1 ===============================
2 Documentation for /proc/sys/vm/
3 ===============================
4
5 kernel version 2.6.29
6
7 Copyright (c) 1998, 1999, Rik van Riel <riel@n
8
9 Copyright (c) 2008 Peter W. Morreale <p
10
11 For general info and legal blurb, please look
12
13 ----------------------------------------------
14
15 This file contains the documentation for the s
16 /proc/sys/vm and is valid for Linux kernel ver
17
18 The files in this directory can be used to tun
19 of the virtual memory (VM) subsystem of the Li
20 the writeout of dirty data to disk.
21
22 Default values and initialization routines for
23 files can be found in mm/swap.c.
24
25 Currently, these files are in /proc/sys/vm:
26
27 - admin_reserve_kbytes
28 - compact_memory
29 - compaction_proactiveness
30 - compact_unevictable_allowed
31 - dirty_background_bytes
32 - dirty_background_ratio
33 - dirty_bytes
34 - dirty_expire_centisecs
35 - dirty_ratio
36 - dirtytime_expire_seconds
37 - dirty_writeback_centisecs
38 - drop_caches
39 - enable_soft_offline
40 - extfrag_threshold
41 - highmem_is_dirtyable
42 - hugetlb_shm_group
43 - laptop_mode
44 - legacy_va_layout
45 - lowmem_reserve_ratio
46 - max_map_count
47 - mem_profiling (only if CONFIG_MEM_AL
48 - memory_failure_early_kill
49 - memory_failure_recovery
50 - min_free_kbytes
51 - min_slab_ratio
52 - min_unmapped_ratio
53 - mmap_min_addr
54 - mmap_rnd_bits
55 - mmap_rnd_compat_bits
56 - nr_hugepages
57 - nr_hugepages_mempolicy
58 - nr_overcommit_hugepages
59 - nr_trim_pages (only if CONFIG_MMU=n)
60 - numa_zonelist_order
61 - oom_dump_tasks
62 - oom_kill_allocating_task
63 - overcommit_kbytes
64 - overcommit_memory
65 - overcommit_ratio
66 - page-cluster
67 - page_lock_unfairness
68 - panic_on_oom
69 - percpu_pagelist_high_fraction
70 - stat_interval
71 - stat_refresh
72 - numa_stat
73 - swappiness
74 - unprivileged_userfaultfd
75 - user_reserve_kbytes
76 - vfs_cache_pressure
77 - watermark_boost_factor
78 - watermark_scale_factor
79 - zone_reclaim_mode
80
81
82 admin_reserve_kbytes
83 ====================
84
85 The amount of free memory in the system that s
86 with the capability cap_sys_admin.
87
88 admin_reserve_kbytes defaults to min(3% of fre
89
90 That should provide enough for the admin to lo
91 if necessary, under the default overcommit 'gu
92
93 Systems running under overcommit 'never' shoul
94 for the full Virtual Memory Size of programs u
95 root may not be able to log in to recover the
96
97 How do you calculate a minimum useful reserve?
98
99 sshd or login + bash (or some other shell) + t
100
101 For overcommit 'guess', we can sum resident se
102 On x86_64 this is about 8MB.
103
104 For overcommit 'never', we can take the max of
105 and add the sum of their RSS.
106 On x86_64 this is about 128MB.
107
108 Changing this takes effect whenever an applica
109
110
111 compact_memory
112 ==============
113
114 Available only when CONFIG_COMPACTION is set.
115 all zones are compacted such that free memory
116 blocks where possible. This can be important f
117 huge pages although processes will also direct
118
119 compaction_proactiveness
120 ========================
121
122 This tunable takes a value in the range [0, 10
123 20. This tunable determines how aggressively c
124 background. Write of a non zero value to this
125 trigger the proactive compaction. Setting it t
126
127 Note that compaction has a non-trivial system-
128 belonging to different processes are moved aro
129 to latency spikes in unsuspecting applications
130 various heuristics to avoid wasting CPU cycles
131 proactive compaction is not being effective.
