2 .\" Copyright (c) 2013 by Turbo Fredriksson <turbo@bayour.com>. All rights reserved.
3 .\" The contents of this file are subject to the terms of the Common Development
4 .\" and Distribution License (the "License"). You may not use this file except
5 .\" in compliance with the License. You can obtain a copy of the license at
6 .\" usr/src/OPENSOLARIS.LICENSE or http://www.opensolaris.org/os/licensing.
8 .\" See the License for the specific language governing permissions and
9 .\" limitations under the License. When distributing Covered Code, include this
10 .\" CDDL HEADER in each file and include the License file at
11 .\" usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this
12 .\" CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your
13 .\" own identifying information:
14 .\" Portions Copyright [yyyy] [name of copyright owner]
15 .TH ZFS-MODULE-PARAMETERS 5 "Nov 16, 2013"
17 zfs\-module\-parameters \- ZFS module parameters
21 Description of the different parameters to the ZFS module.
23 .SS "Module parameters"
30 \fBignore_hole_birth\fR (int)
33 When set, the hole_birth optimization will not be used, and all holes will
34 always be sent on zfs send. Useful if you suspect your datasets are affected
35 by a bug in hole_birth.
37 Use \fB1\fR for on (default) and \fB0\fR for off.
43 \fBl2arc_feed_again\fR (int)
46 Turbo L2ARC warm-up. When the L2ARC is cold the fill interval will be set as
49 Use \fB1\fR for yes (default) and \fB0\fR to disable.
55 \fBl2arc_feed_min_ms\fR (ulong)
58 Min feed interval in milliseconds. Requires \fBl2arc_feed_again=1\fR and only
59 applicable in related situations.
61 Default value: \fB200\fR.
67 \fBl2arc_feed_secs\fR (ulong)
70 Seconds between L2ARC writing
72 Default value: \fB1\fR.
78 \fBl2arc_headroom\fR (ulong)
81 How far through the ARC lists to search for L2ARC cacheable content, expressed
82 as a multiplier of \fBl2arc_write_max\fR
84 Default value: \fB2\fR.
90 \fBl2arc_headroom_boost\fR (ulong)
93 Scales \fBl2arc_headroom\fR by this percentage when L2ARC contents are being
94 successfully compressed before writing. A value of 100 disables this feature.
96 Default value: \fB200\fR.
102 \fBl2arc_nocompress\fR (int)
105 Skip compressing L2ARC buffers
107 Use \fB1\fR for yes and \fB0\fR for no (default).
113 \fBl2arc_noprefetch\fR (int)
116 Do not write buffers to L2ARC if they were prefetched but not used by
119 Use \fB1\fR for yes (default) and \fB0\fR to disable.
125 \fBl2arc_norw\fR (int)
128 No reads during writes
130 Use \fB1\fR for yes and \fB0\fR for no (default).
136 \fBl2arc_write_boost\fR (ulong)
139 Cold L2ARC devices will have \fBl2arc_write_max\fR increased by this amount
140 while they remain cold.
142 Default value: \fB8,388,608\fR.
148 \fBl2arc_write_max\fR (ulong)
151 Max write bytes per interval
153 Default value: \fB8,388,608\fR.
159 \fBmetaslab_aliquot\fR (ulong)
162 Metaslab granularity, in bytes. This is roughly similar to what would be
163 referred to as the "stripe size" in traditional RAID arrays. In normal
164 operation, ZFS will try to write this amount of data to a top-level vdev
165 before moving on to the next one.
167 Default value: \fB524,288\fR.
173 \fBmetaslab_bias_enabled\fR (int)
176 Enable metaslab group biasing based on its vdev's over- or under-utilization
177 relative to the pool.
179 Use \fB1\fR for yes (default) and \fB0\fR for no.
185 \fBzfs_metaslab_segment_weight_enabled\fR (int)
188 Enable/disable segment-based metaslab selection.
190 Use \fB1\fR for yes (default) and \fB0\fR for no.
196 \fBzfs_metaslab_switch_threshold\fR (int)
199 When using segment-based metaslab selection, continue allocating
200 from the active metaslab until \fBzfs_metaslab_switch_threshold\fR
201 worth of buckets have been exhausted.
203 Default value: \fB2\fR.
209 \fBmetaslab_debug_load\fR (int)
212 Load all metaslabs during pool import.
214 Use \fB1\fR for yes and \fB0\fR for no (default).
220 \fBmetaslab_debug_unload\fR (int)
223 Prevent metaslabs from being unloaded.
225 Use \fB1\fR for yes and \fB0\fR for no (default).
231 \fBmetaslab_fragmentation_factor_enabled\fR (int)
234 Enable use of the fragmentation metric in computing metaslab weights.
236 Use \fB1\fR for yes (default) and \fB0\fR for no.
242 \fBmetaslabs_per_vdev\fR (int)
245 When a vdev is added, it will be divided into approximately (but no more than) this number of metaslabs.
247 Default value: \fB200\fR.
253 \fBmetaslab_preload_enabled\fR (int)
256 Enable metaslab group preloading.
258 Use \fB1\fR for yes (default) and \fB0\fR for no.
264 \fBmetaslab_lba_weighting_enabled\fR (int)
267 Give more weight to metaslabs with lower LBAs, assuming they have
268 greater bandwidth as is typically the case on a modern constant
269 angular velocity disk drive.
271 Use \fB1\fR for yes (default) and \fB0\fR for no.
277 \fBspa_config_path\fR (charp)
282 Default value: \fB/etc/zfs/zpool.cache\fR.
288 \fBspa_asize_inflation\fR (int)
291 Multiplication factor used to estimate actual disk consumption from the
292 size of data being written. The default value is a worst case estimate,
293 but lower values may be valid for a given pool depending on its
294 configuration. Pool administrators who understand the factors involved
295 may wish to specify a more realistic inflation factor, particularly if
296 they operate close to quota or capacity limits.
298 Default value: \fB24\fR.
304 \fBspa_load_verify_data\fR (int)
307 Whether to traverse data blocks during an "extreme rewind" (\fB-X\fR)
308 import. Use 0 to disable and 1 to enable.
310 An extreme rewind import normally performs a full traversal of all
311 blocks in the pool for verification. If this parameter is set to 0,
312 the traversal skips non-metadata blocks. It can be toggled once the
313 import has started to stop or start the traversal of non-metadata blocks.
315 Default value: \fB1\fR.
321 \fBspa_load_verify_metadata\fR (int)
324 Whether to traverse blocks during an "extreme rewind" (\fB-X\fR)
325 pool import. Use 0 to disable and 1 to enable.
