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_nax\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 \fBlzfs_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_arc_num_sublists_per_state\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.
633 Default value: \fB1\fR or the number of online CPUs, whichever is greater
639 \fBzfs_arc_overflow_shift\fR (int)
642 The ARC size is considered to be overflowing if it exceeds the current
643 ARC target size (arc_c) by a threshold determined by this parameter.
644 The threshold is calculated as a fraction of arc_c using the formula
645 "arc_c >> \fBzfs_arc_overflow_shift\fR".
647 The default value of 8 causes the ARC to be considered to be overflowing
648 if it exceeds the target size by 1/256th (0.3%) of the target size.
650 When the ARC is overflowing, new buffer allocations are stalled until
651 the reclaim thread catches up and the overflow condition no longer exists.
653 Default value: \fB8\fR.
660 \fBzfs_arc_p_min_shift\fR (int)
663 arc_c shift to calc min/max arc_p
665 Default value: \fB4\fR.
671 \fBzfs_arc_p_aggressive_disable\fR (int)
674 Disable aggressive arc_p growth
676 Use \fB1\fR for yes (default) and \fB0\fR to disable.
682 \fBzfs_arc_p_dampener_disable\fR (int)
685 Disable arc_p adapt dampener
687 Use \fB1\fR for yes (default) and \fB0\fR to disable.
693 \fBzfs_arc_shrink_shift\fR (int)
696 log2(fraction of arc to reclaim)
698 Default value: \fB5\fR.
704 \fBzfs_arc_sys_free\fR (ulong)
707 The target number of bytes the ARC should leave as free memory on the system.
708 Defaults to the larger of 1/64 of physical memory or 512K. Setting this
709 option to a non-zero value will override the default.
711 Default value: \fB0\fR.
717 \fBzfs_autoimport_disable\fR (int)
720 Disable pool import at module load by ignoring the cache file (typically \fB/etc/zfs/zpool.cache\fR).
722 Use \fB1\fR for yes (default) and \fB0\fR for no.
728 \fBzfs_dbgmsg_enable\fR (int)
731 Internally ZFS keeps a small log to facilitate debugging. By default the log
732 is disabled, to enable it set this option to 1. The contents of the log can
733 be accessed by reading the /proc/spl/kstat/zfs/dbgmsg file. Writing 0 to
734 this proc file clears the log.
736 Default value: \fB0\fR.
742 \fBzfs_dbgmsg_maxsize\fR (int)
745 The maximum size in bytes of the internal ZFS debug log.
747 Default value: \fB4M\fR.
753 \fBzfs_dbuf_state_index\fR (int)
756 This feature is currently unused. It is normally used for controlling what
757 reporting is available under /proc/spl/kstat/zfs.
759 Default value: \fB0\fR.
765 \fBzfs_deadman_enabled\fR (int)
768 When a pool sync operation takes longer than \fBzfs_deadman_synctime_ms\fR
769 milliseconds, a "slow spa_sync" message is logged to the debug log
770 (see \fBzfs_dbgmsg_enable\fR). If \fBzfs_deadman_enabled\fR is set,
771 all pending IO operations are also checked and if any haven't completed
772 within \fBzfs_deadman_synctime_ms\fR milliseconds, a "SLOW IO" message
773 is logged to the debug log and a "delay" system event with the details of
774 the hung IO is posted.
776 Use \fB1\fR (default) to enable the slow IO check and \fB0\fR to disable.
782 \fBzfs_deadman_checktime_ms\fR (int)
785 Once a pool sync operation has taken longer than
786 \fBzfs_deadman_synctime_ms\fR milliseconds, continue to check for slow
787 operations every \fBzfs_deadman_checktime_ms\fR milliseconds.
789 Default value: \fB5,000\fR.
795 \fBzfs_deadman_synctime_ms\fR (ulong)
798 Interval in milliseconds after which the deadman is triggered and also
799 the interval after which an IO operation is considered to be "hung"
800 if \fBzfs_deadman_enabled\fR is set.
802 See \fBzfs_deadman_enabled\fR.
804 Default value: \fB1,000,000\fR.
810 \fBzfs_dedup_prefetch\fR (int)
813 Enable prefetching dedup-ed blks
815 Use \fB1\fR for yes and \fB0\fR to disable (default).
821 \fBzfs_delay_min_dirty_percent\fR (int)
824 Start to delay each transaction once there is this amount of dirty data,
825 expressed as a percentage of \fBzfs_dirty_data_max\fR.
826 This value should be >= zfs_vdev_async_write_active_max_dirty_percent.
827 See the section "ZFS TRANSACTION DELAY".
829 Default value: \fB60\fR.
835 \fBzfs_delay_scale\fR (int)
838 This controls how quickly the transaction delay approaches infinity.
839 Larger values cause longer delays for a given amount of dirty data.
841 For the smoothest delay, this value should be about 1 billion divided
842 by the maximum number of operations per second. This will smoothly
843 handle between 10x and 1/10th this number.