132
133 Be careful when setting it to extreme values l
134 cause excessive background compaction activity
135
136 compact_unevictable_allowed
137 ===========================
138
139 Available only when CONFIG_COMPACTION is set.
140 allowed to examine the unevictable lru (mlocke
141 This should be used on systems where stalls fo
142 acceptable trade for large contiguous free mem
143 compaction from moving pages that are unevicta
144 On CONFIG_PREEMPT_RT the default value is 0 in
145 to compaction, which would block the task from
146 is resolved.
147
148
149 dirty_background_bytes
150 ======================
151
152 Contains the amount of dirty memory at which t
153 flusher threads will start writeback.
154
155 Note:
156 dirty_background_bytes is the counterpart of
157 one of them may be specified at a time. When
158 immediately taken into account to evaluate t
159 other appears as 0 when read.
160
161
162 dirty_background_ratio
163 ======================
164
165 Contains, as a percentage of total available m
166 and reclaimable pages, the number of pages at
167 flusher threads will start writing out dirty d
168
169 The total available memory is not equal to tot
170
171
172 dirty_bytes
173 ===========
174
175 Contains the amount of dirty memory at which a
176 will itself start writeback.
177
178 Note: dirty_bytes is the counterpart of dirty_
179 specified at a time. When one sysctl is writte
180 account to evaluate the dirty memory limits an
181 read.
182
183 Note: the minimum value allowed for dirty_byte
184 value lower than this limit will be ignored an
185 retained.
186
187
188 dirty_expire_centisecs
189 ======================
190
191 This tunable is used to define when dirty data
192 for writeout by the kernel flusher threads. I
193 of a second. Data which has been dirty in-mem
194 interval will be written out next time a flush
195
196
197 dirty_ratio
198 ===========
199
200 Contains, as a percentage of total available m
201 and reclaimable pages, the number of pages at
202 generating disk writes will itself start writi
203
204 The total available memory is not equal to tot
205
206
207 dirtytime_expire_seconds
208 ========================
209
210 When a lazytime inode is constantly having its
211 an updated timestamp will never get chance to
212 only thing that has happened on the file syste
213 by an atime update, a worker will be scheduled
214 eventually gets pushed out to disk. This tuna
215 inode is old enough to be eligible for writeba
216 And, it is also used as the interval to wakeup
217
218
219 dirty_writeback_centisecs
220 =========================
221
222 The kernel flusher threads will periodically w
223 out to disk. This tunable expresses the inter
224 100'ths of a second.
225
226 Setting this to zero disables periodic writeba
227
228
229 drop_caches
230 ===========
231
232 Writing to this will cause the kernel to drop
233 reclaimable slab objects like dentries and ino
234 memory becomes free.
235
236 To free pagecache::
237
238 echo 1 > /proc/sys/vm/drop_caches
239
240 To free reclaimable slab objects (includes den
241
242 echo 2 > /proc/sys/vm/drop_caches
243
244 To free slab objects and pagecache::
245
246 echo 3 > /proc/sys/vm/drop_caches
247
248 This is a non-destructive operation and will n
249 To increase the number of objects freed by thi
250 `sync` prior to writing to /proc/sys/vm/drop_c
251 number of dirty objects on the system and crea
252 dropped.
253
254 This file is not a means to control the growth
255 (inodes, dentries, pagecache, etc...) These o
256 reclaimed by the kernel when memory is needed
257
258 Use of this file can cause performance problem
259 objects, it may cost a significant amount of I
260 dropped objects, especially if they were under
261 use outside of a testing or debugging environm
262
263 You may see informational messages in your ker
264 used::
265
266 cat (1234): drop_caches: 3
267
268 These are informational only. They do not mea
269 with your system. To disable them, echo 4 (bi
270
271 enable_soft_offline
272 ===================
273 Correctable memory errors are very common on s
274 solution for memory pages having (excessive) c
275
276 For different types of page, soft-offline has
277
278 - For a raw error page, soft-offline migrates
279 a new raw page.
280
281 - For a page that is part of a transparent hug
282 transparent hugepage into raw pages, then mi
283 As a result, user is transparently backed by
284 memory access performance.
