswap pool. This setting has no effect if zswap is disabled, and swapping is allowed unless memory.swap.max is set to 0.”…””}”(hj”-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhju-ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjc-ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjiubj·)”}”(hŒmemory.pressure A read-only nested-keyed file. Shows pressure stall information for memory. See :ref:`Documentation/accounting/psi.rst ` for details. ”h]”(j½)”}”(hŒmemory.pressure”h]”hŒmemory.pressure”…””}”(hj²-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhj®-ubjÍ)”}”(hhh]”(hæ)”}”(hŒA read-only nested-keyed file.”h]”hŒA read-only nested-keyed file.”…””}”(hjÃ-hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÀ-ubhæ)”}”(hŒkShows pressure stall information for memory. See :ref:`Documentation/accounting/psi.rst ` for details.”h]”(hŒ1Shows pressure stall information for memory. See ”…””}”(hjÑ-hžhhŸNh Nubh)”}”(hŒ-:ref:`Documentation/accounting/psi.rst `”h]”jX)”}”(hjÛ-h]”hŒ Documentation/accounting/psi.rst”…””}”(hjÝ-hžhhŸNh Nubah}”(h]”h ]”(jcŒstd”Œstd-ref”eh"]”h$]”h&]”uh1jWhjÙ-ubah}”(h]”h ]”h"]”h$]”h&]”Œrefdoc”jpŒ refdomain”jç-Œreftype”Œref”Œrefexplicit”ˆŒrefwarn”ˆjvŒpsi”uh1hhŸh¯h MhjÑ-ubhŒ for details.”…””}”(hjÑ-hžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÀ-ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj®-ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjiubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjÐubeh}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M>hjzhžhubeh}”(h]”Œmemory-interface-files”ah ]”h"]”Œmemory interface files”ah$]”h&]”uh1h°hjîhžhhŸh¯h MÊubh±)”}”(hhh]”(h¶)”}”(hŒUsage Guidelines”h]”hŒUsage Guidelines”…””}”(hj&.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj#.hžhhŸh¯h Mubhæ)”}”(hŒã"memory.high" is the main mechanism to control memory usage. Over-committing on high limit (sum of high limits > available memory) and letting global memory pressure to distribute memory according to usage is a viable strategy.”h]”hŒçâ€œmemory.highâ€ is the main mechanism to control memory usage. Over-committing on high limit (sum of high limits > available memory) and letting global memory pressure to distribute memory according to usage is a viable strategy.”…””}”(hj4.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj#.hžhubhæ)”}”(hŒðBecause breach of the high limit doesn't trigger the OOM killer but throttles the offending cgroup, a management agent has ample opportunities to monitor and take appropriate actions such as granting more memory or terminating the workload.”h]”hŒòBecause breach of the high limit doesnâ€™t trigger the OOM killer but throttles the offending cgroup, a management agent has ample opportunities to monitor and take appropriate actions such as granting more memory or terminating the workload.”…””}”(hjB.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M$hj#.hžhubhæ)”}”(hX&Determining whether a cgroup has enough memory is not trivial as memory usage doesn't indicate whether the workload can benefit from more memory. For example, a workload which writes data received from network to a file can use all available memory but can also operate as performant with a small amount of memory. A measure of memory pressure - how much the workload is being impacted due to lack of memory - is necessary to determine whether a workload needs more memory; unfortunately, memory pressure monitoring mechanism isn't implemented yet.”h]”hX*Determining whether a cgroup has enough memory is not trivial as memory usage doesnâ€™t indicate whether the workload can benefit from more memory. For example, a workload which writes data received from network to a file can use all available memory but can also operate as performant with a small amount of memory. A measure of memory pressure - how much the workload is being impacted due to lack of memory - is necessary to determine whether a workload needs more memory; unfortunately, memory pressure monitoring mechanism isnâ€™t implemented yet.”…””}”(hjP.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M)hj#.hžhubeh}”(h]”Œusage-guidelines”ah ]”h"]”Œusage guidelines”ah$]”h&]”uh1h°hjîhžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒMemory Ownership”h]”hŒMemory Ownership”…””}”(hji.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjf.hžhhŸh¯h M5ubhæ)”}”(hXA memory area is charged to the cgroup which instantiated it and stays charged to the cgroup until the area is released. Migrating a process to a different cgroup doesn't move the memory usages that it instantiated while in the previous cgroup to the new cgroup.”h]”hX A memory area is charged to the cgroup which instantiated it and stays charged to the cgroup until the area is released. Migrating a process to a different cgroup doesnâ€™t move the memory usages that it instantiated while in the previous cgroup to the new cgroup.”…””}”(hjw.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M7hjf.hžhubhæ)”}”(hX A memory area may be used by processes belonging to different cgroups. To which cgroup the area will be charged is in-deterministic; however, over time, the memory area is likely to end up in a cgroup which has enough memory allowance to avoid high reclaim pressure.”h]”hX A memory area may be used by processes belonging to different cgroups. To which cgroup the area will be charged is in-deterministic; however, over time, the memory area is likely to end up in a cgroup which has enough memory allowance to avoid high reclaim pressure.”…””}”(hj….hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M<hjf.hžhubhæ)”}”(hX If a cgroup sweeps a considerable amount of memory which is expected to be accessed repeatedly by other cgroups, it may make sense to use POSIX_FADV_DONTNEED to relinquish the ownership of memory areas belonging to the affected files to ensure correct memory ownership.”h]”hX If a cgroup sweeps a considerable amount of memory which is expected to be accessed repeatedly by other cgroups, it may make sense to use POSIX_FADV_DONTNEED to relinquish the ownership of memory areas belonging to the affected files to ensure correct memory ownership.”…””}”(hj“.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MAhjf.hžhubeh}”(h]”Œmemory-ownership”ah ]”h"]”Œmemory ownership”ah$]”h&]”uh1h°hjîhžhhŸh¯h M5ubeh}”(h]”Œmemory”ah ]”h"]”h$]”Œmemory”ah&]”uh1h°hj«hžhhŸh¯h M³Œ referenced”Kubh±)”}”(hhh]”(h¶)”}”(hŒIO”h]”hŒIO”…””}”(hjµ.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj².hžhhŸh¯h MHubhæ)”}”(hX"The "io" controller regulates the distribution of IO resources. This controller implements both weight based and absolute bandwidth or IOPS limit distribution; however, weight based distribution is available only if cfq-iosched is in use and neither scheme is available for blk-mq devices.”h]”hX&The â€œioâ€ controller regulates the distribution of IO resources. This controller implements both weight based and absolute bandwidth or IOPS limit distribution; however, weight based distribution is available only if cfq-iosched is in use and neither scheme is available for blk-mq devices.”…””}”(hjÃ.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MJhj².hžhubh±)”}”(hhh]”(h¶)”}”(hŒIO Interface Files”h]”hŒIO Interface Files”…””}”(hjÔ.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjÑ.hžhhŸh¯h MRubj¬)”}”(hXio.stat A read-only nested-keyed file. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The following nested keys are defined. ====== ===================== rbytes Bytes read wbytes Bytes written rios Number of read IOs wios Number of write IOs dbytes Bytes discarded dios Number of discard IOs ====== ===================== An example read output follows:: 8:16 rbytes=1459200 wbytes=314773504 rios=192 wios=353 dbytes=0 dios=0 8:0 rbytes=90430464 wbytes=299008000 rios=8950 wios=1252 dbytes=50331648 dios=3021 io.cost.qos A read-write nested-keyed file which exists only on the root cgroup. This file configures the Quality of Service of the IO cost model based controller (CONFIG_BLK_CGROUP_IOCOST) which currently implements "io.weight" proportional control. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The line for a given device is populated on the first write for the device on "io.cost.qos" or "io.cost.model". The following nested keys are defined. ====== ===================================== enable Weight-based control enable ctrl "auto" or "user" rpct Read latency percentile [0, 100] rlat Read latency threshold wpct Write latency percentile [0, 100] wlat Write latency threshold min Minimum scaling percentage [1, 10000] max Maximum scaling percentage [1, 10000] ====== ===================================== The controller is disabled by default and can be enabled by setting "enable" to 1. "rpct" and "wpct" parameters default to zero and the controller uses internal device saturation state to adjust the overall IO rate between "min" and "max". When a better control quality is needed, latency QoS parameters can be configured. For example:: 8:16 enable=1 ctrl=auto rpct=95.00 rlat=75000 wpct=95.00 wlat=150000 min=50.00 max=150.0 shows that on sdb, the controller is enabled, will consider the device saturated if the 95th percentile of read completion latencies is above 75ms or write 150ms, and adjust the overall IO issue rate between 50% and 150% accordingly. The lower the saturation point, the better the latency QoS at the cost of aggregate bandwidth. The narrower the allowed adjustment range between "min" and "max", the more conformant to the cost model the IO behavior. Note that the IO issue base rate may be far off from 100% and setting "min" and "max" blindly can lead to a significant loss of device capacity or control quality. "min" and "max" are useful for regulating devices which show wide temporary behavior changes - e.g. a ssd which accepts writes at the line speed for a while and then completely stalls for multiple seconds. When "ctrl" is "auto", the parameters are controlled by the kernel and may change automatically. Setting "ctrl" to "user" or setting any of the percentile and latency parameters puts it into "user" mode and disables the automatic changes. The automatic mode can be restored by setting "ctrl" to "auto". io.cost.model A read-write nested-keyed file which exists only on the root cgroup. This file configures the cost model of the IO cost model based controller (CONFIG_BLK_CGROUP_IOCOST) which currently implements "io.weight" proportional control. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The line for a given device is populated on the first write for the device on "io.cost.qos" or "io.cost.model". The following nested keys are defined. ===== ================================ ctrl "auto" or "user" model The cost model in use - "linear" ===== ================================ When "ctrl" is "auto", the kernel may change all parameters dynamically. When "ctrl" is set to "user" or any other parameters are written to, "ctrl" become "user" and the automatic changes are disabled. When "model" is "linear", the following model parameters are defined. ============= ======================================== [r|w]bps The maximum sequential IO throughput [r|w]seqiops The maximum 4k sequential IOs per second [r|w]randiops The maximum 4k random IOs per second ============= ======================================== From the above, the builtin linear model determines the base costs of a sequential and random IO and the cost coefficient for the IO size. While simple, this model can cover most common device classes acceptably. The IO cost model isn't expected to be accurate in absolute sense and is scaled to the device behavior dynamically. If needed, tools/cgroup/iocost_coef_gen.py can be used to generate device-specific coefficients. io.weight A read-write flat-keyed file which exists on non-root cgroups. The default is "default 100". The first line is the default weight applied to devices without specific override. The rest are overrides keyed by $MAJ:$MIN device numbers and not ordered. The weights are in the range [1, 10000] and specifies the relative amount IO time the cgroup can use in relation to its siblings. The default weight can be updated by writing either "default $WEIGHT" or simply "$WEIGHT". Overrides can be set by writing "$MAJ:$MIN $WEIGHT" and unset by writing "$MAJ:$MIN default". An example read output follows:: default 100 8:16 200 8:0 50 io.max A read-write nested-keyed file which exists on non-root cgroups. BPS and IOPS based IO limit. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The following nested keys are defined. ===== ================================== rbps Max read bytes per second wbps Max write bytes per second riops Max read IO operations per second wiops Max write IO operations per second ===== ================================== When writing, any number of nested key-value pairs can be specified in any order. "max" can be specified as the value to remove a specific limit. If the same key is specified multiple times, the outcome is undefined. BPS and IOPS are measured in each IO direction and IOs are delayed if limit is reached. Temporary bursts are allowed. Setting read limit at 2M BPS and write at 120 IOPS for 8:16:: echo "8:16 rbps=2097152 wiops=120" > io.max Reading returns the following:: 8:16 rbps=2097152 wbps=max riops=max wiops=120 Write IOPS limit can be removed by writing the following:: echo "8:16 wiops=max" > io.max Reading now returns the following:: 8:16 rbps=2097152 wbps=max riops=max wiops=max io.pressure A read-only nested-keyed file. Shows pressure stall information for IO. See :ref:`Documentation/accounting/psi.rst ` for details. ”h]”j²)”}”(hhh]”(j·)”}”(hX`io.stat A read-only nested-keyed file. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The following nested keys are defined. ====== ===================== rbytes Bytes read wbytes Bytes written rios Number of read IOs wios Number of write IOs dbytes Bytes discarded dios Number of discard IOs ====== ===================== An example read output follows:: 8:16 rbytes=1459200 wbytes=314773504 rios=192 wios=353 dbytes=0 dios=0 8:0 rbytes=90430464 wbytes=299008000 rios=8950 wios=1252 dbytes=50331648 dios=3021 ”h]”(j½)”}”(hŒio.stat”h]”hŒio.stat”…””}”(hjí.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mfhjé.ubjÍ)”}”(hhh]”(hæ)”}”(hŒA read-only nested-keyed file.”h]”hŒA read-only nested-keyed file.”…””}”(hjþ.hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MUhjû.ubhæ)”}”(hŒcLines are keyed by $MAJ:$MIN device numbers and not ordered. The following nested keys are defined.”h]”hŒcLines are keyed by $MAJ:$MIN device numbers and not ordered. The following nested keys are defined.”…””}”(hj/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MWhjû.ubj¬)”}”(hX====== ===================== rbytes Bytes read wbytes Bytes written rios Number of read IOs wios Number of write IOs dbytes Bytes discarded dios Number of discard IOs ====== ===================== ”h]”jÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhj!/ubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhj!/ubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒrbytes”h]”hŒrbytes”…””}”(hjA/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M[hj>/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj;/ubj)”}”(hhh]”hæ)”}”(hŒ Bytes read”h]”hŒ Bytes read”…””}”(hjX/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M[hjU/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj;/ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj8/ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒwbytes”h]”hŒwbytes”…””}”(hjx/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M\hju/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjr/ubj)”}”(hhh]”hæ)”}”(hŒ Bytes written”h]”hŒ Bytes written”…””}”(hj/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M\hjŒ/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjr/ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj8/ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒrios”h]”hŒrios”…””}”(hj¯/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M]hj¬/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj©/ubj)”}”(hhh]”hæ)”}”(hŒNumber of read IOs”h]”hŒNumber of read IOs”…””}”(hjÆ/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M]hjÃ/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj©/ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj8/ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒwios”h]”hŒwios”…””}”(hjæ/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M^hjã/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjà/ubj)”}”(hhh]”hæ)”}”(hŒNumber of write IOs”h]”hŒNumber of write IOs”…””}”(hjý/hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M^hjú/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjà/ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj8/ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒdbytes”h]”hŒdbytes”…””}”(hj0hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M_hj0ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj0ubj)”}”(hhh]”hæ)”}”(hŒBytes discarded”h]”hŒBytes discarded”…””}”(hj40hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M_hj10ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj0ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj8/ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒdios”h]”hŒdios”…””}”(hjT0hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M`hjQ0ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjN0ubj)”}”(hhh]”hæ)”}”(hŒNumber of discard IOs”h]”hŒNumber of discard IOs”…””}”(hjk0hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M`hjh0ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjN0ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj8/ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühj!/ubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhj/ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhj/ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MZhjû.ubhæ)”}”(hŒ An example read output follows::”h]”hŒAn example read output follows:”…””}”(hjž0hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mchjû.ubjV)”}”(hŒ™8:16 rbytes=1459200 wbytes=314773504 rios=192 wios=353 dbytes=0 dios=0 8:0 rbytes=90430464 wbytes=299008000 rios=8950 wios=1252 dbytes=50331648 dios=3021”h]”hŒ™8:16 rbytes=1459200 wbytes=314773504 rios=192 wios=353 dbytes=0 dios=0 8:0 rbytes=90430464 wbytes=299008000 rios=8950 wios=1252 dbytes=50331648 dios=3021”…””}”hj¬0sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mehjû.ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjé.ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mfhjæ.ubj·)”}”(hXÍ io.cost.qos A read-write nested-keyed file which exists only on the root cgroup. This file configures the Quality of Service of the IO cost model based controller (CONFIG_BLK_CGROUP_IOCOST) which currently implements "io.weight" proportional control. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The line for a given device is populated on the first write for the device on "io.cost.qos" or "io.cost.model". The following nested keys are defined. ====== ===================================== enable Weight-based control enable ctrl "auto" or "user" rpct Read latency percentile [0, 100] rlat Read latency threshold wpct Write latency percentile [0, 100] wlat Write latency threshold min Minimum scaling percentage [1, 10000] max Maximum scaling percentage [1, 10000] ====== ===================================== The controller is disabled by default and can be enabled by setting "enable" to 1. "rpct" and "wpct" parameters default to zero and the controller uses internal device saturation state to adjust the overall IO rate between "min" and "max". When a better control quality is needed, latency QoS parameters can be configured. For example:: 8:16 enable=1 ctrl=auto rpct=95.00 rlat=75000 wpct=95.00 wlat=150000 min=50.00 max=150.0 shows that on sdb, the controller is enabled, will consider the device saturated if the 95th percentile of read completion latencies is above 75ms or write 150ms, and adjust the overall IO issue rate between 50% and 150% accordingly. The lower the saturation point, the better the latency QoS at the cost of aggregate bandwidth. The narrower the allowed adjustment range between "min" and "max", the more conformant to the cost model the IO behavior. Note that the IO issue base rate may be far off from 100% and setting "min" and "max" blindly can lead to a significant loss of device capacity or control quality. "min" and "max" are useful for regulating devices which show wide temporary behavior changes - e.g. a ssd which accepts writes at the line speed for a while and then completely stalls for multiple seconds. When "ctrl" is "auto", the parameters are controlled by the kernel and may change automatically. Setting "ctrl" to "user" or setting any of the percentile and latency parameters puts it into "user" mode and disables the automatic changes. The automatic mode can be restored by setting "ctrl" to "auto". ”h]”(j½)”}”(hŒio.cost.qos”h]”hŒio.cost.qos”…””}”(hjÊ0hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MhjÆ0ubjÍ)”}”(hhh]”(hæ)”}”(hŒDA read-write nested-keyed file which exists only on the root cgroup.”h]”hŒDA read-write nested-keyed file which exists only on the root cgroup.”…””}”(hjÛ0hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MihjØ0ubhæ)”}”(hX€This file configures the Quality of Service of the IO cost model based controller (CONFIG_BLK_CGROUP_IOCOST) which currently implements "io.weight" proportional control. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The line for a given device is populated on the first write for the device on "io.cost.qos" or "io.cost.model". The following nested keys are defined.”h]”hXŒThis file configures the Quality of Service of the IO cost model based controller (CONFIG_BLK_CGROUP_IOCOST) which currently implements â€œio.weightâ€ proportional control. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The line for a given device is populated on the first write for the device on â€œio.cost.qosâ€ or â€œio.cost.modelâ€. The following nested keys are defined.”…””}”(hjé0hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MlhjØ0ubj¬)”}”(hXÈ====== ===================================== enable Weight-based control enable ctrl "auto" or "user" rpct Read latency percentile [0, 100] rlat Read latency threshold wpct Write latency percentile [0, 100] wlat Write latency threshold min Minimum scaling percentage [1, 10000] max Maximum scaling percentage [1, 10000] ====== ===================================== ”h]”jÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhjþ0ubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K%uh1jæhjþ0ubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒenable”h]”hŒenable”…””}”(hj1hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Muhj1ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubj)”}”(hhh]”hæ)”}”(hŒWeight-based control enable”h]”hŒWeight-based control enable”…””}”(hj51hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Muhj21ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒctrl”h]”hŒctrl”…””}”(hjU1hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MvhjR1ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjO1ubj)”}”(hhh]”hæ)”}”(hŒ"auto" or "user"”h]”hŒâ€œautoâ€ or â€œuserâ€”…””}”(hjl1hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mvhji1ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjO1ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒrpct”h]”hŒrpct”…””}”(hjŒ1hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mwhj‰1ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj†1ubj)”}”(hhh]”hæ)”}”(hŒ#Read latency percentile [0, 100]”h]”hŒ#Read latency percentile [0, 100]”…””}”(hj£1hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mwhj 1ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj†1ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒrlat”h]”hŒrlat”…””}”(hjÃ1hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MxhjÀ1ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj½1ubj)”}”(hhh]”hæ)”}”(hŒRead latency threshold”h]”hŒRead latency threshold”…””}”(hjÚ1hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mxhj×1ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj½1ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒwpct”h]”hŒwpct”…””}”(hjú1hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Myhj÷1ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjô1ubj)”}”(hhh]”hæ)”}”(hŒ#Write latency percentile [0, 100]”h]”hŒ#Write latency percentile [0, 100]”…””}”(hj2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Myhj2ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjô1ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒwlat”h]”hŒwlat”…””}”(hj12hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mzhj.2ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj+2ubj)”}”(hhh]”hæ)”}”(hŒWrite latency threshold”h]”hŒWrite latency threshold”…””}”(hjH2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MzhjE2ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj+2ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒmin”h]”hŒmin”…””}”(hjh2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M{hje2ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjb2ubj)”}”(hhh]”hæ)”}”(hŒ%Minimum scaling percentage [1, 10000]”h]”hŒ%Minimum scaling percentage [1, 10000]”…””}”(hj2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M{hj|2ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjb2ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒmax”h]”hŒmax”…””}”(hjŸ2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M|hjœ2ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj™2ubj)”}”(hhh]”hæ)”}”(hŒ%Maximum scaling percentage [1, 10000]”h]”hŒ%Maximum scaling percentage [1, 10000]”…””}”(hj¶2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M|hj³2ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj™2ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj1ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühjþ0ubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhjû0ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhj÷0ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MthjØ0ubhæ)”}”(hŒðThe controller is disabled by default and can be enabled by setting "enable" to 1. "rpct" and "wpct" parameters default to zero and the controller uses internal device saturation state to adjust the overall IO rate between "min" and "max".”h]”hXThe controller is disabled by default and can be enabled by setting â€œenableâ€ to 1. â€œrpctâ€ and â€œwpctâ€ parameters default to zero and the controller uses internal device saturation state to adjust the overall IO rate between â€œminâ€ and â€œmaxâ€.”