doc: add text about possibly verify errors with norandommap
[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**
994 Sequential trims (Linux block devices only).
f80dba8d
MT
995 **randread**
996 Random reads.
2831be97
SW
997 **randwrite**
998 Random writes.
f80dba8d
MT
999 **randtrim**
1000 Random trims (Linux block devices only).
1001 **rw,readwrite**
1002 Sequential mixed reads and writes.
1003 **randrw**
1004 Random mixed reads and writes.
1005 **trimwrite**
1006 Sequential trim+write sequences. Blocks will be trimmed first,
1007 then the same blocks will be written to.
1008
1009 Fio defaults to read if the option is not specified. For the mixed I/O
1010 types, the default is to split them 50/50. For certain types of I/O the
730bd7d9
SW
1011 result may still be skewed a bit, since the speed may be different.
1012
1013 It is possible to specify the number of I/Os to do before getting a new
1014 offset by appending ``:<nr>`` to the end of the string given. For a
f80dba8d
MT
1015 random read, it would look like ``rw=randread:8`` for passing in an offset
1016 modifier with a value of 8. If the suffix is used with a sequential I/O
730bd7d9
SW
1017 pattern, then the *<nr>* value specified will be **added** to the generated
1018 offset for each I/O turning sequential I/O into sequential I/O with holes.
1019 For instance, using ``rw=write:4k`` will skip 4k for every write. Also see
1020 the :option:`rw_sequencer` option.
f80dba8d
MT
1021
1022.. option:: rw_sequencer=str
1023
1024 If an offset modifier is given by appending a number to the ``rw=<str>``
1025 line, then this option controls how that number modifies the I/O offset
1026 being generated. Accepted values are:
1027
1028 **sequential**
1029 Generate sequential offset.
1030 **identical**
1031 Generate the same offset.
1032
1033 ``sequential`` is only useful for random I/O, where fio would normally
c60ebc45 1034 generate a new random offset for every I/O. If you append e.g. 8 to randread,
007c7be9
SW
1035 you would get a new random offset for every 8 I/Os. The result would be a
1036 seek for only every 8 I/Os, instead of for every I/O. Use ``rw=randread:8``
f80dba8d
MT
1037 to specify that. As sequential I/O is already sequential, setting
1038 ``sequential`` for that would not result in any differences. ``identical``
1039 behaves in a similar fashion, except it sends the same offset 8 number of
1040 times before generating a new offset.
1041
1042.. option:: unified_rw_reporting=bool
1043
1044 Fio normally reports statistics on a per data direction basis, meaning that
1045 reads, writes, and trims are accounted and reported separately. If this
1046 option is set fio sums the results and report them as "mixed" instead.
1047
1048.. option:: randrepeat=bool
1049
1050 Seed the random number generator used for random I/O patterns in a
1051 predictable way so the pattern is repeatable across runs. Default: true.
1052
1053.. option:: allrandrepeat=bool
1054
1055 Seed all random number generators in a predictable way so results are
1056 repeatable across runs. Default: false.
1057
1058.. option:: randseed=int
1059
1060 Seed the random number generators based on this seed value, to be able to
1061 control what sequence of output is being generated. If not set, the random
1062 sequence depends on the :option:`randrepeat` setting.
1063
1064.. option:: fallocate=str
1065
1066 Whether pre-allocation is performed when laying down files.
1067 Accepted values are:
1068
1069 **none**
1070 Do not pre-allocate space.
1071
2c3e17be
SW
1072 **native**
1073 Use a platform's native pre-allocation call but fall back to
1074 **none** behavior if it fails/is not implemented.
1075
f80dba8d
MT
1076 **posix**
1077 Pre-allocate via :manpage:`posix_fallocate(3)`.
1078
1079 **keep**
1080 Pre-allocate via :manpage:`fallocate(2)` with
1081 FALLOC_FL_KEEP_SIZE set.
1082
1083 **0**
1084 Backward-compatible alias for **none**.
1085
1086 **1**
1087 Backward-compatible alias for **posix**.
1088
1089 May not be available on all supported platforms. **keep** is only available
2c3e17be
SW
1090 on Linux. If using ZFS on Solaris this cannot be set to **posix**
1091 because ZFS doesn't support pre-allocation. Default: **native** if any
1092 pre-allocation methods are available, **none** if not.
f80dba8d
MT
1093
1094.. option:: fadvise_hint=str
1095
c712c97a
JA
1096 Use :manpage:`posix_fadvise(2)` or :manpage:`posix_fadvise(2)` to
1097 advise the kernel on what I/O patterns are likely to be issued.
1098 Accepted values are:
f80dba8d
MT
1099
1100 **0**
1101 Backwards-compatible hint for "no hint".
1102
1103 **1**
1104 Backwards compatible hint for "advise with fio workload type". This
1105 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1106 for a sequential workload.
1107
1108 **sequential**
1109 Advise using **FADV_SEQUENTIAL**.
1110
1111 **random**
1112 Advise using **FADV_RANDOM**.
1113
8f4b9f24 1114.. option:: write_hint=str
f80dba8d 1115
8f4b9f24
JA
1116 Use :manpage:`fcntl(2)` to advise the kernel what life time to expect
1117 from a write. Only supported on Linux, as of version 4.13. Accepted
1118 values are:
1119
1120 **none**
1121 No particular life time associated with this file.
1122
1123 **short**
1124 Data written to this file has a short life time.
1125
1126 **medium**
1127 Data written to this file has a medium life time.
1128
1129 **long**
1130 Data written to this file has a long life time.
1131
1132 **extreme**
1133 Data written to this file has a very long life time.
1134
1135 The values are all relative to each other, and no absolute meaning
1136 should be associated with them.
f80dba8d
MT
1137
1138.. option:: offset=int
1139
82dbb8cb 1140 Start I/O at the provided offset in the file, given as either a fixed size in
83c8b093
JF
1141 bytes or a percentage. If a percentage is given, the generated offset will be
1142 aligned to the minimum ``blocksize`` or to the value of ``offset_align`` if
1143 provided. Data before the given offset will not be touched. This
89978a6b
BW
1144 effectively caps the file size at `real_size - offset`. Can be combined with
1145 :option:`size` to constrain the start and end range of the I/O workload.
44bb1142
TK
1146 A percentage can be specified by a number between 1 and 100 followed by '%',
1147 for example, ``offset=20%`` to specify 20%.
f80dba8d 1148
83c8b093
JF
1149.. option:: offset_align=int
1150
1151 If set to non-zero value, the byte offset generated by a percentage ``offset``
1152 is aligned upwards to this value. Defaults to 0 meaning that a percentage
1153 offset is aligned to the minimum block size.
1154
f80dba8d
MT
1155.. option:: offset_increment=int
1156
1157 If this is provided, then the real offset becomes `offset + offset_increment
1158 * thread_number`, where the thread number is a counter that starts at 0 and
1159 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1160 specified). This option is useful if there are several jobs which are
1161 intended to operate on a file in parallel disjoint segments, with even
1162 spacing between the starting points.
1163
1164.. option:: number_ios=int
1165
c60ebc45 1166 Fio will normally perform I/Os until it has exhausted the size of the region
f80dba8d
MT
1167 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1168 condition). With this setting, the range/size can be set independently of
c60ebc45 1169 the number of I/Os to perform. When fio reaches this number, it will exit
f80dba8d
MT
1170 normally and report status. Note that this does not extend the amount of I/O
1171 that will be done, it will only stop fio if this condition is met before
1172 other end-of-job criteria.
1173
1174.. option:: fsync=int
1175
730bd7d9
SW
1176 If writing to a file, issue an :manpage:`fsync(2)` (or its equivalent) of
1177 the dirty data for every number of blocks given. For example, if you give 32
1178 as a parameter, fio will sync the file after every 32 writes issued. If fio is
1179 using non-buffered I/O, we may not sync the file. The exception is the sg
1180 I/O engine, which synchronizes the disk cache anyway. Defaults to 0, which
1181 means fio does not periodically issue and wait for a sync to complete. Also
1182 see :option:`end_fsync` and :option:`fsync_on_close`.
f80dba8d
MT
1183
1184.. option:: fdatasync=int
1185
1186 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
000a5f1c 1187 not metadata blocks. In Windows, FreeBSD, and DragonFlyBSD there is no
730bd7d9
SW
1188 :manpage:`fdatasync(2)` so this falls back to using :manpage:`fsync(2)`.
1189 Defaults to 0, which means fio does not periodically issue and wait for a
1190 data-only sync to complete.
f80dba8d
MT
1191
1192.. option:: write_barrier=int
1193
2831be97 1194 Make every `N-th` write a barrier write.
f80dba8d 1195
f50fbdda 1196.. option:: sync_file_range=str:int
f80dba8d 1197
f50fbdda 1198 Use :manpage:`sync_file_range(2)` for every `int` number of write
f80dba8d
MT
1199 operations. Fio will track range of writes that have happened since the last
1200 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1201
1202 **wait_before**
1203 SYNC_FILE_RANGE_WAIT_BEFORE
1204 **write**
1205 SYNC_FILE_RANGE_WRITE
1206 **wait_after**
1207 SYNC_FILE_RANGE_WAIT_AFTER
1208
1209 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1210 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1211 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1212 Linux specific.
1213
1214.. option:: overwrite=bool
1215
1216 If true, writes to a file will always overwrite existing data. If the file
1217 doesn't already exist, it will be created before the write phase begins. If
1218 the file exists and is large enough for the specified write phase, nothing
a47b697c 1219 will be done. Default: false.
f80dba8d
MT
1220
1221.. option:: end_fsync=bool
1222
a47b697c
SW
1223 If true, :manpage:`fsync(2)` file contents when a write stage has completed.
1224 Default: false.
f80dba8d
MT
1225
1226.. option:: fsync_on_close=bool
1227
1228 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
a47b697c
SW
1229 from :option:`end_fsync` in that it will happen on every file close, not
1230 just at the end of the job. Default: false.
f80dba8d
MT
1231
1232.. option:: rwmixread=int
1233
1234 Percentage of a mixed workload that should be reads. Default: 50.
1235
1236.. option:: rwmixwrite=int
1237
1238 Percentage of a mixed workload that should be writes. If both
1239 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1240 add up to 100%, the latter of the two will be used to override the
1241 first. This may interfere with a given rate setting, if fio is asked to
1242 limit reads or writes to a certain rate. If that is the case, then the
1243 distribution may be skewed. Default: 50.
