1 Memory Resource Controller
3 NOTE: This document is hopelessly outdated and it asks for a complete
4 rewrite. It still contains a useful information so we are keeping it
5 here but make sure to check the current code if you need a deeper
8 NOTE: The Memory Resource Controller has generically been referred to as the
9 memory controller in this document. Do not confuse memory controller
10 used here with the memory controller that is used in hardware.
14 When we mention a cgroup (cgroupfs's directory) with memory controller,
15 we call it "memory cgroup". When you see git-log and source code, you'll
16 see patch's title and function names tend to use "memcg".
17 In this document, we avoid using it.
19 Benefits and Purpose of the memory controller
21 The memory controller isolates the memory behaviour of a group of tasks
22 from the rest of the system. The article on LWN [12] mentions some probable
23 uses of the memory controller. The memory controller can be used to
25 a. Isolate an application or a group of applications
26 Memory-hungry applications can be isolated and limited to a smaller
28 b. Create a cgroup with a limited amount of memory; this can be used
29 as a good alternative to booting with mem=XXXX.
30 c. Virtualization solutions can control the amount of memory they want
31 to assign to a virtual machine instance.
32 d. A CD/DVD burner could control the amount of memory used by the
33 rest of the system to ensure that burning does not fail due to lack
35 e. There are several other use cases; find one or use the controller just
36 for fun (to learn and hack on the VM subsystem).
38 Current Status: linux-2.6.34-mmotm(development version of 2010/April)
41 - accounting anonymous pages, file caches, swap caches usage and limiting them.
42 - pages are linked to per-memcg LRU exclusively, and there is no global LRU.
43 - optionally, memory+swap usage can be accounted and limited.
44 - hierarchical accounting
46 - moving (recharging) account at moving a task is selectable.
47 - usage threshold notifier
48 - memory pressure notifier
49 - oom-killer disable knob and oom-notifier
50 - Root cgroup has no limit controls.
52 Kernel memory support is a work in progress, and the current version provides
53 basically functionality. (See Section 2.7)
55 Brief summary of control files.
57 tasks # attach a task(thread) and show list of threads
58 cgroup.procs # show list of processes
59 cgroup.event_control # an interface for event_fd()
60 memory.usage_in_bytes # show current usage for memory
62 memory.memsw.usage_in_bytes # show current usage for memory+Swap
64 memory.limit_in_bytes # set/show limit of memory usage
65 memory.memsw.limit_in_bytes # set/show limit of memory+Swap usage
66 memory.failcnt # show the number of memory usage hits limits
67 memory.memsw.failcnt # show the number of memory+Swap hits limits
68 memory.max_usage_in_bytes # show max memory usage recorded
69 memory.memsw.max_usage_in_bytes # show max memory+Swap usage recorded
70 memory.soft_limit_in_bytes # set/show soft limit of memory usage
71 memory.stat # show various statistics
72 memory.use_hierarchy # set/show hierarchical account enabled
73 memory.force_empty # trigger forced page reclaim
74 memory.pressure_level # set memory pressure notifications
75 memory.swappiness # set/show swappiness parameter of vmscan
76 (See sysctl's vm.swappiness)
77 memory.move_charge_at_immigrate # set/show controls of moving charges
78 memory.oom_control # set/show oom controls.
79 memory.numa_stat # show the number of memory usage per numa node
81 memory.kmem.limit_in_bytes # set/show hard limit for kernel memory
82 memory.kmem.usage_in_bytes # show current kernel memory allocation
83 memory.kmem.failcnt # show the number of kernel memory usage hits limits
84 memory.kmem.max_usage_in_bytes # show max kernel memory usage recorded
86 memory.kmem.tcp.limit_in_bytes # set/show hard limit for tcp buf memory
87 memory.kmem.tcp.usage_in_bytes # show current tcp buf memory allocation
88 memory.kmem.tcp.failcnt # show the number of tcp buf memory usage hits limits
89 memory.kmem.tcp.max_usage_in_bytes # show max tcp buf memory usage recorded
93 The memory controller has a long history. A request for comments for the memory
94 controller was posted by Balbir Singh [1]. At the time the RFC was posted
95 there were several implementations for memory control. The goal of the
96 RFC was to build consensus and agreement for the minimal features required
97 for memory control. The first RSS controller was posted by Balbir Singh[2]
98 in Feb 2007. Pavel Emelianov [3][4][5] has since posted three versions of the
99 RSS controller. At OLS, at the resource management BoF, everyone suggested
100 that we handle both page cache and RSS together. Another request was raised
101 to allow user space handling of OOM. The current memory controller is
102 at version 6; it combines both mapped (RSS) and unmapped Page
107 Memory is a unique resource in the sense that it is present in a limited
108 amount. If a task requires a lot of CPU processing, the task can spread
109 its processing over a period of hours, days, months or years, but with
110 memory, the same physical memory needs to be reused to accomplish the task.
