8 :Author: Tejun Heo <tj@kernel.org>
10 This is the authoritative documentation on the design, interface and
11 conventions of cgroup v2. It describes all userland-visible aspects
12 of cgroup including core and specific controller behaviors. All
13 future changes must be reflected in this document. Documentation for
14 v1 is available under :ref:`Documentation/admin-guide/cgroup-v1/index.rst <cgroup-v1>`.
23 2-2. Organizing Processes and Threads
26 2-3. [Un]populated Notification
27 2-4. Controlling Controllers
28 2-4-1. Enabling and Disabling
29 2-4-2. Top-down Constraint
30 2-4-3. No Internal Process Constraint
32 2-5-1. Model of Delegation
33 2-5-2. Delegation Containment
35 2-6-1. Organize Once and Control
36 2-6-2. Avoid Name Collisions
37 3. Resource Distribution Models
45 4-3. Core Interface Files
48 5-1-1. CPU Interface Files
50 5-2-1. Memory Interface Files
51 5-2-2. Usage Guidelines
52 5-2-3. Memory Ownership
54 5-3-1. IO Interface Files
57 5-3-3-1. How IO Latency Throttling Works
58 5-3-3-2. IO Latency Interface Files
61 5-4-1. PID Interface Files
63 5.5-1. Cpuset Interface Files
66 5-7-1. RDMA Interface Files
68 5.8-1. HugeTLB Interface Files
70 5.9-1 Miscellaneous cgroup Interface Files
71 5.9-2 Migration and Ownership
74 5-N. Non-normative information
75 5-N-1. CPU controller root cgroup process behaviour
76 5-N-2. IO controller root cgroup process behaviour
79 6-2. The Root and Views
80 6-3. Migration and setns(2)
81 6-4. Interaction with Other Namespaces
82 P. Information on Kernel Programming
83 P-1. Filesystem Support for Writeback
84 D. Deprecated v1 Core Features
85 R. Issues with v1 and Rationales for v2
86 R-1. Multiple Hierarchies
87 R-2. Thread Granularity
88 R-3. Competition Between Inner Nodes and Threads
89 R-4. Other Interface Issues
90 R-5. Controller Issues and Remedies
100 "cgroup" stands for "control group" and is never capitalized. The
101 singular form is used to designate the whole feature and also as a
102 qualifier as in "cgroup controllers". When explicitly referring to
103 multiple individual control groups, the plural form "cgroups" is used.
109 cgroup is a mechanism to organize processes hierarchically and
110 distribute system resources along the hierarchy in a controlled and
113 cgroup is largely composed of two parts - the core and controllers.
114 cgroup core is primarily responsible for hierarchically organizing
115 processes. A cgroup controller is usually responsible for
116 distributing a specific type of system resource along the hierarchy
117 although there are utility controllers which serve purposes other than
118 resource distribution.
120 cgroups form a tree structure and every process in the system belongs
121 to one and only one cgroup. All threads of a process belong to the
122 same cgroup. On creation, all processes are put in the cgroup that
123 the parent process belongs to at the time. A process can be migrated
124 to another cgroup. Migration of a process doesn't affect already
125 existing descendant processes.
127 Following certain structural constraints, controllers may be enabled or
128 disabled selectively on a cgroup. All controller behaviors are
129 hierarchical - if a controller is enabled on a cgroup, it affects all
130 processes which belong to the cgroups consisting the inclusive
131 sub-hierarchy of the cgroup. When a controller is enabled on a nested
132 cgroup, it always restricts the resource distribution further. The
133 restrictions set closer to the root in the hierarchy can not be
134 overridden from further away.
143 Unlike v1, cgroup v2 has only single hierarchy. The cgroup v2
144 hierarchy can be mounted with the following mount command::
146 # mount -t cgroup2 none $MOUNT_POINT
148 cgroup2 filesystem has the magic number 0x63677270 ("cgrp"). All
149 controllers which support v2 and are not bound to a v1 hierarchy are
150 automatically bound to the v2 hierarchy and show up at the root.
151 Controllers which are not in active use in the v2 hierarchy can be
152 bound to other hierarchies. This allows mixing v2 hierarchy with the
153 legacy v1 multiple hierarchies in a fully backward compatible way.
155 A controller can be moved across hierarchies only after the controller
156 is no longer referenced in its current hierarchy. Because per-cgroup
157 controller states are destroyed asynchronously and controllers may
158 have lingering references, a controller may not show up immediately on
159 the v2 hierarchy after the final umount of the previous hierarchy.
160 Similarly, a controller should be fully disabled to be moved out of
161 the unified hierarchy and it may take some time for the disabled
162 controller to become available for other hierarchies; furthermore, due
163 to inter-controller dependencies, other controllers may need to be
166 While useful for development and manual configurations, moving
167 controllers dynamically between the v2 and other hierarchies is
168 strongly discouraged for production use. It is recommended to decide
169 the hierarchies and controller associations before starting using the
170 controllers after system boot.
172 During transition to v2, system management software might still
173 automount the v1 cgroup filesystem and so hijack all controllers
174 during boot, before manual intervention is possible. To make testing
175 and experimenting easier, the kernel parameter cgroup_no_v1= allows
176 disabling controllers in v1 and make them always available in v2.
178 cgroup v2 currently supports the following mount options.
181 Consider cgroup namespaces as delegation boundaries. This
182 option is system wide and can only be set on mount or modified
183 through remount from the init namespace. The mount option is
184 ignored on non-init namespace mounts. Please refer to the
185 Delegation section for details.
188 Reduce the latencies of dynamic cgroup modifications such as
189 task migrations and controller on/offs at the cost of making
190 hot path operations such as forks and exits more expensive.
191 The static usage pattern of creating a cgroup, enabling
192 controllers, and then seeding it with CLONE_INTO_CGROUP is
193 not affected by this option.
196 Only populate memory.events with data for the current cgroup,
197 and not any subtrees. This is legacy behaviour, the default
198 behaviour without this option is to include subtree counts.
199 This option is system wide and can only be set on mount or
200 modified through remount from the init namespace. The mount
201 option is ignored on non-init namespace mounts.
204 Recursively apply memory.min and memory.low protection to
205 entire subtrees, without requiring explicit downward
206 propagation into leaf cgroups. This allows protecting entire
207 subtrees from one another, while retaining free competition
208 within those subtrees. This should have been the default
209 behavior but is a mount-option to avoid regressing setups
210 relying on the original semantics (e.g. specifying bogusly
211 high 'bypass' protection values at higher tree levels).
214 Organizing Processes and Threads
215 --------------------------------
220 Initially, only the root cgroup exists to which all processes belong.
221 A child cgroup can be created by creating a sub-directory::
225 A given cgroup may have multiple child cgroups forming a tree
226 structure. Each cgroup has a read-writable interface file
227 "cgroup.procs". When read, it lists the PIDs of all processes which
228 belong to the cgroup one-per-line. The PIDs are not ordered and the
229 same PID may show up more than once if the process got moved to
230 another cgroup and then back or the PID got recycled while reading.
232 A process can be migrated into a cgroup by writing its PID to the
233 target cgroup's "cgroup.procs" file. Only one process can be migrated
234 on a single write(2) call. If a process is composed of multiple
235 threads, writing the PID of any thread migrates all threads of the
238 When a process forks a child process, the new process is born into the
239 cgroup that the forking process belongs to at the time of the
240 operation. After exit, a process stays associated with the cgroup
241 that it belonged to at the time of exit until it's reaped; however, a
242 zombie process does not appear in "cgroup.procs" and thus can't be
243 moved to another cgroup.
245 A cgroup which doesn't have any children or live processes can be
246 destroyed by removing the directory. Note that a cgroup which doesn't
247 have any children and is associated only with zombie processes is
248 considered empty and can be removed::
252 "/proc/$PID/cgroup" lists a process's cgroup membership. If legacy
253 cgroup is in use in the system, this file may contain multiple lines,
254 one for each hierarchy. The entry for cgroup v2 is always in the
257 # cat /proc/842/cgroup
259 0::/test-cgroup/test-cgroup-nested
261 If the process becomes a zombie and the cgroup it was associated with
262 is removed subsequently, " (deleted)" is appended to the path::
264 # cat /proc/842/cgroup
266 0::/test-cgroup/test-cgroup-nested (deleted)
272 cgroup v2 supports thread granularity for a subset of controllers to
273 support use cases requiring hierarchical resource distribution across
274 the threads of a group of processes. By default, all threads of a
275 process belong to the same cgroup, which also serves as the resource
276 domain to host resource consumptions which are not specific to a
277 process or thread. The thread mode allows threads to be spread across
278 a subtree while still maintaining the common resource domain for them.
280 Controllers which support thread mode are called threaded controllers.
281 The ones which don't are called domain controllers.
283 Marking a cgroup threaded makes it join the resource domain of its
284 parent as a threaded cgroup. The parent may be another threaded
285 cgroup whose resource domain is further up in the hierarchy. The root
286 of a threaded subtree, that is, the nearest ancestor which is not
287 threaded, is called threaded domain or thread root interchangeably and
288 serves as the resource domain for the entire subtree.
290 Inside a threaded subtree, threads of a process can be put in
291 different cgroups and are not subject to the no internal process
292 constraint - threaded controllers can be enabled on non-leaf cgroups
293 whether they have threads in them or not.
295 As the threaded domain cgroup hosts all the domain resource
296 consumptions of the subtree, it is considered to have internal
297 resource consumptions whether there are processes in it or not and
298 can't have populated child cgroups which aren't threaded. Because the
299 root cgroup is not subject to no internal process constraint, it can
300 serve both as a threaded domain and a parent to domain cgroups.
302 The current operation mode or type of the cgroup is shown in the
303 "cgroup.type" file which indicates whether the cgroup is a normal
304 domain, a domain which is serving as the domain of a threaded subtree,
305 or a threaded cgroup.
307 On creation, a cgroup is always a domain cgroup and can be made
308 threaded by writing "threaded" to the "cgroup.type" file. The
309 operation is single direction::
311 # echo threaded > cgroup.type
313 Once threaded, the cgroup can't be made a domain again. To enable the
314 thread mode, the following conditions must be met.
316 - As the cgroup will join the parent's resource domain. The parent
317 must either be a valid (threaded) domain or a threaded cgroup.
319 - When the parent is an unthreaded domain, it must not have any domain
320 controllers enabled or populated domain children. The root is
321 exempt from this requirement.
323 Topology-wise, a cgroup can be in an invalid state. Please consider
324 the following topology::
326 A (threaded domain) - B (threaded) - C (domain, just created)
328 C is created as a domain but isn't connected to a parent which can
329 host child domains. C can't be used until it is turned into a
330 threaded cgroup. "cgroup.type" file will report "domain (invalid)" in
331 these cases. Operations which fail due to invalid topology use
332 EOPNOTSUPP as the errno.
334 A domain cgroup is turned into a threaded domain when one of its child
335 cgroup becomes threaded or threaded controllers are enabled in the
336 "cgroup.subtree_control" file while there are processes in the cgroup.
337 A threaded domain reverts to a normal domain when the conditions
340 When read, "cgroup.threads" contains the list of the thread IDs of all
341 threads in the cgroup. Except that the operations are per-thread
342 instead of per-process, "cgroup.threads" has the same format and
343 behaves the same way as "cgroup.procs". While "cgroup.threads" can be
344 written to in any cgroup, as it can only move threads inside the same
345 threaded domain, its operations are confined inside each threaded
348 The threaded domain cgroup serves as the resource domain for the whole
349 subtree, and, while the threads can be scattered across the subtree,
350 all the processes are considered to be in the threaded domain cgroup.
351 "cgroup.procs" in a threaded domain cgroup contains the PIDs of all
352 processes in the subtree and is not readable in the subtree proper.
353 However, "cgroup.procs" can be written to from anywhere in the subtree
354 to migrate all threads of the matching process to the cgroup.
356 Only threaded controllers can be enabled in a threaded subtree. When
357 a threaded controller is enabled inside a threaded subtree, it only
358 accounts for and controls resource consumptions associated with the
359 threads in the cgroup and its descendants. All consumptions which
360 aren't tied to a specific thread belong to the threaded domain cgroup.