845 See the section "ZFS TRANSACTION DELAY".
847 Note: \fBzfs_delay_scale\fR * \fBzfs_dirty_data_max\fR must be < 2^64.
849 Default value: \fB500,000\fR.
855 \fBzfs_delete_blocks\fR (ulong)
858 This is the used to define a large file for the purposes of delete. Files
859 containing more than \fBzfs_delete_blocks\fR will be deleted asynchronously
860 while smaller files are deleted synchronously. Decreasing this value will
861 reduce the time spent in an unlink(2) system call at the expense of a longer
862 delay before the freed space is available.
864 Default value: \fB20,480\fR.
870 \fBzfs_dirty_data_max\fR (int)
873 Determines the dirty space limit in bytes. Once this limit is exceeded, new
874 writes are halted until space frees up. This parameter takes precedence
875 over \fBzfs_dirty_data_max_percent\fR.
876 See the section "ZFS TRANSACTION DELAY".
878 Default value: 10 percent of all memory, capped at \fBzfs_dirty_data_max_max\fR.
884 \fBzfs_dirty_data_max_max\fR (int)
887 Maximum allowable value of \fBzfs_dirty_data_max\fR, expressed in bytes.
888 This limit is only enforced at module load time, and will be ignored if
889 \fBzfs_dirty_data_max\fR is later changed. This parameter takes
890 precedence over \fBzfs_dirty_data_max_max_percent\fR. See the section
891 "ZFS TRANSACTION DELAY".
893 Default value: 25% of physical RAM.
899 \fBzfs_dirty_data_max_max_percent\fR (int)
902 Maximum allowable value of \fBzfs_dirty_data_max\fR, expressed as a
903 percentage of physical RAM. This limit is only enforced at module load
904 time, and will be ignored if \fBzfs_dirty_data_max\fR is later changed.
905 The parameter \fBzfs_dirty_data_max_max\fR takes precedence over this
906 one. See the section "ZFS TRANSACTION DELAY".
908 Default value: \fB25\fR.
914 \fBzfs_dirty_data_max_percent\fR (int)
917 Determines the dirty space limit, expressed as a percentage of all
918 memory. Once this limit is exceeded, new writes are halted until space frees
919 up. The parameter \fBzfs_dirty_data_max\fR takes precedence over this
920 one. See the section "ZFS TRANSACTION DELAY".
922 Default value: 10%, subject to \fBzfs_dirty_data_max_max\fR.
928 \fBzfs_dirty_data_sync\fR (int)
931 Start syncing out a transaction group if there is at least this much dirty data.
933 Default value: \fB67,108,864\fR.
939 \fBzfs_fletcher_4_impl\fR (string)
942 Select a fletcher 4 implementation.
944 Supported selectors are: \fBfastest\fR, \fBscalar\fR, \fBsse2\fR, \fBssse3\fR,
945 \fBavx2\fR, \fBavx512f\fR, and \fBaarch64_neon\fR.
946 All of the selectors except \fBfastest\fR and \fBscalar\fR require instruction
947 set extensions to be available and will only appear if ZFS detects that they are
948 present at runtime. If multiple implementations of fletcher 4 are available,
949 the \fBfastest\fR will be chosen using a micro benchmark. Selecting \fBscalar\fR
950 results in the original, CPU based calculation, being used. Selecting any option
951 other than \fBfastest\fR and \fBscalar\fR results in vector instructions from
952 the respective CPU instruction set being used.
954 Default value: \fBfastest\fR.
960 \fBzfs_free_bpobj_enabled\fR (int)
963 Enable/disable the processing of the free_bpobj object.
965 Default value: \fB1\fR.
971 \fBzfs_free_max_blocks\fR (ulong)
974 Maximum number of blocks freed in a single txg.
976 Default value: \fB100,000\fR.
982 \fBzfs_vdev_async_read_max_active\fR (int)
985 Maximum asynchronous read I/Os active to each device.
986 See the section "ZFS I/O SCHEDULER".
988 Default value: \fB3\fR.
994 \fBzfs_vdev_async_read_min_active\fR (int)
997 Minimum asynchronous read I/Os active to each device.
998 See the section "ZFS I/O SCHEDULER".
1000 Default value: \fB1\fR.
1006 \fBzfs_vdev_async_write_active_max_dirty_percent\fR (int)
1009 When the pool has more than
1010 \fBzfs_vdev_async_write_active_max_dirty_percent\fR dirty data, use
1011 \fBzfs_vdev_async_write_max_active\fR to limit active async writes. If
1012 the dirty data is between min and max, the active I/O limit is linearly
1013 interpolated. See the section "ZFS I/O SCHEDULER".
