Lines Matching +full:memory +full:- +full:to +full:- +full:memory
2 Memory Resource Controller
8 here but make sure to check the current code if you need a deeper
12 The Memory Resource Controller has generically been referred to as the
13 memory controller in this document. Do not confuse memory controller
14 used here with the memory controller that is used in hardware.
17 When we mention a cgroup (cgroupfs's directory) with memory controller,
18 we call it "memory cgroup". When you see git-log and source code, you'll
19 see patch's title and function names tend to use "memcg".
22 Benefits and Purpose of the memory controller
25 The memory controller isolates the memory behaviour of a group of tasks
27 uses of the memory controller. The memory controller can be used to
30 Memory-hungry applications can be isolated and limited to a smaller
31 amount of memory.
32 b. Create a cgroup with a limited amount of memory; this can be used
33 as a good alternative to booting with mem=XXXX.
34 c. Virtualization solutions can control the amount of memory they want
35 to assign to a virtual machine instance.
36 d. A CD/DVD burner could control the amount of memory used by the
37 rest of the system to ensure that burning does not fail due to lack
38 of available memory.
40 for fun (to learn and hack on the VM subsystem).
42 Current Status: linux-2.6.34-mmotm(development version of 2010/April)
46 - accounting anonymous pages, file caches, swap caches usage and limiting them.
47 - pages are linked to per-memcg LRU exclusively, and there is no global LRU.
48 - optionally, memory+swap usage can be accounted and limited.
49 - hierarchical accounting
50 - soft limit
51 - moving (recharging) account at moving a task is selectable.
52 - usage threshold notifier
53 - memory pressure notifier
54 - oom-killer disable knob and oom-notifier
55 - Root cgroup has no limit controls.
57 Kernel memory support is a work in progress, and the current version provides
67 memory.usage_in_bytes show current usage for memory
69 memory.memsw.usage_in_bytes show current usage for memory+Swap
71 memory.limit_in_bytes set/show limit of memory usage
72 memory.memsw.limit_in_bytes set/show limit of memory+Swap usage
73 memory.failcnt show the number of memory usage hits limits
74 memory.memsw.failcnt show the number of memory+Swap hits limits
75 memory.max_usage_in_bytes show max memory usage recorded
76 memory.memsw.max_usage_in_bytes show max memory+Swap usage recorded
77 memory.soft_limit_in_bytes set/show soft limit of memory usage
78 memory.stat show various statistics
79 memory.use_hierarchy set/show hierarchical account enabled
80 memory.force_empty trigger forced page reclaim
81 memory.pressure_level set memory pressure notifications
82 memory.swappiness set/show swappiness parameter of vmscan
84 memory.move_charge_at_immigrate set/show controls of moving charges
85 memory.oom_control set/show oom controls.
86 memory.numa_stat show the number of memory usage per numa
88 memory.kmem.limit_in_bytes set/show hard limit for kernel memory
92 memory.kmem.usage_in_bytes show current kernel memory allocation
93 memory.kmem.failcnt show the number of kernel memory usage
95 memory.kmem.max_usage_in_bytes show max kernel memory usage recorded
97 memory.kmem.tcp.limit_in_bytes set/show hard limit for tcp buf memory
98 memory.kmem.tcp.usage_in_bytes show current tcp buf memory allocation
99 memory.kmem.tcp.failcnt show the number of tcp buf memory usage
101 memory.kmem.tcp.max_usage_in_bytes show max tcp buf memory usage recorded
107 The memory controller has a long history. A request for comments for the memory
109 there were several implementations for memory control. The goal of the
110 RFC was to build consensus and agreement for the minimal features required
111 for memory control. The first RSS controller was posted by Balbir Singh[2]
115 to allow user space handling of OOM. The current memory controller is
119 2. Memory Control
122 Memory is a unique resource in the sense that it is present in a limited
125 memory, the same physical memory needs to be reused to accomplish the task.