…””}”(hjé2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjØ0ubhæ)”}”(hŒaWhen a better control quality is needed, latency QoS parameters can be configured. For example::”h]”hŒ`When a better control quality is needed, latency QoS parameters can be configured. For example:”…””}”(hj÷2hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M„hjØ0ubjV)”}”(hŒX8:16 enable=1 ctrl=auto rpct=95.00 rlat=75000 wpct=95.00 wlat=150000 min=50.00 max=150.0”h]”hŒX8:16 enable=1 ctrl=auto rpct=95.00 rlat=75000 wpct=95.00 wlat=150000 min=50.00 max=150.0”…””}”hj3sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M‡hjØ0ubhæ)”}”(hŒéshows that on sdb, the controller is enabled, will consider the device saturated if the 95th percentile of read completion latencies is above 75ms or write 150ms, and adjust the overall IO issue rate between 50% and 150% accordingly.”h]”hŒéshows that on sdb, the controller is enabled, will consider the device saturated if the 95th percentile of read completion latencies is above 75ms or write 150ms, and adjust the overall IO issue rate between 50% and 150% accordingly.”…””}”(hj3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‰hjØ0ubhæ)”}”(hXMThe lower the saturation point, the better the latency QoS at the cost of aggregate bandwidth. The narrower the allowed adjustment range between "min" and "max", the more conformant to the cost model the IO behavior. Note that the IO issue base rate may be far off from 100% and setting "min" and "max" blindly can lead to a significant loss of device capacity or control quality. "min" and "max" are useful for regulating devices which show wide temporary behavior changes - e.g. a ssd which accepts writes at the line speed for a while and then completely stalls for multiple seconds.”h]”hXeThe lower the saturation point, the better the latency QoS at the cost of aggregate bandwidth. The narrower the allowed adjustment range between â€œminâ€ and â€œmaxâ€, the more conformant to the cost model the IO behavior. Note that the IO issue base rate may be far off from 100% and setting â€œminâ€ and â€œmaxâ€ blindly can lead to a significant loss of device capacity or control quality. â€œminâ€ and â€œmaxâ€ are useful for regulating devices which show wide temporary behavior changes - e.g. a ssd which accepts writes at the line speed for a while and then completely stalls for multiple seconds.”…””}”(hj!3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŽhjØ0ubhæ)”}”(hX0When "ctrl" is "auto", the parameters are controlled by the kernel and may change automatically. Setting "ctrl" to "user" or setting any of the percentile and latency parameters puts it into "user" mode and disables the automatic changes. The automatic mode can be restored by setting "ctrl" to "auto".”h]”hXLWhen â€œctrlâ€ is â€œautoâ€, the parameters are controlled by the kernel and may change automatically. Setting â€œctrlâ€ to â€œuserâ€ or setting any of the percentile and latency parameters puts it into â€œuserâ€ mode and disables the automatic changes. The automatic mode can be restored by setting â€œctrlâ€ to â€œautoâ€.”…””}”(hj/3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M™hjØ0ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÆ0ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjæ.ubj·)”}”(hXZio.cost.model A read-write nested-keyed file which exists only on the root cgroup. This file configures the cost model of the IO cost model based controller (CONFIG_BLK_CGROUP_IOCOST) which currently implements "io.weight" proportional control. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The line for a given device is populated on the first write for the device on "io.cost.qos" or "io.cost.model". The following nested keys are defined. ===== ================================ ctrl "auto" or "user" model The cost model in use - "linear" ===== ================================ When "ctrl" is "auto", the kernel may change all parameters dynamically. When "ctrl" is set to "user" or any other parameters are written to, "ctrl" become "user" and the automatic changes are disabled. When "model" is "linear", the following model parameters are defined. ============= ======================================== [r|w]bps The maximum sequential IO throughput [r|w]seqiops The maximum 4k sequential IOs per second [r|w]randiops The maximum 4k random IOs per second ============= ======================================== From the above, the builtin linear model determines the base costs of a sequential and random IO and the cost coefficient for the IO size. While simple, this model can cover most common device classes acceptably. The IO cost model isn't expected to be accurate in absolute sense and is scaled to the device behavior dynamically. If needed, tools/cgroup/iocost_coef_gen.py can be used to generate device-specific coefficients. ”h]”(j½)”}”(hŒ io.cost.model”h]”hŒ io.cost.model”…””}”(hjM3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÇhjI3ubjÍ)”}”(hhh]”(hæ)”}”(hŒDA read-write nested-keyed file which exists only on the root cgroup.”h]”hŒDA read-write nested-keyed file which exists only on the root cgroup.”…””}”(hj^3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hj[3ubhæ)”}”(hXxThis file configures the cost model of the IO cost model based controller (CONFIG_BLK_CGROUP_IOCOST) which currently implements "io.weight" proportional control. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The line for a given device is populated on the first write for the device on "io.cost.qos" or "io.cost.model". The following nested keys are defined.”h]”hX„This file configures the cost model of the IO cost model based controller (CONFIG_BLK_CGROUP_IOCOST) which currently implements â€œio.weightâ€ proportional control. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The line for a given device is populated on the first write for the device on â€œio.cost.qosâ€ or â€œio.cost.modelâ€. The following nested keys are defined.”…””}”(hjl3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M£hj[3ubj¬)”}”(hŒ¬===== ================================ ctrl "auto" or "user" model The cost model in use - "linear" ===== ================================ ”h]”jÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhj3ubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K uh1jæhj3ubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒctrl”h]”hŒctrl”…””}”(hj¡3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¬hjž3ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj›3ubj)”}”(hhh]”hæ)”}”(hŒ"auto" or "user"”h]”hŒâ€œautoâ€ or â€œuserâ€”…””}”(hj¸3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¬hjµ3ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj›3ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj˜3ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒmodel”h]”hŒmodel”…””}”(hjØ3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÕ3ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÒ3ubj)”}”(hhh]”hæ)”}”(hŒ The cost model in use - "linear"”h]”hŒ$The cost model in use - â€œlinearâ€”…””}”(hjï3hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhjì3ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÒ3ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj˜3ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühj3ubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhj~3ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhjz3ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M«hj[3ubhæ)”}”(hŒËWhen "ctrl" is "auto", the kernel may change all parameters dynamically. When "ctrl" is set to "user" or any other parameters are written to, "ctrl" become "user" and the automatic changes are disabled.”•h]”hŒãWhen â€œctrlâ€ is â€œautoâ€, the kernel may change all parameters dynamically. When â€œctrlâ€ is set to â€œuserâ€ or any other parameters are written to, â€œctrlâ€ become â€œuserâ€ and the automatic changes are disabled.”…””}”(hj"4hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M°hj[3ubhæ)”}”(hŒEWhen "model" is "linear", the following model parameters are defined.”h]”hŒMWhen â€œmodelâ€ is â€œlinearâ€, the following model parameters are defined.”…””}”(hj04hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mµhj[3ubj¬)”}”(hX============= ======================================== [r|w]bps The maximum sequential IO throughput [r|w]seqiops The maximum 4k sequential IOs per second [r|w]randiops The maximum 4k random IOs per second ============= ======================================== ”h]”jÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K uh1jæhjE4ubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K(uh1jæhjE4ubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ[r|w]bps”h]”hŒ[r|w]bps”…””}”(hje4hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹hjb4ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj_4ubj)”}”(hhh]”hæ)”}”(hŒ$The maximum sequential IO throughput”h]”hŒ$The maximum sequential IO throughput”…””}”(hj|4hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹hjy4ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj_4ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj\4ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ[r|w]seqiops”h]”hŒ[r|w]seqiops”…””}”(hjœ4hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mºhj™4ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj–4ubj)”}”(hhh]”hæ)”}”(hŒ(The maximum 4k sequential IOs per second”h]”hŒ(The maximum 4k sequential IOs per second”…””}”(hj³4hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mºhj°4ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj–4ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj\4ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ [r|w]randiops”h]”hŒ [r|w]randiops”…””}”(hjÓ4hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M»hjÐ4ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÍ4ubj)”}”(hhh]”hæ)”}”(hŒ$The maximum 4k random IOs per second”h]”hŒ$The maximum 4k random IOs per second”…””}”(hjê4hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M»hjç4ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÍ4ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj\4ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühjE4ubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhjB4ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhj>4ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M¸hj[3ubhæ)”}”(hŒÕFrom the above, the builtin linear model determines the base costs of a sequential and random IO and the cost coefficient for the IO size. While simple, this model can cover most common device classes acceptably.”h]”hŒÕFrom the above, the builtin linear model determines the base costs of a sequential and random IO and the cost coefficient for the IO size. While simple, this model can cover most common device classes acceptably.”…””}”(hj5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¾hj[3ubhæ)”}”(hŒsThe IO cost model isn't expected to be accurate in absolute sense and is scaled to the device behavior dynamically.”h]”hŒuThe IO cost model isnâ€™t expected to be accurate in absolute sense and is scaled to the device behavior dynamically.”…””}”(hj+5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÃhj[3ubhæ)”}”(hŒ`If needed, tools/cgroup/iocost_coef_gen.py can be used to generate device-specific coefficients.”h]”hŒ`If needed, tools/cgroup/iocost_coef_gen.py can be used to generate device-specific coefficients.”…””}”(hj95hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÆhj[3ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjI3ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÇhjæ.ubj·)”}”(hX‰io.weight A read-write flat-keyed file which exists on non-root cgroups. The default is "default 100". The first line is the default weight applied to devices without specific override. The rest are overrides keyed by $MAJ:$MIN device numbers and not ordered. The weights are in the range [1, 10000] and specifies the relative amount IO time the cgroup can use in relation to its siblings. The default weight can be updated by writing either "default $WEIGHT" or simply "$WEIGHT". Overrides can be set by writing "$MAJ:$MIN $WEIGHT" and unset by writing "$MAJ:$MIN default". An example read output follows:: default 100 8:16 200 8:0 50 ”h]”(j½)”}”(hŒ io.weight”h]”hŒ io.weight”…””}”(hjW5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÛhjS5ubjÍ)”}”(hhh]”(hæ)”}”(hŒ\A read-write flat-keyed file which exists on non-root cgroups. The default is "default 100".”h]”hŒ`A read-write flat-keyed file which exists on non-root cgroups. The default is â€œdefault 100â€.”…””}”(hjh5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÊhje5ubhæ)”}”(hX The first line is the default weight applied to devices without specific override. The rest are overrides keyed by $MAJ:$MIN device numbers and not ordered. The weights are in the range [1, 10000] and specifies the relative amount IO time the cgroup can use in relation to its siblings.”h]”hX The first line is the default weight applied to devices without specific override. The rest are overrides keyed by $MAJ:$MIN device numbers and not ordered. The weights are in the range [1, 10000] and specifies the relative amount IO time the cgroup can use in relation to its siblings.”…””}”(hjv5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÍhje5ubhæ)”}”(hŒ¹The default weight can be updated by writing either "default $WEIGHT" or simply "$WEIGHT". Overrides can be set by writing "$MAJ:$MIN $WEIGHT" and unset by writing "$MAJ:$MIN default".”h]”hŒÉThe default weight can be updated by writing either â€œdefault $WEIGHTâ€ or simply â€œ$WEIGHTâ€. Overrides can be set by writing â€œ$MAJ:$MIN $WEIGHTâ€ and unset by writing â€œ$MAJ:$MIN defaultâ€.”…””}”(hj„5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÓhje5ubhæ)”}”(hŒ An example read output follows::”h]”hŒAn example read output follows:”…””}”(hj’5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M×hje5ubjV)”}”(hŒdefault 100 8:16 200 8:0 50”h]”hŒdefault 100 8:16 200 8:0 50”…””}”hj 5sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MÙhje5ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjS5ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÛhjæ.ubj·)”}”(hX·io.max A read-write nested-keyed file which exists on non-root cgroups. BPS and IOPS based IO limit. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The following nested keys are defined. ===== ================================== rbps Max read bytes per second wbps Max write bytes per second riops Max read IO operations per second wiops Max write IO operations per second ===== ================================== When writing, any number of nested key-value pairs can be specified in any order. "max" can be specified as the value to remove a specific limit. If the same key is specified multiple times, the outcome is undefined. BPS and IOPS are measured in each IO direction and IOs are delayed if limit is reached. Temporary bursts are allowed. Setting read limit at 2M BPS and write at 120 IOPS for 8:16:: echo "8:16 rbps=2097152 wiops=120" > io.max Reading returns the following:: 8:16 rbps=2097152 wbps=max riops=max wiops=120 Write IOPS limit can be removed by writing the following:: echo "8:16 wiops=max" > io.max Reading now returns the following:: 8:16 rbps=2097152 wbps=max riops=max wiops=max ”h]”(j½)”}”(hŒio.max”h]”hŒio.max”…””}”(hj¾5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhjº5ubjÍ)”}”(hhh]”(hæ)”}”(hŒ@A read-write nested-keyed file which exists on non-root cgroups.”h]”hŒ@A read-write nested-keyed file which exists on non-root cgroups.”…””}”(hjÏ5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÞhjÌ5ubhæ)”}”(hŒ‚BPS and IOPS based IO limit. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The following nested keys are defined.”h]”hŒ‚BPS and IOPS based IO limit. Lines are keyed by $MAJ:$MIN device numbers and not ordered. The following nested keys are defined.”…””}”(hjÝ5hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MáhjÌ5ubj¬)”}”(hX===== ================================== rbps Max read bytes per second wbps Max write bytes per second riops Max read IO operations per second wiops Max write IO operations per second ===== ================================== ”h]”jÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhjò5ubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K"uh1jæhjò5ubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒrbps”h]”hŒrbps”…””}”(hj6hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mæhj6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj6ubj)”}”(hhh]”hæ)”}”(hŒMax read bytes per second”h]”hŒMax read bytes per second”…””}”(hj)6hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mæhj&6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj6ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj 6ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒwbps”h]”hŒwbps”…””}”(hjI6hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MçhjF6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjC6ubj)”}”(hhh]”hæ)”}”(hŒMax write bytes per second”h]”hŒMax write bytes per second”…””}”(hj`6hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mçhj]6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjC6ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj 6ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒriops”h]”hŒriops”…””}”(hj€6hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mèhj}6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjz6ubj)”}”(hhh]”hæ)”}”(hŒ!Max read IO operations per second”h]”hŒ!Max read IO operations per second”…””}”(hj—6hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mèhj”6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjz6ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj 6ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒwiops”h]”hŒwiops”…””}”(hj·6hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Méhj´6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj±6ubj)”}”(hhh]”hæ)”}”(hŒ"Max write IO operations per second”h]”hŒ"Max write IO operations per second”…””}”(hjÎ6hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MéhjË6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj±6ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj 6ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühjò5ubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhjï5ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhjë5ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MåhjÌ5ubhæ)”}”(hŒÚWhen writing, any number of nested key-value pairs can be specified in any order. "max" can be specified as the value to remove a specific limit. If the same key is specified multiple times, the outcome is undefined.”h]”hŒÞWhen writing, any number of nested key-value pairs can be specified in any order. â€œmaxâ€ can be specified as the value to remove a specific limit. If the same key is specified multiple times, the outcome is undefined.”…””}”(hj7hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MìhjÌ5ubhæ)”}”(hŒvBPS and IOPS are measured in each IO direction and IOs are delayed if limit is reached. Temporary bursts are allowed.”h]”hŒvBPS and IOPS are measured in each IO direction and IOs are delayed if limit is reached. Temporary bursts are allowed.”…””}”(hj7hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MñhjÌ5ubhæ)”}”(hŒ=Setting read limit at 2M BPS and write at 120 IOPS for 8:16::”h]”hŒ io.max”h]”hŒ+echo "8:16 rbps=2097152 wiops=120" > io.max”…””}”hj+7sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MöhjÌ5ubhæ)”}”(hŒReading returns the following::”h]”hŒReading returns the following:”…””}”(hj97hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MøhjÌ5ubjV)”}”(hŒ.8:16 rbps=2097152 wbps=max riops=max wiops=120”h]”hŒ.8:16 rbps=2097152 wbps=max riops=max wiops=120”…””}”hjG7sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MúhjÌ5ubhæ)”}”(hŒ:Write IOPS limit can be removed by writing the following::”h]”hŒ9Write IOPS limit can be removed by writing the following:”…””}”(hjU7hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MühjÌ5ubjV)”}”(hŒecho "8:16 wiops=max" > io.max”h]”hŒecho "8:16 wiops=max" > io.max”…””}”hjc7sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MþhjÌ5ubhæ)”}”(hŒ#Reading now returns the following::”h]”hŒ"Reading now returns the following:”…””}”(hjq7hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÌ5ubjV)”}”(hŒ.8:16 rbps=2097152 wbps=max riops=max wiops=max”h]”hŒ.8:16 rbps=2097152 wbps=max riops=max wiops=max”…””}”hj7sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MhjÌ5ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjº5ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhjæ.ubj·)”}”(hŒ•io.pressure A read-only nested-keyed file. Shows pressure stall information for IO. See :ref:`Documentation/accounting/psi.rst ` for details. ”h]”(j½)”}”(hŒio.pressure”h]”hŒio.pressure”…””}”(hj7hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M hj™7ubjÍ)”}”(hhh]”(hæ)”}”(hŒA read-only nested-keyed file.”h]”hŒA read-only nested-keyed file.”…””}”(hj®7hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj«7ubhæ)”}”(hŒgShows pressure stall information for IO. See :ref:`Documentation/accounting/psi.rst ` for details.”h]”(hŒ-Shows pressure stall information for IO. See ”…””}”(hj¼7hžhhŸNh Nubh)”}”(hŒ-:ref:`Documentation/accounting/psi.rst `”h]”jX)”}”(hjÆ7h]”hŒ Documentation/accounting/psi.rst”…””}”(hjÈ7hžhhŸNh Nubah}”(h]”h ]”(jcŒstd”Œstd-ref”eh"]”h$]”h&]”uh1jWhjÄ7ubah}”(h]”h ]”h"]”h$]”h&]”Œrefdoc”jpŒ refdomain”jÒ7Œreftype”Œref”Œrefexplicit”ˆŒrefwarn”ˆjvŒpsi”uh1hhŸh¯h Mhj¼7ubhŒ for details.”…””}”(hj¼7hžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj«7ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj™7ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M hjæ.ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjâ.ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MThjÑ.hžhubeh}”(h]”Œio-interface-files”ah ]”h"]”Œio interface files”ah$]”h&]”uh1h°hj².hžhhŸh¯h MRubh±)”}”(hhh]”(h¶)”}”(hŒ Writeback”h]”hŒ Writeback”…””}”(hj8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj8hžhhŸh¯h Mubhæ)”}”(hXPage cache is dirtied through buffered writes and shared mmaps and written asynchronously to the backing filesystem by the writeback mechanism. Writeback sits between the memory and IO domains and regulates the proportion of dirty memory by balancing dirtying and write IOs.”h]”hXPage cache is dirtied through buffered writes and shared mmaps and written asynchronously to the backing filesystem by the writeback mechanism. Writeback sits between the memory and IO domains and regulates the proportion of dirty memory by balancing dirtying and write IOs.”…””}”(hj8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj8hžhubhæ)”}”(hXªThe io controller, in conjunction with the memory controller, implements control of page cache writeback IOs. The memory controller defines the memory domain that dirty memory ratio is calculated and maintained for and the io controller defines the io domain which writes out dirty pages for the memory domain. Both system-wide and per-cgroup dirty memory states are examined and the more restrictive of the two is enforced.”h]”hXªThe io controller, in conjunction with the memory controller, implements control of page cache writeback IOs. The memory controller defines the memory domain that dirty memory ratio is calculated and maintained for and the io controller defines the io domain which writes out dirty pages for the memory domain. Both system-wide and per-cgroup dirty memory states are examined and the more restrictive of the two is enforced.”…””}”(hj-8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj8hžhubhæ)”}”(hŒçcgroup writeback requires explicit support from the underlying filesystem. Currently, cgroup writeback is implemented on ext2, ext4, btrfs, f2fs, and xfs. On other filesystems, all writeback IOs are attributed to the root cgroup.”h]”hŒçcgroup writeback requires explicit support from the underlying filesystem. Currently, cgroup writeback is implemented on ext2, ext4, btrfs, f2fs, and xfs. On other filesystems, all writeback IOs are attributed to the root cgroup.”…””}”(hj;8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj8hžhubhæ)”}”(hX>There are inherent differences in memory and writeback management which affects how cgroup ownership is tracked. Memory is tracked per page while writeback per inode. For the purpose of writeback, an inode is assigned to a cgroup and all IO requests to write dirty pages from the inode are attributed to that cgroup.”h]”hX>There are inherent differences in memory and writeback management which affects how cgroup ownership is tracked. Memory is tracked per page while writeback per inode. For the purpose of writeback, an inode is assigned to a cgroup and all IO requests to write dirty pages from the inode are attributed to that cgroup.”…””}”(hjI8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M!hj8hžhubhæ)”}”(hXAs cgroup ownership for memory is tracked per page, there can be pages which are associated with different cgroups than the one the inode is associated with. These are called foreign pages. The writeback constantly keeps track of foreign pages and, if a particular foreign cgroup becomes the majority over a certain period of time, switches the ownership of the inode to that cgroup.”h]”hXAs cgroup ownership for memory is tracked per page, there can be pages which are associated with different cgroups than the one the inode is associated with. These are called foreign pages. The writeback constantly keeps track of foreign pages and, if a particular foreign cgroup becomes the majority over a certain period of time, switches the ownership of the inode to that cgroup.”…””}”(hjW8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M'hj8hžhubhæ)”}”(hXrWhile this model is enough for most use cases where a given inode is mostly dirtied by a single cgroup even when the main writing cgroup changes over time, use cases where multiple cgroups write to a single inode simultaneously are not supported well. In such circumstances, a significant portion of IOs are likely to be attributed incorrectly. As memory controller assigns page ownership on the first use and doesn't update it until the page is released, even if writeback strictly follows page ownership, multiple cgroups dirtying overlapping areas wouldn't work as expected. It's recommended to avoid such usage patterns.”h]”hXxWhile this model is enough for most use cases where a given inode is mostly dirtied by a single cgroup even when the main writing cgroup changes over time, use cases where multiple cgroups write to a single inode simultaneously are not supported well. In such circumstances, a significant portion of IOs are likely to be attributed incorrectly. As memory controller assigns page ownership on the first use and doesnâ€™t update it until the page is released, even if writeback strictly follows page ownership, multiple cgroups dirtying overlapping areas wouldnâ€™t work as expected. Itâ€™s recommended to avoid such usage patterns.”…””}”(hje8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M.hj8hžhubhæ)”}”(hŒ\The sysctl knobs which affect writeback behavior are applied to cgroup writeback as follows.”h]”hŒ\The sysctl knobs which affect writeback behavior are applied to cgroup writeback as follows.”