327 An extreme rewind import normally performs a full traversal of all
328 blocks in the pool for verification. If this parameter is set to 0,
329 the traversal is not performed. It can be toggled once the import has
330 started to stop or start the traversal.
332 Default value: \fB1\fR.
338 \fBspa_load_verify_maxinflight\fR (int)
341 Maximum concurrent I/Os during the traversal performed during an "extreme
342 rewind" (\fB-X\fR) pool import.
344 Default value: \fB10000\fR.
350 \fBspa_slop_shift\fR (int)
353 Normally, we don't allow the last 3.2% (1/(2^spa_slop_shift)) of space
354 in the pool to be consumed. This ensures that we don't run the pool
355 completely out of space, due to unaccounted changes (e.g. to the MOS).
356 It also limits the worst-case time to allocate space. If we have
357 less than this amount of free space, most ZPL operations (e.g. write,
358 create) will return ENOSPC.
360 Default value: \fB5\fR.
366 \fBzfetch_array_rd_sz\fR (ulong)
369 If prefetching is enabled, disable prefetching for reads larger than this size.
371 Default value: \fB1,048,576\fR.
377 \fBzfetch_max_distance\fR (uint)
380 Max bytes to prefetch per stream (default 8MB).
382 Default value: \fB8,388,608\fR.
388 \fBzfetch_max_streams\fR (uint)
391 Max number of streams per zfetch (prefetch streams per file).
393 Default value: \fB8\fR.
399 \fBzfetch_min_sec_reap\fR (uint)
402 Min time before an active prefetch stream can be reclaimed
404 Default value: \fB2\fR.
410 \fBzfs_arc_dnode_limit\fR (ulong)
413 When the number of bytes consumed by dnodes in the ARC exceeds this number of
414 bytes, try to unpin some of it in response to demand for non-metadata. This
415 value acts as a floor to the amount of dnode metadata, and defaults to 0 which
416 indicates that a percent which is based on \fBzfs_arc_dnode_limit_percent\fR of
417 the ARC meta buffers that may be used for dnodes.
419 See also \fBzfs_arc_meta_prune\fR which serves a similar purpose but is used
420 when the amount of metadata in the ARC exceeds \fBzfs_arc_meta_limit\fR rather
421 than in response to overall demand for non-metadata.
424 Default value: \fB0\fR.
430 \fBzfs_arc_dnode_limit_percent\fR (ulong)
433 Percentage that can be consumed by dnodes of ARC meta buffers.
435 See also \fBzfs_arc_dnode_limit\fR which serves a similar purpose but has a
436 higher priority if set to nonzero value.
438 Default value: \fB10\fR.
444 \fBzfs_arc_dnode_reduce_percent\fR (ulong)
447 Percentage of ARC dnodes to try to scan in response to demand for non-metadata
448 when the number of bytes consumed by dnodes exceeds \fBzfs_arc_dnode_limit\fR.
451 Default value: \fB10% of the number of dnodes in the ARC\fR.
457 \fBzfs_arc_average_blocksize\fR (int)
460 The ARC's buffer hash table is sized based on the assumption of an average
461 block size of \fBzfs_arc_average_blocksize\fR (default 8K). This works out
462 to roughly 1MB of hash table per 1GB of physical memory with 8-byte pointers.
463 For configurations with a known larger average block size this value can be
464 increased to reduce the memory footprint.
467 Default value: \fB8192\fR.
473 \fBzfs_arc_evict_batch_limit\fR (int)
476 Number ARC headers to evict per sub-list before proceeding to another sub-list.
477 This batch-style operation prevents entire sub-lists from being evicted at once
478 but comes at a cost of additional unlocking and locking.
480 Default value: \fB10\fR.
486 \fBzfs_arc_grow_retry\fR (int)
489 After a memory pressure event the ARC will wait this many seconds before trying
492 Default value: \fB5\fR.
498 \fBzfs_arc_lotsfree_percent\fR (int)
501 Throttle I/O when free system memory drops below this percentage of total
502 system memory. Setting this value to 0 will disable the throttle.
504 Default value: \fB10\fR.
510 \fBzfs_arc_max\fR (ulong)
513 Max arc size of ARC in bytes. If set to 0 then it will consume 1/2 of system
514 RAM. This value must be at least 67108864 (64 megabytes).
516 This value can be changed dynamically with some caveats. It cannot be set back
517 to 0 while running and reducing it below the current ARC size will not cause
518 the ARC to shrink without memory pressure to induce shrinking.
520 Default value: \fB0\fR.
526 \fBzfs_arc_meta_limit\fR (ulong)
529 The maximum allowed size in bytes that meta data buffers are allowed to
530 consume in the ARC. When this limit is reached meta data buffers will
531 be reclaimed even if the overall arc_c_max has not been reached. This
532 value defaults to 0 which indicates that a percent which is based on
533 \fBzfs_arc_meta_limit_percent\fR of the ARC may be used for meta data.
535 This value my be changed dynamically except that it cannot be set back to 0
536 for a specific percent of the ARC; it must be set to an explicit value.
538 Default value: \fB0\fR.
544 \fBzfs_arc_meta_limit_percent\fR (ulong)
547 Percentage of ARC buffers that can be used for meta data.
549 See also \fBzfs_arc_meta_limit\fR which serves a similar purpose but has a
550 higher priority if set to nonzero value.
553 Default value: \fB75\fR.
559 \fBzfs_arc_meta_min\fR (ulong)
562 The minimum allowed size in bytes that meta data buffers may consume in
563 the ARC. This value defaults to 0 which disables a floor on the amount
564 of the ARC devoted meta data.
566 Default value: \fB0\fR.
572 \fBzfs_arc_meta_prune\fR (int)
575 The number of dentries and inodes to be scanned looking for entries
576 which can be dropped. This may be required when the ARC reaches the
577 \fBzfs_arc_meta_limit\fR because dentries and inodes can pin buffers
578 in the ARC. Increasing this value will cause to dentry and inode caches
579 to be pruned more aggressively. Setting this value to 0 will disable
580 pruning the inode and dentry caches.
582 Default value: \fB10,000\fR.
588 \fBzfs_arc_meta_adjust_restarts\fR (ulong)
591 The number of restart passes to make while scanning the ARC attempting
592 the free buffers in order to stay below the \fBzfs_arc_meta_limit\fR.
593 This value should not need to be tuned but is available to facilitate
594 performance analysis.