285
286 - For a page that is part of a HugeTLB hugepag
287 the entire HugeTLB hugepage, during which a
288 as migration target. Then the original huge
289 pages without compensation, reducing the cap
290
291 It is user's call to choose between reliabilit
292 physical memory) vs performance / capacity imp
293 HugeTLB cases.
294
295 For all architectures, enable_soft_offline con
296 memory pages. When set to 1, kernel attempts
297 whenever it thinks needed. When set to 0, ker
298 the request to soft offline the pages. Its de
299
300 It is worth mentioning that after setting enab
301 following requests to soft offline pages will
302
303 - Request to soft offline pages from RAS Corre
304
305 - On ARM, the request to soft offline pages fr
306
307 - On PARISC, the request to soft offline pages
308
309 extfrag_threshold
310 =================
311
312 This parameter affects whether the kernel will
313 reclaim to satisfy a high-order allocation. Th
314 debugfs shows what the fragmentation index for
315 the system. Values tending towards 0 imply all
316 of memory, values towards 1000 imply failures
317 implies that the allocation will succeed as lo
318
319 The kernel will not compact memory in a zone i
320 fragmentation index is <= extfrag_threshold. T
321
322
323 highmem_is_dirtyable
324 ====================
325
326 Available only for systems with CONFIG_HIGHMEM
327
328 This parameter controls whether the high memor
329 writers throttling. This is not the case by d
330 only the amount of memory directly visible/usa
331 be dirtied. As a result, on systems with a lar
332 lowmem basically depleted writers might be thr
333 streaming writes can get very slow.
334
335 Changing the value to non zero would allow mor
336 and thus allow writers to write more data whic
337 storage more effectively. Note this also comes
338 OOM killer because some writers (e.g. direct b
339 only use the low memory and they can fill it u
340 any throttling.
341
342
343 hugetlb_shm_group
344 =================
345
346 hugetlb_shm_group contains group id that is al
347 shared memory segment using hugetlb page.
348
349
350 laptop_mode
351 ===========
352
353 laptop_mode is a knob that controls "laptop mo
354 controlled by this knob are discussed in Docum
355
356
357 legacy_va_layout
358 ================
359
360 If non-zero, this sysctl disables the new 32-b
361 will use the legacy (2.4) layout for all proce
362
363
364 lowmem_reserve_ratio
365 ====================
366
367 For some specialised workloads on highmem mach
368 the kernel to allow process memory to be alloc
369 zone. This is because that memory could then
370 system call, or by unavailability of swapspace
371
372 And on large highmem machines this lack of rec
373 can be fatal.
374
375 So the Linux page allocator has a mechanism wh
376 which *could* use highmem from using too much
377 a certain amount of lowmem is defended from th
378 captured into pinned user memory.
379
380 (The same argument applies to the old 16 megab
381 mechanism will also defend that region from al
382 highmem or lowmem).
383
384 The `lowmem_reserve_ratio` tunable determines
385 in defending these lower zones.
386
387 If you have a machine which uses highmem or IS
388 applications are using mlock(), or if you are
389 you probably should change the lowmem_reserve_
390
391 The lowmem_reserve_ratio is an array. You can
392
393 % cat /proc/sys/vm/lowmem_reserve_rati
394 256 256 32
395
396 But, these values are not used directly. The k
397 pages for each zones from them. These are show
398 in /proc/zoneinfo like the following. (This is
399 Each zone has an array of protection pages lik
400
401 Node 0, zone DMA
402 pages free 1355
403 min 3
404 low 3
405 high 4
406 :
407 :
408 numa_other 0
409 protection: (0, 2004, 2004, 2004)
410 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
411 pagesets
412 cpu: 0 pcp: 0
413 :
414
415 These protections are added to score to judge
416 for page allocation or should be reclaimed.
417
418 In this example, if normal pages (index=2) are
419 watermark[WMARK_HIGH] is used for watermark, t
420 not be used because pages_free(1355) is smalle
421 (4 + 2004 = 2008). If this protection value is
422 normal page requirement. If requirement is DMA
423 (=0) is used.