362 Because a threaded subtree is exempt from no internal process
363 constraint, a threaded controller must be able to handle competition
364 between threads in a non-leaf cgroup and its child cgroups. Each
365 threaded controller defines how such competitions are handled.
368 [Un]populated Notification
369 --------------------------
371 Each non-root cgroup has a "cgroup.events" file which contains
372 "populated" field indicating whether the cgroup's sub-hierarchy has
373 live processes in it. Its value is 0 if there is no live process in
374 the cgroup and its descendants; otherwise, 1. poll and [id]notify
375 events are triggered when the value changes. This can be used, for
376 example, to start a clean-up operation after all processes of a given
377 sub-hierarchy have exited. The populated state updates and
378 notifications are recursive. Consider the following sub-hierarchy
379 where the numbers in the parentheses represent the numbers of processes
385 A, B and C's "populated" fields would be 1 while D's 0. After the one
386 process in C exits, B and C's "populated" fields would flip to "0" and
387 file modified events will be generated on the "cgroup.events" files of
391 Controlling Controllers
392 -----------------------
394 Enabling and Disabling
395 ~~~~~~~~~~~~~~~~~~~~~~
397 Each cgroup has a "cgroup.controllers" file which lists all
398 controllers available for the cgroup to enable::
400 # cat cgroup.controllers
403 No controller is enabled by default. Controllers can be enabled and
404 disabled by writing to the "cgroup.subtree_control" file::
406 # echo "+cpu +memory -io" > cgroup.subtree_control
408 Only controllers which are listed in "cgroup.controllers" can be
409 enabled. When multiple operations are specified as above, either they
410 all succeed or fail. If multiple operations on the same controller
411 are specified, the last one is effective.
413 Enabling a controller in a cgroup indicates that the distribution of
414 the target resource across its immediate children will be controlled.
415 Consider the following sub-hierarchy. The enabled controllers are
416 listed in parentheses::
418 A(cpu,memory) - B(memory) - C()
421 As A has "cpu" and "memory" enabled, A will control the distribution
422 of CPU cycles and memory to its children, in this case, B. As B has
423 "memory" enabled but not "CPU", C and D will compete freely on CPU
424 cycles but their division of memory available to B will be controlled.
426 As a controller regulates the distribution of the target resource to
427 the cgroup's children, enabling it creates the controller's interface
428 files in the child cgroups. In the above example, enabling "cpu" on B
429 would create the "cpu." prefixed controller interface files in C and
430 D. Likewise, disabling "memory" from B would remove the "memory."
431 prefixed controller interface files from C and D. This means that the
432 controller interface files - anything which doesn't start with
433 "cgroup." are owned by the parent rather than the cgroup itself.
439 Resources are distributed top-down and a cgroup can further distribute
440 a resource only if the resource has been distributed to it from the
441 parent. This means that all non-root "cgroup.subtree_control" files
442 can only contain controllers which are enabled in the parent's
443 "cgroup.subtree_control" file. A controller can be enabled only if
444 the parent has the controller enabled and a controller can't be
445 disabled if one or more children have it enabled.
448 No Internal Process Constraint
449 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
451 Non-root cgroups can distribute domain resources to their children
452 only when they don't have any processes of their own. In other words,
453 only domain cgroups which don't contain any processes can have domain
454 controllers enabled in their "cgroup.subtree_control" files.
456 This guarantees that, when a domain controller is looking at the part
457 of the hierarchy which has it enabled, processes are always only on
458 the leaves. This rules out situations where child cgroups compete
459 against internal processes of the parent.
461 The root cgroup is exempt from this restriction. Root contains
462 processes and anonymous resource consumption which can't be associated
463 with any other cgroups and requires special treatment from most
464 controllers. How resource consumption in the root cgroup is governed
465 is up to each controller (for more information on this topic please
466 refer to the Non-normative information section in the Controllers
469 Note that the restriction doesn't get in the way if there is no
470 enabled controller in the cgroup's "cgroup.subtree_control". This is
471 important as otherwise it wouldn't be possible to create children of a
472 populated cgroup. To control resource distribution of a cgroup, the
473 cgroup must create children and transfer all its processes to the
474 children before enabling controllers in its "cgroup.subtree_control"
484 A cgroup can be delegated in two ways. First, to a less privileged
485 user by granting write access of the directory and its "cgroup.procs",
486 "cgroup.threads" and "cgroup.subtree_control" files to the user.
487 Second, if the "nsdelegate" mount option is set, automatically to a
488 cgroup namespace on namespace creation.
490 Because the resource control interface files in a given directory
491 control the distribution of the parent's resources, the delegatee
492 shouldn't be allowed to write to them. For the first method, this is
493 achieved by not granting access to these files. For the second, the
494 kernel rejects writes to all files other than "cgroup.procs" and
495 "cgroup.subtree_control" on a namespace root from inside the
498 The end results are equivalent for both delegation types. Once
499 delegated, the user can build sub-hierarchy under the directory,
500 organize processes inside it as it sees fit and further distribute the
501 resources it received from the parent. The limits and other settings
502 of all resource controllers are hierarchical and regardless of what
503 happens in the delegated sub-hierarchy, nothing can escape the
504 resource restrictions imposed by the parent.
506 Currently, cgroup doesn't impose any restrictions on the number of
507 cgroups in or nesting depth of a delegated sub-hierarchy; however,
508 this may be limited explicitly in the future.
511 Delegation Containment
512 ~~~~~~~~~~~~~~~~~~~~~~
514 A delegated sub-hierarchy is contained in the sense that processes
515 can't be moved into or out of the sub-hierarchy by the delegatee.
517 For delegations to a less privileged user, this is achieved by
518 requiring the following conditions for a process with a non-root euid
519 to migrate a target process into a cgroup by writing its PID to the
522 - The writer must have write access to the "cgroup.procs" file.
524 - The writer must have write access to the "cgroup.procs" file of the
525 common ancestor of the source and destination cgroups.
527 The above two constraints ensure that while a delegatee may migrate
528 processes around freely in the delegated sub-hierarchy it can't pull
529 in from or push out to outside the sub-hierarchy.
531 For an example, let's assume cgroups C0 and C1 have been delegated to
532 user U0 who created C00, C01 under C0 and C10 under C1 as follows and
533 all processes under C0 and C1 belong to U0::
535 ~~~~~~~~~~~~~ - C0 - C00
538 ~~~~~~~~~~~~~ - C1 - C10
540 Let's also say U0 wants to write the PID of a process which is
541 currently in C10 into "C00/cgroup.procs". U0 has write access to the
542 file; however, the common ancestor of the source cgroup C10 and the
543 destination cgroup C00 is above the points of delegation and U0 would
544 not have write access to its "cgroup.procs" files and thus the write
545 will be denied with -EACCES.
547 For delegations to namespaces, containment is achieved by requiring
548 that both the source and destination cgroups are reachable from the
549 namespace of the process which is attempting the migration. If either
550 is not reachable, the migration is rejected with -ENOENT.
556 Organize Once and Control
557 ~~~~~~~~~~~~~~~~~~~~~~~~~
559 Migrating a process across cgroups is a relatively expensive operation
560 and stateful resources such as memory are not moved together with the
561 process. This is an explicit design decision as there often exist
562 inherent trade-offs between migration and various hot paths in terms
563 of synchronization cost.
565 As such, migrating processes across cgroups frequently as a means to
566 apply different resource restrictions is discouraged. A workload
567 should be assigned to a cgroup according to the system's logical and
568 resource structure once on start-up. Dynamic adjustments to resource
569 distribution can be made by changing controller configuration through
573 Avoid Name Collisions
574 ~~~~~~~~~~~~~~~~~~~~~
576 Interface files for a cgroup and its children cgroups occupy the same
577 directory and it is possible to create children cgroups which collide
578 with interface files.
580 All cgroup core interface files are prefixed with "cgroup." and each
581 controller's interface files are prefixed with the controller name and
582 a dot. A controller's name is composed of lower case alphabets and
583 '_'s but never begins with an '_' so it can be used as the prefix
584 character for collision avoidance. Also, interface file names won't
585 start or end with terms which are often used in categorizing workloads
586 such as job, service, slice, unit or workload.
588 cgroup doesn't do anything to prevent name collisions and it's the
589 user's responsibility to avoid them.
592 Resource Distribution Models
593 ============================
595 cgroup controllers implement several resource distribution schemes
596 depending on the resource type and expected use cases. This section
597 describes major schemes in use along with their expected behaviors.
603 A parent's resource is distributed by adding up the weights of all
604 active children and giving each the fraction matching the ratio of its
605 weight against the sum. As only children which can make use of the
606 resource at the moment participate in the distribution, this is
607 work-conserving. Due to the dynamic nature, this model is usually
608 used for stateless resources.
610 All weights are in the range [1, 10000] with the default at 100. This
611 allows symmetric multiplicative biases in both directions at fine
612 enough granularity while staying in the intuitive range.
614 As long as the weight is in range, all configuration combinations are
615 valid and there is no reason to reject configuration changes or
618 "cpu.weight" proportionally distributes CPU cycles to active children
619 and is an example of this type.
622 .. _cgroupv2-limits-distributor:
627 A child can only consume up to the configured amount of the resource.
628 Limits can be over-committed - the sum of the limits of children can
629 exceed the amount of resource available to the parent.
631 Limits are in the range [0, max] and defaults to "max", which is noop.
633 As limits can be over-committed, all configuration combinations are
634 valid and there is no reason to reject configuration changes or
637 "io.max" limits the maximum BPS and/or IOPS that a cgroup can consume
638 on an IO device and is an example of this type.
640 .. _cgroupv2-protections-distributor:
645 A cgroup is protected up to the configured amount of the resource
646 as long as the usages of all its ancestors are under their
647 protected levels. Protections can be hard guarantees or best effort
648 soft boundaries. Protections can also be over-committed in which case
649 only up to the amount available to the parent is protected among
652 Protections are in the range [0, max] and defaults to 0, which is
655 As protections can be over-committed, all configuration combinations
656 are valid and there is no reason to reject configuration changes or
659 "memory.low" implements best-effort memory protection and is an
660 example of this type.
666 A cgroup is exclusively allocated a certain amount of a finite
667 resource. Allocations can't be over-committed - the sum of the
668 allocations of children can not exceed the amount of resource
669 available to the parent.
671 Allocations are in the range [0, max] and defaults to 0, which is no
674 As allocations can't be over-committed, some configuration
675 combinations are invalid and should be rejected. Also, if the
676 resource is mandatory for execution of processes, process migrations
679 "cpu.rt.max" hard-allocates realtime slices and is an example of this
689 All interface files should be in one of the following formats whenever
692 New-line separated values
693 (when only one value can be written at once)
699 Space separated values
700 (when read-only or multiple values can be written at once)
712 KEY0 SUB_KEY0=VAL00 SUB_KEY1=VAL01...
713 KEY1 SUB_KEY0=VAL10 SUB_KEY1=VAL11...
716 For a writable file, the format for writing should generally match
717 reading; however, controllers may allow omitting later fields or
718 implement restricted shortcuts for most common use cases.
720 For both flat and nested keyed files, only the values for a single key
721 can be written at a time. For nested keyed files, the sub key pairs
722 may be specified in any order and not all pairs have to be specified.
728 - Settings for a single feature should be contained in a single file.
730 - The root cgroup should be exempt from resource control and thus
731 shouldn't have resource control interface files.
733 - The default time unit is microseconds. If a different unit is ever
734 used, an explicit unit suffix must be present.
736 - A parts-per quantity should use a percentage decimal with at least
737 two digit fractional part - e.g. 13.40.
739 - If a controller implements weight based resource distribution, its
740 interface file should be named "weight" and have the range [1,
741 10000] with 100 as the default. The values are chosen to allow
742 enough and symmetric bias in both directions while keeping it
743 intuitive (the default is 100%).
745 - If a controller implements an absolute resource guarantee and/or
746 limit, the interface files should be named "min" and "max"
747 respectively. If a controller implements best effort resource
748 guarantee and/or limit, the interface files should be named "low"
749 and "high" respectively.
751 In the above four control files, the special token "max" should be
752 used to represent upward infinity for both reading and writing.
754 - If a setting has a configurable default value and keyed specific
755 overrides, the default entry should be keyed with "default" and
756 appear as the first entry in the file.
