client: support --status-interval option in client/server mode
[fio.git] / HOWTO
CommitLineData
f80dba8d
MT
1How fio works
2-------------
3
4The first step in getting fio to simulate a desired I/O workload, is writing a
5job file describing that specific setup. A job file may contain any number of
6threads and/or files -- the typical contents of the job file is a *global*
7section defining shared parameters, and one or more job sections describing the
8jobs involved. When run, fio parses this file and sets everything up as
9described. If we break down a job from top to bottom, it contains the following
10basic parameters:
11
12`I/O type`_
13
14 Defines the I/O pattern issued to the file(s). We may only be reading
15 sequentially from this file(s), or we may be writing randomly. Or even
16 mixing reads and writes, sequentially or randomly.
17 Should we be doing buffered I/O, or direct/raw I/O?
18
19`Block size`_
20
21 In how large chunks are we issuing I/O? This may be a single value,
22 or it may describe a range of block sizes.
23
24`I/O size`_
25
26 How much data are we going to be reading/writing.
27
28`I/O engine`_
29
30 How do we issue I/O? We could be memory mapping the file, we could be
31 using regular read/write, we could be using splice, async I/O, or even
32 SG (SCSI generic sg).
33
34`I/O depth`_
35
36 If the I/O engine is async, how large a queuing depth do we want to
37 maintain?
38
39
40`Target file/device`_
41
42 How many files are we spreading the workload over.
43
44`Threads, processes and job synchronization`_
45
46 How many threads or processes should we spread this workload over.
47
48The above are the basic parameters defined for a workload, in addition there's a
49multitude of parameters that modify other aspects of how this job behaves.
50
51
52Command line options
53--------------------
54
55.. option:: --debug=type
56
f50fbdda 57 Enable verbose tracing `type` of various fio actions. May be ``all`` for all types
b034c0dd
TK
58 or individual types separated by a comma (e.g. ``--debug=file,mem`` will
59 enable file and memory debugging). Currently, additional logging is
60 available for:
f80dba8d 61
b034c0dd 62 *process*
f80dba8d 63 Dump info related to processes.
b034c0dd 64 *file*
f80dba8d 65 Dump info related to file actions.
b034c0dd 66 *io*
f80dba8d 67 Dump info related to I/O queuing.
b034c0dd 68 *mem*
f80dba8d 69 Dump info related to memory allocations.
b034c0dd 70 *blktrace*
f80dba8d 71 Dump info related to blktrace setup.
b034c0dd 72 *verify*
f80dba8d 73 Dump info related to I/O verification.
b034c0dd 74 *all*
f80dba8d 75 Enable all debug options.
b034c0dd 76 *random*
f80dba8d 77 Dump info related to random offset generation.
b034c0dd 78 *parse*
f80dba8d 79 Dump info related to option matching and parsing.
b034c0dd 80 *diskutil*
f80dba8d 81 Dump info related to disk utilization updates.
b034c0dd 82 *job:x*
f80dba8d 83 Dump info only related to job number x.
b034c0dd 84 *mutex*
f80dba8d 85 Dump info only related to mutex up/down ops.
b034c0dd 86 *profile*
f80dba8d 87 Dump info related to profile extensions.
b034c0dd 88 *time*
f80dba8d 89 Dump info related to internal time keeping.
b034c0dd 90 *net*
f80dba8d 91 Dump info related to networking connections.
b034c0dd 92 *rate*
f80dba8d 93 Dump info related to I/O rate switching.
b034c0dd 94 *compress*
f80dba8d 95 Dump info related to log compress/decompress.
b034c0dd 96 *?* or *help*
f80dba8d
MT
97 Show available debug options.
98
99.. option:: --parse-only
100
25cd4b95 101 Parse options only, don't start any I/O.
f80dba8d
MT
102
103.. option:: --output=filename
104
105 Write output to file `filename`.
106
f50fbdda 107.. option:: --output-format=format
b8f7e412 108
f50fbdda 109 Set the reporting `format` to `normal`, `terse`, `json`, or `json+`. Multiple
b8f7e412
TK
110 formats can be selected, separated by a comma. `terse` is a CSV based
111 format. `json+` is like `json`, except it adds a full dump of the latency
112 buckets.
113
f80dba8d
MT
114.. option:: --bandwidth-log
115
116 Generate aggregate bandwidth logs.
117
118.. option:: --minimal
119
120 Print statistics in a terse, semicolon-delimited format.
121
122.. option:: --append-terse
123
b034c0dd
TK
124 Print statistics in selected mode AND terse, semicolon-delimited format.
125 **Deprecated**, use :option:`--output-format` instead to select multiple
126 formats.
f80dba8d 127
f50fbdda 128.. option:: --terse-version=version
f80dba8d 129
f50fbdda 130 Set terse `version` output format (default 3, or 2 or 4 or 5).
f80dba8d
MT
131
132.. option:: --version
133
b8f7e412 134 Print version information and exit.
f80dba8d
MT
135
136.. option:: --help
137
113f0e7c 138 Print a summary of the command line options and exit.
f80dba8d
MT
139
140.. option:: --cpuclock-test
141
142 Perform test and validation of internal CPU clock.
143
113f0e7c 144.. option:: --crctest=[test]
f80dba8d 145
b034c0dd
TK
146 Test the speed of the built-in checksumming functions. If no argument is
147 given, all of them are tested. Alternatively, a comma separated list can
148 be passed, in which case the given ones are tested.
f80dba8d
MT
149
150.. option:: --cmdhelp=command
151
152 Print help information for `command`. May be ``all`` for all commands.
153
154.. option:: --enghelp=[ioengine[,command]]
155
f50fbdda
TK
156 List all commands defined by `ioengine`, or print help for `command`
157 defined by `ioengine`. If no `ioengine` is given, list all
b034c0dd 158 available ioengines.
f80dba8d
MT
159
160.. option:: --showcmd=jobfile
161
b8f7e412 162 Convert `jobfile` to a set of command-line options.
f80dba8d
MT
163
164.. option:: --readonly
165
4027b2a1
VF
166 Turn on safety read-only checks, preventing writes and trims. The
167 ``--readonly`` option is an extra safety guard to prevent users from
168 accidentally starting a write or trim workload when that is not desired.
169 Fio will only modify the device under test if
170 `rw=write/randwrite/rw/randrw/trim/randtrim/trimwrite` is given. This
171 safety net can be used as an extra precaution.
f80dba8d
MT
172
173.. option:: --eta=when
174
b8f7e412 175 Specifies when real-time ETA estimate should be printed. `when` may be
db37d890
JA
176 `always`, `never` or `auto`. `auto` is the default, it prints ETA
177 when requested if the output is a TTY. `always` disregards the output
178 type, and prints ETA when requested. `never` never prints ETA.
179
180.. option:: --eta-interval=time
181
182 By default, fio requests client ETA status roughly every second. With
183 this option, the interval is configurable. Fio imposes a minimum
184 allowed time to avoid flooding the console, less than 250 msec is
185 not supported.
f80dba8d
MT
186
187.. option:: --eta-newline=time
188
947e0fe0
SW
189 Force a new line for every `time` period passed. When the unit is omitted,
190 the value is interpreted in seconds.
f80dba8d
MT
191
192.. option:: --status-interval=time
193
aa6cb459
VF
194 Force a full status dump of cumulative (from job start) values at `time`
195 intervals. This option does *not* provide per-period measurements. So
196 values such as bandwidth are running averages. When the time unit is omitted,
197 `time` is interpreted in seconds.
f80dba8d
MT
198
199.. option:: --section=name
200
b034c0dd
TK
201 Only run specified section `name` in job file. Multiple sections can be specified.
202 The ``--section`` option allows one to combine related jobs into one file.
203 E.g. one job file could define light, moderate, and heavy sections. Tell
204 fio to run only the "heavy" section by giving ``--section=heavy``
205 command line option. One can also specify the "write" operations in one
206 section and "verify" operation in another section. The ``--section`` option
207 only applies to job sections. The reserved *global* section is always
208 parsed and used.
f80dba8d
MT
209
210.. option:: --alloc-size=kb
211
b034c0dd
TK
212 Set the internal smalloc pool size to `kb` in KiB. The
213 ``--alloc-size`` switch allows one to use a larger pool size for smalloc.
214 If running large jobs with randommap enabled, fio can run out of memory.
215 Smalloc is an internal allocator for shared structures from a fixed size
216 memory pool and can grow to 16 pools. The pool size defaults to 16MiB.
f80dba8d 217
b034c0dd
TK
218 NOTE: While running :file:`.fio_smalloc.*` backing store files are visible
219 in :file:`/tmp`.
f80dba8d
MT
220
221.. option:: --warnings-fatal
222
b034c0dd
TK
223 All fio parser warnings are fatal, causing fio to exit with an
224 error.
f80dba8d
MT
225
226.. option:: --max-jobs=nr
227
f50fbdda 228 Set the maximum number of threads/processes to support to `nr`.
818322cc 229 NOTE: On Linux, it may be necessary to increase the shared-memory
71aa48eb 230 limit (:file:`/proc/sys/kernel/shmmax`) if fio runs into errors while
818322cc 231 creating jobs.
f80dba8d
MT
232
233.. option:: --server=args
234
b034c0dd
TK
235 Start a backend server, with `args` specifying what to listen to.
236 See `Client/Server`_ section.
f80dba8d
MT
237
238.. option:: --daemonize=pidfile
239
b034c0dd 240 Background a fio server, writing the pid to the given `pidfile` file.
f80dba8d
MT
241
242.. option:: --client=hostname
243
f50fbdda 244 Instead of running the jobs locally, send and run them on the given `hostname`
71aa48eb 245 or set of `hostname`\s. See `Client/Server`_ section.
f80dba8d
MT
246
247.. option:: --remote-config=file
248
f50fbdda 249 Tell fio server to load this local `file`.
f80dba8d
MT
250
251.. option:: --idle-prof=option
252
b8f7e412 253 Report CPU idleness. `option` is one of the following:
113f0e7c
SW
254
255 **calibrate**
256 Run unit work calibration only and exit.
257
258 **system**
259 Show aggregate system idleness and unit work.
260
261 **percpu**
262 As **system** but also show per CPU idleness.
f80dba8d
MT
263
264.. option:: --inflate-log=log
265
f50fbdda 266 Inflate and output compressed `log`.
f80dba8d
MT
267
268.. option:: --trigger-file=file
269
f50fbdda 270 Execute trigger command when `file` exists.
f80dba8d 271
f50fbdda 272.. option:: --trigger-timeout=time
f80dba8d 273
f50fbdda 274 Execute trigger at this `time`.
f80dba8d 275
f50fbdda 276.. option:: --trigger=command
f80dba8d 277
f50fbdda 278 Set this `command` as local trigger.
f80dba8d 279
f50fbdda 280.. option:: --trigger-remote=command
f80dba8d 281
f50fbdda 282 Set this `command` as remote trigger.
f80dba8d
MT
283
284.. option:: --aux-path=path
285
f50fbdda 286 Use this `path` for fio state generated files.
f80dba8d
MT
287
288Any parameters following the options will be assumed to be job files, unless
289they match a job file parameter. Multiple job files can be listed and each job
290file will be regarded as a separate group. Fio will :option:`stonewall`
291execution between each group.
292
293
294Job file format
295---------------
296
297As previously described, fio accepts one or more job files describing what it is
298supposed to do. The job file format is the classic ini file, where the names
c60ebc45 299enclosed in [] brackets define the job name. You are free to use any ASCII name
f80dba8d
MT
300you want, except *global* which has special meaning. Following the job name is
301a sequence of zero or more parameters, one per line, that define the behavior of
302the job. If the first character in a line is a ';' or a '#', the entire line is
303discarded as a comment.
304
305A *global* section sets defaults for the jobs described in that file. A job may
306override a *global* section parameter, and a job file may even have several
307*global* sections if so desired. A job is only affected by a *global* section
308residing above it.
309
f50fbdda
TK
310The :option:`--cmdhelp` option also lists all options. If used with a `command`
311argument, :option:`--cmdhelp` will detail the given `command`.
f80dba8d
MT
312
313See the `examples/` directory for inspiration on how to write job files. Note
314the copyright and license requirements currently apply to `examples/` files.
315
316So let's look at a really simple job file that defines two processes, each
317randomly reading from a 128MiB file:
318
319.. code-block:: ini
320
321 ; -- start job file --
322 [global]
323 rw=randread
324 size=128m
325
326 [job1]
327
328 [job2]
329
330 ; -- end job file --
331
332As you can see, the job file sections themselves are empty as all the described
333parameters are shared. As no :option:`filename` option is given, fio makes up a
334`filename` for each of the jobs as it sees fit. On the command line, this job
335would look as follows::
336
337$ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
338
339
340Let's look at an example that has a number of processes writing randomly to
341files:
342
343.. code-block:: ini
344
345 ; -- start job file --
346 [random-writers]
347 ioengine=libaio
348 iodepth=4
349 rw=randwrite
350 bs=32k
351 direct=0
352 size=64m
353 numjobs=4
354 ; -- end job file --
355
356Here we have no *global* section, as we only have one job defined anyway. We
357want to use async I/O here, with a depth of 4 for each file. We also increased
358the buffer size used to 32KiB and define numjobs to 4 to fork 4 identical
359jobs. The result is 4 processes each randomly writing to their own 64MiB
360file. Instead of using the above job file, you could have given the parameters
361on the command line. For this case, you would specify::
362
363$ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
364
365When fio is utilized as a basis of any reasonably large test suite, it might be
366desirable to share a set of standardized settings across multiple job files.
367Instead of copy/pasting such settings, any section may pull in an external
368:file:`filename.fio` file with *include filename* directive, as in the following
369example::
370
371 ; -- start job file including.fio --
372 [global]
373 filename=/tmp/test
374 filesize=1m
375 include glob-include.fio
376
377 [test]
378 rw=randread
379 bs=4k
380 time_based=1
381 runtime=10
382 include test-include.fio
383 ; -- end job file including.fio --
384
385.. code-block:: ini
386
387 ; -- start job file glob-include.fio --
388 thread=1
389 group_reporting=1
390 ; -- end job file glob-include.fio --
391
392.. code-block:: ini
393
394 ; -- start job file test-include.fio --
395 ioengine=libaio
396 iodepth=4
397 ; -- end job file test-include.fio --
398
399Settings pulled into a section apply to that section only (except *global*
400section). Include directives may be nested in that any included file may contain
401further include directive(s). Include files may not contain [] sections.
402
403
404Environment variables
405~~~~~~~~~~~~~~~~~~~~~
406
407Fio also supports environment variable expansion in job files. Any sub-string of
408the form ``${VARNAME}`` as part of an option value (in other words, on the right
409of the '='), will be expanded to the value of the environment variable called
410`VARNAME`. If no such environment variable is defined, or `VARNAME` is the
411empty string, the empty string will be substituted.
412
413As an example, let's look at a sample fio invocation and job file::
414
415$ SIZE=64m NUMJOBS=4 fio jobfile.fio
416
417.. code-block:: ini
418
419 ; -- start job file --
420 [random-writers]
421 rw=randwrite
422 size=${SIZE}
423 numjobs=${NUMJOBS}
424 ; -- end job file --
425
426This will expand to the following equivalent job file at runtime:
427
428.. code-block:: ini
429
430 ; -- start job file --
431 [random-writers]
432 rw=randwrite
433 size=64m
434 numjobs=4
435 ; -- end job file --
436
437Fio ships with a few example job files, you can also look there for inspiration.
438
439Reserved keywords
440~~~~~~~~~~~~~~~~~
441
442Additionally, fio has a set of reserved keywords that will be replaced
443internally with the appropriate value. Those keywords are:
444
445**$pagesize**
446
447 The architecture page size of the running system.
448
449**$mb_memory**
450
451 Megabytes of total memory in the system.
452
453**$ncpus**
454
455 Number of online available CPUs.
456
457These can be used on the command line or in the job file, and will be
458automatically substituted with the current system values when the job is
459run. Simple math is also supported on these keywords, so you can perform actions
460like::
461
b034c0dd 462 size=8*$mb_memory
f80dba8d
MT
463
464and get that properly expanded to 8 times the size of memory in the machine.
465
466
467Job file parameters
468-------------------
469
470This section describes in details each parameter associated with a job. Some
471parameters take an option of a given type, such as an integer or a
472string. Anywhere a numeric value is required, an arithmetic expression may be
473used, provided it is surrounded by parentheses. Supported operators are:
474
475 - addition (+)
476 - subtraction (-)
477 - multiplication (*)
478 - division (/)
479 - modulus (%)
480 - exponentiation (^)
481
482For time values in expressions, units are microseconds by default. This is
483different than for time values not in expressions (not enclosed in
484parentheses). The following types are used:
485
486
487Parameter types
488~~~~~~~~~~~~~~~
489
490**str**
b034c0dd 491 String: A sequence of alphanumeric characters.
f80dba8d
MT
492
493**time**
008d0feb
SW
494 Integer with possible time suffix. Without a unit value is interpreted as
495 seconds unless otherwise specified. Accepts a suffix of 'd' for days, 'h' for
496 hours, 'm' for minutes, 's' for seconds, 'ms' (or 'msec') for milliseconds and
497 'us' (or 'usec') for microseconds. For example, use 10m for 10 minutes.
f80dba8d
MT
498
499.. _int:
500
501**int**
502 Integer. A whole number value, which may contain an integer prefix
503 and an integer suffix:
504
b034c0dd 505 [*integer prefix*] **number** [*integer suffix*]
f80dba8d
MT
506
507 The optional *integer prefix* specifies the number's base. The default
508 is decimal. *0x* specifies hexadecimal.
509
510 The optional *integer suffix* specifies the number's units, and includes an
511 optional unit prefix and an optional unit. For quantities of data, the
947e0fe0
SW
512 default unit is bytes. For quantities of time, the default unit is seconds
513 unless otherwise specified.
f80dba8d 514
9207a0cb 515 With :option:`kb_base`\=1000, fio follows international standards for unit
f80dba8d
MT
516 prefixes. To specify power-of-10 decimal values defined in the
517 International System of Units (SI):
518
eccce61a
TK
519 * *K* -- means kilo (K) or 1000
520 * *M* -- means mega (M) or 1000**2
521 * *G* -- means giga (G) or 1000**3
522 * *T* -- means tera (T) or 1000**4
523 * *P* -- means peta (P) or 1000**5
f80dba8d
MT
524
525 To specify power-of-2 binary values defined in IEC 80000-13:
526
eccce61a
TK
527 * *Ki* -- means kibi (Ki) or 1024
528 * *Mi* -- means mebi (Mi) or 1024**2
529 * *Gi* -- means gibi (Gi) or 1024**3
530 * *Ti* -- means tebi (Ti) or 1024**4
531 * *Pi* -- means pebi (Pi) or 1024**5
f80dba8d 532
9207a0cb 533 With :option:`kb_base`\=1024 (the default), the unit prefixes are opposite
f80dba8d
MT
534 from those specified in the SI and IEC 80000-13 standards to provide
535 compatibility with old scripts. For example, 4k means 4096.
536
537 For quantities of data, an optional unit of 'B' may be included
b8f7e412 538 (e.g., 'kB' is the same as 'k').
f80dba8d
MT
539
540 The *integer suffix* is not case sensitive (e.g., m/mi mean mebi/mega,
541 not milli). 'b' and 'B' both mean byte, not bit.