1015 Default value: \fB60\fR.
1021 \fBzfs_vdev_async_write_active_min_dirty_percent\fR (int)
1024 When the pool has less than
1025 \fBzfs_vdev_async_write_active_min_dirty_percent\fR dirty data, use
1026 \fBzfs_vdev_async_write_min_active\fR to limit active async writes. If
1027 the dirty data is between min and max, the active I/O limit is linearly
1028 interpolated. See the section "ZFS I/O SCHEDULER".
1030 Default value: \fB30\fR.
1036 \fBzfs_vdev_async_write_max_active\fR (int)
1039 Maximum asynchronous write I/Os active to each device.
1040 See the section "ZFS I/O SCHEDULER".
1042 Default value: \fB10\fR.
1048 \fBzfs_vdev_async_write_min_active\fR (int)
1051 Minimum asynchronous write I/Os active to each device.
1052 See the section "ZFS I/O SCHEDULER".
1054 Default value: \fB1\fR.
1060 \fBzfs_vdev_max_active\fR (int)
1063 The maximum number of I/Os active to each device. Ideally, this will be >=
1064 the sum of each queue's max_active. It must be at least the sum of each
1065 queue's min_active. See the section "ZFS I/O SCHEDULER".
1067 Default value: \fB1,000\fR.
1073 \fBzfs_vdev_scrub_max_active\fR (int)
1076 Maximum scrub I/Os active to each device.
1077 See the section "ZFS I/O SCHEDULER".
1079 Default value: \fB2\fR.
1085 \fBzfs_vdev_scrub_min_active\fR (int)
1088 Minimum scrub I/Os active to each device.
1089 See the section "ZFS I/O SCHEDULER".
1091 Default value: \fB1\fR.
1097 \fBzfs_vdev_sync_read_max_active\fR (int)
1100 Maximum synchronous read I/Os active to each device.
1101 See the section "ZFS I/O SCHEDULER".
1103 Default value: \fB10\fR.
1109 \fBzfs_vdev_sync_read_min_active\fR (int)
1112 Minimum synchronous read I/Os active to each device.
1113 See the section "ZFS I/O SCHEDULER".
1115 Default value: \fB10\fR.
1121 \fBzfs_vdev_sync_write_max_active\fR (int)
1124 Maximum synchronous write I/Os active to each device.
1125 See the section "ZFS I/O SCHEDULER".
1127 Default value: \fB10\fR.
1133 \fBzfs_vdev_sync_write_min_active\fR (int)
1136 Minimum synchronous write I/Os active to each device.
1137 See the section "ZFS I/O SCHEDULER".
1139 Default value: \fB10\fR.
1145 \fBzfs_vdev_queue_depth_pct\fR (int)
1148 The queue depth percentage for each top-level virtual device.
1149 Used in conjunction with zfs_vdev_async_max_active.
1151 Default value: \fB1000\fR.
1157 \fBzfs_disable_dup_eviction\fR (int)
1160 Disable duplicate buffer eviction
1162 Use \fB1\fR for yes and \fB0\fR for no (default).
1168 \fBzfs_expire_snapshot\fR (int)
1171 Seconds to expire .zfs/snapshot
1173 Default value: \fB300\fR.
1179 \fBzfs_admin_snapshot\fR (int)
1182 Allow the creation, removal, or renaming of entries in the .zfs/snapshot
1183 directory to cause the creation, destruction, or renaming of snapshots.
1184 When enabled this functionality works both locally and over NFS exports
1185 which have the 'no_root_squash' option set. This functionality is disabled
1188 Use \fB1\fR for yes and \fB0\fR for no (default).
1194 \fBzfs_flags\fR (int)
1197 Set additional debugging flags. The following flags may be bitwise-or'd
1209 Enable dprintf entries in the debug log.
1211 2 ZFS_DEBUG_DBUF_VERIFY *
1212 Enable extra dbuf verifications.
1214 4 ZFS_DEBUG_DNODE_VERIFY *
1215 Enable extra dnode verifications.
1217 8 ZFS_DEBUG_SNAPNAMES
1218 Enable snapshot name verification.
1221 Check for illegally modified ARC buffers.
1224 Enable spa_dbgmsg entries in the debug log.
1226 64 ZFS_DEBUG_ZIO_FREE
1227 Enable verification of block frees.
1229 128 ZFS_DEBUG_HISTOGRAM_VERIFY
1230 Enable extra spacemap histogram verifications.
1233 * Requires debug build.
1235 Default value: \fB0\fR.
1241 \fBzfs_free_leak_on_eio\fR (int)
1244 If destroy encounters an EIO while reading metadata (e.g. indirect
1245 blocks), space referenced by the missing metadata can not be freed.