596 Default value: \fB4096\fR.
602 \fBzfs_arc_min\fR (ulong)
607 Default value: \fB100\fR.
613 \fBzfs_arc_min_prefetch_lifespan\fR (int)
616 Minimum time prefetched blocks are locked in the ARC, specified in jiffies.
617 A value of 0 will default to 1 second.
619 Default value: \fB0\fR.
625 \fBzfs_multilist_num_sublists\fR (int)
628 To allow more fine-grained locking, each ARC state contains a series
629 of lists for both data and meta data objects. Locking is performed at
630 the level of these "sub-lists". This parameters controls the number of
631 sub-lists per ARC state, and also applies to other uses of the
632 multilist data structure.
634 Default value: \fB4\fR or the number of online CPUs, whichever is greater
640 \fBzfs_arc_overflow_shift\fR (int)
643 The ARC size is considered to be overflowing if it exceeds the current
644 ARC target size (arc_c) by a threshold determined by this parameter.
645 The threshold is calculated as a fraction of arc_c using the formula
646 "arc_c >> \fBzfs_arc_overflow_shift\fR".
648 The default value of 8 causes the ARC to be considered to be overflowing
649 if it exceeds the target size by 1/256th (0.3%) of the target size.
651 When the ARC is overflowing, new buffer allocations are stalled until
652 the reclaim thread catches up and the overflow condition no longer exists.
654 Default value: \fB8\fR.
661 \fBzfs_arc_p_min_shift\fR (int)
664 arc_c shift to calc min/max arc_p
666 Default value: \fB4\fR.
672 \fBzfs_arc_p_aggressive_disable\fR (int)
675 Disable aggressive arc_p growth
677 Use \fB1\fR for yes (default) and \fB0\fR to disable.
683 \fBzfs_arc_p_dampener_disable\fR (int)
686 Disable arc_p adapt dampener
688 Use \fB1\fR for yes (default) and \fB0\fR to disable.
694 \fBzfs_arc_shrink_shift\fR (int)
697 log2(fraction of arc to reclaim)
699 Default value: \fB5\fR.
705 \fBzfs_arc_sys_free\fR (ulong)
708 The target number of bytes the ARC should leave as free memory on the system.
709 Defaults to the larger of 1/64 of physical memory or 512K. Setting this
710 option to a non-zero value will override the default.
712 Default value: \fB0\fR.
718 \fBzfs_autoimport_disable\fR (int)
721 Disable pool import at module load by ignoring the cache file (typically \fB/etc/zfs/zpool.cache\fR).
723 Use \fB1\fR for yes (default) and \fB0\fR for no.
729 \fBzfs_dbgmsg_enable\fR (int)
732 Internally ZFS keeps a small log to facilitate debugging. By default the log
733 is disabled, to enable it set this option to 1. The contents of the log can
734 be accessed by reading the /proc/spl/kstat/zfs/dbgmsg file. Writing 0 to
735 this proc file clears the log.
737 Default value: \fB0\fR.
743 \fBzfs_dbgmsg_maxsize\fR (int)
746 The maximum size in bytes of the internal ZFS debug log.
748 Default value: \fB4M\fR.
754 \fBzfs_dbuf_state_index\fR (int)
757 This feature is currently unused. It is normally used for controlling what
758 reporting is available under /proc/spl/kstat/zfs.
760 Default value: \fB0\fR.
766 \fBzfs_deadman_enabled\fR (int)
769 When a pool sync operation takes longer than \fBzfs_deadman_synctime_ms\fR
770 milliseconds, a "slow spa_sync" message is logged to the debug log
771 (see \fBzfs_dbgmsg_enable\fR). If \fBzfs_deadman_enabled\fR is set,
772 all pending IO operations are also checked and if any haven't completed
773 within \fBzfs_deadman_synctime_ms\fR milliseconds, a "SLOW IO" message
774 is logged to the debug log and a "delay" system event with the details of
775 the hung IO is posted.
777 Use \fB1\fR (default) to enable the slow IO check and \fB0\fR to disable.
783 \fBzfs_deadman_checktime_ms\fR (int)
786 Once a pool sync operation has taken longer than
787 \fBzfs_deadman_synctime_ms\fR milliseconds, continue to check for slow
788 operations every \fBzfs_deadman_checktime_ms\fR milliseconds.
790 Default value: \fB5,000\fR.
796 \fBzfs_deadman_synctime_ms\fR (ulong)
799 Interval in milliseconds after which the deadman is triggered and also
800 the interval after which an IO operation is considered to be "hung"
801 if \fBzfs_deadman_enabled\fR is set.
803 See \fBzfs_deadman_enabled\fR.
805 Default value: \fB1,000,000\fR.
811 \fBzfs_dedup_prefetch\fR (int)
814 Enable prefetching dedup-ed blks
816 Use \fB1\fR for yes and \fB0\fR to disable (default).
822 \fBzfs_delay_min_dirty_percent\fR (int)
825 Start to delay each transaction once there is this amount of dirty data,
826 expressed as a percentage of \fBzfs_dirty_data_max\fR.
827 This value should be >= zfs_vdev_async_write_active_max_dirty_percent.
828 See the section "ZFS TRANSACTION DELAY".
830 Default value: \fB60\fR.
836 \fBzfs_delay_scale\fR (int)
839 This controls how quickly the transaction delay approaches infinity.
840 Larger values cause longer delays for a given amount of dirty data.
842 For the smoothest delay, this value should be about 1 billion divided
843 by the maximum number of operations per second. This will smoothly
844 handle between 10x and 1/10th this number.
846 See the section "ZFS TRANSACTION DELAY".
848 Note: \fBzfs_delay_scale\fR * \fBzfs_dirty_data_max\fR must be < 2^64.
850 Default value: \fB500,000\fR.
856 \fBzfs_delete_blocks\fR (ulong)
859 This is the used to define a large file for the purposes of delete. Files
860 containing more than \fBzfs_delete_blocks\fR will be deleted asynchronously
861 while smaller files are deleted synchronously. Decreasing this value will
862 reduce the time spent in an unlink(2) system call at the expense of a longer
863 delay before the freed space is available.
865 Default value: \fB20,480\fR.
871 \fBzfs_dirty_data_max\fR (int)
874 Determines the dirty space limit in bytes. Once this limit is exceeded, new
875 writes are halted until space frees up. This parameter takes precedence
876 over \fBzfs_dirty_data_max_percent\fR.
877 See the section "ZFS TRANSACTION DELAY".
879 Default value: 10 percent of all memory, capped at \fBzfs_dirty_data_max_max\fR.
885 \fBzfs_dirty_data_max_max\fR (int)
888 Maximum allowable value of \fBzfs_dirty_data_max\fR, expressed in bytes.
889 This limit is only enforced at module load time, and will be ignored if
890 \fBzfs_dirty_data_max\fR is later changed. This parameter takes
891 precedence over \fBzfs_dirty_data_max_max_percent\fR. See the section
892 "ZFS TRANSACTION DELAY".