542
9207a0cb 543 Examples with :option:`kb_base`\=1000:
f80dba8d
MT
544
545 * *4 KiB*: 4096, 4096b, 4096B, 4ki, 4kib, 4kiB, 4Ki, 4KiB
546 * *1 MiB*: 1048576, 1mi, 1024ki
547 * *1 MB*: 1000000, 1m, 1000k
548 * *1 TiB*: 1099511627776, 1ti, 1024gi, 1048576mi
549 * *1 TB*: 1000000000, 1t, 1000m, 1000000k
550
9207a0cb 551 Examples with :option:`kb_base`\=1024 (default):
f80dba8d
MT
552
553 * *4 KiB*: 4096, 4096b, 4096B, 4k, 4kb, 4kB, 4K, 4KB
554 * *1 MiB*: 1048576, 1m, 1024k
555 * *1 MB*: 1000000, 1mi, 1000ki
556 * *1 TiB*: 1099511627776, 1t, 1024g, 1048576m
557 * *1 TB*: 1000000000, 1ti, 1000mi, 1000000ki
558
559 To specify times (units are not case sensitive):
560
561 * *D* -- means days
562 * *H* -- means hours
4502cb42 563 * *M* -- means minutes
f80dba8d
MT
564 * *s* -- or sec means seconds (default)
565 * *ms* -- or *msec* means milliseconds
566 * *us* -- or *usec* means microseconds
567
568 If the option accepts an upper and lower range, use a colon ':' or
569 minus '-' to separate such values. See :ref:`irange <irange>`.
4502cb42
SW
570 If the lower value specified happens to be larger than the upper value
571 the two values are swapped.
f80dba8d
MT
572
573.. _bool:
574
575**bool**
576 Boolean. Usually parsed as an integer, however only defined for
577 true and false (1 and 0).
578
579.. _irange:
580
581**irange**
582 Integer range with suffix. Allows value range to be given, such as
c60ebc45 583 1024-4096. A colon may also be used as the separator, e.g. 1k:4k. If the
f80dba8d
MT
584 option allows two sets of ranges, they can be specified with a ',' or '/'
585 delimiter: 1k-4k/8k-32k. Also see :ref:`int <int>`.
586
587**float_list**
588 A list of floating point numbers, separated by a ':' character.
589
f5c3bcf2
TK
590With the above in mind, here follows the complete list of fio job parameters.
591
f80dba8d
MT
592
593Units
594~~~~~
595
596.. option:: kb_base=int
597
598 Select the interpretation of unit prefixes in input parameters.
599
600 **1000**
601 Inputs comply with IEC 80000-13 and the International
602 System of Units (SI). Use:
603
604 - power-of-2 values with IEC prefixes (e.g., KiB)
605 - power-of-10 values with SI prefixes (e.g., kB)
606
607 **1024**
608 Compatibility mode (default). To avoid breaking old scripts:
609
610 - power-of-2 values with SI prefixes
611 - power-of-10 values with IEC prefixes
612
613 See :option:`bs` for more details on input parameters.
614
615 Outputs always use correct prefixes. Most outputs include both
616 side-by-side, like::
617
618 bw=2383.3kB/s (2327.4KiB/s)
619
620 If only one value is reported, then kb_base selects the one to use:
621
622 **1000** -- SI prefixes
623
624 **1024** -- IEC prefixes
625
626.. option:: unit_base=int
627
628 Base unit for reporting. Allowed values are:
629
630 **0**
631 Use auto-detection (default).
632 **8**
633 Byte based.
634 **1**
635 Bit based.
636
637
f80dba8d
MT
638Job description
639~~~~~~~~~~~~~~~
640
641.. option:: name=str
642
643 ASCII name of the job. This may be used to override the name printed by fio
644 for this job. Otherwise the job name is used. On the command line this
645 parameter has the special purpose of also signaling the start of a new job.
646
647.. option:: description=str
648
649 Text description of the job. Doesn't do anything except dump this text
650 description when this job is run. It's not parsed.
651
652.. option:: loops=int
653
654 Run the specified number of iterations of this job. Used to repeat the same
655 workload a given number of times. Defaults to 1.
656
657.. option:: numjobs=int
658
79591fa9
TK
659 Create the specified number of clones of this job. Each clone of job
660 is spawned as an independent thread or process. May be used to setup a
f80dba8d
MT
661 larger number of threads/processes doing the same thing. Each thread is
662 reported separately; to see statistics for all clones as a whole, use
663 :option:`group_reporting` in conjunction with :option:`new_group`.
a47b697c 664 See :option:`--max-jobs`. Default: 1.
f80dba8d
MT
665
666
667Time related parameters
668~~~~~~~~~~~~~~~~~~~~~~~
669
670.. option:: runtime=time
671
f75ede1d 672 Tell fio to terminate processing after the specified period of time. It
f80dba8d 673 can be quite hard to determine for how long a specified job will run, so
f75ede1d 674 this parameter is handy to cap the total runtime to a given time. When
804c0839 675 the unit is omitted, the value is interpreted in seconds.
f80dba8d
MT
676
677.. option:: time_based
678
679 If set, fio will run for the duration of the :option:`runtime` specified
680 even if the file(s) are completely read or written. It will simply loop over
681 the same workload as many times as the :option:`runtime` allows.
682
a881438b 683.. option:: startdelay=irange(time)
f80dba8d 684
947e0fe0
SW
685 Delay the start of job for the specified amount of time. Can be a single
686 value or a range. When given as a range, each thread will choose a value
687 randomly from within the range. Value is in seconds if a unit is omitted.
f80dba8d
MT
688
689.. option:: ramp_time=time
690
691 If set, fio will run the specified workload for this amount of time before
692 logging any performance numbers. Useful for letting performance settle
693 before logging results, thus minimizing the runtime required for stable
694 results. Note that the ``ramp_time`` is considered lead in time for a job,
695 thus it will increase the total runtime if a special timeout or
f75ede1d
SW
696 :option:`runtime` is specified. When the unit is omitted, the value is
697 given in seconds.
f80dba8d
MT
698
699.. option:: clocksource=str
700
701 Use the given clocksource as the base of timing. The supported options are:
702
703 **gettimeofday**
704 :manpage:`gettimeofday(2)`
705
706 **clock_gettime**
707 :manpage:`clock_gettime(2)`
708
709 **cpu**
710 Internal CPU clock source
711
712 cpu is the preferred clocksource if it is reliable, as it is very fast (and
713 fio is heavy on time calls). Fio will automatically use this clocksource if
714 it's supported and considered reliable on the system it is running on,
715 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
716 means supporting TSC Invariant.
717
718.. option:: gtod_reduce=bool
719
720 Enable all of the :manpage:`gettimeofday(2)` reducing options
f75ede1d 721 (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus
f80dba8d
MT
722 reduce precision of the timeout somewhat to really shrink the
723 :manpage:`gettimeofday(2)` call count. With this option enabled, we only do
724 about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all
725 time keeping was enabled.
726
727.. option:: gtod_cpu=int
728
729 Sometimes it's cheaper to dedicate a single thread of execution to just
730 getting the current time. Fio (and databases, for instance) are very
731 intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set
732 one CPU aside for doing nothing but logging current time to a shared memory
733 location. Then the other threads/processes that run I/O workloads need only
734 copy that segment, instead of entering the kernel with a
735 :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time
736 calls will be excluded from other uses. Fio will manually clear it from the
737 CPU mask of other jobs.
738
739
740Target file/device
741~~~~~~~~~~~~~~~~~~
742
743.. option:: directory=str
744
745 Prefix filenames with this directory. Used to place files in a different
746 location than :file:`./`. You can specify a number of directories by
747 separating the names with a ':' character. These directories will be
02dd2689 748 assigned equally distributed to job clones created by :option:`numjobs` as
f80dba8d
MT
749 long as they are using generated filenames. If specific `filename(s)` are
750 set fio will use the first listed directory, and thereby matching the
751 `filename` semantic which generates a file each clone if not specified, but
752 let all clones use the same if set.
753
02dd2689
SW
754 See the :option:`filename` option for information on how to escape "``:``" and
755 "``\``" characters within the directory path itself.
f80dba8d
MT
756
757.. option:: filename=str
758
759 Fio normally makes up a `filename` based on the job name, thread number, and
02dd2689
SW
760 file number (see :option:`filename_format`). If you want to share files
761 between threads in a job or several
79591fa9
TK
762 jobs with fixed file paths, specify a `filename` for each of them to override
763 the default. If the ioengine is file based, you can specify a number of files
764 by separating the names with a ':' colon. So if you wanted a job to open
765 :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use
766 ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is
767 specified, :option:`nrfiles` is ignored. The size of regular files specified
02dd2689 768 by this option will be :option:`size` divided by number of files unless an
79591fa9
TK
769 explicit size is specified by :option:`filesize`.
770
02dd2689
SW
771 Each colon and backslash in the wanted path must be escaped with a ``\``
772 character. For instance, if the path is :file:`/dev/dsk/foo@3,0:c` then you
773 would use ``filename=/dev/dsk/foo@3,0\:c`` and if the path is
774 :file:`F:\\filename` then you would use ``filename=F\:\\filename``.
775
f80dba8d
MT
776 On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for
777 the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc.
778 Note: Windows and FreeBSD prevent write access to areas
02dd2689
SW
779 of the disk containing in-use data (e.g. filesystems).
780
781 The filename "`-`" is a reserved name, meaning *stdin* or *stdout*. Which
782 of the two depends on the read/write direction set.
f80dba8d
MT
783
784.. option:: filename_format=str
785
786 If sharing multiple files between jobs, it is usually necessary to have fio
787 generate the exact names that you want. By default, fio will name a file
788 based on the default file format specification of
789 :file:`jobname.jobnumber.filenumber`. With this option, that can be
790 customized. Fio will recognize and replace the following keywords in this
791 string:
792
793 **$jobname**
794 The name of the worker thread or process.
795 **$jobnum**
796 The incremental number of the worker thread or process.
797 **$filenum**
798 The incremental number of the file for that worker thread or
799 process.
800
801 To have dependent jobs share a set of files, this option can be set to have
802 fio generate filenames that are shared between the two. For instance, if
803 :file:`testfiles.$filenum` is specified, file number 4 for any job will be
804 named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum`
805 will be used if no other format specifier is given.
806
645943c0
JB
807 If you specify a path then the directories will be created up to the
808 main directory for the file. So for example if you specify
809 ``filename_format=a/b/c/$jobnum`` then the directories a/b/c will be
810 created before the file setup part of the job. If you specify
811 :option:`directory` then the path will be relative that directory,
812 otherwise it is treated as the absolute path.
813
f80dba8d
MT
814.. option:: unique_filename=bool
815
816 To avoid collisions between networked clients, fio defaults to prefixing any
817 generated filenames (with a directory specified) with the source of the
818 client connecting. To disable this behavior, set this option to 0.
819
820.. option:: opendir=str
821
822 Recursively open any files below directory `str`.
823
824.. option:: lockfile=str
825
826 Fio defaults to not locking any files before it does I/O to them. If a file
827 or file descriptor is shared, fio can serialize I/O to that file to make the
828 end result consistent. This is usual for emulating real workloads that share
829 files. The lock modes are:
830
831 **none**
832 No locking. The default.
833 **exclusive**
834 Only one thread or process may do I/O at a time, excluding all
835 others.
836 **readwrite**
837 Read-write locking on the file. Many readers may
838 access the file at the same time, but writes get exclusive access.
839
840.. option:: nrfiles=int
841
79591fa9
TK
842 Number of files to use for this job. Defaults to 1. The size of files
843 will be :option:`size` divided by this unless explicit size is specified by
844 :option:`filesize`. Files are created for each thread separately, and each
845 file will have a file number within its name by default, as explained in
846 :option:`filename` section.
847
f80dba8d
MT
848
849.. option:: openfiles=int
850
851 Number of files to keep open at the same time. Defaults to the same as
852 :option:`nrfiles`, can be set smaller to limit the number simultaneous
853 opens.
854
855.. option:: file_service_type=str
856
857 Defines how fio decides which file from a job to service next. The following
858 types are defined:
859
860 **random**
861 Choose a file at random.
862
863 **roundrobin**
864 Round robin over opened files. This is the default.
865
866 **sequential**
867 Finish one file before moving on to the next. Multiple files can
f50fbdda 868 still be open depending on :option:`openfiles`.
f80dba8d
MT
869
870 **zipf**
c60ebc45 871 Use a *Zipf* distribution to decide what file to access.
f80dba8d
MT
872
873 **pareto**
c60ebc45 874 Use a *Pareto* distribution to decide what file to access.
f80dba8d 875
dd3503d3 876 **normal**
c60ebc45 877 Use a *Gaussian* (normal) distribution to decide what file to
f80dba8d
MT
878 access.
879
dd3503d3
SW
880 **gauss**
881 Alias for normal.
882
f80dba8d
MT
883 For *random*, *roundrobin*, and *sequential*, a postfix can be appended to
884 tell fio how many I/Os to issue before switching to a new file. For example,
885 specifying ``file_service_type=random:8`` would cause fio to issue
886 8 I/Os before selecting a new file at random. For the non-uniform
887 distributions, a floating point postfix can be given to influence how the
888 distribution is skewed. See :option:`random_distribution` for a description
889 of how that would work.
890
891.. option:: ioscheduler=str
892
893 Attempt to switch the device hosting the file to the specified I/O scheduler
894 before running.
895
896.. option:: create_serialize=bool
897
898 If true, serialize the file creation for the jobs. This may be handy to
899 avoid interleaving of data files, which may greatly depend on the filesystem
a47b697c 900 used and even the number of processors in the system. Default: true.
f80dba8d
MT
901
902.. option:: create_fsync=bool
903
22413915 904 :manpage:`fsync(2)` the data file after creation. This is the default.
f80dba8d
MT
905
906.. option:: create_on_open=bool
907
730bd7d9
SW
908 If true, don't pre-create files but allow the job's open() to create a file
909 when it's time to do I/O. Default: false -- pre-create all necessary files
910 when the job starts.
f80dba8d
MT
911
912.. option:: create_only=bool
913
914 If true, fio will only run the setup phase of the job. If files need to be
4502cb42 915 laid out or updated on disk, only that will be done -- the actual job contents
a47b697c 916 are not executed. Default: false.
f80dba8d
MT
917
918.. option:: allow_file_create=bool
919
730bd7d9
SW
920 If true, fio is permitted to create files as part of its workload. If this
921 option is false, then fio will error out if
f80dba8d
MT
922 the files it needs to use don't already exist. Default: true.
923
924.. option:: allow_mounted_write=bool
925
c60ebc45 926 If this isn't set, fio will abort jobs that are destructive (e.g. that write)
f80dba8d
MT
927 to what appears to be a mounted device or partition. This should help catch
928 creating inadvertently destructive tests, not realizing that the test will
b1db0375
TK
929 destroy data on the mounted file system. Note that some platforms don't allow
930 writing against a mounted device regardless of this option. Default: false.
f80dba8d
MT
931
932.. option:: pre_read=bool
933
934 If this is given, files will be pre-read into memory before starting the
935 given I/O operation. This will also clear the :option:`invalidate` flag,
936 since it is pointless to pre-read and then drop the cache. This will only
937 work for I/O engines that are seek-able, since they allow you to read the
a47b697c
SW
938 same data multiple times. Thus it will not work on non-seekable I/O engines
939 (e.g. network, splice). Default: false.
f80dba8d
MT
940
941.. option:: unlink=bool
942
943 Unlink the job files when done. Not the default, as repeated runs of that
a47b697c
SW
944 job would then waste time recreating the file set again and again. Default:
945 false.
f80dba8d
MT
946
947.. option:: unlink_each_loop=bool
948
a47b697c 949 Unlink job files after each iteration or loop. Default: false.
f80dba8d
MT
950
951.. option:: zonesize=int
952
953 Divide a file into zones of the specified size. See :option:`zoneskip`.
954
955.. option:: zonerange=int
956
957 Give size of an I/O zone. See :option:`zoneskip`.
958
959.. option:: zoneskip=int
960
961 Skip the specified number of bytes when :option:`zonesize` data has been
962 read. The two zone options can be used to only do I/O on zones of a file.
963
964
965I/O type
966~~~~~~~~
967
968.. option:: direct=bool
969
970 If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that
8e889110 971 OpenBSD and ZFS on Solaris don't support direct I/O. On Windows the synchronous
f80dba8d
MT
972 ioengines don't support direct I/O. Default: false.
973
974.. option:: atomic=bool
975
976 If value is true, attempt to use atomic direct I/O. Atomic writes are
977 guaranteed to be stable once acknowledged by the operating system. Only
978 Linux supports O_ATOMIC right now.
979
980.. option:: buffered=bool
981
982 If value is true, use buffered I/O. This is the opposite of the
983 :option:`direct` option. Defaults to true.
984
985.. option:: readwrite=str, rw=str
986
987 Type of I/O pattern. Accepted values are:
988
989 **read**
990 Sequential reads.
991 **write**
992 Sequential writes.
993 **trim**
3740cfc8
VF
994 Sequential trims (Linux block devices and SCSI
995 character devices only).
f80dba8d
MT
996 **randread**
997 Random reads.
2831be97
SW
998 **randwrite**
999 Random writes.
f80dba8d 1000 **randtrim**
3740cfc8
VF
1001 Random trims (Linux block devices and SCSI
1002 character devices only).
f80dba8d
MT
1003 **rw,readwrite**
1004 Sequential mixed reads and writes.
1005 **randrw**
1006 Random mixed reads and writes.
1007 **trimwrite**
1008 Sequential trim+write sequences. Blocks will be trimmed first,
1009 then the same blocks will be written to.
1010
1011 Fio defaults to read if the option is not specified. For the mixed I/O
1012 types, the default is to split them 50/50. For certain types of I/O the
730bd7d9
SW
1013 result may still be skewed a bit, since the speed may be different.
1014
1015 It is possible to specify the number of I/Os to do before getting a new
1016 offset by appending ``:<nr>`` to the end of the string given. For a
f80dba8d
MT
1017 random read, it would look like ``rw=randread:8`` for passing in an offset
1018 modifier with a value of 8. If the suffix is used with a sequential I/O
730bd7d9
SW
1019 pattern, then the *<nr>* value specified will be **added** to the generated
1020 offset for each I/O turning sequential I/O into sequential I/O with holes.
1021 For instance, using ``rw=write:4k`` will skip 4k for every write. Also see
1022 the :option:`rw_sequencer` option.
f80dba8d
MT
1023
1024.. option:: rw_sequencer=str
1025
1026 If an offset modifier is given by appending a number to the ``rw=<str>``
1027 line, then this option controls how that number modifies the I/O offset
1028 being generated. Accepted values are:
1029
1030 **sequential**
1031 Generate sequential offset.
1032 **identical**
1033 Generate the same offset.
1034
1035 ``sequential`` is only useful for random I/O, where fio would normally
c60ebc45 1036 generate a new random offset for every I/O. If you append e.g. 8 to randread,
007c7be9
SW
1037 you would get a new random offset for every 8 I/Os. The result would be a
1038 seek for only every 8 I/Os, instead of for every I/O. Use ``rw=randread:8``
f80dba8d
MT
1039 to specify that. As sequential I/O is already sequential, setting
1040 ``sequential`` for that would not result in any differences. ``identical``
1041 behaves in a similar fashion, except it sends the same offset 8 number of
1042 times before generating a new offset.
1043
1044.. option:: unified_rw_reporting=bool
1045
1046 Fio normally reports statistics on a per data direction basis, meaning that
1047 reads, writes, and trims are accounted and reported separately. If this
1048 option is set fio sums the results and report them as "mixed" instead.
1049
1050.. option:: randrepeat=bool
1051
1052 Seed the random number generator used for random I/O patterns in a
1053 predictable way so the pattern is repeatable across runs. Default: true.
1054
1055.. option:: allrandrepeat=bool
1056
1057 Seed all random number generators in a predictable way so results are
1058 repeatable across runs. Default: false.
1059
1060.. option:: randseed=int
1061
1062 Seed the random number generators based on this seed value, to be able to
1063 control what sequence of output is being generated. If not set, the random
1064 sequence depends on the :option:`randrepeat` setting.
1065
1066.. option:: fallocate=str
1067
1068 Whether pre-allocation is performed when laying down files.
1069 Accepted values are:
1070
1071 **none**
1072 Do not pre-allocate space.
1073
2c3e17be
SW
1074 **native**
1075 Use a platform's native pre-allocation call but fall back to
1076 **none** behavior if it fails/is not implemented.