1246 Normally this causes the background destroy to become "stalled", as
1247 it is unable to make forward progress. While in this stalled state,
1248 all remaining space to free from the error-encountering filesystem is
1249 "temporarily leaked". Set this flag to cause it to ignore the EIO,
1250 permanently leak the space from indirect blocks that can not be read,
1251 and continue to free everything else that it can.
1253 The default, "stalling" behavior is useful if the storage partially
1254 fails (i.e. some but not all i/os fail), and then later recovers. In
1255 this case, we will be able to continue pool operations while it is
1256 partially failed, and when it recovers, we can continue to free the
1257 space, with no leaks. However, note that this case is actually
1260 Typically pools either (a) fail completely (but perhaps temporarily,
1261 e.g. a top-level vdev going offline), or (b) have localized,
1262 permanent errors (e.g. disk returns the wrong data due to bit flip or
1263 firmware bug). In case (a), this setting does not matter because the
1264 pool will be suspended and the sync thread will not be able to make
1265 forward progress regardless. In case (b), because the error is
1266 permanent, the best we can do is leak the minimum amount of space,
1267 which is what setting this flag will do. Therefore, it is reasonable
1268 for this flag to normally be set, but we chose the more conservative
1269 approach of not setting it, so that there is no possibility of
1270 leaking space in the "partial temporary" failure case.
1272 Default value: \fB0\fR.
1278 \fBzfs_free_min_time_ms\fR (int)
1281 During a \fBzfs destroy\fR operation using \fBfeature@async_destroy\fR a minimum
1282 of this much time will be spent working on freeing blocks per txg.
1284 Default value: \fB1,000\fR.
1290 \fBzfs_immediate_write_sz\fR (long)
1293 Largest data block to write to zil. Larger blocks will be treated as if the
1294 dataset being written to had the property setting \fBlogbias=throughput\fR.
1296 Default value: \fB32,768\fR.
1302 \fBzfs_max_recordsize\fR (int)
1305 We currently support block sizes from 512 bytes to 16MB. The benefits of
1306 larger blocks, and thus larger IO, need to be weighed against the cost of
1307 COWing a giant block to modify one byte. Additionally, very large blocks
1308 can have an impact on i/o latency, and also potentially on the memory
1309 allocator. Therefore, we do not allow the recordsize to be set larger than
1310 zfs_max_recordsize (default 1MB). Larger blocks can be created by changing
1311 this tunable, and pools with larger blocks can always be imported and used,
1312 regardless of this setting.
1314 Default value: \fB1,048,576\fR.
1320 \fBzfs_mdcomp_disable\fR (int)
1323 Disable meta data compression
1325 Use \fB1\fR for yes and \fB0\fR for no (default).
1331 \fBzfs_metaslab_fragmentation_threshold\fR (int)
1334 Allow metaslabs to keep their active state as long as their fragmentation
1335 percentage is less than or equal to this value. An active metaslab that
1336 exceeds this threshold will no longer keep its active status allowing
1337 better metaslabs to be selected.
1339 Default value: \fB70\fR.
1345 \fBzfs_mg_fragmentation_threshold\fR (int)
1348 Metaslab groups are considered eligible for allocations if their
1349 fragmentation metric (measured as a percentage) is less than or equal to
1350 this value. If a metaslab group exceeds this threshold then it will be
1351 skipped unless all metaslab groups within the metaslab class have also
1352 crossed this threshold.
1354 Default value: \fB85\fR.
1360 \fBzfs_mg_noalloc_threshold\fR (int)
1363 Defines a threshold at which metaslab groups should be eligible for
1364 allocations. The value is expressed as a percentage of free space
1365 beyond which a metaslab group is always eligible for allocations.
1366 If a metaslab group's free space is less than or equal to the
1367 threshold, the allocator will avoid allocating to that group
1368 unless all groups in the pool have reached the threshold. Once all
1369 groups have reached the threshold, all groups are allowed to accept
1370 allocations. The default value of 0 disables the feature and causes
1371 all metaslab groups to be eligible for allocations.
1373 This parameter allows to deal with pools having heavily imbalanced
1374 vdevs such as would be the case when a new vdev has been added.
1375 Setting the threshold to a non-zero percentage will stop allocations
1376 from being made to vdevs that aren't filled to the specified percentage
1377 and allow lesser filled vdevs to acquire more allocations than they
1378 otherwise would under the old \fBzfs_mg_alloc_failures\fR facility.
1380 Default value: \fB0\fR.
1386 \fBzfs_no_scrub_io\fR (int)
1389 Set for no scrub I/O. This results in scrubs not actually scrubbing data and
1390 simply doing a metadata crawl of the pool instead.
1392 Use \fB1\fR for yes and \fB0\fR for no (default).
1398 \fBzfs_no_scrub_prefetch\fR (int)
1401 Set to disable block prefetching for scrubs.
1403 Use \fB1\fR for yes and \fB0\fR for no (default).
1409 \fBzfs_nocacheflush\fR (int)
1412 Disable cache flush operations on disks when writing. Beware, this may cause
1413 corruption if disks re-order writes.