…””}”(hjs8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M9hj8hžhubj¬)”}”(hXªvm.dirty_background_ratio, vm.dirty_ratio These ratios apply the same to cgroup writeback with the amount of available memory capped by limits imposed by the memory controller and system-wide clean memory. vm.dirty_background_bytes, vm.dirty_bytes For cgroup writeback, this is calculated into ratio against total available memory and applied the same way as vm.dirty[_background]_ratio. ”h]”j²)”}”(hhh]”(j·)”}”(hŒÎvm.dirty_background_ratio, vm.dirty_ratio These ratios apply the same to cgroup writeback with the amount of available memory capped by limits imposed by the memory controller and system-wide clean memory. ”h]”(j½)”}”(hŒ)vm.dirty_background_ratio, vm.dirty_ratio”h]”hŒ)vm.dirty_background_ratio, vm.dirty_ratio”…””}”(hjŒ8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M?hjˆ8ubjÍ)”}”(hhh]”hæ)”}”(hŒ£These ratios apply the same to cgroup writeback with the amount of available memory capped by limits imposed by the memory controller and system-wide clean memory.”h]”hŒ£These ratios apply the same to cgroup writeback with the amount of available memory capped by limits imposed by the memory controller and system-wide clean memory.”…””}”(hj8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M=hjš8ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjˆ8ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M?hj…8ubj·)”}”(hŒ·vm.dirty_background_bytes, vm.dirty_bytes For cgroup writeback, this is calculated into ratio against total available memory and applied the same way as vm.dirty[_background]_ratio. ”h]”(j½)”}”(hŒ)vm.dirty_background_bytes, vm.dirty_bytes”h]”hŒ)vm.dirty_background_bytes, vm.dirty_bytes”…””}”(hj»8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MEhj·8ubjÍ)”}”(hhh]”hæ)”}”(hŒ‹For cgroup writeback, this is calculated into ratio against total available memory and applied the same way as vm.dirty[_background]_ratio.”h]”hŒ‹For cgroup writeback, this is calculated into ratio against total available memory and applied the same way as vm.dirty[_background]_ratio.”…””}”(hjÌ8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MBhjÉ8ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj·8ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MEhj…8ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj8ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M<hj8hžhubeh}”(h]”Œ writeback”ah ]”h"]”Œ writeback”ah$]”h&]”uh1h°hj².hžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒ IO Latency”h]”hŒ IO Latency”…””}”(hjý8hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjú8hžhhŸh¯h MHubhæ)”}”(hŒúThis is a cgroup v2 controller for IO workload protection. You provide a group with a latency target, and if the average latency exceeds that target the controller will throttle any peers that have a lower latency target than the protected workload.”h]”hŒúThis is a cgroup v2 controller for IO workload protection. You provide a group with a latency target, and if the average latency exceeds that target the controller will throttle any peers that have a lower latency target than the protected workload.”…””}”(hj9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MJhjú8hžhubhæ)”}”(hŒèThe limits are only applied at the peer level in the hierarchy. This means that in the diagram below, only groups A, B, and C will influence each other, and groups D and F will influence each other. Group G will influence nobody::”h]”hŒçThe limits are only applied at the peer level in the hierarchy. This means that in the diagram below, only groups A, B, and C will influence each other, and groups D and F will influence each other. Group G will influence nobody:”…””}”(hj9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MOhjú8hžhubjV)”}”(hŒf [root] / | \ A B C / \ | D F G”h]”hŒf [root] / | \ A B C / \ | D F G”…””}”hj'9sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MShjú8hžhubhæ)”}”(hX So the ideal way to configure this is to set io.latency in groups A, B, and C. Generally you do not want to set a value lower than the latency your device supports. Experiment to find the value that works best for your workload. Start at higher than the expected latency for your device and watch the avg_lat value in io.stat for your workload group to get an idea of the latency you see during normal operation. Use the avg_lat value as a basis for your real setting, setting at 10-15% higher than the value in io.stat.”h]”hX So the ideal way to configure this is to set io.latency in groups A, B, and C. Generally you do not want to set a value lower than the latency your device supports. Experiment to find the value that works best for your workload. Start at higher than the expected latency for your device and watch the avg_lat value in io.stat for your workload group to get an idea of the latency you see during normal operation. Use the avg_lat value as a basis for your real setting, setting at 10-15% higher than the value in io.stat.”…””}”(hj59hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MZhjú8hžhubeh}”(h]”Œ io-latency”ah ]”h"]”Œ io latency”ah$]”h&]”uh1h°hj².hžhhŸh¯h MHubh±)”}”(hhh]”(h¶)”}”(hŒHow IO Latency Throttling Works”h]”hŒHow IO Latency Throttling Works”…””}”(hjN9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjK9hžhhŸh¯h Mcubhæ)”}”(hXio.latency is work conserving; so as long as everybody is meeting their latency target the controller doesn't do anything. Once a group starts missing its target it begins throttling any peer group that has a higher target than itself. This throttling takes 2 forms:”h]”hX io.latency is work conserving; so as long as everybody is meeting their latency target the controller doesnâ€™t do anything. Once a group starts missing its target it begins throttling any peer group that has a higher target than itself. This throttling takes 2 forms:”…””}”(hj\9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MehjK9hžhubjª)”}”(hhh]”(j¯)”}”(hŒÇQueue depth throttling. This is the number of outstanding IO's a group is allowed to have. We will clamp down relatively quickly, starting at no limit and going all the way down to 1 IO at a time. ”h]”hæ)”}”(hŒÆQueue depth throttling. This is the number of outstanding IO's a group is allowed to have. We will clamp down relatively quickly, starting at no limit and going all the way down to 1 IO at a time.”h]”hŒÈQueue depth throttling. This is the number of outstanding IOâ€™s a group is allowed to have. We will clamp down relatively quickly, starting at no limit and going all the way down to 1 IO at a time.”…””}”(hjq9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mjhjm9ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjj9hžhhŸh¯h Nubj¯)”}”(hX“Artificial delay induction. There are certain types of IO that cannot be throttled without possibly adversely affecting higher priority groups. This includes swapping and metadata IO. These types of IO are allowed to occur normally, however they are "charged" to the originating group. If the originating group is being throttled you will see the use_delay and delay fields in io.stat increase. The delay value is how many microseconds that are being added to any process that runs in this group. Because this number can grow quite large if there is a lot of swapping or metadata IO occurring we limit the individual delay events to 1 second at a time. ”h]”hæ)”}”(hX’Artificial delay induction. There are certain types of IO that cannot be throttled without possibly adversely affecting higher priority groups. This includes swapping and metadata IO. These types of IO are allowed to occur normally, however they are "charged" to the originating group. If the originating group is being throttled you will see the use_delay and delay fields in io.stat increase. The delay value is how many microseconds that are being added to any process that runs in this group. Because this number can grow quite large if there is a lot of swapping or metadata IO occurring we limit the individual delay events to 1 second at a time.”h]”hX–Artificial delay induction. There are certain types of IO that cannot be throttled without possibly adversely affecting higher priority groups. This includes swapping and metadata IO. These types of IO are allowed to occur normally, however they are â€œchargedâ€ to the originating group. If the originating group is being throttled you will see the use_delay and delay fields in io.stat increase. The delay value is how many microseconds that are being added to any process that runs in this group. Because this number can grow quite large if there is a lot of swapping or metadata IO occurring we limit the individual delay events to 1 second at a time.”…””}”(hj‰9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mnhj…9ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjj9hžhhŸh¯h Nubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h MjhjK9hžhubhæ)”}”(hŒíOnce the victimized group starts meeting its latency target again it will start unthrottling any peer groups that were throttled previously. If the victimized group simply stops doing IO the global counter will unthrottle appropriately.”h]”hŒíOnce the victimized group starts meeting its latency target again it will start unthrottling any peer groups that were throttled previously. If the victimized group simply stops doing IO the global counter will unthrottle appropriately.”…””}”(hj£9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MxhjK9hžhubeh}”(h]”Œhow-io-latency-throttling-works”ah ]”h"]”Œhow io latency throttling works”ah$]”h&]”uh1h°hj².hžhhŸh¯h Mcubh±)”}”(hhh]”(h¶)”}”(hŒIO Latency Interface Files”h]”hŒIO Latency Interface Files”…””}”(hj¼9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj¹9hžhhŸh¯h M}ubj¬)”}”(hX}io.latency This takes a similar format as the other controllers. "MAJOR:MINOR target=" io.stat If the controller is enabled you will see extra stats in io.stat in addition to the normal ones. depth This is the current queue depth for the group. avg_lat This is an exponential moving average with a decay rate of 1/exp bound by the sampling interval. The decay rate interval can be calculated by multiplying the win value in io.stat by the corresponding number of samples based on the win value. win The sampling window size in milliseconds. This is the minimum duration of time between evaluation events. Windows only elapse with IO activity. Idle periods extend the most recent window. ”h]”j²)”}”(hhh]”(j·)”}”(hŒ}io.latency This takes a similar format as the other controllers. "MAJOR:MINOR target=" ”h]”(j½)”}”(hŒ io.latency”h]”hŒ io.latency”…””}”(hjÕ9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M‚hjÑ9ubjÍ)”}”(hhh]”(hæ)”}”(hŒ5This takes a similar format as the other controllers.”h]”hŒ5This takes a similar format as the other controllers.”…””}”(hjæ9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M€hjã9ubj¬)”}”(hŒ3"MAJOR:MINOR target=" ”h]”hæ)”}”(hŒ2"MAJOR:MINOR target="”h]”hŒ6â€œMAJOR:MINOR target=â€”…””}”(hjø9hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‚hjô9ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M‚hjã9ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÑ9ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M‚hjÎ9ubj·)”}”(hX¥io.stat If the controller is enabled you will see extra stats in io.stat in addition to the normal ones. depth This is the current queue depth for the group. avg_lat This is an exponential moving average with a decay rate of 1/exp bound by the sampling interval. The decay rate interval can be calculated by multiplying the win value in io.stat by the corresponding number of samples based on the win value. win The sampling window size in milliseconds. This is the minimum duration of time between evaluation events. Windows only elapse with IO activity. Idle periods extend the most recent window. ”h]”(j½)”}”(hŒio.stat”h]”hŒio.stat”…””}”(hj:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M”hj:ubjÍ)”}”(hhh]”(hæ)”}”(hŒ`If the controller is enabled you will see extra stats in io.stat in addition to the normal ones.”h]”hŒ`If the controller is enabled you will see extra stats in io.stat in addition to the normal ones.”…””}”(hj-:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M…hj*:ubj¬)”}”(hX%depth This is the current queue depth for the group. avg_lat This is an exponential moving average with a decay rate of 1/exp bound by the sampling interval. The decay rate interval can be calculated by multiplying the win value in io.stat by the corresponding number of samples based on the win value. win The sampling window size in milliseconds. This is the minimum duration of time between evaluation events. Windows only elapse with IO activity. Idle periods extend the most recent window. ”h]”j²)”}”(hhh]”(j·)”}”(hŒ5depth This is the current queue depth for the group. ”h]”(j½)”}”(hŒdepth”h]”hŒdepth”…””}”(hjF:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M‰hjB:ubjÍ)”}”(hhh]”hæ)”}”(hŒ.This is the current queue depth for the group.”h]”hŒ.This is the current queue depth for the group.”…””}”(hjW:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‰hjT:ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjB:ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M‰hj?:ubj·)”}”(hŒûavg_lat This is an exponential moving average with a decay rate of 1/exp bound by the sampling interval. The decay rate interval can be calculated by multiplying the win value in io.stat by the corresponding number of samples based on the win value. ”h]”(j½)”}”(hŒavg_lat”h]”hŒavg_lat”…””}”(hju:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhjq:ubjÍ)”}”(hhh]”hæ)”}”(hŒòThis is an exponential moving average with a decay rate of 1/exp bound by the sampling interval. The decay rate interval can be calculated by multiplying the win value in io.stat by the corresponding number of samples based on the win value.”h]”hŒòThis is an exponential moving average with a decay rate of 1/exp bound by the sampling interval. The decay rate interval can be calculated by multiplying the win value in io.stat by the corresponding number of samples based on the win value.”…””}”(hj†:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŒhjƒ:ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjq:ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhj?:ubj·)”}”(hŒÃwin The sampling window size in milliseconds. This is the minimum duration of time between evaluation events. Windows only elapse with IO activity. Idle periods extend the most recent window. ”h]”(j½)”}”(hŒwin”h]”hŒwin”…””}”(hj¤:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M”hj :ubjÍ)”}”(hhh]”hæ)”}”(hŒ¾The sampling window size in milliseconds. This is the minimum duration of time between evaluation events. Windows only elapse with IO activity. Idle periods extend the most recent window.”h]”hŒ¾The sampling window size in milliseconds. This is the minimum duration of time between evaluation events. Windows only elapse with IO activity. Idle periods extend the most recent window.”…””}”(hjµ:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M’hj²:ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj :ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M”hj?:ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj;:ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mˆhj*:ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj:ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M”hjÎ9ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjÊ9ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mhj¹9hžhubeh}”(h]”Œio-latency-interface-files”ah ]”h"]”Œio latency interface files”ah$]”h&]”uh1h°hj².hžhhŸh¯h M}ubh±)”}”(hhh]”(h¶)”}”(hŒIO Priority”h]”hŒIO Priority”…””}”(hjþ:hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjû:hžhhŸh¯h M—ubhæ)”}”(hŒ¥A single attribute controls the behavior of the I/O priority cgroup policy, namely the io.prio.class attribute. The following values are accepted for that attribute:”h]”hŒ¥A single attribute controls the behavior of the I/O priority cgroup policy, namely the io.prio.class attribute. The following values are accepted for that attribute:”…””}”(hj;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M™hjû:hžhubj¬)”}”(hXàno-change Do not modify the I/O priority class. promote-to-rt For requests that have a non-RT I/O priority class, change it into RT. Also change the priority level of these requests to 4. Do not modify the I/O priority of requests that have priority class RT. restrict-to-be For requests that do not have an I/O priority class or that have I/O priority class RT, change it into BE. Also change the priority level of these requests to 0. Do not modify the I/O priority class of requests that have priority class IDLE. idle Change the I/O priority class of all requests into IDLE, the lowest I/O priority class. none-to-rt Deprecated. Just an alias for promote-to-rt. ”h]”j²)”}”(hhh]”(j·)”}”(hŒ0no-change Do not modify the I/O priority class. ”h]”(j½)”}”(hŒ no-change”h]”hŒ no-change”…””}”(hj%;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mžhj!;ubjÍ)”}”(hhh]”hæ)”}”(hŒ%Do not modify the I/O priority class.”h]”hŒ%Do not modify the I/O priority class.”…””}”(hj6;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mžhj3;ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj!;ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mžhj;ubj·)”}”(hŒÔpromote-to-rt For requests that have a non-RT I/O priority class, change it into RT. Also change the priority level of these requests to 4. Do not modify the I/O priority of requests that have priority class RT. ”h]”(j½)”}”(hŒ promote-to-rt”h]”hŒ promote-to-rt”…””}”(hjT;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M£hjP;ubjÍ)”}”(hhh]”hæ)”}”(hŒÅFor requests that have a non-RT I/O priority class, change it into RT. Also change the priority level of these requests to 4. Do not modify the I/O priority of requests that have priority class RT.”h]”hŒÅFor requests that have a non-RT I/O priority class, change it into RT. Also change the priority level of these requests to 4. Do not modify the I/O priority of requests that have priority class RT.”…””}”(hje;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¡hjb;ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjP;ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M£hj;ubj·)”}”(hXrestrict-to-be For requests that do not have an I/O priority class or that have I/O priority class RT, change it into BE. Also change the priority level of these requests to 0. Do not modify the I/O priority class of requests that have priority class IDLE. ”h]”(j½)”}”(hŒrestrict-to-be”h]”hŒrestrict-to-be”…””}”(hjƒ;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M©hj;ubjÍ)”}”(hhh]”hæ)”}”(hŒñFor requests that do not have an I/O priority class or that have I/O priority class RT, change it into BE. Also change the priority level of these requests to 0. Do not modify the I/O priority class of requests that have priority class IDLE.”h]”hŒñFor requests that do not have an I/O priority class or that have I/O priority class RT, change it into BE. Also change the priority level of these requests to 0. Do not modify the I/O priority class of requests that have priority class IDLE.”…””}”(hj”;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¦hj‘;ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj;ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M©hj;ubj·)”}”(hŒ]idle Change the I/O priority class of all requests into IDLE, the lowest I/O priority class. ”h]”(j½)”}”(hŒidle”h]”hŒidle”…””}”(hj²;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mhj®;ubjÍ)”}”(hhh]”hæ)”}”(hŒWChange the I/O priority class of all requests into IDLE, the lowest I/O priority class.”h]”hŒWChange the I/O priority class of all requests into IDLE, the lowest I/O priority class.”…””}”(hjÃ;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¬hjÀ;ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj®;ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mhj;ubj·)”}”(hŒ8none-to-rt Deprecated. Just an alias for promote-to-rt. ”h]”(j½)”}”(hŒ none-to-rt”h]”hŒ none-to-rt”…””}”(hjá;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M°hjÝ;ubjÍ)”}”(hhh]”hæ)”}”(hŒ,Deprecated. Just an alias for promote-to-rt.”h]”hŒ,Deprecated. Just an alias for promote-to-rt.”…””}”(hjò;hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M°hjï;ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÝ;ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M°hj;ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj;ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mhjû:hžhubhæ)”}”(hŒMThe following numerical values are associated with the I/O priority policies:”h]”hŒMThe following numerical values are associated with the I/O priority policies:”…””}”(hj<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M²hjû:hžhubjÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhj)<ubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhj)<ubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ no-change”h]”hŒ no-change”…””}”(hjI<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MµhjF<ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjC<ubj)”}”(hhh]”hæ)”}”(hŒ0”h]”hŒ0”…””}”(hj`<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mµhj]<ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjC<ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj@<ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ promote-to-rt”h]”hŒ promote-to-rt”…””}”(hj€<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M·hj}<ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjz<ubj)”}”(hhh]”hæ)”}”(hŒ1”h]”hŒ1”…””}”(hj—<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M·hj”<ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjz<ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj@<ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒrestrict-to-be”h]”hŒrestrict-to-be”…””}”(hj·<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹hj´<ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj±<ubj)”}”(hhh]”hæ)”}”(hŒ2”h]”hŒ2”…””}”(hjÎ<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹hjË<ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj±<ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj@<ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒidle”h]”hŒidle”…””}”(hjî<hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M»hjë<ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjè<ubj)”}”(hhh]”hæ)”}”(hŒ3”h]”hŒ3”…””}”(hj=hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M»hj=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjè<ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj@<ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühj)<ubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhj&<ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhjû:hžhhŸh¯h Nubhæ)”}”(hŒNThe numerical value that corresponds to each I/O priority class is as follows:”h]”hŒNThe numerical value that corresponds to each I/O priority class is as follows:”…””}”(hj2=hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¾hjû:hžhubjÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhjC=ubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhjC=ubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒIOPRIO_CLASS_NONE”h]”hŒIOPRIO_CLASS_NONE”…””}”(hjc=hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÁhj`=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj]=ubj)”}”(hhh]”hæ)”}”(hjb<h]”hŒ0”…””}”(hjz=hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÁhjw=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj]=ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjZ=ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒIOPRIO_CLASS_RT (real-time)”h]”hŒIOPRIO_CLASS_RT (real-time)”…””}”(hj™=hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÃhj–=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj“=ubj)”}”(hhh]”hæ)”}”(hj™<h]”hŒ1”…””}”(hj°=hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÃhj=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj“=ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjZ=ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒIOPRIO_CLASS_BE (best effort)”h]”hŒIOPRIO_CLASS_BE (best effort)”…””}”(hjÏ=hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÅhjÌ=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÉ=ubj)”}”(hhh]”hæ)”}”(hjÐ<h]”hŒ2”…””}”(hjæ=hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÅhjã=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÉ=ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjZ=ubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒIOPRIO_CLASS_IDLE”h]”hŒIOPRIO_CLASS_IDLE”…””}”(hj>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÇhj>ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÿ=ubj)”}”(hhh]”hæ)”}”(hj=h]”hŒ3”…””}”(hj>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÇhj>ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÿ=ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjZ=ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühjC=ubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhj@=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhjû:hžhhŸh¯h Nubhæ)”}”(hŒHThe algorithm to set the I/O priority class for a request is as follows:”h]”hŒHThe algorithm to set the I/O priority class for a request is as follows:”…””}”(hjH>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÊhjû:hžhubjª)”}”(hhh]”(j¯)”}”(hŒ—If I/O priority class policy is promote-to-rt, change the request I/O priority class to IOPRIO_CLASS_RT and change the request I/O priority level to 4.”h]”hæ)”}”(hŒ—If I/O priority class policy is promote-to-rt, change the request I/O priority class to IOPRIO_CLASS_RT and change the request I/O priority level to 4.”h]”hŒ—If I/O priority class policy is promote-to-rt, change the request I/O priority class to IOPRIO_CLASS_RT and change the request I/O priority level to 4.”…””}”(hj]>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÌhjY>ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjV>hžhhŸh¯h Nubj¯)”}”(hŒôIf I/O priority class policy is not promote-to-rt, translate the I/O priority class policy into a number, then change the request I/O priority class into the maximum of the I/O priority class policy number and the numerical I/O priority class. ”h]”hæ)”}”(hŒóIf I/O priority class policy is not promote-to-rt, translate the I/O priority class policy into a number, then change the request I/O priority class into the maximum of the I/O priority class policy number and the numerical I/O priority class.”h]”hŒóIf I/O priority class policy is not promote-to-rt, translate the I/O priority class policy into a number, then change the request I/O priority class into the maximum of the I/O priority class policy number and the numerical I/O priority class.”