1077
f80dba8d
MT
1078 **posix**
1079 Pre-allocate via :manpage:`posix_fallocate(3)`.
1080
1081 **keep**
1082 Pre-allocate via :manpage:`fallocate(2)` with
1083 FALLOC_FL_KEEP_SIZE set.
1084
1085 **0**
1086 Backward-compatible alias for **none**.
1087
1088 **1**
1089 Backward-compatible alias for **posix**.
1090
1091 May not be available on all supported platforms. **keep** is only available
2c3e17be
SW
1092 on Linux. If using ZFS on Solaris this cannot be set to **posix**
1093 because ZFS doesn't support pre-allocation. Default: **native** if any
1094 pre-allocation methods are available, **none** if not.
f80dba8d
MT
1095
1096.. option:: fadvise_hint=str
1097
c712c97a
JA
1098 Use :manpage:`posix_fadvise(2)` or :manpage:`posix_fadvise(2)` to
1099 advise the kernel on what I/O patterns are likely to be issued.
1100 Accepted values are:
f80dba8d
MT
1101
1102 **0**
1103 Backwards-compatible hint for "no hint".
1104
1105 **1**
1106 Backwards compatible hint for "advise with fio workload type". This
1107 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1108 for a sequential workload.
1109
1110 **sequential**
1111 Advise using **FADV_SEQUENTIAL**.
1112
1113 **random**
1114 Advise using **FADV_RANDOM**.
1115
8f4b9f24 1116.. option:: write_hint=str
f80dba8d 1117
8f4b9f24
JA
1118 Use :manpage:`fcntl(2)` to advise the kernel what life time to expect
1119 from a write. Only supported on Linux, as of version 4.13. Accepted
1120 values are:
1121
1122 **none**
1123 No particular life time associated with this file.
1124
1125 **short**
1126 Data written to this file has a short life time.
1127
1128 **medium**
1129 Data written to this file has a medium life time.
1130
1131 **long**
1132 Data written to this file has a long life time.
1133
1134 **extreme**
1135 Data written to this file has a very long life time.
1136
1137 The values are all relative to each other, and no absolute meaning
1138 should be associated with them.
f80dba8d
MT
1139
1140.. option:: offset=int
1141
82dbb8cb 1142 Start I/O at the provided offset in the file, given as either a fixed size in
83c8b093
JF
1143 bytes or a percentage. If a percentage is given, the generated offset will be
1144 aligned to the minimum ``blocksize`` or to the value of ``offset_align`` if
1145 provided. Data before the given offset will not be touched. This
89978a6b
BW
1146 effectively caps the file size at `real_size - offset`. Can be combined with
1147 :option:`size` to constrain the start and end range of the I/O workload.
44bb1142
TK
1148 A percentage can be specified by a number between 1 and 100 followed by '%',
1149 for example, ``offset=20%`` to specify 20%.
f80dba8d 1150
83c8b093
JF
1151.. option:: offset_align=int
1152
1153 If set to non-zero value, the byte offset generated by a percentage ``offset``
1154 is aligned upwards to this value. Defaults to 0 meaning that a percentage
1155 offset is aligned to the minimum block size.
1156
f80dba8d
MT
1157.. option:: offset_increment=int
1158
1159 If this is provided, then the real offset becomes `offset + offset_increment
1160 * thread_number`, where the thread number is a counter that starts at 0 and
1161 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1162 specified). This option is useful if there are several jobs which are
1163 intended to operate on a file in parallel disjoint segments, with even
1164 spacing between the starting points.
1165
1166.. option:: number_ios=int
1167
c60ebc45 1168 Fio will normally perform I/Os until it has exhausted the size of the region
f80dba8d
MT
1169 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1170 condition). With this setting, the range/size can be set independently of
c60ebc45 1171 the number of I/Os to perform. When fio reaches this number, it will exit
f80dba8d
MT
1172 normally and report status. Note that this does not extend the amount of I/O
1173 that will be done, it will only stop fio if this condition is met before
1174 other end-of-job criteria.
1175
1176.. option:: fsync=int
1177
730bd7d9
SW
1178 If writing to a file, issue an :manpage:`fsync(2)` (or its equivalent) of
1179 the dirty data for every number of blocks given. For example, if you give 32
1180 as a parameter, fio will sync the file after every 32 writes issued. If fio is
1181 using non-buffered I/O, we may not sync the file. The exception is the sg
1182 I/O engine, which synchronizes the disk cache anyway. Defaults to 0, which
1183 means fio does not periodically issue and wait for a sync to complete. Also
1184 see :option:`end_fsync` and :option:`fsync_on_close`.
f80dba8d
MT
1185
1186.. option:: fdatasync=int
1187
1188 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
000a5f1c 1189 not metadata blocks. In Windows, FreeBSD, and DragonFlyBSD there is no
730bd7d9
SW
1190 :manpage:`fdatasync(2)` so this falls back to using :manpage:`fsync(2)`.
1191 Defaults to 0, which means fio does not periodically issue and wait for a
1192 data-only sync to complete.
f80dba8d
MT
1193
1194.. option:: write_barrier=int
1195
2831be97 1196 Make every `N-th` write a barrier write.
f80dba8d 1197
f50fbdda 1198.. option:: sync_file_range=str:int
f80dba8d 1199
f50fbdda 1200 Use :manpage:`sync_file_range(2)` for every `int` number of write
f80dba8d
MT
1201 operations. Fio will track range of writes that have happened since the last
1202 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1203
1204 **wait_before**
1205 SYNC_FILE_RANGE_WAIT_BEFORE
1206 **write**
1207 SYNC_FILE_RANGE_WRITE
1208 **wait_after**
1209 SYNC_FILE_RANGE_WAIT_AFTER
1210
1211 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1212 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1213 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1214 Linux specific.
1215
1216.. option:: overwrite=bool
1217
1218 If true, writes to a file will always overwrite existing data. If the file
1219 doesn't already exist, it will be created before the write phase begins. If
1220 the file exists and is large enough for the specified write phase, nothing
a47b697c 1221 will be done. Default: false.
f80dba8d
MT
1222
1223.. option:: end_fsync=bool
1224
a47b697c
SW
1225 If true, :manpage:`fsync(2)` file contents when a write stage has completed.
1226 Default: false.
f80dba8d
MT
1227
1228.. option:: fsync_on_close=bool
1229
1230 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
a47b697c
SW
1231 from :option:`end_fsync` in that it will happen on every file close, not
1232 just at the end of the job. Default: false.
f80dba8d
MT
1233
1234.. option:: rwmixread=int
1235
1236 Percentage of a mixed workload that should be reads. Default: 50.
1237
1238.. option:: rwmixwrite=int
1239
1240 Percentage of a mixed workload that should be writes. If both
1241 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1242 add up to 100%, the latter of the two will be used to override the
1243 first. This may interfere with a given rate setting, if fio is asked to
1244 limit reads or writes to a certain rate. If that is the case, then the
1245 distribution may be skewed. Default: 50.
1246
1247.. option:: random_distribution=str:float[,str:float][,str:float]
1248
1249 By default, fio will use a completely uniform random distribution when asked
1250 to perform random I/O. Sometimes it is useful to skew the distribution in
1251 specific ways, ensuring that some parts of the data is more hot than others.
1252 fio includes the following distribution models:
1253
1254 **random**
1255 Uniform random distribution
1256
1257 **zipf**
1258 Zipf distribution
1259
1260 **pareto**
1261 Pareto distribution
1262
b2f4b559 1263 **normal**
c60ebc45 1264 Normal (Gaussian) distribution
f80dba8d
MT
1265
1266 **zoned**
1267 Zoned random distribution
1268
59466396
JA
1269 **zoned_abs**
1270 Zone absolute random distribution
1271
f80dba8d 1272 When using a **zipf** or **pareto** distribution, an input value is also
f50fbdda 1273 needed to define the access pattern. For **zipf**, this is the `Zipf
c60ebc45 1274 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
f50fbdda 1275 program, :command:`fio-genzipf`, that can be used visualize what the given input
f80dba8d
MT
1276 values will yield in terms of hit rates. If you wanted to use **zipf** with
1277 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1278 option. If a non-uniform model is used, fio will disable use of the random
b2f4b559
SW
1279 map. For the **normal** distribution, a normal (Gaussian) deviation is
1280 supplied as a value between 0 and 100.
f80dba8d
MT
1281
1282 For a **zoned** distribution, fio supports specifying percentages of I/O
1283 access that should fall within what range of the file or device. For
1284 example, given a criteria of:
1285
f50fbdda
TK
1286 * 60% of accesses should be to the first 10%
1287 * 30% of accesses should be to the next 20%
1288 * 8% of accesses should be to the next 30%
1289 * 2% of accesses should be to the next 40%
f80dba8d
MT
1290
1291 we can define that through zoning of the random accesses. For the above
1292 example, the user would do::
1293
1294 random_distribution=zoned:60/10:30/20:8/30:2/40
1295
59466396
JA
1296 A **zoned_abs** distribution works exactly like the **zoned**, except
1297 that it takes absolute sizes. For example, let's say you wanted to
1298 define access according to the following criteria:
1299
1300 * 60% of accesses should be to the first 20G
1301 * 30% of accesses should be to the next 100G
1302 * 10% of accesses should be to the next 500G
1303
1304 we can define an absolute zoning distribution with:
1305
1306 random_distribution=zoned_abs=60/20G:30/100G:10/500g
1307
6a16ece8
JA
1308 For both **zoned** and **zoned_abs**, fio supports defining up to
1309 256 separate zones.
1310
59466396
JA
1311 Similarly to how :option:`bssplit` works for setting ranges and
1312 percentages of block sizes. Like :option:`bssplit`, it's possible to
1313 specify separate zones for reads, writes, and trims. If just one set
1314 is given, it'll apply to all of them. This goes for both **zoned**
1315 **zoned_abs** distributions.
f80dba8d
MT
1316
1317.. option:: percentage_random=int[,int][,int]
1318
1319 For a random workload, set how big a percentage should be random. This
1320 defaults to 100%, in which case the workload is fully random. It can be set
1321 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1322 sequential. Any setting in between will result in a random mix of sequential
1323 and random I/O, at the given percentages. Comma-separated values may be
1324 specified for reads, writes, and trims as described in :option:`blocksize`.
1325
1326.. option:: norandommap
1327
1328 Normally fio will cover every block of the file when doing random I/O. If
1329 this option is given, fio will just get a new random offset without looking
1330 at past I/O history. This means that some blocks may not be read or written,
1331 and that some blocks may be read/written more than once. If this option is
1332 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1333 only intact blocks are verified, i.e., partially-overwritten blocks are
47e6a6e5
SW
1334 ignored. With an async I/O engine and an I/O depth > 1, it is possible for
1335 the same block to be overwritten, which can cause verification errors. Either
1336 do not use norandommap in this case, or also use the lfsr random generator.
f80dba8d
MT
1337
1338.. option:: softrandommap=bool
1339
1340 See :option:`norandommap`. If fio runs with the random block map enabled and
1341 it fails to allocate the map, if this option is set it will continue without
1342 a random block map. As coverage will not be as complete as with random maps,
1343 this option is disabled by default.
1344
1345.. option:: random_generator=str
1346
f50fbdda 1347 Fio supports the following engines for generating I/O offsets for random I/O:
f80dba8d
MT
1348
1349 **tausworthe**
f50fbdda 1350 Strong 2^88 cycle random number generator.
f80dba8d 1351 **lfsr**
f50fbdda 1352 Linear feedback shift register generator.
f80dba8d 1353 **tausworthe64**
f50fbdda 1354 Strong 64-bit 2^258 cycle random number generator.
f80dba8d
MT
1355
1356 **tausworthe** is a strong random number generator, but it requires tracking
1357 on the side if we want to ensure that blocks are only read or written
f50fbdda 1358 once. **lfsr** guarantees that we never generate the same offset twice, and
f80dba8d 1359 it's also less computationally expensive. It's not a true random generator,
f50fbdda 1360 however, though for I/O purposes it's typically good enough. **lfsr** only
f80dba8d
MT
1361 works with single block sizes, not with workloads that use multiple block
1362 sizes. If used with such a workload, fio may read or write some blocks
1363 multiple times. The default value is **tausworthe**, unless the required
1364 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1365 selected automatically.
1366
1367
1368Block size
1369~~~~~~~~~~
1370
1371.. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1372
1373 The block size in bytes used for I/O units. Default: 4096. A single value
1374 applies to reads, writes, and trims. Comma-separated values may be
1375 specified for reads, writes, and trims. A value not terminated in a comma
1376 applies to subsequent types.
1377
1378 Examples:
1379
1380 **bs=256k**
1381 means 256k for reads, writes and trims.
1382
1383 **bs=8k,32k**
1384 means 8k for reads, 32k for writes and trims.
1385
1386 **bs=8k,32k,**
1387 means 8k for reads, 32k for writes, and default for trims.
1388
1389 **bs=,8k**
1390 means default for reads, 8k for writes and trims.
1391
1392 **bs=,8k,**
b443ae44 1393 means default for reads, 8k for writes, and default for trims.
f80dba8d
MT
1394
1395.. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1396
1397 A range of block sizes in bytes for I/O units. The issued I/O unit will
1398 always be a multiple of the minimum size, unless
1399 :option:`blocksize_unaligned` is set.
1400
1401 Comma-separated ranges may be specified for reads, writes, and trims as
1402 described in :option:`blocksize`.
1403
1404 Example: ``bsrange=1k-4k,2k-8k``.
1405
1406.. option:: bssplit=str[,str][,str]
1407
6a16ece8
JA
1408 Sometimes you want even finer grained control of the block sizes
1409 issued, not just an even split between them. This option allows you to
1410 weight various block sizes, so that you are able to define a specific
1411 amount of block sizes issued. The format for this option is::
f80dba8d
MT
1412
1413 bssplit=blocksize/percentage:blocksize/percentage
1414
6a16ece8
JA
1415 for as many block sizes as needed. So if you want to define a workload
1416 that has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would
1417 write::
f80dba8d
MT
1418
1419 bssplit=4k/10:64k/50:32k/40
1420
6a16ece8
JA
1421 Ordering does not matter. If the percentage is left blank, fio will
1422 fill in the remaining values evenly. So a bssplit option like this one::
f80dba8d
MT
1423
1424 bssplit=4k/50:1k/:32k/
1425
6a16ece8
JA
1426 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always
1427 add up to 100, if bssplit is given a range that adds up to more, it
1428 will error out.
f80dba8d
MT
1429
1430 Comma-separated values may be specified for reads, writes, and trims as
1431 described in :option:`blocksize`.
1432
6a16ece8
JA
1433 If you want a workload that has 50% 2k reads and 50% 4k reads, while
1434 having 90% 4k writes and 10% 8k writes, you would specify::
f80dba8d 1435
cf04b906 1436 bssplit=2k/50:4k/50,4k/90:8k/10
f80dba8d 1437
6a16ece8
JA
1438 Fio supports defining up to 64 different weights for each data
1439 direction.
1440
f80dba8d
MT
1441.. option:: blocksize_unaligned, bs_unaligned
1442
1443 If set, fio will issue I/O units with any size within
1444 :option:`blocksize_range`, not just multiples of the minimum size. This
1445 typically won't work with direct I/O, as that normally requires sector
1446 alignment.
1447
589e88b7 1448.. option:: bs_is_seq_rand=bool
f80dba8d
MT
1449
1450 If this option is set, fio will use the normal read,write blocksize settings
1451 as sequential,random blocksize settings instead. Any random read or write
1452 will use the WRITE blocksize settings, and any sequential read or write will
1453 use the READ blocksize settings.
1454
1455.. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1456
1457 Boundary to which fio will align random I/O units. Default:
1458 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1459 I/O, though it usually depends on the hardware block size. This option is
1460 mutually exclusive with using a random map for files, so it will turn off
1461 that option. Comma-separated values may be specified for reads, writes, and
1462 trims as described in :option:`blocksize`.
1463
1464
1465Buffers and memory
1466~~~~~~~~~~~~~~~~~~
1467
1468.. option:: zero_buffers
1469
1470 Initialize buffers with all zeros. Default: fill buffers with random data.
1471
1472.. option:: refill_buffers
1473
1474 If this option is given, fio will refill the I/O buffers on every
72592780
SW
1475 submit. Only makes sense if :option:`zero_buffers` isn't specified,
1476 naturally. Defaults to being unset i.e., the buffer is only filled at
1477 init time and the data in it is reused when possible but if any of
1478 :option:`verify`, :option:`buffer_compress_percentage` or
1479 :option:`dedupe_percentage` are enabled then `refill_buffers` is also
1480 automatically enabled.
f80dba8d
MT
1481
1482.. option:: scramble_buffers=bool
1483
1484 If :option:`refill_buffers` is too costly and the target is using data
1485 deduplication, then setting this option will slightly modify the I/O buffer
1486 contents to defeat normal de-dupe attempts. This is not enough to defeat
1487 more clever block compression attempts, but it will stop naive dedupe of
1488 blocks. Default: true.
1489
1490.. option:: buffer_compress_percentage=int
1491
72592780
SW
1492 If this is set, then fio will attempt to provide I/O buffer content
1493 (on WRITEs) that compresses to the specified level. Fio does this by
1494 providing a mix of random data followed by fixed pattern data. The
1495 fixed pattern is either zeros, or the pattern specified by
1496 :option:`buffer_pattern`. If the `buffer_pattern` option is used, it
1497 might skew the compression ratio slightly. Setting
1498 `buffer_compress_percentage` to a value other than 100 will also
1499 enable :option:`refill_buffers` in order to reduce the likelihood that
1500 adjacent blocks are so similar that they over compress when seen
1501 together. See :option:`buffer_compress_chunk` for how to set a finer or
1502 coarser granularity for the random/fixed data region. Defaults to unset
1503 i.e., buffer data will not adhere to any compression level.
f80dba8d
MT
1504
1505.. option:: buffer_compress_chunk=int
1506
72592780
SW
1507 This setting allows fio to manage how big the random/fixed data region
1508 is when using :option:`buffer_compress_percentage`. When
1509 `buffer_compress_chunk` is set to some non-zero value smaller than the
1510 block size, fio can repeat the random/fixed region throughout the I/O
1511 buffer at the specified interval (which particularly useful when
1512 bigger block sizes are used for a job). When set to 0, fio will use a
1513 chunk size that matches the block size resulting in a single
1514 random/fixed region within the I/O buffer. Defaults to 512. When the
1515 unit is omitted, the value is interpreted in bytes.
f80dba8d
MT
1516
1517.. option:: buffer_pattern=str
1518
a1554f65
SB
1519 If set, fio will fill the I/O buffers with this pattern or with the contents
1520 of a file. If not set, the contents of I/O buffers are defined by the other
1521 options related to buffer contents. The setting can be any pattern of bytes,
1522 and can be prefixed with 0x for hex values. It may also be a string, where
1523 the string must then be wrapped with ``""``. Or it may also be a filename,
1524 where the filename must be wrapped with ``''`` in which case the file is
1525 opened and read. Note that not all the file contents will be read if that
1526 would cause the buffers to overflow. So, for example::
1527
1528 buffer_pattern='filename'
1529
1530 or::
f80dba8d
MT
1531
1532 buffer_pattern="abcd"
1533
1534 or::
1535
1536 buffer_pattern=-12
1537
1538 or::
1539
1540 buffer_pattern=0xdeadface
1541
1542 Also you can combine everything together in any order::
1543
a1554f65 1544 buffer_pattern=0xdeadface"abcd"-12'filename'
f80dba8d
MT
1545
1546.. option:: dedupe_percentage=int
1547
1548 If set, fio will generate this percentage of identical buffers when
1549 writing. These buffers will be naturally dedupable. The contents of the
1550 buffers depend on what other buffer compression settings have been set. It's
1551 possible to have the individual buffers either fully compressible, or not at
72592780
SW
1552 all -- this option only controls the distribution of unique buffers. Setting
1553 this option will also enable :option:`refill_buffers` to prevent every buffer
1554 being identical.
f80dba8d
MT
1555
1556.. option:: invalidate=bool
1557
730bd7d9
SW
1558 Invalidate the buffer/page cache parts of the files to be used prior to
1559 starting I/O if the platform and file type support it. Defaults to true.