1415 Use \fB1\fR for yes and \fB0\fR for no (default).
1421 \fBzfs_nopwrite_enabled\fR (int)
1426 Use \fB1\fR for yes (default) and \fB0\fR to disable.
1432 \fBzfs_pd_bytes_max\fR (int)
1435 The number of bytes which should be prefetched during a pool traversal
1436 (eg: \fBzfs send\fR or other data crawling operations)
1438 Default value: \fB52,428,800\fR.
1444 \fBzfs_per_txg_dirty_frees_percent \fR (ulong)
1447 Tunable to control percentage of dirtied blocks from frees in one TXG.
1448 After this threshold is crossed, additional dirty blocks from frees
1449 wait until the next TXG.
1450 A value of zero will disable this throttle.
1452 Default value: \fB30\fR and \fB0\fR to disable.
1460 \fBzfs_prefetch_disable\fR (int)
1463 This tunable disables predictive prefetch. Note that it leaves "prescient"
1464 prefetch (e.g. prefetch for zfs send) intact. Unlike predictive prefetch,
1465 prescient prefetch never issues i/os that end up not being needed, so it
1466 can't hurt performance.
1468 Use \fB1\fR for yes and \fB0\fR for no (default).
1474 \fBzfs_read_chunk_size\fR (long)
1477 Bytes to read per chunk
1479 Default value: \fB1,048,576\fR.
1485 \fBzfs_read_history\fR (int)
1488 Historic statistics for the last N reads will be available in
1489 \fR/proc/spl/kstat/zfs/POOLNAME/reads\fB
1491 Default value: \fB0\fR (no data is kept).
1497 \fBzfs_read_history_hits\fR (int)
1500 Include cache hits in read history
1502 Use \fB1\fR for yes and \fB0\fR for no (default).
1508 \fBzfs_recover\fR (int)
1511 Set to attempt to recover from fatal errors. This should only be used as a
1512 last resort, as it typically results in leaked space, or worse.
1514 Use \fB1\fR for yes and \fB0\fR for no (default).
1520 \fBzfs_resilver_delay\fR (int)
1523 Number of ticks to delay prior to issuing a resilver I/O operation when
1524 a non-resilver or non-scrub I/O operation has occurred within the past
1525 \fBzfs_scan_idle\fR ticks.
1527 Default value: \fB2\fR.
1533 \fBzfs_resilver_min_time_ms\fR (int)
1536 Resilvers are processed by the sync thread. While resilvering it will spend
1537 at least this much time working on a resilver between txg flushes.
1539 Default value: \fB3,000\fR.
1545 \fBzfs_scan_idle\fR (int)
1548 Idle window in clock ticks. During a scrub or a resilver, if
1549 a non-scrub or non-resilver I/O operation has occurred during this
1550 window, the next scrub or resilver operation is delayed by, respectively
1551 \fBzfs_scrub_delay\fR or \fBzfs_resilver_delay\fR ticks.
1553 Default value: \fB50\fR.
1559 \fBzfs_scan_min_time_ms\fR (int)
1562 Scrubs are processed by the sync thread. While scrubbing it will spend
1563 at least this much time working on a scrub between txg flushes.
1565 Default value: \fB1,000\fR.
1571 \fBzfs_scrub_delay\fR (int)
1574 Number of ticks to delay prior to issuing a scrub I/O operation when
1575 a non-scrub or non-resilver I/O operation has occurred within the past
1576 \fBzfs_scan_idle\fR ticks.
1578 Default value: \fB4\fR.
1584 \fBzfs_send_corrupt_data\fR (int)
1587 Allow sending of corrupt data (ignore read/checksum errors when sending data)
1589 Use \fB1\fR for yes and \fB0\fR for no (default).
1595 \fBzfs_sync_pass_deferred_free\fR (int)
1598 Flushing of data to disk is done in passes. Defer frees starting in this pass
1600 Default value: \fB2\fR.
1606 \fBzfs_sync_pass_dont_compress\fR (int)
1609 Don't compress starting in this pass
1611 Default value: \fB5\fR.
1617 \fBzfs_sync_pass_rewrite\fR (int)
1620 Rewrite new block pointers starting in this pass
1622 Default value: \fB2\fR.
1628 \fBzfs_top_maxinflight\fR (int)
1631 Max concurrent I/Os per top-level vdev (mirrors or raidz arrays) allowed during
1632 scrub or resilver operations.