…””}”(hju>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÏhjq>ubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjV>hžhhŸh¯h Nubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h MÌhjû:hžhubeh}”(h]”Œio-priority”ah ]”h"]”Œio priority”ah$]”h&]”uh1h°hj².hžhhŸh¯h M—ubeh}”(h]”Œio”ah ]”h"]”Œio”ah$]”h&]”uh1h°hj«hžhhŸh¯h MHubh±)”}”(hhh]”(h¶)”}”(hŒPID”h]”hŒPID”…””}”(hj¢>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjŸ>hžhhŸh¯h MÕubhæ)”}”(hŒ’The process number controller is used to allow a cgroup to stop any new tasks from being fork()'d or clone()'d after a specified limit is reached.”h]”hŒ–The process number controller is used to allow a cgroup to stop any new tasks from being fork()â€™d or clone()â€™d after a specified limit is reached.”…””}”(hj°>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M×hjŸ>hžhubhæ)”}”(hŒëThe number of tasks in a cgroup can be exhausted in ways which other controllers cannot prevent, thus warranting its own controller. For example, a fork bomb is likely to exhaust the number of tasks before hitting memory restrictions.”h]”hŒëThe number of tasks in a cgroup can be exhausted in ways which other controllers cannot prevent, thus warranting its own controller. For example, a fork bomb is likely to exhaust the number of tasks before hitting memory restrictions.”…””}”(hj¾>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÛhjŸ>hžhubhæ)”}”(hŒXNote that PIDs used in this controller refer to TIDs, process IDs as used by the kernel.”h]”hŒXNote that PIDs used in this controller refer to TIDs, process IDs as used by the kernel.”…””}”(hjÌ>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MàhjŸ>hžhubh±)”}”(hhh]”(h¶)”}”(hŒPID Interface Files”h]”hŒPID Interface Files”…””}”(hjÝ>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjÚ>hžhhŸh¯h Måubj¬)”}”(hXDpids.max A read-write single value file which exists on non-root cgroups. The default is "max". Hard limit of number of processes. pids.current A read-only single value file which exists on non-root cgroups. The number of processes currently in the cgroup and its descendants. pids.peak A read-only single value file which exists on non-root cgroups. The maximum value that the number of processes in the cgroup and its descendants has ever reached. pids.events A read-only flat-keyed file which exists on non-root cgroups. Unless specified otherwise, a value change in this file generates a file modified event. The following entries are defined. max The number of times the cgroup's total number of processes hit the pids.max limit (see also pids_localevents). pids.events.local Similar to pids.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events. ”h]”j²)”}”(hhh]”(j·)”}”(hŒ…pids.max A read-write single value file which exists on non-root cgroups. The default is "max". Hard limit of number of processes. ”h]”(j½)”}”(hŒpids.max”h]”hŒpids.max”…””}”(hjö>hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mëhjò>ubjÍ)”}”(hhh]”(hæ)”}”(hŒWA read-write single value file which exists on non-root cgroups. The default is "max".”h]”hŒ[A read-write single value file which exists on non-root cgroups. The default is â€œmaxâ€.”…””}”(hj?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mèhj?ubhæ)”}”(hŒ"Hard limit of number of processes.”h]”hŒ"Hard limit of number of processes.”…””}”(hj?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mëhj?ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjò>ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mëhjï>ubj·)”}”(hŒ“pids.current A read-only single value file which exists on non-root cgroups. The number of processes currently in the cgroup and its descendants. ”h]”(j½)”}”(hŒpids.current”h]”hŒpids.current”…””}”(hj3?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mñhj/?ubjÍ)”}”(hhh]”(hæ)”}”(hŒ?A read-only single value file which exists on non-root cgroups.”h]”hŒ?A read-only single value file which exists on non-root cgroups.”…””}”(hjD?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MîhjA?ubhæ)”}”(hŒDThe number of processes currently in the cgroup and its descendants.”h]”hŒDThe number of processes currently in the cgroup and its descendants.”…””}”(hjR?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MðhjA?ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj/?ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mñhjï>ubj·)”}”(hŒ®pids.peak A read-only single value file which exists on non-root cgroups. The maximum value that the number of processes in the cgroup and its descendants has ever reached. ”h]”(j½)”}”(hŒ pids.peak”h]”hŒ pids.peak”…””}”(hjp?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M÷hjl?ubjÍ)”}”(hhh]”(hæ)”}”(hŒ?A read-only single value file which exists on non-root cgroups.”h]”hŒ?A read-only single value file which exists on non-root cgroups.”…””}”(hj?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Môhj~?ubhæ)”}”(hŒbThe maximum value that the number of processes in the cgroup and its descendants has ever reached.”h]”hŒbThe maximum value that the number of processes in the cgroup and its descendants has ever reached.”…””}”(hj?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Möhj~?ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjl?ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M÷hjï>ubj·)”}”(hXLpids.events A read-only flat-keyed file which exists on non-root cgroups. Unless specified otherwise, a value change in this file generates a file modified event. The following entries are defined. max The number of times the cgroup's total number of processes hit the pids.max limit (see also pids_localevents). ”h]”(j½)”}”(hŒpids.events”h]”hŒpids.events”…””}”(hj?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M hj©?ubjÍ)”}”(hhh]”(hæ)”}”(hŒ¹A read-only flat-keyed file which exists on non-root cgroups. Unless specified otherwise, a value change in this file generates a file modified event. The following entries are defined.”h]”hŒ¹A read-only flat-keyed file which exists on non-root cgroups. Unless specified otherwise, a value change in this file generates a file modified event. The following entries are defined.”…””}”(hj¾?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Múhj»?ubj¬)”}”(hŒmax The number of times the cgroup's total number of processes hit the pids.max limit (see also pids_localevents). ”h]”j²)”}”(hhh]”j·)”}”(hŒsmax The number of times the cgroup's total number of processes hit the pids.max limit (see also pids_localevents). ”h]”(j½)”}”(hŒmax”h]”hŒmax”…””}”(hj×?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M hjÓ?ubjÍ)”}”(hhh]”hæ)”}”(hŒnThe number of times the cgroup's total number of processes hit the pids.max limit (see also pids_localevents).”h]”hŒpThe number of times the cgroupâ€™s total number of processes hit the pids.max limit (see also pids_localevents).”…””}”(hjè?hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mÿhjå?ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÓ?ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M hjÐ?ubah}”(h]”h ]”h"]”h$]”h&]”uh1j±hjÌ?ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mþhj»?ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj©?ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M hjï>ubj·)”}”(hŒÂpids.events.local Similar to pids.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events. ”h]”(j½)”}”(hŒpids.events.local”h]”hŒpids.events.local”…””}”(hj@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M hj@ubjÍ)”}”(hhh]”hæ)”}”(hŒ¯Similar to pids.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events.”h]”hŒ¯Similar to pids.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events.”…””}”(hj/@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hj,@ubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj@ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M hjï>ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjë>ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MçhjÚ>hžhubhæ)”}”(hXÙOrganisational operations are not blocked by cgroup policies, so it is possible to have pids.current > pids.max. This can be done by either setting the limit to be smaller than pids.current, or attaching enough processes to the cgroup such that pids.current is larger than pids.max. However, it is not possible to violate a cgroup PID policy through fork() or clone(). These will return -EAGAIN if the creation of a new process would cause a cgroup policy to be violated.”h]”hXÙOrganisational operations are not blocked by cgroup policies, so it is possible to have pids.current > pids.max. This can be done by either setting the limit to be smaller than pids.current, or attaching enough processes to the cgroup such that pids.current is larger than pids.max. However, it is not possible to violate a cgroup PID policy through fork() or clone(). These will return -EAGAIN if the creation of a new process would cause a cgroup policy to be violated.”…””}”(hjU@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjÚ>hžhubeh}”(h]”Œpid-interface-files”ah ]”h"]”Œpid interface files”ah$]”h&]”uh1h°hjŸ>hžhhŸh¯h Måubeh}”(h]”Œpid”ah ]”h"]”Œpid”ah$]”h&]”uh1h°hj«hžhhŸh¯h MÕubh±)”}”(hhh]”(h¶)”}”(hŒCpuset”h]”hŒCpuset”…””}”(hjv@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjs@hžhhŸh¯h M ubhæ)”}”(hX¶The "cpuset" controller provides a mechanism for constraining the CPU and memory node placement of tasks to only the resources specified in the cpuset interface files in a task's current cgroup. This is especially valuable on large NUMA systems where placing jobs on properly sized subsets of the systems with careful processor and memory placement to reduce cross-node memory access and contention can improve overall system performance.”h]”hX¼The â€œcpusetâ€ controller provides a mechanism for constraining the CPU and memory node placement of tasks to only the resources specified in the cpuset interface files in a taskâ€™s current cgroup. This is especially valuable on large NUMA systems where placing jobs on properly sized subsets of the systems with careful processor and memory placement to reduce cross-node memory access and contention can improve overall system performance.”…””}”(hj„@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjs@hžhubhæ)”}”(hŒ~The "cpuset" controller is hierarchical. That means the controller cannot use CPUs or memory nodes not allowed in its parent.”h]”hŒ‚The â€œcpusetâ€ controller is hierarchical. That means the controller cannot use CPUs or memory nodes not allowed in its parent.”…””}”(hj’@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjs@hžhubh±)”}”(hhh]”(h¶)”}”(hŒCpuset Interface Files”h]”hŒCpuset Interface Files”…””}”(hj£@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj @hžhhŸh¯h M ubj¬)”}”(hXn-cpuset.cpus A read-write multiple values file which exists on non-root cpuset-enabled cgroups. It lists the requested CPUs to be used by tasks within this cgroup. The actual list of CPUs to be granted, however, is subjected to constraints imposed by its parent and can differ from the requested CPUs. The CPU numbers are comma-separated numbers or ranges. For example:: # cat cpuset.cpus 0-4,6,8-10 An empty value indicates that the cgroup is using the same setting as the nearest cgroup ancestor with a non-empty "cpuset.cpus" or all the available CPUs if none is found. The value of "cpuset.cpus" stays constant until the next update and won't be affected by any CPU hotplug events. cpuset.cpus.effective A read-only multiple values file which exists on all cpuset-enabled cgroups. It lists the onlined CPUs that are actually granted to this cgroup by its parent. These CPUs are allowed to be used by tasks within the current cgroup. If "cpuset.cpus" is empty, the "cpuset.cpus.effective" file shows all the CPUs from the parent cgroup that can be available to be used by this cgroup. Otherwise, it should be a subset of "cpuset.cpus" unless none of the CPUs listed in "cpuset.cpus" can be granted. In this case, it will be treated just like an empty "cpuset.cpus". Its value will be affected by CPU hotplug events. cpuset.mems A read-write multiple values file which exists on non-root cpuset-enabled cgroups. It lists the requested memory nodes to be used by tasks within this cgroup. The actual list of memory nodes granted, however, is subjected to constraints imposed by its parent and can differ from the requested memory nodes. The memory node numbers are comma-separated numbers or ranges. For example:: # cat cpuset.mems 0-1,3 An empty value indicates that the cgroup is using the same setting as the nearest cgroup ancestor with a non-empty "cpuset.mems" or all the available memory nodes if none is found. The value of "cpuset.mems" stays constant until the next update and won't be affected by any memory nodes hotplug events. Setting a non-empty value to "cpuset.mems" causes memory of tasks within the cgroup to be migrated to the designated nodes if they are currently using memory outside of the designated nodes. There is a cost for this memory migration. The migration may not be complete and some memory pages may be left behind. So it is recommended that "cpuset.mems" should be set properly before spawning new tasks into the cpuset. Even if there is a need to change "cpuset.mems" with active tasks, it shouldn't be done frequently. cpuset.mems.effective A read-only multiple values file which exists on all cpuset-enabled cgroups. It lists the onlined memory nodes that are actually granted to this cgroup by its parent. These memory nodes are allowed to be used by tasks within the current cgroup. If "cpuset.mems" is empty, it shows all the memory nodes from the parent cgroup that will be available to be used by this cgroup. Otherwise, it should be a subset of "cpuset.mems" unless none of the memory nodes listed in "cpuset.mems" can be granted. In this case, it will be treated just like an empty "cpuset.mems". Its value will be affected by memory nodes hotplug events. cpuset.cpus.exclusive A read-write multiple values file which exists on non-root cpuset-enabled cgroups. It lists all the exclusive CPUs that are allowed to be used to create a new cpuset partition. Its value is not used unless the cgroup becomes a valid partition root. See the "cpuset.cpus.partition" section below for a description of what a cpuset partition is. When the cgroup becomes a partition root, the actual exclusive CPUs that are allocated to that partition are listed in "cpuset.cpus.exclusive.effective" which may be different from "cpuset.cpus.exclusive". If "cpuset.cpus.exclusive" has previously been set, "cpuset.cpus.exclusive.effective" is always a subset of it. Users can manually set it to a value that is different from "cpuset.cpus". One constraint in setting it is that the list of CPUs must be exclusive with respect to "cpuset.cpus.exclusive" of its sibling. If "cpuset.cpus.exclusive" of a sibling cgroup isn't set, its "cpuset.cpus" value, if set, cannot be a subset of it to leave at least one CPU available when the exclusive CPUs are taken away. For a parent cgroup, any one of its exclusive CPUs can only be distributed to at most one of its child cgroups. Having an exclusive CPU appearing in two or more of its child cgroups is not allowed (the exclusivity rule). A value that violates the exclusivity rule will be rejected with a write error. The root cgroup is a partition root and all its available CPUs are in its exclusive CPU set. cpuset.cpus.exclusive.effective A read-only multiple values file which exists on all non-root cpuset-enabled cgroups. This file shows the effective set of exclusive CPUs that can be used to create a partition root. The content of this file will always be a subset of its parent's "cpuset.cpus.exclusive.effective" if its parent is not the root cgroup. It will also be a subset of "cpuset.cpus.exclusive" if it is set. If "cpuset.cpus.exclusive" is not set, it is treated to have an implicit value of "cpuset.cpus" in the formation of local partition. cpuset.cpus.isolated A read-only and root cgroup only multiple values file. This file shows the set of all isolated CPUs used in existing isolated partitions. It will be empty if no isolated partition is created. cpuset.cpus.partition A read-write single value file which exists on non-root cpuset-enabled cgroups. This flag is owned by the parent cgroup and is not delegatable. It accepts only the following input values when written to. ========== ===================================== "member" Non-root member of a partition "root" Partition root "isolated" Partition root without load balancing ========== ===================================== A cpuset partition is a collection of cpuset-enabled cgroups with a partition root at the top of the hierarchy and its descendants except those that are separate partition roots themselves and their descendants. A partition has exclusive access to the set of exclusive CPUs allocated to it. Other cgroups outside of that partition cannot use any CPUs in that set. There are two types of partitions - local and remote. A local partition is one whose parent cgroup is also a valid partition root. A remote partition is one whose parent cgroup is not a valid partition root itself. Writing to "cpuset.cpus.exclusive" is optional for the creation of a local partition as its "cpuset.cpus.exclusive" file will assume an implicit value that is the same as "cpuset.cpus" if it is not set. Writing the proper "cpuset.cpus.exclusive" values down the cgroup hierarchy before the target partition root is mandatory for the creation of a remote partition. Currently, a remote partition cannot be created under a local partition. All the ancestors of a remote partition root except the root cgroup cannot be a partition root. The root cgroup is always a partition root and its state cannot be changed. All other non-root cgroups start out as "member". When set to "root", the current cgroup is the root of a new partition or scheduling domain. The set of exclusive CPUs is determined by the value of its "cpuset.cpus.exclusive.effective". When set to "isolated", the CPUs in that partition will be in an isolated state without any load balancing from the scheduler and excluded from the unbound workqueues. Tasks placed in such a partition with multiple CPUs should be carefully distributed and bound to each of the individual CPUs for optimal performance. A partition root ("root" or "isolated") can be in one of the two possible states - valid or invalid. An invalid partition root is in a degraded state where some state information may be retained, but behaves more like a "member". All possible state transitions among "member", "root" and "isolated" are allowed. On read, the "cpuset.cpus.partition" file can show the following values. ============================= ===================================== "member" Non-root member of a partition "root" Partition root "isolated" Partition root without load balancing "root invalid ()" Invalid partition root "isolated invalid ()" Invalid isolated partition root ============================= ===================================== In the case of an invalid partition root, a descriptive string on why the partition is invalid is included within parentheses. For a local partition root to be valid, the following conditions must be met. 1) The parent cgroup is a valid partition root. 2) The "cpuset.cpus.exclusive.effective" file cannot be empty, though it may contain offline CPUs. 3) The "cpuset.cpus.effective" cannot be empty unless there is no task associated with this partition. For a remote partition root to be valid, all the above conditions except the first one must be met. External events like hotplug or changes to "cpuset.cpus" or "cpuset.cpus.exclusive" can cause a valid partition root to become invalid and vice versa. Note that a task cannot be moved to a cgroup with empty "cpuset.cpus.effective". A valid non-root parent partition may distribute out all its CPUs to its child local partitions when there is no task associated with it. Care must be taken to change a valid partition root to "member" as all its child local partitions, if present, will become invalid causing disruption to tasks running in those child partitions. These inactivated partitions could be recovered if their parent is switched back to a partition root with a proper value in "cpuset.cpus" or "cpuset.cpus.exclusive". Poll and inotify events are triggered whenever the state of "cpuset.cpus.partition" changes. That includes changes caused by write to "cpuset.cpus.partition", cpu hotplug or other changes that modify the validity status of the partition. This will allow user space agents to monitor unexpected changes to "cpuset.cpus.partition" without the need to do continuous polling. A user can pre-configure certain CPUs to an isolated state with load balancing disabled at boot time with the "isolcpus" kernel boot command line option. If those CPUs are to be put into a partition, they have to be used in an isolated partition. ”•‹h]”j²)”}”(hhh]”(j·)”}”(hX·cpuset.cpus A read-write multiple values file which exists on non-root cpuset-enabled cgroups. It lists the requested CPUs to be used by tasks within this cgroup. The actual list of CPUs to be granted, however, is subjected to constraints imposed by its parent and can differ from the requested CPUs. The CPU numbers are comma-separated numbers or ranges. For example:: # cat cpuset.cpus 0-4,6,8-10 An empty value indicates that the cgroup is using the same setting as the nearest cgroup ancestor with a non-empty "cpuset.cpus" or all the available CPUs if none is found. The value of "cpuset.cpus" stays constant until the next update and won't be affected by any CPU hotplug events. ”h]”(j½)”}”(hŒcpuset.cpus”h]”hŒcpuset.cpus”…””}”(hj¼@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M6 hj¸@ubjÍ)”}”(hhh]”(hæ)”}”(hŒRA read-write multiple values file which exists on non-root cpuset-enabled cgroups.”h]”hŒRA read-write multiple values file which exists on non-root cpuset-enabled cgroups.”…””}”(hjÍ@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M# hjÊ@ubhæ)”}”(hŒÎIt lists the requested CPUs to be used by tasks within this cgroup. The actual list of CPUs to be granted, however, is subjected to constraints imposed by its parent and can differ from the requested CPUs.”h]”hŒÎIt lists the requested CPUs to be used by tasks within this cgroup. The actual list of CPUs to be granted, however, is subjected to constraints imposed by its parent and can differ from the requested CPUs.”…””}”(hjÛ@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M& hjÊ@ubhæ)”}”(hŒDThe CPU numbers are comma-separated numbers or ranges. For example::”h]”hŒCThe CPU numbers are comma-separated numbers or ranges. For example:”…””}”(hjé@hžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M+ hjÊ@ubjV)”}”(hŒ# cat cpuset.cpus 0-4,6,8-10”h]”hŒ# cat cpuset.cpus 0-4,6,8-10”…””}”hj÷@sbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M. hjÊ@ubhæ)”}”(hŒ¬An empty value indicates that the cgroup is using the same setting as the nearest cgroup ancestor with a non-empty "cpuset.cpus" or all the available CPUs if none is found.”h]”hŒ°An empty value indicates that the cgroup is using the same setting as the nearest cgroup ancestor with a non-empty â€œcpuset.cpusâ€ or all the available CPUs if none is found.”…””}”(hjAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M1 hjÊ@ubhæ)”}”(hŒpThe value of "cpuset.cpus" stays constant until the next update and won't be affected by any CPU hotplug events.”h]”hŒvThe value of â€œcpuset.cpusâ€ stays constant until the next update and wonâ€™t be affected by any CPU hotplug events.”…””}”(hjAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M5 hjÊ@ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj¸@ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M6 hjµ@ubj·)”}”(hXcpuset.cpus.effective A read-only multiple values file which exists on all cpuset-enabled cgroups. It lists the onlined CPUs that are actually granted to this cgroup by its parent. These CPUs are allowed to be used by tasks within the current cgroup. If "cpuset.cpus" is empty, the "cpuset.cpus.effective" file shows all the CPUs from the parent cgroup that can be available to be used by this cgroup. Otherwise, it should be a subset of "cpuset.cpus" unless none of the CPUs listed in "cpuset.cpus" can be granted. In this case, it will be treated just like an empty "cpuset.cpus". Its value will be affected by CPU hotplug events. ”h]”(j½)”}”(hŒcpuset.cpus.effective”h]”hŒcpuset.cpus.effective”…””}”(hj1AhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MG hj-AubjÍ)”}”(hhh]”(hæ)”}”(hŒLA read-only multiple values file which exists on all cpuset-enabled cgroups.”h]”hŒLA read-only multiple values file which exists on all cpuset-enabled cgroups.”…””}”(hjBAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M9 hj?Aubhæ)”}”(hŒ˜It lists the onlined CPUs that are actually granted to this cgroup by its parent. These CPUs are allowed to be used by tasks within the current cgroup.”h]”hŒ˜It lists the onlined CPUs that are actually granted to this cgroup by its parent. These CPUs are allowed to be used by tasks within the current cgroup.”…””}”(hjPAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M< hj?Aubhæ)”}”(hXMIf "cpuset.cpus" is empty, the "cpuset.cpus.effective" file shows all the CPUs from the parent cgroup that can be available to be used by this cgroup. Otherwise, it should be a subset of "cpuset.cpus" unless none of the CPUs listed in "cpuset.cpus" can be granted. In this case, it will be treated just like an empty "cpuset.cpus".”h]”hXaIf â€œcpuset.cpusâ€ is empty, the â€œcpuset.cpus.effectiveâ€ file shows all the CPUs from the parent cgroup that can be available to be used by this cgroup. Otherwise, it should be a subset of â€œcpuset.cpusâ€ unless none of the CPUs listed in â€œcpuset.cpusâ€ can be granted. In this case, it will be treated just like an empty â€œcpuset.cpusâ€.”…””}”(hj^AhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M@ hj?Aubhæ)”}”(hŒ1Its value will be affected by CPU hotplug events.”h]”hŒ1Its value will be affected by CPU hotplug events.”…””}”(hjlAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MG hj?Aubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj-Aubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MG hjµ@ubj·)”}”(hXåcpuset.mems A read-write multiple values file which exists on non-root cpuset-enabled cgroups. It lists the requested memory nodes to be used by tasks within this cgroup. The actual list of memory nodes granted, however, is subjected to constraints imposed by its parent and can differ from the requested memory nodes. The memory node numbers are comma-separated numbers or ranges. For example:: # cat cpuset.mems 0-1,3 An empty value indicates that the cgroup is using the same setting as the nearest cgroup ancestor with a non-empty "cpuset.mems" or all the available memory nodes if none is found. The value of "cpuset.mems" stays constant until the next update and won't be affected by any memory nodes hotplug events. Setting a non-empty value to "cpuset.mems" causes memory of tasks within the cgroup to be migrated to the designated nodes if they are currently using memory outside of the designated nodes. There is a cost for this memory migration. The migration may not be complete and some memory pages may be left behind. So it is recommended that "cpuset.mems" should be set properly before spawning new tasks into the cpuset. Even if there is a need to change "cpuset.mems" with active tasks, it shouldn't be done frequently. ”h]”(j½)”}”(hŒcpuset.mems”h]”hŒcpuset.mems”…””}”(hjŠAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mi hj†AubjÍ)”}”(hhh]”(hæ)”}”(hŒRA read-write multiple values file which exists on non-root cpuset-enabled cgroups.”h]”hŒRA read-write multiple values file which exists on non-root cpuset-enabled cgroups.”