1244
1245.. option:: random_distribution=str:float[,str:float][,str:float]
1246
1247 By default, fio will use a completely uniform random distribution when asked
1248 to perform random I/O. Sometimes it is useful to skew the distribution in
1249 specific ways, ensuring that some parts of the data is more hot than others.
1250 fio includes the following distribution models:
1251
1252 **random**
1253 Uniform random distribution
1254
1255 **zipf**
1256 Zipf distribution
1257
1258 **pareto**
1259 Pareto distribution
1260
b2f4b559 1261 **normal**
c60ebc45 1262 Normal (Gaussian) distribution
f80dba8d
MT
1263
1264 **zoned**
1265 Zoned random distribution
1266
59466396
JA
1267 **zoned_abs**
1268 Zone absolute random distribution
1269
f80dba8d 1270 When using a **zipf** or **pareto** distribution, an input value is also
f50fbdda 1271 needed to define the access pattern. For **zipf**, this is the `Zipf
c60ebc45 1272 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
f50fbdda 1273 program, :command:`fio-genzipf`, that can be used visualize what the given input
f80dba8d
MT
1274 values will yield in terms of hit rates. If you wanted to use **zipf** with
1275 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1276 option. If a non-uniform model is used, fio will disable use of the random
b2f4b559
SW
1277 map. For the **normal** distribution, a normal (Gaussian) deviation is
1278 supplied as a value between 0 and 100.
f80dba8d
MT
1279
1280 For a **zoned** distribution, fio supports specifying percentages of I/O
1281 access that should fall within what range of the file or device. For
1282 example, given a criteria of:
1283
f50fbdda
TK
1284 * 60% of accesses should be to the first 10%
1285 * 30% of accesses should be to the next 20%
1286 * 8% of accesses should be to the next 30%
1287 * 2% of accesses should be to the next 40%
f80dba8d
MT
1288
1289 we can define that through zoning of the random accesses. For the above
1290 example, the user would do::
1291
1292 random_distribution=zoned:60/10:30/20:8/30:2/40
1293
59466396
JA
1294 A **zoned_abs** distribution works exactly like the **zoned**, except
1295 that it takes absolute sizes. For example, let's say you wanted to
1296 define access according to the following criteria:
1297
1298 * 60% of accesses should be to the first 20G
1299 * 30% of accesses should be to the next 100G
1300 * 10% of accesses should be to the next 500G
1301
1302 we can define an absolute zoning distribution with:
1303
1304 random_distribution=zoned_abs=60/20G:30/100G:10/500g
1305
6a16ece8
JA
1306 For both **zoned** and **zoned_abs**, fio supports defining up to
1307 256 separate zones.
1308
59466396
JA
1309 Similarly to how :option:`bssplit` works for setting ranges and
1310 percentages of block sizes. Like :option:`bssplit`, it's possible to
1311 specify separate zones for reads, writes, and trims. If just one set
1312 is given, it'll apply to all of them. This goes for both **zoned**
1313 **zoned_abs** distributions.
f80dba8d
MT
1314
1315.. option:: percentage_random=int[,int][,int]
1316
1317 For a random workload, set how big a percentage should be random. This
1318 defaults to 100%, in which case the workload is fully random. It can be set
1319 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1320 sequential. Any setting in between will result in a random mix of sequential
1321 and random I/O, at the given percentages. Comma-separated values may be
1322 specified for reads, writes, and trims as described in :option:`blocksize`.
1323
1324.. option:: norandommap
1325
1326 Normally fio will cover every block of the file when doing random I/O. If
1327 this option is given, fio will just get a new random offset without looking
1328 at past I/O history. This means that some blocks may not be read or written,
1329 and that some blocks may be read/written more than once. If this option is
1330 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1331 only intact blocks are verified, i.e., partially-overwritten blocks are
47e6a6e5
SW
1332 ignored. With an async I/O engine and an I/O depth > 1, it is possible for
1333 the same block to be overwritten, which can cause verification errors. Either
1334 do not use norandommap in this case, or also use the lfsr random generator.
f80dba8d
MT
1335
1336.. option:: softrandommap=bool
1337
1338 See :option:`norandommap`. If fio runs with the random block map enabled and
1339 it fails to allocate the map, if this option is set it will continue without
1340 a random block map. As coverage will not be as complete as with random maps,
1341 this option is disabled by default.
1342
1343.. option:: random_generator=str
1344
f50fbdda 1345 Fio supports the following engines for generating I/O offsets for random I/O:
f80dba8d
MT
1346
1347 **tausworthe**
f50fbdda 1348 Strong 2^88 cycle random number generator.
f80dba8d 1349 **lfsr**
f50fbdda 1350 Linear feedback shift register generator.
f80dba8d 1351 **tausworthe64**
f50fbdda 1352 Strong 64-bit 2^258 cycle random number generator.
f80dba8d
MT
1353
1354 **tausworthe** is a strong random number generator, but it requires tracking
1355 on the side if we want to ensure that blocks are only read or written
f50fbdda 1356 once. **lfsr** guarantees that we never generate the same offset twice, and
f80dba8d 1357 it's also less computationally expensive. It's not a true random generator,
f50fbdda 1358 however, though for I/O purposes it's typically good enough. **lfsr** only
f80dba8d
MT
1359 works with single block sizes, not with workloads that use multiple block
1360 sizes. If used with such a workload, fio may read or write some blocks
1361 multiple times. The default value is **tausworthe**, unless the required
1362 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1363 selected automatically.
1364
1365
1366Block size
1367~~~~~~~~~~
1368
1369.. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1370
1371 The block size in bytes used for I/O units. Default: 4096. A single value
1372 applies to reads, writes, and trims. Comma-separated values may be
1373 specified for reads, writes, and trims. A value not terminated in a comma
1374 applies to subsequent types.
1375
1376 Examples:
1377
1378 **bs=256k**
1379 means 256k for reads, writes and trims.
1380
1381 **bs=8k,32k**
1382 means 8k for reads, 32k for writes and trims.
1383
1384 **bs=8k,32k,**
1385 means 8k for reads, 32k for writes, and default for trims.
1386
1387 **bs=,8k**
1388 means default for reads, 8k for writes and trims.
1389
1390 **bs=,8k,**
b443ae44 1391 means default for reads, 8k for writes, and default for trims.
f80dba8d
MT
1392
1393.. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1394
1395 A range of block sizes in bytes for I/O units. The issued I/O unit will
1396 always be a multiple of the minimum size, unless
1397 :option:`blocksize_unaligned` is set.
1398
1399 Comma-separated ranges may be specified for reads, writes, and trims as
1400 described in :option:`blocksize`.
1401
1402 Example: ``bsrange=1k-4k,2k-8k``.
1403
1404.. option:: bssplit=str[,str][,str]
1405
6a16ece8
JA
1406 Sometimes you want even finer grained control of the block sizes
1407 issued, not just an even split between them. This option allows you to
1408 weight various block sizes, so that you are able to define a specific
1409 amount of block sizes issued. The format for this option is::
f80dba8d
MT
1410
1411 bssplit=blocksize/percentage:blocksize/percentage
1412
6a16ece8
JA
1413 for as many block sizes as needed. So if you want to define a workload
1414 that has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would
1415 write::
f80dba8d
MT
1416
1417 bssplit=4k/10:64k/50:32k/40
1418
6a16ece8
JA
1419 Ordering does not matter. If the percentage is left blank, fio will
1420 fill in the remaining values evenly. So a bssplit option like this one::
f80dba8d
MT
1421
1422 bssplit=4k/50:1k/:32k/
1423
6a16ece8
JA
1424 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always
1425 add up to 100, if bssplit is given a range that adds up to more, it
1426 will error out.
f80dba8d
MT
1427
1428 Comma-separated values may be specified for reads, writes, and trims as
1429 described in :option:`blocksize`.
1430
6a16ece8
JA
1431 If you want a workload that has 50% 2k reads and 50% 4k reads, while
1432 having 90% 4k writes and 10% 8k writes, you would specify::
f80dba8d 1433
cf04b906 1434 bssplit=2k/50:4k/50,4k/90:8k/10
f80dba8d 1435
6a16ece8
JA
1436 Fio supports defining up to 64 different weights for each data
1437 direction.
1438
f80dba8d
MT
1439.. option:: blocksize_unaligned, bs_unaligned
1440
1441 If set, fio will issue I/O units with any size within
1442 :option:`blocksize_range`, not just multiples of the minimum size. This
1443 typically won't work with direct I/O, as that normally requires sector
1444 alignment.
1445
589e88b7 1446.. option:: bs_is_seq_rand=bool
f80dba8d
MT
1447
1448 If this option is set, fio will use the normal read,write blocksize settings
1449 as sequential,random blocksize settings instead. Any random read or write
1450 will use the WRITE blocksize settings, and any sequential read or write will
1451 use the READ blocksize settings.
1452
1453.. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1454
1455 Boundary to which fio will align random I/O units. Default:
1456 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1457 I/O, though it usually depends on the hardware block size. This option is
1458 mutually exclusive with using a random map for files, so it will turn off
1459 that option. Comma-separated values may be specified for reads, writes, and
1460 trims as described in :option:`blocksize`.
1461
1462
1463Buffers and memory
1464~~~~~~~~~~~~~~~~~~
1465
1466.. option:: zero_buffers
1467
1468 Initialize buffers with all zeros. Default: fill buffers with random data.
1469
1470.. option:: refill_buffers
1471
1472 If this option is given, fio will refill the I/O buffers on every
72592780
SW
1473 submit. Only makes sense if :option:`zero_buffers` isn't specified,
1474 naturally. Defaults to being unset i.e., the buffer is only filled at
1475 init time and the data in it is reused when possible but if any of
1476 :option:`verify`, :option:`buffer_compress_percentage` or
1477 :option:`dedupe_percentage` are enabled then `refill_buffers` is also
1478 automatically enabled.
f80dba8d
MT
1479
1480.. option:: scramble_buffers=bool
1481
1482 If :option:`refill_buffers` is too costly and the target is using data
1483 deduplication, then setting this option will slightly modify the I/O buffer
1484 contents to defeat normal de-dupe attempts. This is not enough to defeat
1485 more clever block compression attempts, but it will stop naive dedupe of
1486 blocks. Default: true.