758 The default value can be updated by writing either "default $VAL" or
761 When writing to update a specific override, "default" can be used as
762 the value to indicate removal of the override. Override entries
763 with "default" as the value must not appear when read.
765 For example, a setting which is keyed by major:minor device numbers
766 with integer values may look like the following::
768 # cat cgroup-example-interface-file
772 The default value can be updated by::
774 # echo 125 > cgroup-example-interface-file
778 # echo "default 125" > cgroup-example-interface-file
780 An override can be set by::
782 # echo "8:16 170" > cgroup-example-interface-file
786 # echo "8:0 default" > cgroup-example-interface-file
787 # cat cgroup-example-interface-file
791 - For events which are not very high frequency, an interface file
792 "events" should be created which lists event key value pairs.
793 Whenever a notifiable event happens, file modified event should be
794 generated on the file.
800 All cgroup core files are prefixed with "cgroup."
803 A read-write single value file which exists on non-root
806 When read, it indicates the current type of the cgroup, which
807 can be one of the following values.
809 - "domain" : A normal valid domain cgroup.
811 - "domain threaded" : A threaded domain cgroup which is
812 serving as the root of a threaded subtree.
814 - "domain invalid" : A cgroup which is in an invalid state.
815 It can't be populated or have controllers enabled. It may
816 be allowed to become a threaded cgroup.
818 - "threaded" : A threaded cgroup which is a member of a
821 A cgroup can be turned into a threaded cgroup by writing
822 "threaded" to this file.
825 A read-write new-line separated values file which exists on
828 When read, it lists the PIDs of all processes which belong to
829 the cgroup one-per-line. The PIDs are not ordered and the
830 same PID may show up more than once if the process got moved
831 to another cgroup and then back or the PID got recycled while
834 A PID can be written to migrate the process associated with
835 the PID to the cgroup. The writer should match all of the
836 following conditions.
838 - It must have write access to the "cgroup.procs" file.
840 - It must have write access to the "cgroup.procs" file of the
841 common ancestor of the source and destination cgroups.
843 When delegating a sub-hierarchy, write access to this file
844 should be granted along with the containing directory.
846 In a threaded cgroup, reading this file fails with EOPNOTSUPP
847 as all the processes belong to the thread root. Writing is
848 supported and moves every thread of the process to the cgroup.
851 A read-write new-line separated values file which exists on
854 When read, it lists the TIDs of all threads which belong to
855 the cgroup one-per-line. The TIDs are not ordered and the
856 same TID may show up more than once if the thread got moved to
857 another cgroup and then back or the TID got recycled while
860 A TID can be written to migrate the thread associated with the
861 TID to the cgroup. The writer should match all of the
862 following conditions.
864 - It must have write access to the "cgroup.threads" file.
866 - The cgroup that the thread is currently in must be in the
867 same resource domain as the destination cgroup.
869 - It must have write access to the "cgroup.procs" file of the
870 common ancestor of the source and destination cgroups.
872 When delegating a sub-hierarchy, write access to this file
873 should be granted along with the containing directory.
876 A read-only space separated values file which exists on all
879 It shows space separated list of all controllers available to
880 the cgroup. The controllers are not ordered.
882 cgroup.subtree_control
883 A read-write space separated values file which exists on all
884 cgroups. Starts out empty.
886 When read, it shows space separated list of the controllers
887 which are enabled to control resource distribution from the
888 cgroup to its children.
890 Space separated list of controllers prefixed with '+' or '-'
891 can be written to enable or disable controllers. A controller
892 name prefixed with '+' enables the controller and '-'
893 disables. If a controller appears more than once on the list,
894 the last one is effective. When multiple enable and disable
895 operations are specified, either all succeed or all fail.
898 A read-only flat-keyed file which exists on non-root cgroups.
899 The following entries are defined. Unless specified
900 otherwise, a value change in this file generates a file
904 1 if the cgroup or its descendants contains any live
905 processes; otherwise, 0.
907 1 if the cgroup is frozen; otherwise, 0.
909 cgroup.max.descendants
910 A read-write single value files. The default is "max".
912 Maximum allowed number of descent cgroups.
913 If the actual number of descendants is equal or larger,
914 an attempt to create a new cgroup in the hierarchy will fail.
917 A read-write single value files. The default is "max".
919 Maximum allowed descent depth below the current cgroup.
920 If the actual descent depth is equal or larger,
921 an attempt to create a new child cgroup will fail.
924 A read-only flat-keyed file with the following entries:
927 Total number of visible descendant cgroups.
930 Total number of dying descendant cgroups. A cgroup becomes
931 dying after being deleted by a user. The cgroup will remain
932 in dying state for some time undefined time (which can depend
933 on system load) before being completely destroyed.
935 A process can't enter a dying cgroup under any circumstances,
936 a dying cgroup can't revive.
938 A dying cgroup can consume system resources not exceeding
939 limits, which were active at the moment of cgroup deletion.
942 A read-write single value file which exists on non-root cgroups.
943 Allowed values are "0" and "1". The default is "0".
945 Writing "1" to the file causes freezing of the cgroup and all
946 descendant cgroups. This means that all belonging processes will
947 be stopped and will not run until the cgroup will be explicitly
948 unfrozen. Freezing of the cgroup may take some time; when this action
949 is completed, the "frozen" value in the cgroup.events control file
950 will be updated to "1" and the corresponding notification will be
953 A cgroup can be frozen either by its own settings, or by settings
954 of any ancestor cgroups. If any of ancestor cgroups is frozen, the
955 cgroup will remain frozen.
957 Processes in the frozen cgroup can be killed by a fatal signal.
958 They also can enter and leave a frozen cgroup: either by an explicit
959 move by a user, or if freezing of the cgroup races with fork().
960 If a process is moved to a frozen cgroup, it stops. If a process is
961 moved out of a frozen cgroup, it becomes running.
963 Frozen status of a cgroup doesn't affect any cgroup tree operations:
964 it's possible to delete a frozen (and empty) cgroup, as well as
965 create new sub-cgroups.
968 A write-only single value file which exists in non-root cgroups.
969 The only allowed value is "1".
971 Writing "1" to the file causes the cgroup and all descendant cgroups to
972 be killed. This means that all processes located in the affected cgroup
973 tree will be killed via SIGKILL.
975 Killing a cgroup tree will deal with concurrent forks appropriately and
976 is protected against migrations.
978 In a threaded cgroup, writing this file fails with EOPNOTSUPP as
979 killing cgroups is a process directed operation, i.e. it affects
980 the whole thread-group.
983 A read-write single value file that allowed values are "0" and "1".
986 Writing "0" to the file will disable the cgroup PSI accounting.
987 Writing "1" to the file will re-enable the cgroup PSI accounting.
989 This control attribute is not hierarchical, so disable or enable PSI
990 accounting in a cgroup does not affect PSI accounting in descendants
991 and doesn't need pass enablement via ancestors from root.
993 The reason this control attribute exists is that PSI accounts stalls for
994 each cgroup separately and aggregates it at each level of the hierarchy.
995 This may cause non-negligible overhead for some workloads when under
996 deep level of the hierarchy, in which case this control attribute can
997 be used to disable PSI accounting in the non-leaf cgroups.
1000 A read-write nested-keyed file.
1002 Shows pressure stall information for IRQ/SOFTIRQ. See
1003 :ref:`Documentation/accounting/psi.rst <psi>` for details.
1013 The "cpu" controllers regulates distribution of CPU cycles. This
1014 controller implements weight and absolute bandwidth limit models for
1015 normal scheduling policy and absolute bandwidth allocation model for
1016 realtime scheduling policy.
1018 In all the above models, cycles distribution is defined only on a temporal
1019 base and it does not account for the frequency at which tasks are executed.
1020 The (optional) utilization clamping support allows to hint the schedutil
1021 cpufreq governor about the minimum desired frequency which should always be
1022 provided by a CPU, as well as the maximum desired frequency, which should not
1023 be exceeded by a CPU.
1025 WARNING: cgroup2 doesn't yet support control of realtime processes and
1026 the cpu controller can only be enabled when all RT processes are in
1027 the root cgroup. Be aware that system management software may already
1028 have placed RT processes into nonroot cgroups during the system boot
1029 process, and these processes may need to be moved to the root cgroup
1030 before the cpu controller can be enabled.
1036 All time durations are in microseconds.
1039 A read-only flat-keyed file.
1040 This file exists whether the controller is enabled or not.
1042 It always reports the following three stats:
1048 and the following three when the controller is enabled:
1057 A read-write single value file which exists on non-root
1058 cgroups. The default is "100".
1060 The weight in the range [1, 10000].
1063 A read-write single value file which exists on non-root
1064 cgroups. The default is "0".
1066 The nice value is in the range [-20, 19].
1068 This interface file is an alternative interface for
1069 "cpu.weight" and allows reading and setting weight using the
1070 same values used by nice(2). Because the range is smaller and
1071 granularity is coarser for the nice values, the read value is
1072 the closest approximation of the current weight.
1075 A read-write two value file which exists on non-root cgroups.
1076 The default is "max 100000".
1078 The maximum bandwidth limit. It's in the following format::
1082 which indicates that the group may consume up to $MAX in each
1083 $PERIOD duration. "max" for $MAX indicates no limit. If only
1084 one number is written, $MAX is updated.
1087 A read-write single value file which exists on non-root
1088 cgroups. The default is "0".
1090 The burst in the range [0, $MAX].
1093 A read-write nested-keyed file.
1095 Shows pressure stall information for CPU. See
1096 :ref:`Documentation/accounting/psi.rst <psi>` for details.
1099 A read-write single value file which exists on non-root cgroups.
1100 The default is "0", i.e. no utilization boosting.
1102 The requested minimum utilization (protection) as a percentage
1103 rational number, e.g. 12.34 for 12.34%.
1105 This interface allows reading and setting minimum utilization clamp
1106 values similar to the sched_setattr(2). This minimum utilization
1107 value is used to clamp the task specific minimum utilization clamp.
1109 The requested minimum utilization (protection) is always capped by
1110 the current value for the maximum utilization (limit), i.e.
1114 A read-write single value file which exists on non-root cgroups.
1115 The default is "max". i.e. no utilization capping
1117 The requested maximum utilization (limit) as a percentage rational
1118 number, e.g. 98.76 for 98.76%.
1120 This interface allows reading and setting maximum utilization clamp
1121 values similar to the sched_setattr(2). This maximum utilization
1122 value is used to clamp the task specific maximum utilization clamp.
1129 The "memory" controller regulates distribution of memory. Memory is
1130 stateful and implements both limit and protection models. Due to the
1131 intertwining between memory usage and reclaim pressure and the
1132 stateful nature of memory, the distribution model is relatively
1135 While not completely water-tight, all major memory usages by a given
1136 cgroup are tracked so that the total memory consumption can be
1137 accounted and controlled to a reasonable extent. Currently, the
1138 following types of memory usages are tracked.
1140 - Userland memory - page cache and anonymous memory.
1142 - Kernel data structures such as dentries and inodes.
1144 - TCP socket buffers.
1146 The above list may expand in the future for better coverage.
1149 Memory Interface Files
1150 ~~~~~~~~~~~~~~~~~~~~~~
1152 All memory amounts are in bytes. If a value which is not aligned to
1153 PAGE_SIZE is written, the value may be rounded up to the closest
1154 PAGE_SIZE multiple when read back.
1157 A read-only single value file which exists on non-root
1160 The total amount of memory currently being used by the cgroup
1161 and its descendants.
1164 A read-write single value file which exists on non-root
1165 cgroups. The default is "0".
1167 Hard memory protection. If the memory usage of a cgroup
1168 is within its effective min boundary, the cgroup's memory
1169 won't be reclaimed under any conditions. If there is no
1170 unprotected reclaimable memory available, OOM killer
1171 is invoked. Above the effective min boundary (or
1172 effective low boundary if it is higher), pages are reclaimed
1173 proportionally to the overage, reducing reclaim pressure for
1176 Effective min boundary is limited by memory.min values of
1177 all ancestor cgroups. If there is memory.min overcommitment
1178 (child cgroup or cgroups are requiring more protected memory
1179 than parent will allow), then each child cgroup will get
1180 the part of parent's protection proportional to its
1181 actual memory usage below memory.min.