…””}”(hj›AhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MJ hj˜Aubhæ)”}”(hŒàIt lists the requested memory nodes to be used by tasks within this cgroup. The actual list of memory nodes granted, however, is subjected to constraints imposed by its parent and can differ from the requested memory nodes.”h]”hŒàIt lists the requested memory nodes to be used by tasks within this cgroup. The actual list of memory nodes granted, however, is subjected to constraints imposed by its parent and can differ from the requested memory nodes.”…””}”(hj©AhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MM hj˜Aubhæ)”}”(hŒLThe memory node numbers are comma-separated numbers or ranges. For example::”h]”hŒKThe memory node numbers are comma-separated numbers or ranges. For example:”…””}”(hj·AhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MR hj˜AubjV)”}”(hŒ# cat cpuset.mems 0-1,3”h]”hŒ# cat cpuset.mems 0-1,3”…””}”hjÅAsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MU hj˜Aubhæ)”}”(hŒ´An empty value indicates that the cgroup is using the same setting as the nearest cgroup ancestor with a non-empty "cpuset.mems" or all the available memory nodes if none is found.”h]”hŒ¸An empty value indicates that the cgroup is using the same setting as the nearest cgroup ancestor with a non-empty â€œcpuset.memsâ€ or all the available memory nodes if none is found.”…””}”(hjÓAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MX hj˜Aubhæ)”}”(hŒyThe value of "cpuset.mems" stays constant until the next update and won't be affected by any memory nodes hotplug events.”h]”hŒThe value of â€œcpuset.memsâ€ stays constant until the next update and wonâ€™t be affected by any memory nodes hotplug events.”…””}”(hjáAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M] hj˜Aubhæ)”}”(hŒ¾Setting a non-empty value to "cpuset.mems" causes memory of tasks within the cgroup to be migrated to the designated nodes if they are currently using memory outside of the designated nodes.”h]”hŒÂSetting a non-empty value to â€œcpuset.memsâ€ causes memory of tasks within the cgroup to be migrated to the designated nodes if they are currently using memory outside of the designated nodes.”…””}”(hjïAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M` hj˜Aubhæ)”}”(hXFThere is a cost for this memory migration. The migration may not be complete and some memory pages may be left behind. So it is recommended that "cpuset.mems" should be set properly before spawning new tasks into the cpuset. Even if there is a need to change "cpuset.mems" with active tasks, it shouldn't be done frequently.”h]”hXPThere is a cost for this memory migration. The migration may not be complete and some memory pages may be left behind. So it is recommended that â€œcpuset.memsâ€ should be set properly before spawning new tasks into the cpuset. Even if there is a need to change â€œcpuset.memsâ€ with active tasks, it shouldnâ€™t be done frequently.”…””}”(hjýAhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Md hj˜Aubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj†Aubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mi hjµ@ubj·)”}”(hX‰cpuset.mems.effective A read-only multiple values file which exists on all cpuset-enabled cgroups. It lists the onlined memory nodes that are actually granted to this cgroup by its parent. These memory nodes are allowed to be used by tasks within the current cgroup. If "cpuset.mems" is empty, it shows all the memory nodes from the parent cgroup that will be available to be used by this cgroup. Otherwise, it should be a subset of "cpuset.mems" unless none of the memory nodes listed in "cpuset.mems" can be granted. In this case, it will be treated just like an empty "cpuset.mems". Its value will be affected by memory nodes hotplug events. ”h]”(j½)”}”(hŒcpuset.mems.effective”h]”hŒcpuset.mems.effective”…””}”(hjBhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h My hjBubjÍ)”}”(hhh]”(hæ)”}”(hŒLA read-only multiple values file which exists on all cpuset-enabled cgroups.”h]”hŒLA read-only multiple values file which exists on all cpuset-enabled cgroups.”…””}”(hj,BhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Ml hj)Bubhæ)”}”(hŒ§It lists the onlined memory nodes that are actually granted to this cgroup by its parent. These memory nodes are allowed to be used by tasks within the current cgroup.”h]”hŒ§It lists the onlined memory nodes that are actually granted to this cgroup by its parent. These memory nodes are allowed to be used by tasks within the current cgroup.”…””}”(hj:BhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mo hj)Bubhæ)”}”(hX?If "cpuset.mems" is empty, it shows all the memory nodes from the parent cgroup that will be available to be used by this cgroup. Otherwise, it should be a subset of "cpuset.mems" unless none of the memory nodes listed in "cpuset.mems" can be granted. In this case, it will be treated just like an empty "cpuset.mems".”h]”hXOIf â€œcpuset.memsâ€ is empty, it shows all the memory nodes from the parent cgroup that will be available to be used by this cgroup. Otherwise, it should be a subset of â€œcpuset.memsâ€ unless none of the memory nodes listed in â€œcpuset.memsâ€ can be granted. In this case, it will be treated just like an empty â€œcpuset.memsâ€.”…””}”(hjHBhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Ms hj)Bubhæ)”}”(hŒ:Its value will be affected by memory nodes hotplug events.”h]”hŒ:Its value will be affected by memory nodes hotplug events.”…””}”(hjVBhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h My hj)Bubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjBubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h My hjµ@ubj·)”}”(hXÍcpuset.cpus.exclusive A read-write multiple values file which exists on non-root cpuset-enabled cgroups. It lists all the exclusive CPUs that are allowed to be used to create a new cpuset partition. Its value is not used unless the cgroup becomes a valid partition root. See the "cpuset.cpus.partition" section below for a description of what a cpuset partition is. When the cgroup becomes a partition root, the actual exclusive CPUs that are allocated to that partition are listed in "cpuset.cpus.exclusive.effective" which may be different from "cpuset.cpus.exclusive". If "cpuset.cpus.exclusive" has previously been set, "cpuset.cpus.exclusive.effective" is always a subset of it. Users can manually set it to a value that is different from "cpuset.cpus". One constraint in setting it is that the list of CPUs must be exclusive with respect to "cpuset.cpus.exclusive" of its sibling. If "cpuset.cpus.exclusive" of a sibling cgroup isn't set, its "cpuset.cpus" value, if set, cannot be a subset of it to leave at least one CPU available when the exclusive CPUs are taken away. For a parent cgroup, any one of its exclusive CPUs can only be distributed to at most one of its child cgroups. Having an exclusive CPU appearing in two or more of its child cgroups is not allowed (the exclusivity rule). A value that violates the exclusivity rule will be rejected with a write error. The root cgroup is a partition root and all its available CPUs are in its exclusive CPU set. ”h]”(j½)”}”(hŒcpuset.cpus.exclusive”h]”hŒcpuset.cpus.exclusive”…””}”(hjtBhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M› hjpBubjÍ)”}”(hhh]”(hæ)”}”(hŒRA read-write multiple values file which exists on non-root cpuset-enabled cgroups.”h]”hŒRA read-write multiple values file which exists on non-root cpuset-enabled cgroups.”…””}”(hj…BhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M| hj‚Bubhæ)”}”(hXIt lists all the exclusive CPUs that are allowed to be used to create a new cpuset partition. Its value is not used unless the cgroup becomes a valid partition root. See the "cpuset.cpus.partition" section below for a description of what a cpuset partition is.”h]”hX It lists all the exclusive CPUs that are allowed to be used to create a new cpuset partition. Its value is not used unless the cgroup becomes a valid partition root. See the â€œcpuset.cpus.partitionâ€ section below for a description of what a cpuset partition is.”…””}”(hj“BhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hj‚Bubhæ)”}”(hX>When the cgroup becomes a partition root, the actual exclusive CPUs that are allocated to that partition are listed in "cpuset.cpus.exclusive.effective" which may be different from "cpuset.cpus.exclusive". If "cpuset.cpus.exclusive" has previously been set, "cpuset.cpus.exclusive.effective" is always a subset of it.”h]”hXNWhen the cgroup becomes a partition root, the actual exclusive CPUs that are allocated to that partition are listed in â€œcpuset.cpus.exclusive.effectiveâ€ which may be different from â€œcpuset.cpus.exclusiveâ€. If â€œcpuset.cpus.exclusiveâ€ has previously been set, â€œcpuset.cpus.exclusive.effectiveâ€ is always a subset of it.”…””}”(hj¡BhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M… hj‚Bubhæ)”}”(hXŒUsers can manually set it to a value that is different from "cpuset.cpus". One constraint in setting it is that the list of CPUs must be exclusive with respect to "cpuset.cpus.exclusive" of its sibling. If "cpuset.cpus.exclusive" of a sibling cgroup isn't set, its "cpuset.cpus" value, if set, cannot be a subset of it to leave at least one CPU available when the exclusive CPUs are taken away.”h]”hXžUsers can manually set it to a value that is different from â€œcpuset.cpusâ€. One constraint in setting it is that the list of CPUs must be exclusive with respect to â€œcpuset.cpus.exclusiveâ€ of its sibling. If â€œcpuset.cpus.exclusiveâ€ of a sibling cgroup isnâ€™t set, its â€œcpuset.cpusâ€ value, if set, cannot be a subset of it to leave at least one CPU available when the exclusive CPUs are taken away.”…””}”(hj¯BhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŒ hj‚Bubhæ)”}”(hX.For a parent cgroup, any one of its exclusive CPUs can only be distributed to at most one of its child cgroups. Having an exclusive CPU appearing in two or more of its child cgroups is not allowed (the exclusivity rule). A value that violates the exclusivity rule will be rejected with a write error.”h]”hX.For a parent cgroup, any one of its exclusive CPUs can only be distributed to at most one of its child cgroups. Having an exclusive CPU appearing in two or more of its child cgroups is not allowed (the exclusivity rule). A value that violates the exclusivity rule will be rejected with a write error.”…””}”(hj½BhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M” hj‚Bubhæ)”}”(hŒ\The root cgroup is a partition root and all its available CPUs are in its exclusive CPU set.”h]”hŒ\The root cgroup is a partition root and all its available CPUs are in its exclusive CPU set.”…””}”(hjËBhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mš hj‚Bubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjpBubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M› hjµ@ubj·)”}”(hX+cpuset.cpus.exclusive.effective A read-only multiple values file which exists on all non-root cpuset-enabled cgroups. This file shows the effective set of exclusive CPUs that can be used to create a partition root. The content of this file will always be a subset of its parent's "cpuset.cpus.exclusive.effective" if its parent is not the root cgroup. It will also be a subset of "cpuset.cpus.exclusive" if it is set. If "cpuset.cpus.exclusive" is not set, it is treated to have an implicit value of "cpuset.cpus" in the formation of local partition. ”h]”(j½)”}”(hŒcpuset.cpus.exclusive.effective”h]”hŒcpuset.cpus.exclusive.effective”…””}”(hjéBhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¨ hjåBubjÍ)”}”(hhh]”(hæ)”}”(hŒUA read-only multiple values file which exists on all non-root cpuset-enabled cgroups.”h]”hŒUA read-only multiple values file which exists on all non-root cpuset-enabled cgroups.”…””}”(hjúBhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mž hj÷Bubhæ)”}”(hX³This file shows the effective set of exclusive CPUs that can be used to create a partition root. The content of this file will always be a subset of its parent's "cpuset.cpus.exclusive.effective" if its parent is not the root cgroup. It will also be a subset of "cpuset.cpus.exclusive" if it is set. If "cpuset.cpus.exclusive" is not set, it is treated to have an implicit value of "cpuset.cpus" in the formation of local partition.”h]”hXÅThis file shows the effective set of exclusive CPUs that can be used to create a partition root. The content of this file will always be a subset of its parentâ€™s â€œcpuset.cpus.exclusive.effectiveâ€ if its parent is not the root cgroup. It will also be a subset of â€œcpuset.cpus.exclusiveâ€ if it is set. If â€œcpuset.cpus.exclusiveâ€ is not set, it is treated to have an implicit value of â€œcpuset.cpusâ€ in the formation of local partition.”…””}”(hjChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¡ hj÷Bubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjåBubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¨ hjµ@ubj·)”}”(hŒÖcpuset.cpus.isolated A read-only and root cgroup only multiple values file. This file shows the set of all isolated CPUs used in existing isolated partitions. It will be empty if no isolated partition is created. ”h]”(j½)”}”(hŒcpuset.cpus.isolated”h]”hŒcpuset.cpus.isolated”…””}”(hj&ChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¯ hj"CubjÍ)”}”(hhh]”(hæ)”}”(hŒ6A read-only and root cgroup only multiple values file.”h]”hŒ6A read-only and root cgroup only multiple values file.”…””}”(hj7ChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M« hj4Cubhæ)”}”(hŒˆThis file shows the set of all isolated CPUs used in existing isolated partitions. It will be empty if no isolated partition is created.”h]”hŒˆThis file shows the set of all isolated CPUs used in existing isolated partitions. It will be empty if no isolated partition is created.”…””}”(hjEChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hj4Cubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj"Cubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¯ hjµ@ubj·)”}”(hXucpuset.cpus.partition A read-write single value file which exists on non-root cpuset-enabled cgroups. This flag is owned by the parent cgroup and is not delegatable. It accepts only the following input values when written to. ========== ===================================== "member" Non-root member of a partition "root" Partition root "isolated" Partition root without load balancing ========== ===================================== A cpuset partition is a collection of cpuset-enabled cgroups with a partition root at the top of the hierarchy and its descendants except those that are separate partition roots themselves and their descendants. A partition has exclusive access to the set of exclusive CPUs allocated to it. Other cgroups outside of that partition cannot use any CPUs in that set. There are two types of partitions - local and remote. A local partition is one whose parent cgroup is also a valid partition root. A remote partition is one whose parent cgroup is not a valid partition root itself. Writing to "cpuset.cpus.exclusive" is optional for the creation of a local partition as its "cpuset.cpus.exclusive" file will assume an implicit value that is the same as "cpuset.cpus" if it is not set. Writing the proper "cpuset.cpus.exclusive" values down the cgroup hierarchy before the target partition root is mandatory for the creation of a remote partition. Currently, a remote partition cannot be created under a local partition. All the ancestors of a remote partition root except the root cgroup cannot be a partition root. The root cgroup is always a partition root and its state cannot be changed. All other non-root cgroups start out as "member". When set to "root", the current cgroup is the root of a new partition or scheduling domain. The set of exclusive CPUs is determined by the value of its "cpuset.cpus.exclusive.effective". When set to "isolated", the CPUs in that partition will be in an isolated state without any load balancing from the scheduler and excluded from the unbound workqueues. Tasks placed in such a partition with multiple CPUs should be carefully distributed and bound to each of the individual CPUs for optimal performance. A partition root ("root" or "isolated") can be in one of the two possible states - valid or invalid. An invalid partition root is in a degraded state where some state information may be retained, but behaves more like a "member". All possible state transitions among "member", "root" and "isolated" are allowed. On read, the "cpuset.cpus.partition" file can show the following values. ============================= ===================================== "member" Non-root member of a partition "root" Partition root "isolated" Partition root without load balancing "root invalid ()" Invalid partition root "isolated invalid ()" Invalid isolated partition root ============================= ===================================== In the case of an invalid partition root, a descriptive string on why the partition is invalid is included within parentheses. For a local partition root to be valid, the following conditions must be met. 1) The parent cgroup is a valid partition root. 2) The "cpuset.cpus.exclusive.effective" file cannot be empty, though it may contain offline CPUs. 3) The "cpuset.cpus.effective" cannot be empty unless there is no task associated with this partition. For a remote partition root to be valid, all the above conditions except the first one must be met. External events like hotplug or changes to "cpuset.cpus" or "cpuset.cpus.exclusive" can cause a valid partition root to become invalid and vice versa. Note that a task cannot be moved to a cgroup with empty "cpuset.cpus.effective". A valid non-root parent partition may distribute out all its CPUs to its child local partitions when there is no task associated with it. Care must be taken to change a valid partition root to "member" as all its child local partitions, if present, will become invalid causing disruption to tasks running in those child partitions. These inactivated partitions could be recovered if their parent is switched back to a partition root with a proper value in "cpuset.cpus" or "cpuset.cpus.exclusive". Poll and inotify events are triggered whenever the state of "cpuset.cpus.partition" changes. That includes changes caused by write to "cpuset.cpus.partition", cpu hotplug or other changes that modify the validity status of the partition. This will allow user space agents to monitor unexpected changes to "cpuset.cpus.partition" without the need to do continuous polling. A user can pre-configure certain CPUs to an isolated state with load balancing disabled at boot time with the "isolcpus" kernel boot command line option. If those CPUs are to be put into a partition, they have to be used in an isolated partition. ”h]”(j½)”}”(hŒcpuset.cpus.partition”h]”hŒcpuset.cpus.partition”…””}”(hjcChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M hj_CubjÍ)”}”(hhh]”(hæ)”}”(hŒA read-write single value file which exists on non-root cpuset-enabled cgroups. This flag is owned by the parent cgroup and is not delegatable.”h]”hŒA read-write single value file which exists on non-root cpuset-enabled cgroups. This flag is owned by the parent cgroup and is not delegatable.”…””}”(hjtChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M² hjqCubhæ)”}”(hŒ;It accepts only the following input values when written to.”h]”hŒ;It accepts only the following input values when written to.”…””}”(hj‚ChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¶ hjqCubj¬)”}”(hŒæ========== ===================================== "member" Non-root member of a partition "root" Partition root "isolated" Partition root without load balancing ========== ===================================== ”h]”jÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K uh1jæhj—Cubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K%uh1jæhj—Cubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ"member"”h]”hŒâ€œmemberâ€”…””}”(hj·ChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹ hj´Cubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj±Cubj)”}”(hhh]”hæ)”}”(hŒNon-root member of a partition”h]”hŒNon-root member of a partition”…””}”(hjÎChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹ hjËCubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj±Cubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj®Cubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ"root"”h]”hŒ â€œrootâ€”…””}”(hjîChžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mº hjëCubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjèCubj)”}”(hhh]”hæ)”}”(hŒPartition root”h]”hŒPartition root”…””}”(hjDhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mº hjDubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjèCubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhj®Cubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ "isolated"”h]”hŒâ€œisolatedâ€”…””}”(hj%DhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M» hj"Dubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjDubj)”}”(hhh]”hæ)”}”(hŒ%Partition root without load balancing”h]”hŒ%Partition root without load balancing”…””}”(hjWhen set to "isolated", the CPUs in that partition will be in an isolated state without any load balancing from the scheduler and excluded from the unbound workqueues. Tasks placed in such a partition with multiple CPUs should be carefully distributed and bound to each of the individual CPUs for optimal performance.”h]”hXBWhen set to â€œisolatedâ€, the CPUs in that partition will be in an isolated state without any load balancing from the scheduler and excluded from the unbound workqueues. Tasks placed in such a partition with multiple CPUs should be carefully distributed and bound to each of the individual CPUs for optimal performance.”…””}”(hjµDhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÛ hjqCubhæ)”}”(hŒæA partition root ("root" or "isolated") can be in one of the two possible states - valid or invalid. An invalid partition root is in a degraded state where some state information may be retained, but behaves more like a "member".”h]”hŒòA partition root (â€œrootâ€ or â€œisolatedâ€) can be in one of the two possible states - valid or invalid. An invalid partition root is in a degraded state where some state information may be retained, but behaves more like a â€œmemberâ€.”…””}”(hjÃDhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Má hjqCubhæ)”}”(hŒQAll possible state transitions among "member", "root" and "isolated" are allowed.”h]”hŒ]All possible state transitions among â€œmemberâ€, â€œrootâ€ and â€œisolatedâ€ are allowed.”…””}”(hjÑDhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mæ hjqCubhæ)”}”(hŒHOn read, the "cpuset.cpus.partition" file can show the following values.”h]”hŒLOn read, the â€œcpuset.cpus.partitionâ€ file can show the following values.”…””}”(hjßDhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mé hjqCubj¬)”}”(hX©============================= ===================================== "member" Non-root member of a partition "root" Partition root "isolated" Partition root without load balancing "root invalid ()" Invalid partition root "isolated invalid ()" Invalid isolated partition root ============================= ===================================== ”h]”jÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhjôDubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K%uh1jæhjôDubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ"member"”h]”hŒâ€œmemberâ€”…””}”(hjEhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mí hjEubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjEubj)”}”(hhh]”hæ)”}”(hŒNon-root member of a partition”h]”hŒNon-root member of a partition”…””}”(hj+EhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mí hj(Eubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjEubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjEubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ"root"”h]”hŒ â€œrootâ€”…””}”(hjKEhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mî hjHEubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjEEubj)”}”(hhh]”hæ)”}”(hŒPartition root”h]”hŒPartition root”…””}”(hjbEhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mî hj_Eubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjEEubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjEubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ "isolated"”h]”hŒâ€œisolatedâ€”…””}”(hj‚EhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mï hjEubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj|Eubj)”}”(hhh]”hæ)”}”(hŒ%Partition root without load balancing”h]”hŒ%Partition root without load balancing”…””}”(hj™EhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mï hj–Eubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj|Eubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjEubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ"root invalid ()"”h]”hŒâ€œroot invalid ()â€”…””}”(hj¹EhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mð hj¶Eubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj³Eubj)”}”(hhh]”hæ)”}”(hŒInvalid partition root”h]”hŒInvalid partition root”…””}”(hjÐEhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mð hjÍEubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj³Eubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjEubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ"isolated invalid ()"”h]”hŒ!â€œisolated invalid ()â€”…””}”(hjðEhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mñ hjíEubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjêEubj)”}”(hhh]”hæ)”}”(hŒInvalid isolated partition root”h]”hŒInvalid isolated partition root”…””}”(hjFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mñ hjFubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjêEubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjEubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühjôDubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhjñDubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhjíDubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mì hjqCubhæ)”}”(hŒ~In the case of an invalid partition root, a descriptive string on why the partition is invalid is included within parentheses.”h]”hŒ~In the case of an invalid partition root, a descriptive string on why the partition is invalid is included within parentheses.”