1487
1488.. option:: buffer_compress_percentage=int
1489
72592780
SW
1490 If this is set, then fio will attempt to provide I/O buffer content
1491 (on WRITEs) that compresses to the specified level. Fio does this by
1492 providing a mix of random data followed by fixed pattern data. The
1493 fixed pattern is either zeros, or the pattern specified by
1494 :option:`buffer_pattern`. If the `buffer_pattern` option is used, it
1495 might skew the compression ratio slightly. Setting
1496 `buffer_compress_percentage` to a value other than 100 will also
1497 enable :option:`refill_buffers` in order to reduce the likelihood that
1498 adjacent blocks are so similar that they over compress when seen
1499 together. See :option:`buffer_compress_chunk` for how to set a finer or
1500 coarser granularity for the random/fixed data region. Defaults to unset
1501 i.e., buffer data will not adhere to any compression level.
f80dba8d
MT
1502
1503.. option:: buffer_compress_chunk=int
1504
72592780
SW
1505 This setting allows fio to manage how big the random/fixed data region
1506 is when using :option:`buffer_compress_percentage`. When
1507 `buffer_compress_chunk` is set to some non-zero value smaller than the
1508 block size, fio can repeat the random/fixed region throughout the I/O
1509 buffer at the specified interval (which particularly useful when
1510 bigger block sizes are used for a job). When set to 0, fio will use a
1511 chunk size that matches the block size resulting in a single
1512 random/fixed region within the I/O buffer. Defaults to 512. When the
1513 unit is omitted, the value is interpreted in bytes.
f80dba8d
MT
1514
1515.. option:: buffer_pattern=str
1516
a1554f65
SB
1517 If set, fio will fill the I/O buffers with this pattern or with the contents
1518 of a file. If not set, the contents of I/O buffers are defined by the other
1519 options related to buffer contents. The setting can be any pattern of bytes,
1520 and can be prefixed with 0x for hex values. It may also be a string, where
1521 the string must then be wrapped with ``""``. Or it may also be a filename,
1522 where the filename must be wrapped with ``''`` in which case the file is
1523 opened and read. Note that not all the file contents will be read if that
1524 would cause the buffers to overflow. So, for example::
1525
1526 buffer_pattern='filename'
1527
1528 or::
f80dba8d
MT
1529
1530 buffer_pattern="abcd"
1531
1532 or::
1533
1534 buffer_pattern=-12
1535
1536 or::
1537
1538 buffer_pattern=0xdeadface
1539
1540 Also you can combine everything together in any order::
1541
a1554f65 1542 buffer_pattern=0xdeadface"abcd"-12'filename'
f80dba8d
MT
1543
1544.. option:: dedupe_percentage=int
1545
1546 If set, fio will generate this percentage of identical buffers when
1547 writing. These buffers will be naturally dedupable. The contents of the
1548 buffers depend on what other buffer compression settings have been set. It's
1549 possible to have the individual buffers either fully compressible, or not at
72592780
SW
1550 all -- this option only controls the distribution of unique buffers. Setting
1551 this option will also enable :option:`refill_buffers` to prevent every buffer
1552 being identical.
f80dba8d
MT
1553
1554.. option:: invalidate=bool
1555
730bd7d9
SW
1556 Invalidate the buffer/page cache parts of the files to be used prior to
1557 starting I/O if the platform and file type support it. Defaults to true.
21c1b29e
TK
1558 This will be ignored if :option:`pre_read` is also specified for the
1559 same job.
f80dba8d
MT
1560
1561.. option:: sync=bool
1562
1563 Use synchronous I/O for buffered writes. For the majority of I/O engines,
1564 this means using O_SYNC. Default: false.
1565
1566.. option:: iomem=str, mem=str
1567
1568 Fio can use various types of memory as the I/O unit buffer. The allowed
1569 values are:
1570
1571 **malloc**
1572 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1573 type.
1574
1575 **shm**
1576 Use shared memory as the buffers. Allocated through
1577 :manpage:`shmget(2)`.
1578
1579 **shmhuge**
1580 Same as shm, but use huge pages as backing.
1581
1582 **mmap**
22413915 1583 Use :manpage:`mmap(2)` to allocate buffers. May either be anonymous memory, or can
f80dba8d
MT
1584 be file backed if a filename is given after the option. The format
1585 is `mem=mmap:/path/to/file`.
1586
1587 **mmaphuge**
1588 Use a memory mapped huge file as the buffer backing. Append filename
1589 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1590
1591 **mmapshared**
1592 Same as mmap, but use a MMAP_SHARED mapping.
1593
03553853
YR
1594 **cudamalloc**
1595 Use GPU memory as the buffers for GPUDirect RDMA benchmark.
f50fbdda 1596 The :option:`ioengine` must be `rdma`.
03553853 1597
f80dba8d
MT
1598 The area allocated is a function of the maximum allowed bs size for the job,
1599 multiplied by the I/O depth given. Note that for **shmhuge** and
1600 **mmaphuge** to work, the system must have free huge pages allocated. This
1601 can normally be checked and set by reading/writing
1602 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1603 is 4MiB in size. So to calculate the number of huge pages you need for a
1604 given job file, add up the I/O depth of all jobs (normally one unless
1605 :option:`iodepth` is used) and multiply by the maximum bs set. Then divide
1606 that number by the huge page size. You can see the size of the huge pages in
1607 :file:`/proc/meminfo`. If no huge pages are allocated by having a non-zero
1608 number in `nr_hugepages`, using **mmaphuge** or **shmhuge** will fail. Also
1609 see :option:`hugepage-size`.
1610
1611 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1612 should point there. So if it's mounted in :file:`/huge`, you would use
1613 `mem=mmaphuge:/huge/somefile`.
1614
f50fbdda 1615.. option:: iomem_align=int, mem_align=int
f80dba8d
MT
1616
1617 This indicates the memory alignment of the I/O memory buffers. Note that
1618 the given alignment is applied to the first I/O unit buffer, if using
1619 :option:`iodepth` the alignment of the following buffers are given by the
1620 :option:`bs` used. In other words, if using a :option:`bs` that is a
1621 multiple of the page sized in the system, all buffers will be aligned to
1622 this value. If using a :option:`bs` that is not page aligned, the alignment
1623 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1624 :option:`bs` used.
1625
1626.. option:: hugepage-size=int
1627
1628 Defines the size of a huge page. Must at least be equal to the system
1629 setting, see :file:`/proc/meminfo`. Defaults to 4MiB. Should probably
1630 always be a multiple of megabytes, so using ``hugepage-size=Xm`` is the
1631 preferred way to set this to avoid setting a non-pow-2 bad value.
1632
1633.. option:: lockmem=int
1634
1635 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1636 simulate a smaller amount of memory. The amount specified is per worker.
1637
1638
1639I/O size
1640~~~~~~~~
1641
1642.. option:: size=int
1643
79591fa9
TK
1644 The total size of file I/O for each thread of this job. Fio will run until
1645 this many bytes has been transferred, unless runtime is limited by other options
1646 (such as :option:`runtime`, for instance, or increased/decreased by :option:`io_size`).
1647 Fio will divide this size between the available files determined by options
1648 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1649 specified by the job. If the result of division happens to be 0, the size is
c4aa2d08 1650 set to the physical size of the given files or devices if they exist.
79591fa9 1651 If this option is not specified, fio will use the full size of the given
f80dba8d
MT
1652 files or devices. If the files do not exist, size must be given. It is also
1653 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1654 given, fio will use 20% of the full size of the given files or devices.
9d25d068
SW
1655 Can be combined with :option:`offset` to constrain the start and end range
1656 that I/O will be done within.
f80dba8d
MT
1657
1658.. option:: io_size=int, io_limit=int
1659
1660 Normally fio operates within the region set by :option:`size`, which means
1661 that the :option:`size` option sets both the region and size of I/O to be
1662 performed. Sometimes that is not what you want. With this option, it is
1663 possible to define just the amount of I/O that fio should do. For instance,
1664 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1665 will perform I/O within the first 20GiB but exit when 5GiB have been
1666 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1667 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1668 the 0..20GiB region.
1669
7fdd97ca 1670.. option:: filesize=irange(int)
f80dba8d
MT
1671
1672 Individual file sizes. May be a range, in which case fio will select sizes
1673 for files at random within the given range and limited to :option:`size` in
1674 total (if that is given). If not given, each created file is the same size.
79591fa9
TK
1675 This option overrides :option:`size` in terms of file size, which means
1676 this value is used as a fixed size or possible range of each file.
f80dba8d
MT
1677
1678.. option:: file_append=bool
1679
1680 Perform I/O after the end of the file. Normally fio will operate within the
1681 size of a file. If this option is set, then fio will append to the file
1682 instead. This has identical behavior to setting :option:`offset` to the size
1683 of a file. This option is ignored on non-regular files.
1684
1685.. option:: fill_device=bool, fill_fs=bool
1686
1687 Sets size to something really large and waits for ENOSPC (no space left on
1688 device) as the terminating condition. Only makes sense with sequential
1689 write. For a read workload, the mount point will be filled first then I/O
1690 started on the result. This option doesn't make sense if operating on a raw
1691 device node, since the size of that is already known by the file system.
1692 Additionally, writing beyond end-of-device will not return ENOSPC there.
1693
1694
1695I/O engine
1696~~~~~~~~~~
1697
1698.. option:: ioengine=str
1699
1700 Defines how the job issues I/O to the file. The following types are defined:
1701
1702 **sync**
1703 Basic :manpage:`read(2)` or :manpage:`write(2)`
1704 I/O. :manpage:`lseek(2)` is used to position the I/O location.
54227e6b 1705 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
f80dba8d
MT
1706
1707 **psync**
1708 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1709 all supported operating systems except for Windows.