127 The memory controller implementation has been divided into phases. These
130 1. Memory controller
132 3. Kernel user memory accounting and slab control
135 The memory controller is the first controller developed.
138 -----------
141 page_counter tracks the current memory usage and limit of the group of
142 processes associated with the controller. Each cgroup has a memory controller
146 ---------------
150 +--------------------+
153 +--------------------+
156 +---------------+ | +---------------+
159 +---------------+ | +---------------+
161 + --------------+
163 +---------------+ +------+--------+
164 | page +----------> page_cgroup|
166 +---------------+ +---------------+
174 2. Each mm_struct knows about which cgroup it belongs to
175 3. Each page has a pointer to the page_cgroup, which in turn knows the
176 cgroup it belongs to
178 The accounting is done as follows: mem_cgroup_charge_common() is invoked to
182 If everything goes well, a page meta-data-structure called page_cgroup is
184 (*) page_cgroup structure is allocated at boot/memory-hotplug time.
187 ------------------------
195 inserted into inode (radix-tree). While it's mapped into the page tables of
199 unaccounted when it's removed from radix-tree. Even if RSS pages are fully
202 A swapped-in page is accounted after adding into swapcache.
204 Note: The kernel does swapin-readahead and reads multiple swaps at once.
210 Note: we just account pages-on-LRU because our purpose is to control amount
211 of used pages; not-on-LRU pages tend to be out-of-control from VM view.
214 --------------------------
220 the cgroup that brought it in -- this will happen on memory pressure).
222 But see section 8.2: when moving a task to another cgroup, its pages may
223 be recharged to the new cgroup, if move_charge_at_immigrate has been chosen.
226 --------------------------------------
228 Swap usage is always recorded for each of cgroup. Swap Extension allows you to
233 - memory.memsw.usage_in_bytes.
234 - memory.memsw.limit_in_bytes.
236 memsw means memory+swap. Usage of memory+swap is limited by
239 Example: Assume a system with 4G of swap. A task which allocates 6G of memory
240 (by mistake) under 2G memory limitation will use all swap.
245 **why 'memory+swap' rather than swap**
247 The global LRU(kswapd) can swap out arbitrary pages. Swap-out means
248 to move account from memory to swap...there is no change in usage of
249 memory+swap. In other words, when we want to limit the usage of swap without
250 affecting global LRU, memory+swap limit is better than just limiting swap from
253 **What happens when a cgroup hits memory.memsw.limit_in_bytes**
255 When a cgroup hits memory.memsw.limit_in_bytes, it's useless to do swap-out
256 in this cgroup. Then, swap-out will not be done by cgroup routine and file
257 caches are dropped. But as mentioned above, global LRU can do swapout memory
258 from it for sanity of the system's memory management state. You can't forbid
262 -----------
266 to reclaim memory from the cgroup so as to make space for the new
268 an OOM routine is invoked to select and kill the bulkiest task in the
272 pages that are selected for reclaiming come from the per-cgroup LRU
280 When panic_on_oom is set to "2", the whole system will panic.
286 -----------
294 mm->page_table_lock
295 pgdat->lru_lock
300 per-zone-per-cgroup LRU (cgroup's private LRU) is just guarded by
301 pgdat->lru_lock, it has no lock of its own.
303 2.7 Kernel Memory Extension (CONFIG_MEMCG_KMEM)
304 -----------------------------------------------
306 With the Kernel memory extension, the Memory Controller is able to limit
307 the amount of kernel memory used by the system. Kernel memory is fundamentally
308 different than user memory, since it can't be swapped out, which makes it
309 possible to DoS the system by consuming too much of this precious resource.
311 Kernel memory accounting is enabled for all memory cgroups by default. But
312 it can be disabled system-wide by passing cgroup.memory=nokmem to the kernel
313 at boot time. In this case, kernel memory will not be accounted at all.
315 Kernel memory limits are not imposed for the root cgroup. Usage for the root
316 cgroup may or may not be accounted. The memory used is accumulated into
317 memory.kmem.usage_in_bytes, or in a separate counter when it makes sense.
323 Currently no soft limit is implemented for kernel memory. It is future work
324 to trigger slab reclaim when those limits are reached.
326 2.7.1 Current Kernel Memory resources accounted
327 -----------------------------------------------
331 kernel memory, we prevent new processes from being created when the kernel
332 memory usage is too high.
339 belong to the same memcg. This only fails to hold when a task is migrated to a
342 sockets memory pressure:
343 some sockets protocols have memory pressure
344 thresholds. The Memory Controller allows them to be controlled individually
347 tcp memory pressure:
348 sockets memory pressure for the tcp protocol.
351 ----------------------
353 Because the "kmem" counter is fed to the main user counter, kernel memory can
354 never be limited completely independently of user memory. Say "U" is the user
360 accounting. Kernel memory is completely ignored.