112 The memory controller implementation has been divided into phases. These
116 2. mlock(2) controller
117 3. Kernel user memory accounting and slab control
118 4. user mappings length controller
120 The memory controller is the first controller developed.
124 The core of the design is a counter called the page_counter. The
125 page_counter tracks the current memory usage and limit of the group of
126 processes associated with the controller. Each cgroup has a memory controller
127 specific data structure (mem_cgroup) associated with it.
131 +--------------------+
134 +--------------------+
137 +---------------+ | +---------------+
138 | mm_struct | |.... | mm_struct |
140 +---------------+ | +---------------+
144 +---------------+ +------+--------+
145 | page +----------> page_cgroup|
147 +---------------+ +---------------+
149 (Figure 1: Hierarchy of Accounting)
152 Figure 1 shows the important aspects of the controller
154 1. Accounting happens per cgroup
155 2. Each mm_struct knows about which cgroup it belongs to
156 3. Each page has a pointer to the page_cgroup, which in turn knows the
159 The accounting is done as follows: mem_cgroup_charge_common() is invoked to
160 set up the necessary data structures and check if the cgroup that is being
161 charged is over its limit. If it is, then reclaim is invoked on the cgroup.
162 More details can be found in the reclaim section of this document.
163 If everything goes well, a page meta-data-structure called page_cgroup is
164 updated. page_cgroup has its own LRU on cgroup.
165 (*) page_cgroup structure is allocated at boot/memory-hotplug time.
167 2.2.1 Accounting details
169 All mapped anon pages (RSS) and cache pages (Page Cache) are accounted.
170 Some pages which are never reclaimable and will not be on the LRU
171 are not accounted. We just account pages under usual VM management.
173 RSS pages are accounted at page_fault unless they've already been accounted
174 for earlier. A file page will be accounted for as Page Cache when it's
175 inserted into inode (radix-tree). While it's mapped into the page tables of
176 processes, duplicate accounting is carefully avoided.
178 An RSS page is unaccounted when it's fully unmapped. A PageCache page is
179 unaccounted when it's removed from radix-tree. Even if RSS pages are fully
180 unmapped (by kswapd), they may exist as SwapCache in the system until they
181 are really freed. Such SwapCaches are also accounted.
182 A swapped-in page is not accounted until it's mapped.
184 Note: The kernel does swapin-readahead and reads multiple swaps at once.
185 This means swapped-in pages may contain pages for other tasks than a task
186 causing page fault. So, we avoid accounting at swap-in I/O.
188 At page migration, accounting information is kept.
190 Note: we just account pages-on-LRU because our purpose is to control amount
191 of used pages; not-on-LRU pages tend to be out-of-control from VM view.
193 2.3 Shared Page Accounting
195 Shared pages are accounted on the basis of the first touch approach. The
196 cgroup that first touches a page is accounted for the page. The principle
197 behind this approach is that a cgroup that aggressively uses a shared
198 page will eventually get charged for it (once it is uncharged from
199 the cgroup that brought it in -- this will happen on memory pressure).
201 But see section 8.2: when moving a task to another cgroup, its pages may
202 be recharged to the new cgroup, if move_charge_at_immigrate has been chosen.
204 Exception: If CONFIG_MEMCG_SWAP is not used.
205 When you do swapoff and make swapped-out pages of shmem(tmpfs) to
206 be backed into memory in force, charges for pages are accounted against the
207 caller of swapoff rather than the users of shmem.
209 2.4 Swap Extension (CONFIG_MEMCG_SWAP)
211 Swap Extension allows you to record charge for swap. A swapped-in page is
212 charged back to original page allocator if possible.
214 When swap is accounted, following files are added.
215 - memory.memsw.usage_in_bytes.
216 - memory.memsw.limit_in_bytes.
218 memsw means memory+swap. Usage of memory+swap is limited by
219 memsw.limit_in_bytes.