1710
1711 **vsync**
1712 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
c60ebc45 1713 queuing by coalescing adjacent I/Os into a single submission.
f80dba8d
MT
1714
1715 **pvsync**
1716 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1717
1718 **pvsync2**
1719 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1720
1721 **libaio**
1722 Linux native asynchronous I/O. Note that Linux may only support
22413915 1723 queued behavior with non-buffered I/O (set ``direct=1`` or
f80dba8d
MT
1724 ``buffered=0``).
1725 This engine defines engine specific options.
1726
1727 **posixaio**
1728 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
1729 :manpage:`aio_write(3)`.
1730
1731 **solarisaio**
1732 Solaris native asynchronous I/O.
1733
1734 **windowsaio**
1735 Windows native asynchronous I/O. Default on Windows.
1736
1737 **mmap**
1738 File is memory mapped with :manpage:`mmap(2)` and data copied
1739 to/from using :manpage:`memcpy(3)`.
1740
1741 **splice**
1742 :manpage:`splice(2)` is used to transfer the data and
1743 :manpage:`vmsplice(2)` to transfer data from user space to the
1744 kernel.
1745
1746 **sg**
1747 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
1748 ioctl, or if the target is an sg character device we use
1749 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
f50fbdda
TK
1750 I/O. Requires :option:`filename` option to specify either block or
1751 character devices.
52b81b7c 1752 The sg engine includes engine specific options.
f80dba8d
MT
1753
1754 **null**
1755 Doesn't transfer any data, just pretends to. This is mainly used to
1756 exercise fio itself and for debugging/testing purposes.
1757
1758 **net**
1759 Transfer over the network to given ``host:port``. Depending on the
1760 :option:`protocol` used, the :option:`hostname`, :option:`port`,
1761 :option:`listen` and :option:`filename` options are used to specify
1762 what sort of connection to make, while the :option:`protocol` option
1763 determines which protocol will be used. This engine defines engine
1764 specific options.
1765
1766 **netsplice**
1767 Like **net**, but uses :manpage:`splice(2)` and
1768 :manpage:`vmsplice(2)` to map data and send/receive.
1769 This engine defines engine specific options.
1770
1771 **cpuio**
1772 Doesn't transfer any data, but burns CPU cycles according to the
1773 :option:`cpuload` and :option:`cpuchunks` options. Setting
9207a0cb 1774 :option:`cpuload`\=85 will cause that job to do nothing but burn 85%
71aa48eb 1775 of the CPU. In case of SMP machines, use :option:`numjobs`\=<nr_of_cpu>
f50fbdda 1776 to get desired CPU usage, as the cpuload only loads a
f80dba8d
MT
1777 single CPU at the desired rate. A job never finishes unless there is
1778 at least one non-cpuio job.
1779
1780 **guasi**
804c0839 1781 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall
f80dba8d
MT
1782 Interface approach to async I/O. See
1783
1784 http://www.xmailserver.org/guasi-lib.html
1785
1786 for more info on GUASI.
1787
1788 **rdma**
1789 The RDMA I/O engine supports both RDMA memory semantics
1790 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
609ac152
SB
1791 InfiniBand, RoCE and iWARP protocols. This engine defines engine
1792 specific options.
f80dba8d
MT
1793
1794 **falloc**
1795 I/O engine that does regular fallocate to simulate data transfer as
1796 fio ioengine.
1797
1798 DDIR_READ
1799 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
1800
1801 DDIR_WRITE
1802 does fallocate(,mode = 0).
1803
1804 DDIR_TRIM
1805 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
1806
761cd093
SW
1807 **ftruncate**
1808 I/O engine that sends :manpage:`ftruncate(2)` operations in response
1809 to write (DDIR_WRITE) events. Each ftruncate issued sets the file's
f50fbdda 1810 size to the current block offset. :option:`blocksize` is ignored.
761cd093 1811
f80dba8d
MT
1812 **e4defrag**
1813 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
1814 defragment activity in request to DDIR_WRITE event.
1815
f3f96717
IF
1816 **rados**
1817 I/O engine supporting direct access to Ceph Reliable Autonomic
1818 Distributed Object Store (RADOS) via librados. This ioengine
1819 defines engine specific options.
1820
f80dba8d
MT
1821 **rbd**
1822 I/O engine supporting direct access to Ceph Rados Block Devices
1823 (RBD) via librbd without the need to use the kernel rbd driver. This
1824 ioengine defines engine specific options.
1825
1826 **gfapi**
ac8ca2af
SW
1827 Using GlusterFS libgfapi sync interface to direct access to
1828 GlusterFS volumes without having to go through FUSE. This ioengine
f80dba8d
MT
1829 defines engine specific options.
1830
1831 **gfapi_async**
ac8ca2af
SW
1832 Using GlusterFS libgfapi async interface to direct access to
1833 GlusterFS volumes without having to go through FUSE. This ioengine
f80dba8d
MT
1834 defines engine specific options.
1835
1836 **libhdfs**
f50fbdda 1837 Read and write through Hadoop (HDFS). The :option:`filename` option
f80dba8d
MT
1838 is used to specify host,port of the hdfs name-node to connect. This
1839 engine interprets offsets a little differently. In HDFS, files once
e25c0c91
SW
1840 created cannot be modified so random writes are not possible. To
1841 imitate this the libhdfs engine expects a bunch of small files to be
1842 created over HDFS and will randomly pick a file from them
1843 based on the offset generated by fio backend (see the example
f80dba8d 1844 job file to create such files, use ``rw=write`` option). Please
e25c0c91
SW
1845 note, it may be necessary to set environment variables to work
1846 with HDFS/libhdfs properly. Each job uses its own connection to
f80dba8d
MT
1847 HDFS.
1848
1849 **mtd**
1850 Read, write and erase an MTD character device (e.g.,
1851 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
1852 underlying device type, the I/O may have to go in a certain pattern,
1853 e.g., on NAND, writing sequentially to erase blocks and discarding
c298ee71 1854 before overwriting. The `trimwrite` mode works well for this
f80dba8d
MT
1855 constraint.
1856
1857 **pmemblk**
1858 Read and write using filesystem DAX to a file on a filesystem
363a5f65 1859 mounted with DAX on a persistent memory device through the PMDK
f80dba8d
MT
1860 libpmemblk library.
1861
1862 **dev-dax**
1863 Read and write using device DAX to a persistent memory device (e.g.,
363a5f65 1864 /dev/dax0.0) through the PMDK libpmem library.
f80dba8d
MT
1865
1866 **external**
1867 Prefix to specify loading an external I/O engine object file. Append
c60ebc45 1868 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
d243fd6d
TK
1869 ioengine :file:`foo.o` in :file:`/tmp`. The path can be either
1870 absolute or relative. See :file:`engines/skeleton_external.c` for
1871 details of writing an external I/O engine.
f80dba8d 1872
1216cc5a 1873 **filecreate**
b71968b1 1874 Simply create the files and do no I/O to them. You still need to
1216cc5a 1875 set `filesize` so that all the accounting still occurs, but no
b71968b1 1876 actual I/O will be done other than creating the file.
f80dba8d 1877
ae0db592
TI
1878 **libpmem**
1879 Read and write using mmap I/O to a file on a filesystem
363a5f65 1880 mounted with DAX on a persistent memory device through the PMDK
ae0db592
TI
1881 libpmem library.
1882
f80dba8d
MT
1883I/O engine specific parameters
1884~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1885
1886In addition, there are some parameters which are only valid when a specific
f50fbdda
TK
1887:option:`ioengine` is in use. These are used identically to normal parameters,
1888with the caveat that when used on the command line, they must come after the
f80dba8d
MT
1889:option:`ioengine` that defines them is selected.
1890
1891.. option:: userspace_reap : [libaio]
1892
1893 Normally, with the libaio engine in use, fio will use the
1894 :manpage:`io_getevents(2)` system call to reap newly returned events. With
1895 this flag turned on, the AIO ring will be read directly from user-space to
1896 reap events. The reaping mode is only enabled when polling for a minimum of
c60ebc45 1897 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
f80dba8d 1898
9d25d068 1899.. option:: hipri : [pvsync2]
f80dba8d
MT
1900
1901 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
1902 than normal.
1903
a0679ce5
SB
1904.. option:: hipri_percentage : [pvsync2]
1905
f50fbdda 1906 When hipri is set this determines the probability of a pvsync2 I/O being high
a0679ce5
SB
1907 priority. The default is 100%.
1908
f80dba8d
MT
1909.. option:: cpuload=int : [cpuio]
1910
da19cdb4
TK
1911 Attempt to use the specified percentage of CPU cycles. This is a mandatory
1912 option when using cpuio I/O engine.
f80dba8d
MT
1913
1914.. option:: cpuchunks=int : [cpuio]
1915
1916 Split the load into cycles of the given time. In microseconds.
1917
1918.. option:: exit_on_io_done=bool : [cpuio]
1919
1920 Detect when I/O threads are done, then exit.
1921
f80dba8d
MT
1922.. option:: namenode=str : [libhdfs]
1923
22413915 1924 The hostname or IP address of a HDFS cluster namenode to contact.
f80dba8d
MT
1925
1926.. option:: port=int
1927
f50fbdda
TK
1928 [libhdfs]
1929
1930 The listening port of the HFDS cluster namenode.
1931
f80dba8d
MT
1932 [netsplice], [net]
1933
1934 The TCP or UDP port to bind to or connect to. If this is used with
1935 :option:`numjobs` to spawn multiple instances of the same job type, then
1936 this will be the starting port number since fio will use a range of
1937 ports.
1938
609ac152
SB
1939 [rdma]
1940
1941 The port to use for RDMA-CM communication. This should be the same value
1942 on the client and the server side.
1943
1944.. option:: hostname=str : [netsplice] [net] [rdma]
f80dba8d 1945
609ac152
SB
1946 The hostname or IP address to use for TCP, UDP or RDMA-CM based I/O. If the job
1947 is a TCP listener or UDP reader, the hostname is not used and must be omitted
f50fbdda 1948 unless it is a valid UDP multicast address.
f80dba8d
MT
1949
1950.. option:: interface=str : [netsplice] [net]
1951
1952 The IP address of the network interface used to send or receive UDP
1953 multicast.