1183 Putting more memory than generally available under this
1184 protection is discouraged and may lead to constant OOMs.
1186 If a memory cgroup is not populated with processes,
1187 its memory.min is ignored.
1190 A read-write single value file which exists on non-root
1191 cgroups. The default is "0".
1193 Best-effort memory protection. If the memory usage of a
1194 cgroup is within its effective low boundary, the cgroup's
1195 memory won't be reclaimed unless there is no reclaimable
1196 memory available in unprotected cgroups.
1197 Above the effective low boundary (or
1198 effective min boundary if it is higher), pages are reclaimed
1199 proportionally to the overage, reducing reclaim pressure for
1202 Effective low boundary is limited by memory.low values of
1203 all ancestor cgroups. If there is memory.low overcommitment
1204 (child cgroup or cgroups are requiring more protected memory
1205 than parent will allow), then each child cgroup will get
1206 the part of parent's protection proportional to its
1207 actual memory usage below memory.low.
1209 Putting more memory than generally available under this
1210 protection is discouraged.
1213 A read-write single value file which exists on non-root
1214 cgroups. The default is "max".
1216 Memory usage throttle limit. This is the main mechanism to
1217 control memory usage of a cgroup. If a cgroup's usage goes
1218 over the high boundary, the processes of the cgroup are
1219 throttled and put under heavy reclaim pressure.
1221 Going over the high limit never invokes the OOM killer and
1222 under extreme conditions the limit may be breached.
1225 A read-write single value file which exists on non-root
1226 cgroups. The default is "max".
1228 Memory usage hard limit. This is the final protection
1229 mechanism. If a cgroup's memory usage reaches this limit and
1230 can't be reduced, the OOM killer is invoked in the cgroup.
1231 Under certain circumstances, the usage may go over the limit
1234 In default configuration regular 0-order allocations always
1235 succeed unless OOM killer chooses current task as a victim.
1237 Some kinds of allocations don't invoke the OOM killer.
1238 Caller could retry them differently, return into userspace
1239 as -ENOMEM or silently ignore in cases like disk readahead.
1241 This is the ultimate protection mechanism. As long as the
1242 high limit is used and monitored properly, this limit's
1243 utility is limited to providing the final safety net.
1246 A write-only nested-keyed file which exists for all cgroups.
1248 This is a simple interface to trigger memory reclaim in the
1251 This file accepts a single key, the number of bytes to reclaim.
1252 No nested keys are currently supported.
1256 echo "1G" > memory.reclaim
1258 The interface can be later extended with nested keys to
1259 configure the reclaim behavior. For example, specify the
1260 type of memory to reclaim from (anon, file, ..).
1262 Please note that the kernel can over or under reclaim from
1263 the target cgroup. If less bytes are reclaimed than the
1264 specified amount, -EAGAIN is returned.
1266 Please note that the proactive reclaim (triggered by this
1267 interface) is not meant to indicate memory pressure on the
1268 memory cgroup. Therefore socket memory balancing triggered by
1269 the memory reclaim normally is not exercised in this case.
1270 This means that the networking layer will not adapt based on
1271 reclaim induced by memory.reclaim.
1274 A read-only single value file which exists on non-root
1277 The max memory usage recorded for the cgroup and its
1278 descendants since the creation of the cgroup.
1281 A read-write single value file which exists on non-root
1282 cgroups. The default value is "0".
1284 Determines whether the cgroup should be treated as
1285 an indivisible workload by the OOM killer. If set,
1286 all tasks belonging to the cgroup or to its descendants
1287 (if the memory cgroup is not a leaf cgroup) are killed
1288 together or not at all. This can be used to avoid
1289 partial kills to guarantee workload integrity.
1291 Tasks with the OOM protection (oom_score_adj set to -1000)
1292 are treated as an exception and are never killed.
1294 If the OOM killer is invoked in a cgroup, it's not going
1295 to kill any tasks outside of this cgroup, regardless
1296 memory.oom.group values of ancestor cgroups.
1299 A read-only flat-keyed file which exists on non-root cgroups.
1300 The following entries are defined. Unless specified
1301 otherwise, a value change in this file generates a file
1304 Note that all fields in this file are hierarchical and the
1305 file modified event can be generated due to an event down the
1306 hierarchy. For the local events at the cgroup level see
1307 memory.events.local.
1310 The number of times the cgroup is reclaimed due to
1311 high memory pressure even though its usage is under
1312 the low boundary. This usually indicates that the low
1313 boundary is over-committed.
1316 The number of times processes of the cgroup are
1317 throttled and routed to perform direct memory reclaim
1318 because the high memory boundary was exceeded. For a
1319 cgroup whose memory usage is capped by the high limit
1320 rather than global memory pressure, this event's
1321 occurrences are expected.
1324 The number of times the cgroup's memory usage was
1325 about to go over the max boundary. If direct reclaim
1326 fails to bring it down, the cgroup goes to OOM state.
1329 The number of time the cgroup's memory usage was
1330 reached the limit and allocation was about to fail.
1332 This event is not raised if the OOM killer is not
1333 considered as an option, e.g. for failed high-order
1334 allocations or if caller asked to not retry attempts.
1337 The number of processes belonging to this cgroup
1338 killed by any kind of OOM killer.
1341 The number of times a group OOM has occurred.
1344 Similar to memory.events but the fields in the file are local
1345 to the cgroup i.e. not hierarchical. The file modified event
1346 generated on this file reflects only the local events.
1349 A read-only flat-keyed file which exists on non-root cgroups.
1351 This breaks down the cgroup's memory footprint into different
1352 types of memory, type-specific details, and other information
1353 on the state and past events of the memory management system.
1355 All memory amounts are in bytes.
1357 The entries are ordered to be human readable, and new entries
1358 can show up in the middle. Don't rely on items remaining in a
1359 fixed position; use the keys to look up specific values!
1361 If the entry has no per-node counter (or not show in the
1362 memory.numa_stat). We use 'npn' (non-per-node) as the tag
1363 to indicate that it will not show in the memory.numa_stat.
1366 Amount of memory used in anonymous mappings such as
1367 brk(), sbrk(), and mmap(MAP_ANONYMOUS)
1370 Amount of memory used to cache filesystem data,
1371 including tmpfs and shared memory.
1374 Amount of total kernel memory, including
1375 (kernel_stack, pagetables, percpu, vmalloc, slab) in
1376 addition to other kernel memory use cases.
1379 Amount of memory allocated to kernel stacks.
1382 Amount of memory allocated for page tables.
1385 Amount of memory allocated for secondary page tables,
1386 this currently includes KVM mmu allocations on x86
1390 Amount of memory used for storing per-cpu kernel
1394 Amount of memory used in network transmission buffers
1397 Amount of memory used for vmap backed memory.
1400 Amount of cached filesystem data that is swap-backed,
1401 such as tmpfs, shm segments, shared anonymous mmap()s
1404 Amount of memory consumed by the zswap compression backend.
1407 Amount of application memory swapped out to zswap.
1410 Amount of cached filesystem data mapped with mmap()
1413 Amount of cached filesystem data that was modified but
1414 not yet written back to disk
1417 Amount of cached filesystem data that was modified and
1418 is currently being written back to disk
1421 Amount of swap cached in memory. The swapcache is accounted
1422 against both memory and swap usage.
1425 Amount of memory used in anonymous mappings backed by
1426 transparent hugepages
1429 Amount of cached filesystem data backed by transparent
1433 Amount of shm, tmpfs, shared anonymous mmap()s backed by
1434 transparent hugepages
1436 inactive_anon, active_anon, inactive_file, active_file, unevictable
1437 Amount of memory, swap-backed and filesystem-backed,
1438 on the internal memory management lists used by the
1439 page reclaim algorithm.
1441 As these represent internal list state (eg. shmem pages are on anon
1442 memory management lists), inactive_foo + active_foo may not be equal to
1443 the value for the foo counter, since the foo counter is type-based, not
1447 Part of "slab" that might be reclaimed, such as
1448 dentries and inodes.
1451 Part of "slab" that cannot be reclaimed on memory
1455 Amount of memory used for storing in-kernel data
1458 workingset_refault_anon
1459 Number of refaults of previously evicted anonymous pages.
1461 workingset_refault_file
1462 Number of refaults of previously evicted file pages.
1464 workingset_activate_anon
1465 Number of refaulted anonymous pages that were immediately
1468 workingset_activate_file
1469 Number of refaulted file pages that were immediately activated.
1471 workingset_restore_anon
1472 Number of restored anonymous pages which have been detected as
1473 an active workingset before they got reclaimed.
1475 workingset_restore_file
1476 Number of restored file pages which have been detected as an
1477 active workingset before they got reclaimed.
1479 workingset_nodereclaim
1480 Number of times a shadow node has been reclaimed
1483 Amount of scanned pages (in an inactive LRU list)
1486 Amount of reclaimed pages
1489 Amount of scanned pages by kswapd (in an inactive LRU list)
1492 Amount of scanned pages directly (in an inactive LRU list)
1494 pgscan_khugepaged (npn)
1495 Amount of scanned pages by khugepaged (in an inactive LRU list)
1497 pgsteal_kswapd (npn)
1498 Amount of reclaimed pages by kswapd
1500 pgsteal_direct (npn)
1501 Amount of reclaimed pages directly
1503 pgsteal_khugepaged (npn)
1504 Amount of reclaimed pages by khugepaged
1507 Total number of page faults incurred
1510 Number of major page faults incurred
1513 Amount of scanned pages (in an active LRU list)
1516 Amount of pages moved to the active LRU list
1519 Amount of pages moved to the inactive LRU list
1522 Amount of pages postponed to be freed under memory pressure
1525 Amount of reclaimed lazyfree pages
1527 thp_fault_alloc (npn)
1528 Number of transparent hugepages which were allocated to satisfy
1529 a page fault. This counter is not present when CONFIG_TRANSPARENT_HUGEPAGE
1532 thp_collapse_alloc (npn)
1533 Number of transparent hugepages which were allocated to allow
1534 collapsing an existing range of pages. This counter is not
1535 present when CONFIG_TRANSPARENT_HUGEPAGE is not set.
1538 A read-only nested-keyed file which exists on non-root cgroups.
1540 This breaks down the cgroup's memory footprint into different
1541 types of memory, type-specific details, and other information
1542 per node on the state of the memory management system.
1544 This is useful for providing visibility into the NUMA locality
1545 information within an memcg since the pages are allowed to be
1546 allocated from any physical node. One of the use case is evaluating
1547 application performance by combining this information with the
1548 application's CPU allocation.
1550 All memory amounts are in bytes.
1552 The output format of memory.numa_stat is::
1554 type N0=<bytes in node 0> N1=<bytes in node 1> ...
1556 The entries are ordered to be human readable, and new entries
1557 can show up in the middle. Don't rely on items remaining in a
1558 fixed position; use the keys to look up specific values!
1560 The entries can refer to the memory.stat.
1563 A read-only single value file which exists on non-root
1566 The total amount of swap currently being used by the cgroup
1567 and its descendants.
1570 A read-write single value file which exists on non-root
1571 cgroups. The default is "max".
1573 Swap usage throttle limit. If a cgroup's swap usage exceeds
1574 this limit, all its further allocations will be throttled to
1575 allow userspace to implement custom out-of-memory procedures.
1577 This limit marks a point of no return for the cgroup. It is NOT
1578 designed to manage the amount of swapping a workload does
1579 during regular operation. Compare to memory.swap.max, which
1580 prohibits swapping past a set amount, but lets the cgroup
1581 continue unimpeded as long as other memory can be reclaimed.
1583 Healthy workloads are not expected to reach this limit.
1586 A read-write single value file which exists on non-root
1587 cgroups. The default is "max".
1589 Swap usage hard limit. If a cgroup's swap usage reaches this
1590 limit, anonymous memory of the cgroup will not be swapped out.
1593 A read-only flat-keyed file which exists on non-root cgroups.
1594 The following entries are defined. Unless specified
1595 otherwise, a value change in this file generates a file
1599 The number of times the cgroup's swap usage was over
1603 The number of times the cgroup's swap usage was about
1604 to go over the max boundary and swap allocation
1608 The number of times swap allocation failed either
1609 because of running out of swap system-wide or max
1612 When reduced under the current usage, the existing swap
1613 entries are reclaimed gradually and the swap usage may stay
1614 higher than the limit for an extended period of time. This
1615 reduces the impact on the workload and memory management.