221 Example: Assume a system with 4G of swap. A task which allocates 6G of memory
222 (by mistake) under 2G memory limitation will use all swap.
223 In this case, setting memsw.limit_in_bytes=3G will prevent bad use of swap.
224 By using the memsw limit, you can avoid system OOM which can be caused by swap
227 * why 'memory+swap' rather than swap.
228 The global LRU(kswapd) can swap out arbitrary pages. Swap-out means
229 to move account from memory to swap...there is no change in usage of
230 memory+swap. In other words, when we want to limit the usage of swap without
231 affecting global LRU, memory+swap limit is better than just limiting swap from
234 * What happens when a cgroup hits memory.memsw.limit_in_bytes
235 When a cgroup hits memory.memsw.limit_in_bytes, it's useless to do swap-out
236 in this cgroup. Then, swap-out will not be done by cgroup routine and file
237 caches are dropped. But as mentioned above, global LRU can do swapout memory
238 from it for sanity of the system's memory management state. You can't forbid
243 Each cgroup maintains a per cgroup LRU which has the same structure as
244 global VM. When a cgroup goes over its limit, we first try
245 to reclaim memory from the cgroup so as to make space for the new
246 pages that the cgroup has touched. If the reclaim is unsuccessful,
247 an OOM routine is invoked to select and kill the bulkiest task in the
248 cgroup. (See 10. OOM Control below.)
250 The reclaim algorithm has not been modified for cgroups, except that
251 pages that are selected for reclaiming come from the per-cgroup LRU
254 NOTE: Reclaim does not work for the root cgroup, since we cannot set any
255 limits on the root cgroup.
257 Note2: When panic_on_oom is set to "2", the whole system will panic.
259 When oom event notifier is registered, event will be delivered.
260 (See oom_control section)
264 lock_page_cgroup()/unlock_page_cgroup() should not be called under
267 Other lock order is following:
272 In many cases, just lock_page_cgroup() is called.
273 per-zone-per-cgroup LRU (cgroup's private LRU) is just guarded by
274 pgdat->lru_lock, it has no lock of its own.
276 2.7 Kernel Memory Extension (CONFIG_MEMCG_KMEM)
278 With the Kernel memory extension, the Memory Controller is able to limit
279 the amount of kernel memory used by the system. Kernel memory is fundamentally
280 different than user memory, since it can't be swapped out, which makes it
281 possible to DoS the system by consuming too much of this precious resource.
283 Kernel memory accounting is enabled for all memory cgroups by default. But
284 it can be disabled system-wide by passing cgroup.memory=nokmem to the kernel
285 at boot time. In this case, kernel memory will not be accounted at all.
287 Kernel memory limits are not imposed for the root cgroup. Usage for the root
288 cgroup may or may not be accounted. The memory used is accumulated into
289 memory.kmem.usage_in_bytes, or in a separate counter when it makes sense.
290 (currently only for tcp).
291 The main "kmem" counter is fed into the main counter, so kmem charges will
292 also be visible from the user counter.
294 Currently no soft limit is implemented for kernel memory. It is future work
295 to trigger slab reclaim when those limits are reached.
297 2.7.1 Current Kernel Memory resources accounted
299 * stack pages: every process consumes some stack pages. By accounting into
300 kernel memory, we prevent new processes from being created when the kernel
301 memory usage is too high.
303 * slab pages: pages allocated by the SLAB or SLUB allocator are tracked. A copy
304 of each kmem_cache is created every time the cache is touched by the first time
305 from inside the memcg. The creation is done lazily, so some objects can still be
306 skipped while the cache is being created. All objects in a slab page should
307 belong to the same memcg. This only fails to hold when a task is migrated to a
308 different memcg during the page allocation by the cache.
310 * sockets memory pressure: some sockets protocols have memory pressure
311 thresholds. The Memory Controller allows them to be controlled individually
312 per cgroup, instead of globally.
314 * tcp memory pressure: sockets memory pressure for the tcp protocol.
316 2.7.2 Common use cases
318 Because the "kmem" counter is fed to the main user counter, kernel memory can
319 never be limited completely independently of user memory. Say "U" is the user
320 limit, and "K" the kernel limit. There are three possible ways limits can be
323 U != 0, K = unlimited:
324 This is the standard memcg limitation mechanism already present before kmem
325 accounting. Kernel memory is completely ignored.