21c1b29e
TK
1560 This will be ignored if :option:`pre_read` is also specified for the
1561 same job.
f80dba8d
MT
1562
1563.. option:: sync=bool
1564
1565 Use synchronous I/O for buffered writes. For the majority of I/O engines,
1566 this means using O_SYNC. Default: false.
1567
1568.. option:: iomem=str, mem=str
1569
1570 Fio can use various types of memory as the I/O unit buffer. The allowed
1571 values are:
1572
1573 **malloc**
1574 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1575 type.
1576
1577 **shm**
1578 Use shared memory as the buffers. Allocated through
1579 :manpage:`shmget(2)`.
1580
1581 **shmhuge**
1582 Same as shm, but use huge pages as backing.
1583
1584 **mmap**
22413915 1585 Use :manpage:`mmap(2)` to allocate buffers. May either be anonymous memory, or can
f80dba8d
MT
1586 be file backed if a filename is given after the option. The format
1587 is `mem=mmap:/path/to/file`.
1588
1589 **mmaphuge**
1590 Use a memory mapped huge file as the buffer backing. Append filename
1591 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1592
1593 **mmapshared**
1594 Same as mmap, but use a MMAP_SHARED mapping.
1595
03553853
YR
1596 **cudamalloc**
1597 Use GPU memory as the buffers for GPUDirect RDMA benchmark.
f50fbdda 1598 The :option:`ioengine` must be `rdma`.
03553853 1599
f80dba8d
MT
1600 The area allocated is a function of the maximum allowed bs size for the job,
1601 multiplied by the I/O depth given. Note that for **shmhuge** and
1602 **mmaphuge** to work, the system must have free huge pages allocated. This
1603 can normally be checked and set by reading/writing
1604 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1605 is 4MiB in size. So to calculate the number of huge pages you need for a
1606 given job file, add up the I/O depth of all jobs (normally one unless
1607 :option:`iodepth` is used) and multiply by the maximum bs set. Then divide
1608 that number by the huge page size. You can see the size of the huge pages in
1609 :file:`/proc/meminfo`. If no huge pages are allocated by having a non-zero
1610 number in `nr_hugepages`, using **mmaphuge** or **shmhuge** will fail. Also
1611 see :option:`hugepage-size`.
1612
1613 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1614 should point there. So if it's mounted in :file:`/huge`, you would use
1615 `mem=mmaphuge:/huge/somefile`.
1616
f50fbdda 1617.. option:: iomem_align=int, mem_align=int
f80dba8d
MT
1618
1619 This indicates the memory alignment of the I/O memory buffers. Note that
1620 the given alignment is applied to the first I/O unit buffer, if using
1621 :option:`iodepth` the alignment of the following buffers are given by the
1622 :option:`bs` used. In other words, if using a :option:`bs` that is a
1623 multiple of the page sized in the system, all buffers will be aligned to
1624 this value. If using a :option:`bs` that is not page aligned, the alignment
1625 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1626 :option:`bs` used.
1627
1628.. option:: hugepage-size=int
1629
1630 Defines the size of a huge page. Must at least be equal to the system
1631 setting, see :file:`/proc/meminfo`. Defaults to 4MiB. Should probably
1632 always be a multiple of megabytes, so using ``hugepage-size=Xm`` is the
1633 preferred way to set this to avoid setting a non-pow-2 bad value.
1634
1635.. option:: lockmem=int
1636
1637 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1638 simulate a smaller amount of memory. The amount specified is per worker.
1639
1640
1641I/O size
1642~~~~~~~~
1643
1644.. option:: size=int
1645
79591fa9
TK
1646 The total size of file I/O for each thread of this job. Fio will run until
1647 this many bytes has been transferred, unless runtime is limited by other options
1648 (such as :option:`runtime`, for instance, or increased/decreased by :option:`io_size`).
1649 Fio will divide this size between the available files determined by options
1650 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1651 specified by the job. If the result of division happens to be 0, the size is
c4aa2d08 1652 set to the physical size of the given files or devices if they exist.
79591fa9 1653 If this option is not specified, fio will use the full size of the given
f80dba8d
MT
1654 files or devices. If the files do not exist, size must be given. It is also
1655 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1656 given, fio will use 20% of the full size of the given files or devices.
9d25d068
SW
1657 Can be combined with :option:`offset` to constrain the start and end range
1658 that I/O will be done within.
f80dba8d
MT
1659
1660.. option:: io_size=int, io_limit=int
1661
1662 Normally fio operates within the region set by :option:`size`, which means
1663 that the :option:`size` option sets both the region and size of I/O to be
1664 performed. Sometimes that is not what you want. With this option, it is
1665 possible to define just the amount of I/O that fio should do. For instance,
1666 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1667 will perform I/O within the first 20GiB but exit when 5GiB have been
1668 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1669 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1670 the 0..20GiB region.
1671
7fdd97ca 1672.. option:: filesize=irange(int)
f80dba8d
MT
1673
1674 Individual file sizes. May be a range, in which case fio will select sizes
1675 for files at random within the given range and limited to :option:`size` in
1676 total (if that is given). If not given, each created file is the same size.
79591fa9
TK
1677 This option overrides :option:`size` in terms of file size, which means
1678 this value is used as a fixed size or possible range of each file.
f80dba8d
MT
1679
1680.. option:: file_append=bool
1681
1682 Perform I/O after the end of the file. Normally fio will operate within the
1683 size of a file. If this option is set, then fio will append to the file
1684 instead. This has identical behavior to setting :option:`offset` to the size
1685 of a file. This option is ignored on non-regular files.
1686
1687.. option:: fill_device=bool, fill_fs=bool
1688
1689 Sets size to something really large and waits for ENOSPC (no space left on
1690 device) as the terminating condition. Only makes sense with sequential
1691 write. For a read workload, the mount point will be filled first then I/O
1692 started on the result. This option doesn't make sense if operating on a raw
1693 device node, since the size of that is already known by the file system.
1694 Additionally, writing beyond end-of-device will not return ENOSPC there.
1695
1696
1697I/O engine
1698~~~~~~~~~~
1699
1700.. option:: ioengine=str
1701
1702 Defines how the job issues I/O to the file. The following types are defined:
1703
1704 **sync**
1705 Basic :manpage:`read(2)` or :manpage:`write(2)`
1706 I/O. :manpage:`lseek(2)` is used to position the I/O location.
54227e6b 1707 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
f80dba8d
MT
1708
1709 **psync**
1710 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1711 all supported operating systems except for Windows.
1712
1713 **vsync**
1714 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
c60ebc45 1715 queuing by coalescing adjacent I/Os into a single submission.
f80dba8d
MT
1716
1717 **pvsync**
1718 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1719
1720 **pvsync2**
1721 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1722
1723 **libaio**
1724 Linux native asynchronous I/O. Note that Linux may only support
22413915 1725 queued behavior with non-buffered I/O (set ``direct=1`` or
f80dba8d
MT
1726 ``buffered=0``).
1727 This engine defines engine specific options.
1728
1729 **posixaio**
1730 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
1731 :manpage:`aio_write(3)`.
1732
1733 **solarisaio**
1734 Solaris native asynchronous I/O.
1735
1736 **windowsaio**
1737 Windows native asynchronous I/O. Default on Windows.
1738
1739 **mmap**
1740 File is memory mapped with :manpage:`mmap(2)` and data copied
1741 to/from using :manpage:`memcpy(3)`.
1742
1743 **splice**
1744 :manpage:`splice(2)` is used to transfer the data and
1745 :manpage:`vmsplice(2)` to transfer data from user space to the
1746 kernel.
1747
1748 **sg**
1749 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
1750 ioctl, or if the target is an sg character device we use
1751 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
f50fbdda 1752 I/O. Requires :option:`filename` option to specify either block or
3740cfc8 1753 character devices. This engine supports trim operations.
52b81b7c 1754 The sg engine includes engine specific options.
f80dba8d
MT
1755
1756 **null**
1757 Doesn't transfer any data, just pretends to. This is mainly used to
1758 exercise fio itself and for debugging/testing purposes.
1759
1760 **net**
1761 Transfer over the network to given ``host:port``. Depending on the
1762 :option:`protocol` used, the :option:`hostname`, :option:`port`,
1763 :option:`listen` and :option:`filename` options are used to specify
1764 what sort of connection to make, while the :option:`protocol` option
1765 determines which protocol will be used. This engine defines engine
1766 specific options.
1767
1768 **netsplice**
1769 Like **net**, but uses :manpage:`splice(2)` and
1770 :manpage:`vmsplice(2)` to map data and send/receive.
1771 This engine defines engine specific options.
1772
1773 **cpuio**
1774 Doesn't transfer any data, but burns CPU cycles according to the
1775 :option:`cpuload` and :option:`cpuchunks` options. Setting
9207a0cb 1776 :option:`cpuload`\=85 will cause that job to do nothing but burn 85%
71aa48eb 1777 of the CPU. In case of SMP machines, use :option:`numjobs`\=<nr_of_cpu>
f50fbdda 1778 to get desired CPU usage, as the cpuload only loads a
f80dba8d
MT
1779 single CPU at the desired rate. A job never finishes unless there is
1780 at least one non-cpuio job.
1781
1782 **guasi**
804c0839 1783 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall
f80dba8d
MT
1784 Interface approach to async I/O. See
1785
1786 http://www.xmailserver.org/guasi-lib.html
1787
1788 for more info on GUASI.
1789
1790 **rdma**
1791 The RDMA I/O engine supports both RDMA memory semantics
1792 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
609ac152
SB
1793 InfiniBand, RoCE and iWARP protocols. This engine defines engine
1794 specific options.
f80dba8d
MT
1795
1796 **falloc**
1797 I/O engine that does regular fallocate to simulate data transfer as
1798 fio ioengine.
1799
1800 DDIR_READ
1801 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
1802
1803 DDIR_WRITE
1804 does fallocate(,mode = 0).
1805
1806 DDIR_TRIM
1807 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
1808
761cd093
SW
1809 **ftruncate**
1810 I/O engine that sends :manpage:`ftruncate(2)` operations in response
1811 to write (DDIR_WRITE) events. Each ftruncate issued sets the file's
f50fbdda 1812 size to the current block offset. :option:`blocksize` is ignored.
761cd093 1813
f80dba8d
MT
1814 **e4defrag**
1815 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
1816 defragment activity in request to DDIR_WRITE event.
1817
f3f96717
IF
1818 **rados**
1819 I/O engine supporting direct access to Ceph Reliable Autonomic
1820 Distributed Object Store (RADOS) via librados. This ioengine
1821 defines engine specific options.
1822
f80dba8d
MT
1823 **rbd**
1824 I/O engine supporting direct access to Ceph Rados Block Devices
1825 (RBD) via librbd without the need to use the kernel rbd driver. This
1826 ioengine defines engine specific options.
1827
1828 **gfapi**
ac8ca2af
SW
1829 Using GlusterFS libgfapi sync interface to direct access to
1830 GlusterFS volumes without having to go through FUSE. This ioengine
f80dba8d
MT
1831 defines engine specific options.
1832
1833 **gfapi_async**
ac8ca2af
SW
1834 Using GlusterFS libgfapi async interface to direct access to
1835 GlusterFS volumes without having to go through FUSE. This ioengine
f80dba8d
MT
1836 defines engine specific options.
1837
1838 **libhdfs**
f50fbdda 1839 Read and write through Hadoop (HDFS). The :option:`filename` option
f80dba8d
MT
1840 is used to specify host,port of the hdfs name-node to connect. This
1841 engine interprets offsets a little differently. In HDFS, files once
e25c0c91
SW
1842 created cannot be modified so random writes are not possible. To
1843 imitate this the libhdfs engine expects a bunch of small files to be
1844 created over HDFS and will randomly pick a file from them
1845 based on the offset generated by fio backend (see the example
f80dba8d 1846 job file to create such files, use ``rw=write`` option). Please
e25c0c91
SW
1847 note, it may be necessary to set environment variables to work
1848 with HDFS/libhdfs properly. Each job uses its own connection to
f80dba8d
MT
1849 HDFS.
1850
1851 **mtd**
1852 Read, write and erase an MTD character device (e.g.,
1853 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
1854 underlying device type, the I/O may have to go in a certain pattern,
1855 e.g., on NAND, writing sequentially to erase blocks and discarding
c298ee71 1856 before overwriting. The `trimwrite` mode works well for this
f80dba8d
MT
1857 constraint.
1858
1859 **pmemblk**
1860 Read and write using filesystem DAX to a file on a filesystem
363a5f65 1861 mounted with DAX on a persistent memory device through the PMDK
f80dba8d
MT
1862 libpmemblk library.
1863
1864 **dev-dax**
1865 Read and write using device DAX to a persistent memory device (e.g.,
363a5f65 1866 /dev/dax0.0) through the PMDK libpmem library.
f80dba8d
MT
1867
1868 **external**
1869 Prefix to specify loading an external I/O engine object file. Append
c60ebc45 1870 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
d243fd6d
TK
1871 ioengine :file:`foo.o` in :file:`/tmp`. The path can be either
1872 absolute or relative. See :file:`engines/skeleton_external.c` for
1873 details of writing an external I/O engine.
f80dba8d 1874
1216cc5a 1875 **filecreate**
b71968b1 1876 Simply create the files and do no I/O to them. You still need to
1216cc5a 1877 set `filesize` so that all the accounting still occurs, but no
b71968b1 1878 actual I/O will be done other than creating the file.
f80dba8d 1879
ae0db592
TI
1880 **libpmem**
1881 Read and write using mmap I/O to a file on a filesystem
363a5f65 1882 mounted with DAX on a persistent memory device through the PMDK
ae0db592
TI
1883 libpmem library.
1884
f80dba8d
MT
1885I/O engine specific parameters
1886~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1887
1888In addition, there are some parameters which are only valid when a specific
f50fbdda
TK
1889:option:`ioengine` is in use. These are used identically to normal parameters,
1890with the caveat that when used on the command line, they must come after the
f80dba8d
MT
1891:option:`ioengine` that defines them is selected.
1892
1893.. option:: userspace_reap : [libaio]
1894
1895 Normally, with the libaio engine in use, fio will use the
1896 :manpage:`io_getevents(2)` system call to reap newly returned events. With
1897 this flag turned on, the AIO ring will be read directly from user-space to
1898 reap events. The reaping mode is only enabled when polling for a minimum of
c60ebc45 1899 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
f80dba8d 1900
9d25d068 1901.. option:: hipri : [pvsync2]
f80dba8d
MT
1902
1903 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
1904 than normal.
1905
a0679ce5
SB
1906.. option:: hipri_percentage : [pvsync2]
1907
f50fbdda 1908 When hipri is set this determines the probability of a pvsync2 I/O being high
a0679ce5
SB
1909 priority. The default is 100%.
1910
f80dba8d
MT
1911.. option:: cpuload=int : [cpuio]
1912
da19cdb4
TK
1913 Attempt to use the specified percentage of CPU cycles. This is a mandatory
1914 option when using cpuio I/O engine.
f80dba8d
MT
1915
1916.. option:: cpuchunks=int : [cpuio]
1917
1918 Split the load into cycles of the given time. In microseconds.
1919
1920.. option:: exit_on_io_done=bool : [cpuio]
1921
1922 Detect when I/O threads are done, then exit.
1923
f80dba8d
MT
1924.. option:: namenode=str : [libhdfs]
1925
22413915 1926 The hostname or IP address of a HDFS cluster namenode to contact.
f80dba8d
MT
1927
1928.. option:: port=int
1929
f50fbdda
TK
1930 [libhdfs]
1931
1932 The listening port of the HFDS cluster namenode.
1933
f80dba8d
MT
1934 [netsplice], [net]
1935
1936 The TCP or UDP port to bind to or connect to. If this is used with
1937 :option:`numjobs` to spawn multiple instances of the same job type, then
1938 this will be the starting port number since fio will use a range of
1939 ports.
1940
609ac152
SB
1941 [rdma]
1942
1943 The port to use for RDMA-CM communication. This should be the same value
1944 on the client and the server side.
1945
1946.. option:: hostname=str : [netsplice] [net] [rdma]
f80dba8d 1947
609ac152
SB
1948 The hostname or IP address to use for TCP, UDP or RDMA-CM based I/O. If the job
1949 is a TCP listener or UDP reader, the hostname is not used and must be omitted
f50fbdda 1950 unless it is a valid UDP multicast address.
f80dba8d
MT
1951
1952.. option:: interface=str : [netsplice] [net]
1953
1954 The IP address of the network interface used to send or receive UDP
1955 multicast.
1956
1957.. option:: ttl=int : [netsplice] [net]
1958
1959 Time-to-live value for outgoing UDP multicast packets. Default: 1.
1960
1961.. option:: nodelay=bool : [netsplice] [net]
1962
1963 Set TCP_NODELAY on TCP connections.
1964
f50fbdda 1965.. option:: protocol=str, proto=str : [netsplice] [net]
f80dba8d
MT
1966
1967 The network protocol to use. Accepted values are:
1968
1969 **tcp**
1970 Transmission control protocol.
1971 **tcpv6**
1972 Transmission control protocol V6.
1973 **udp**
1974 User datagram protocol.
1975 **udpv6**
1976 User datagram protocol V6.
1977 **unix**
1978 UNIX domain socket.
1979
1980 When the protocol is TCP or UDP, the port must also be given, as well as the
1981 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
f50fbdda 1982 normal :option:`filename` option should be used and the port is invalid.
f80dba8d 1983
e9184ec1 1984.. option:: listen : [netsplice] [net]
f80dba8d
MT
1985
1986 For TCP network connections, tell fio to listen for incoming connections
1987 rather than initiating an outgoing connection. The :option:`hostname` must
1988 be omitted if this option is used.
1989
e9184ec1 1990.. option:: pingpong : [netsplice] [net]
f80dba8d
MT
1991
1992 Normally a network writer will just continue writing data, and a network
1993 reader will just consume packages. If ``pingpong=1`` is set, a writer will
1994 send its normal payload to the reader, then wait for the reader to send the
1995 same payload back. This allows fio to measure network latencies. The
1996 submission and completion latencies then measure local time spent sending or
1997 receiving, and the completion latency measures how long it took for the
1998 other end to receive and send back. For UDP multicast traffic
1999 ``pingpong=1`` should only be set for a single reader when multiple readers
2000 are listening to the same address.
2001
e9184ec1 2002.. option:: window_size : [netsplice] [net]
f80dba8d
MT
2003
2004 Set the desired socket buffer size for the connection.
2005
e9184ec1 2006.. option:: mss : [netsplice] [net]
f80dba8d
MT
2007
2008 Set the TCP maximum segment size (TCP_MAXSEG).
2009
2010.. option:: donorname=str : [e4defrag]
2011
730bd7d9 2012 File will be used as a block donor (swap extents between files).
f80dba8d
MT
2013
2014.. option:: inplace=int : [e4defrag]
2015
2016 Configure donor file blocks allocation strategy:
2017
2018 **0**
2019 Default. Preallocate donor's file on init.
2020 **1**
2b455dbf 2021 Allocate space immediately inside defragment event, and free right
f80dba8d
MT
2022 after event.
2023
f3f96717 2024.. option:: clustername=str : [rbd,rados]
f80dba8d
MT
2025
2026 Specifies the name of the Ceph cluster.
2027
2028.. option:: rbdname=str : [rbd]
2029
2030 Specifies the name of the RBD.
2031
f3f96717 2032.. option:: pool=str : [rbd,rados]
f80dba8d 2033
f3f96717 2034 Specifies the name of the Ceph pool containing RBD or RADOS data.
f80dba8d 2035
f3f96717 2036.. option:: clientname=str : [rbd,rados]
f80dba8d
MT
2037
2038 Specifies the username (without the 'client.' prefix) used to access the
2039 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
2040 the full *type.id* string. If no type. prefix is given, fio will add
2041 'client.' by default.
2042
f3f96717
IF
2043.. option:: busy_poll=bool : [rbd,rados]
2044
2045 Poll store instead of waiting for completion. Usually this provides better
2046 throughput at cost of higher(up to 100%) CPU utilization.
2047
f80dba8d
MT
2048.. option:: skip_bad=bool : [mtd]
2049
2050 Skip operations against known bad blocks.
2051
2052.. option:: hdfsdirectory : [libhdfs]
2053
2054 libhdfs will create chunk in this HDFS directory.