1617 memory.zswap.current
1618 A read-only single value file which exists on non-root
1621 The total amount of memory consumed by the zswap compression
1625 A read-write single value file which exists on non-root
1626 cgroups. The default is "max".
1628 Zswap usage hard limit. If a cgroup's zswap pool reaches this
1629 limit, it will refuse to take any more stores before existing
1630 entries fault back in or are written out to disk.
1633 A read-only nested-keyed file.
1635 Shows pressure stall information for memory. See
1636 :ref:`Documentation/accounting/psi.rst <psi>` for details.
1642 "memory.high" is the main mechanism to control memory usage.
1643 Over-committing on high limit (sum of high limits > available memory)
1644 and letting global memory pressure to distribute memory according to
1645 usage is a viable strategy.
1647 Because breach of the high limit doesn't trigger the OOM killer but
1648 throttles the offending cgroup, a management agent has ample
1649 opportunities to monitor and take appropriate actions such as granting
1650 more memory or terminating the workload.
1652 Determining whether a cgroup has enough memory is not trivial as
1653 memory usage doesn't indicate whether the workload can benefit from
1654 more memory. For example, a workload which writes data received from
1655 network to a file can use all available memory but can also operate as
1656 performant with a small amount of memory. A measure of memory
1657 pressure - how much the workload is being impacted due to lack of
1658 memory - is necessary to determine whether a workload needs more
1659 memory; unfortunately, memory pressure monitoring mechanism isn't
1666 A memory area is charged to the cgroup which instantiated it and stays
1667 charged to the cgroup until the area is released. Migrating a process
1668 to a different cgroup doesn't move the memory usages that it
1669 instantiated while in the previous cgroup to the new cgroup.
1671 A memory area may be used by processes belonging to different cgroups.
1672 To which cgroup the area will be charged is in-deterministic; however,
1673 over time, the memory area is likely to end up in a cgroup which has
1674 enough memory allowance to avoid high reclaim pressure.
1676 If a cgroup sweeps a considerable amount of memory which is expected
1677 to be accessed repeatedly by other cgroups, it may make sense to use
1678 POSIX_FADV_DONTNEED to relinquish the ownership of memory areas
1679 belonging to the affected files to ensure correct memory ownership.
1685 The "io" controller regulates the distribution of IO resources. This
1686 controller implements both weight based and absolute bandwidth or IOPS
1687 limit distribution; however, weight based distribution is available
1688 only if cfq-iosched is in use and neither scheme is available for
1696 A read-only nested-keyed file.
1698 Lines are keyed by $MAJ:$MIN device numbers and not ordered.
1699 The following nested keys are defined.
1701 ====== =====================
1703 wbytes Bytes written
1704 rios Number of read IOs
1705 wios Number of write IOs
1706 dbytes Bytes discarded
1707 dios Number of discard IOs
1708 ====== =====================
1710 An example read output follows::
1712 8:16 rbytes=1459200 wbytes=314773504 rios=192 wios=353 dbytes=0 dios=0
1713 8:0 rbytes=90430464 wbytes=299008000 rios=8950 wios=1252 dbytes=50331648 dios=3021
1716 A read-write nested-keyed file which exists only on the root
1719 This file configures the Quality of Service of the IO cost
1720 model based controller (CONFIG_BLK_CGROUP_IOCOST) which
1721 currently implements "io.weight" proportional control. Lines
1722 are keyed by $MAJ:$MIN device numbers and not ordered. The
1723 line for a given device is populated on the first write for
1724 the device on "io.cost.qos" or "io.cost.model". The following
1725 nested keys are defined.
1727 ====== =====================================
1728 enable Weight-based control enable
1729 ctrl "auto" or "user"
1730 rpct Read latency percentile [0, 100]
1731 rlat Read latency threshold
1732 wpct Write latency percentile [0, 100]
1733 wlat Write latency threshold
1734 min Minimum scaling percentage [1, 10000]
1735 max Maximum scaling percentage [1, 10000]
1736 ====== =====================================
1738 The controller is disabled by default and can be enabled by
1739 setting "enable" to 1. "rpct" and "wpct" parameters default
1740 to zero and the controller uses internal device saturation
1741 state to adjust the overall IO rate between "min" and "max".
1743 When a better control quality is needed, latency QoS
1744 parameters can be configured. For example::
1746 8:16 enable=1 ctrl=auto rpct=95.00 rlat=75000 wpct=95.00 wlat=150000 min=50.00 max=150.0
1748 shows that on sdb, the controller is enabled, will consider
1749 the device saturated if the 95th percentile of read completion
1750 latencies is above 75ms or write 150ms, and adjust the overall
1751 IO issue rate between 50% and 150% accordingly.
1753 The lower the saturation point, the better the latency QoS at
1754 the cost of aggregate bandwidth. The narrower the allowed
1755 adjustment range between "min" and "max", the more conformant
1756 to the cost model the IO behavior. Note that the IO issue
1757 base rate may be far off from 100% and setting "min" and "max"
1758 blindly can lead to a significant loss of device capacity or
1759 control quality. "min" and "max" are useful for regulating
1760 devices which show wide temporary behavior changes - e.g. a
1761 ssd which accepts writes at the line speed for a while and
1762 then completely stalls for multiple seconds.
1764 When "ctrl" is "auto", the parameters are controlled by the
1765 kernel and may change automatically. Setting "ctrl" to "user"
1766 or setting any of the percentile and latency parameters puts
1767 it into "user" mode and disables the automatic changes. The
1768 automatic mode can be restored by setting "ctrl" to "auto".
1771 A read-write nested-keyed file which exists only on the root
1774 This file configures the cost model of the IO cost model based
1775 controller (CONFIG_BLK_CGROUP_IOCOST) which currently
1776 implements "io.weight" proportional control. Lines are keyed
1777 by $MAJ:$MIN device numbers and not ordered. The line for a
1778 given device is populated on the first write for the device on
1779 "io.cost.qos" or "io.cost.model". The following nested keys
1782 ===== ================================
1783 ctrl "auto" or "user"
1784 model The cost model in use - "linear"
1785 ===== ================================
1787 When "ctrl" is "auto", the kernel may change all parameters
1788 dynamically. When "ctrl" is set to "user" or any other
1789 parameters are written to, "ctrl" become "user" and the
1790 automatic changes are disabled.
1792 When "model" is "linear", the following model parameters are
1795 ============= ========================================
1796 [r|w]bps The maximum sequential IO throughput
1797 [r|w]seqiops The maximum 4k sequential IOs per second
1798 [r|w]randiops The maximum 4k random IOs per second
1799 ============= ========================================
1801 From the above, the builtin linear model determines the base
1802 costs of a sequential and random IO and the cost coefficient
1803 for the IO size. While simple, this model can cover most
1804 common device classes acceptably.
1806 The IO cost model isn't expected to be accurate in absolute
1807 sense and is scaled to the device behavior dynamically.
1809 If needed, tools/cgroup/iocost_coef_gen.py can be used to
1810 generate device-specific coefficients.
1813 A read-write flat-keyed file which exists on non-root cgroups.
1814 The default is "default 100".
1816 The first line is the default weight applied to devices
1817 without specific override. The rest are overrides keyed by
1818 $MAJ:$MIN device numbers and not ordered. The weights are in
1819 the range [1, 10000] and specifies the relative amount IO time
1820 the cgroup can use in relation to its siblings.
1822 The default weight can be updated by writing either "default
1823 $WEIGHT" or simply "$WEIGHT". Overrides can be set by writing
1824 "$MAJ:$MIN $WEIGHT" and unset by writing "$MAJ:$MIN default".
1826 An example read output follows::
1833 A read-write nested-keyed file which exists on non-root
1836 BPS and IOPS based IO limit. Lines are keyed by $MAJ:$MIN
1837 device numbers and not ordered. The following nested keys are
1840 ===== ==================================
1841 rbps Max read bytes per second
1842 wbps Max write bytes per second
1843 riops Max read IO operations per second
1844 wiops Max write IO operations per second
1845 ===== ==================================
1847 When writing, any number of nested key-value pairs can be
1848 specified in any order. "max" can be specified as the value
1849 to remove a specific limit. If the same key is specified
1850 multiple times, the outcome is undefined.
1852 BPS and IOPS are measured in each IO direction and IOs are
1853 delayed if limit is reached. Temporary bursts are allowed.
1855 Setting read limit at 2M BPS and write at 120 IOPS for 8:16::
1857 echo "8:16 rbps=2097152 wiops=120" > io.max
1859 Reading returns the following::
1861 8:16 rbps=2097152 wbps=max riops=max wiops=120
1863 Write IOPS limit can be removed by writing the following::
1865 echo "8:16 wiops=max" > io.max
1867 Reading now returns the following::
1869 8:16 rbps=2097152 wbps=max riops=max wiops=max
1872 A read-only nested-keyed file.
1874 Shows pressure stall information for IO. See
1875 :ref:`Documentation/accounting/psi.rst <psi>` for details.
1881 Page cache is dirtied through buffered writes and shared mmaps and
1882 written asynchronously to the backing filesystem by the writeback
1883 mechanism. Writeback sits between the memory and IO domains and
1884 regulates the proportion of dirty memory by balancing dirtying and
1887 The io controller, in conjunction with the memory controller,
1888 implements control of page cache writeback IOs. The memory controller
1889 defines the memory domain that dirty memory ratio is calculated and
1890 maintained for and the io controller defines the io domain which
1891 writes out dirty pages for the memory domain. Both system-wide and
1892 per-cgroup dirty memory states are examined and the more restrictive
1893 of the two is enforced.
1895 cgroup writeback requires explicit support from the underlying
1896 filesystem. Currently, cgroup writeback is implemented on ext2, ext4,
1897 btrfs, f2fs, and xfs. On other filesystems, all writeback IOs are
1898 attributed to the root cgroup.
1900 There are inherent differences in memory and writeback management
1901 which affects how cgroup ownership is tracked. Memory is tracked per
1902 page while writeback per inode. For the purpose of writeback, an
1903 inode is assigned to a cgroup and all IO requests to write dirty pages
1904 from the inode are attributed to that cgroup.
1906 As cgroup ownership for memory is tracked per page, there can be pages
1907 which are associated with different cgroups than the one the inode is
1908 associated with. These are called foreign pages. The writeback
1909 constantly keeps track of foreign pages and, if a particular foreign
1910 cgroup becomes the majority over a certain period of time, switches
1911 the ownership of the inode to that cgroup.
1913 While this model is enough for most use cases where a given inode is
1914 mostly dirtied by a single cgroup even when the main writing cgroup
1915 changes over time, use cases where multiple cgroups write to a single
1916 inode simultaneously are not supported well. In such circumstances, a
1917 significant portion of IOs are likely to be attributed incorrectly.
1918 As memory controller assigns page ownership on the first use and
1919 doesn't update it until the page is released, even if writeback
1920 strictly follows page ownership, multiple cgroups dirtying overlapping
1921 areas wouldn't work as expected. It's recommended to avoid such usage
1924 The sysctl knobs which affect writeback behavior are applied to cgroup
1925 writeback as follows.
1927 vm.dirty_background_ratio, vm.dirty_ratio
1928 These ratios apply the same to cgroup writeback with the
1929 amount of available memory capped by limits imposed by the
1930 memory controller and system-wide clean memory.
1932 vm.dirty_background_bytes, vm.dirty_bytes
1933 For cgroup writeback, this is calculated into ratio against
1934 total available memory and applied the same way as
1935 vm.dirty[_background]_ratio.
1941 This is a cgroup v2 controller for IO workload protection. You provide a group
1942 with a latency target, and if the average latency exceeds that target the
1943 controller will throttle any peers that have a lower latency target than the
1946 The limits are only applied at the peer level in the hierarchy. This means that
1947 in the diagram below, only groups A, B, and C will influence each other, and
1948 groups D and F will influence each other. Group G will influence nobody::
1957 So the ideal way to configure this is to set io.latency in groups A, B, and C.
1958 Generally you do not want to set a value lower than the latency your device
1959 supports. Experiment to find the value that works best for your workload.