363 Kernel memory is a subset of the user memory. This setup is useful in
364 deployments where the total amount of memory per-cgroup is overcommited.
365 Overcommiting kernel memory limits is definitely not recommended, since the
366 box can still run out of non-reclaimable memory.
368 never greater than the total memory, and freely set U at the cost of his
372 In the current implementation, memory reclaim will NOT be
377 Since kmem charges will also be fed to the user counter and reclaim will be
378 triggered for the cgroup for both kinds of memory. This setup gives the
379 admin a unified view of memory, and it is also useful for people who just
380 want to track kernel memory usage.
386 ------------------
390 c. Enable CONFIG_MEMCG_SWAP (to use swap extension)
391 d. Enable CONFIG_MEMCG_KMEM (to use kmem extension)
394 -------------------------------------------------------------------
398 # mount -t tmpfs none /sys/fs/cgroup
399 # mkdir /sys/fs/cgroup/memory
400 # mount -t cgroup none /sys/fs/cgroup/memory -o memory
404 # mkdir /sys/fs/cgroup/memory/0
405 # echo $$ > /sys/fs/cgroup/memory/0/tasks
407 Since now we're in the 0 cgroup, we can alter the memory limit::
409 # echo 4M > /sys/fs/cgroup/memory/0/memory.limit_in_bytes
412 We can use a suffix (k, K, m, M, g or G) to indicate values in kilo,
417 We can write "-1" to reset the ``*.limit_in_bytes(unlimited)``.
424 # cat /sys/fs/cgroup/memory/0/memory.limit_in_bytes
429 # cat /sys/fs/cgroup/memory/0/memory.usage_in_bytes
432 A successful write to this file does not guarantee a successful setting of
433 this limit to the value written into the file. This can be due to a
434 number of factors, such as rounding up to page boundaries or the total
435 availability of memory on the system. The user is required to re-read
436 this file after a write to guarantee the value committed by the kernel::
438 # echo 1 > memory.limit_in_bytes
439 # cat memory.limit_in_bytes
442 The memory.failcnt field gives the number of times that the cgroup limit was
445 The memory.stat file gives accounting information. Now, the number of
453 Performance test is also important. To see pure memory controller's overhead,
457 Page-fault scalability is also important. At measuring parallel
458 page fault test, multi-process test may be better than multi-thread
462 Trying usual test under memory controller is always helpful.
465 -------------------
470 1. The cgroup limit is too low (just too low to do anything useful)
471 2. The user is using anonymous memory and swap is turned off or too low
476 To know what happens, disabling OOM_Kill as per "10. OOM Control" (below) and
480 ------------------
482 When a task migrates from one cgroup to another, its charge is not
484 remain charged to it, the charge is dropped when the page is freed or
491 ---------------------
498 We move the stats to root (if use_hierarchy==0) or parent (if
512 ---------------
513 memory.force_empty interface is provided to make cgroup's memory usage empty.
514 When writing anything to this::
516 # echo 0 > memory.force_empty
521 Though rmdir() offlines memcg, but the memcg may still stay there due to
522 charged file caches. Some out-of-use page caches may keep charged until
523 memory pressure happens. If you want to avoid that, force_empty will be useful.
525 Also, note that when memory.kmem.limit_in_bytes is set the charges due to
528 memory.kmem.usage_in_bytes == memory.usage_in_bytes.
533 -------------
535 memory.stat file includes following statistics
537 per-memory cgroup local status
541 cache # of bytes of page cache memory.
542 rss # of bytes of anonymous and swap cache memory (includes
546 pgpgin # of charging events to the memory cgroup. The charging
548 anon page(RSS) or cache page(Page Cache) to the cgroup.
549 pgpgout # of uncharging events to the memory cgroup. The uncharging
552 dirty # of bytes that are waiting to get written back to the disk.
553 writeback # of bytes of file/anon cache that are queued for syncing to
555 inactive_anon # of bytes of anonymous and swap cache memory on inactive
557 active_anon # of bytes of anonymous and swap cache memory on active
559 inactive_file # of bytes of file-backed memory on inactive LRU list.
560 active_file # of bytes of file-backed memory on active LRU list.
561 unevictable # of bytes of memory that cannot be reclaimed (mlocked etc).