328 Kernel memory is a subset of the user memory. This setup is useful in
329 deployments where the total amount of memory per-cgroup is overcommited.
330 Overcommiting kernel memory limits is definitely not recommended, since the
331 box can still run out of non-reclaimable memory.
332 In this case, the admin could set up K so that the sum of all groups is
333 never greater than the total memory, and freely set U at the cost of his
335 WARNING: In the current implementation, memory reclaim will NOT be
336 triggered for a cgroup when it hits K while staying below U, which makes
337 this setup impractical.
340 Since kmem charges will also be fed to the user counter and reclaim will be
341 triggered for the cgroup for both kinds of memory. This setup gives the
342 admin a unified view of memory, and it is also useful for people who just
343 want to track kernel memory usage.
349 a. Enable CONFIG_CGROUPS
350 b. Enable CONFIG_MEMCG
351 c. Enable CONFIG_MEMCG_SWAP (to use swap extension)
352 d. Enable CONFIG_MEMCG_KMEM (to use kmem extension)
354 3.1. Prepare the cgroups (see cgroups.txt, Why are cgroups needed?)
355 # mount -t tmpfs none /sys/fs/cgroup
356 # mkdir /sys/fs/cgroup/memory
357 # mount -t cgroup none /sys/fs/cgroup/memory -o memory
359 3.2. Make the new group and move bash into it
360 # mkdir /sys/fs/cgroup/memory/0
361 # echo $$ > /sys/fs/cgroup/memory/0/tasks
363 Since now we're in the 0 cgroup, we can alter the memory limit:
364 # echo 4M > /sys/fs/cgroup/memory/0/memory.limit_in_bytes
366 NOTE: We can use a suffix (k, K, m, M, g or G) to indicate values in kilo,
367 mega or gigabytes. (Here, Kilo, Mega, Giga are Kibibytes, Mebibytes, Gibibytes.)
369 NOTE: We can write "-1" to reset the *.limit_in_bytes(unlimited).
370 NOTE: We cannot set limits on the root cgroup any more.
372 # cat /sys/fs/cgroup/memory/0/memory.limit_in_bytes
375 We can check the usage:
376 # cat /sys/fs/cgroup/memory/0/memory.usage_in_bytes
379 A successful write to this file does not guarantee a successful setting of
380 this limit to the value written into the file. This can be due to a
381 number of factors, such as rounding up to page boundaries or the total
382 availability of memory on the system. The user is required to re-read
383 this file after a write to guarantee the value committed by the kernel.
385 # echo 1 > memory.limit_in_bytes
386 # cat memory.limit_in_bytes
389 The memory.failcnt field gives the number of times that the cgroup limit was
392 The memory.stat file gives accounting information. Now, the number of
393 caches, RSS and Active pages/Inactive pages are shown.
397 For testing features and implementation, see memcg_test.txt.
399 Performance test is also important. To see pure memory controller's overhead,
400 testing on tmpfs will give you good numbers of small overheads.
401 Example: do kernel make on tmpfs.
403 Page-fault scalability is also important. At measuring parallel
404 page fault test, multi-process test may be better than multi-thread
405 test because it has noise of shared objects/status.
407 But the above two are testing extreme situations.
408 Trying usual test under memory controller is always helpful.
412 Sometimes a user might find that the application under a cgroup is
413 terminated by the OOM killer. There are several causes for this:
415 1. The cgroup limit is too low (just too low to do anything useful)
416 2. The user is using anonymous memory and swap is turned off or too low
418 A sync followed by echo 1 > /proc/sys/vm/drop_caches will help get rid of
419 some of the pages cached in the cgroup (page cache pages).
421 To know what happens, disabling OOM_Kill as per "10. OOM Control" (below) and
422 seeing what happens will be helpful.
426 When a task migrates from one cgroup to another, its charge is not
427 carried forward by default. The pages allocated from the original cgroup still
428 remain charged to it, the charge is dropped when the page is freed or
431 You can move charges of a task along with task migration.
432 See 8. "Move charges at task migration"
434 4.3 Removing a cgroup
436 A cgroup can be removed by rmdir, but as discussed in sections 4.1 and 4.2, a
437 cgroup might have some charge associated with it, even though all
438 tasks have migrated away from it. (because we charge against pages, not
441 We move the stats to root (if use_hierarchy==0) or parent (if
442 use_hierarchy==1), and no change on the charge except uncharging
445 Charges recorded in swap information is not updated at removal of cgroup.
446 Recorded information is discarded and a cgroup which uses swap (swapcache)
447 will be charged as a new owner of it.