894 Default value: 25% of physical RAM.
900 \fBzfs_dirty_data_max_max_percent\fR (int)
903 Maximum allowable value of \fBzfs_dirty_data_max\fR, expressed as a
904 percentage of physical RAM. This limit is only enforced at module load
905 time, and will be ignored if \fBzfs_dirty_data_max\fR is later changed.
906 The parameter \fBzfs_dirty_data_max_max\fR takes precedence over this
907 one. See the section "ZFS TRANSACTION DELAY".
909 Default value: \fB25\fR.
915 \fBzfs_dirty_data_max_percent\fR (int)
918 Determines the dirty space limit, expressed as a percentage of all
919 memory. Once this limit is exceeded, new writes are halted until space frees
920 up. The parameter \fBzfs_dirty_data_max\fR takes precedence over this
921 one. See the section "ZFS TRANSACTION DELAY".
923 Default value: 10%, subject to \fBzfs_dirty_data_max_max\fR.
929 \fBzfs_dirty_data_sync\fR (int)
932 Start syncing out a transaction group if there is at least this much dirty data.
934 Default value: \fB67,108,864\fR.
940 \fBzfs_fletcher_4_impl\fR (string)
943 Select a fletcher 4 implementation.
945 Supported selectors are: \fBfastest\fR, \fBscalar\fR, \fBsse2\fR, \fBssse3\fR,
946 \fBavx2\fR, \fBavx512f\fR, and \fBaarch64_neon\fR.
947 All of the selectors except \fBfastest\fR and \fBscalar\fR require instruction
948 set extensions to be available and will only appear if ZFS detects that they are
949 present at runtime. If multiple implementations of fletcher 4 are available,
950 the \fBfastest\fR will be chosen using a micro benchmark. Selecting \fBscalar\fR
951 results in the original, CPU based calculation, being used. Selecting any option
952 other than \fBfastest\fR and \fBscalar\fR results in vector instructions from
953 the respective CPU instruction set being used.
955 Default value: \fBfastest\fR.
961 \fBzfs_free_bpobj_enabled\fR (int)
964 Enable/disable the processing of the free_bpobj object.
966 Default value: \fB1\fR.
972 \fBzfs_free_max_blocks\fR (ulong)
975 Maximum number of blocks freed in a single txg.
977 Default value: \fB100,000\fR.
983 \fBzfs_vdev_async_read_max_active\fR (int)
986 Maximum asynchronous read I/Os active to each device.
987 See the section "ZFS I/O SCHEDULER".
989 Default value: \fB3\fR.
995 \fBzfs_vdev_async_read_min_active\fR (int)
998 Minimum asynchronous read I/Os active to each device.
999 See the section "ZFS I/O SCHEDULER".
1001 Default value: \fB1\fR.
1007 \fBzfs_vdev_async_write_active_max_dirty_percent\fR (int)
1010 When the pool has more than
1011 \fBzfs_vdev_async_write_active_max_dirty_percent\fR dirty data, use
1012 \fBzfs_vdev_async_write_max_active\fR to limit active async writes. If
1013 the dirty data is between min and max, the active I/O limit is linearly
1014 interpolated. See the section "ZFS I/O SCHEDULER".
1016 Default value: \fB60\fR.
1022 \fBzfs_vdev_async_write_active_min_dirty_percent\fR (int)
1025 When the pool has less than
1026 \fBzfs_vdev_async_write_active_min_dirty_percent\fR dirty data, use
1027 \fBzfs_vdev_async_write_min_active\fR to limit active async writes. If
1028 the dirty data is between min and max, the active I/O limit is linearly
1029 interpolated. See the section "ZFS I/O SCHEDULER".
1031 Default value: \fB30\fR.
1037 \fBzfs_vdev_async_write_max_active\fR (int)
1040 Maximum asynchronous write I/Os active to each device.
1041 See the section "ZFS I/O SCHEDULER".
1043 Default value: \fB10\fR.
1049 \fBzfs_vdev_async_write_min_active\fR (int)
1052 Minimum asynchronous write I/Os active to each device.
1053 See the section "ZFS I/O SCHEDULER".
1055 Lower values are associated with better latency on rotational media but poorer
1056 resilver performance. The default value of 2 was chosen as a compromise. A
1057 value of 3 has been shown to improve resilver performance further at a cost of
1058 further increasing latency.
1060 Default value: \fB2\fR.
1066 \fBzfs_vdev_max_active\fR (int)
1069 The maximum number of I/Os active to each device. Ideally, this will be >=
1070 the sum of each queue's max_active. It must be at least the sum of each
1071 queue's min_active. See the section "ZFS I/O SCHEDULER".
1073 Default value: \fB1,000\fR.
1079 \fBzfs_vdev_scrub_max_active\fR (int)
1082 Maximum scrub I/Os active to each device.
1083 See the section "ZFS I/O SCHEDULER".
1085 Default value: \fB2\fR.
1091 \fBzfs_vdev_scrub_min_active\fR (int)
1094 Minimum scrub I/Os active to each device.
1095 See the section "ZFS I/O SCHEDULER".
1097 Default value: \fB1\fR.
1103 \fBzfs_vdev_sync_read_max_active\fR (int)
1106 Maximum synchronous read I/Os active to each device.
1107 See the section "ZFS I/O SCHEDULER".
1109 Default value: \fB10\fR.
1115 \fBzfs_vdev_sync_read_min_active\fR (int)
1118 Minimum synchronous read I/Os active to each device.
1119 See the section "ZFS I/O SCHEDULER".
1121 Default value: \fB10\fR.
1127 \fBzfs_vdev_sync_write_max_active\fR (int)
1130 Maximum synchronous write I/Os active to each device.
1131 See the section "ZFS I/O SCHEDULER".
1133 Default value: \fB10\fR.
1139 \fBzfs_vdev_sync_write_min_active\fR (int)
1142 Minimum synchronous write I/Os active to each device.
1143 See the section "ZFS I/O SCHEDULER".
1145 Default value: \fB10\fR.
1151 \fBzfs_vdev_queue_depth_pct\fR (int)
1154 Maximum number of queued allocations per top-level vdev expressed as
1155 a percentage of \fBzfs_vdev_async_write_max_active\fR which allows the
1156 system to detect devices that are more capable of handling allocations
1157 and to allocate more blocks to those devices. It allows for dynamic
1158 allocation distribution when devices are imbalanced as fuller devices
1159 will tend to be slower than empty devices.