424
425 zone[i]'s protection[j] is calculated by follo
426
427 (i < j):
428 zone[i]->protection[j]
429 = (total sums of managed_pages from zone[i
430 / lowmem_reserve_ratio[i];
431 (i = j):
432 (should not be protected. = 0;
433 (i > j):
434 (not necessary, but looks 0)
435
436 The default values of lowmem_reserve_ratio[i]
437
438 === ====================================
439 256 (if zone[i] means DMA or DMA32 zone)
440 32 (others)
441 === ====================================
442
443 As above expression, they are reciprocal numbe
444 256 means 1/256. # of protection pages becomes
445 pages of higher zones on the node.
446
447 If you would like to protect more pages, small
448 The minimum value is 1 (1/1 -> 100%). The valu
449 disables protection of the pages.
450
451
452 max_map_count:
453 ==============
454
455 This file contains the maximum number of memor
456 may have. Memory map areas are used as a side-
457 malloc, directly by mmap, mprotect, and madvis
458 shared libraries.
459
460 While most applications need less than a thous
461 programs, particularly malloc debuggers, may c
462 e.g., up to one or two maps per allocation.
463
464 The default value is 65530.
465
466
467 mem_profiling
468 ==============
469
470 Enable memory profiling (when CONFIG_MEM_ALLOC
471
472 1: Enable memory profiling.
473
474 0: Disable memory profiling.
475
476 Enabling memory profiling introduces a small p
477 memory allocations.
478
479 The default value depends on CONFIG_MEM_ALLOC_
480
481
482 memory_failure_early_kill:
483 ==========================
484
485 Control how to kill processes when uncorrected
486 a 2bit error in a memory module) is detected i
487 that cannot be handled by the kernel. In some
488 still having a valid copy on disk) the kernel
489 transparently without affecting any applicatio
490 no other up-to-date copy of the data it will k
491 corruptions from propagating.
492
493 1: Kill all processes that have the corrupted
494 as soon as the corruption is detected. Note t
495 for a few types of pages, like kernel internal
496 the swap cache, but works for the majority of
497
498 0: Only unmap the corrupted page from all proc
499 who tries to access it.
500
501 The kill is done using a catchable SIGBUS with
502 handle this if they want to.
503
504 This is only active on architectures/platforms
505 check handling and depends on the hardware cap
506
507 Applications can override this setting individ
508
509
510 memory_failure_recovery
511 =======================
512
513 Enable memory failure recovery (when supported
514
515 1: Attempt recovery.
516
517 0: Always panic on a memory failure.
518
519
520 min_free_kbytes
521 ===============
522
523 This is used to force the Linux VM to keep a m
524 of kilobytes free. The VM uses this number to
525 watermark[WMARK_MIN] value for each lowmem zon
526 Each lowmem zone gets a number of reserved fre
527 proportionally on its size.
528
529 Some minimal amount of memory is needed to sat
530 allocations; if you set this to lower than 102
531 become subtly broken, and prone to deadlock un
532
533 Setting this too high will OOM your machine in
534
535
536 min_slab_ratio
537 ==============
538
539 This is available only on NUMA kernels.
540
541 A percentage of the total pages in each zone.
542 (fallback from the local zone occurs) slabs wi
543 than this percentage of pages in a zone are re
544 This insures that the slab growth stays under
545 systems that rarely perform global reclaim.
546
547 The default is 5 percent.
548
549 Note that slab reclaim is triggered in a per z
550 The process of reclaiming slab memory is curre
551 and may not be fast.
552
553
554 min_unmapped_ratio
555 ==================
556
557 This is available only on NUMA kernels.
558
559 This is a percentage of the total pages in eac
560 only occur if more than this percentage of pag
561 zone_reclaim_mode allows to be reclaimed.
562
563 If zone_reclaim_mode has the value 4 OR'd, the
564 against all file-backed unmapped pages includi
565 files. Otherwise, only unmapped pages backed b
566 files and similar are considered.
567
568 The default is 1 percent.
569
570
571 mmap_min_addr
572 =============
573
574 This file indicates the amount of address spac
575 be restricted from mmapping. Since kernel nul
576 accidentally operate based on the information
577 of memory userspace processes should not be al
578 default this value is set to 0 and no protecti
579 security module. Setting this value to someth
580 vast majority of applications to work correctl
581 against future potential kernel bugs.