2055
2056.. option:: chunk_size : [libhdfs]
2057
2b455dbf 2058 The size of the chunk to use for each file.
f80dba8d 2059
609ac152
SB
2060.. option:: verb=str : [rdma]
2061
2062 The RDMA verb to use on this side of the RDMA ioengine connection. Valid
2063 values are write, read, send and recv. These correspond to the equivalent
2064 RDMA verbs (e.g. write = rdma_write etc.). Note that this only needs to be
2065 specified on the client side of the connection. See the examples folder.
2066
2067.. option:: bindname=str : [rdma]
2068
2069 The name to use to bind the local RDMA-CM connection to a local RDMA device.
2070 This could be a hostname or an IPv4 or IPv6 address. On the server side this
2071 will be passed into the rdma_bind_addr() function and on the client site it
2072 will be used in the rdma_resolve_add() function. This can be useful when
2073 multiple paths exist between the client and the server or in certain loopback
2074 configurations.
f80dba8d 2075
52b81b7c
KD
2076.. option:: readfua=bool : [sg]
2077
2078 With readfua option set to 1, read operations include
2079 the force unit access (fua) flag. Default is 0.
2080
2081.. option:: writefua=bool : [sg]
2082
2083 With writefua option set to 1, write operations include
2084 the force unit access (fua) flag. Default is 0.
2085
2c3a9150 2086.. option:: sg_write_mode=str : [sg]
3740cfc8 2087
2c3a9150
VF
2088 Specify the type of write commands to issue. This option can take three values:
2089
2090 **write**
2091 This is the default where write opcodes are issued as usual.
2092 **verify**
2093 Issue WRITE AND VERIFY commands. The BYTCHK bit is set to 0. This
2094 directs the device to carry out a medium verification with no data
2095 comparison. The writefua option is ignored with this selection.
2096 **same**
2097 Issue WRITE SAME commands. This transfers a single block to the device
2098 and writes this same block of data to a contiguous sequence of LBAs
2099 beginning at the specified offset. fio's block size parameter specifies
2100 the amount of data written with each command. However, the amount of data
2101 actually transferred to the device is equal to the device's block
2102 (sector) size. For a device with 512 byte sectors, blocksize=8k will
2103 write 16 sectors with each command. fio will still generate 8k of data
2104 for each command but only the first 512 bytes will be used and
2105 transferred to the device. The writefua option is ignored with this
2106 selection.
52b81b7c 2107
f80dba8d
MT
2108I/O depth
2109~~~~~~~~~
2110
2111.. option:: iodepth=int
2112
2113 Number of I/O units to keep in flight against the file. Note that
2114 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
c60ebc45 2115 for small degrees when :option:`verify_async` is in use). Even async
f80dba8d
MT
2116 engines may impose OS restrictions causing the desired depth not to be
2117 achieved. This may happen on Linux when using libaio and not setting
9207a0cb 2118 :option:`direct`\=1, since buffered I/O is not async on that OS. Keep an
f80dba8d
MT
2119 eye on the I/O depth distribution in the fio output to verify that the
2120 achieved depth is as expected. Default: 1.
2121
2122.. option:: iodepth_batch_submit=int, iodepth_batch=int
2123
2124 This defines how many pieces of I/O to submit at once. It defaults to 1
2125 which means that we submit each I/O as soon as it is available, but can be
2126 raised to submit bigger batches of I/O at the time. If it is set to 0 the
2127 :option:`iodepth` value will be used.
2128
2129.. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
2130
2131 This defines how many pieces of I/O to retrieve at once. It defaults to 1
2132 which means that we'll ask for a minimum of 1 I/O in the retrieval process
2133 from the kernel. The I/O retrieval will go on until we hit the limit set by
2134 :option:`iodepth_low`. If this variable is set to 0, then fio will always
2135 check for completed events before queuing more I/O. This helps reduce I/O
2136 latency, at the cost of more retrieval system calls.
2137
2138.. option:: iodepth_batch_complete_max=int
2139
2140 This defines maximum pieces of I/O to retrieve at once. This variable should
9207a0cb 2141 be used along with :option:`iodepth_batch_complete_min`\=int variable,
f80dba8d 2142 specifying the range of min and max amount of I/O which should be
730bd7d9 2143 retrieved. By default it is equal to the :option:`iodepth_batch_complete_min`
f80dba8d
MT
2144 value.
2145
2146 Example #1::
2147
2148 iodepth_batch_complete_min=1
2149 iodepth_batch_complete_max=<iodepth>
2150
2151 which means that we will retrieve at least 1 I/O and up to the whole
2152 submitted queue depth. If none of I/O has been completed yet, we will wait.
2153
2154 Example #2::
2155
2156 iodepth_batch_complete_min=0
2157 iodepth_batch_complete_max=<iodepth>
2158
2159 which means that we can retrieve up to the whole submitted queue depth, but
2160 if none of I/O has been completed yet, we will NOT wait and immediately exit
2161 the system call. In this example we simply do polling.
2162
2163.. option:: iodepth_low=int
2164
2165 The low water mark indicating when to start filling the queue
2166 again. Defaults to the same as :option:`iodepth`, meaning that fio will
2167 attempt to keep the queue full at all times. If :option:`iodepth` is set to
c60ebc45 2168 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
f80dba8d
MT
2169 16 requests, it will let the depth drain down to 4 before starting to fill
2170 it again.
2171
997b5680
SW
2172.. option:: serialize_overlap=bool
2173
2174 Serialize in-flight I/Os that might otherwise cause or suffer from data races.
2175 When two or more I/Os are submitted simultaneously, there is no guarantee that
2176 the I/Os will be processed or completed in the submitted order. Further, if
2177 two or more of those I/Os are writes, any overlapping region between them can
2178 become indeterminate/undefined on certain storage. These issues can cause
2179 verification to fail erratically when at least one of the racing I/Os is
2180 changing data and the overlapping region has a non-zero size. Setting
2181 ``serialize_overlap`` tells fio to avoid provoking this behavior by explicitly
2182 serializing in-flight I/Os that have a non-zero overlap. Note that setting
ee21ebee 2183 this option can reduce both performance and the :option:`iodepth` achieved.
997b5680
SW
2184 Additionally this option does not work when :option:`io_submit_mode` is set to
2185 offload. Default: false.
2186
f80dba8d
MT
2187.. option:: io_submit_mode=str
2188
2189 This option controls how fio submits the I/O to the I/O engine. The default
2190 is `inline`, which means that the fio job threads submit and reap I/O
2191 directly. If set to `offload`, the job threads will offload I/O submission
2192 to a dedicated pool of I/O threads. This requires some coordination and thus
2193 has a bit of extra overhead, especially for lower queue depth I/O where it
2194 can increase latencies. The benefit is that fio can manage submission rates
2195 independently of the device completion rates. This avoids skewed latency
730bd7d9 2196 reporting if I/O gets backed up on the device side (the coordinated omission
f80dba8d
MT
2197 problem).
2198
2199
2200I/O rate
2201~~~~~~~~
2202
a881438b 2203.. option:: thinktime=time
f80dba8d 2204
f75ede1d
SW
2205 Stall the job for the specified period of time after an I/O has completed before issuing the
2206 next. May be used to simulate processing being done by an application.
947e0fe0 2207 When the unit is omitted, the value is interpreted in microseconds. See
f80dba8d
MT
2208 :option:`thinktime_blocks` and :option:`thinktime_spin`.
2209
a881438b 2210.. option:: thinktime_spin=time
f80dba8d
MT
2211
2212 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
2213 something with the data received, before falling back to sleeping for the
f75ede1d 2214 rest of the period specified by :option:`thinktime`. When the unit is
947e0fe0 2215 omitted, the value is interpreted in microseconds.
f80dba8d
MT
2216
2217.. option:: thinktime_blocks=int
2218
2219 Only valid if :option:`thinktime` is set - control how many blocks to issue,
f50fbdda
TK
2220 before waiting :option:`thinktime` usecs. If not set, defaults to 1 which will make
2221 fio wait :option:`thinktime` usecs after every block. This effectively makes any
f80dba8d 2222 queue depth setting redundant, since no more than 1 I/O will be queued
f50fbdda 2223 before we have to complete it and do our :option:`thinktime`. In other words, this
f80dba8d 2224 setting effectively caps the queue depth if the latter is larger.
71bfa161 2225
f80dba8d 2226.. option:: rate=int[,int][,int]
71bfa161 2227
f80dba8d
MT
2228 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
2229 suffix rules apply. Comma-separated values may be specified for reads,
2230 writes, and trims as described in :option:`blocksize`.
71bfa161 2231
b25b3464
SW
2232 For example, using `rate=1m,500k` would limit reads to 1MiB/sec and writes to
2233 500KiB/sec. Capping only reads or writes can be done with `rate=,500k` or
2234 `rate=500k,` where the former will only limit writes (to 500KiB/sec) and the
2235 latter will only limit reads.
2236
f80dba8d 2237.. option:: rate_min=int[,int][,int]
71bfa161 2238
f80dba8d
MT
2239 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
2240 to meet this requirement will cause the job to exit. Comma-separated values
2241 may be specified for reads, writes, and trims as described in
2242 :option:`blocksize`.
71bfa161 2243
f80dba8d 2244.. option:: rate_iops=int[,int][,int]
71bfa161 2245
f80dba8d
MT
2246 Cap the bandwidth to this number of IOPS. Basically the same as
2247 :option:`rate`, just specified independently of bandwidth. If the job is
2248 given a block size range instead of a fixed value, the smallest block size
2249 is used as the metric. Comma-separated values may be specified for reads,
2250 writes, and trims as described in :option:`blocksize`.
71bfa161 2251
f80dba8d 2252.. option:: rate_iops_min=int[,int][,int]
71bfa161 2253
f80dba8d
MT
2254 If fio doesn't meet this rate of I/O, it will cause the job to exit.
2255 Comma-separated values may be specified for reads, writes, and trims as
2256 described in :option:`blocksize`.
71bfa161 2257
f80dba8d 2258.. option:: rate_process=str
66c098b8 2259
f80dba8d
MT
2260 This option controls how fio manages rated I/O submissions. The default is
2261 `linear`, which submits I/O in a linear fashion with fixed delays between
c60ebc45 2262 I/Os that gets adjusted based on I/O completion rates. If this is set to
f80dba8d
MT
2263 `poisson`, fio will submit I/O based on a more real world random request
2264 flow, known as the Poisson process
2265 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
2266 10^6 / IOPS for the given workload.
71bfa161 2267
1a9bf814
JA
2268.. option:: rate_ignore_thinktime=bool
2269
2270 By default, fio will attempt to catch up to the specified rate setting,
2271 if any kind of thinktime setting was used. If this option is set, then
2272 fio will ignore the thinktime and continue doing IO at the specified
2273 rate, instead of entering a catch-up mode after thinktime is done.
2274
71bfa161 2275
f80dba8d
MT
2276I/O latency
2277~~~~~~~~~~~
71bfa161 2278
a881438b 2279.. option:: latency_target=time
71bfa161 2280
f80dba8d 2281 If set, fio will attempt to find the max performance point that the given
f75ede1d 2282 workload will run at while maintaining a latency below this target. When
947e0fe0 2283 the unit is omitted, the value is interpreted in microseconds. See
f75ede1d 2284 :option:`latency_window` and :option:`latency_percentile`.
71bfa161 2285
a881438b 2286.. option:: latency_window=time
71bfa161 2287
f80dba8d 2288 Used with :option:`latency_target` to specify the sample window that the job
f75ede1d 2289 is run at varying queue depths to test the performance. When the unit is
947e0fe0 2290 omitted, the value is interpreted in microseconds.
b4692828 2291
f80dba8d 2292.. option:: latency_percentile=float
71bfa161 2293
c60ebc45 2294 The percentage of I/Os that must fall within the criteria specified by
f80dba8d 2295 :option:`latency_target` and :option:`latency_window`. If not set, this
c60ebc45 2296 defaults to 100.0, meaning that all I/Os must be equal or below to the value
f80dba8d 2297 set by :option:`latency_target`.
71bfa161 2298
a881438b 2299.. option:: max_latency=time
71bfa161 2300
f75ede1d 2301 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
947e0fe0 2302 maximum latency. When the unit is omitted, the value is interpreted in
f75ede1d 2303 microseconds.
71bfa161 2304
f80dba8d 2305.. option:: rate_cycle=int
71bfa161 2306
f80dba8d 2307 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
a47b697c 2308 of milliseconds. Defaults to 1000.
71bfa161 2309
71bfa161 2310
f80dba8d
MT
2311I/O replay
2312~~~~~~~~~~
71bfa161 2313
f80dba8d 2314.. option:: write_iolog=str
c2b1e753 2315
f80dba8d
MT
2316 Write the issued I/O patterns to the specified file. See
2317 :option:`read_iolog`. Specify a separate file for each job, otherwise the
2318 iologs will be interspersed and the file may be corrupt.
c2b1e753 2319
f80dba8d 2320.. option:: read_iolog=str
71bfa161 2321
22413915 2322 Open an iolog with the specified filename and replay the I/O patterns it
f80dba8d
MT
2323 contains. This can be used to store a workload and replay it sometime
2324 later. The iolog given may also be a blktrace binary file, which allows fio
2325 to replay a workload captured by :command:`blktrace`. See
2326 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
2327 replay, the file needs to be turned into a blkparse binary data file first
2328 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
71bfa161 2329
77be374d
AK
2330.. option:: read_iolog_chunked=bool
2331
2332 Determines how iolog is read. If false(default) entire :option:`read_iolog`
2333 will be read at once. If selected true, input from iolog will be read
2334 gradually. Useful when iolog is very large, or it is generated.
2335
589e88b7 2336.. option:: replay_no_stall=bool
71bfa161 2337
f80dba8d 2338 When replaying I/O with :option:`read_iolog` the default behavior is to
22413915 2339 attempt to respect the timestamps within the log and replay them with the
f80dba8d
MT
2340 appropriate delay between IOPS. By setting this variable fio will not
2341 respect the timestamps and attempt to replay them as fast as possible while
2342 still respecting ordering. The result is the same I/O pattern to a given
2343 device, but different timings.
71bfa161 2344
6dd7fa77
JA
2345.. option:: replay_time_scale=int
2346
2347 When replaying I/O with :option:`read_iolog`, fio will honor the
2348 original timing in the trace. With this option, it's possible to scale
2349 the time. It's a percentage option, if set to 50 it means run at 50%
2350 the original IO rate in the trace. If set to 200, run at twice the
2351 original IO rate. Defaults to 100.
2352
f80dba8d 2353.. option:: replay_redirect=str
b4692828 2354
f80dba8d
MT
2355 While replaying I/O patterns using :option:`read_iolog` the default behavior
2356 is to replay the IOPS onto the major/minor device that each IOP was recorded
2357 from. This is sometimes undesirable because on a different machine those
2358 major/minor numbers can map to a different device. Changing hardware on the
2359 same system can also result in a different major/minor mapping.
730bd7d9 2360 ``replay_redirect`` causes all I/Os to be replayed onto the single specified
f80dba8d 2361 device regardless of the device it was recorded
9207a0cb 2362 from. i.e. :option:`replay_redirect`\= :file:`/dev/sdc` would cause all I/O
f80dba8d
MT
2363 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
2364 multiple devices will be replayed onto a single device, if the trace
2365 contains multiple devices. If you want multiple devices to be replayed
2366 concurrently to multiple redirected devices you must blkparse your trace
2367 into separate traces and replay them with independent fio invocations.
2368 Unfortunately this also breaks the strict time ordering between multiple
2369 device accesses.
71bfa161 2370
f80dba8d 2371.. option:: replay_align=int
74929ac2 2372
f80dba8d
MT
2373 Force alignment of I/O offsets and lengths in a trace to this power of 2
2374 value.
3c54bc46 2375
f80dba8d 2376.. option:: replay_scale=int
3c54bc46 2377
f80dba8d 2378 Scale sector offsets down by this factor when replaying traces.
3c54bc46 2379
38f68906
JA
2380.. option:: replay_skip=str
2381
2382 Sometimes it's useful to skip certain IO types in a replay trace.
2383 This could be, for instance, eliminating the writes in the trace.
2384 Or not replaying the trims/discards, if you are redirecting to
2385 a device that doesn't support them. This option takes a comma
2386 separated list of read, write, trim, sync.
2387
3c54bc46 2388
f80dba8d
MT
2389Threads, processes and job synchronization
2390~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3c54bc46 2391
f80dba8d 2392.. option:: thread
3c54bc46 2393
730bd7d9
SW
2394 Fio defaults to creating jobs by using fork, however if this option is
2395 given, fio will create jobs by using POSIX Threads' function
2396 :manpage:`pthread_create(3)` to create threads instead.
71bfa161 2397
f80dba8d 2398.. option:: wait_for=str
74929ac2 2399
730bd7d9
SW
2400 If set, the current job won't be started until all workers of the specified
2401 waitee job are done.
74929ac2 2402
f80dba8d
MT
2403 ``wait_for`` operates on the job name basis, so there are a few
2404 limitations. First, the waitee must be defined prior to the waiter job
2405 (meaning no forward references). Second, if a job is being referenced as a
2406 waitee, it must have a unique name (no duplicate waitees).
74929ac2 2407
f80dba8d 2408.. option:: nice=int
892a6ffc 2409
f80dba8d 2410 Run the job with the given nice value. See man :manpage:`nice(2)`.
892a6ffc 2411
f80dba8d
MT
2412 On Windows, values less than -15 set the process class to "High"; -1 through
2413 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
2414 priority class.
74929ac2 2415
f80dba8d 2416.. option:: prio=int
71bfa161 2417
f80dba8d
MT
2418 Set the I/O priority value of this job. Linux limits us to a positive value
2419 between 0 and 7, with 0 being the highest. See man
2420 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
2421 systems since meaning of priority may differ.
71bfa161 2422
f80dba8d 2423.. option:: prioclass=int
d59aa780 2424
f80dba8d 2425 Set the I/O priority class. See man :manpage:`ionice(1)`.
d59aa780 2426
f80dba8d 2427.. option:: cpus_allowed=str
6d500c2e 2428
730bd7d9 2429 Controls the same options as :option:`cpumask`, but accepts a textual
b570e037
SW
2430 specification of the permitted CPUs instead and CPUs are indexed from 0. So
2431 to use CPUs 0 and 5 you would specify ``cpus_allowed=0,5``. This option also
2432 allows a range of CPUs to be specified -- say you wanted a binding to CPUs
2433 0, 5, and 8 to 15, you would set ``cpus_allowed=0,5,8-15``.
2434
2435 On Windows, when ``cpus_allowed`` is unset only CPUs from fio's current
2436 processor group will be used and affinity settings are inherited from the
2437 system. An fio build configured to target Windows 7 makes options that set
2438 CPUs processor group aware and values will set both the processor group
2439 and a CPU from within that group. For example, on a system where processor
2440 group 0 has 40 CPUs and processor group 1 has 32 CPUs, ``cpus_allowed``
2441 values between 0 and 39 will bind CPUs from processor group 0 and
2442 ``cpus_allowed`` values between 40 and 71 will bind CPUs from processor
2443 group 1. When using ``cpus_allowed_policy=shared`` all CPUs specified by a
2444 single ``cpus_allowed`` option must be from the same processor group. For
2445 Windows fio builds not built for Windows 7, CPUs will only be selected from
2446 (and be relative to) whatever processor group fio happens to be running in
2447 and CPUs from other processor groups cannot be used.
6d500c2e 2448
f80dba8d 2449.. option:: cpus_allowed_policy=str
6d500c2e 2450
f80dba8d 2451 Set the policy of how fio distributes the CPUs specified by
730bd7d9 2452 :option:`cpus_allowed` or :option:`cpumask`. Two policies are supported:
6d500c2e 2453
f80dba8d
MT
2454 **shared**
2455 All jobs will share the CPU set specified.
2456 **split**
2457 Each job will get a unique CPU from the CPU set.
6d500c2e 2458
22413915 2459 **shared** is the default behavior, if the option isn't specified. If
f80dba8d
MT
2460 **split** is specified, then fio will will assign one cpu per job. If not
2461 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
2462 in the set.