1161 See also \fBzio_dva_throttle_enabled\fR.
1163 Default value: \fB1000\fR.
1169 \fBzfs_disable_dup_eviction\fR (int)
1172 Disable duplicate buffer eviction
1174 Use \fB1\fR for yes and \fB0\fR for no (default).
1180 \fBzfs_expire_snapshot\fR (int)
1183 Seconds to expire .zfs/snapshot
1185 Default value: \fB300\fR.
1191 \fBzfs_admin_snapshot\fR (int)
1194 Allow the creation, removal, or renaming of entries in the .zfs/snapshot
1195 directory to cause the creation, destruction, or renaming of snapshots.
1196 When enabled this functionality works both locally and over NFS exports
1197 which have the 'no_root_squash' option set. This functionality is disabled
1200 Use \fB1\fR for yes and \fB0\fR for no (default).
1206 \fBzfs_flags\fR (int)
1209 Set additional debugging flags. The following flags may be bitwise-or'd
1221 Enable dprintf entries in the debug log.
1223 2 ZFS_DEBUG_DBUF_VERIFY *
1224 Enable extra dbuf verifications.
1226 4 ZFS_DEBUG_DNODE_VERIFY *
1227 Enable extra dnode verifications.
1229 8 ZFS_DEBUG_SNAPNAMES
1230 Enable snapshot name verification.
1233 Check for illegally modified ARC buffers.
1236 Enable spa_dbgmsg entries in the debug log.
1238 64 ZFS_DEBUG_ZIO_FREE
1239 Enable verification of block frees.
1241 128 ZFS_DEBUG_HISTOGRAM_VERIFY
1242 Enable extra spacemap histogram verifications.
1245 * Requires debug build.
1247 Default value: \fB0\fR.
1253 \fBzfs_free_leak_on_eio\fR (int)
1256 If destroy encounters an EIO while reading metadata (e.g. indirect
1257 blocks), space referenced by the missing metadata can not be freed.
1258 Normally this causes the background destroy to become "stalled", as
1259 it is unable to make forward progress. While in this stalled state,
1260 all remaining space to free from the error-encountering filesystem is
1261 "temporarily leaked". Set this flag to cause it to ignore the EIO,
1262 permanently leak the space from indirect blocks that can not be read,
1263 and continue to free everything else that it can.
1265 The default, "stalling" behavior is useful if the storage partially
1266 fails (i.e. some but not all i/os fail), and then later recovers. In
1267 this case, we will be able to continue pool operations while it is
1268 partially failed, and when it recovers, we can continue to free the
1269 space, with no leaks. However, note that this case is actually
1272 Typically pools either (a) fail completely (but perhaps temporarily,
1273 e.g. a top-level vdev going offline), or (b) have localized,
1274 permanent errors (e.g. disk returns the wrong data due to bit flip or
1275 firmware bug). In case (a), this setting does not matter because the
1276 pool will be suspended and the sync thread will not be able to make
1277 forward progress regardless. In case (b), because the error is
1278 permanent, the best we can do is leak the minimum amount of space,
1279 which is what setting this flag will do. Therefore, it is reasonable
1280 for this flag to normally be set, but we chose the more conservative
1281 approach of not setting it, so that there is no possibility of
1282 leaking space in the "partial temporary" failure case.
1284 Default value: \fB0\fR.
1290 \fBzfs_free_min_time_ms\fR (int)
1293 During a \fBzfs destroy\fR operation using \fBfeature@async_destroy\fR a minimum
1294 of this much time will be spent working on freeing blocks per txg.
1296 Default value: \fB1,000\fR.
1302 \fBzfs_immediate_write_sz\fR (long)
1305 Largest data block to write to zil. Larger blocks will be treated as if the
1306 dataset being written to had the property setting \fBlogbias=throughput\fR.
1308 Default value: \fB32,768\fR.
1314 \fBzfs_max_recordsize\fR (int)
1317 We currently support block sizes from 512 bytes to 16MB. The benefits of
1318 larger blocks, and thus larger IO, need to be weighed against the cost of
1319 COWing a giant block to modify one byte. Additionally, very large blocks
1320 can have an impact on i/o latency, and also potentially on the memory
1321 allocator. Therefore, we do not allow the recordsize to be set larger than
1322 zfs_max_recordsize (default 1MB). Larger blocks can be created by changing
1323 this tunable, and pools with larger blocks can always be imported and used,
1324 regardless of this setting.
1326 Default value: \fB1,048,576\fR.
1332 \fBzfs_mdcomp_disable\fR (int)
1335 Disable meta data compression
1337 Use \fB1\fR for yes and \fB0\fR for no (default).
1343 \fBzfs_metaslab_fragmentation_threshold\fR (int)
1346 Allow metaslabs to keep their active state as long as their fragmentation
1347 percentage is less than or equal to this value. An active metaslab that
1348 exceeds this threshold will no longer keep its active status allowing
1349 better metaslabs to be selected.
1351 Default value: \fB70\fR.
1357 \fBzfs_mg_fragmentation_threshold\fR (int)
1360 Metaslab groups are considered eligible for allocations if their
1361 fragmentation metric (measured as a percentage) is less than or equal to
1362 this value. If a metaslab group exceeds this threshold then it will be
1363 skipped unless all metaslab groups within the metaslab class have also
1364 crossed this threshold.
1366 Default value: \fB85\fR.
1372 \fBzfs_mg_noalloc_threshold\fR (int)
1375 Defines a threshold at which metaslab groups should be eligible for
1376 allocations. The value is expressed as a percentage of free space
1377 beyond which a metaslab group is always eligible for allocations.
1378 If a metaslab group's free space is less than or equal to the
1379 threshold, the allocator will avoid allocating to that group
1380 unless all groups in the pool have reached the threshold. Once all
1381 groups have reached the threshold, all groups are allowed to accept
1382 allocations. The default value of 0 disables the feature and causes
1383 all metaslab groups to be eligible for allocations.
1385 This parameter allows to deal with pools having heavily imbalanced
1386 vdevs such as would be the case when a new vdev has been added.
1387 Setting the threshold to a non-zero percentage will stop allocations
1388 from being made to vdevs that aren't filled to the specified percentage
1389 and allow lesser filled vdevs to acquire more allocations than they
1390 otherwise would under the old \fBzfs_mg_alloc_failures\fR facility.
1392 Default value: \fB0\fR.
1398 \fBzfs_no_scrub_io\fR (int)
1401 Set for no scrub I/O. This results in scrubs not actually scrubbing data and
1402 simply doing a metadata crawl of the pool instead.
1404 Use \fB1\fR for yes and \fB0\fR for no (default).
1410 \fBzfs_no_scrub_prefetch\fR (int)
1413 Set to disable block prefetching for scrubs.
1415 Use \fB1\fR for yes and \fB0\fR for no (default).
1421 \fBzfs_nocacheflush\fR (int)
1424 Disable cache flush operations on disks when writing. Beware, this may cause
1425 corruption if disks re-order writes.