…””}”(hj:FhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mô hjqCubhæ)”}”(hŒMFor a local partition root to be valid, the following conditions must be met.”h]”hŒMFor a local partition root to be valid, the following conditions must be met.”…””}”(hjHFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M÷ hjqCubhŒenumerated_list”“”)”}”(hhh]”(j¯)”}”(hŒ,The parent cgroup is a valid partition root.”h]”hæ)”}”(hj]Fh]”hŒ,The parent cgroup is a valid partition root.”…””}”(hj_FhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mú hj[Fubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjXFubj¯)”}”(hŒ_The "cpuset.cpus.exclusive.effective" file cannot be empty, though it may contain offline CPUs.”h]”hæ)”}”(hŒ_The "cpuset.cpus.exclusive.effective" file cannot be empty, though it may contain offline CPUs.”h]”hŒcThe â€œcpuset.cpus.exclusive.effectiveâ€ file cannot be empty, though it may contain offline CPUs.”…””}”(hjvFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mû hjrFubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjXFubj¯)”}”(hŒdThe "cpuset.cpus.effective" cannot be empty unless there is no task associated with this partition. ”h]”hæ)”}”(hŒcThe "cpuset.cpus.effective" cannot be empty unless there is no task associated with this partition.”h]”hŒgThe â€œcpuset.cpus.effectiveâ€ cannot be empty unless there is no task associated with this partition.”…””}”(hjŽFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mý hjŠFubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjXFubeh}”(h]”h ]”h"]”h$]”h&]”Œenumtype”Œarabic”Œprefix”hŒsuffix”Œ)”uh1jVFhjqCubhæ)”}”(hŒcFor a remote partition root to be valid, all the above conditions except the first one must be met.”h]”hŒcFor a remote partition root to be valid, all the above conditions except the first one must be met.”…””}”(hjFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjqCubhæ)”}”(hŒèExternal events like hotplug or changes to "cpuset.cpus" or "cpuset.cpus.exclusive" can cause a valid partition root to become invalid and vice versa. Note that a task cannot be moved to a cgroup with empty "cpuset.cpus.effective".”h]”hŒôExternal events like hotplug or changes to â€œcpuset.cpusâ€ or â€œcpuset.cpus.exclusiveâ€ can cause a valid partition root to become invalid and vice versa. Note that a task cannot be moved to a cgroup with empty â€œcpuset.cpus.effectiveâ€.”…””}”(hj»FhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjqCubhæ)”}”(hŒ‰A valid non-root parent partition may distribute out all its CPUs to its child local partitions when there is no task associated with it.”h]”hŒ‰A valid non-root parent partition may distribute out all its CPUs to its child local partitions when there is no task associated with it.”…””}”(hjÉFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjqCubhæ)”}”(hXgCare must be taken to change a valid partition root to "member" as all its child local partitions, if present, will become invalid causing disruption to tasks running in those child partitions. These inactivated partitions could be recovered if their parent is switched back to a partition root with a proper value in "cpuset.cpus" or "cpuset.cpus.exclusive".”h]”hXsCare must be taken to change a valid partition root to â€œmemberâ€ as all its child local partitions, if present, will become invalid causing disruption to tasks running in those child partitions. These inactivated partitions could be recovered if their parent is switched back to a partition root with a proper value in â€œcpuset.cpusâ€ or â€œcpuset.cpus.exclusiveâ€.”…””}”(hj×FhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjqCubhæ)”}”(hXtPoll and inotify events are triggered whenever the state of "cpuset.cpus.partition" changes. That includes changes caused by write to "cpuset.cpus.partition", cpu hotplug or other changes that modify the validity status of the partition. This will allow user space agents to monitor unexpected changes to "cpuset.cpus.partition" without the need to do continuous polling.”h]”hX€Poll and inotify events are triggered whenever the state of â€œcpuset.cpus.partitionâ€ changes. That includes changes caused by write to â€œcpuset.cpus.partitionâ€, cpu hotplug or other changes that modify the validity status of the partition. This will allow user space agents to monitor unexpected changes to â€œcpuset.cpus.partitionâ€ without the need to do continuous polling.”…””}”(hjåFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjqCubhæ)”}”(hŒ÷A user can pre-configure certain CPUs to an isolated state with load balancing disabled at boot time with the "isolcpus" kernel boot command line option. If those CPUs are to be put into a partition, they have to be used in an isolated partition.”h]”hŒûA user can pre-configure certain CPUs to an isolated state with load balancing disabled at boot time with the â€œisolcpusâ€ kernel boot command line option. If those CPUs are to be put into a partition, they have to be used in an isolated partition.”…””}”(hjóFhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjqCubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj_Cubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M hjµ@ubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj±@ubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M" hj @hžhubeh}”(h]”Œcpuset-interface-files”ah ]”h"]”Œcpuset interface files”ah$]”h&]”uh1h°hjs@hžhhŸh¯h M ubeh}”(h]”Œcpuset”ah ]”h"]”Œcpuset”ah$]”h&]”uh1h°hj«hžhhŸh¯h M ubh±)”}”(hhh]”(h¶)”}”(hŒDevice controller”h]”hŒDevice controller”…””}”(hj,GhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj)GhžhhŸh¯h M" ubhæ)”}”(hŒ—Device controller manages access to device files. It includes both creation of new device files (using mknod), and access to the existing device files.”h]”hŒ—Device controller manages access to device files. It includes both creation of new device files (using mknod), and access to the existing device files.”…””}”(hj:GhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M$ hj)Ghžhubhæ)”}”(hXCgroup v2 device controller has no interface files and is implemented on top of cgroup BPF. To control access to device files, a user may create bpf programs of type BPF_PROG_TYPE_CGROUP_DEVICE and attach them to cgroups with BPF_CGROUP_DEVICE flag. On an attempt to access a device file, corresponding BPF programs will be executed, and depending on the return value the attempt will succeed or fail with -EPERM.”h]”hXCgroup v2 device controller has no interface files and is implemented on top of cgroup BPF. To control access to device files, a user may create bpf programs of type BPF_PROG_TYPE_CGROUP_DEVICE and attach them to cgroups with BPF_CGROUP_DEVICE flag. On an attempt to access a device file, corresponding BPF programs will be executed, and depending on the return value the attempt will succeed or fail with -EPERM.”…””}”(hjHGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M( hj)Ghžhubhæ)”}”(hX A BPF_PROG_TYPE_CGROUP_DEVICE program takes a pointer to the bpf_cgroup_dev_ctx structure, which describes the device access attempt: access type (mknod/read/write) and device (type, major and minor numbers). If the program returns 0, the attempt fails with -EPERM, otherwise it succeeds.”h]”hX A BPF_PROG_TYPE_CGROUP_DEVICE program takes a pointer to the bpf_cgroup_dev_ctx structure, which describes the device access attempt: access type (mknod/read/write) and device (type, major and minor numbers). If the program returns 0, the attempt fails with -EPERM, otherwise it succeeds.”…””}”(hjVGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M/ hj)Ghžhubhæ)”}”(hŒ‹An example of BPF_PROG_TYPE_CGROUP_DEVICE program may be found in tools/testing/selftests/bpf/progs/dev_cgroup.c in the kernel source tree.”h]”hŒ‹An example of BPF_PROG_TYPE_CGROUP_DEVICE program may be found in tools/testing/selftests/bpf/progs/dev_cgroup.c in the kernel source tree.”…””}”(hjdGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M5 hj)Ghžhubeh}”(h]”Œdevice-controller”ah ]”h"]”Œdevice controller”ah$]”h&]”uh1h°hj«hžhhŸh¯h M" ubh±)”}”(hhh]”(h¶)”}”(hŒRDMA”h]”hŒRDMA”…””}”(hj}GhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjzGhžhhŸh¯h M: ubhæ)”}”(hŒRThe "rdma" controller regulates the distribution and accounting of RDMA resources.”h]”hŒVThe â€œrdmaâ€ controller regulates the distribution and accounting of RDMA resources.”…””}”(hj‹GhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M< hjzGhžhubh±)”}”(hhh]”(h¶)”}”(hŒRDMA Interface Files”h]”hŒRDMA Interface Files”…””}”(hjœGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj™GhžhhŸh¯h M@ ubj¬)”}”(hXîrdma.max A readwrite nested-keyed file that exists for all the cgroups except root that describes current configured resource limit for a RDMA/IB device. Lines are keyed by device name and are not ordered. Each line contains space separated resource name and its configured limit that can be distributed. The following nested keys are defined. ========== ============================= hca_handle Maximum number of HCA Handles hca_object Maximum number of HCA Objects ========== ============================= An example for mlx4 and ocrdma device follows:: mlx4_0 hca_handle=2 hca_object=2000 ocrdma1 hca_handle=3 hca_object=max rdma.current A read-only file that describes current resource usage. It exists for all the cgroup except root. An example for mlx4 and ocrdma device follows:: mlx4_0 hca_handle=1 hca_object=20 ocrdma1 hca_handle=1 hca_object=23 ”h]”j²)”}”(hhh]”(j·)”}”(hX‘rdma.max A readwrite nested-keyed file that exists for all the cgroups except root that describes current configured resource limit for a RDMA/IB device. Lines are keyed by device name and are not ordered. Each line contains space separated resource name and its configured limit that can be distributed. The following nested keys are defined. ========== ============================= hca_handle Maximum number of HCA Handles hca_object Maximum number of HCA Objects ========== ============================= An example for mlx4 and ocrdma device follows:: mlx4_0 hca_handle=2 hca_object=2000 ocrdma1 hca_handle=3 hca_object=max ”h]”(j½)”}”(hŒrdma.max”h]”hŒrdma.max”…””}”(hjµGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MU hj±GubjÍ)”}”(hhh]”(hæ)”}”(hŒA readwrite nested-keyed file that exists for all the cgroups except root that describes current configured resource limit for a RDMA/IB device.”h]”hŒA readwrite nested-keyed file that exists for all the cgroups except root that describes current configured resource limit for a RDMA/IB device.”…””}”(hjÆGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MC hjÃGubhæ)”}”(hŒ–Lines are keyed by device name and are not ordered. Each line contains space separated resource name and its configured limit that can be distributed.”h]”hŒ–Lines are keyed by device name and are not ordered. Each line contains space separated resource name and its configured limit that can be distributed.”…””}”(hjÔGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MG hjÃGubhæ)”}”(hŒ&The following nested keys are defined.”h]”hŒ&The following nested keys are defined.”…””}”(hjâGhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MK hjÃGubj¬)”}”(hŒ°========== ============================= hca_handle Maximum number of HCA Handles hca_object Maximum number of HCA Objects ========== ============================= ”h]”jÝ)”}”(hhh]”jâ)”}”(hhh]”(jç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K uh1jæhj÷Gubjç)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”Kuh1jæhj÷Gubjý)”}”(hhh]”(j)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ hca_handle”h]”hŒ hca_handle”…””}”(hjHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MN hjHubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjHubj)”}”(hhh]”hæ)”}”(hŒMaximum number of HCA Handles”h]”hŒMaximum number of HCA Handles”…””}”(hj.HhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MN hj+Hubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjHubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjHubj)”}”(hhh]”(j)”}”(hhh]”hæ)”}”(hŒ hca_object”h]”hŒ hca_object”…””}”(hjNHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MO hjKHubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjHHubj)”}”(hhh]”hæ)”}”(hŒMaximum number of HCA Objects”h]”hŒMaximum number of HCA Objects”…””}”(hjeHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MO hjbHubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjHHubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjHubeh}”(h]”h ]”h"]”h$]”h&]”uh1jühj÷Gubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1jáhjôGubah}”(h]”h ]”h"]”h$]”h&]”uh1jÜhjðGubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MM hjÃGubhæ)”}”(hŒ/An example for mlx4 and ocrdma device follows::”h]”hŒ.An example for mlx4 and ocrdma device follows:”…””}”(hj˜HhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MR hjÃGubjV)”}”(hŒGmlx4_0 hca_handle=2 hca_object=2000 ocrdma1 hca_handle=3 hca_object=max”h]”hŒGmlx4_0 hca_handle=2 hca_object=2000 ocrdma1 hca_handle=3 hca_object=max”…””}”hj¦Hsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MT hjÃGubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj±Gubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MU hj®Gubj·)”}”(hŒêrdma.current A read-only file that describes current resource usage. It exists for all the cgroup except root. An example for mlx4 and ocrdma device follows:: mlx4_0 hca_handle=1 hca_object=20 ocrdma1 hca_handle=1 hca_object=23 ”h]”(j½)”}”(hŒrdma.current”h]”hŒrdma.current”…””}”(hjÄHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M^ hjÀHubjÍ)”}”(hhh]”(hæ)”}”(hŒaA read-only file that describes current resource usage. It exists for all the cgroup except root.”h]”hŒaA read-only file that describes current resource usage. It exists for all the cgroup except root.”…””}”(hjÕHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MX hjÒHubhæ)”}”(hŒ/An example for mlx4 and ocrdma device follows::”h]”hŒ.An example for mlx4 and ocrdma device follows:”…””}”(hjãHhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M[ hjÒHubjV)”}”(hŒDmlx4_0 hca_handle=1 hca_object=20 ocrdma1 hca_handle=1 hca_object=23”h]”hŒDmlx4_0 hca_handle=1 hca_object=20 ocrdma1 hca_handle=1 hca_object=23”…””}”hjñHsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M] hjÒHubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÀHubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M^ hj®Gubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjªGubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MB hj™Ghžhubeh}”(h]”Œrdma-interface-files”ah ]”h"]”Œrdma interface files”ah$]”h&]”uh1h°hjzGhžhhŸh¯h M@ ubeh}”(h]”Œrdma”ah ]”h"]”Œrdma”ah$]”h&]”uh1h°hj«hžhhŸh¯h M: ubh±)”}”(hhh]”(h¶)”}”(hŒDMEM”h]”hŒDMEM”…””}”(hj*IhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj'IhžhhŸh¯h Ma ubhæ)”}”(hŒåThe "dmem" controller regulates the distribution and accounting of device memory regions. Because each memory region may have its own page size, which does not have to be equal to the system page size, the units are always bytes.”h]”hŒéThe â€œdmemâ€ controller regulates the distribution and accounting of device memory regions. Because each memory region may have its own page size, which does not have to be equal to the system page size, the units are always bytes.”…””}”(hj8IhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mc hj'Ihžhubh±)”}”(hhh]”(h¶)”}”(hŒDMEM Interface Files”h]”hŒDMEM Interface Files”…””}”(hjIIhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjFIhžhhŸh¯h Mh ubj¬)”}”(hXßdmem.max, dmem.min, dmem.low A readwrite nested-keyed file that exists for all the cgroups except root that describes current configured resource limit for a region. An example for xe follows:: drm/0000:03:00.0/vram0 1073741824 drm/0000:03:00.0/stolen max The semantics are the same as for the memory cgroup controller, and are calculated in the same way. dmem.capacity A read-only file that describes maximum region capacity. It only exists on the root cgroup. Not all memory can be allocated by cgroups, as the kernel reserves some for internal use. An example for xe follows:: drm/0000:03:00.0/vram0 8514437120 drm/0000:03:00.0/stolen 67108864 dmem.current A read-only file that describes current resource usage. It exists for all the cgroup except root. An example for xe follows:: drm/0000:03:00.0/vram0 12550144 drm/0000:03:00.0/stolen 8650752 ”h]”j²)”}”(hhh]”(j·)”}”(hXkdmem.max, dmem.min, dmem.low A readwrite nested-keyed file that exists for all the cgroups except root that describes current configured resource limit for a region. An example for xe follows:: drm/0000:03:00.0/vram0 1073741824 drm/0000:03:00.0/stolen max The semantics are the same as for the memory cgroup controller, and are calculated in the same way. ”h]”(j½)”}”(hŒdmem.max, dmem.min, dmem.low”h]”hŒdmem.max, dmem.min, dmem.low”…””}”(hjbIhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mu hj^IubjÍ)”}”(hhh]”(hæ)”}”(hŒˆA readwrite nested-keyed file that exists for all the cgroups except root that describes current configured resource limit for a region.”h]”hŒˆA readwrite nested-keyed file that exists for all the cgroups except root that describes current configured resource limit for a region.”…””}”(hjsIhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mk hjpIubhæ)”}”(hŒAn example for xe follows::”h]”hŒAn example for xe follows:”…””}”(hjIhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mo hjpIubjV)”}”(hŒ=drm/0000:03:00.0/vram0 1073741824 drm/0000:03:00.0/stolen max”h]”hŒ=drm/0000:03:00.0/vram0 1073741824 drm/0000:03:00.0/stolen max”…””}”hjIsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mq hjpIubhæ)”}”(hŒcThe semantics are the same as for the memory cgroup controller, and are calculated in the same way.”h]”hŒcThe semantics are the same as for the memory cgroup controller, and are calculated in the same way.”…””}”(hjIhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mt hjpIubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj^Iubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mu hj[Iubj·)”}”(hX)dmem.capacity A read-only file that describes maximum region capacity. It only exists on the root cgroup. Not all memory can be allocated by cgroups, as the kernel reserves some for internal use. An example for xe follows:: drm/0000:03:00.0/vram0 8514437120 drm/0000:03:00.0/stolen 67108864 ”h]”(j½)”}”(hŒ dmem.capacity”h]”hŒ dmem.capacity”…””}”(hj»IhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M€ hj·IubjÍ)”}”(hhh]”(hæ)”}”(hŒµA read-only file that describes maximum region capacity. It only exists on the root cgroup. Not all memory can be allocated by cgroups, as the kernel reserves some for internal use.”h]”hŒµA read-only file that describes maximum region capacity. It only exists on the root cgroup. Not all memory can be allocated by cgroups, as the kernel reserves some for internal use.”…””}”(hjÌIhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mx hjÉIubhæ)”}”(hŒAn example for xe follows::”h]”hŒAn example for xe follows:”…””}”(hjÚIhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M} hjÉIubjV)”}”(hŒBdrm/0000:03:00.0/vram0 8514437120 drm/0000:03:00.0/stolen 67108864”h]”hŒBdrm/0000:03:00.0/vram0 8514437120 drm/0000:03:00.0/stolen 67108864”…””}”hjèIsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M hjÉIubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj·Iubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M€ hj[Iubj·)”}”(hŒÑdmem.current A read-only file that describes current resource usage. It exists for all the cgroup except root. An example for xe follows:: drm/0000:03:00.0/vram0 12550144 drm/0000:03:00.0/stolen 8650752 ”h]”(j½)”}”(hŒdmem.current”h]”hŒdmem.current”…””}”(hjJhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M‰ hjJubjÍ)”}”(hhh]”(hæ)”}”(hŒaA read-only file that describes current resource usage. It exists for all the cgroup except root.”h]”hŒaA read-only file that describes current resource usage. It exists for all the cgroup except root.”…””}”(hjJhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mƒ hjJubhæ)”}”(hŒAn example for xe follows::”h]”hŒAn example for xe follows:”…””}”(hj%JhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M† hjJubjV)”}”(hŒ?drm/0000:03:00.0/vram0 12550144 drm/0000:03:00.0/stolen 8650752”h]”hŒ?drm/0000:03:00.0/vram0 12550144 drm/0000:03:00.0/stolen 8650752”…””}”hj3Jsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mˆ hjJubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjJubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M‰ hj[Iubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjWIubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mj hjFIhžhubeh}”(h]”Œdmem-interface-files”ah ]”h"]”Œdmem interface files”ah$]”h&]”uh1h°hj'IhžhhŸh¯h Mh ubeh}”(h]”Œdmem”ah ]”h"]”Œdmem”ah$]”h&]”uh1h°hj«hžhhŸh¯h Ma ubh±)”}”(hhh]”(h¶)”}”(hŒHugeTLB”h]”hŒHugeTLB”…””}”(hjlJhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjiJhžhhŸh¯h MŒ ubhæ)”}”(hŒThe HugeTLB controller allows to limit the HugeTLB usage per control group and enforces the controller limit during page fault.”h]”hŒThe HugeTLB controller allows to limit the HugeTLB usage per control group and enforces the controller limit during page fault.”…””}”(hjzJhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MŽ hjiJhžhubh±)”}”(hhh]”(h¶)”}”(hŒHugeTLB Interface Files”h]”hŒHugeTLB Interface Files”…””}”(hj‹JhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjˆJhžhhŸh¯h M’ ubj¬)”}”(hXÓhugetlb..current Show current usage for "hugepagesize" hugetlb. It exists for all the cgroup except root. hugetlb..max Set/show the hard limit of "hugepagesize" hugetlb usage. The default value is "max". It exists for all the cgroup except root. hugetlb..events A read-only flat-keyed file which exists on non-root cgroups. max The number of allocation failure due to HugeTLB limit hugetlb..events.local Similar to hugetlb..events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events. hugetlb..numa_stat Similar to memory.numa_stat, it shows the numa information of the hugetlb pages of in this cgroup. Only active in use hugetlb pages are included. The per-node values are in bytes. ”h]”j²)”}”(hhh]”(j·)”}”(hŒyhugetlb..current Show current usage for "hugepagesize" hugetlb. It exists for all the cgroup except root. ”h]”(j½)”}”(hŒhugetlb..current”h]”hŒhugetlb..current”…””}”(hj¤JhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M– hj JubjÍ)”}”(hhh]”hæ)”}”(hŒYShow current usage for "hugepagesize" hugetlb. It exists for all the cgroup except root.”h]”hŒ]Show current usage for â€œhugepagesizeâ€ hugetlb. It exists for all the cgroup except root.”…””}”(hjµJhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M• hj²Jubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj Jubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M– hjJubj·)”}”(hŒ›hugetlb..max Set/show the hard limit of "hugepagesize" hugetlb usage. The default value is "max". It exists for all the cgroup except root. ”h]”(j½)”}”(hŒhugetlb..max”h]”hŒhugetlb..max”…””}”(hjÓJhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mš hjÏJubjÍ)”}”(hhh]”hæ)”}”(hŒSet/show the hard limit of "hugepagesize" hugetlb usage. The default value is "max". It exists for all the cgroup except root.”h]”hŒ‡Set/show the hard limit of â€œhugepagesizeâ€ hugetlb usage. The default value is â€œmaxâ€. It exists for all the cgroup except root.”…””}”(hjäJhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M™ hjáJubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÏJubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mš hjJubj·)”}”(hŒ¡hugetlb..events A read-only flat-keyed file which exists on non-root cgroups. max The number of allocation failure due to HugeTLB limit ”h]”(j½)”}”(hŒhugetlb..events”h]”hŒhugetlb..events”…””}”(hjKhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M hjþJubjÍ)”}”(hhh]”(hæ)”}”(hŒ=A read-only flat-keyed file which exists on non-root cgroups.”h]”hŒ=A read-only flat-keyed file which exists on non-root cgroups.”…””}”(hjKhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hjKubj¬)”}”(hŒ@max The number of allocation failure due to HugeTLB limit ”h]”j²)”}”(hhh]”j·)”}”(hŒ:max The number of allocation failure due to HugeTLB limit ”h]”(j½)”}”(hŒmax”h]”hŒmax”…””}”(hj,KhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M hj(KubjÍ)”}”(hhh]”hæ)”}”(hŒ5The number of allocation failure due to HugeTLB limit”h]”hŒ5The number of allocation failure due to HugeTLB limit”…””}”(hj=KhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M hj:Kubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj(Kubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M hj%Kubah}”(h]”h ]”h"]”h$]”h&]”uh1j±hj!Kubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MŸ hjKubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjþJubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M hjJubj·)”}”(hŒæhugetlb..events.local Similar to hugetlb..events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events. ”h]”(j½)”}”(hŒ#hugetlb..events.local”h]”hŒ#hugetlb..events.local”…””}”(hjsKhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M¥ hjoKubjÍ)”}”(hhh]”hæ)”}”(hŒÁSimilar to hugetlb..events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events.”h]”hŒÁSimilar to hugetlb..events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events.”…””}”(hj„KhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M£ hjKubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjoKubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M¥ hjJubj·)”}”(hŒæhugetlb..numa_stat Similar to memory.numa_stat, it shows the numa information of the hugetlb pages of in this cgroup. Only active in use hugetlb pages are included. The per-node values are in bytes. ”h]”(j½)”}”(hŒ hugetlb..numa_stat”h]”hŒ hugetlb..numa_stat”…””}”(hj¢KhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mª hjžKubjÍ)”}”(hhh]”hæ)”}”(hŒÄSimilar to memory.numa_stat, it shows the numa information of the hugetlb pages of in this cgroup. Only active in use hugetlb pages are included. The per-node values are in bytes.”h]”hŒÄSimilar to memory.numa_stat, it shows the numa information of the hugetlb pages of in this cgroup. Only active in use hugetlb pages are included. The per-node values are in bytes.”…””•ˆ}”(hj³KhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¨ hj°Kubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjžKubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mª hjJubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj™Jubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h M” hjˆJhžhubeh}”(h]”Œhugetlb-interface-files”ah ]”h"]”Œhugetlb interface files”ah$]”h&]”uh1h°hjiJhžhhŸh¯h M’ ubeh}”(h]”Œhugetlb”ah ]”h"]”Œhugetlb”ah$]”h&]”uh1h°hj«hžhhŸh¯h MŒ ubh±)”}”(hhh]”(h¶)”}”(hŒMisc”h]”hŒMisc”…””}”(hjìKhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjéKhžhhŸh¯h M ubhæ)”}”(hŒâThe Miscellaneous cgroup provides the resource limiting and tracking mechanism for the scalar resources which cannot be abstracted like the other cgroup resources. Controller is enabled by the CONFIG_CGROUP_MISC config option.”h]”hŒâThe Miscellaneous cgroup provides the resource limiting and tracking mechanism for the scalar resources which cannot be abstracted like the other cgroup resources. Controller is enabled by the CONFIG_CGROUP_MISC config option.”