6d500c2e 2463
b570e037
SW
2464.. option:: cpumask=int
2465
2466 Set the CPU affinity of this job. The parameter given is a bit mask of
2467 allowed CPUs the job may run on. So if you want the allowed CPUs to be 1
2468 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
2469 :manpage:`sched_setaffinity(2)`. This may not work on all supported
2470 operating systems or kernel versions. This option doesn't work well for a
2471 higher CPU count than what you can store in an integer mask, so it can only
2472 control cpus 1-32. For boxes with larger CPU counts, use
2473 :option:`cpus_allowed`.
2474
f80dba8d 2475.. option:: numa_cpu_nodes=str
6d500c2e 2476
f80dba8d
MT
2477 Set this job running on specified NUMA nodes' CPUs. The arguments allow
2478 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
ac8ca2af 2479 NUMA options support, fio must be built on a system with libnuma-dev(el)
f80dba8d 2480 installed.
61b9861d 2481
f80dba8d 2482.. option:: numa_mem_policy=str
61b9861d 2483
f80dba8d
MT
2484 Set this job's memory policy and corresponding NUMA nodes. Format of the
2485 arguments::
5c94b008 2486
f80dba8d 2487 <mode>[:<nodelist>]
ce35b1ec 2488
804c0839 2489 ``mode`` is one of the following memory policies: ``default``, ``prefer``,
730bd7d9
SW
2490 ``bind``, ``interleave`` or ``local``. For ``default`` and ``local`` memory
2491 policies, no node needs to be specified. For ``prefer``, only one node is
2492 allowed. For ``bind`` and ``interleave`` the ``nodelist`` may be as
2493 follows: a comma delimited list of numbers, A-B ranges, or `all`.
71bfa161 2494
f80dba8d 2495.. option:: cgroup=str
390b1537 2496
f80dba8d
MT
2497 Add job to this control group. If it doesn't exist, it will be created. The
2498 system must have a mounted cgroup blkio mount point for this to work. If
2499 your system doesn't have it mounted, you can do so with::
5af1c6f3 2500
f80dba8d 2501 # mount -t cgroup -o blkio none /cgroup
5af1c6f3 2502
f80dba8d 2503.. option:: cgroup_weight=int
5af1c6f3 2504
f80dba8d
MT
2505 Set the weight of the cgroup to this value. See the documentation that comes
2506 with the kernel, allowed values are in the range of 100..1000.
a086c257 2507
f80dba8d 2508.. option:: cgroup_nodelete=bool
8c07860d 2509
f80dba8d
MT
2510 Normally fio will delete the cgroups it has created after the job
2511 completion. To override this behavior and to leave cgroups around after the
2512 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
2513 to inspect various cgroup files after job completion. Default: false.
8c07860d 2514
f80dba8d 2515.. option:: flow_id=int
8c07860d 2516
f80dba8d
MT
2517 The ID of the flow. If not specified, it defaults to being a global
2518 flow. See :option:`flow`.
1907dbc6 2519
f80dba8d 2520.. option:: flow=int
71bfa161 2521
f80dba8d
MT
2522 Weight in token-based flow control. If this value is used, then there is a
2523 'flow counter' which is used to regulate the proportion of activity between
2524 two or more jobs. Fio attempts to keep this flow counter near zero. The
2525 ``flow`` parameter stands for how much should be added or subtracted to the
2526 flow counter on each iteration of the main I/O loop. That is, if one job has
2527 ``flow=8`` and another job has ``flow=-1``, then there will be a roughly 1:8
2528 ratio in how much one runs vs the other.
71bfa161 2529
f80dba8d 2530.. option:: flow_watermark=int
a31041ea 2531
f80dba8d
MT
2532 The maximum value that the absolute value of the flow counter is allowed to
2533 reach before the job must wait for a lower value of the counter.
82407585 2534
f80dba8d 2535.. option:: flow_sleep=int
82407585 2536
f80dba8d
MT
2537 The period of time, in microseconds, to wait after the flow watermark has
2538 been exceeded before retrying operations.
82407585 2539
f80dba8d 2540.. option:: stonewall, wait_for_previous
82407585 2541
f80dba8d
MT
2542 Wait for preceding jobs in the job file to exit, before starting this
2543 one. Can be used to insert serialization points in the job file. A stone
2544 wall also implies starting a new reporting group, see
2545 :option:`group_reporting`.
2546
2547.. option:: exitall
2548
730bd7d9
SW
2549 By default, fio will continue running all other jobs when one job finishes
2550 but sometimes this is not the desired action. Setting ``exitall`` will
2551 instead make fio terminate all other jobs when one job finishes.
f80dba8d
MT
2552
2553.. option:: exec_prerun=str
2554
2555 Before running this job, issue the command specified through
2556 :manpage:`system(3)`. Output is redirected in a file called
2557 :file:`jobname.prerun.txt`.
2558
2559.. option:: exec_postrun=str
2560
2561 After the job completes, issue the command specified though
2562 :manpage:`system(3)`. Output is redirected in a file called
2563 :file:`jobname.postrun.txt`.
2564
2565.. option:: uid=int
2566
2567 Instead of running as the invoking user, set the user ID to this value
2568 before the thread/process does any work.
2569
2570.. option:: gid=int
2571
2572 Set group ID, see :option:`uid`.
2573
2574
2575Verification
2576~~~~~~~~~~~~
2577
2578.. option:: verify_only
2579
2580 Do not perform specified workload, only verify data still matches previous
2581 invocation of this workload. This option allows one to check data multiple
2582 times at a later date without overwriting it. This option makes sense only
2583 for workloads that write data, and does not support workloads with the
2584 :option:`time_based` option set.
2585
2586.. option:: do_verify=bool
2587
2588 Run the verify phase after a write phase. Only valid if :option:`verify` is
2589 set. Default: true.
2590
2591.. option:: verify=str
2592
2593 If writing to a file, fio can verify the file contents after each iteration
2594 of the job. Each verification method also implies verification of special
2595 header, which is written to the beginning of each block. This header also
2596 includes meta information, like offset of the block, block number, timestamp
2597 when block was written, etc. :option:`verify` can be combined with
2598 :option:`verify_pattern` option. The allowed values are:
2599
2600 **md5**
2601 Use an md5 sum of the data area and store it in the header of
2602 each block.
2603
2604 **crc64**
2605 Use an experimental crc64 sum of the data area and store it in the
2606 header of each block.
2607
2608 **crc32c**
a5896300
SW
2609 Use a crc32c sum of the data area and store it in the header of
2610 each block. This will automatically use hardware acceleration
2611 (e.g. SSE4.2 on an x86 or CRC crypto extensions on ARM64) but will
2612 fall back to software crc32c if none is found. Generally the
804c0839 2613 fastest checksum fio supports when hardware accelerated.
f80dba8d
MT
2614
2615 **crc32c-intel**
a5896300 2616 Synonym for crc32c.
f80dba8d
MT
2617
2618 **crc32**
2619 Use a crc32 sum of the data area and store it in the header of each
2620 block.
2621
2622 **crc16**
2623 Use a crc16 sum of the data area and store it in the header of each
2624 block.
2625
2626 **crc7**
2627 Use a crc7 sum of the data area and store it in the header of each
2628 block.
2629
2630 **xxhash**
2631 Use xxhash as the checksum function. Generally the fastest software
2632 checksum that fio supports.
2633
2634 **sha512**
2635 Use sha512 as the checksum function.
2636
2637 **sha256**
2638 Use sha256 as the checksum function.
2639
2640 **sha1**
2641 Use optimized sha1 as the checksum function.
82407585 2642
ae3a5acc
JA
2643 **sha3-224**
2644 Use optimized sha3-224 as the checksum function.
2645
2646 **sha3-256**
2647 Use optimized sha3-256 as the checksum function.
2648
2649 **sha3-384**
2650 Use optimized sha3-384 as the checksum function.
2651
2652 **sha3-512**
2653 Use optimized sha3-512 as the checksum function.
2654
f80dba8d
MT
2655 **meta**
2656 This option is deprecated, since now meta information is included in
2657 generic verification header and meta verification happens by
2658 default. For detailed information see the description of the
2659 :option:`verify` setting. This option is kept because of
2660 compatibility's sake with old configurations. Do not use it.
2661
2662 **pattern**
2663 Verify a strict pattern. Normally fio includes a header with some
2664 basic information and checksumming, but if this option is set, only
2665 the specific pattern set with :option:`verify_pattern` is verified.
2666
2667 **null**
2668 Only pretend to verify. Useful for testing internals with
9207a0cb 2669 :option:`ioengine`\=null, not for much else.
f80dba8d
MT
2670
2671 This option can be used for repeated burn-in tests of a system to make sure
2672 that the written data is also correctly read back. If the data direction
2673 given is a read or random read, fio will assume that it should verify a
2674 previously written file. If the data direction includes any form of write,
2675 the verify will be of the newly written data.
2676
47e6a6e5
SW
2677 To avoid false verification errors, do not use the norandommap option when
2678 verifying data with async I/O engines and I/O depths > 1. Or use the
2679 norandommap and the lfsr random generator together to avoid writing to the
2680 same offset with muliple outstanding I/Os.
2681
f80dba8d
MT
2682.. option:: verify_offset=int
2683
2684 Swap the verification header with data somewhere else in the block before
2685 writing. It is swapped back before verifying.
2686
2687.. option:: verify_interval=int
2688
2689 Write the verification header at a finer granularity than the
2690 :option:`blocksize`. It will be written for chunks the size of
2691 ``verify_interval``. :option:`blocksize` should divide this evenly.
2692
2693.. option:: verify_pattern=str
2694
2695 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
2696 filling with totally random bytes, but sometimes it's interesting to fill
2697 with a known pattern for I/O verification purposes. Depending on the width
730bd7d9 2698 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time (it can
f80dba8d
MT
2699 be either a decimal or a hex number). The ``verify_pattern`` if larger than
2700 a 32-bit quantity has to be a hex number that starts with either "0x" or
2701 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
2702 format, which means that for each block offset will be written and then
2703 verified back, e.g.::
61b9861d
RP
2704
2705 verify_pattern=%o
2706
f80dba8d
MT
2707 Or use combination of everything::
2708
61b9861d 2709 verify_pattern=0xff%o"abcd"-12
e28218f3 2710
f80dba8d
MT
2711.. option:: verify_fatal=bool
2712
2713 Normally fio will keep checking the entire contents before quitting on a
2714 block verification failure. If this option is set, fio will exit the job on
2715 the first observed failure. Default: false.
2716
2717.. option:: verify_dump=bool
2718
2719 If set, dump the contents of both the original data block and the data block
2720 we read off disk to files. This allows later analysis to inspect just what
2721 kind of data corruption occurred. Off by default.
2722
2723.. option:: verify_async=int
2724
2725 Fio will normally verify I/O inline from the submitting thread. This option
2726 takes an integer describing how many async offload threads to create for I/O
2727 verification instead, causing fio to offload the duty of verifying I/O
2728 contents to one or more separate threads. If using this offload option, even
2729 sync I/O engines can benefit from using an :option:`iodepth` setting higher
2730 than 1, as it allows them to have I/O in flight while verifies are running.
d7e6ea1c 2731 Defaults to 0 async threads, i.e. verification is not asynchronous.
f80dba8d
MT
2732
2733.. option:: verify_async_cpus=str
2734
2735 Tell fio to set the given CPU affinity on the async I/O verification
2736 threads. See :option:`cpus_allowed` for the format used.
2737
2738.. option:: verify_backlog=int
2739
2740 Fio will normally verify the written contents of a job that utilizes verify
2741 once that job has completed. In other words, everything is written then
2742 everything is read back and verified. You may want to verify continually
2743 instead for a variety of reasons. Fio stores the meta data associated with
2744 an I/O block in memory, so for large verify workloads, quite a bit of memory
2745 would be used up holding this meta data. If this option is enabled, fio will
2746 write only N blocks before verifying these blocks.
2747
2748.. option:: verify_backlog_batch=int
2749
2750 Control how many blocks fio will verify if :option:`verify_backlog` is
2751 set. If not set, will default to the value of :option:`verify_backlog`
2752 (meaning the entire queue is read back and verified). If
2753 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
2754 blocks will be verified, if ``verify_backlog_batch`` is larger than
2755 :option:`verify_backlog`, some blocks will be verified more than once.
2756
2757.. option:: verify_state_save=bool
2758
2759 When a job exits during the write phase of a verify workload, save its
2760 current state. This allows fio to replay up until that point, if the verify
2761 state is loaded for the verify read phase. The format of the filename is,
2762 roughly::
2763
f50fbdda 2764 <type>-<jobname>-<jobindex>-verify.state.
f80dba8d
MT
2765
2766 <type> is "local" for a local run, "sock" for a client/server socket
2767 connection, and "ip" (192.168.0.1, for instance) for a networked
d7e6ea1c 2768 client/server connection. Defaults to true.
f80dba8d
MT
2769
2770.. option:: verify_state_load=bool
2771
2772 If a verify termination trigger was used, fio stores the current write state
2773 of each thread. This can be used at verification time so that fio knows how
2774 far it should verify. Without this information, fio will run a full
a47b697c
SW
2775 verification pass, according to the settings in the job file used. Default
2776 false.
f80dba8d
MT
2777
2778.. option:: trim_percentage=int
2779
2780 Number of verify blocks to discard/trim.
2781
2782.. option:: trim_verify_zero=bool
2783
22413915 2784 Verify that trim/discarded blocks are returned as zeros.
f80dba8d
MT
2785
2786.. option:: trim_backlog=int
2787
5cfd1e9a 2788 Trim after this number of blocks are written.
f80dba8d
MT
2789
2790.. option:: trim_backlog_batch=int
2791
2792 Trim this number of I/O blocks.
2793
2794.. option:: experimental_verify=bool
2795
2796 Enable experimental verification.
2797
f80dba8d
MT
2798Steady state
2799~~~~~~~~~~~~
2800
2801.. option:: steadystate=str:float, ss=str:float
2802
2803 Define the criterion and limit for assessing steady state performance. The
2804 first parameter designates the criterion whereas the second parameter sets
2805 the threshold. When the criterion falls below the threshold for the
2806 specified duration, the job will stop. For example, `iops_slope:0.1%` will
2807 direct fio to terminate the job when the least squares regression slope
2808 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
2809 this will apply to all jobs in the group. Below is the list of available
2810 steady state assessment criteria. All assessments are carried out using only
2811 data from the rolling collection window. Threshold limits can be expressed
2812 as a fixed value or as a percentage of the mean in the collection window.
2813
2814 **iops**
2815 Collect IOPS data. Stop the job if all individual IOPS measurements
2816 are within the specified limit of the mean IOPS (e.g., ``iops:2``
2817 means that all individual IOPS values must be within 2 of the mean,
2818 whereas ``iops:0.2%`` means that all individual IOPS values must be
2819 within 0.2% of the mean IOPS to terminate the job).
2820
2821 **iops_slope**
2822 Collect IOPS data and calculate the least squares regression
2823 slope. Stop the job if the slope falls below the specified limit.
2824
2825 **bw**
2826 Collect bandwidth data. Stop the job if all individual bandwidth
2827 measurements are within the specified limit of the mean bandwidth.
2828
2829 **bw_slope**
2830 Collect bandwidth data and calculate the least squares regression
2831 slope. Stop the job if the slope falls below the specified limit.
2832
2833.. option:: steadystate_duration=time, ss_dur=time
2834
2835 A rolling window of this duration will be used to judge whether steady state
2836 has been reached. Data will be collected once per second. The default is 0
f75ede1d 2837 which disables steady state detection. When the unit is omitted, the
947e0fe0 2838 value is interpreted in seconds.
f80dba8d
MT
2839
2840.. option:: steadystate_ramp_time=time, ss_ramp=time
2841
2842 Allow the job to run for the specified duration before beginning data
2843 collection for checking the steady state job termination criterion. The
947e0fe0 2844 default is 0. When the unit is omitted, the value is interpreted in seconds.
f80dba8d
MT
2845
2846
2847Measurements and reporting
2848~~~~~~~~~~~~~~~~~~~~~~~~~~
2849
2850.. option:: per_job_logs=bool
2851
2852 If set, this generates bw/clat/iops log with per file private filenames. If
2853 not set, jobs with identical names will share the log filename. Default:
2854 true.
2855
2856.. option:: group_reporting
2857
2858 It may sometimes be interesting to display statistics for groups of jobs as
2859 a whole instead of for each individual job. This is especially true if
2860 :option:`numjobs` is used; looking at individual thread/process output
2861 quickly becomes unwieldy. To see the final report per-group instead of
2862 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
2863 same reporting group, unless if separated by a :option:`stonewall`, or by
2864 using :option:`new_group`.
2865
2866.. option:: new_group
2867
2868 Start a new reporting group. See: :option:`group_reporting`. If not given,
2869 all jobs in a file will be part of the same reporting group, unless
2870 separated by a :option:`stonewall`.
2871
589e88b7 2872.. option:: stats=bool
8243be59
JA
2873
2874 By default, fio collects and shows final output results for all jobs
2875 that run. If this option is set to 0, then fio will ignore it in
2876 the final stat output.
2877
f80dba8d
MT
2878.. option:: write_bw_log=str
2879
2880 If given, write a bandwidth log for this job. Can be used to store data of
074f0817 2881 the bandwidth of the jobs in their lifetime.
f80dba8d 2882
074f0817
SW
2883 If no str argument is given, the default filename of
2884 :file:`jobname_type.x.log` is used. Even when the argument is given, fio
2885 will still append the type of log. So if one specifies::
2886
2887 write_bw_log=foo
f80dba8d 2888
074f0817
SW
2889 The actual log name will be :file:`foo_bw.x.log` where `x` is the index
2890 of the job (`1..N`, where `N` is the number of jobs). If
2891 :option:`per_job_logs` is false, then the filename will not include the
2892 `.x` job index.
e3cedca7 2893
074f0817
SW
2894 The included :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
2895 text files into nice graphs. See `Log File Formats`_ for how data is
2896 structured within the file.
2897
2898.. option:: write_lat_log=str
e3cedca7 2899
074f0817 2900 Same as :option:`write_bw_log`, except this option creates I/O
77b7e675
SW
2901 submission (e.g., :file:`name_slat.x.log`), completion (e.g.,
2902 :file:`name_clat.x.log`), and total (e.g., :file:`name_lat.x.log`)
074f0817
SW
2903 latency files instead. See :option:`write_bw_log` for details about
2904 the filename format and `Log File Formats`_ for how data is structured
2905 within the files.
be4ecfdf 2906
f80dba8d 2907.. option:: write_hist_log=str
06842027 2908
074f0817 2909 Same as :option:`write_bw_log` but writes an I/O completion latency
77b7e675 2910 histogram file (e.g., :file:`name_hist.x.log`) instead. Note that this
074f0817
SW
2911 file will be empty unless :option:`log_hist_msec` has also been set.
2912 See :option:`write_bw_log` for details about the filename format and
2913 `Log File Formats`_ for how data is structured within the file.
06842027 2914
f80dba8d 2915.. option:: write_iops_log=str
06842027 2916
074f0817 2917 Same as :option:`write_bw_log`, but writes an IOPS file (e.g.
77b7e675 2918 :file:`name_iops.x.log`) instead. See :option:`write_bw_log` for
074f0817
SW
2919 details about the filename format and `Log File Formats`_ for how data
2920 is structured within the file.
06842027 2921
f80dba8d 2922.. option:: log_avg_msec=int
06842027 2923
f80dba8d
MT
2924 By default, fio will log an entry in the iops, latency, or bw log for every
2925 I/O that completes. When writing to the disk log, that can quickly grow to a
2926 very large size. Setting this option makes fio average the each log entry
2927 over the specified period of time, reducing the resolution of the log. See
2928 :option:`log_max_value` as well. Defaults to 0, logging all entries.
6fc82095 2929 Also see `Log File Formats`_.