1427 Use \fB1\fR for yes and \fB0\fR for no (default).
1433 \fBzfs_nopwrite_enabled\fR (int)
1438 Use \fB1\fR for yes (default) and \fB0\fR to disable.
1444 \fBzfs_dmu_offset_next_sync\fR (int)
1447 Enable forcing txg sync to find holes. When enabled forces ZFS to act
1448 like prior versions when SEEK_HOLE or SEEK_DATA flags are used, which
1449 when a dnode is dirty causes txg's to be synced so that this data can be
1452 Use \fB1\fR for yes and \fB0\fR to disable (default).
1458 \fBzfs_pd_bytes_max\fR (int)
1461 The number of bytes which should be prefetched during a pool traversal
1462 (eg: \fBzfs send\fR or other data crawling operations)
1464 Default value: \fB52,428,800\fR.
1470 \fBzfs_per_txg_dirty_frees_percent \fR (ulong)
1473 Tunable to control percentage of dirtied blocks from frees in one TXG.
1474 After this threshold is crossed, additional dirty blocks from frees
1475 wait until the next TXG.
1476 A value of zero will disable this throttle.
1478 Default value: \fB30\fR and \fB0\fR to disable.
1486 \fBzfs_prefetch_disable\fR (int)
1489 This tunable disables predictive prefetch. Note that it leaves "prescient"
1490 prefetch (e.g. prefetch for zfs send) intact. Unlike predictive prefetch,
1491 prescient prefetch never issues i/os that end up not being needed, so it
1492 can't hurt performance.
1494 Use \fB1\fR for yes and \fB0\fR for no (default).
1500 \fBzfs_read_chunk_size\fR (long)
1503 Bytes to read per chunk
1505 Default value: \fB1,048,576\fR.
1511 \fBzfs_read_history\fR (int)
1514 Historic statistics for the last N reads will be available in
1515 \fR/proc/spl/kstat/zfs/POOLNAME/reads\fB
1517 Default value: \fB0\fR (no data is kept).
1523 \fBzfs_read_history_hits\fR (int)
1526 Include cache hits in read history
1528 Use \fB1\fR for yes and \fB0\fR for no (default).
1534 \fBzfs_recover\fR (int)
1537 Set to attempt to recover from fatal errors. This should only be used as a
1538 last resort, as it typically results in leaked space, or worse.
1540 Use \fB1\fR for yes and \fB0\fR for no (default).
1546 \fBzfs_resilver_delay\fR (int)
1549 Number of ticks to delay prior to issuing a resilver I/O operation when
1550 a non-resilver or non-scrub I/O operation has occurred within the past
1551 \fBzfs_scan_idle\fR ticks.
1553 Default value: \fB2\fR.
1559 \fBzfs_resilver_min_time_ms\fR (int)
1562 Resilvers are processed by the sync thread. While resilvering it will spend
1563 at least this much time working on a resilver between txg flushes.
1565 Default value: \fB3,000\fR.
1571 \fBzfs_scan_idle\fR (int)
1574 Idle window in clock ticks. During a scrub or a resilver, if
1575 a non-scrub or non-resilver I/O operation has occurred during this
1576 window, the next scrub or resilver operation is delayed by, respectively
1577 \fBzfs_scrub_delay\fR or \fBzfs_resilver_delay\fR ticks.
1579 Default value: \fB50\fR.
1585 \fBzfs_scan_min_time_ms\fR (int)
1588 Scrubs are processed by the sync thread. While scrubbing it will spend
1589 at least this much time working on a scrub between txg flushes.
1591 Default value: \fB1,000\fR.
1597 \fBzfs_scrub_delay\fR (int)
1600 Number of ticks to delay prior to issuing a scrub I/O operation when
1601 a non-scrub or non-resilver I/O operation has occurred within the past
1602 \fBzfs_scan_idle\fR ticks.
1604 Default value: \fB4\fR.
1610 \fBzfs_send_corrupt_data\fR (int)
1613 Allow sending of corrupt data (ignore read/checksum errors when sending data)
1615 Use \fB1\fR for yes and \fB0\fR for no (default).
1621 \fBzfs_sync_pass_deferred_free\fR (int)
1624 Flushing of data to disk is done in passes. Defer frees starting in this pass
1626 Default value: \fB2\fR.
1632 \fBzfs_sync_pass_dont_compress\fR (int)
1635 Don't compress starting in this pass
1637 Default value: \fB5\fR.
1643 \fBzfs_sync_pass_rewrite\fR (int)
1646 Rewrite new block pointers starting in this pass
1648 Default value: \fB2\fR.
1654 \fBzfs_top_maxinflight\fR (int)
1657 Max concurrent I/Os per top-level vdev (mirrors or raidz arrays) allowed during
1658 scrub or resilver operations.
1660 Default value: \fB32\fR.
1666 \fBzfs_txg_history\fR (int)
1669 Historic statistics for the last N txgs will be available in
1670 \fR/proc/spl/kstat/zfs/POOLNAME/txgs\fB
1672 Default value: \fB0\fR.
1678 \fBzfs_txg_timeout\fR (int)
1681 Flush dirty data to disk at least every N seconds (maximum txg duration)
1683 Default value: \fB5\fR.
1689 \fBzfs_vdev_aggregation_limit\fR (int)
1692 Max vdev I/O aggregation size
1694 Default value: \fB131,072\fR.
1700 \fBzfs_vdev_cache_bshift\fR (int)
1703 Shift size to inflate reads too
1705 Default value: \fB16\fR (effectively 65536).
1711 \fBzfs_vdev_cache_max\fR (int)
1714 Inflate reads small than this value to meet the \fBzfs_vdev_cache_bshift\fR
1717 Default value: \fB16384\fR.
1723 \fBzfs_vdev_cache_size\fR (int)
1726 Total size of the per-disk cache in bytes.
1728 Currently this feature is disabled as it has been found to not be helpful
1729 for performance and in some cases harmful.
1731 Default value: \fB0\fR.
1737 \fBzfs_vdev_mirror_rotating_inc\fR (int)
1740 A number by which the balancing algorithm increments the load calculation for
1741 the purpose of selecting the least busy mirror member when an I/O immediately
1742 follows its predecessor on rotational vdevs for the purpose of making decisions
1745 Default value: \fB0\fR.
1751 \fBzfs_vdev_mirror_rotating_seek_inc\fR (int)
1754 A number by which the balancing algorithm increments the load calculation for
1755 the purpose of selecting the least busy mirror member when an I/O lacks
1756 locality as defined by the zfs_vdev_mirror_rotating_seek_offset. I/Os within
1757 this that are not immediately following the previous I/O are incremented by
1760 Default value: \fB5\fR.
1766 \fBzfs_vdev_mirror_rotating_seek_offset\fR (int)
1769 The maximum distance for the last queued I/O in which the balancing algorithm
1770 considers an I/O to have locality.