1960 Start at higher than the expected latency for your device and watch the
1961 avg_lat value in io.stat for your workload group to get an idea of the
1962 latency you see during normal operation. Use the avg_lat value as a basis for
1963 your real setting, setting at 10-15% higher than the value in io.stat.
1965 How IO Latency Throttling Works
1966 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1968 io.latency is work conserving; so as long as everybody is meeting their latency
1969 target the controller doesn't do anything. Once a group starts missing its
1970 target it begins throttling any peer group that has a higher target than itself.
1971 This throttling takes 2 forms:
1973 - Queue depth throttling. This is the number of outstanding IO's a group is
1974 allowed to have. We will clamp down relatively quickly, starting at no limit
1975 and going all the way down to 1 IO at a time.
1977 - Artificial delay induction. There are certain types of IO that cannot be
1978 throttled without possibly adversely affecting higher priority groups. This
1979 includes swapping and metadata IO. These types of IO are allowed to occur
1980 normally, however they are "charged" to the originating group. If the
1981 originating group is being throttled you will see the use_delay and delay
1982 fields in io.stat increase. The delay value is how many microseconds that are
1983 being added to any process that runs in this group. Because this number can
1984 grow quite large if there is a lot of swapping or metadata IO occurring we
1985 limit the individual delay events to 1 second at a time.
1987 Once the victimized group starts meeting its latency target again it will start
1988 unthrottling any peer groups that were throttled previously. If the victimized
1989 group simply stops doing IO the global counter will unthrottle appropriately.
1991 IO Latency Interface Files
1992 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1995 This takes a similar format as the other controllers.
1997 "MAJOR:MINOR target=<target time in microseconds>"
2000 If the controller is enabled you will see extra stats in io.stat in
2001 addition to the normal ones.
2004 This is the current queue depth for the group.
2007 This is an exponential moving average with a decay rate of 1/exp
2008 bound by the sampling interval. The decay rate interval can be
2009 calculated by multiplying the win value in io.stat by the
2010 corresponding number of samples based on the win value.
2013 The sampling window size in milliseconds. This is the minimum
2014 duration of time between evaluation events. Windows only elapse
2015 with IO activity. Idle periods extend the most recent window.
2020 A single attribute controls the behavior of the I/O priority cgroup policy,
2021 namely the blkio.prio.class attribute. The following values are accepted for
2025 Do not modify the I/O priority class.
2028 For requests that do not have an I/O priority class (NONE),
2029 change the I/O priority class into RT. Do not modify
2030 the I/O priority class of other requests.
2033 For requests that do not have an I/O priority class or that have I/O
2034 priority class RT, change it into BE. Do not modify the I/O priority
2035 class of requests that have priority class IDLE.
2038 Change the I/O priority class of all requests into IDLE, the lowest
2041 The following numerical values are associated with the I/O priority policies:
2053 The numerical value that corresponds to each I/O priority class is as follows:
2055 +-------------------------------+---+
2056 | IOPRIO_CLASS_NONE | 0 |
2057 +-------------------------------+---+
2058 | IOPRIO_CLASS_RT (real-time) | 1 |
2059 +-------------------------------+---+
2060 | IOPRIO_CLASS_BE (best effort) | 2 |
2061 +-------------------------------+---+
2062 | IOPRIO_CLASS_IDLE | 3 |
2063 +-------------------------------+---+
2065 The algorithm to set the I/O priority class for a request is as follows:
2067 - Translate the I/O priority class policy into a number.
2068 - Change the request I/O priority class into the maximum of the I/O priority
2069 class policy number and the numerical I/O priority class.
2074 The process number controller is used to allow a cgroup to stop any
2075 new tasks from being fork()'d or clone()'d after a specified limit is
2078 The number of tasks in a cgroup can be exhausted in ways which other
2079 controllers cannot prevent, thus warranting its own controller. For
2080 example, a fork bomb is likely to exhaust the number of tasks before
2081 hitting memory restrictions.
2083 Note that PIDs used in this controller refer to TIDs, process IDs as
2091 A read-write single value file which exists on non-root
2092 cgroups. The default is "max".
2094 Hard limit of number of processes.
2097 A read-only single value file which exists on all cgroups.
2099 The number of processes currently in the cgroup and its
2102 Organisational operations are not blocked by cgroup policies, so it is
2103 possible to have pids.current > pids.max. This can be done by either
2104 setting the limit to be smaller than pids.current, or attaching enough
2105 processes to the cgroup such that pids.current is larger than
2106 pids.max. However, it is not possible to violate a cgroup PID policy
2107 through fork() or clone(). These will return -EAGAIN if the creation
2108 of a new process would cause a cgroup policy to be violated.
2114 The "cpuset" controller provides a mechanism for constraining
2115 the CPU and memory node placement of tasks to only the resources
2116 specified in the cpuset interface files in a task's current cgroup.
2117 This is especially valuable on large NUMA systems where placing jobs
2118 on properly sized subsets of the systems with careful processor and
2119 memory placement to reduce cross-node memory access and contention
2120 can improve overall system performance.
2122 The "cpuset" controller is hierarchical. That means the controller
2123 cannot use CPUs or memory nodes not allowed in its parent.
2126 Cpuset Interface Files
2127 ~~~~~~~~~~~~~~~~~~~~~~
2130 A read-write multiple values file which exists on non-root
2131 cpuset-enabled cgroups.
2133 It lists the requested CPUs to be used by tasks within this
2134 cgroup. The actual list of CPUs to be granted, however, is
2135 subjected to constraints imposed by its parent and can differ
2136 from the requested CPUs.
2138 The CPU numbers are comma-separated numbers or ranges.
2144 An empty value indicates that the cgroup is using the same
2145 setting as the nearest cgroup ancestor with a non-empty
2146 "cpuset.cpus" or all the available CPUs if none is found.
2148 The value of "cpuset.cpus" stays constant until the next update
2149 and won't be affected by any CPU hotplug events.
2151 cpuset.cpus.effective
2152 A read-only multiple values file which exists on all
2153 cpuset-enabled cgroups.
2155 It lists the onlined CPUs that are actually granted to this
2156 cgroup by its parent. These CPUs are allowed to be used by
2157 tasks within the current cgroup.
2159 If "cpuset.cpus" is empty, the "cpuset.cpus.effective" file shows
2160 all the CPUs from the parent cgroup that can be available to
2161 be used by this cgroup. Otherwise, it should be a subset of
2162 "cpuset.cpus" unless none of the CPUs listed in "cpuset.cpus"
2163 can be granted. In this case, it will be treated just like an
2164 empty "cpuset.cpus".
2166 Its value will be affected by CPU hotplug events.
2169 A read-write multiple values file which exists on non-root
2170 cpuset-enabled cgroups.
2172 It lists the requested memory nodes to be used by tasks within
2173 this cgroup. The actual list of memory nodes granted, however,
2174 is subjected to constraints imposed by its parent and can differ
2175 from the requested memory nodes.
2177 The memory node numbers are comma-separated numbers or ranges.
2183 An empty value indicates that the cgroup is using the same
2184 setting as the nearest cgroup ancestor with a non-empty
2185 "cpuset.mems" or all the available memory nodes if none
2188 The value of "cpuset.mems" stays constant until the next update
2189 and won't be affected by any memory nodes hotplug events.
2191 Setting a non-empty value to "cpuset.mems" causes memory of
2192 tasks within the cgroup to be migrated to the designated nodes if
2193 they are currently using memory outside of the designated nodes.
2195 There is a cost for this memory migration. The migration
2196 may not be complete and some memory pages may be left behind.
2197 So it is recommended that "cpuset.mems" should be set properly
2198 before spawning new tasks into the cpuset. Even if there is
2199 a need to change "cpuset.mems" with active tasks, it shouldn't
2202 cpuset.mems.effective
2203 A read-only multiple values file which exists on all
2204 cpuset-enabled cgroups.
2206 It lists the onlined memory nodes that are actually granted to
2207 this cgroup by its parent. These memory nodes are allowed to
2208 be used by tasks within the current cgroup.
2210 If "cpuset.mems" is empty, it shows all the memory nodes from the
2211 parent cgroup that will be available to be used by this cgroup.
2212 Otherwise, it should be a subset of "cpuset.mems" unless none of
2213 the memory nodes listed in "cpuset.mems" can be granted. In this
2214 case, it will be treated just like an empty "cpuset.mems".
2216 Its value will be affected by memory nodes hotplug events.
2218 cpuset.cpus.partition
2219 A read-write single value file which exists on non-root
2220 cpuset-enabled cgroups. This flag is owned by the parent cgroup
2221 and is not delegatable.
2223 It accepts only the following input values when written to.
2225 ========== =====================================
2226 "member" Non-root member of a partition
2227 "root" Partition root
2228 "isolated" Partition root without load balancing
2229 ========== =====================================
2231 The root cgroup is always a partition root and its state
2232 cannot be changed. All other non-root cgroups start out as
2235 When set to "root", the current cgroup is the root of a new
2236 partition or scheduling domain that comprises itself and all
2237 its descendants except those that are separate partition roots
2238 themselves and their descendants.
2240 When set to "isolated", the CPUs in that partition root will
2241 be in an isolated state without any load balancing from the
2242 scheduler. Tasks placed in such a partition with multiple
2243 CPUs should be carefully distributed and bound to each of the
2244 individual CPUs for optimal performance.
2246 The value shown in "cpuset.cpus.effective" of a partition root
2247 is the CPUs that the partition root can dedicate to a potential
2248 new child partition root. The new child subtracts available
2249 CPUs from its parent "cpuset.cpus.effective".
2251 A partition root ("root" or "isolated") can be in one of the
2252 two possible states - valid or invalid. An invalid partition
2253 root is in a degraded state where some state information may
2254 be retained, but behaves more like a "member".
2256 All possible state transitions among "member", "root" and
2257 "isolated" are allowed.
2259 On read, the "cpuset.cpus.partition" file can show the following
2262 ============================= =====================================
2263 "member" Non-root member of a partition
2264 "root" Partition root
2265 "isolated" Partition root without load balancing
2266 "root invalid (<reason>)" Invalid partition root
2267 "isolated invalid (<reason>)" Invalid isolated partition root
2268 ============================= =====================================
2270 In the case of an invalid partition root, a descriptive string on
2271 why the partition is invalid is included within parentheses.
2273 For a partition root to become valid, the following conditions
2276 1) The "cpuset.cpus" is exclusive with its siblings , i.e. they
2277 are not shared by any of its siblings (exclusivity rule).
2278 2) The parent cgroup is a valid partition root.
2279 3) The "cpuset.cpus" is not empty and must contain at least
2280 one of the CPUs from parent's "cpuset.cpus", i.e. they overlap.
2281 4) The "cpuset.cpus.effective" cannot be empty unless there is
2282 no task associated with this partition.
2284 External events like hotplug or changes to "cpuset.cpus" can
2285 cause a valid partition root to become invalid and vice versa.
2286 Note that a task cannot be moved to a cgroup with empty
2287 "cpuset.cpus.effective".
2289 For a valid partition root with the sibling cpu exclusivity
2290 rule enabled, changes made to "cpuset.cpus" that violate the
2291 exclusivity rule will invalidate the partition as well as its
2292 sibling partitions with conflicting cpuset.cpus values. So
2293 care must be taking in changing "cpuset.cpus".
2295 A valid non-root parent partition may distribute out all its CPUs
2296 to its child partitions when there is no task associated with it.
2298 Care must be taken to change a valid partition root to
2299 "member" as all its child partitions, if present, will become
2300 invalid causing disruption to tasks running in those child
2301 partitions. These inactivated partitions could be recovered if
2302 their parent is switched back to a partition root with a proper
2303 set of "cpuset.cpus".
2305 Poll and inotify events are triggered whenever the state of
2306 "cpuset.cpus.partition" changes. That includes changes caused
2307 by write to "cpuset.cpus.partition", cpu hotplug or other
2308 changes that modify the validity status of the partition.
2309 This will allow user space agents to monitor unexpected changes
2310 to "cpuset.cpus.partition" without the need to do continuous
2317 Device controller manages access to device files. It includes both
2318 creation of new device files (using mknod), and access to the
2319 existing device files.