06842027 2930
f80dba8d 2931.. option:: log_hist_msec=int
06842027 2932
f80dba8d
MT
2933 Same as :option:`log_avg_msec`, but logs entries for completion latency
2934 histograms. Computing latency percentiles from averages of intervals using
c60ebc45 2935 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
f80dba8d
MT
2936 histogram entries over the specified period of time, reducing log sizes for
2937 high IOPS devices while retaining percentile accuracy. See
074f0817
SW
2938 :option:`log_hist_coarseness` and :option:`write_hist_log` as well.
2939 Defaults to 0, meaning histogram logging is disabled.
06842027 2940
f80dba8d 2941.. option:: log_hist_coarseness=int
06842027 2942
f80dba8d
MT
2943 Integer ranging from 0 to 6, defining the coarseness of the resolution of
2944 the histogram logs enabled with :option:`log_hist_msec`. For each increment
2945 in coarseness, fio outputs half as many bins. Defaults to 0, for which
074f0817
SW
2946 histogram logs contain 1216 latency bins. See :option:`write_hist_log`
2947 and `Log File Formats`_.
8b28bd41 2948
f80dba8d 2949.. option:: log_max_value=bool
66c098b8 2950
f80dba8d
MT
2951 If :option:`log_avg_msec` is set, fio logs the average over that window. If
2952 you instead want to log the maximum value, set this option to 1. Defaults to
2953 0, meaning that averaged values are logged.
a696fa2a 2954
589e88b7 2955.. option:: log_offset=bool
a696fa2a 2956
f80dba8d 2957 If this is set, the iolog options will include the byte offset for the I/O
5a83478f
SW
2958 entry as well as the other data values. Defaults to 0 meaning that
2959 offsets are not present in logs. Also see `Log File Formats`_.
71bfa161 2960
f80dba8d 2961.. option:: log_compression=int
7de87099 2962
f80dba8d
MT
2963 If this is set, fio will compress the I/O logs as it goes, to keep the
2964 memory footprint lower. When a log reaches the specified size, that chunk is
2965 removed and compressed in the background. Given that I/O logs are fairly
2966 highly compressible, this yields a nice memory savings for longer runs. The
2967 downside is that the compression will consume some background CPU cycles, so
2968 it may impact the run. This, however, is also true if the logging ends up
2969 consuming most of the system memory. So pick your poison. The I/O logs are
2970 saved normally at the end of a run, by decompressing the chunks and storing
2971 them in the specified log file. This feature depends on the availability of
2972 zlib.
e0b0d892 2973
f80dba8d 2974.. option:: log_compression_cpus=str
e0b0d892 2975
f80dba8d
MT
2976 Define the set of CPUs that are allowed to handle online log compression for
2977 the I/O jobs. This can provide better isolation between performance
0cf90a62
SW
2978 sensitive jobs, and background compression work. See
2979 :option:`cpus_allowed` for the format used.
9e684a49 2980
f80dba8d 2981.. option:: log_store_compressed=bool
9e684a49 2982
f80dba8d
MT
2983 If set, fio will store the log files in a compressed format. They can be
2984 decompressed with fio, using the :option:`--inflate-log` command line
2985 parameter. The files will be stored with a :file:`.fz` suffix.
9e684a49 2986
f80dba8d 2987.. option:: log_unix_epoch=bool
9e684a49 2988
f80dba8d
MT
2989 If set, fio will log Unix timestamps to the log files produced by enabling
2990 write_type_log for each log type, instead of the default zero-based
2991 timestamps.
2992
2993.. option:: block_error_percentiles=bool
2994
2995 If set, record errors in trim block-sized units from writes and trims and
2996 output a histogram of how many trims it took to get to errors, and what kind
2997 of error was encountered.
2998
2999.. option:: bwavgtime=int
3000
3001 Average the calculated bandwidth over the given time. Value is specified in
3002 milliseconds. If the job also does bandwidth logging through
3003 :option:`write_bw_log`, then the minimum of this option and
3004 :option:`log_avg_msec` will be used. Default: 500ms.
3005
3006.. option:: iopsavgtime=int
3007
3008 Average the calculated IOPS over the given time. Value is specified in
3009 milliseconds. If the job also does IOPS logging through
3010 :option:`write_iops_log`, then the minimum of this option and
3011 :option:`log_avg_msec` will be used. Default: 500ms.
3012
3013.. option:: disk_util=bool
3014
3015 Generate disk utilization statistics, if the platform supports it.
3016 Default: true.
3017
3018.. option:: disable_lat=bool
3019
3020 Disable measurements of total latency numbers. Useful only for cutting back
3021 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
3022 performance at really high IOPS rates. Note that to really get rid of a
3023 large amount of these calls, this option must be used with
f75ede1d 3024 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
f80dba8d
MT
3025
3026.. option:: disable_clat=bool
3027
3028 Disable measurements of completion latency numbers. See
3029 :option:`disable_lat`.
3030
3031.. option:: disable_slat=bool
3032
3033 Disable measurements of submission latency numbers. See
f50fbdda 3034 :option:`disable_lat`.
f80dba8d 3035
f75ede1d 3036.. option:: disable_bw_measurement=bool, disable_bw=bool
f80dba8d
MT
3037
3038 Disable measurements of throughput/bandwidth numbers. See
3039 :option:`disable_lat`.
3040
3041.. option:: clat_percentiles=bool
3042
b599759b
JA
3043 Enable the reporting of percentiles of completion latencies. This
3044 option is mutually exclusive with :option:`lat_percentiles`.
3045
3046.. option:: lat_percentiles=bool
3047
b71968b1 3048 Enable the reporting of percentiles of I/O latencies. This is similar
b599759b
JA
3049 to :option:`clat_percentiles`, except that this includes the
3050 submission latency. This option is mutually exclusive with
3051 :option:`clat_percentiles`.
f80dba8d
MT
3052
3053.. option:: percentile_list=float_list
3054
c32ba107
JA
3055 Overwrite the default list of percentiles for completion latencies and
3056 the block error histogram. Each number is a floating number in the
3057 range (0,100], and the maximum length of the list is 20. Use ``:`` to
3058 separate the numbers, and list the numbers in ascending order. For
3059 example, ``--percentile_list=99.5:99.9`` will cause fio to report the
3060 values of completion latency below which 99.5% and 99.9% of the observed
3061 latencies fell, respectively.
f80dba8d 3062
e883cb35
JF
3063.. option:: significant_figures=int
3064
c32ba107
JA
3065 If using :option:`--output-format` of `normal`, set the significant
3066 figures to this value. Higher values will yield more precise IOPS and
3067 throughput units, while lower values will round. Requires a minimum
3068 value of 1 and a maximum value of 10. Defaults to 4.
e883cb35 3069
f80dba8d
MT
3070
3071Error handling
3072~~~~~~~~~~~~~~
3073
3074.. option:: exitall_on_error
3075
3076 When one job finishes in error, terminate the rest. The default is to wait
3077 for each job to finish.
3078
3079.. option:: continue_on_error=str
3080
3081 Normally fio will exit the job on the first observed failure. If this option
3082 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
3083 EILSEQ) until the runtime is exceeded or the I/O size specified is
3084 completed. If this option is used, there are two more stats that are
3085 appended, the total error count and the first error. The error field given
3086 in the stats is the first error that was hit during the run.
3087
3088 The allowed values are:
3089
3090 **none**
3091 Exit on any I/O or verify errors.
3092
3093 **read**
3094 Continue on read errors, exit on all others.
3095
3096 **write**
3097 Continue on write errors, exit on all others.
3098
3099 **io**
3100 Continue on any I/O error, exit on all others.
3101
3102 **verify**
3103 Continue on verify errors, exit on all others.
3104
3105 **all**
3106 Continue on all errors.
3107
3108 **0**
3109 Backward-compatible alias for 'none'.
3110
3111 **1**
3112 Backward-compatible alias for 'all'.
3113
3114.. option:: ignore_error=str
3115
3116 Sometimes you want to ignore some errors during test in that case you can
a35ef7cb
TK
3117 specify error list for each error type, instead of only being able to
3118 ignore the default 'non-fatal error' using :option:`continue_on_error`.
f80dba8d
MT
3119 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
3120 given error type is separated with ':'. Error may be symbol ('ENOSPC',
3121 'ENOMEM') or integer. Example::
3122
3123 ignore_error=EAGAIN,ENOSPC:122
3124
3125 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
a35ef7cb
TK
3126 WRITE. This option works by overriding :option:`continue_on_error` with
3127 the list of errors for each error type if any.
f80dba8d
MT
3128
3129.. option:: error_dump=bool
3130
3131 If set dump every error even if it is non fatal, true by default. If
3132 disabled only fatal error will be dumped.
3133
f75ede1d
SW
3134Running predefined workloads
3135----------------------------
3136
3137Fio includes predefined profiles that mimic the I/O workloads generated by
3138other tools.
3139
3140.. option:: profile=str
3141
3142 The predefined workload to run. Current profiles are:
3143
3144 **tiobench**
3145 Threaded I/O bench (tiotest/tiobench) like workload.
3146
3147 **act**
3148 Aerospike Certification Tool (ACT) like workload.
3149
3150To view a profile's additional options use :option:`--cmdhelp` after specifying
3151the profile. For example::
3152
f50fbdda 3153 $ fio --profile=act --cmdhelp
f75ede1d
SW
3154
3155Act profile options
3156~~~~~~~~~~~~~~~~~~~
3157
3158.. option:: device-names=str
3159 :noindex:
3160
3161 Devices to use.
3162
3163.. option:: load=int
3164 :noindex:
3165
3166 ACT load multiplier. Default: 1.
3167
3168.. option:: test-duration=time
3169 :noindex:
3170
947e0fe0
SW
3171 How long the entire test takes to run. When the unit is omitted, the value
3172 is given in seconds. Default: 24h.
f75ede1d
SW
3173
3174.. option:: threads-per-queue=int
3175 :noindex:
3176
f50fbdda 3177 Number of read I/O threads per device. Default: 8.
f75ede1d
SW
3178
3179.. option:: read-req-num-512-blocks=int
3180 :noindex:
3181
3182 Number of 512B blocks to read at the time. Default: 3.
3183
3184.. option:: large-block-op-kbytes=int
3185 :noindex:
3186
3187 Size of large block ops in KiB (writes). Default: 131072.
3188
3189.. option:: prep
3190 :noindex:
3191
3192 Set to run ACT prep phase.
3193
3194Tiobench profile options
3195~~~~~~~~~~~~~~~~~~~~~~~~
3196
3197.. option:: size=str
3198 :noindex:
3199
f50fbdda 3200 Size in MiB.
f75ede1d
SW
3201
3202.. option:: block=int
3203 :noindex:
3204
3205 Block size in bytes. Default: 4096.
3206
3207.. option:: numruns=int
3208 :noindex:
3209
3210 Number of runs.
3211
3212.. option:: dir=str
3213 :noindex:
3214
3215 Test directory.
3216
3217.. option:: threads=int
3218 :noindex:
3219
3220 Number of threads.
f80dba8d
MT
3221
3222Interpreting the output
3223-----------------------
3224
36214730
SW
3225..
3226 Example output was based on the following:
3227 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --time_based \
3228 --rate=1256k --bs=14K --name=quick --runtime=1s --name=mixed \
3229 --runtime=2m --rw=rw
3230
f80dba8d
MT
3231Fio spits out a lot of output. While running, fio will display the status of the
3232jobs created. An example of that would be::
3233
9d25d068 3234 Jobs: 1 (f=1): [_(1),M(1)][24.8%][r=20.5MiB/s,w=23.5MiB/s][r=82,w=94 IOPS][eta 01m:31s]
f80dba8d 3235
36214730
SW
3236The characters inside the first set of square brackets denote the current status of
3237each thread. The first character is the first job defined in the job file, and so
3238forth. The possible values (in typical life cycle order) are:
f80dba8d
MT
3239
3240+------+-----+-----------------------------------------------------------+
3241| Idle | Run | |
3242+======+=====+===========================================================+
3243| P | | Thread setup, but not started. |
3244+------+-----+-----------------------------------------------------------+
3245| C | | Thread created. |
3246+------+-----+-----------------------------------------------------------+
3247| I | | Thread initialized, waiting or generating necessary data. |
3248+------+-----+-----------------------------------------------------------+
3249| | p | Thread running pre-reading file(s). |
3250+------+-----+-----------------------------------------------------------+
36214730
SW
3251| | / | Thread is in ramp period. |
3252+------+-----+-----------------------------------------------------------+
f80dba8d
MT
3253| | R | Running, doing sequential reads. |
3254+------+-----+-----------------------------------------------------------+
3255| | r | Running, doing random reads. |
3256+------+-----+-----------------------------------------------------------+
3257| | W | Running, doing sequential writes. |
3258+------+-----+-----------------------------------------------------------+
3259| | w | Running, doing random writes. |
3260+------+-----+-----------------------------------------------------------+
3261| | M | Running, doing mixed sequential reads/writes. |
3262+------+-----+-----------------------------------------------------------+
3263| | m | Running, doing mixed random reads/writes. |
3264+------+-----+-----------------------------------------------------------+
36214730
SW
3265| | D | Running, doing sequential trims. |
3266+------+-----+-----------------------------------------------------------+
3267| | d | Running, doing random trims. |
3268+------+-----+-----------------------------------------------------------+
3269| | F | Running, currently waiting for :manpage:`fsync(2)`. |
f80dba8d
MT
3270+------+-----+-----------------------------------------------------------+
3271| | V | Running, doing verification of written data. |
3272+------+-----+-----------------------------------------------------------+
36214730
SW
3273| f | | Thread finishing. |
3274+------+-----+-----------------------------------------------------------+
f80dba8d
MT
3275| E | | Thread exited, not reaped by main thread yet. |
3276+------+-----+-----------------------------------------------------------+
36214730 3277| _ | | Thread reaped. |
f80dba8d
MT
3278+------+-----+-----------------------------------------------------------+
3279| X | | Thread reaped, exited with an error. |
3280+------+-----+-----------------------------------------------------------+
3281| K | | Thread reaped, exited due to signal. |
3282+------+-----+-----------------------------------------------------------+
3283
36214730
SW
3284..
3285 Example output was based on the following:
3286 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --runtime=58m \
3287 --time_based --rate=2512k --bs=256K --numjobs=10 \
3288 --name=readers --rw=read --name=writers --rw=write
3289
f80dba8d 3290Fio will condense the thread string as not to take up more space on the command
36214730 3291line than needed. For instance, if you have 10 readers and 10 writers running,
f80dba8d
MT
3292the output would look like this::
3293
9d25d068 3294 Jobs: 20 (f=20): [R(10),W(10)][4.0%][r=20.5MiB/s,w=23.5MiB/s][r=82,w=94 IOPS][eta 57m:36s]
f80dba8d 3295
36214730
SW
3296Note that the status string is displayed in order, so it's possible to tell which of
3297the jobs are currently doing what. In the example above this means that jobs 1--10
3298are readers and 11--20 are writers.
f80dba8d
MT
3299
3300The other values are fairly self explanatory -- number of threads currently
36214730
SW
3301running and doing I/O, the number of currently open files (f=), the estimated
3302completion percentage, the rate of I/O since last check (read speed listed first,
f50fbdda
TK
3303then write speed and optionally trim speed) in terms of bandwidth and IOPS,
3304and time to completion for the current running group. It's impossible to estimate
3305runtime of the following groups (if any).
36214730
SW
3306
3307..
3308 Example output was based on the following:
3309 TZ=UTC fio --iodepth=16 --ioengine=posixaio --filename=/tmp/fiofile \
3310 --direct=1 --size=100M --time_based --runtime=50s --rate_iops=89 \
3311 --bs=7K --name=Client1 --rw=write
3312
3313When fio is done (or interrupted by :kbd:`Ctrl-C`), it will show the data for
3314each thread, group of threads, and disks in that order. For each overall thread (or
3315group) the output looks like::
3316
3317 Client1: (groupid=0, jobs=1): err= 0: pid=16109: Sat Jun 24 12:07:54 2017
3318 write: IOPS=88, BW=623KiB/s (638kB/s)(30.4MiB/50032msec)
3319 slat (nsec): min=500, max=145500, avg=8318.00, stdev=4781.50
3320 clat (usec): min=170, max=78367, avg=4019.02, stdev=8293.31
3321 lat (usec): min=174, max=78375, avg=4027.34, stdev=8291.79
3322 clat percentiles (usec):
3323 | 1.00th=[ 302], 5.00th=[ 326], 10.00th=[ 343], 20.00th=[ 363],
3324 | 30.00th=[ 392], 40.00th=[ 404], 50.00th=[ 416], 60.00th=[ 445],
3325 | 70.00th=[ 816], 80.00th=[ 6718], 90.00th=[12911], 95.00th=[21627],
3326 | 99.00th=[43779], 99.50th=[51643], 99.90th=[68682], 99.95th=[72877],
3327 | 99.99th=[78119]
3328 bw ( KiB/s): min= 532, max= 686, per=0.10%, avg=622.87, stdev=24.82, samples= 100
3329 iops : min= 76, max= 98, avg=88.98, stdev= 3.54, samples= 100
29092211
VF
3330 lat (usec) : 250=0.04%, 500=64.11%, 750=4.81%, 1000=2.79%
3331 lat (msec) : 2=4.16%, 4=1.84%, 10=4.90%, 20=11.33%, 50=5.37%
3332 lat (msec) : 100=0.65%
36214730
SW
3333 cpu : usr=0.27%, sys=0.18%, ctx=12072, majf=0, minf=21
3334 IO depths : 1=85.0%, 2=13.1%, 4=1.8%, 8=0.1%, 16=0.0%, 32=0.0%, >=64=0.0%
3335 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3336 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3337 issued rwt: total=0,4450,0, short=0,0,0, dropped=0,0,0
3338 latency : target=0, window=0, percentile=100.00%, depth=8
3339
3340The job name (or first job's name when using :option:`group_reporting`) is printed,
3341along with the group id, count of jobs being aggregated, last error id seen (which
3342is 0 when there are no errors), pid/tid of that thread and the time the job/group
3343completed. Below are the I/O statistics for each data direction performed (showing
3344writes in the example above). In the order listed, they denote:
3345
3346**read/write/trim**
3347 The string before the colon shows the I/O direction the statistics
3348 are for. **IOPS** is the average I/Os performed per second. **BW**
3349 is the average bandwidth rate shown as: value in power of 2 format
3350 (value in power of 10 format). The last two values show: (**total
3351 I/O performed** in power of 2 format / **runtime** of that thread).
f80dba8d
MT
3352
3353**slat**
36214730
SW
3354 Submission latency (**min** being the minimum, **max** being the
3355 maximum, **avg** being the average, **stdev** being the standard
3356 deviation). This is the time it took to submit the I/O. For
3357 sync I/O this row is not displayed as the slat is really the
3358 completion latency (since queue/complete is one operation there).
3359 This value can be in nanoseconds, microseconds or milliseconds ---
3360 fio will choose the most appropriate base and print that (in the
3361 example above nanoseconds was the best scale). Note: in :option:`--minimal` mode
0d237712 3362 latencies are always expressed in microseconds.
f80dba8d
MT
3363
3364**clat**
3365 Completion latency. Same names as slat, this denotes the time from
3366 submission to completion of the I/O pieces. For sync I/O, clat will
3367 usually be equal (or very close) to 0, as the time from submit to
3368 complete is basically just CPU time (I/O has already been done, see slat
3369 explanation).
3370
29092211
VF
3371**lat**
3372 Total latency. Same names as slat and clat, this denotes the time from
3373 when fio created the I/O unit to completion of the I/O operation.
3374
f80dba8d 3375**bw**
36214730
SW
3376 Bandwidth statistics based on samples. Same names as the xlat stats,
3377 but also includes the number of samples taken (**samples**) and an
3378 approximate percentage of total aggregate bandwidth this thread
3379 received in its group (**per**). This last value is only really
3380 useful if the threads in this group are on the same disk, since they
3381 are then competing for disk access.