1634 Default value: \fB32\fR.
1640 \fBzfs_txg_history\fR (int)
1643 Historic statistics for the last N txgs will be available in
1644 \fR/proc/spl/kstat/zfs/POOLNAME/txgs\fB
1646 Default value: \fB0\fR.
1652 \fBzfs_txg_timeout\fR (int)
1655 Flush dirty data to disk at least every N seconds (maximum txg duration)
1657 Default value: \fB5\fR.
1663 \fBzfs_vdev_aggregation_limit\fR (int)
1666 Max vdev I/O aggregation size
1668 Default value: \fB131,072\fR.
1674 \fBzfs_vdev_cache_bshift\fR (int)
1677 Shift size to inflate reads too
1679 Default value: \fB16\fR (effectively 65536).
1685 \fBzfs_vdev_cache_max\fR (int)
1688 Inflate reads small than this value to meet the \fBzfs_vdev_cache_bshift\fR
1691 Default value: \fB16384\fR.
1697 \fBzfs_vdev_cache_size\fR (int)
1700 Total size of the per-disk cache in bytes.
1702 Currently this feature is disabled as it has been found to not be helpful
1703 for performance and in some cases harmful.
1705 Default value: \fB0\fR.
1711 \fBzfs_vdev_mirror_rotating_inc\fR (int)
1714 A number by which the balancing algorithm increments the load calculation for
1715 the purpose of selecting the least busy mirror member when an I/O immediately
1716 follows its predecessor on rotational vdevs for the purpose of making decisions
1719 Default value: \fB0\fR.
1725 \fBzfs_vdev_mirror_rotating_seek_inc\fR (int)
1728 A number by which the balancing algorithm increments the load calculation for
1729 the purpose of selecting the least busy mirror member when an I/O lacks
1730 locality as defined by the zfs_vdev_mirror_rotating_seek_offset. I/Os within
1731 this that are not immediately following the previous I/O are incremented by
1734 Default value: \fB5\fR.
1740 \fBzfs_vdev_mirror_rotating_seek_offset\fR (int)
1743 The maximum distance for the last queued I/O in which the balancing algorithm
1744 considers an I/O to have locality.
1745 See the section "ZFS I/O SCHEDULER".
1747 Default value: \fB1048576\fR.
1753 \fBzfs_vdev_mirror_non_rotating_inc\fR (int)
1756 A number by which the balancing algorithm increments the load calculation for
1757 the purpose of selecting the least busy mirror member on non-rotational vdevs
1758 when I/Os do not immediately follow one another.
1760 Default value: \fB0\fR.
1766 \fBzfs_vdev_mirror_non_rotating_seek_inc\fR (int)
1769 A number by which the balancing algorithm increments the load calculation for
1770 the purpose of selecting the least busy mirror member when an I/O lacks
1771 locality as defined by the zfs_vdev_mirror_rotating_seek_offset. I/Os within
1772 this that are not immediately following the previous I/O are incremented by
1775 Default value: \fB1\fR.
1781 \fBzfs_vdev_read_gap_limit\fR (int)
1784 Aggregate read I/O operations if the gap on-disk between them is within this
1787 Default value: \fB32,768\fR.
1793 \fBzfs_vdev_scheduler\fR (charp)
1796 Set the Linux I/O scheduler on whole disk vdevs to this scheduler
1798 Default value: \fBnoop\fR.
1804 \fBzfs_vdev_write_gap_limit\fR (int)
1807 Aggregate write I/O over gap
1809 Default value: \fB4,096\fR.
1815 \fBzfs_vdev_raidz_impl\fR (string)
1818 Parameter for selecting raidz parity implementation to use.
1820 Options marked (always) below may be selected on module load as they are
1821 supported on all systems.
1822 The remaining options may only be set after the module is loaded, as they
1823 are available only if the implementations are compiled in and supported
1824 on the running system.
1826 Once the module is loaded, the content of
1827 /sys/module/zfs/parameters/zfs_vdev_raidz_impl will show available options
1828 with the currently selected one enclosed in [].
1829 Possible options are:
1830 fastest - (always) implementation selected using built-in benchmark
1831 original - (always) original raidz implementation
1832 scalar - (always) scalar raidz implementation
1833 sse2 - implementation using SSE2 instruction set (64bit x86 only)
1834 ssse3 - implementation using SSSE3 instruction set (64bit x86 only)
1835 avx2 - implementation using AVX2 instruction set (64bit x86 only)
1836 avx512f - implementation using AVX512F instruction set (64bit x86 only)
1837 avx512bw - implementation using AVX512F & AVX512BW instruction sets (64bit x86 only)
1838 aarch64_neon - implementation using NEON (Aarch64/64 bit ARMv8 only)
1839 aarch64_neonx2 - implementation using NEON with more unrolling (Aarch64/64 bit ARMv8 only)
1841 Default value: \fBfastest\fR.
1847 \fBzfs_zevent_cols\fR (int)
1850 When zevents are logged to the console use this as the word wrap width.
1852 Default value: \fB80\fR.
1858 \fBzfs_zevent_console\fR (int)
1861 Log events to the console
1863 Use \fB1\fR for yes and \fB0\fR for no (default).
1869 \fBzfs_zevent_len_max\fR (int)
1872 Max event queue length. A value of 0 will result in a calculated value which
1873 increases with the number of CPUs in the system (minimum 64 events). Events
1874 in the queue can be viewed with the \fBzpool events\fR command.