1954
1955.. option:: ttl=int : [netsplice] [net]
1956
1957 Time-to-live value for outgoing UDP multicast packets. Default: 1.
1958
1959.. option:: nodelay=bool : [netsplice] [net]
1960
1961 Set TCP_NODELAY on TCP connections.
1962
f50fbdda 1963.. option:: protocol=str, proto=str : [netsplice] [net]
f80dba8d
MT
1964
1965 The network protocol to use. Accepted values are:
1966
1967 **tcp**
1968 Transmission control protocol.
1969 **tcpv6**
1970 Transmission control protocol V6.
1971 **udp**
1972 User datagram protocol.
1973 **udpv6**
1974 User datagram protocol V6.
1975 **unix**
1976 UNIX domain socket.
1977
1978 When the protocol is TCP or UDP, the port must also be given, as well as the
1979 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
f50fbdda 1980 normal :option:`filename` option should be used and the port is invalid.
f80dba8d 1981
e9184ec1 1982.. option:: listen : [netsplice] [net]
f80dba8d
MT
1983
1984 For TCP network connections, tell fio to listen for incoming connections
1985 rather than initiating an outgoing connection. The :option:`hostname` must
1986 be omitted if this option is used.
1987
e9184ec1 1988.. option:: pingpong : [netsplice] [net]
f80dba8d
MT
1989
1990 Normally a network writer will just continue writing data, and a network
1991 reader will just consume packages. If ``pingpong=1`` is set, a writer will
1992 send its normal payload to the reader, then wait for the reader to send the
1993 same payload back. This allows fio to measure network latencies. The
1994 submission and completion latencies then measure local time spent sending or
1995 receiving, and the completion latency measures how long it took for the
1996 other end to receive and send back. For UDP multicast traffic
1997 ``pingpong=1`` should only be set for a single reader when multiple readers
1998 are listening to the same address.
1999
e9184ec1 2000.. option:: window_size : [netsplice] [net]
f80dba8d
MT
2001
2002 Set the desired socket buffer size for the connection.
2003
e9184ec1 2004.. option:: mss : [netsplice] [net]
f80dba8d
MT
2005
2006 Set the TCP maximum segment size (TCP_MAXSEG).
2007
2008.. option:: donorname=str : [e4defrag]
2009
730bd7d9 2010 File will be used as a block donor (swap extents between files).
f80dba8d
MT
2011
2012.. option:: inplace=int : [e4defrag]
2013
2014 Configure donor file blocks allocation strategy:
2015
2016 **0**
2017 Default. Preallocate donor's file on init.
2018 **1**
2b455dbf 2019 Allocate space immediately inside defragment event, and free right
f80dba8d
MT
2020 after event.
2021
f3f96717 2022.. option:: clustername=str : [rbd,rados]
f80dba8d
MT
2023
2024 Specifies the name of the Ceph cluster.
2025
2026.. option:: rbdname=str : [rbd]
2027
2028 Specifies the name of the RBD.
2029
f3f96717 2030.. option:: pool=str : [rbd,rados]
f80dba8d 2031
f3f96717 2032 Specifies the name of the Ceph pool containing RBD or RADOS data.
f80dba8d 2033
f3f96717 2034.. option:: clientname=str : [rbd,rados]
f80dba8d
MT
2035
2036 Specifies the username (without the 'client.' prefix) used to access the
2037 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
2038 the full *type.id* string. If no type. prefix is given, fio will add
2039 'client.' by default.
2040
f3f96717
IF
2041.. option:: busy_poll=bool : [rbd,rados]
2042
2043 Poll store instead of waiting for completion. Usually this provides better
2044 throughput at cost of higher(up to 100%) CPU utilization.
2045
f80dba8d
MT
2046.. option:: skip_bad=bool : [mtd]
2047
2048 Skip operations against known bad blocks.
2049
2050.. option:: hdfsdirectory : [libhdfs]
2051
2052 libhdfs will create chunk in this HDFS directory.
2053
2054.. option:: chunk_size : [libhdfs]
2055
2b455dbf 2056 The size of the chunk to use for each file.
f80dba8d 2057
609ac152
SB
2058.. option:: verb=str : [rdma]
2059
2060 The RDMA verb to use on this side of the RDMA ioengine connection. Valid
2061 values are write, read, send and recv. These correspond to the equivalent
2062 RDMA verbs (e.g. write = rdma_write etc.). Note that this only needs to be
2063 specified on the client side of the connection. See the examples folder.
2064
2065.. option:: bindname=str : [rdma]
2066
2067 The name to use to bind the local RDMA-CM connection to a local RDMA device.
2068 This could be a hostname or an IPv4 or IPv6 address. On the server side this
2069 will be passed into the rdma_bind_addr() function and on the client site it
2070 will be used in the rdma_resolve_add() function. This can be useful when
2071 multiple paths exist between the client and the server or in certain loopback
2072 configurations.
f80dba8d 2073
52b81b7c
KD
2074.. option:: readfua=bool : [sg]
2075
2076 With readfua option set to 1, read operations include
2077 the force unit access (fua) flag. Default is 0.
2078
2079.. option:: writefua=bool : [sg]
2080
2081 With writefua option set to 1, write operations include
2082 the force unit access (fua) flag. Default is 0.
2083
2c3a9150
VF
2084.. option:: sg_write_mode=str : [sg]
2085 Specify the type of write commands to issue. This option can take three values:
2086
2087 **write**
2088 This is the default where write opcodes are issued as usual.
2089 **verify**
2090 Issue WRITE AND VERIFY commands. The BYTCHK bit is set to 0. This
2091 directs the device to carry out a medium verification with no data
2092 comparison. The writefua option is ignored with this selection.
2093 **same**
2094 Issue WRITE SAME commands. This transfers a single block to the device
2095 and writes this same block of data to a contiguous sequence of LBAs
2096 beginning at the specified offset. fio's block size parameter specifies
2097 the amount of data written with each command. However, the amount of data
2098 actually transferred to the device is equal to the device's block
2099 (sector) size. For a device with 512 byte sectors, blocksize=8k will
2100 write 16 sectors with each command. fio will still generate 8k of data
2101 for each command but only the first 512 bytes will be used and
2102 transferred to the device. The writefua option is ignored with this
2103 selection.
52b81b7c 2104
f80dba8d
MT
2105I/O depth
2106~~~~~~~~~
2107
2108.. option:: iodepth=int
2109
2110 Number of I/O units to keep in flight against the file. Note that
2111 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
c60ebc45 2112 for small degrees when :option:`verify_async` is in use). Even async
f80dba8d
MT
2113 engines may impose OS restrictions causing the desired depth not to be
2114 achieved. This may happen on Linux when using libaio and not setting
9207a0cb 2115 :option:`direct`\=1, since buffered I/O is not async on that OS. Keep an
f80dba8d
MT
2116 eye on the I/O depth distribution in the fio output to verify that the
2117 achieved depth is as expected. Default: 1.
2118
2119.. option:: iodepth_batch_submit=int, iodepth_batch=int
2120
2121 This defines how many pieces of I/O to submit at once. It defaults to 1
2122 which means that we submit each I/O as soon as it is available, but can be
2123 raised to submit bigger batches of I/O at the time. If it is set to 0 the
2124 :option:`iodepth` value will be used.
2125
2126.. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
2127
2128 This defines how many pieces of I/O to retrieve at once. It defaults to 1
2129 which means that we'll ask for a minimum of 1 I/O in the retrieval process
2130 from the kernel. The I/O retrieval will go on until we hit the limit set by
2131 :option:`iodepth_low`. If this variable is set to 0, then fio will always
2132 check for completed events before queuing more I/O. This helps reduce I/O
2133 latency, at the cost of more retrieval system calls.
2134
2135.. option:: iodepth_batch_complete_max=int
2136
2137 This defines maximum pieces of I/O to retrieve at once. This variable should
9207a0cb 2138 be used along with :option:`iodepth_batch_complete_min`\=int variable,
f80dba8d 2139 specifying the range of min and max amount of I/O which should be
730bd7d9 2140 retrieved. By default it is equal to the :option:`iodepth_batch_complete_min`
f80dba8d
MT
2141 value.
2142
2143 Example #1::
2144
2145 iodepth_batch_complete_min=1
2146 iodepth_batch_complete_max=<iodepth>
2147
2148 which means that we will retrieve at least 1 I/O and up to the whole
2149 submitted queue depth. If none of I/O has been completed yet, we will wait.
2150
2151 Example #2::
2152
2153 iodepth_batch_complete_min=0
2154 iodepth_batch_complete_max=<iodepth>
2155
2156 which means that we can retrieve up to the whole submitted queue depth, but
2157 if none of I/O has been completed yet, we will NOT wait and immediately exit
2158 the system call. In this example we simply do polling.
2159
2160.. option:: iodepth_low=int
2161
2162 The low water mark indicating when to start filling the queue
2163 again. Defaults to the same as :option:`iodepth`, meaning that fio will
2164 attempt to keep the queue full at all times. If :option:`iodepth` is set to
c60ebc45 2165 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
f80dba8d
MT
2166 16 requests, it will let the depth drain down to 4 before starting to fill
2167 it again.
2168
997b5680
SW
2169.. option:: serialize_overlap=bool
2170
2171 Serialize in-flight I/Os that might otherwise cause or suffer from data races.
2172 When two or more I/Os are submitted simultaneously, there is no guarantee that
2173 the I/Os will be processed or completed in the submitted order. Further, if
2174 two or more of those I/Os are writes, any overlapping region between them can
2175 become indeterminate/undefined on certain storage. These issues can cause
2176 verification to fail erratically when at least one of the racing I/Os is
2177 changing data and the overlapping region has a non-zero size. Setting
2178 ``serialize_overlap`` tells fio to avoid provoking this behavior by explicitly
2179 serializing in-flight I/Os that have a non-zero overlap. Note that setting
ee21ebee 2180 this option can reduce both performance and the :option:`iodepth` achieved.