3382
3383**iops**
3384 IOPS statistics based on samples. Same names as bw.
f80dba8d 3385
29092211
VF
3386**lat (nsec/usec/msec)**
3387 The distribution of I/O completion latencies. This is the time from when
3388 I/O leaves fio and when it gets completed. Unlike the separate
3389 read/write/trim sections above, the data here and in the remaining
3390 sections apply to all I/Os for the reporting group. 250=0.04% means that
3391 0.04% of the I/Os completed in under 250us. 500=64.11% means that 64.11%
3392 of the I/Os required 250 to 499us for completion.
3393
f80dba8d
MT
3394**cpu**
3395 CPU usage. User and system time, along with the number of context
3396 switches this thread went through, usage of system and user time, and
3397 finally the number of major and minor page faults. The CPU utilization
3398 numbers are averages for the jobs in that reporting group, while the
23a8e176 3399 context and fault counters are summed.
f80dba8d
MT
3400
3401**IO depths**
a2140525
SW
3402 The distribution of I/O depths over the job lifetime. The numbers are
3403 divided into powers of 2 and each entry covers depths from that value
3404 up to those that are lower than the next entry -- e.g., 16= covers
3405 depths from 16 to 31. Note that the range covered by a depth
3406 distribution entry can be different to the range covered by the
3407 equivalent submit/complete distribution entry.
f80dba8d
MT
3408
3409**IO submit**
3410 How many pieces of I/O were submitting in a single submit call. Each
c60ebc45 3411 entry denotes that amount and below, until the previous entry -- e.g.,
a2140525
SW
3412 16=100% means that we submitted anywhere between 9 to 16 I/Os per submit
3413 call. Note that the range covered by a submit distribution entry can
3414 be different to the range covered by the equivalent depth distribution
3415 entry.
f80dba8d
MT
3416
3417**IO complete**
3418 Like the above submit number, but for completions instead.
3419
36214730
SW
3420**IO issued rwt**
3421 The number of read/write/trim requests issued, and how many of them were
3422 short or dropped.
f80dba8d 3423
29092211 3424**IO latency**
ee21ebee 3425 These values are for :option:`latency_target` and related options. When
29092211
VF
3426 these options are engaged, this section describes the I/O depth required
3427 to meet the specified latency target.
71bfa161 3428
36214730
SW
3429..
3430 Example output was based on the following:
3431 TZ=UTC fio --ioengine=null --iodepth=2 --size=100M --numjobs=2 \
3432 --rate_process=poisson --io_limit=32M --name=read --bs=128k \
3433 --rate=11M --name=write --rw=write --bs=2k --rate=700k
3434
71bfa161 3435After each client has been listed, the group statistics are printed. They
f80dba8d 3436will look like this::
71bfa161 3437
f80dba8d 3438 Run status group 0 (all jobs):
36214730
SW
3439 READ: bw=20.9MiB/s (21.9MB/s), 10.4MiB/s-10.8MiB/s (10.9MB/s-11.3MB/s), io=64.0MiB (67.1MB), run=2973-3069msec
3440 WRITE: bw=1231KiB/s (1261kB/s), 616KiB/s-621KiB/s (630kB/s-636kB/s), io=64.0MiB (67.1MB), run=52747-53223msec
71bfa161 3441
36214730 3442For each data direction it prints:
71bfa161 3443
36214730
SW
3444**bw**
3445 Aggregate bandwidth of threads in this group followed by the
3446 minimum and maximum bandwidth of all the threads in this group.
3447 Values outside of brackets are power-of-2 format and those
3448 within are the equivalent value in a power-of-10 format.
f80dba8d 3449**io**
36214730
SW
3450 Aggregate I/O performed of all threads in this group. The
3451 format is the same as bw.
3452**run**
3453 The smallest and longest runtimes of the threads in this group.
71bfa161 3454
f50fbdda 3455And finally, the disk statistics are printed. This is Linux specific. They will look like this::
71bfa161 3456
f80dba8d
MT
3457 Disk stats (read/write):
3458 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
71bfa161
JA
3459
3460Each value is printed for both reads and writes, with reads first. The
3461numbers denote:
3462
f80dba8d 3463**ios**
c60ebc45 3464 Number of I/Os performed by all groups.
f80dba8d 3465**merge**
007c7be9 3466 Number of merges performed by the I/O scheduler.
f80dba8d
MT
3467**ticks**
3468 Number of ticks we kept the disk busy.
36214730 3469**in_queue**
f80dba8d
MT
3470 Total time spent in the disk queue.
3471**util**
3472 The disk utilization. A value of 100% means we kept the disk
71bfa161
JA
3473 busy constantly, 50% would be a disk idling half of the time.
3474
f80dba8d
MT
3475It is also possible to get fio to dump the current output while it is running,
3476without terminating the job. To do that, send fio the **USR1** signal. You can
3477also get regularly timed dumps by using the :option:`--status-interval`
3478parameter, or by creating a file in :file:`/tmp` named
3479:file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
3480current output status.
8423bd11 3481
71bfa161 3482
f80dba8d
MT
3483Terse output
3484------------
71bfa161 3485
f80dba8d
MT
3486For scripted usage where you typically want to generate tables or graphs of the
3487results, fio can output the results in a semicolon separated format. The format
3488is one long line of values, such as::
71bfa161 3489
f80dba8d
MT
3490 2;card0;0;0;7139336;121836;60004;1;10109;27.932460;116.933948;220;126861;3495.446807;1085.368601;226;126864;3523.635629;1089.012448;24063;99944;50.275485%;59818.274627;5540.657370;7155060;122104;60004;1;8338;29.086342;117.839068;388;128077;5032.488518;1234.785715;391;128085;5061.839412;1236.909129;23436;100928;50.287926%;59964.832030;5644.844189;14.595833%;19.394167%;123706;0;7313;0.1%;0.1%;0.1%;0.1%;0.1%;0.1%;100.0%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.01%;0.02%;0.05%;0.16%;6.04%;40.40%;52.68%;0.64%;0.01%;0.00%;0.01%;0.00%;0.00%;0.00%;0.00%;0.00%
3491 A description of this job goes here.
562c2d2f
DN
3492
3493The job description (if provided) follows on a second line.
71bfa161 3494
a7f77fa6
SW
3495To enable terse output, use the :option:`--minimal` or
3496:option:`--output-format`\=terse command line options. The
f80dba8d
MT
3497first value is the version of the terse output format. If the output has to be
3498changed for some reason, this number will be incremented by 1 to signify that
3499change.
6820cb3b 3500
a2c95580 3501Split up, the format is as follows (comments in brackets denote when a
007c7be9 3502field was introduced or whether it's specific to some terse version):
71bfa161 3503
f80dba8d
MT
3504 ::
3505
f50fbdda 3506 terse version, fio version [v3], jobname, groupid, error
f80dba8d
MT
3507
3508 READ status::
3509
3510 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3511 Submission latency: min, max, mean, stdev (usec)
3512 Completion latency: min, max, mean, stdev (usec)
3513 Completion latency percentiles: 20 fields (see below)
3514 Total latency: min, max, mean, stdev (usec)
f50fbdda
TK
3515 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3516 IOPS [v5]: min, max, mean, stdev, number of samples
f80dba8d
MT
3517
3518 WRITE status:
3519
3520 ::
3521
3522 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3523 Submission latency: min, max, mean, stdev (usec)
247823cc 3524 Completion latency: min, max, mean, stdev (usec)
f80dba8d
MT
3525 Completion latency percentiles: 20 fields (see below)
3526 Total latency: min, max, mean, stdev (usec)
f50fbdda
TK
3527 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3528 IOPS [v5]: min, max, mean, stdev, number of samples
a2c95580
AH
3529
3530 TRIM status [all but version 3]:
3531
f50fbdda 3532 Fields are similar to READ/WRITE status.
f80dba8d
MT
3533
3534 CPU usage::
3535
3536 user, system, context switches, major faults, minor faults
3537
3538 I/O depths::
3539
3540 <=1, 2, 4, 8, 16, 32, >=64
3541
3542 I/O latencies microseconds::
3543
3544 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
3545
3546 I/O latencies milliseconds::
3547
3548 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
3549
a2c95580 3550 Disk utilization [v3]::
f80dba8d 3551
f50fbdda
TK
3552 disk name, read ios, write ios, read merges, write merges, read ticks, write ticks,
3553 time spent in queue, disk utilization percentage
f80dba8d
MT
3554
3555 Additional Info (dependent on continue_on_error, default off)::
3556
3557 total # errors, first error code
3558
3559 Additional Info (dependent on description being set)::
3560
3561 Text description
3562
3563Completion latency percentiles can be a grouping of up to 20 sets, so for the
3564terse output fio writes all of them. Each field will look like this::
1db92cb6 3565
f50fbdda 3566 1.00%=6112
1db92cb6 3567
f80dba8d 3568which is the Xth percentile, and the `usec` latency associated with it.
1db92cb6 3569
f50fbdda 3570For `Disk utilization`, all disks used by fio are shown. So for each disk there
f80dba8d 3571will be a disk utilization section.
f2f788dd 3572
2fc26c3d 3573Below is a single line containing short names for each of the fields in the
2831be97 3574minimal output v3, separated by semicolons::
2fc26c3d 3575
f50fbdda 3576 terse_version_3;fio_version;jobname;groupid;error;read_kb;read_bandwidth;read_iops;read_runtime_ms;read_slat_min;read_slat_max;read_slat_mean;read_slat_dev;read_clat_min;read_clat_max;read_clat_mean;read_clat_dev;read_clat_pct01;read_clat_pct02;read_clat_pct03;read_clat_pct04;read_clat_pct05;read_clat_pct06;read_clat_pct07;read_clat_pct08;read_clat_pct09;read_clat_pct10;read_clat_pct11;read_clat_pct12;read_clat_pct13;read_clat_pct14;read_clat_pct15;read_clat_pct16;read_clat_pct17;read_clat_pct18;read_clat_pct19;read_clat_pct20;read_tlat_min;read_lat_max;read_lat_mean;read_lat_dev;read_bw_min;read_bw_max;read_bw_agg_pct;read_bw_mean;read_bw_dev;write_kb;write_bandwidth;write_iops;write_runtime_ms;write_slat_min;write_slat_max;write_slat_mean;write_slat_dev;write_clat_min;write_clat_max;write_clat_mean;write_clat_dev;write_clat_pct01;write_clat_pct02;write_clat_pct03;write_clat_pct04;write_clat_pct05;write_clat_pct06;write_clat_pct07;write_clat_pct08;write_clat_pct09;write_clat_pct10;write_clat_pct11;write_clat_pct12;write_clat_pct13;write_clat_pct14;write_clat_pct15;write_clat_pct16;write_clat_pct17;write_clat_pct18;write_clat_pct19;write_clat_pct20;write_tlat_min;write_lat_max;write_lat_mean;write_lat_dev;write_bw_min;write_bw_max;write_bw_agg_pct;write_bw_mean;write_bw_dev;cpu_user;cpu_sys;cpu_csw;cpu_mjf;cpu_minf;iodepth_1;iodepth_2;iodepth_4;iodepth_8;iodepth_16;iodepth_32;iodepth_64;lat_2us;lat_4us;lat_10us;lat_20us;lat_50us;lat_100us;lat_250us;lat_500us;lat_750us;lat_1000us;lat_2ms;lat_4ms;lat_10ms;lat_20ms;lat_50ms;lat_100ms;lat_250ms;lat_500ms;lat_750ms;lat_1000ms;lat_2000ms;lat_over_2000ms;disk_name;disk_read_iops;disk_write_iops;disk_read_merges;disk_write_merges;disk_read_ticks;write_ticks;disk_queue_time;disk_util
2fc26c3d 3577
25c8b9d7 3578
44c82dba
VF
3579JSON output
3580------------
3581
3582The `json` output format is intended to be both human readable and convenient
3583for automated parsing. For the most part its sections mirror those of the
3584`normal` output. The `runtime` value is reported in msec and the `bw` value is
3585reported in 1024 bytes per second units.
3586
3587
d29c4a91
VF
3588JSON+ output
3589------------
3590
3591The `json+` output format is identical to the `json` output format except that it
3592adds a full dump of the completion latency bins. Each `bins` object contains a
3593set of (key, value) pairs where keys are latency durations and values count how
3594many I/Os had completion latencies of the corresponding duration. For example,
3595consider:
3596
3597 "bins" : { "87552" : 1, "89600" : 1, "94720" : 1, "96768" : 1, "97792" : 1, "99840" : 1, "100864" : 2, "103936" : 6, "104960" : 534, "105984" : 5995, "107008" : 7529, ... }
3598
3599This data indicates that one I/O required 87,552ns to complete, two I/Os required
3600100,864ns to complete, and 7529 I/Os required 107,008ns to complete.
3601
3602Also included with fio is a Python script `fio_jsonplus_clat2csv` that takes
3603json+ output and generates CSV-formatted latency data suitable for plotting.
3604
3605The latency durations actually represent the midpoints of latency intervals.
f50fbdda 3606For details refer to :file:`stat.h`.
d29c4a91
VF
3607
3608
f80dba8d
MT
3609Trace file format
3610-----------------
3611
3612There are two trace file format that you can encounter. The older (v1) format is
3613unsupported since version 1.20-rc3 (March 2008). It will still be described
25c8b9d7
PD
3614below in case that you get an old trace and want to understand it.
3615
3616In any case the trace is a simple text file with a single action per line.
3617
3618
f80dba8d
MT
3619Trace file format v1
3620~~~~~~~~~~~~~~~~~~~~
3621
3622Each line represents a single I/O action in the following format::
3623
3624 rw, offset, length
25c8b9d7 3625
f50fbdda 3626where `rw=0/1` for read/write, and the `offset` and `length` entries being in bytes.
25c8b9d7 3627
22413915 3628This format is not supported in fio versions >= 1.20-rc3.
25c8b9d7 3629
25c8b9d7 3630
f80dba8d
MT
3631Trace file format v2
3632~~~~~~~~~~~~~~~~~~~~
25c8b9d7 3633
f80dba8d
MT
3634The second version of the trace file format was added in fio version 1.17. It
3635allows to access more then one file per trace and has a bigger set of possible
3636file actions.
25c8b9d7 3637
f80dba8d 3638The first line of the trace file has to be::
25c8b9d7 3639
f80dba8d 3640 fio version 2 iolog
25c8b9d7
PD
3641
3642Following this can be lines in two different formats, which are described below.
3643
f80dba8d 3644The file management format::
25c8b9d7 3645
f80dba8d 3646 filename action
25c8b9d7 3647
f50fbdda 3648The `filename` is given as an absolute path. The `action` can be one of these:
25c8b9d7 3649
f80dba8d 3650**add**
f50fbdda 3651 Add the given `filename` to the trace.
f80dba8d 3652**open**
f50fbdda 3653 Open the file with the given `filename`. The `filename` has to have
f80dba8d
MT
3654 been added with the **add** action before.
3655**close**
f50fbdda 3656 Close the file with the given `filename`. The file has to have been
f80dba8d
MT
3657 opened before.
3658
3659
3660The file I/O action format::
3661
3662 filename action offset length
3663
3664The `filename` is given as an absolute path, and has to have been added and
3665opened before it can be used with this format. The `offset` and `length` are
3666given in bytes. The `action` can be one of these:
3667
3668**wait**
3669 Wait for `offset` microseconds. Everything below 100 is discarded.
3670 The time is relative to the previous `wait` statement.
3671**read**
3672 Read `length` bytes beginning from `offset`.
3673**write**
3674 Write `length` bytes beginning from `offset`.
3675**sync**
3676 :manpage:`fsync(2)` the file.
3677**datasync**
3678 :manpage:`fdatasync(2)` the file.
3679**trim**
3680 Trim the given file from the given `offset` for `length` bytes.
3681
3682CPU idleness profiling
3683----------------------
3684
3685In some cases, we want to understand CPU overhead in a test. For example, we
3686test patches for the specific goodness of whether they reduce CPU usage.
3687Fio implements a balloon approach to create a thread per CPU that runs at idle
3688priority, meaning that it only runs when nobody else needs the cpu.
3689By measuring the amount of work completed by the thread, idleness of each CPU
3690can be derived accordingly.
3691
3692An unit work is defined as touching a full page of unsigned characters. Mean and
3693standard deviation of time to complete an unit work is reported in "unit work"
3694section. Options can be chosen to report detailed percpu idleness or overall
3695system idleness by aggregating percpu stats.
3696
3697
3698Verification and triggers
3699-------------------------
3700
3701Fio is usually run in one of two ways, when data verification is done. The first
3702is a normal write job of some sort with verify enabled. When the write phase has
3703completed, fio switches to reads and verifies everything it wrote. The second
3704model is running just the write phase, and then later on running the same job
3705(but with reads instead of writes) to repeat the same I/O patterns and verify
3706the contents. Both of these methods depend on the write phase being completed,
3707as fio otherwise has no idea how much data was written.
3708
3709With verification triggers, fio supports dumping the current write state to
3710local files. Then a subsequent read verify workload can load this state and know
3711exactly where to stop. This is useful for testing cases where power is cut to a
3712server in a managed fashion, for instance.
99b9a85a
JA
3713
3714A verification trigger consists of two things:
3715
f80dba8d
MT
37161) Storing the write state of each job.
37172) Executing a trigger command.
99b9a85a 3718
f80dba8d
MT
3719The write state is relatively small, on the order of hundreds of bytes to single
3720kilobytes. It contains information on the number of completions done, the last X
3721completions, etc.
99b9a85a 3722
f80dba8d
MT
3723A trigger is invoked either through creation ('touch') of a specified file in
3724the system, or through a timeout setting. If fio is run with
9207a0cb 3725:option:`--trigger-file`\= :file:`/tmp/trigger-file`, then it will continually
f80dba8d
MT
3726check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
3727will fire off the trigger (thus saving state, and executing the trigger
99b9a85a
JA
3728command).
3729
f80dba8d
MT
3730For client/server runs, there's both a local and remote trigger. If fio is
3731running as a server backend, it will send the job states back to the client for
3732safe storage, then execute the remote trigger, if specified. If a local trigger
3733is specified, the server will still send back the write state, but the client
3734will then execute the trigger.
99b9a85a 3735
f80dba8d
MT
3736Verification trigger example
3737~~~~~~~~~~~~~~~~~~~~~~~~~~~~
99b9a85a 3738
f50fbdda
TK
3739Let's say we want to run a powercut test on the remote Linux machine 'server'.
3740Our write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
f80dba8d
MT
3741some point during the run, and we'll run this test from the safety or our local
3742machine, 'localbox'. On the server, we'll start the fio backend normally::
99b9a85a 3743
f80dba8d 3744 server# fio --server
99b9a85a 3745
f80dba8d 3746and on the client, we'll fire off the workload::
99b9a85a 3747
f80dba8d 3748 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
99b9a85a 3749
f80dba8d 3750We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
99b9a85a 3751
f80dba8d 3752 echo b > /proc/sysrq-trigger
99b9a85a 3753
f80dba8d
MT
3754on the server once it has received the trigger and sent us the write state. This
3755will work, but it's not **really** cutting power to the server, it's merely
3756abruptly rebooting it. If we have a remote way of cutting power to the server
3757through IPMI or similar, we could do that through a local trigger command
4502cb42 3758instead. Let's assume we have a script that does IPMI reboot of a given hostname,
f80dba8d
MT
3759ipmi-reboot. On localbox, we could then have run fio with a local trigger
3760instead::
99b9a85a 3761
f80dba8d 3762 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
99b9a85a 3763
f80dba8d
MT
3764For this case, fio would wait for the server to send us the write state, then
3765execute ``ipmi-reboot server`` when that happened.
3766
3767Loading verify state
3768~~~~~~~~~~~~~~~~~~~~
3769
4502cb42 3770To load stored write state, a read verification job file must contain the
f80dba8d 3771:option:`verify_state_load` option. If that is set, fio will load the previously
99b9a85a 3772stored state. For a local fio run this is done by loading the files directly,
f80dba8d
MT
3773and on a client/server run, the server backend will ask the client to send the
3774files over and load them from there.
a3ae5b05
JA
3775
3776
f80dba8d
MT
3777Log File Formats