1876 Default value: \fB0\fR.
1882 \fBzil_replay_disable\fR (int)
1885 Disable intent logging replay. Can be disabled for recovery from corrupted
1888 Use \fB1\fR for yes and \fB0\fR for no (default).
1894 \fBzil_slog_limit\fR (ulong)
1897 Max commit bytes to separate log device
1899 Default value: \fB1,048,576\fR.
1905 \fBzio_delay_max\fR (int)
1908 A zevent will be logged if a ZIO operation takes more than N milliseconds to
1909 complete. Note that this is only a logging facility, not a timeout on
1912 Default value: \fB30,000\fR.
1918 \fBzio_dva_throttle_enabled\fR (int)
1921 Throttle block allocations in the ZIO pipeline. This allows for
1922 dynamic allocation distribution when devices are imbalanced.
1924 Default value: \fB1\fR.
1930 \fBzio_requeue_io_start_cut_in_line\fR (int)
1933 Prioritize requeued I/O
1935 Default value: \fB0\fR.
1941 \fBzio_taskq_batch_pct\fR (uint)
1944 Percentage of online CPUs (or CPU cores, etc) which will run a worker thread
1945 for IO. These workers are responsible for IO work such as compression and
1946 checksum calculations. Fractional number of CPUs will be rounded down.
1948 The default value of 75 was chosen to avoid using all CPUs which can result in
1949 latency issues and inconsistent application performance, especially when high
1950 compression is enabled.
1952 Default value: \fB75\fR.
1958 \fBzvol_inhibit_dev\fR (uint)
1961 Do not create zvol device nodes. This may slightly improve startup time on
1962 systems with a very large number of zvols.
1964 Use \fB1\fR for yes and \fB0\fR for no (default).
1970 \fBzvol_major\fR (uint)
1973 Major number for zvol block devices
1975 Default value: \fB230\fR.
1981 \fBzvol_max_discard_blocks\fR (ulong)
1984 Discard (aka TRIM) operations done on zvols will be done in batches of this
1985 many blocks, where block size is determined by the \fBvolblocksize\fR property
1988 Default value: \fB16,384\fR.
1994 \fBzvol_prefetch_bytes\fR (uint)
1997 When adding a zvol to the system prefetch \fBzvol_prefetch_bytes\fR
1998 from the start and end of the volume. Prefetching these regions
1999 of the volume is desirable because they are likely to be accessed
2000 immediately by \fBblkid(8)\fR or by the kernel scanning for a partition
2003 Default value: \fB131,072\fR.
2006 .SH ZFS I/O SCHEDULER
2007 ZFS issues I/O operations to leaf vdevs to satisfy and complete I/Os.
2008 The I/O scheduler determines when and in what order those operations are
2009 issued. The I/O scheduler divides operations into five I/O classes
2010 prioritized in the following order: sync read, sync write, async read,
2011 async write, and scrub/resilver. Each queue defines the minimum and
2012 maximum number of concurrent operations that may be issued to the
2013 device. In addition, the device has an aggregate maximum,
2014 \fBzfs_vdev_max_active\fR. Note that the sum of the per-queue minimums
2015 must not exceed the aggregate maximum. If the sum of the per-queue
2016 maximums exceeds the aggregate maximum, then the number of active I/Os
2017 may reach \fBzfs_vdev_max_active\fR, in which case no further I/Os will
2018 be issued regardless of whether all per-queue minimums have been met.
2020 For many physical devices, throughput increases with the number of
2021 concurrent operations, but latency typically suffers. Further, physical
2022 devices typically have a limit at which more concurrent operations have no
2023 effect on throughput or can actually cause it to decrease.
2025 The scheduler selects the next operation to issue by first looking for an
2026 I/O class whose minimum has not been satisfied. Once all are satisfied and
2027 the aggregate maximum has not been hit, the scheduler looks for classes
2028 whose maximum has not been satisfied. Iteration through the I/O classes is
2029 done in the order specified above. No further operations are issued if the
2030 aggregate maximum number of concurrent operations has been hit or if there
2031 are no operations queued for an I/O class that has not hit its maximum.
2032 Every time an I/O is queued or an operation completes, the I/O scheduler
2033 looks for new operations to issue.
2035 In general, smaller max_active's will lead to lower latency of synchronous
2036 operations. Larger max_active's may lead to higher overall throughput,
2037 depending on underlying storage.