1771 See the section "ZFS I/O SCHEDULER".
1773 Default value: \fB1048576\fR.
1779 \fBzfs_vdev_mirror_non_rotating_inc\fR (int)
1782 A number by which the balancing algorithm increments the load calculation for
1783 the purpose of selecting the least busy mirror member on non-rotational vdevs
1784 when I/Os do not immediately follow one another.
1786 Default value: \fB0\fR.
1792 \fBzfs_vdev_mirror_non_rotating_seek_inc\fR (int)
1795 A number by which the balancing algorithm increments the load calculation for
1796 the purpose of selecting the least busy mirror member when an I/O lacks
1797 locality as defined by the zfs_vdev_mirror_rotating_seek_offset. I/Os within
1798 this that are not immediately following the previous I/O are incremented by
1801 Default value: \fB1\fR.
1807 \fBzfs_vdev_read_gap_limit\fR (int)
1810 Aggregate read I/O operations if the gap on-disk between them is within this
1813 Default value: \fB32,768\fR.
1819 \fBzfs_vdev_scheduler\fR (charp)
1822 Set the Linux I/O scheduler on whole disk vdevs to this scheduler
1824 Default value: \fBnoop\fR.
1830 \fBzfs_vdev_write_gap_limit\fR (int)
1833 Aggregate write I/O over gap
1835 Default value: \fB4,096\fR.
1841 \fBzfs_vdev_raidz_impl\fR (string)
1844 Parameter for selecting raidz parity implementation to use.
1846 Options marked (always) below may be selected on module load as they are
1847 supported on all systems.
1848 The remaining options may only be set after the module is loaded, as they
1849 are available only if the implementations are compiled in and supported
1850 on the running system.
1852 Once the module is loaded, the content of
1853 /sys/module/zfs/parameters/zfs_vdev_raidz_impl will show available options
1854 with the currently selected one enclosed in [].
1855 Possible options are:
1856 fastest - (always) implementation selected using built-in benchmark
1857 original - (always) original raidz implementation
1858 scalar - (always) scalar raidz implementation
1859 sse2 - implementation using SSE2 instruction set (64bit x86 only)
1860 ssse3 - implementation using SSSE3 instruction set (64bit x86 only)
1861 avx2 - implementation using AVX2 instruction set (64bit x86 only)
1862 avx512f - implementation using AVX512F instruction set (64bit x86 only)
1863 avx512bw - implementation using AVX512F & AVX512BW instruction sets (64bit x86 only)
1864 aarch64_neon - implementation using NEON (Aarch64/64 bit ARMv8 only)
1865 aarch64_neonx2 - implementation using NEON with more unrolling (Aarch64/64 bit ARMv8 only)
1867 Default value: \fBfastest\fR.
1873 \fBzfs_zevent_cols\fR (int)
1876 When zevents are logged to the console use this as the word wrap width.
1878 Default value: \fB80\fR.
1884 \fBzfs_zevent_console\fR (int)
1887 Log events to the console
1889 Use \fB1\fR for yes and \fB0\fR for no (default).
1895 \fBzfs_zevent_len_max\fR (int)
1898 Max event queue length. A value of 0 will result in a calculated value which
1899 increases with the number of CPUs in the system (minimum 64 events). Events
1900 in the queue can be viewed with the \fBzpool events\fR command.
1902 Default value: \fB0\fR.
1908 \fBzil_replay_disable\fR (int)
1911 Disable intent logging replay. Can be disabled for recovery from corrupted
1914 Use \fB1\fR for yes and \fB0\fR for no (default).
1920 \fBzil_slog_limit\fR (ulong)
1923 Max commit bytes to separate log device
1925 Default value: \fB1,048,576\fR.
1931 \fBzio_delay_max\fR (int)
1934 A zevent will be logged if a ZIO operation takes more than N milliseconds to
1935 complete. Note that this is only a logging facility, not a timeout on
1938 Default value: \fB30,000\fR.
1944 \fBzio_dva_throttle_enabled\fR (int)
1947 Throttle block allocations in the ZIO pipeline. This allows for
1948 dynamic allocation distribution when devices are imbalanced.
1949 When enabled, the maximum number of pending allocations per top-level vdev
1950 is limited by \fBzfs_vdev_queue_depth_pct\fR.
1952 Default value: \fB1\fR.
1958 \fBzio_requeue_io_start_cut_in_line\fR (int)
1961 Prioritize requeued I/O
1963 Default value: \fB0\fR.
1969 \fBzio_taskq_batch_pct\fR (uint)
1972 Percentage of online CPUs (or CPU cores, etc) which will run a worker thread
1973 for IO. These workers are responsible for IO work such as compression and
1974 checksum calculations. Fractional number of CPUs will be rounded down.
1976 The default value of 75 was chosen to avoid using all CPUs which can result in
1977 latency issues and inconsistent application performance, especially when high
1978 compression is enabled.
1980 Default value: \fB75\fR.
1986 \fBzvol_inhibit_dev\fR (uint)
1989 Do not create zvol device nodes. This may slightly improve startup time on
1990 systems with a very large number of zvols.
1992 Use \fB1\fR for yes and \fB0\fR for no (default).
1998 \fBzvol_major\fR (uint)
2001 Major number for zvol block devices
2003 Default value: \fB230\fR.
2009 \fBzvol_max_discard_blocks\fR (ulong)
2012 Discard (aka TRIM) operations done on zvols will be done in batches of this
2013 many blocks, where block size is determined by the \fBvolblocksize\fR property
2016 Default value: \fB16,384\fR.
2022 \fBzvol_prefetch_bytes\fR (uint)
2025 When adding a zvol to the system prefetch \fBzvol_prefetch_bytes\fR
2026 from the start and end of the volume. Prefetching these regions
2027 of the volume is desirable because they are likely to be accessed
2028 immediately by \fBblkid(8)\fR or by the kernel scanning for a partition
2031 Default value: \fB131,072\fR.
2037 \fBzfs_qat_disable\fR (int)
2040 This tunable disables qat hardware acceleration for gzip compression.
2041 It is available only if qat acceleration is compiled in and qat driver
2044 Use \fB1\fR for yes and \fB0\fR for no (default).
2047 .SH ZFS I/O SCHEDULER
2048 ZFS issues I/O operations to leaf vdevs to satisfy and complete I/Os.
2049 The I/O scheduler determines when and in what order those operations are
2050 issued. The I/O scheduler divides operations into five I/O classes
2051 prioritized in the following order: sync read, sync write, async read,
2052 async write, and scrub/resilver. Each queue defines the minimum and
2053 maximum number of concurrent operations that may be issued to the
2054 device. In addition, the device has an aggregate maximum,
2055 \fBzfs_vdev_max_active\fR. Note that the sum of the per-queue minimums
2056 must not exceed the aggregate maximum. If the sum of the per-queue
2057 maximums exceeds the aggregate maximum, then the number of active I/Os
2058 may reach \fBzfs_vdev_max_active\fR, in which case no further I/Os will
2059 be issued regardless of whether all per-queue minimums have been met.