997b5680
SW
2181 Additionally this option does not work when :option:`io_submit_mode` is set to
2182 offload. Default: false.
2183
f80dba8d
MT
2184.. option:: io_submit_mode=str
2185
2186 This option controls how fio submits the I/O to the I/O engine. The default
2187 is `inline`, which means that the fio job threads submit and reap I/O
2188 directly. If set to `offload`, the job threads will offload I/O submission
2189 to a dedicated pool of I/O threads. This requires some coordination and thus
2190 has a bit of extra overhead, especially for lower queue depth I/O where it
2191 can increase latencies. The benefit is that fio can manage submission rates
2192 independently of the device completion rates. This avoids skewed latency
730bd7d9 2193 reporting if I/O gets backed up on the device side (the coordinated omission
f80dba8d
MT
2194 problem).
2195
2196
2197I/O rate
2198~~~~~~~~
2199
a881438b 2200.. option:: thinktime=time
f80dba8d 2201
f75ede1d
SW
2202 Stall the job for the specified period of time after an I/O has completed before issuing the
2203 next. May be used to simulate processing being done by an application.
947e0fe0 2204 When the unit is omitted, the value is interpreted in microseconds. See
f80dba8d
MT
2205 :option:`thinktime_blocks` and :option:`thinktime_spin`.
2206
a881438b 2207.. option:: thinktime_spin=time
f80dba8d
MT
2208
2209 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
2210 something with the data received, before falling back to sleeping for the
f75ede1d 2211 rest of the period specified by :option:`thinktime`. When the unit is
947e0fe0 2212 omitted, the value is interpreted in microseconds.
f80dba8d
MT
2213
2214.. option:: thinktime_blocks=int
2215
2216 Only valid if :option:`thinktime` is set - control how many blocks to issue,
f50fbdda
TK
2217 before waiting :option:`thinktime` usecs. If not set, defaults to 1 which will make
2218 fio wait :option:`thinktime` usecs after every block. This effectively makes any
f80dba8d 2219 queue depth setting redundant, since no more than 1 I/O will be queued
f50fbdda 2220 before we have to complete it and do our :option:`thinktime`. In other words, this
f80dba8d 2221 setting effectively caps the queue depth if the latter is larger.
71bfa161 2222
f80dba8d 2223.. option:: rate=int[,int][,int]
71bfa161 2224
f80dba8d
MT
2225 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
2226 suffix rules apply. Comma-separated values may be specified for reads,
2227 writes, and trims as described in :option:`blocksize`.
71bfa161 2228
b25b3464
SW
2229 For example, using `rate=1m,500k` would limit reads to 1MiB/sec and writes to
2230 500KiB/sec. Capping only reads or writes can be done with `rate=,500k` or
2231 `rate=500k,` where the former will only limit writes (to 500KiB/sec) and the
2232 latter will only limit reads.
2233
f80dba8d 2234.. option:: rate_min=int[,int][,int]
71bfa161 2235
f80dba8d
MT
2236 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
2237 to meet this requirement will cause the job to exit. Comma-separated values
2238 may be specified for reads, writes, and trims as described in
2239 :option:`blocksize`.
71bfa161 2240
f80dba8d 2241.. option:: rate_iops=int[,int][,int]
71bfa161 2242
f80dba8d
MT
2243 Cap the bandwidth to this number of IOPS. Basically the same as
2244 :option:`rate`, just specified independently of bandwidth. If the job is
2245 given a block size range instead of a fixed value, the smallest block size
2246 is used as the metric. Comma-separated values may be specified for reads,
2247 writes, and trims as described in :option:`blocksize`.
71bfa161 2248
f80dba8d 2249.. option:: rate_iops_min=int[,int][,int]
71bfa161 2250
f80dba8d
MT
2251 If fio doesn't meet this rate of I/O, it will cause the job to exit.
2252 Comma-separated values may be specified for reads, writes, and trims as
2253 described in :option:`blocksize`.
71bfa161 2254
f80dba8d 2255.. option:: rate_process=str
66c098b8 2256
f80dba8d
MT
2257 This option controls how fio manages rated I/O submissions. The default is
2258 `linear`, which submits I/O in a linear fashion with fixed delays between
c60ebc45 2259 I/Os that gets adjusted based on I/O completion rates. If this is set to
f80dba8d
MT
2260 `poisson`, fio will submit I/O based on a more real world random request
2261 flow, known as the Poisson process
2262 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
2263 10^6 / IOPS for the given workload.
71bfa161 2264
1a9bf814
JA
2265.. option:: rate_ignore_thinktime=bool
2266
2267 By default, fio will attempt to catch up to the specified rate setting,
2268 if any kind of thinktime setting was used. If this option is set, then
2269 fio will ignore the thinktime and continue doing IO at the specified
2270 rate, instead of entering a catch-up mode after thinktime is done.
2271
71bfa161 2272
f80dba8d
MT
2273I/O latency
2274~~~~~~~~~~~
71bfa161 2275
a881438b 2276.. option:: latency_target=time
71bfa161 2277
f80dba8d 2278 If set, fio will attempt to find the max performance point that the given
f75ede1d 2279 workload will run at while maintaining a latency below this target. When
947e0fe0 2280 the unit is omitted, the value is interpreted in microseconds. See
f75ede1d 2281 :option:`latency_window` and :option:`latency_percentile`.
71bfa161 2282
a881438b 2283.. option:: latency_window=time
71bfa161 2284
f80dba8d 2285 Used with :option:`latency_target` to specify the sample window that the job
f75ede1d 2286 is run at varying queue depths to test the performance. When the unit is
947e0fe0 2287 omitted, the value is interpreted in microseconds.
b4692828 2288
f80dba8d 2289.. option:: latency_percentile=float
71bfa161 2290
c60ebc45 2291 The percentage of I/Os that must fall within the criteria specified by
f80dba8d 2292 :option:`latency_target` and :option:`latency_window`. If not set, this
c60ebc45 2293 defaults to 100.0, meaning that all I/Os must be equal or below to the value
f80dba8d 2294 set by :option:`latency_target`.
71bfa161 2295
a881438b 2296.. option:: max_latency=time
71bfa161 2297
f75ede1d 2298 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
947e0fe0 2299 maximum latency. When the unit is omitted, the value is interpreted in
f75ede1d 2300 microseconds.
71bfa161 2301
f80dba8d 2302.. option:: rate_cycle=int
71bfa161 2303
f80dba8d 2304 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
a47b697c 2305 of milliseconds. Defaults to 1000.
71bfa161 2306
71bfa161 2307
f80dba8d
MT
2308I/O replay
2309~~~~~~~~~~
71bfa161 2310
f80dba8d 2311.. option:: write_iolog=str
c2b1e753 2312
f80dba8d
MT
2313 Write the issued I/O patterns to the specified file. See
2314 :option:`read_iolog`. Specify a separate file for each job, otherwise the
2315 iologs will be interspersed and the file may be corrupt.
c2b1e753 2316
f80dba8d 2317.. option:: read_iolog=str
71bfa161 2318
22413915 2319 Open an iolog with the specified filename and replay the I/O patterns it
f80dba8d
MT
2320 contains. This can be used to store a workload and replay it sometime
2321 later. The iolog given may also be a blktrace binary file, which allows fio
2322 to replay a workload captured by :command:`blktrace`. See
2323 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
2324 replay, the file needs to be turned into a blkparse binary data file first
2325 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
71bfa161 2326
589e88b7 2327.. option:: replay_no_stall=bool
71bfa161 2328
f80dba8d 2329 When replaying I/O with :option:`read_iolog` the default behavior is to
22413915 2330 attempt to respect the timestamps within the log and replay them with the
f80dba8d
MT
2331 appropriate delay between IOPS. By setting this variable fio will not
2332 respect the timestamps and attempt to replay them as fast as possible while
2333 still respecting ordering. The result is the same I/O pattern to a given
2334 device, but different timings.
71bfa161 2335
6dd7fa77
JA
2336.. option:: replay_time_scale=int
2337
2338 When replaying I/O with :option:`read_iolog`, fio will honor the
2339 original timing in the trace. With this option, it's possible to scale
2340 the time. It's a percentage option, if set to 50 it means run at 50%
2341 the original IO rate in the trace. If set to 200, run at twice the
2342 original IO rate. Defaults to 100.
2343
f80dba8d 2344.. option:: replay_redirect=str
b4692828 2345
f80dba8d
MT
2346 While replaying I/O patterns using :option:`read_iolog` the default behavior
2347 is to replay the IOPS onto the major/minor device that each IOP was recorded
2348 from. This is sometimes undesirable because on a different machine those
2349 major/minor numbers can map to a different device. Changing hardware on the
2350 same system can also result in a different major/minor mapping.
730bd7d9 2351 ``replay_redirect`` causes all I/Os to be replayed onto the single specified
f80dba8d 2352 device regardless of the device it was recorded
9207a0cb 2353 from. i.e. :option:`replay_redirect`\= :file:`/dev/sdc` would cause all I/O
f80dba8d
MT
2354 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
2355 multiple devices will be replayed onto a single device, if the trace
2356 contains multiple devices. If you want multiple devices to be replayed
2357 concurrently to multiple redirected devices you must blkparse your trace
2358 into separate traces and replay them with independent fio invocations.
2359 Unfortunately this also breaks the strict time ordering between multiple
2360 device accesses.
71bfa161 2361
f80dba8d 2362.. option:: replay_align=int
74929ac2 2363
f80dba8d
MT
2364 Force alignment of I/O offsets and lengths in a trace to this power of 2
2365 value.
3c54bc46 2366
f80dba8d 2367.. option:: replay_scale=int
3c54bc46 2368
f80dba8d 2369 Scale sector offsets down by this factor when replaying traces.
3c54bc46 2370
38f68906
JA
2371.. option:: replay_skip=str
2372
2373 Sometimes it's useful to skip certain IO types in a replay trace.
2374 This could be, for instance, eliminating the writes in the trace.
2375 Or not replaying the trims/discards, if you are redirecting to
2376 a device that doesn't support them. This option takes a comma
2377 separated list of read, write, trim, sync.