2039 The ratio of the queues' max_actives determines the balance of performance
2040 between reads, writes, and scrubs. E.g., increasing
2041 \fBzfs_vdev_scrub_max_active\fR will cause the scrub or resilver to complete
2042 more quickly, but reads and writes to have higher latency and lower throughput.
2044 All I/O classes have a fixed maximum number of outstanding operations
2045 except for the async write class. Asynchronous writes represent the data
2046 that is committed to stable storage during the syncing stage for
2047 transaction groups. Transaction groups enter the syncing state
2048 periodically so the number of queued async writes will quickly burst up
2049 and then bleed down to zero. Rather than servicing them as quickly as
2050 possible, the I/O scheduler changes the maximum number of active async
2051 write I/Os according to the amount of dirty data in the pool. Since
2052 both throughput and latency typically increase with the number of
2053 concurrent operations issued to physical devices, reducing the
2054 burstiness in the number of concurrent operations also stabilizes the
2055 response time of operations from other -- and in particular synchronous
2056 -- queues. In broad strokes, the I/O scheduler will issue more
2057 concurrent operations from the async write queue as there's more dirty
2062 The number of concurrent operations issued for the async write I/O class
2063 follows a piece-wise linear function defined by a few adjustable points.
2066 | o---------| <-- zfs_vdev_async_write_max_active
2073 |-------o | | <-- zfs_vdev_async_write_min_active
2074 0|_______^______|_________|
2075 0% | | 100% of zfs_dirty_data_max
2077 | `-- zfs_vdev_async_write_active_max_dirty_percent
2078 `--------- zfs_vdev_async_write_active_min_dirty_percent
2081 Until the amount of dirty data exceeds a minimum percentage of the dirty
2082 data allowed in the pool, the I/O scheduler will limit the number of
2083 concurrent operations to the minimum. As that threshold is crossed, the
2084 number of concurrent operations issued increases linearly to the maximum at
2085 the specified maximum percentage of the dirty data allowed in the pool.
2087 Ideally, the amount of dirty data on a busy pool will stay in the sloped
2088 part of the function between \fBzfs_vdev_async_write_active_min_dirty_percent\fR
2089 and \fBzfs_vdev_async_write_active_max_dirty_percent\fR. If it exceeds the
2090 maximum percentage, this indicates that the rate of incoming data is
2091 greater than the rate that the backend storage can handle. In this case, we
2092 must further throttle incoming writes, as described in the next section.
2094 .SH ZFS TRANSACTION DELAY
2095 We delay transactions when we've determined that the backend storage
2096 isn't able to accommodate the rate of incoming writes.
2098 If there is already a transaction waiting, we delay relative to when
2099 that transaction will finish waiting. This way the calculated delay time
2100 is independent of the number of threads concurrently executing
2103 If we are the only waiter, wait relative to when the transaction
2104 started, rather than the current time. This credits the transaction for
2105 "time already served", e.g. reading indirect blocks.
2107 The minimum time for a transaction to take is calculated as:
2109 min_time = zfs_delay_scale * (dirty - min) / (max - dirty)
2110 min_time is then capped at 100 milliseconds.
2113 The delay has two degrees of freedom that can be adjusted via tunables. The
2114 percentage of dirty data at which we start to delay is defined by
2115 \fBzfs_delay_min_dirty_percent\fR. This should typically be at or above
2116 \fBzfs_vdev_async_write_active_max_dirty_percent\fR so that we only start to
2117 delay after writing at full speed has failed to keep up with the incoming write
2118 rate. The scale of the curve is defined by \fBzfs_delay_scale\fR. Roughly speaking,
2119 this variable determines the amount of delay at the midpoint of the curve.
2123 10ms +-------------------------------------------------------------*+
2139 2ms + (midpoint) * +
2142 | zfs_delay_scale ----------> ******** |
2143 0 +-------------------------------------*********----------------+
2144 0% <- zfs_dirty_data_max -> 100%
2147 Note that since the delay is added to the outstanding time remaining on the
2148 most recent transaction, the delay is effectively the inverse of IOPS.
2149 Here the midpoint of 500us translates to 2000 IOPS. The shape of the curve
2150 was chosen such that small changes in the amount of accumulated dirty data
2151 in the first 3/4 of the curve yield relatively small differences in the
2154 The effects can be easier to understand when the amount of delay is
2155 represented on a log scale:
2159 100ms +-------------------------------------------------------------++
2168 + zfs_delay_scale ----------> ***** +
2179 +--------------------------------------------------------------+
2180 0% <- zfs_dirty_data_max -> 100%
2183 Note here that only as the amount of dirty data approaches its limit does
2184 the delay start to increase rapidly. The goal of a properly tuned system
2185 should be to keep the amount of dirty data out of that range by first
2186 ensuring that the appropriate limits are set for the I/O scheduler to reach
2187 optimal throughput on the backend storage, and then by changing the value
2188 of \fBzfs_delay_scale\fR to increase the steepness of the curve.