582
583
584 mmap_rnd_bits
585 =============
586
587 This value can be used to select the number of
588 determine the random offset to the base addres
589 resulting from mmap allocations on architectur
590 tuning address space randomization. This valu
591 by the architecture's minimum and maximum supp
592
593 This value can be changed after boot using the
594 /proc/sys/vm/mmap_rnd_bits tunable
595
596
597 mmap_rnd_compat_bits
598 ====================
599
600 This value can be used to select the number of
601 determine the random offset to the base addres
602 resulting from mmap allocations for applicatio
603 compatibility mode on architectures which supp
604 space randomization. This value will be bound
605 architecture's minimum and maximum supported v
606
607 This value can be changed after boot using the
608 /proc/sys/vm/mmap_rnd_compat_bits tunable
609
610
611 nr_hugepages
612 ============
613
614 Change the minimum size of the hugepage pool.
615
616 See Documentation/admin-guide/mm/hugetlbpage.r
617
618
619 hugetlb_optimize_vmemmap
620 ========================
621
622 This knob is not available when the size of 's
623 in include/linux/mm_types.h) is not power of t
624 result in this).
625
626 Enable (set to 1) or disable (set to 0) HugeTL
627
628 Once enabled, the vmemmap pages of subsequent
629 buddy allocator will be optimized (7 pages per
630 per 1GB HugeTLB page), whereas already allocat
631 optimized. When those optimized HugeTLB pages
632 to the buddy allocator, the vmemmap pages repr
633 remapped again and the vmemmap pages discarded
634 again. If your use case is that HugeTLB pages
635 never explicitly allocating HugeTLB pages with
636 'nr_overcommit_hugepages', those overcommitted
637 the fly') instead of being pulled from the Hug
638 benefits of memory savings against the more ov
639 of allocation or freeing HugeTLB pages between
640 allocator. Another behavior to note is that i
641 pressure, it could prevent the user from freei
642 pool to the buddy allocator since the allocati
643 failed, you have to retry later if your system
644
645 Once disabled, the vmemmap pages of subsequent
646 buddy allocator will not be optimized meaning
647 time from buddy allocator disappears, whereas
648 will not be affected. If you want to make sur
649 pages, you can set "nr_hugepages" to 0 first a
650 writing 0 to nr_hugepages will make any "in us
651 pages. So, those surplus pages are still opti
652 in use. You would need to wait for those surp
653 there are no optimized pages in the system.
654
655
656 nr_hugepages_mempolicy
657 ======================
658
659 Change the size of the hugepage pool at run-ti
660 set of NUMA nodes.
661
662 See Documentation/admin-guide/mm/hugetlbpage.r
663
664
665 nr_overcommit_hugepages
666 =======================
667
668 Change the maximum size of the hugepage pool.
669 nr_hugepages + nr_overcommit_hugepages.
670
671 See Documentation/admin-guide/mm/hugetlbpage.r
672
673
674 nr_trim_pages
675 =============
676
677 This is available only on NOMMU kernels.
678
679 This value adjusts the excess page trimming be
680 NOMMU mmap allocations.
681
682 A value of 0 disables trimming of allocations
683 trims excess pages aggressively. Any value >=
684 trimming of allocations is initiated.
685
686 The default value is 1.
687
688 See Documentation/admin-guide/mm/nommu-mmap.rs
689
690
691 numa_zonelist_order
692 ===================
693
694 This sysctl is only for NUMA and it is depreca
695 Node order will fail!
696
697 'where the memory is allocated from' is contro
698
699 (This documentation ignores ZONE_HIGHMEM/ZONE_
700 you may be able to read ZONE_DMA as ZONE_DMA32
701
702 In non-NUMA case, a zonelist for GFP_KERNEL is
703 ZONE_NORMAL -> ZONE_DMA
704 This means that a memory allocation request fo
705 get memory from ZONE_DMA only when ZONE_NORMAL
706
707 In NUMA case, you can think of following 2 typ
708 Assume 2 node NUMA and below is zonelist of No
709
710 (A) Node(0) ZONE_NORMAL -> Node(0) ZONE_DMA
711 (B) Node(0) ZONE_NORMAL -> Node(1) ZONE_NORM
712
713 Type(A) offers the best locality for processes
714 will be used before ZONE_NORMAL exhaustion. Th
715 out-of-memory(OOM) of ZONE_DMA because ZONE_DM
716
717 Type(B) cannot offer the best locality but is
718 the DMA zone.