2378
3c54bc46 2379
f80dba8d
MT
2380Threads, processes and job synchronization
2381~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3c54bc46 2382
f80dba8d 2383.. option:: thread
3c54bc46 2384
730bd7d9
SW
2385 Fio defaults to creating jobs by using fork, however if this option is
2386 given, fio will create jobs by using POSIX Threads' function
2387 :manpage:`pthread_create(3)` to create threads instead.
71bfa161 2388
f80dba8d 2389.. option:: wait_for=str
74929ac2 2390
730bd7d9
SW
2391 If set, the current job won't be started until all workers of the specified
2392 waitee job are done.
74929ac2 2393
f80dba8d
MT
2394 ``wait_for`` operates on the job name basis, so there are a few
2395 limitations. First, the waitee must be defined prior to the waiter job
2396 (meaning no forward references). Second, if a job is being referenced as a
2397 waitee, it must have a unique name (no duplicate waitees).
74929ac2 2398
f80dba8d 2399.. option:: nice=int
892a6ffc 2400
f80dba8d 2401 Run the job with the given nice value. See man :manpage:`nice(2)`.
892a6ffc 2402
f80dba8d
MT
2403 On Windows, values less than -15 set the process class to "High"; -1 through
2404 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
2405 priority class.
74929ac2 2406
f80dba8d 2407.. option:: prio=int
71bfa161 2408
f80dba8d
MT
2409 Set the I/O priority value of this job. Linux limits us to a positive value
2410 between 0 and 7, with 0 being the highest. See man
2411 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
2412 systems since meaning of priority may differ.
71bfa161 2413
f80dba8d 2414.. option:: prioclass=int
d59aa780 2415
f80dba8d 2416 Set the I/O priority class. See man :manpage:`ionice(1)`.
d59aa780 2417
f80dba8d 2418.. option:: cpus_allowed=str
6d500c2e 2419
730bd7d9 2420 Controls the same options as :option:`cpumask`, but accepts a textual
b570e037
SW
2421 specification of the permitted CPUs instead and CPUs are indexed from 0. So
2422 to use CPUs 0 and 5 you would specify ``cpus_allowed=0,5``. This option also
2423 allows a range of CPUs to be specified -- say you wanted a binding to CPUs
2424 0, 5, and 8 to 15, you would set ``cpus_allowed=0,5,8-15``.
2425
2426 On Windows, when ``cpus_allowed`` is unset only CPUs from fio's current
2427 processor group will be used and affinity settings are inherited from the
2428 system. An fio build configured to target Windows 7 makes options that set
2429 CPUs processor group aware and values will set both the processor group
2430 and a CPU from within that group. For example, on a system where processor
2431 group 0 has 40 CPUs and processor group 1 has 32 CPUs, ``cpus_allowed``
2432 values between 0 and 39 will bind CPUs from processor group 0 and
2433 ``cpus_allowed`` values between 40 and 71 will bind CPUs from processor
2434 group 1. When using ``cpus_allowed_policy=shared`` all CPUs specified by a
2435 single ``cpus_allowed`` option must be from the same processor group. For
2436 Windows fio builds not built for Windows 7, CPUs will only be selected from
2437 (and be relative to) whatever processor group fio happens to be running in
2438 and CPUs from other processor groups cannot be used.
6d500c2e 2439
f80dba8d 2440.. option:: cpus_allowed_policy=str
6d500c2e 2441
f80dba8d 2442 Set the policy of how fio distributes the CPUs specified by
730bd7d9 2443 :option:`cpus_allowed` or :option:`cpumask`. Two policies are supported:
6d500c2e 2444
f80dba8d
MT
2445 **shared**
2446 All jobs will share the CPU set specified.
2447 **split**
2448 Each job will get a unique CPU from the CPU set.
6d500c2e 2449
22413915 2450 **shared** is the default behavior, if the option isn't specified. If
f80dba8d
MT
2451 **split** is specified, then fio will will assign one cpu per job. If not
2452 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
2453 in the set.
6d500c2e 2454
b570e037
SW
2455.. option:: cpumask=int
2456
2457 Set the CPU affinity of this job. The parameter given is a bit mask of
2458 allowed CPUs the job may run on. So if you want the allowed CPUs to be 1
2459 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
2460 :manpage:`sched_setaffinity(2)`. This may not work on all supported
2461 operating systems or kernel versions. This option doesn't work well for a
2462 higher CPU count than what you can store in an integer mask, so it can only
2463 control cpus 1-32. For boxes with larger CPU counts, use
2464 :option:`cpus_allowed`.
2465
f80dba8d 2466.. option:: numa_cpu_nodes=str
6d500c2e 2467
f80dba8d
MT
2468 Set this job running on specified NUMA nodes' CPUs. The arguments allow
2469 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
ac8ca2af 2470 NUMA options support, fio must be built on a system with libnuma-dev(el)
f80dba8d 2471 installed.
61b9861d 2472
f80dba8d 2473.. option:: numa_mem_policy=str
61b9861d 2474
f80dba8d
MT
2475 Set this job's memory policy and corresponding NUMA nodes. Format of the
2476 arguments::
5c94b008 2477
f80dba8d 2478 <mode>[:<nodelist>]
ce35b1ec 2479
804c0839 2480 ``mode`` is one of the following memory policies: ``default``, ``prefer``,
730bd7d9
SW
2481 ``bind``, ``interleave`` or ``local``. For ``default`` and ``local`` memory
2482 policies, no node needs to be specified. For ``prefer``, only one node is
2483 allowed. For ``bind`` and ``interleave`` the ``nodelist`` may be as
2484 follows: a comma delimited list of numbers, A-B ranges, or `all`.
71bfa161 2485
f80dba8d 2486.. option:: cgroup=str
390b1537 2487
f80dba8d
MT
2488 Add job to this control group. If it doesn't exist, it will be created. The
2489 system must have a mounted cgroup blkio mount point for this to work. If
2490 your system doesn't have it mounted, you can do so with::
5af1c6f3 2491
f80dba8d 2492 # mount -t cgroup -o blkio none /cgroup
5af1c6f3 2493
f80dba8d 2494.. option:: cgroup_weight=int
5af1c6f3 2495
f80dba8d
MT
2496 Set the weight of the cgroup to this value. See the documentation that comes
2497 with the kernel, allowed values are in the range of 100..1000.
a086c257 2498
f80dba8d 2499.. option:: cgroup_nodelete=bool
8c07860d 2500
f80dba8d
MT
2501 Normally fio will delete the cgroups it has created after the job
2502 completion. To override this behavior and to leave cgroups around after the
2503 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
2504 to inspect various cgroup files after job completion. Default: false.
8c07860d 2505
f80dba8d 2506.. option:: flow_id=int
8c07860d 2507
f80dba8d
MT
2508 The ID of the flow. If not specified, it defaults to being a global
2509 flow. See :option:`flow`.
1907dbc6 2510
f80dba8d 2511.. option:: flow=int
71bfa161 2512
f80dba8d
MT
2513 Weight in token-based flow control. If this value is used, then there is a
2514 'flow counter' which is used to regulate the proportion of activity between
2515 two or more jobs. Fio attempts to keep this flow counter near zero. The
2516 ``flow`` parameter stands for how much should be added or subtracted to the
2517 flow counter on each iteration of the main I/O loop. That is, if one job has
2518 ``flow=8`` and another job has ``flow=-1``, then there will be a roughly 1:8
2519 ratio in how much one runs vs the other.
71bfa161 2520
f80dba8d 2521.. option:: flow_watermark=int
a31041ea 2522
f80dba8d
MT
2523 The maximum value that the absolute value of the flow counter is allowed to
2524 reach before the job must wait for a lower value of the counter.
82407585 2525
f80dba8d 2526.. option:: flow_sleep=int
82407585 2527
f80dba8d
MT
2528 The period of time, in microseconds, to wait after the flow watermark has
2529 been exceeded before retrying operations.
82407585 2530
f80dba8d 2531.. option:: stonewall, wait_for_previous
82407585 2532
f80dba8d
MT
2533 Wait for preceding jobs in the job file to exit, before starting this
2534 one. Can be used to insert serialization points in the job file. A stone
2535 wall also implies starting a new reporting group, see
2536 :option:`group_reporting`.
2537
2538.. option:: exitall
2539
730bd7d9
SW
2540 By default, fio will continue running all other jobs when one job finishes
2541 but sometimes this is not the desired action. Setting ``exitall`` will
2542 instead make fio terminate all other jobs when one job finishes.
f80dba8d
MT
2543
2544.. option:: exec_prerun=str
2545
2546 Before running this job, issue the command specified through
2547 :manpage:`system(3)`. Output is redirected in a file called
2548 :file:`jobname.prerun.txt`.
2549
2550.. option:: exec_postrun=str
2551
2552 After the job completes, issue the command specified though
2553 :manpage:`system(3)`. Output is redirected in a file called
2554 :file:`jobname.postrun.txt`.
2555
2556.. option:: uid=int
2557
2558 Instead of running as the invoking user, set the user ID to this value
2559 before the thread/process does any work.
2560
2561.. option:: gid=int
2562
2563 Set group ID, see :option:`uid`.
2564
2565
2566Verification
2567~~~~~~~~~~~~
2568
2569.. option:: verify_only
2570
2571 Do not perform specified workload, only verify data still matches previous
2572 invocation of this workload. This option allows one to check data multiple
2573 times at a later date without overwriting it. This option makes sense only
2574 for workloads that write data, and does not support workloads with the
2575 :option:`time_based` option set.
2576
2577.. option:: do_verify=bool
2578
2579 Run the verify phase after a write phase. Only valid if :option:`verify` is
2580 set. Default: true.
2581
2582.. option:: verify=str
2583
2584 If writing to a file, fio can verify the file contents after each iteration
2585 of the job. Each verification method also implies verification of special
2586 header, which is written to the beginning of each block. This header also
2587 includes meta information, like offset of the block, block number, timestamp
2588 when block was written, etc. :option:`verify` can be combined with
2589 :option:`verify_pattern` option. The allowed values are:
2590
2591 **md5**
2592 Use an md5 sum of the data area and store it in the header of
2593 each block.