564 status considering hierarchy (see memory.use_hierarchy settings)
568 hierarchical_memory_limit # of bytes of memory limit with regard to hierarchy
569 under which the memory cgroup is
570 hierarchical_memsw_limit # of bytes of memory+swap limit with regard to
571 hierarchy under which memory cgroup is.
574 addition to the cgroup's own value includes the
591 recent_scanned means recent # of scans to LRU.
595 Only anonymous and swap cache memory is listed as part of 'rss' stat.
597 amount of physical memory used by the cgroup.
602 mapped_file is accounted only when the memory cgroup is owner of page
606 --------------
609 in the root cgroup corresponds to the global swappiness setting.
613 there is a swap storage available. This might lead to memcg OOM killer
614 if there are no file pages to reclaim.
617 -----------
619 A memory cgroup provides memory.failcnt and memory.memsw.failcnt files.
621 hit its limit. When a memory cgroup hits a limit, failcnt increases and
622 memory under it will be reclaimed.
624 You can reset failcnt by writing 0 to failcnt file::
626 # echo 0 > .../memory.failcnt
629 ------------------
631 For efficiency, as other kernel components, memory cgroup uses some optimization
632 to avoid unnecessary cacheline false sharing. usage_in_bytes is affected by the
633 method and doesn't show 'exact' value of memory (and swap) usage, it's a fuzz
635 If you want to know more exact memory usage, you should use RSS+CACHE(+SWAP)
636 value in memory.stat(see 5.2).
639 -------------
641 This is similar to numa_maps but operates on a per-memcg basis. This is
643 an memcg since the pages are allowed to be allocated from any physical
648 per-node page counts including "hierarchical_<counter>" which sums up all
649 hierarchical children's values in addition to the memcg's own value.
651 The output format of memory.numa_stat is::
664 The memory controller supports a deep hierarchy and hierarchical accounting.
677 In the diagram above, with hierarchical accounting enabled, all memory
678 usage of e, is accounted to its ancestors up until the root (i.e, c and root),
679 that has memory.use_hierarchy enabled. If one of the ancestors goes over its
684 ------------------------------------------------
686 A memory cgroup by default disables the hierarchy feature. Support
687 can be enabled by writing 1 to memory.use_hierarchy file of the root cgroup::
689 # echo 1 > memory.use_hierarchy
693 # echo 0 > memory.use_hierarchy
701 When panic_on_oom is set to "2", the whole system will panic in
707 Soft limits allow for greater sharing of memory. The idea behind soft limits
708 is to allow control groups to use as much of the memory as needed, provided
710 a. There is no memory contention
713 When the system detects memory contention or low memory, control groups
714 are pushed back to their soft limits. If the soft limit of each control
715 group is very high, they are pushed back as much as possible to make
716 sure that one control group does not starve the others of memory.
718 Please note that soft limits is a best-effort feature; it comes with
719 no guarantees, but it does its best to make sure that when memory is
720 heavily contended for, memory is allocated based on the soft limit
725 -------------
730 # echo 256M > memory.soft_limit_in_bytes
732 If we want to change this to 1G, we can at any time use::
734 # echo 1G > memory.soft_limit_in_bytes
738 reclaiming memory for balancing between memory cgroups
740 It is recommended to set the soft limit always below the hard limit,
747 is, uncharge task's pages from the old cgroup and charge them to the new cgroup.
752 -------------
755 writing to memory.move_charge_at_immigrate of the destination cgroup.
757 If you want to enable it::
759 # echo (some positive value) > memory.move_charge_at_immigrate
765 Charges are moved only when you move mm->owner, in other words,
769 try to make space by reclaiming memory. Task migration may fail if we
776 # echo 0 > memory.move_charge_at_immigrate
779 --------------------------------------
783 a page or a swap can be moved only when it is charged to the task's current
784 (old) memory cgroup.
786 +---+--------------------------------------------------------------------------+
790 | | You must enable Swap Extension (see 2.4) to enable move of swap charges. |
791 +---+--------------------------------------------------------------------------+
792 | 1 | A charge of file pages (normal file, tmpfs file (e.g. ipc shared memory) |
798 | | page_mapcount(page) > 1). You must enable Swap Extension (see 2.4) to |
800 +---+--------------------------------------------------------------------------+
803 --------
805 - All of moving charge operations are done under cgroup_mutex. It's not good
806 behavior to hold the mutex too long, so we may need some trick.