719
720 Type(A) is called as "Node" order. Type (B) is
721
722 "Node order" orders the zonelists by node, the
723 Specify "[Nn]ode" for node order
724
725 "Zone Order" orders the zonelists by zone type
726 zone. Specify "[Zz]one" for zone order.
727
728 Specify "[Dd]efault" to request automatic conf
729
730 On 32-bit, the Normal zone needs to be preserv
731 by the kernel, so "zone" order will be selecte
732
733 On 64-bit, devices that require DMA32/DMA are
734 order will be selected.
735
736 Default order is recommended unless this is ca
737 system/application.
738
739
740 oom_dump_tasks
741 ==============
742
743 Enables a system-wide task dump (excluding ker
744 when the kernel performs an OOM-killing and in
745 pid, uid, tgid, vm size, rss, pgtables_bytes,
746 score, and name. This is helpful to determine
747 invoked, to identify the rogue task that cause
748 the OOM killer chose the task it did to kill.
749
750 If this is set to zero, this information is su
751 large systems with thousands of tasks it may n
752 the memory state information for each one. Su
753 be forced to incur a performance penalty in OO
754 information may not be desired.
755
756 If this is set to non-zero, this information i
757 OOM killer actually kills a memory-hogging tas
758
759 The default value is 1 (enabled).
760
761
762 oom_kill_allocating_task
763 ========================
764
765 This enables or disables killing the OOM-trigg
766 out-of-memory situations.
767
768 If this is set to zero, the OOM killer will sc
769 tasklist and select a task based on heuristics
770 selects a rogue memory-hogging task that frees
771 memory when killed.
772
773 If this is set to non-zero, the OOM killer sim
774 triggered the out-of-memory condition. This a
775 tasklist scan.
776
777 If panic_on_oom is selected, it takes preceden
778 is used in oom_kill_allocating_task.
779
780 The default value is 0.
781
782
783 overcommit_kbytes
784 =================
785
786 When overcommit_memory is set to 2, the commit
787 permitted to exceed swap plus this amount of p
788
789 Note: overcommit_kbytes is the counterpart of
790 of them may be specified at a time. Setting on
791 then appears as 0 when read).
792
793
794 overcommit_memory
795 =================
796
797 This value contains a flag that enables memory
798
799 When this flag is 0, the kernel compares the u
800 size against total memory plus swap and reject
801
802 When this flag is 1, the kernel pretends there
803 memory until it actually runs out.
804
805 When this flag is 2, the kernel uses a "never
806 policy that attempts to prevent any overcommit
807 Note that user_reserve_kbytes affects this pol
808
809 This feature can be very useful because there
810 programs that malloc() huge amounts of memory
811 and don't use much of it.
812
813 The default value is 0.
814
815 See Documentation/mm/overcommit-accounting.rst
816 mm/util.c::__vm_enough_memory() for more infor
817
818
819 overcommit_ratio
820 ================
821
822 When overcommit_memory is set to 2, the commit
823 space is not permitted to exceed swap plus thi
824 of physical RAM. See above.
825
826
827 page-cluster
828 ============
829
830 page-cluster controls the number of pages up t
831 are read in from swap in a single attempt. Thi
832 to page cache readahead.
833 The mentioned consecutivity is not in terms of
834 but consecutive on swap space - that means the
835
836 It is a logarithmic value - setting it to zero
837 it to 1 means "2 pages", setting it to 2 means
838 Zero disables swap readahead completely.
839
840 The default value is three (eight pages at a t
841 small benefits in tuning this to a different v
842 swap-intensive.