…””}”(hjúKhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¯ hjéKhžhubhæ)”}”(hX)A resource can be added to the controller via enum misc_res_type{} in the include/linux/misc_cgroup.h file and the corresponding name via misc_res_name[] in the kernel/cgroup/misc.c file. Provider of the resource must set its capacity prior to using the resource by calling misc_cg_set_capacity().”h]”hX)A resource can be added to the controller via enum misc_res_type{} in the include/linux/misc_cgroup.h file and the corresponding name via misc_res_name[] in the kernel/cgroup/misc.c file. Provider of the resource must set its capacity prior to using the resource by calling misc_cg_set_capacity().”…””}”(hjLhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M´ hjéKhžhubhæ)”}”(hŒ²Once a capacity is set then the resource usage can be updated using charge and uncharge APIs. All of the APIs to interact with misc controller are in include/linux/misc_cgroup.h.”h]”hŒ²Once a capacity is set then the resource usage can be updated using charge and uncharge APIs. All of the APIs to interact with misc controller are in include/linux/misc_cgroup.h.”…””}”(hjLhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹ hjéKhžhubh±)”}”(hhh]”(h¶)”}”(hŒMisc Interface Files”h]”hŒMisc Interface Files”…””}”(hj'LhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj$LhžhhŸh¯h M¾ ubhæ)”}”(hŒqMiscellaneous controller provides 3 interface files. If two misc resources (res_a and res_b) are registered then:”h]”hŒqMiscellaneous controller provides 3 interface files. If two misc resources (res_a and res_b) are registered then:”…””}”(hj5LhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÀ hj$Lhžhubj¬)”}”(hXÎmisc.capacity A read-only flat-keyed file shown only in the root cgroup. It shows miscellaneous scalar resources available on the platform along with their quantities:: $ cat misc.capacity res_a 50 res_b 10 misc.current A read-only flat-keyed file shown in the all cgroups. It shows the current usage of the resources in the cgroup and its children.:: $ cat misc.current res_a 3 res_b 0 misc.peak A read-only flat-keyed file shown in all cgroups. It shows the historical maximum usage of the resources in the cgroup and its children.:: $ cat misc.peak res_a 10 res_b 8 misc.max A read-write flat-keyed file shown in the non root cgroups. Allowed maximum usage of the resources in the cgroup and its children.:: $ cat misc.max res_a max res_b 4 Limit can be set by:: # echo res_a 1 > misc.max Limit can be set to max by:: # echo res_a max > misc.max Limits can be set higher than the capacity value in the misc.capacity file. misc.events A read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. All fields in this file are hierarchical. max The number of times the cgroup's resource usage was about to go over the max boundary. misc.events.local Similar to misc.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events. ”h]”j²)”}”(hhh]”(j·)”}”(hŒ×misc.capacity A read-only flat-keyed file shown only in the root cgroup. It shows miscellaneous scalar resources available on the platform along with their quantities:: $ cat misc.capacity res_a 50 res_b 10 ”h]”(j½)”}”(hŒ misc.capacity”h]”hŒ misc.capacity”…””}”(hjNLhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÉ hjJLubjÍ)”}”(hhh]”(hæ)”}”(hŒ›A read-only flat-keyed file shown only in the root cgroup. It shows miscellaneous scalar resources available on the platform along with their quantities::”h]”hŒšA read-only flat-keyed file shown only in the root cgroup. It shows miscellaneous scalar resources available on the platform along with their quantities:”…””}”(hj_LhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÃ hj\LubjV)”}”(hŒ%$ cat misc.capacity res_a 50 res_b 10”h]”hŒ%$ cat misc.capacity res_a 50 res_b 10”…””}”hjmLsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MÇ hj\Lubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjJLubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÉ hjGLubj·)”}”(hŒ¼misc.current A read-only flat-keyed file shown in the all cgroups. It shows the current usage of the resources in the cgroup and its children.:: $ cat misc.current res_a 3 res_b 0 ”h]”(j½)”}”(hŒmisc.current”h]”hŒmisc.current”…””}”(hj‹LhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÑ hj‡LubjÍ)”}”(hhh]”(hæ)”}”(hŒ„A read-only flat-keyed file shown in the all cgroups. It shows the current usage of the resources in the cgroup and its children.::”h]”hŒƒA read-only flat-keyed file shown in the all cgroups. It shows the current usage of the resources in the cgroup and its children.:”…””}”(hjœLhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÌ hj™LubjV)”}”(hŒ"$ cat misc.current res_a 3 res_b 0”h]”hŒ"$ cat misc.current res_a 3 res_b 0”…””}”hjªLsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MÏ hj™Lubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj‡Lubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÑ hjGLubj·)”}”(hŒ¾misc.peak A read-only flat-keyed file shown in all cgroups. It shows the historical maximum usage of the resources in the cgroup and its children.:: $ cat misc.peak res_a 10 res_b 8 ”h]”(j½)”}”(hŒ misc.peak”h]”hŒ misc.peak”…””}”(hjÈLhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÚ hjÄLubjÍ)”}”(hhh]”(hæ)”}”(hŒ‹A read-only flat-keyed file shown in all cgroups. It shows the historical maximum usage of the resources in the cgroup and its children.::”h]”hŒŠA read-only flat-keyed file shown in all cgroups. It shows the historical maximum usage of the resources in the cgroup and its children.:”…””}”(hjÙLhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÔ hjÖLubjV)”}”(hŒ $ cat misc.peak res_a 10 res_b 8”h]”hŒ $ cat misc.peak res_a 10 res_b 8”…””}”hjçLsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MØ hjÖLubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÄLubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÚ hjGLubj·)”}”(hXtmisc.max A read-write flat-keyed file shown in the non root cgroups. Allowed maximum usage of the resources in the cgroup and its children.:: $ cat misc.max res_a max res_b 4 Limit can be set by:: # echo res_a 1 > misc.max Limit can be set to max by:: # echo res_a max > misc.max Limits can be set higher than the capacity value in the misc.capacity file. ”h]”(j½)”}”(hŒmisc.max”h]”hŒmisc.max”…””}”(hjMhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mí hjMubjÍ)”}”(hhh]”(hæ)”}”(hŒ„A read-write flat-keyed file shown in the non root cgroups. Allowed maximum usage of the resources in the cgroup and its children.::”h]”hŒƒA read-write flat-keyed file shown in the non root cgroups. Allowed maximum usage of the resources in the cgroup and its children.:”…””}”(hjMhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÝ hjMubjV)”}”(hŒ $ cat misc.max res_a max res_b 4”h]”hŒ $ cat misc.max res_a max res_b 4”…””}”hj$Msbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mà hjMubhæ)”}”(hŒLimit can be set by::”h]”hŒLimit can be set by:”…””}”(hj2MhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mä hjMubjV)”}”(hŒ# echo res_a 1 > misc.max”h]”hŒ# echo res_a 1 > misc.max”…””}”hj@Msbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mæ hjMubhæ)”}”(hŒLimit can be set to max by::”h]”hŒLimit can be set to max by:”…””}”(hjNMhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mè hjMubjV)”}”(hŒ# echo res_a max > misc.max”h]”hŒ# echo res_a max > misc.max”…””}”hj\Msbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h Mê hjMubhæ)”}”(hŒKLimits can be set higher than the capacity value in the misc.capacity file.”h]”hŒKLimits can be set higher than the capacity value in the misc.capacity file.”…””}”(hjjMhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mì hjMubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjMubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mí hjGLubj·)”}”(hX^misc.events A read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. All fields in this file are hierarchical. max The number of times the cgroup's resource usage was about to go over the max boundary. ”h]”(j½)”}”(hŒmisc.events”h]”hŒmisc.events”…””}”(hjˆMhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M÷ hj„MubjÍ)”}”(hhh]”(hæ)”}”(hŒãA read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. All fields in this file are hierarchical.”h]”hŒãA read-only flat-keyed file which exists on non-root cgroups. The following entries are defined. Unless specified otherwise, a value change in this file generates a file modified event. All fields in this file are hierarchical.”…””}”(hj™MhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mð hj–Mubj¬)”}”(hŒgmax The number of times the cgroup's resource usage was about to go over the max boundary. ”h]”j²)”}”(hhh]”j·)”}”(hŒ[max The number of times the cgroup's resource usage was about to go over the max boundary. ”h]”(j½)”}”(hŒmax”h]”hŒmax”…””}”(hj²MhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h M÷ hj®MubjÍ)”}”(hhh]”hæ)”}”(hŒVThe number of times the cgroup's resource usage was about to go over the max boundary.”h]”hŒXThe number of times the cgroupâ€™s resource usage was about to go over the max boundary.”…””}”(hjÃMhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mö hjÀMubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj®Mubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M÷ hj«Mubah}”(h]”h ]”h"]”h$]”h&]”uh1j±hj§Mubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h Mõ hj–Mubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÌhj„Mubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h M÷ hjGLubj·)”}”(hŒÂmisc.events.local Similar to misc.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events. ”h]”(j½)”}”(hŒmisc.events.local”h]”hŒmisc.events.local”…””}”(hjùMhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h Mü hjõMubjÍ)”}”(hhh]”hæ)”}”(hŒ¯Similar to misc.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events.”h]”hŒ¯Similar to misc.events but the fields in the file are local to the cgroup i.e. not hierarchical. The file modified event generated on this file reflects only the local events.”…””}”(hj NhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mú hjNubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjõMubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h Mü hjGLubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hjCLubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MÂ hj$Lhžhubeh}”(h]”Œmisc-interface-files”ah ]”h"]”Œmisc interface files”ah$]”h&]”uh1h°hjéKhžhhŸh¯h M¾ ubh±)”}”(hhh]”(h¶)”}”(hŒMigration and Ownership”h]”hŒMigration and Ownership”…””}”(hj;NhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj8NhžhhŸh¯h Mÿ ubhæ)”}”(hXA miscellaneous scalar resource is charged to the cgroup in which it is used first, and stays charged to that cgroup until that resource is freed. Migrating a process to a different cgroup does not move the charge to the destination cgroup where the process has moved.”h]”hXA miscellaneous scalar resource is charged to the cgroup in which it is used first, and stays charged to that cgroup until that resource is freed. Migrating a process to a different cgroup does not move the charge to the destination cgroup where the process has moved.”…””}”(hjINhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj8Nhžhubeh}”(h]”Œmigration-and-ownership”ah ]”h"]”Œmigration and ownership”ah$]”h&]”uh1h°hjéKhžhhŸh¯h Mÿ ubeh}”(h]”Œmisc”ah ]”h"]”Œmisc”ah$]”h&]”uh1h°hj«hžhhŸh¯h M ubh±)”}”(hhh]”(h¶)”}”(hŒOthers”h]”hŒOthers”…””}”(hjjNhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjgNhžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒ perf_event”h]”hŒ perf_event”…””}”(hj{NhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjxNhžhhŸh¯h M ubhæ)”}”(hXperf_event controller, if not mounted on a legacy hierarchy, is automatically enabled on the v2 hierarchy so that perf events can always be filtered by cgroup v2 path. The controller can still be moved to a legacy hierarchy after v2 hierarchy is populated.”h]”hXperf_event controller, if not mounted on a legacy hierarchy, is automatically enabled on the v2 hierarchy so that perf events can always be filtered by cgroup v2 path. The controller can still be moved to a legacy hierarchy after v2 hierarchy is populated.”…””}”(hj‰NhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjxNhžhubeh}”(h]”Œ perf-event”ah ]”h"]”Œ perf_event”ah$]”h&]”uh1h°hjgNhžhhŸh¯h M ubeh}”(h]”Œothers”ah ]”h"]”Œothers”ah$]”h&]”uh1h°hj«hžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒNon-normative information”h]”hŒNon-normative information”…””}”(hjªNhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj§NhžhhŸh¯h Mubhæ)”}”(hŒzThis section contains information that isn't considered to be a part of the stable kernel API and so is subject to change.”h]”hŒ|This section contains information that isnâ€™t considered to be a part of the stable kernel API and so is subject to change.”…””}”(hj¸NhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Mhj§Nhžhubh±)”}”(hhh]”(h¶)”}”(hŒ,CPU controller root cgroup process behaviour”h]”hŒ,CPU controller root cgroup process behaviour”…””}”(hjÉNhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjÆNhžhhŸh¯h Mubhæ)”}”(hŒØWhen distributing CPU cycles in the root cgroup each thread in this cgroup is treated as if it was hosted in a separate child cgroup of the root cgroup. This child cgroup weight is dependent on its thread nice level.”h]”hŒØWhen distributing CPU cycles in the root cgroup each thread in this cgroup is treated as if it was hosted in a separate child cgroup of the root cgroup. This child cgroup weight is dependent on its thread nice level.”…””}”(hj×NhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÆNhžhubhæ)”}”(hŒÆFor details of this mapping see sched_prio_to_weight array in kernel/sched/core.c file (values from this array should be scaled appropriately so the neutral - nice 0 - value is 100 instead of 1024).”h]”hŒÆFor details of this mapping see sched_prio_to_weight array in kernel/sched/core.c file (values from this array should be scaled appropriately so the neutral - nice 0 - value is 100 instead of 1024).”…””}”(hjåNhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M!hjÆNhžhubeh}”(h]”Œ,cpu-controller-root-cgroup-process-behaviour”ah ]”h"]”Œ,cpu controller root cgroup process behaviour”ah$]”h&]”uh1h°hj§NhžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒ+IO controller root cgroup process behaviour”h]”hŒ+IO controller root cgroup process behaviour”…””}”(hjþNhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjûNhžhhŸh¯h M'ubhæ)”}”(hŒàRoot cgroup processes are hosted in an implicit leaf child node. When distributing IO resources this implicit child node is taken into account as if it was a normal child cgroup of the root cgroup with a weight value of 200.”h]”hŒàRoot cgroup processes are hosted in an implicit leaf child node. When distributing IO resources this implicit child node is taken into account as if it was a normal child cgroup of the root cgroup with a weight value of 200.”…””}”(hjOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M)hjûNhžhubeh}”(h]”Œ+io-controller-root-cgroup-process-behaviour”ah ]”h"]”Œ+io controller root cgroup process behaviour”ah$]”h&]”uh1h°hj§NhžhhŸh¯h M'ubeh}”(h]”Œnon-normative-information”ah ]”h"]”Œnon-normative information”ah$]”h&]”uh1h°hj«hžhhŸh¯h Mubeh}”(h]”Œcontrollers”ah ]”h"]”Œcontrollers”ah$]”h&]”uh1h°hh²hžhhŸh¯h M$ubh±)”}”(hhh]”(h¶)”}”(hŒ Namespace”h]”hŒ Namespace”…””}”(hj5OhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj2OhžhhŸh¯h M0ubh±)”}”(hhh]”(h¶)”}”(hŒBasics”h]”hŒBasics”…””}”(hjFOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjCOhžhhŸh¯h M3ubhæ)”}”(hX²cgroup namespace provides a mechanism to virtualize the view of the "/proc/$PID/cgroup" file and cgroup mounts. The CLONE_NEWCGROUP clone flag can be used with clone(2) and unshare(2) to create a new cgroup namespace. The process running inside the cgroup namespace will have its "/proc/$PID/cgroup" output restricted to cgroupns root. The cgroupns root is the cgroup of the process at the time of creation of the cgroup namespace.”h]”hXºcgroup namespace provides a mechanism to virtualize the view of the â€œ/proc/$PID/cgroupâ€ file and cgroup mounts. The CLONE_NEWCGROUP clone flag can be used with clone(2) and unshare(2) to create a new cgroup namespace. The process running inside the cgroup namespace will have its â€œ/proc/$PID/cgroupâ€ output restricted to cgroupns root. The cgroupns root is the cgroup of the process at the time of creation of the cgroup namespace.”…””}”(hjTOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M5hjCOhžhubhæ)”}”(hX;Without cgroup namespace, the "/proc/$PID/cgroup" file shows the complete path of the cgroup of a process. In a container setup where a set of cgroups and namespaces are intended to isolate processes the "/proc/$PID/cgroup" file may leak potential system level information to the isolated processes. For example::”h]”hXBWithout cgroup namespace, the â€œ/proc/$PID/cgroupâ€ file shows the complete path of the cgroup of a process. In a container setup where a set of cgroups and namespaces are intended to isolate processes the â€œ/proc/$PID/cgroupâ€ file may leak potential system level information to the isolated processes. For example:”…””}”(hjbOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M=hjCOhžhubjV)”}”(hŒ3# cat /proc/self/cgroup 0::/batchjobs/container_id1”h]”hŒ3# cat /proc/self/cgroup 0::/batchjobs/container_id1”…””}”hjpOsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MChjCOhžhubhæ)”}”(hŒþThe path '/batchjobs/container_id1' can be considered as system-data and undesirable to expose to the isolated processes. cgroup namespace can be used to restrict visibility of this path. For example, before creating a cgroup namespace, one would see::”h]”hXThe path â€˜/batchjobs/container_id1â€™ can be considered as system-data and undesirable to expose to the isolated processes. cgroup namespace can be used to restrict visibility of this path. For example, before creating a cgroup namespace, one would see:”…””}”(hj~OhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MFhjCOhžhubjV)”}”(hŒ¦# ls -l /proc/self/ns/cgroup lrwxrwxrwx 1 root root 0 2014-07-15 10:37 /proc/self/ns/cgroup -> cgroup:[4026531835] # cat /proc/self/cgroup 0::/batchjobs/container_id1”h]”hŒ¦# ls -l /proc/self/ns/cgroup lrwxrwxrwx 1 root root 0 2014-07-15 10:37 /proc/self/ns/cgroup -> cgroup:[4026531835] # cat /proc/self/cgroup 0::/batchjobs/container_id1”…””}”hjŒOsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MKhjCOhžhubhæ)”}”(hŒ3After unsharing a new namespace, the view changes::”h]”hŒ2After unsharing a new namespace, the view changes:”…””}”(hjšOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MPhjCOhžhubjV)”}”(hŒ# ls -l /proc/self/ns/cgroup lrwxrwxrwx 1 root root 0 2014-07-15 10:35 /proc/self/ns/cgroup -> cgroup:[4026532183] # cat /proc/self/cgroup 0::/”h]”hŒ# ls -l /proc/self/ns/cgroup lrwxrwxrwx 1 root root 0 2014-07-15 10:35 /proc/self/ns/cgroup -> cgroup:[4026532183] # cat /proc/self/cgroup 0::/”…””}”hj¨Osbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MRhjCOhžhubhæ)”}”(hŒøWhen some thread from a multi-threaded process unshares its cgroup namespace, the new cgroupns gets applied to the entire process (all the threads). This is natural for the v2 hierarchy; however, for the legacy hierarchies, this may be unexpected.”h]”hŒøWhen some thread from a multi-threaded process unshares its cgroup namespace, the new cgroupns gets applied to the entire process (all the threads). This is natural for the v2 hierarchy; however, for the legacy hierarchies, this may be unexpected.”…””}”(hj¶OhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MWhjCOhžhubhæ)”}”(hŒÌA cgroup namespace is alive as long as there are processes inside or mounts pinning it. When the last usage goes away, the cgroup namespace is destroyed. The cgroupns root and the actual cgroups remain.”h]”hŒÌA cgroup namespace is alive as long as there are processes inside or mounts pinning it. When the last usage goes away, the cgroup namespace is destroyed. The cgroupns root and the actual cgroups remain.”…””}”(hjÄOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M\hjCOhžhubeh}”(h]”Œbasics”ah ]”h"]”Œbasics”ah$]”h&]”uh1h°hj2OhžhhŸh¯h M3ubh±)”}”(hhh]”(h¶)”}”(hŒThe Root and Views”h]”hŒThe Root and Views”…””}”(hjÝOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjÚOhžhhŸh¯h Mcubhæ)”}”(hX1The 'cgroupns root' for a cgroup namespace is the cgroup in which the process calling unshare(2) is running. For example, if a process in /batchjobs/container_id1 cgroup calls unshare, cgroup /batchjobs/container_id1 becomes the cgroupns root. For the init_cgroup_ns, this is the real root ('/') cgroup.”h]”hX9The â€˜cgroupns rootâ€™ for a cgroup namespace is the cgroup in which the process calling unshare(2) is running. For example, if a process in /batchjobs/container_id1 cgroup calls unshare, cgroup /batchjobs/container_id1 becomes the cgroupns root. For the init_cgroup_ns, this is the real root (â€˜/â€™) cgroup.”…””}”(hjëOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MehjÚOhžhubhæ)”}”(hŒrThe cgroupns root cgroup does not change even if the namespace creator process later moves to a different cgroup::”h]”hŒqThe cgroupns root cgroup does not change even if the namespace creator process later moves to a different cgroup:”…””}”(hjùOhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MkhjÚOhžhubjV)”}”(hŒª# ~/unshare -c # unshare cgroupns in some cgroup # cat /proc/self/cgroup 0::/ # mkdir sub_cgrp_1 # echo 0 > sub_cgrp_1/cgroup.procs # cat /proc/self/cgroup 0::/sub_cgrp_1”h]”hŒª# ~/unshare -c # unshare cgroupns in some cgroup # cat /proc/self/cgroup 0::/ # mkdir sub_cgrp_1 # echo 0 > sub_cgrp_1/cgroup.procs # cat /proc/self/cgroup 0::/sub_cgrp_1”…””}”hjPsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MnhjÚOhžhubhæ)”}”(hŒDEach process gets its namespace-specific view of "/proc/$PID/cgroup"”h]”hŒHEach process gets its namespace-specific view of â€œ/proc/$PID/cgroupâ€”…””}”(hjPhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MvhjÚOhžhubhæ)”}”(hŒ§Processes running inside the cgroup namespace will be able to see cgroup paths (in /proc/self/cgroup) only inside their root cgroup. From within an unshared cgroupns::”h]”hŒ¦Processes running inside the cgroup namespace will be able to see cgroup paths (in /proc/self/cgroup) only inside their root cgroup. From within an unshared cgroupns:”…””}”(hj#PhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MxhjÚOhžhubjV)”}”(hŒf# sleep 100000 & [1] 7353 # echo 7353 > sub_cgrp_1/cgroup.procs # cat /proc/7353/cgroup 0::/sub_cgrp_1”h]”hŒf# sleep 100000 & [1] 7353 # echo 7353 > sub_cgrp_1/cgroup.procs # cat /proc/7353/cgroup 0::/sub_cgrp_1”…””}”hj1Psbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M|hjÚOhžhubhæ)”}”(hŒIFrom the initial cgroup namespace, the real cgroup path will be visible::”h]”hŒHFrom the initial cgroup namespace, the real cgroup path will be visible:”…””}”(hj?PhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M‚hjÚOhžhubjV)”}”(hŒ>$ cat /proc/7353/cgroup 0::/batchjobs/container_id1/sub_cgrp_1”h]”hŒ>$ cat /proc/7353/cgroup 0::/batchjobs/container_id1/sub_cgrp_1”…””}”hjMPsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M…hjÚOhžhubhæ)”}”(hXFrom a sibling cgroup namespace (that is, a namespace rooted at a different cgroup), the cgroup path relative to its own cgroup namespace root will be shown. For instance, if PID 7353's cgroup namespace root is at '/batchjobs/container_id2', then it will see::”h]”hX From a sibling cgroup namespace (that is, a namespace rooted at a different cgroup), the cgroup path relative to its own cgroup namespace root will be shown. For instance, if PID 7353â€™s cgroup namespace root is at â€˜/batchjobs/container_id2â€™, then it will see:”…””}”(hj[PhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MˆhjÚOhžhubjV)”}”(hŒ7# cat /proc/7353/cgroup 0::/../container_id2/sub_cgrp_1”h]”hŒ7# cat /proc/7353/cgroup 0::/../container_id2/sub_cgrp_1”…””}”hjiPsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MhjÚOhžhubhæ)”}”(hŒ|Note that the relative path always starts with '/' to indicate that its relative to the cgroup namespace root of the caller.”h]”hŒ€Note that the relative path always starts with â€˜/â€™ to indicate that its relative to the cgroup namespace root of the caller.”…””}”(hjwPhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjÚOhžhubeh}”(h]”Œthe-root-and-views”ah ]”h"]”Œthe root and views”ah$]”h&]”uh1h°hj2OhžhhŸh¯h Mcubh±)”}”(hhh]”(h¶)”}”(hŒMigration and setns(2)”h]”hŒMigration and setns(2)”…””}”(hjPhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjPhžhhŸh¯h M•ubhæ)”}”(hX"Processes inside a cgroup namespace can move into and out of the namespace root if they have proper access to external cgroups. For example, from inside a namespace with cgroupns root at /batchjobs/container_id1, and assuming that the global hierarchy is still accessible inside cgroupns::”h]”hX!Processes inside a cgroup namespace can move into and out of the namespace root if they have proper access to external cgroups. For example, from inside a namespace with cgroupns root at /batchjobs/container_id1, and assuming that the global hierarchy is still accessible inside cgroupns:”…””}”(hjžPhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M—hjPhžhubjV)”}”(hŒ†# cat /proc/7353/cgroup 0::/sub_cgrp_1 # echo 7353 > batchjobs/container_id2/cgroup.procs # cat /proc/7353/cgroup 0::/../container_id2”h]”hŒ†# cat /proc/7353/cgroup 0::/sub_cgrp_1 # echo 7353 > batchjobs/container_id2/cgroup.procs # cat /proc/7353/cgroup 0::/../container_id2”…””}”hj¬Psbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h MhjPhžhubhæ)”}”(hŒ…Note that this kind of setup is not encouraged. A task inside cgroup namespace should only be exposed to its own cgroupns hierarchy.”h]”hŒ…Note that this kind of setup is not encouraged. A task inside cgroup namespace should only be exposed to its own cgroupns hierarchy.”…””}”(hjºPhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M£hjPhžhubhæ)”}”(hŒ5setns(2) to another cgroup namespace is allowed when:”h]”hŒ5setns(2) to another cgroup namespace is allowed when:”…””}”(hjÈPhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¦hjPhžhubjWF)”}”(hhh]”(j¯)”}”(hŒ@the process has CAP_SYS_ADMIN against its current user namespace”h]”hæ)”}”(hjÛPh]”hŒ@the process has CAP_SYS_ADMIN against its current user namespace”…””}”(hjÝPhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¨hjÙPubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjÖPhžhhŸh¯h Nubj¯)”}”(hŒKthe process has CAP_SYS_ADMIN against the target cgroup namespace's userns ”h]”hæ)”}”(hŒJthe process has CAP_SYS_ADMIN against the target cgroup namespace's userns”h]”hŒLthe process has CAP_SYS_ADMIN against the target cgroup namespaceâ€™s userns”…””}”(hjôPhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M©hjðPubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjÖPhžhhŸh¯h Nubeh}”(h]”h ]”h"]”h$]”h&]”j¨FŒ loweralpha”jªFŒ(”j«Fj¬Fuh1jVFhjPhžhhŸh¯h M¨ubhæ)”}”(hŒ²No implicit cgroup changes happen with attaching to another cgroup namespace. It is expected that the someone moves the attaching process under the target cgroup namespace root.”h]”hŒ²No implicit cgroup changes happen with attaching to another cgroup namespace. It is expected that the someone moves the attaching process under the target cgroup namespace root.”