808 9. Memory thresholds
811 Memory cgroup implements memory thresholds using the cgroups notification
812 API (see cgroups.txt). It allows to register multiple memory and memsw
815 To register a threshold, an application must:
817 - create an eventfd using eventfd(2);
818 - open memory.usage_in_bytes or memory.memsw.usage_in_bytes;
819 - write string like "<event_fd> <fd of memory.usage_in_bytes> <threshold>" to
822 Application will be notified through eventfd when memory usage crosses
825 It's applicable for root and non-root cgroup.
830 memory.oom_control file is for OOM notification and other controls.
832 Memory cgroup implements OOM notifier using the cgroup notification
833 API (See cgroups.txt). It allows to register multiple OOM notification
836 To register a notifier, an application must:
838 - create an eventfd using eventfd(2)
839 - open memory.oom_control file
840 - write string like "<event_fd> <fd of memory.oom_control>" to
846 You can disable the OOM-killer by writing "1" to memory.oom_control file, as:
848 #echo 1 > memory.oom_control
850 If OOM-killer is disabled, tasks under cgroup will hang/sleep
851 in memory cgroup's OOM-waitqueue when they request accountable memory.
853 For running them, you have to relax the memory cgroup's OOM status by
857 To reduce usage,
860 * move some tasks to other group with account migration.
867 - oom_kill_disable 0 or 1
868 (if 1, oom-killer is disabled)
869 - under_oom 0 or 1
870 (if 1, the memory cgroup is under OOM, tasks may be stopped.)
872 11. Memory Pressure
875 The pressure level notifications can be used to monitor the memory
877 different strategies of managing their memory resources. The pressure
880 The "low" level means that the system is reclaiming memory for new
886 The "medium" level means that the system is experiencing medium memory
888 etc. Upon this event applications may decide to further analyze
889 vmstat/zoneinfo/memcg or internal memory usage statistics and free any
890 resources that can be easily reconstructed or re-read from a disk.
893 about to out of memory (OOM) or even the in-kernel OOM killer is on its
894 way to trigger. Applications should do whatever they can to help the
895 system. It might be too late to consult with vmstat or any other
896 statistics, so it's advisable to take an immediate action.
899 events are not pass-through. For example, you have three cgroups: A->B->C. Now
902 notification, i.e. groups A and B will not receive it. This is done to avoid
904 especially bad if we are low on memory or thrashing. Group B, will receive
909 - "default": this is the default behavior specified above. This mode is the
913 - "hierarchy": events always propagate up to the root, similar to the default
916 example, groups A, B, and C will receive notification of memory pressure.
918 - "local": events are pass-through, i.e. they only receive notifications when
919 memory pressure is experienced in the memcg for which the notification is
921 registered for "local" notification and the group experiences memory
927 specified by a comma-delimited string, i.e. "low,hierarchy" specifies
928 hierarchical, pass-through, notification for all ancestor memcgs. Notification
929 that is the default, non pass-through behavior, does not specify a mode.
930 "medium,local" specifies pass-through notification for the medium level.
932 The file memory.pressure_level is only used to setup an eventfd. To
935 - create an eventfd using eventfd(2);
936 - open memory.pressure_level;
937 - write string as "<event_fd> <fd of memory.pressure_level> <level[,mode]>"
938 to cgroup.event_control.
940 Application will be notified through eventfd when memory pressure is at
941 the specific level (or higher). Read/write operations to
942 memory.pressure_level are no implemented.
947 memory limit, sets up a notification in the cgroup and then makes child
950 # cd /sys/fs/cgroup/memory/
953 # cgroup_event_listener memory.pressure_level low,hierarchy &
954 # echo 8000000 > memory.limit_in_bytes
955 # echo 8000000 > memory.memsw.limit_in_bytes
959 (Expect a bunch of notifications, and eventually, the oom-killer will
965 1. Make per-cgroup scanner reclaim not-shared pages first
966 2. Teach controller to account for shared-pages
973 Overall, the memory controller has been a stable controller and has been
979 1. Singh, Balbir. RFC: Memory Controller, http://lwn.net/Articles/206697/
980 2. Singh, Balbir. Memory Controller (RSS Control),
995 10. Singh, Balbir. Memory controller v6 test results,
997 11. Singh, Balbir. Memory controller introduction (v6),
999 12. Corbet, Jonathan, Controlling memory use in cgroups,