2321 Cgroup v2 device controller has no interface files and is implemented
2322 on top of cgroup BPF. To control access to device files, a user may
2323 create bpf programs of type BPF_PROG_TYPE_CGROUP_DEVICE and attach
2324 them to cgroups with BPF_CGROUP_DEVICE flag. On an attempt to access a
2325 device file, corresponding BPF programs will be executed, and depending
2326 on the return value the attempt will succeed or fail with -EPERM.
2328 A BPF_PROG_TYPE_CGROUP_DEVICE program takes a pointer to the
2329 bpf_cgroup_dev_ctx structure, which describes the device access attempt:
2330 access type (mknod/read/write) and device (type, major and minor numbers).
2331 If the program returns 0, the attempt fails with -EPERM, otherwise it
2334 An example of BPF_PROG_TYPE_CGROUP_DEVICE program may be found in
2335 tools/testing/selftests/bpf/progs/dev_cgroup.c in the kernel source tree.
2341 The "rdma" controller regulates the distribution and accounting of
2344 RDMA Interface Files
2345 ~~~~~~~~~~~~~~~~~~~~
2348 A readwrite nested-keyed file that exists for all the cgroups
2349 except root that describes current configured resource limit
2350 for a RDMA/IB device.
2352 Lines are keyed by device name and are not ordered.
2353 Each line contains space separated resource name and its configured
2354 limit that can be distributed.
2356 The following nested keys are defined.
2358 ========== =============================
2359 hca_handle Maximum number of HCA Handles
2360 hca_object Maximum number of HCA Objects
2361 ========== =============================
2363 An example for mlx4 and ocrdma device follows::
2365 mlx4_0 hca_handle=2 hca_object=2000
2366 ocrdma1 hca_handle=3 hca_object=max
2369 A read-only file that describes current resource usage.
2370 It exists for all the cgroup except root.
2372 An example for mlx4 and ocrdma device follows::
2374 mlx4_0 hca_handle=1 hca_object=20
2375 ocrdma1 hca_handle=1 hca_object=23
2380 The HugeTLB controller allows to limit the HugeTLB usage per control group and
2381 enforces the controller limit during page fault.
2383 HugeTLB Interface Files
2384 ~~~~~~~~~~~~~~~~~~~~~~~
2386 hugetlb.<hugepagesize>.current
2387 Show current usage for "hugepagesize" hugetlb. It exists for all
2388 the cgroup except root.
2390 hugetlb.<hugepagesize>.max
2391 Set/show the hard limit of "hugepagesize" hugetlb usage.
2392 The default value is "max". It exists for all the cgroup except root.
2394 hugetlb.<hugepagesize>.events
2395 A read-only flat-keyed file which exists on non-root cgroups.
2398 The number of allocation failure due to HugeTLB limit
2400 hugetlb.<hugepagesize>.events.local
2401 Similar to hugetlb.<hugepagesize>.events but the fields in the file
2402 are local to the cgroup i.e. not hierarchical. The file modified event
2403 generated on this file reflects only the local events.
2405 hugetlb.<hugepagesize>.numa_stat
2406 Similar to memory.numa_stat, it shows the numa information of the
2407 hugetlb pages of <hugepagesize> in this cgroup. Only active in
2408 use hugetlb pages are included. The per-node values are in bytes.
2413 The Miscellaneous cgroup provides the resource limiting and tracking
2414 mechanism for the scalar resources which cannot be abstracted like the other
2415 cgroup resources. Controller is enabled by the CONFIG_CGROUP_MISC config
2418 A resource can be added to the controller via enum misc_res_type{} in the
2419 include/linux/misc_cgroup.h file and the corresponding name via misc_res_name[]
2420 in the kernel/cgroup/misc.c file. Provider of the resource must set its
2421 capacity prior to using the resource by calling misc_cg_set_capacity().
2423 Once a capacity is set then the resource usage can be updated using charge and
2424 uncharge APIs. All of the APIs to interact with misc controller are in
2425 include/linux/misc_cgroup.h.
2427 Misc Interface Files
2428 ~~~~~~~~~~~~~~~~~~~~
2430 Miscellaneous controller provides 3 interface files. If two misc resources (res_a and res_b) are registered then:
2433 A read-only flat-keyed file shown only in the root cgroup. It shows
2434 miscellaneous scalar resources available on the platform along with
2442 A read-only flat-keyed file shown in the non-root cgroups. It shows
2443 the current usage of the resources in the cgroup and its children.::
2450 A read-write flat-keyed file shown in the non root cgroups. Allowed
2451 maximum usage of the resources in the cgroup and its children.::
2457 Limit can be set by::
2459 # echo res_a 1 > misc.max
2461 Limit can be set to max by::
2463 # echo res_a max > misc.max
2465 Limits can be set higher than the capacity value in the misc.capacity
2469 A read-only flat-keyed file which exists on non-root cgroups. The
2470 following entries are defined. Unless specified otherwise, a value
2471 change in this file generates a file modified event. All fields in
2472 this file are hierarchical.
2475 The number of times the cgroup's resource usage was
2476 about to go over the max boundary.
2478 Migration and Ownership
2479 ~~~~~~~~~~~~~~~~~~~~~~~
2481 A miscellaneous scalar resource is charged to the cgroup in which it is used
2482 first, and stays charged to that cgroup until that resource is freed. Migrating
2483 a process to a different cgroup does not move the charge to the destination
2484 cgroup where the process has moved.
2492 perf_event controller, if not mounted on a legacy hierarchy, is
2493 automatically enabled on the v2 hierarchy so that perf events can
2494 always be filtered by cgroup v2 path. The controller can still be
2495 moved to a legacy hierarchy after v2 hierarchy is populated.
2498 Non-normative information
2499 -------------------------
2501 This section contains information that isn't considered to be a part of
2502 the stable kernel API and so is subject to change.
2505 CPU controller root cgroup process behaviour
2506 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2508 When distributing CPU cycles in the root cgroup each thread in this
2509 cgroup is treated as if it was hosted in a separate child cgroup of the
2510 root cgroup. This child cgroup weight is dependent on its thread nice
2513 For details of this mapping see sched_prio_to_weight array in
2514 kernel/sched/core.c file (values from this array should be scaled
2515 appropriately so the neutral - nice 0 - value is 100 instead of 1024).
2518 IO controller root cgroup process behaviour
2519 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2521 Root cgroup processes are hosted in an implicit leaf child node.
2522 When distributing IO resources this implicit child node is taken into
2523 account as if it was a normal child cgroup of the root cgroup with a
2524 weight value of 200.
2533 cgroup namespace provides a mechanism to virtualize the view of the
2534 "/proc/$PID/cgroup" file and cgroup mounts. The CLONE_NEWCGROUP clone
2535 flag can be used with clone(2) and unshare(2) to create a new cgroup
2536 namespace. The process running inside the cgroup namespace will have
2537 its "/proc/$PID/cgroup" output restricted to cgroupns root. The
2538 cgroupns root is the cgroup of the process at the time of creation of
2539 the cgroup namespace.
2541 Without cgroup namespace, the "/proc/$PID/cgroup" file shows the
2542 complete path of the cgroup of a process. In a container setup where
2543 a set of cgroups and namespaces are intended to isolate processes the
2544 "/proc/$PID/cgroup" file may leak potential system level information
2545 to the isolated processes. For example::
2547 # cat /proc/self/cgroup
2548 0::/batchjobs/container_id1
2550 The path '/batchjobs/container_id1' can be considered as system-data
2551 and undesirable to expose to the isolated processes. cgroup namespace
2552 can be used to restrict visibility of this path. For example, before
2553 creating a cgroup namespace, one would see::
2555 # ls -l /proc/self/ns/cgroup
2556 lrwxrwxrwx 1 root root 0 2014-07-15 10:37 /proc/self/ns/cgroup -> cgroup:[4026531835]
2557 # cat /proc/self/cgroup
2558 0::/batchjobs/container_id1
2560 After unsharing a new namespace, the view changes::
2562 # ls -l /proc/self/ns/cgroup
2563 lrwxrwxrwx 1 root root 0 2014-07-15 10:35 /proc/self/ns/cgroup -> cgroup:[4026532183]
2564 # cat /proc/self/cgroup
2567 When some thread from a multi-threaded process unshares its cgroup
2568 namespace, the new cgroupns gets applied to the entire process (all
2569 the threads). This is natural for the v2 hierarchy; however, for the
2570 legacy hierarchies, this may be unexpected.
2572 A cgroup namespace is alive as long as there are processes inside or
2573 mounts pinning it. When the last usage goes away, the cgroup
2574 namespace is destroyed. The cgroupns root and the actual cgroups
2581 The 'cgroupns root' for a cgroup namespace is the cgroup in which the
2582 process calling unshare(2) is running. For example, if a process in
2583 /batchjobs/container_id1 cgroup calls unshare, cgroup
2584 /batchjobs/container_id1 becomes the cgroupns root. For the
2585 init_cgroup_ns, this is the real root ('/') cgroup.
2587 The cgroupns root cgroup does not change even if the namespace creator
2588 process later moves to a different cgroup::
2590 # ~/unshare -c # unshare cgroupns in some cgroup
2591 # cat /proc/self/cgroup
2594 # echo 0 > sub_cgrp_1/cgroup.procs
2595 # cat /proc/self/cgroup
2598 Each process gets its namespace-specific view of "/proc/$PID/cgroup"
2600 Processes running inside the cgroup namespace will be able to see
2601 cgroup paths (in /proc/self/cgroup) only inside their root cgroup.
2602 From within an unshared cgroupns::
2606 # echo 7353 > sub_cgrp_1/cgroup.procs
2607 # cat /proc/7353/cgroup
2610 From the initial cgroup namespace, the real cgroup path will be
2613 $ cat /proc/7353/cgroup
2614 0::/batchjobs/container_id1/sub_cgrp_1
2616 From a sibling cgroup namespace (that is, a namespace rooted at a
2617 different cgroup), the cgroup path relative to its own cgroup
2618 namespace root will be shown. For instance, if PID 7353's cgroup
2619 namespace root is at '/batchjobs/container_id2', then it will see::
2621 # cat /proc/7353/cgroup
2622 0::/../container_id2/sub_cgrp_1
2624 Note that the relative path always starts with '/' to indicate that
2625 its relative to the cgroup namespace root of the caller.
2628 Migration and setns(2)
2629 ----------------------
2631 Processes inside a cgroup namespace can move into and out of the
2632 namespace root if they have proper access to external cgroups. For
2633 example, from inside a namespace with cgroupns root at
2634 /batchjobs/container_id1, and assuming that the global hierarchy is
2635 still accessible inside cgroupns::
2637 # cat /proc/7353/cgroup
2639 # echo 7353 > batchjobs/container_id2/cgroup.procs
2640 # cat /proc/7353/cgroup
2641 0::/../container_id2
2643 Note that this kind of setup is not encouraged. A task inside cgroup
2644 namespace should only be exposed to its own cgroupns hierarchy.
2646 setns(2) to another cgroup namespace is allowed when:
2648 (a) the process has CAP_SYS_ADMIN against its current user namespace
2649 (b) the process has CAP_SYS_ADMIN against the target cgroup
2652 No implicit cgroup changes happen with attaching to another cgroup
2653 namespace. It is expected that the someone moves the attaching
2654 process under the target cgroup namespace root.
2657 Interaction with Other Namespaces
2658 ---------------------------------
2660 Namespace specific cgroup hierarchy can be mounted by a process
2661 running inside a non-init cgroup namespace::
2663 # mount -t cgroup2 none $MOUNT_POINT
2665 This will mount the unified cgroup hierarchy with cgroupns root as the
2666 filesystem root. The process needs CAP_SYS_ADMIN against its user and
2669 The virtualization of /proc/self/cgroup file combined with restricting
2670 the view of cgroup hierarchy by namespace-private cgroupfs mount
2671 provides a properly isolated cgroup view inside the container.
2674 Information on Kernel Programming
2675 =================================
2677 This section contains kernel programming information in the areas
2678 where interacting with cgroup is necessary. cgroup core and
2679 controllers are not covered.
2682 Filesystem Support for Writeback
2683 --------------------------------
2685 A filesystem can support cgroup writeback by updating
2686 address_space_operations->writepage[s]() to annotate bio's using the
2687 following two functions.