449 About use_hierarchy, see Section 6.
454 memory.force_empty interface is provided to make cgroup's memory usage empty.
455 When writing anything to this
457 # echo 0 > memory.force_empty
459 the cgroup will be reclaimed and as many pages reclaimed as possible.
461 The typical use case for this interface is before calling rmdir().
462 Though rmdir() offlines memcg, but the memcg may still stay there due to
463 charged file caches. Some out-of-use page caches may keep charged until
464 memory pressure happens. If you want to avoid that, force_empty will be useful.
466 Also, note that when memory.kmem.limit_in_bytes is set the charges due to
467 kernel pages will still be seen. This is not considered a failure and the
468 write will still return success. In this case, it is expected that
469 memory.kmem.usage_in_bytes == memory.usage_in_bytes.
471 About use_hierarchy, see Section 6.
475 memory.stat file includes following statistics
477 # per-memory cgroup local status
478 cache - # of bytes of page cache memory.
479 rss - # of bytes of anonymous and swap cache memory (includes
480 transparent hugepages).
481 rss_huge - # of bytes of anonymous transparent hugepages.
482 mapped_file - # of bytes of mapped file (includes tmpfs/shmem)
483 pgpgin - # of charging events to the memory cgroup. The charging
484 event happens each time a page is accounted as either mapped
485 anon page(RSS) or cache page(Page Cache) to the cgroup.
486 pgpgout - # of uncharging events to the memory cgroup. The uncharging
487 event happens each time a page is unaccounted from the cgroup.
488 swap - # of bytes of swap usage
489 dirty - # of bytes that are waiting to get written back to the disk.
490 writeback - # of bytes of file/anon cache that are queued for syncing to
492 inactive_anon - # of bytes of anonymous and swap cache memory on inactive
494 active_anon - # of bytes of anonymous and swap cache memory on active
496 inactive_file - # of bytes of file-backed memory on inactive LRU list.
497 active_file - # of bytes of file-backed memory on active LRU list.
498 unevictable - # of bytes of memory that cannot be reclaimed (mlocked etc).
500 # status considering hierarchy (see memory.use_hierarchy settings)
502 hierarchical_memory_limit - # of bytes of memory limit with regard to hierarchy
503 under which the memory cgroup is
504 hierarchical_memsw_limit - # of bytes of memory+swap limit with regard to
505 hierarchy under which memory cgroup is.
507 total_<counter> - # hierarchical version of <counter>, which in
508 addition to the cgroup's own value includes the
509 sum of all hierarchical children's values of
510 <counter>, i.e. total_cache
512 # The following additional stats are dependent on CONFIG_DEBUG_VM.
514 recent_rotated_anon - VM internal parameter. (see mm/vmscan.c)
515 recent_rotated_file - VM internal parameter. (see mm/vmscan.c)
516 recent_scanned_anon - VM internal parameter. (see mm/vmscan.c)
517 recent_scanned_file - VM internal parameter. (see mm/vmscan.c)
520 recent_rotated means recent frequency of LRU rotation.
521 recent_scanned means recent # of scans to LRU.
522 showing for better debug please see the code for meanings.
525 Only anonymous and swap cache memory is listed as part of 'rss' stat.
526 This should not be confused with the true 'resident set size' or the
527 amount of physical memory used by the cgroup.
528 'rss + mapped_file" will give you resident set size of cgroup.
529 (Note: file and shmem may be shared among other cgroups. In that case,
530 mapped_file is accounted only when the memory cgroup is owner of page
535 Overrides /proc/sys/vm/swappiness for the particular group. The tunable
536 in the root cgroup corresponds to the global swappiness setting.
538 Please note that unlike during the global reclaim, limit reclaim
539 enforces that 0 swappiness really prevents from any swapping even if
540 there is a swap storage available. This might lead to memcg OOM killer
541 if there are no file pages to reclaim.
545 A memory cgroup provides memory.failcnt and memory.memsw.failcnt files.
546 This failcnt(== failure count) shows the number of times that a usage counter
547 hit its limit. When a memory cgroup hits a limit, failcnt increases and
548 memory under it will be reclaimed.