…””}”(hjQhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¬hjPhžhubeh}”(h]”Œmigration-and-setns-2”ah ]”h"]”Œmigration and setns(2)”ah$]”h&]”uh1h°hj2OhžhhŸh¯h M•ubh±)”}”(hhh]”(h¶)”}”(hŒ!Interaction with Other Namespaces”h]”hŒ!Interaction with Other Namespaces”…””}”(hj)QhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj&QhžhhŸh¯h M²ubhæ)”}”(hŒlNamespace specific cgroup hierarchy can be mounted by a process running inside a non-init cgroup namespace::”h]”hŒkNamespace specific cgroup hierarchy can be mounted by a process running inside a non-init cgroup namespace:”…””}”(hj7QhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M´hj&QhžhubjV)”}”(hŒ$# mount -t cgroup2 none $MOUNT_POINT”h]”hŒ$# mount -t cgroup2 none $MOUNT_POINT”…””}”hjEQsbah}”(h]”h ]”h"]”h$]”h&]”j’j“uh1jUhŸh¯h M·hj&Qhžhubhæ)”}”(hŒŸThis will mount the unified cgroup hierarchy with cgroupns root as the filesystem root. The process needs CAP_SYS_ADMIN against its user and mount namespaces.”h]”hŒŸThis will mount the unified cgroup hierarchy with cgroupns root as the filesystem root. The process needs CAP_SYS_ADMIN against its user and mount namespaces.”…””}”(hjSQhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M¹hj&Qhžhubhæ)”}”(hŒÆThe virtualization of /proc/self/cgroup file combined with restricting the view of cgroup hierarchy by namespace-private cgroupfs mount provides a properly isolated cgroup view inside the container.”h]”hŒÆThe virtualization of /proc/self/cgroup file combined with restricting the view of cgroup hierarchy by namespace-private cgroupfs mount provides a properly isolated cgroup view inside the container.”…””}”(hjaQhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M½hj&Qhžhubeh}”(h]”Œ!interaction-with-other-namespaces”ah ]”h"]”Œ!interaction with other namespaces”ah$]”h&]”uh1h°hj2OhžhhŸh¯h M²ubeh}”(h]”Œ namespace”ah ]”h"]”Œ namespace”ah$]”h&]”uh1h°hh²hžhhŸh¯h M0ubh±)”}”(hhh]”(h¶)”}”(hŒ!Information on Kernel Programming”h]”hŒ!Information on Kernel Programming”…””}”(hj‚QhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjQhžhhŸh¯h MÃubhæ)”}”(hŒ›This section contains kernel programming information in the areas where interacting with cgroup is necessary. cgroup core and controllers are not covered.”h]”hŒ›This section contains kernel programming information in the areas where interacting with cgroup is necessary. cgroup core and controllers are not covered.”…””}”(hjQhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÅhjQhžhubh±)”}”(hhh]”(h¶)”}”(hŒ Filesystem Support for Writeback”h]”hŒ Filesystem Support for Writeback”…””}”(hj¡QhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjžQhžhhŸh¯h MËubhæ)”}”(hŒ“A filesystem can support cgroup writeback by updating address_space_operations->writepage[s]() to annotate bio's using the following two functions.”h]”hŒ•A filesystem can support cgroup writeback by updating address_space_operations->writepage[s]() to annotate bioâ€™s using the following two functions.”…””}”(hj¯QhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÍhjžQhžhubj¬)”}”(hXWwbc_init_bio(@wbc, @bio) Should be called for each bio carrying writeback data and associates the bio with the inode's owner cgroup and the corresponding request queue. This must be called after a queue (device) has been associated with the bio and before submission. wbc_account_cgroup_owner(@wbc, @folio, @bytes) Should be called for each data segment being written out. While this function doesn't care exactly when it's called during the writeback session, it's the easiest and most natural to call it as data segments are added to a bio. ”h]”j²)”}”(hhh]”(j·)”}”(hX wbc_init_bio(@wbc, @bio) Should be called for each bio carrying writeback data and associates the bio with the inode's owner cgroup and the corresponding request queue. This must be called after a queue (device) has been associated with the bio and before submission. ”h]”(j½)”}”(hŒwbc_init_bio(@wbc, @bio)”h]”hŒwbc_init_bio(@wbc, @bio)”…””}”(hjÈQhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÖhjÄQubjÍ)”}”(hhh]”hæ)”}”(hŒóShould be called for each bio carrying writeback data and associates the bio with the inode's owner cgroup and the corresponding request queue. This must be called after a queue (device) has been associated with the bio and before submission.”h]”hŒõShould be called for each bio carrying writeback data and associates the bio with the inodeâ€™s owner cgroup and the corresponding request queue. This must be called after a queue (device) has been associated with the bio and before submission.”…””}”(hjÙQhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÒhjÖQubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjÄQubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÖhjÁQubj·)”}”(hXwbc_account_cgroup_owner(@wbc, @folio, @bytes) Should be called for each data segment being written out. While this function doesn't care exactly when it's called during the writeback session, it's the easiest and most natural to call it as data segments are added to a bio. ”h]”(j½)”}”(hŒ.wbc_account_cgroup_owner(@wbc, @folio, @bytes)”h]”hŒ.wbc_account_cgroup_owner(@wbc, @folio, @bytes)”…””}”(hj÷QhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j¼hŸh¯h MÜhjóQubjÍ)”}”(hhh]”hæ)”}”(hŒãShould be called for each data segment being written out. While this function doesn't care exactly when it's called during the writeback session, it's the easiest and most natural to call it as data segments are added to a bio.”h]”hŒéShould be called for each data segment being written out. While this function doesnâ€™t care exactly when itâ€™s called during the writeback session, itâ€™s the easiest and most natural to call it as data segments are added to a bio.”…””}”(hjRhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÙhjRubah}”(h]”h ]”h"]”h$]”h&]”uh1jÌhjóQubeh}”(h]”h ]”h"]”h$]”h&]”uh1j¶hŸh¯h MÜhjÁQubeh}”(h]”h ]”h"]”h$]”h&]”uh1j±hj½Qubah}”(h]”h ]”h"]”h$]”h&]”uh1j«hŸh¯h MÑhjžQhžhubhæ)”}”(hXWith writeback bio's annotated, cgroup support can be enabled per super_block by setting SB_I_CGROUPWB in ->s_iflags. This allows for selective disabling of cgroup writeback support which is helpful when certain filesystem features, e.g. journaled data mode, are incompatible.”h]”hXWith writeback bioâ€™s annotated, cgroup support can be enabled per super_block by setting SB_I_CGROUPWB in ->s_iflags. This allows for selective disabling of cgroup writeback support which is helpful when certain filesystem features, e.g. journaled data mode, are incompatible.”…””}”(hj.RhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MÞhjžQhžhubhæ)”}”(hX¡wbc_init_bio() binds the specified bio to its cgroup. Depending on the configuration, the bio may be executed at a lower priority and if the writeback session is holding shared resources, e.g. a journal entry, may lead to priority inversion. There is no one easy solution for the problem. Filesystems can try to work around specific problem cases by skipping wbc_init_bio() and using bio_associate_blkg() directly.”h]”hX¡wbc_init_bio() binds the specified bio to its cgroup. Depending on the configuration, the bio may be executed at a lower priority and if the writeback session is holding shared resources, e.g. a journal entry, may lead to priority inversion. There is no one easy solution for the problem. Filesystems can try to work around specific problem cases by skipping wbc_init_bio() and using bio_associate_blkg() directly.”…””}”(hjThe "tasks" file is removed and "cgroup.procs" is not sorted. ”h]”hæ)”}”(hŒ=The "tasks" file is removed and "cgroup.procs" is not sorted.”h]”hŒEThe â€œtasksâ€ file is removed and â€œcgroup.procsâ€ is not sorted.”…””}”(hj¢RhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MôhjžRubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjkRhžhhŸh¯h Nubj¯)”}”(hŒ$"cgroup.clone_children" is removed. ”h]”hæ)”}”(hŒ#"cgroup.clone_children" is removed.”h]”hŒ'â€œcgroup.clone_childrenâ€ is removed.”…””}”(hjºRhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h Möhj¶Rubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjkRhžhhŸh¯h Nubj¯)”}”(hŒl/proc/cgroups is meaningless for v2. Use "cgroup.controllers" or "cgroup.stat" files at the root instead. ”h]”hæ)”}”(hŒj/proc/cgroups is meaningless for v2. Use "cgroup.controllers" or "cgroup.stat" files at the root instead.”h]”hŒr/proc/cgroups is meaningless for v2. Use â€œcgroup.controllersâ€ or â€œcgroup.statâ€ files at the root instead.”…””}”(hjÒRhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MøhjÎRubah}”(h]”h ]”h"]”h$]”h&]”uh1j®hjkRhžhhŸh¯h Nubeh}”(h]”h ]”h"]”h$]”h&]”jjóuh1j©hŸh¯h MðhjZRhžhubeh}”(h]”Œdeprecated-v1-core-features”ah ]”h"]”Œdeprecated v1 core features”ah$]”h&]”uh1h°hh²hžhhŸh¯h Mîubh±)”}”(hhh]”(h¶)”}”(hŒ$Issues with v1 and Rationales for v2”h]”hŒ$Issues with v1 and Rationales for v2”…””}”(hj÷RhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjôRhžhhŸh¯h Mýubh±)”}”(hhh]”(h¶)”}”(hŒMultiple Hierarchies”h]”hŒMultiple Hierarchies”…””}”(hjShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjShžhhŸh¯h Mubhæ)”}”(hŒÆcgroup v1 allowed an arbitrary number of hierarchies and each hierarchy could host any number of controllers. While this seemed to provide a high level of flexibility, it wasn't useful in practice.”h]”hŒÈcgroup v1 allowed an arbitrary number of hierarchies and each hierarchy could host any number of controllers. While this seemed to provide a high level of flexibility, it wasnâ€™t useful in practice.”…””}”(hjShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjShžhubhæ)”}”(hXóFor example, as there is only one instance of each controller, utility type controllers such as freezer which can be useful in all hierarchies could only be used in one. The issue is exacerbated by the fact that controllers couldn't be moved to another hierarchy once hierarchies were populated. Another issue was that all controllers bound to a hierarchy were forced to have exactly the same view of the hierarchy. It wasn't possible to vary the granularity depending on the specific controller.”h]”hX÷For example, as there is only one instance of each controller, utility type controllers such as freezer which can be useful in all hierarchies could only be used in one. The issue is exacerbated by the fact that controllers couldnâ€™t be moved to another hierarchy once hierarchies were populated. Another issue was that all controllers bound to a hierarchy were forced to have exactly the same view of the hierarchy. It wasnâ€™t possible to vary the granularity depending on the specific controller.”…””}”(hj$ShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjShžhubhæ)”}”(hXÔIn practice, these issues heavily limited which controllers could be put on the same hierarchy and most configurations resorted to putting each controller on its own hierarchy. Only closely related ones, such as the cpu and cpuacct controllers, made sense to be put on the same hierarchy. This often meant that userland ended up managing multiple similar hierarchies repeating the same steps on each hierarchy whenever a hierarchy management operation was necessary.”h]”hXÔIn practice, these issues heavily limited which controllers could be put on the same hierarchy and most configurations resorted to putting each controller on its own hierarchy. Only closely related ones, such as the cpu and cpuacct controllers, made sense to be put on the same hierarchy. This often meant that userland ended up managing multiple similar hierarchies repeating the same steps on each hierarchy whenever a hierarchy management operation was necessary.”…””}”(hj2ShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjShžhubhæ)”}”(hXFurthermore, support for multiple hierarchies came at a steep cost. It greatly complicated cgroup core implementation but more importantly the support for multiple hierarchies restricted how cgroup could be used in general and what controllers was able to do.”h]”hXFurthermore, support for multiple hierarchies came at a steep cost. It greatly complicated cgroup core implementation but more importantly the support for multiple hierarchies restricted how cgroup could be used in general and what controllers was able to do.”…””}”(hj@ShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjShžhubhæ)”}”(hXªThere was no limit on how many hierarchies there might be, which meant that a thread's cgroup membership couldn't be described in finite length. The key might contain any number of entries and was unlimited in length, which made it highly awkward to manipulate and led to addition of controllers which existed only to identify membership, which in turn exacerbated the original problem of proliferating number of hierarchies.”h]”hX®There was no limit on how many hierarchies there might be, which meant that a threadâ€™s cgroup membership couldnâ€™t be described in finite length. The key might contain any number of entries and was unlimited in length, which made it highly awkward to manipulate and led to addition of controllers which existed only to identify membership, which in turn exacerbated the original problem of proliferating number of hierarchies.”…””}”(hjNShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MhjShžhubhæ)”}”(hXLAlso, as a controller couldn't have any expectation regarding the topologies of hierarchies other controllers might be on, each controller had to assume that all other controllers were attached to completely orthogonal hierarchies. This made it impossible, or at least very cumbersome, for controllers to cooperate with each other.”h]”hXNAlso, as a controller couldnâ€™t have any expectation regarding the topologies of hierarchies other controllers might be on, each controller had to assume that all other controllers were attached to completely orthogonal hierarchies. This made it impossible, or at least very cumbersome, for controllers to cooperate with each other.”…””}”(hj\ShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M$hjShžhubhæ)”}”(hXIn most use cases, putting controllers on hierarchies which are completely orthogonal to each other isn't necessary. What usually is called for is the ability to have differing levels of granularity depending on the specific controller. In other words, hierarchy may be collapsed from leaf towards root when viewed from specific controllers. For example, a given configuration might not care about how memory is distributed beyond a certain level while still wanting to control how CPU cycles are distributed.”h]”hXIn most use cases, putting controllers on hierarchies which are completely orthogonal to each other isnâ€™t necessary. What usually is called for is the ability to have differing levels of granularity depending on the specific controller. In other words, hierarchy may be collapsed from leaf towards root when viewed from specific controllers. For example, a given configuration might not care about how memory is distributed beyond a certain level while still wanting to control how CPU cycles are distributed.”…””}”(hjjShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M*hjShžhubeh}”(h]”Œmultiple-hierarchies”ah ]”h"]”Œmultiple hierarchies”ah$]”h&]”uh1h°hjôRhžhhŸh¯h Mubh±)”}”(hhh]”(h¶)”}”(hŒThread Granularity”h]”hŒThread Granularity”…””}”(hjƒShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhj€ShžhhŸh¯h M5ubhæ)”}”(hXEcgroup v1 allowed threads of a process to belong to different cgroups. This didn't make sense for some controllers and those controllers ended up implementing different ways to ignore such situations but much more importantly it blurred the line between API exposed to individual applications and system management interface.”h]”hXGcgroup v1 allowed threads of a process to belong to different cgroups. This didnâ€™t make sense for some controllers and those controllers ended up implementing different ways to ignore such situations but much more importantly it blurred the line between API exposed to individual applications and system management interface.”…””}”(hj‘ShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M7hj€Shžhubhæ)”}”(hŒòGenerally, in-process knowledge is available only to the process itself; thus, unlike service-level organization of processes, categorizing threads of a process requires active participation from the application which owns the target process.”h]”hŒòGenerally, in-process knowledge is available only to the process itself; thus, unlike service-level organization of processes, categorizing threads of a process requires active participation from the application which owns the target process.”…””}”(hjŸShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M=hj€Shžhubhæ)”}”(hXjcgroup v1 had an ambiguously defined delegation model which got abused in combination with thread granularity. cgroups were delegated to individual applications so that they can create and manage their own sub-hierarchies and control resource distributions along them. This effectively raised cgroup to the status of a syscall-like API exposed to lay programs.”h]”hXjcgroup v1 had an ambiguously defined delegation model which got abused in combination with thread granularity. cgroups were delegated to individual applications so that they can create and manage their own sub-hierarchies and control resource distributions along them. This effectively raised cgroup to the status of a syscall-like API exposed to lay programs.”…””}”(hjShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MBhj€Shžhubhæ)”}”(hX)First of all, cgroup has a fundamentally inadequate interface to be exposed this way. For a process to access its own knobs, it has to extract the path on the target hierarchy from /proc/self/cgroup, construct the path by appending the name of the knob to the path, open and then read and/or write to it. This is not only extremely clunky and unusual but also inherently racy. There is no conventional way to define transaction across the required steps and nothing can guarantee that the process would actually be operating on its own sub-hierarchy.”h]”hX)First of all, cgroup has a fundamentally inadequate interface to be exposed this way. For a process to access its own knobs, it has to extract the path on the target hierarchy from /proc/self/cgroup, construct the path by appending the name of the knob to the path, open and then read and/or write to it. This is not only extremely clunky and unusual but also inherently racy. There is no conventional way to define transaction across the required steps and nothing can guarantee that the process would actually be operating on its own sub-hierarchy.”…””}”(hj»ShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MIhj€Shžhubhæ)”}”(hXcgroup controllers implemented a number of knobs which would never be accepted as public APIs because they were just adding control knobs to system-management pseudo filesystem. cgroup ended up with interface knobs which were not properly abstracted or refined and directly revealed kernel internal details. These knobs got exposed to individual applications through the ill-defined delegation mechanism effectively abusing cgroup as a shortcut to implementing public APIs without going through the required scrutiny.”h]”hXcgroup controllers implemented a number of knobs which would never be accepted as public APIs because they were just adding control knobs to system-management pseudo filesystem. cgroup ended up with interface knobs which were not properly abstracted or refined and directly revealed kernel internal details. These knobs got exposed to individual applications through the ill-defined delegation mechanism effectively abusing cgroup as a shortcut to implementing public APIs without going through the required scrutiny.”…””}”(hjÉShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MRhj€Shžhubhæ)”}”(hŒ±This was painful for both userland and kernel. Userland ended up with misbehaving and poorly abstracted interfaces and kernel exposing and locked into constructs inadvertently.”h]”hŒ±This was painful for both userland and kernel. Userland ended up with misbehaving and poorly abstracted interfaces and kernel exposing and locked into constructs inadvertently.”…””}”(hj×ShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h M[hj€Shžhubeh}”(h]”Œthread-granularity”ah ]”h"]”Œthread granularity”ah$]”h&]”uh1h°hjôRhžhhŸh¯h M5ubh±)”}”(hhh]”(h¶)”}”(hŒ+Competition Between Inner Nodes and Threads”h]”hŒ+Competition Between Inner Nodes and Threads”…””}”(hjðShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hµhjíShžhhŸh¯h Maubhæ)”}”(hXCcgroup v1 allowed threads to be in any cgroups which created an interesting problem where threads belonging to a parent cgroup and its children cgroups competed for resources. This was nasty as two different types of entities competed and there was no obvious way to settle it. Different controllers did different things.”h]”hXCcgroup v1 allowed threads to be in any cgroups which created an interesting problem where threads belonging to a parent cgroup and its children cgroups competed for resources. This was nasty as two different types of entities competed and there was no obvious way to settle it. Different controllers did different things.”…””}”(hjþShžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MchjíShžhubhæ)”}”(hXThe cpu controller considered threads and cgroups as equivalents and mapped nice levels to cgroup weights. This worked for some cases but fell flat when children wanted to be allocated specific ratios of CPU cycles and the number of internal threads fluctuated - the ratios constantly changed as the number of competing entities fluctuated. There also were other issues. The mapping from nice level to weight wasn't obvious or universal, and there were various other knobs which simply weren't available for threads.”h]”hX The cpu controller considered threads and cgroups as equivalents and mapped nice levels to cgroup weights. This worked for some cases but fell flat when children wanted to be allocated specific ratios of CPU cycles and the number of internal threads fluctuated - the ratios constantly changed as the number of competing entities fluctuated. There also were other issues. The mapping from nice level to weight wasnâ€™t obvious or universal, and there were various other knobs which simply werenâ€™t available for threads.”…””}”(hjThžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MihjíShžhubhæ)”}”(hX¤The io controller implicitly created a hidden leaf node for each cgroup to host the threads. The hidden leaf had its own copies of all the knobs with ``leaf_`` prefixed. While this allowed equivalent control over internal threads, it was with serious drawbacks. It always added an extra layer of nesting which wouldn't be necessary otherwise, made the interface messy and significantly complicated the implementation.”h]”(hŒ—The io controller implicitly created a hidden leaf node for each cgroup to host the threads. The hidden leaf had its own copies of all the knobs with ”…””}”(hjThžhhŸNh NubhŒliteral”“”)”}”(hŒ ``leaf_``”h]”hŒleaf_”…””}”(hj$ThžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1j"ThjTubhX prefixed. While this allowed equivalent control over internal threads, it was with serious drawbacks. It always added an extra layer of nesting which wouldnâ€™t be necessary otherwise, made the interface messy and significantly complicated the implementation.”…””}”(hjThžhhŸNh Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1håhŸh¯h MrhjíShžhubhæ)”}”(hXIThe memory controller didn't have a way to control what happened between internal tasks and child cgroups and the behavior was not clearly defined. There were attempts to add ad-hoc behaviors and knobs to tailor the behavior to specific workloads which would have led to problems extremely difficult to resolve in the long term.”h]”hXKThe memory controller didnâ€™t have a way to control what happened between internal tasks and child cgroups and the behavior was not clearly defined. There were attempts to add ad-hoc behaviors and knobs to tailor the behavior to specific workloads which would have led to problems extremely difficult to resolve in the long term.”…””}”(hjj™>j8j8j÷8jô8jH9jE9j¶9j³9jø:jõ:j”>j‘>jp@jm@jh@je@j&Gj#GjGjGjwGjtGj$Ij!IjIjIjfJjcJj^Jj[JjæKjãKjÞKjÛKjdNjaNj5Nj2Nj\NjYNj¤Nj¡NjœNj™Nj'Oj$OjøNjõNjOjOj|QjyQj×OjÔOjŠPj‡Pj#Qj QjtQjqQjWRjTRjORjLRjñRjîRjsUjpUj}SjzSjêSjçSjkTjhTj¼Tj¹TjkUjhUuŒ nametypes”}”(j|Uˆj{U‰j"‰jÉ‰j‰j¯ ‰jd‰j‰j(‰j…‰jÐ‰jø‰jx‰jŸ‰jð‰j ‰jZ‰j ‰j§ ‰j\ ‰jŸ ‰j8‰j* ‰j‡ ˆj† ‰jÝ ˆjÜ ‰j0‰j¨‰jš‰j2 ‰j ‰j/O‰jéˆjè‰jà‰jëU‰j .‰jc.‰j¦.‰jœ>‰j8‰j÷8‰jH9‰j¶9‰jø:‰j”>‰jp@‰jh@‰j&G‰jG‰jwG‰j$I‰jI‰jfJ‰j^J‰jæK‰jÞK‰jdN‰j5N‰j\N‰j¤N‰jœN‰j'O‰jøN‰jO‰j|Q‰j×O‰jŠP‰j#Q‰jtQ‰jWR‰jOR‰jñR‰jsU‰j}S‰jêS‰jkT‰j¼T‰jkU‰uh}”(h®h²jxUh²jj”jÆj¥jjÌj¬ j%jaj6jŠjgj%jxj‚j+jÍjjõjÓjujäjœj{jíj¢j jûjWjj j]j¤ j jY j* jœ j_ j5j² j' jÑ j$ j- jƒ j- j€ jŽ jÙ jŽ j-jâ j¥j;j—jLj/ jjj5 j,Oj«jÆjÇjåjÇjÝjj«.jîj.jzj`.j#.j£.jf.j™>j².j8jÑ.jô8j8jE9jú8j³9jK9jõ:j¹9j‘>jû:jm@jŸ>je@jÚ>j#Gjs@jGj @jtGj)Gj!IjzGjIj™GjcJj'Ij[JjFIjãKjiJjÛKjˆJjaNjéKj2Nj$LjYNj8Nj¡NjgNj™NjxNj$Oj§NjõNjÆNjOjûNjyQj2OjÔOjCOj‡PjÚOj QjPjqQj&QjTRjQjLRjžQjîRjZRjpUjôRjzSjSjçSj€SjhTjíSj¹TjnTjhUj¿TjaUjÐTuŒ footnote_refs”}”Œ citation_refs”}”Œ autofootnotes”]”Œautofootnote_refs”]”Œsymbol_footnotes”]”Œsymbol_footnote_refs”]”Œ footnotes”]”Œ citations”]”Œautofootnote_start”KŒsymbol_footnote_start”KŒ id_counter”Œcollections”ŒCounter”“”}”j±UKs…”R”Œparse_messages”]”hŒsystem_message”“”)”}”(hhh]”hæ)”}”(hŒ)Duplicate implicit target name: "memory".”h]”hŒ-Duplicate implicit target name: â€œmemoryâ€.”…””}”(hjVhžhhŸNh Nubah}”(h]”h ]”h"]”h$]”h&]”uh1håhjVubah}”(h]”h ]”h"]”h$]”h&]”jaUaŒlevel”KŒtype”ŒINFO”Œsource”h¯Œline”Mªuh1j VhjÐThžhhŸh¯h MªubaŒtransform_messages”]”(jV)”}”(hhh]”hæ)”}”(hhh]”hŒ/Hyperlink target "cgroup-v2" is not referenced.”…””}”hj-Vsbah}”(h]”h ]”h"]”h$]”h&]”uh1håhj*Vubah}”(h]”h ]”h"]”h$]”h&]”Œlevel”KŒtype”j%VŒsource”h¯Œline”Kuh1j VubjV)”}”(hhh]”hæ)”}”(hhh]”hŒAHyperlink target "cgroupv2-limits-distributor" is not referenced.”…””}”hjGVsbah}”(h]”h ]”h"]”h$]”h&]”uh1håhjDVubah}”(h]”h ]”h"]”h$]”h&]”Œlevel”KŒtype”j%VŒsource”h¯Œline”Mœuh1j VubjV)”}”(hhh]”hæ)”}”(hhh]”hŒFHyperlink target "cgroupv2-protections-distributor" is not referenced.”…””}”hjaVsbah}”(h]”h ]”h"]”h$]”h&]”uh1håhj^Vubah}”(h]”h ]”h"]”h$]”h&]”Œlevel”KŒtype”j%VŒsource”h¯Œline”M®uh1j VubjV)”}”(hhh]”hæ)”}”(hhh]”hŒ3Hyperlink target "cgroup-v2-cpu" is not referenced.”…””}”hj{Vsbah}”(h]”h ]”h"]”h$]”h&]”uh1håhjxVubah}”(h]”h ]”h"]”h$]”h&]”Œlevel”KŒtype”j%VŒsource”h¯Œline”M&uh1j VubeŒtransformer”NŒinclude_log”]”Œ decoration”Nhžhub.