843
844 Lower values mean lower latencies for initial
845 extra faults and I/O delays for following faul
846 that consecutive pages readahead would have br
847
848
849 page_lock_unfairness
850 ====================
851
852 This value determines the number of times that
853 stolen from under a waiter. After the lock is
854 specified in this file (default is 5), the "fa
855 will apply, and the waiter will only be awaken
856
857 panic_on_oom
858 ============
859
860 This enables or disables panic on out-of-memor
861
862 If this is set to 0, the kernel will kill some
863 called oom_killer. Usually, oom_killer can ki
864 system will survive.
865
866 If this is set to 1, the kernel panics when ou
867 However, if a process limits using nodes by me
868 and those nodes become memory exhaustion statu
869 may be killed by oom-killer. No panic occurs i
870 Because other nodes' memory may be free. This
871 may be not fatal yet.
872
873 If this is set to 2, the kernel panics compuls
874 above-mentioned. Even oom happens under memory
875 system panics.
876
877 The default value is 0.
878
879 1 and 2 are for failover of clustering. Please
880 according to your policy of failover.
881
882 panic_on_oom=2+kdump gives you very strong too
883 why oom happens. You can get snapshot.
884
885
886 percpu_pagelist_high_fraction
887 =============================
888
889 This is the fraction of pages in each zone tha
890 per-cpu page lists. It is an upper boundary th
891 on the number of online CPUs. The min value fo
892 that we do not allow more than 1/8th of pages
893 on per-cpu page lists. This entry only changes
894 page lists. A user can specify a number like 1
895 each zone between per-cpu lists.
896
897 The batch value of each per-cpu page list rema
898 the value of the high fraction so allocation l
899
900 The initial value is zero. Kernel uses this va
901 mark based on the low watermark for the zone a
902 online CPUs. If the user writes '0' to this s
903 this default behavior.
904
905
906 stat_interval
907 =============
908
909 The time interval between which vm statistics
910 is 1 second.
911
912
913 stat_refresh
914 ============
915
916 Any read or write (by root only) flushes all t
917 into their global totals, for more accurate re
918 e.g. cat /proc/sys/vm/stat_refresh /proc/memin
919
920 As a side-effect, it also checks for negative
921 as 0) and "fails" with EINVAL if any are found
922 (At time of writing, a few stats are known som
923 with no ill effects: errors and warnings on th
924
925
926 numa_stat
927 =========
928
929 This interface allows runtime configuration of
930
931 When page allocation performance becomes a bot
932 some possible tool breakage and decreased numa
933 do::
934
935 echo 0 > /proc/sys/vm/numa_stat
936
937 When page allocation performance is not a bott
938 tooling to work, you can do::
939
940 echo 1 > /proc/sys/vm/numa_stat
941
942
943 swappiness
944 ==========
945
946 This control is used to define the rough relat
947 and filesystem paging, as a value between 0 an
948 assumes equal IO cost and will thus apply memo
949 cache and swap-backed pages equally; lower val
950 expensive swap IO, higher values indicates che
951
952 Keep in mind that filesystem IO patterns under
953 be more efficient than swap's random IO. An op
954 experimentation and will also be workload-depe
955
956 The default value is 60.
957
958 For in-memory swap, like zram or zswap, as wel
959 have swap on faster devices than the filesyste
960 be considered. For example, if the random IO a
961 is on average 2x faster than IO from the files
962 be 133 (x + 2x = 200, 2x = 133.33).
963
964 At 0, the kernel will not initiate swap until
965 file-backed pages is less than the high waterm
966
967
968 unprivileged_userfaultfd
969 ========================
970
971 This flag controls the mode in which unprivile
972 userfaultfd system calls. Set this to 0 to res
973 to handle page faults in user mode only. In th
974 SYS_CAP_PTRACE must pass UFFD_USER_MODE_ONLY i
975 succeed. Prohibiting use of userfaultfd for ha
976 mode may make certain vulnerabilities more dif
977
978 Set this to 1 to allow unprivileged users to u
979 calls without any restrictions.