550 You can reset failcnt by writing 0 to failcnt file.
551 # echo 0 > .../memory.failcnt
555 For efficiency, as other kernel components, memory cgroup uses some optimization
556 to avoid unnecessary cacheline false sharing. usage_in_bytes is affected by the
557 method and doesn't show 'exact' value of memory (and swap) usage, it's a fuzz
558 value for efficient access. (Of course, when necessary, it's synchronized.)
559 If you want to know more exact memory usage, you should use RSS+CACHE(+SWAP)
560 value in memory.stat(see 5.2).
564 This is similar to numa_maps but operates on a per-memcg basis. This is
565 useful for providing visibility into the numa locality information within
566 an memcg since the pages are allowed to be allocated from any physical
567 node. One of the use cases is evaluating application performance by
568 combining this information with the application's CPU allocation.
570 Each memcg's numa_stat file includes "total", "file", "anon" and "unevictable"
571 per-node page counts including "hierarchical_<counter>" which sums up all
572 hierarchical children's values in addition to the memcg's own value.
574 The output format of memory.numa_stat is:
576 total=<total pages> N0=<node 0 pages> N1=<node 1 pages> ...
577 file=<total file pages> N0=<node 0 pages> N1=<node 1 pages> ...
578 anon=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
579 unevictable=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
580 hierarchical_<counter>=<counter pages> N0=<node 0 pages> N1=<node 1 pages> ...
582 The "total" count is sum of file + anon + unevictable.
586 The memory controller supports a deep hierarchy and hierarchical accounting.
587 The hierarchy is created by creating the appropriate cgroups in the
588 cgroup filesystem. Consider for example, the following cgroup filesystem
599 In the diagram above, with hierarchical accounting enabled, all memory
600 usage of e, is accounted to its ancestors up until the root (i.e, c and root),
601 that has memory.use_hierarchy enabled. If one of the ancestors goes over its
602 limit, the reclaim algorithm reclaims from the tasks in the ancestor and the
603 children of the ancestor.
605 6.1 Enabling hierarchical accounting and reclaim
607 A memory cgroup by default disables the hierarchy feature. Support
608 can be enabled by writing 1 to memory.use_hierarchy file of the root cgroup
610 # echo 1 > memory.use_hierarchy
612 The feature can be disabled by
614 # echo 0 > memory.use_hierarchy
616 NOTE1: Enabling/disabling will fail if either the cgroup already has other
617 cgroups created below it, or if the parent cgroup has use_hierarchy
620 NOTE2: When panic_on_oom is set to "2", the whole system will panic in
621 case of an OOM event in any cgroup.
625 Soft limits allow for greater sharing of memory. The idea behind soft limits
626 is to allow control groups to use as much of the memory as needed, provided
628 a. There is no memory contention
629 b. They do not exceed their hard limit
631 When the system detects memory contention or low memory, control groups
632 are pushed back to their soft limits. If the soft limit of each control
633 group is very high, they are pushed back as much as possible to make
634 sure that one control group does not starve the others of memory.
636 Please note that soft limits is a best-effort feature; it comes with
637 no guarantees, but it does its best to make sure that when memory is
638 heavily contended for, memory is allocated based on the soft limit
639 hints/setup. Currently soft limit based reclaim is set up such that
640 it gets invoked from balance_pgdat (kswapd).
644 Soft limits can be setup by using the following commands (in this example we
645 assume a soft limit of 256 MiB)
647 # echo 256M > memory.soft_limit_in_bytes
649 If we want to change this to 1G, we can at any time use
651 # echo 1G > memory.soft_limit_in_bytes
653 NOTE1: Soft limits take effect over a long period of time, since they involve
654 reclaiming memory for balancing between memory cgroups
655 NOTE2: It is recommended to set the soft limit always below the hard limit,
656 otherwise the hard limit will take precedence.
658 8. Move charges at task migration
660 Users can move charges associated with a task along with task migration, that
661 is, uncharge task's pages from the old cgroup and charge them to the new cgroup.
662 This feature is not supported in !CONFIG_MMU environments because of lack of
667 This feature is disabled by default. It can be enabled (and disabled again) by
668 writing to memory.move_charge_at_immigrate of the destination cgroup.
670 If you want to enable it:
672 # echo (some positive value) > memory.move_charge_at_immigrate
674 Note: Each bits of move_charge_at_immigrate has its own meaning about what type
675 of charges should be moved. See 8.2 for details.
676 Note: Charges are moved only when you move mm->owner, in other words,
677 a leader of a thread group.