2061 For many physical devices, throughput increases with the number of
2062 concurrent operations, but latency typically suffers. Further, physical
2063 devices typically have a limit at which more concurrent operations have no
2064 effect on throughput or can actually cause it to decrease.
2066 The scheduler selects the next operation to issue by first looking for an
2067 I/O class whose minimum has not been satisfied. Once all are satisfied and
2068 the aggregate maximum has not been hit, the scheduler looks for classes
2069 whose maximum has not been satisfied. Iteration through the I/O classes is
2070 done in the order specified above. No further operations are issued if the
2071 aggregate maximum number of concurrent operations has been hit or if there
2072 are no operations queued for an I/O class that has not hit its maximum.
2073 Every time an I/O is queued or an operation completes, the I/O scheduler
2074 looks for new operations to issue.
2076 In general, smaller max_active's will lead to lower latency of synchronous
2077 operations. Larger max_active's may lead to higher overall throughput,
2078 depending on underlying storage.
2080 The ratio of the queues' max_actives determines the balance of performance
2081 between reads, writes, and scrubs. E.g., increasing
2082 \fBzfs_vdev_scrub_max_active\fR will cause the scrub or resilver to complete
2083 more quickly, but reads and writes to have higher latency and lower throughput.
2085 All I/O classes have a fixed maximum number of outstanding operations
2086 except for the async write class. Asynchronous writes represent the data
2087 that is committed to stable storage during the syncing stage for
2088 transaction groups. Transaction groups enter the syncing state
2089 periodically so the number of queued async writes will quickly burst up
2090 and then bleed down to zero. Rather than servicing them as quickly as
2091 possible, the I/O scheduler changes the maximum number of active async
2092 write I/Os according to the amount of dirty data in the pool. Since
2093 both throughput and latency typically increase with the number of
2094 concurrent operations issued to physical devices, reducing the
2095 burstiness in the number of concurrent operations also stabilizes the
2096 response time of operations from other -- and in particular synchronous
2097 -- queues. In broad strokes, the I/O scheduler will issue more
2098 concurrent operations from the async write queue as there's more dirty
2103 The number of concurrent operations issued for the async write I/O class
2104 follows a piece-wise linear function defined by a few adjustable points.
2107 | o---------| <-- zfs_vdev_async_write_max_active
2114 |-------o | | <-- zfs_vdev_async_write_min_active
2115 0|_______^______|_________|
2116 0% | | 100% of zfs_dirty_data_max
2118 | `-- zfs_vdev_async_write_active_max_dirty_percent
2119 `--------- zfs_vdev_async_write_active_min_dirty_percent
2122 Until the amount of dirty data exceeds a minimum percentage of the dirty
2123 data allowed in the pool, the I/O scheduler will limit the number of
2124 concurrent operations to the minimum. As that threshold is crossed, the
2125 number of concurrent operations issued increases linearly to the maximum at
2126 the specified maximum percentage of the dirty data allowed in the pool.
2128 Ideally, the amount of dirty data on a busy pool will stay in the sloped
2129 part of the function between \fBzfs_vdev_async_write_active_min_dirty_percent\fR
2130 and \fBzfs_vdev_async_write_active_max_dirty_percent\fR. If it exceeds the
2131 maximum percentage, this indicates that the rate of incoming data is
2132 greater than the rate that the backend storage can handle. In this case, we
2133 must further throttle incoming writes, as described in the next section.
2135 .SH ZFS TRANSACTION DELAY
2136 We delay transactions when we've determined that the backend storage
2137 isn't able to accommodate the rate of incoming writes.
2139 If there is already a transaction waiting, we delay relative to when
2140 that transaction will finish waiting. This way the calculated delay time
2141 is independent of the number of threads concurrently executing
2144 If we are the only waiter, wait relative to when the transaction
2145 started, rather than the current time. This credits the transaction for
2146 "time already served", e.g. reading indirect blocks.
2148 The minimum time for a transaction to take is calculated as:
2150 min_time = zfs_delay_scale * (dirty - min) / (max - dirty)
2151 min_time is then capped at 100 milliseconds.
2154 The delay has two degrees of freedom that can be adjusted via tunables. The
2155 percentage of dirty data at which we start to delay is defined by
2156 \fBzfs_delay_min_dirty_percent\fR. This should typically be at or above
2157 \fBzfs_vdev_async_write_active_max_dirty_percent\fR so that we only start to
2158 delay after writing at full speed has failed to keep up with the incoming write
2159 rate. The scale of the curve is defined by \fBzfs_delay_scale\fR. Roughly speaking,
2160 this variable determines the amount of delay at the midpoint of the curve.
2164 10ms +-------------------------------------------------------------*+
2180 2ms + (midpoint) * +
2183 | zfs_delay_scale ----------> ******** |
2184 0 +-------------------------------------*********----------------+
2185 0% <- zfs_dirty_data_max -> 100%
2188 Note that since the delay is added to the outstanding time remaining on the
2189 most recent transaction, the delay is effectively the inverse of IOPS.
2190 Here the midpoint of 500us translates to 2000 IOPS. The shape of the curve
2191 was chosen such that small changes in the amount of accumulated dirty data
2192 in the first 3/4 of the curve yield relatively small differences in the
2195 The effects can be easier to understand when the amount of delay is
2196 represented on a log scale:
2200 100ms +-------------------------------------------------------------++
2209 + zfs_delay_scale ----------> ***** +
2220 +--------------------------------------------------------------+
2221 0% <- zfs_dirty_data_max -> 100%
2224 Note here that only as the amount of dirty data approaches its limit does
2225 the delay start to increase rapidly. The goal of a properly tuned system
2226 should be to keep the amount of dirty data out of that range by first
2227 ensuring that the appropriate limits are set for the I/O scheduler to reach
2228 optimal throughput on the backend storage, and then by changing the value
2229 of \fBzfs_delay_scale\fR to increase the steepness of the curve.