980
981 The default value is 0.
982
983 Another way to control permissions for userfau
984 /dev/userfaultfd instead of userfaultfd(2). Se
985 Documentation/admin-guide/mm/userfaultfd.rst.
986
987 user_reserve_kbytes
988 ===================
989
990 When overcommit_memory is set to 2, "never ove
991 min(3% of current process size, user_reserve_k
992 This is intended to prevent a user from starti
993 process, such that they cannot recover (kill t
994
995 user_reserve_kbytes defaults to min(3% of the
996
997 If this is reduced to zero, then the user will
998 all free memory with a single process, minus a
999 Any subsequent attempts to execute a command w
1000 "fork: Cannot allocate memory".
1001
1002 Changing this takes effect whenever an applic
1003
1004
1005 vfs_cache_pressure
1006 ==================
1007
1008 This percentage value controls the tendency o
1009 the memory which is used for caching of direc
1010
1011 At the default value of vfs_cache_pressure=10
1012 reclaim dentries and inodes at a "fair" rate
1013 swapcache reclaim. Decreasing vfs_cache_pres
1014 to retain dentry and inode caches. When vfs_c
1015 never reclaim dentries and inodes due to memo
1016 lead to out-of-memory conditions. Increasing
1017 causes the kernel to prefer to reclaim dentri
1018
1019 Increasing vfs_cache_pressure significantly b
1020 performance impact. Reclaim code needs to tak
1021 directory and inode objects. With vfs_cache_p
1022 ten times more freeable objects than there ar
1023
1024
1025 watermark_boost_factor
1026 ======================
1027
1028 This factor controls the level of reclaim whe
1029 It defines the percentage of the high waterma
1030 reclaimed if pages of different mobility are
1031 The intent is that compaction has less work t
1032 increase the success rate of future high-orde
1033 allocations, THP and hugetlbfs pages.
1034
1035 To make it sensible with respect to the water
1036 parameter, the unit is in fractions of 10,000
1037 15,000 means that up to 150% of the high wate
1038 event of a pageblock being mixed due to fragm
1039 is determined by the number of fragmentation
1040 recent past. If this value is smaller than a
1041 worth of pages will be reclaimed (e.g. 2MB o
1042 of 0 will disable the feature.
1043
1044
1045 watermark_scale_factor
1046 ======================
1047
1048 This factor controls the aggressiveness of ks
1049 amount of memory left in a node/system before
1050 how much memory needs to be free before kswap
1051
1052 The unit is in fractions of 10,000. The defau
1053 distances between watermarks are 0.1% of the
1054 node/system. The maximum value is 3000, or 30
1055
1056 A high rate of threads entering direct reclai
1057 going to sleep prematurely (kswapd_low_wmark_
1058 that the number of free pages kswapd maintain
1059 too small for the allocation bursts occurring
1060 can then be used to tune kswapd aggressivenes
1061
1062
1063 zone_reclaim_mode
1064 =================
1065
1066 Zone_reclaim_mode allows someone to set more
1067 reclaim memory when a zone runs out of memory
1068 zone reclaim occurs. Allocations will be sati
1069 in the system.
1070
1071 This is value OR'ed together of
1072
1073 = ===================================
1074 1 Zone reclaim on
1075 2 Zone reclaim writes dirty pages out
1076 4 Zone reclaim swaps pages
1077 = ===================================
1078
1079 zone_reclaim_mode is disabled by default. Fo
1080 that benefit from having their data cached, z
1081 left disabled as the caching effect is likely
1082 data locality.
1083
1084 Consider enabling one or more zone_reclaim mo
1085 workload is partitioned such that each partit
1086 and that accessing remote memory would cause
1087 reduction. The page allocator will take addi
1088 allocating off node pages.
1089
1090 Allowing zone reclaim to write out pages stop
1091 writing large amounts of data from dirtying p
1092 reclaim will write out dirty pages if a zone
1093 throttle the process. This may decrease the p
1094 since it cannot use all of system memory to b
1095 anymore but it preserve the memory on other n
1096 of other processes running on other nodes wil
1097
1098 Allowing regular swap effectively restricts a
1099 node unless explicitly overridden by memory p
1100 configurations.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.