678 Note: If we cannot find enough space for the task in the destination cgroup, we
679 try to make space by reclaiming memory. Task migration may fail if we
680 cannot make enough space.
681 Note: It can take several seconds if you move charges much.
683 And if you want disable it again:
685 # echo 0 > memory.move_charge_at_immigrate
687 8.2 Type of charges which can be moved
689 Each bit in move_charge_at_immigrate has its own meaning about what type of
690 charges should be moved. But in any case, it must be noted that an account of
691 a page or a swap can be moved only when it is charged to the task's current
694 bit | what type of charges would be moved ?
695 -----+------------------------------------------------------------------------
696 0 | A charge of an anonymous page (or swap of it) used by the target task.
697 | You must enable Swap Extension (see 2.4) to enable move of swap charges.
698 -----+------------------------------------------------------------------------
699 1 | A charge of file pages (normal file, tmpfs file (e.g. ipc shared memory)
700 | and swaps of tmpfs file) mmapped by the target task. Unlike the case of
701 | anonymous pages, file pages (and swaps) in the range mmapped by the task
702 | will be moved even if the task hasn't done page fault, i.e. they might
703 | not be the task's "RSS", but other task's "RSS" that maps the same file.
704 | And mapcount of the page is ignored (the page can be moved even if
705 | page_mapcount(page) > 1). You must enable Swap Extension (see 2.4) to
706 | enable move of swap charges.
710 - All of moving charge operations are done under cgroup_mutex. It's not good
711 behavior to hold the mutex too long, so we may need some trick.
715 Memory cgroup implements memory thresholds using the cgroups notification
716 API (see cgroups.txt). It allows to register multiple memory and memsw
717 thresholds and gets notifications when it crosses.
719 To register a threshold, an application must:
720 - create an eventfd using eventfd(2);
721 - open memory.usage_in_bytes or memory.memsw.usage_in_bytes;
722 - write string like "<event_fd> <fd of memory.usage_in_bytes> <threshold>" to
723 cgroup.event_control.
725 Application will be notified through eventfd when memory usage crosses
726 threshold in any direction.
728 It's applicable for root and non-root cgroup.
732 memory.oom_control file is for OOM notification and other controls.
734 Memory cgroup implements OOM notifier using the cgroup notification
735 API (See cgroups.txt). It allows to register multiple OOM notification
736 delivery and gets notification when OOM happens.
738 To register a notifier, an application must:
739 - create an eventfd using eventfd(2)
740 - open memory.oom_control file
741 - write string like "<event_fd> <fd of memory.oom_control>" to
744 The application will be notified through eventfd when OOM happens.
745 OOM notification doesn't work for the root cgroup.
747 You can disable the OOM-killer by writing "1" to memory.oom_control file, as:
749 #echo 1 > memory.oom_control
751 If OOM-killer is disabled, tasks under cgroup will hang/sleep
752 in memory cgroup's OOM-waitqueue when they request accountable memory.
754 For running them, you have to relax the memory cgroup's OOM status by
755 * enlarge limit or reduce usage.
758 * move some tasks to other group with account migration.
759 * remove some files (on tmpfs?)
761 Then, stopped tasks will work again.
763 At reading, current status of OOM is shown.
764 oom_kill_disable 0 or 1 (if 1, oom-killer is disabled)
765 under_oom 0 or 1 (if 1, the memory cgroup is under OOM, tasks may
770 The pressure level notifications can be used to monitor the memory
771 allocation cost; based on the pressure, applications can implement
772 different strategies of managing their memory resources. The pressure
773 levels are defined as following:
775 The "low" level means that the system is reclaiming memory for new
776 allocations. Monitoring this reclaiming activity might be useful for
777 maintaining cache level. Upon notification, the program (typically
778 "Activity Manager") might analyze vmstat and act in advance (i.e.
779 prematurely shutdown unimportant services).
781 The "medium" level means that the system is experiencing medium memory
782 pressure, the system might be making swap, paging out active file caches,
783 etc. Upon this event applications may decide to further analyze
784 vmstat/zoneinfo/memcg or internal memory usage statistics and free any
785 resources that can be easily reconstructed or re-read from a disk.
787 The "critical" level means that the system is actively thrashing, it is
788 about to out of memory (OOM) or even the in-kernel OOM killer is on its
789 way to trigger. Applications should do whatever they can to help the
790 system. It might be too late to consult with vmstat or any other
791 statistics, so it's advisable to take an immediate action.