2689 wbc_init_bio(@wbc, @bio)
2690 Should be called for each bio carrying writeback data and
2691 associates the bio with the inode's owner cgroup and the
2692 corresponding request queue. This must be called after
2693 a queue (device) has been associated with the bio and
2696 wbc_account_cgroup_owner(@wbc, @page, @bytes)
2697 Should be called for each data segment being written out.
2698 While this function doesn't care exactly when it's called
2699 during the writeback session, it's the easiest and most
2700 natural to call it as data segments are added to a bio.
2702 With writeback bio's annotated, cgroup support can be enabled per
2703 super_block by setting SB_I_CGROUPWB in ->s_iflags. This allows for
2704 selective disabling of cgroup writeback support which is helpful when
2705 certain filesystem features, e.g. journaled data mode, are
2708 wbc_init_bio() binds the specified bio to its cgroup. Depending on
2709 the configuration, the bio may be executed at a lower priority and if
2710 the writeback session is holding shared resources, e.g. a journal
2711 entry, may lead to priority inversion. There is no one easy solution
2712 for the problem. Filesystems can try to work around specific problem
2713 cases by skipping wbc_init_bio() and using bio_associate_blkg()
2717 Deprecated v1 Core Features
2718 ===========================
2720 - Multiple hierarchies including named ones are not supported.
2722 - All v1 mount options are not supported.
2724 - The "tasks" file is removed and "cgroup.procs" is not sorted.
2726 - "cgroup.clone_children" is removed.
2728 - /proc/cgroups is meaningless for v2. Use "cgroup.controllers" file
2729 at the root instead.
2732 Issues with v1 and Rationales for v2
2733 ====================================
2735 Multiple Hierarchies
2736 --------------------
2738 cgroup v1 allowed an arbitrary number of hierarchies and each
2739 hierarchy could host any number of controllers. While this seemed to
2740 provide a high level of flexibility, it wasn't useful in practice.
2742 For example, as there is only one instance of each controller, utility
2743 type controllers such as freezer which can be useful in all
2744 hierarchies could only be used in one. The issue is exacerbated by
2745 the fact that controllers couldn't be moved to another hierarchy once
2746 hierarchies were populated. Another issue was that all controllers
2747 bound to a hierarchy were forced to have exactly the same view of the
2748 hierarchy. It wasn't possible to vary the granularity depending on
2749 the specific controller.
2751 In practice, these issues heavily limited which controllers could be
2752 put on the same hierarchy and most configurations resorted to putting
2753 each controller on its own hierarchy. Only closely related ones, such
2754 as the cpu and cpuacct controllers, made sense to be put on the same
2755 hierarchy. This often meant that userland ended up managing multiple
2756 similar hierarchies repeating the same steps on each hierarchy
2757 whenever a hierarchy management operation was necessary.
2759 Furthermore, support for multiple hierarchies came at a steep cost.
2760 It greatly complicated cgroup core implementation but more importantly
2761 the support for multiple hierarchies restricted how cgroup could be
2762 used in general and what controllers was able to do.
2764 There was no limit on how many hierarchies there might be, which meant
2765 that a thread's cgroup membership couldn't be described in finite
2766 length. The key might contain any number of entries and was unlimited
2767 in length, which made it highly awkward to manipulate and led to
2768 addition of controllers which existed only to identify membership,
2769 which in turn exacerbated the original problem of proliferating number
2772 Also, as a controller couldn't have any expectation regarding the
2773 topologies of hierarchies other controllers might be on, each
2774 controller had to assume that all other controllers were attached to
2775 completely orthogonal hierarchies. This made it impossible, or at
2776 least very cumbersome, for controllers to cooperate with each other.
2778 In most use cases, putting controllers on hierarchies which are
2779 completely orthogonal to each other isn't necessary. What usually is
2780 called for is the ability to have differing levels of granularity
2781 depending on the specific controller. In other words, hierarchy may
2782 be collapsed from leaf towards root when viewed from specific
2783 controllers. For example, a given configuration might not care about
2784 how memory is distributed beyond a certain level while still wanting
2785 to control how CPU cycles are distributed.
2791 cgroup v1 allowed threads of a process to belong to different cgroups.
2792 This didn't make sense for some controllers and those controllers
2793 ended up implementing different ways to ignore such situations but
2794 much more importantly it blurred the line between API exposed to
2795 individual applications and system management interface.
2797 Generally, in-process knowledge is available only to the process
2798 itself; thus, unlike service-level organization of processes,
2799 categorizing threads of a process requires active participation from
2800 the application which owns the target process.
2802 cgroup v1 had an ambiguously defined delegation model which got abused
2803 in combination with thread granularity. cgroups were delegated to
2804 individual applications so that they can create and manage their own
2805 sub-hierarchies and control resource distributions along them. This
2806 effectively raised cgroup to the status of a syscall-like API exposed
2809 First of all, cgroup has a fundamentally inadequate interface to be
2810 exposed this way. For a process to access its own knobs, it has to
2811 extract the path on the target hierarchy from /proc/self/cgroup,
2812 construct the path by appending the name of the knob to the path, open
2813 and then read and/or write to it. This is not only extremely clunky
2814 and unusual but also inherently racy. There is no conventional way to
2815 define transaction across the required steps and nothing can guarantee
2816 that the process would actually be operating on its own sub-hierarchy.
2818 cgroup controllers implemented a number of knobs which would never be
2819 accepted as public APIs because they were just adding control knobs to
2820 system-management pseudo filesystem. cgroup ended up with interface
2821 knobs which were not properly abstracted or refined and directly
2822 revealed kernel internal details. These knobs got exposed to
2823 individual applications through the ill-defined delegation mechanism
2824 effectively abusing cgroup as a shortcut to implementing public APIs
2825 without going through the required scrutiny.
2827 This was painful for both userland and kernel. Userland ended up with
2828 misbehaving and poorly abstracted interfaces and kernel exposing and
2829 locked into constructs inadvertently.
2832 Competition Between Inner Nodes and Threads
2833 -------------------------------------------
2835 cgroup v1 allowed threads to be in any cgroups which created an
2836 interesting problem where threads belonging to a parent cgroup and its
2837 children cgroups competed for resources. This was nasty as two
2838 different types of entities competed and there was no obvious way to
2839 settle it. Different controllers did different things.
2841 The cpu controller considered threads and cgroups as equivalents and
2842 mapped nice levels to cgroup weights. This worked for some cases but
2843 fell flat when children wanted to be allocated specific ratios of CPU
2844 cycles and the number of internal threads fluctuated - the ratios
2845 constantly changed as the number of competing entities fluctuated.
2846 There also were other issues. The mapping from nice level to weight
2847 wasn't obvious or universal, and there were various other knobs which
2848 simply weren't available for threads.
2850 The io controller implicitly created a hidden leaf node for each
2851 cgroup to host the threads. The hidden leaf had its own copies of all
2852 the knobs with ``leaf_`` prefixed. While this allowed equivalent
2853 control over internal threads, it was with serious drawbacks. It
2854 always added an extra layer of nesting which wouldn't be necessary
2855 otherwise, made the interface messy and significantly complicated the
2858 The memory controller didn't have a way to control what happened
2859 between internal tasks and child cgroups and the behavior was not
2860 clearly defined. There were attempts to add ad-hoc behaviors and
2861 knobs to tailor the behavior to specific workloads which would have
2862 led to problems extremely difficult to resolve in the long term.
2864 Multiple controllers struggled with internal tasks and came up with
2865 different ways to deal with it; unfortunately, all the approaches were
2866 severely flawed and, furthermore, the widely different behaviors
2867 made cgroup as a whole highly inconsistent.
2869 This clearly is a problem which needs to be addressed from cgroup core
2873 Other Interface Issues
2874 ----------------------
2876 cgroup v1 grew without oversight and developed a large number of
2877 idiosyncrasies and inconsistencies. One issue on the cgroup core side
2878 was how an empty cgroup was notified - a userland helper binary was
2879 forked and executed for each event. The event delivery wasn't
2880 recursive or delegatable. The limitations of the mechanism also led
2881 to in-kernel event delivery filtering mechanism further complicating
2884 Controller interfaces were problematic too. An extreme example is
2885 controllers completely ignoring hierarchical organization and treating
2886 all cgroups as if they were all located directly under the root
2887 cgroup. Some controllers exposed a large amount of inconsistent
2888 implementation details to userland.
2890 There also was no consistency across controllers. When a new cgroup
2891 was created, some controllers defaulted to not imposing extra
2892 restrictions while others disallowed any resource usage until
2893 explicitly configured. Configuration knobs for the same type of
2894 control used widely differing naming schemes and formats. Statistics
2895 and information knobs were named arbitrarily and used different
2896 formats and units even in the same controller.
2898 cgroup v2 establishes common conventions where appropriate and updates
2899 controllers so that they expose minimal and consistent interfaces.
2902 Controller Issues and Remedies
2903 ------------------------------
2908 The original lower boundary, the soft limit, is defined as a limit
2909 that is per default unset. As a result, the set of cgroups that
2910 global reclaim prefers is opt-in, rather than opt-out. The costs for
2911 optimizing these mostly negative lookups are so high that the
2912 implementation, despite its enormous size, does not even provide the
2913 basic desirable behavior. First off, the soft limit has no
2914 hierarchical meaning. All configured groups are organized in a global
2915 rbtree and treated like equal peers, regardless where they are located
2916 in the hierarchy. This makes subtree delegation impossible. Second,
2917 the soft limit reclaim pass is so aggressive that it not just
2918 introduces high allocation latencies into the system, but also impacts
2919 system performance due to overreclaim, to the point where the feature
2920 becomes self-defeating.
2922 The memory.low boundary on the other hand is a top-down allocated
2923 reserve. A cgroup enjoys reclaim protection when it's within its
2924 effective low, which makes delegation of subtrees possible. It also
2925 enjoys having reclaim pressure proportional to its overage when
2926 above its effective low.
2928 The original high boundary, the hard limit, is defined as a strict
2929 limit that can not budge, even if the OOM killer has to be called.
2930 But this generally goes against the goal of making the most out of the
2931 available memory. The memory consumption of workloads varies during
2932 runtime, and that requires users to overcommit. But doing that with a
2933 strict upper limit requires either a fairly accurate prediction of the
2934 working set size or adding slack to the limit. Since working set size
2935 estimation is hard and error prone, and getting it wrong results in
2936 OOM kills, most users tend to err on the side of a looser limit and
2937 end up wasting precious resources.
2939 The memory.high boundary on the other hand can be set much more
2940 conservatively. When hit, it throttles allocations by forcing them
2941 into direct reclaim to work off the excess, but it never invokes the
2942 OOM killer. As a result, a high boundary that is chosen too
2943 aggressively will not terminate the processes, but instead it will
2944 lead to gradual performance degradation. The user can monitor this
2945 and make corrections until the minimal memory footprint that still
2946 gives acceptable performance is found.
2948 In extreme cases, with many concurrent allocations and a complete
2949 breakdown of reclaim progress within the group, the high boundary can
2950 be exceeded. But even then it's mostly better to satisfy the
2951 allocation from the slack available in other groups or the rest of the
2952 system than killing the group. Otherwise, memory.max is there to
2953 limit this type of spillover and ultimately contain buggy or even
2954 malicious applications.
2956 Setting the original memory.limit_in_bytes below the current usage was
2957 subject to a race condition, where concurrent charges could cause the
2958 limit setting to fail. memory.max on the other hand will first set the
2959 limit to prevent new charges, and then reclaim and OOM kill until the
2960 new limit is met - or the task writing to memory.max is killed.
2962 The combined memory+swap accounting and limiting is replaced by real
2963 control over swap space.
2965 The main argument for a combined memory+swap facility in the original
2966 cgroup design was that global or parental pressure would always be
2967 able to swap all anonymous memory of a child group, regardless of the
2968 child's own (possibly untrusted) configuration. However, untrusted
2969 groups can sabotage swapping by other means - such as referencing its
2970 anonymous memory in a tight loop - and an admin can not assume full
2971 swappability when overcommitting untrusted jobs.
2973 For trusted jobs, on the other hand, a combined counter is not an
2974 intuitive userspace interface, and it flies in the face of the idea
2975 that cgroup controllers should account and limit specific physical
2976 resources. Swap space is a resource like all others in the system,
2977 and that's why unified hierarchy allows distributing it separately.