793 By default, events are propagated upward until the event is handled, i.e. the
794 events are not pass-through. For example, you have three cgroups: A->B->C. Now
795 you set up an event listener on cgroups A, B and C, and suppose group C
796 experiences some pressure. In this situation, only group C will receive the
797 notification, i.e. groups A and B will not receive it. This is done to avoid
798 excessive "broadcasting" of messages, which disturbs the system and which is
799 especially bad if we are low on memory or thrashing. Group B, will receive
800 notification only if there are no event listers for group C.
802 There are three optional modes that specify different propagation behavior:
804 - "default": this is the default behavior specified above. This mode is the
805 same as omitting the optional mode parameter, preserved by backwards
808 - "hierarchy": events always propagate up to the root, similar to the default
809 behavior, except that propagation continues regardless of whether there are
810 event listeners at each level, with the "hierarchy" mode. In the above
811 example, groups A, B, and C will receive notification of memory pressure.
813 - "local": events are pass-through, i.e. they only receive notifications when
814 memory pressure is experienced in the memcg for which the notification is
815 registered. In the above example, group C will receive notification if
816 registered for "local" notification and the group experiences memory
817 pressure. However, group B will never receive notification, regardless if
818 there is an event listener for group C or not, if group B is registered for
821 The level and event notification mode ("hierarchy" or "local", if necessary) are
822 specified by a comma-delimited string, i.e. "low,hierarchy" specifies
823 hierarchical, pass-through, notification for all ancestor memcgs. Notification
824 that is the default, non pass-through behavior, does not specify a mode.
825 "medium,local" specifies pass-through notification for the medium level.
827 The file memory.pressure_level is only used to setup an eventfd. To
828 register a notification, an application must:
830 - create an eventfd using eventfd(2);
831 - open memory.pressure_level;
832 - write string as "<event_fd> <fd of memory.pressure_level> <level[,mode]>"
833 to cgroup.event_control.
835 Application will be notified through eventfd when memory pressure is at
836 the specific level (or higher). Read/write operations to
837 memory.pressure_level are no implemented.
841 Here is a small script example that makes a new cgroup, sets up a
842 memory limit, sets up a notification in the cgroup and then makes child
843 cgroup experience a critical pressure:
845 # cd /sys/fs/cgroup/memory/
848 # cgroup_event_listener memory.pressure_level low,hierarchy &
849 # echo 8000000 > memory.limit_in_bytes
850 # echo 8000000 > memory.memsw.limit_in_bytes
852 # dd if=/dev/zero | read x
854 (Expect a bunch of notifications, and eventually, the oom-killer will
859 1. Make per-cgroup scanner reclaim not-shared pages first
860 2. Teach controller to account for shared-pages
861 3. Start reclamation in the background when the limit is
862 not yet hit but the usage is getting closer
866 Overall, the memory controller has been a stable controller and has been
867 commented and discussed quite extensively in the community.
871 1. Singh, Balbir. RFC: Memory Controller, http://lwn.net/Articles/206697/
872 2. Singh, Balbir. Memory Controller (RSS Control),
873 http://lwn.net/Articles/222762/
874 3. Emelianov, Pavel. Resource controllers based on process cgroups
875 http://lkml.org/lkml/2007/3/6/198
876 4. Emelianov, Pavel. RSS controller based on process cgroups (v2)
877 http://lkml.org/lkml/2007/4/9/78
878 5. Emelianov, Pavel. RSS controller based on process cgroups (v3)
879 http://lkml.org/lkml/2007/5/30/244
880 6. Menage, Paul. Control Groups v10, http://lwn.net/Articles/236032/
881 7. Vaidyanathan, Srinivasan, Control Groups: Pagecache accounting and control
882 subsystem (v3), http://lwn.net/Articles/235534/
883 8. Singh, Balbir. RSS controller v2 test results (lmbench),
884 http://lkml.org/lkml/2007/5/17/232
885 9. Singh, Balbir. RSS controller v2 AIM9 results
886 http://lkml.org/lkml/2007/5/18/1
887 10. Singh, Balbir. Memory controller v6 test results,
888 http://lkml.org/lkml/2007/8/19/36
889 11. Singh, Balbir. Memory controller introduction (v6),
890 http://lkml.org/lkml/2007/8/17/69
891 12. Corbet, Jonathan, Controlling memory use in cgroups,
892 http://lwn.net/Articles/243795/