2594
2595 **crc64**
2596 Use an experimental crc64 sum of the data area and store it in the
2597 header of each block.
2598
2599 **crc32c**
a5896300
SW
2600 Use a crc32c sum of the data area and store it in the header of
2601 each block. This will automatically use hardware acceleration
2602 (e.g. SSE4.2 on an x86 or CRC crypto extensions on ARM64) but will
2603 fall back to software crc32c if none is found. Generally the
804c0839 2604 fastest checksum fio supports when hardware accelerated.
f80dba8d
MT
2605
2606 **crc32c-intel**
a5896300 2607 Synonym for crc32c.
f80dba8d
MT
2608
2609 **crc32**
2610 Use a crc32 sum of the data area and store it in the header of each
2611 block.
2612
2613 **crc16**
2614 Use a crc16 sum of the data area and store it in the header of each
2615 block.
2616
2617 **crc7**
2618 Use a crc7 sum of the data area and store it in the header of each
2619 block.
2620
2621 **xxhash**
2622 Use xxhash as the checksum function. Generally the fastest software
2623 checksum that fio supports.
2624
2625 **sha512**
2626 Use sha512 as the checksum function.
2627
2628 **sha256**
2629 Use sha256 as the checksum function.
2630
2631 **sha1**
2632 Use optimized sha1 as the checksum function.
82407585 2633
ae3a5acc
JA
2634 **sha3-224**
2635 Use optimized sha3-224 as the checksum function.
2636
2637 **sha3-256**
2638 Use optimized sha3-256 as the checksum function.
2639
2640 **sha3-384**
2641 Use optimized sha3-384 as the checksum function.
2642
2643 **sha3-512**
2644 Use optimized sha3-512 as the checksum function.
2645
f80dba8d
MT
2646 **meta**
2647 This option is deprecated, since now meta information is included in
2648 generic verification header and meta verification happens by
2649 default. For detailed information see the description of the
2650 :option:`verify` setting. This option is kept because of
2651 compatibility's sake with old configurations. Do not use it.
2652
2653 **pattern**
2654 Verify a strict pattern. Normally fio includes a header with some
2655 basic information and checksumming, but if this option is set, only
2656 the specific pattern set with :option:`verify_pattern` is verified.
2657
2658 **null**
2659 Only pretend to verify. Useful for testing internals with
9207a0cb 2660 :option:`ioengine`\=null, not for much else.
f80dba8d
MT
2661
2662 This option can be used for repeated burn-in tests of a system to make sure
2663 that the written data is also correctly read back. If the data direction
2664 given is a read or random read, fio will assume that it should verify a
2665 previously written file. If the data direction includes any form of write,
2666 the verify will be of the newly written data.
2667
47e6a6e5
SW
2668 To avoid false verification errors, do not use the norandommap option when
2669 verifying data with async I/O engines and I/O depths > 1. Or use the
2670 norandommap and the lfsr random generator together to avoid writing to the
2671 same offset with muliple outstanding I/Os.
2672
f80dba8d
MT
2673.. option:: verify_offset=int
2674
2675 Swap the verification header with data somewhere else in the block before
2676 writing. It is swapped back before verifying.
2677
2678.. option:: verify_interval=int
2679
2680 Write the verification header at a finer granularity than the
2681 :option:`blocksize`. It will be written for chunks the size of
2682 ``verify_interval``. :option:`blocksize` should divide this evenly.
2683
2684.. option:: verify_pattern=str
2685
2686 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
2687 filling with totally random bytes, but sometimes it's interesting to fill
2688 with a known pattern for I/O verification purposes. Depending on the width
730bd7d9 2689 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time (it can
f80dba8d
MT
2690 be either a decimal or a hex number). The ``verify_pattern`` if larger than
2691 a 32-bit quantity has to be a hex number that starts with either "0x" or
2692 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
2693 format, which means that for each block offset will be written and then
2694 verified back, e.g.::
61b9861d
RP
2695
2696 verify_pattern=%o
2697
f80dba8d
MT
2698 Or use combination of everything::
2699
61b9861d 2700 verify_pattern=0xff%o"abcd"-12
e28218f3 2701
f80dba8d
MT
2702.. option:: verify_fatal=bool
2703
2704 Normally fio will keep checking the entire contents before quitting on a
2705 block verification failure. If this option is set, fio will exit the job on
2706 the first observed failure. Default: false.
2707
2708.. option:: verify_dump=bool
2709
2710 If set, dump the contents of both the original data block and the data block
2711 we read off disk to files. This allows later analysis to inspect just what
2712 kind of data corruption occurred. Off by default.
2713
2714.. option:: verify_async=int
2715
2716 Fio will normally verify I/O inline from the submitting thread. This option
2717 takes an integer describing how many async offload threads to create for I/O
2718 verification instead, causing fio to offload the duty of verifying I/O
2719 contents to one or more separate threads. If using this offload option, even
2720 sync I/O engines can benefit from using an :option:`iodepth` setting higher
2721 than 1, as it allows them to have I/O in flight while verifies are running.
d7e6ea1c 2722 Defaults to 0 async threads, i.e. verification is not asynchronous.
f80dba8d
MT
2723
2724.. option:: verify_async_cpus=str
2725
2726 Tell fio to set the given CPU affinity on the async I/O verification
2727 threads. See :option:`cpus_allowed` for the format used.
2728
2729.. option:: verify_backlog=int
2730
2731 Fio will normally verify the written contents of a job that utilizes verify
2732 once that job has completed. In other words, everything is written then
2733 everything is read back and verified. You may want to verify continually
2734 instead for a variety of reasons. Fio stores the meta data associated with
2735 an I/O block in memory, so for large verify workloads, quite a bit of memory
2736 would be used up holding this meta data. If this option is enabled, fio will
2737 write only N blocks before verifying these blocks.
2738
2739.. option:: verify_backlog_batch=int
2740
2741 Control how many blocks fio will verify if :option:`verify_backlog` is
2742 set. If not set, will default to the value of :option:`verify_backlog`
2743 (meaning the entire queue is read back and verified). If
2744 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
2745 blocks will be verified, if ``verify_backlog_batch`` is larger than
2746 :option:`verify_backlog`, some blocks will be verified more than once.
2747
2748.. option:: verify_state_save=bool
2749
2750 When a job exits during the write phase of a verify workload, save its
2751 current state. This allows fio to replay up until that point, if the verify
2752 state is loaded for the verify read phase. The format of the filename is,
2753 roughly::
2754
f50fbdda 2755 <type>-<jobname>-<jobindex>-verify.state.
f80dba8d
MT
2756
2757 <type> is "local" for a local run, "sock" for a client/server socket
2758 connection, and "ip" (192.168.0.1, for instance) for a networked
d7e6ea1c 2759 client/server connection. Defaults to true.
f80dba8d
MT
2760
2761.. option:: verify_state_load=bool
2762
2763 If a verify termination trigger was used, fio stores the current write state
2764 of each thread. This can be used at verification time so that fio knows how
2765 far it should verify. Without this information, fio will run a full
a47b697c
SW
2766 verification pass, according to the settings in the job file used. Default
2767 false.
f80dba8d
MT
2768
2769.. option:: trim_percentage=int
2770
2771 Number of verify blocks to discard/trim.
2772
2773.. option:: trim_verify_zero=bool
2774
22413915 2775 Verify that trim/discarded blocks are returned as zeros.
f80dba8d
MT
2776
2777.. option:: trim_backlog=int
2778
5cfd1e9a 2779 Trim after this number of blocks are written.
f80dba8d
MT
2780
2781.. option:: trim_backlog_batch=int
2782
2783 Trim this number of I/O blocks.
2784
2785.. option:: experimental_verify=bool
2786
2787 Enable experimental verification.
2788
f80dba8d
MT
2789Steady state
2790~~~~~~~~~~~~
2791
2792.. option:: steadystate=str:float, ss=str:float
2793
2794 Define the criterion and limit for assessing steady state performance. The
2795 first parameter designates the criterion whereas the second parameter sets
2796 the threshold. When the criterion falls below the threshold for the
2797 specified duration, the job will stop. For example, `iops_slope:0.1%` will
2798 direct fio to terminate the job when the least squares regression slope
2799 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
2800 this will apply to all jobs in the group. Below is the list of available
2801 steady state assessment criteria. All assessments are carried out using only
2802 data from the rolling collection window. Threshold limits can be expressed
2803 as a fixed value or as a percentage of the mean in the collection window.
2804
2805 **iops**
2806 Collect IOPS data. Stop the job if all individual IOPS measurements
2807 are within the specified limit of the mean IOPS (e.g., ``iops:2``
2808 means that all individual IOPS values must be within 2 of the mean,
2809 whereas ``iops:0.2%`` means that all individual IOPS values must be
2810 within 0.2% of the mean IOPS to terminate the job).
2811
2812 **iops_slope**
2813 Collect IOPS data and calculate the least squares regression
2814 slope. Stop the job if the slope falls below the specified limit.
2815
2816 **bw**
2817 Collect bandwidth data. Stop the job if all individual bandwidth
2818 measurements are within the specified limit of the mean bandwidth.
2819
2820 **bw_slope**
2821 Collect bandwidth data and calculate the least squares regression
2822 slope. Stop the job if the slope falls below the specified limit.
2823
2824.. option:: steadystate_duration=time, ss_dur=time
2825
2826 A rolling window of this duration will be used to judge whether steady state
2827 has been reached. Data will be collected once per second. The default is 0
f75ede1d 2828 which disables steady state detection. When the unit is omitted, the
947e0fe0 2829 value is interpreted in seconds.
f80dba8d
MT
2830
2831.. option:: steadystate_ramp_time=time, ss_ramp=time
2832
2833 Allow the job to run for the specified duration before beginning data
2834 collection for checking the steady state job termination criterion. The
947e0fe0 2835 default is 0. When the unit is omitted, the value is interpreted in seconds.
f80dba8d
MT
2836
2837
2838Measurements and reporting
2839~~~~~~~~~~~~~~~~~~~~~~~~~~
2840
2841.. option:: per_job_logs=bool
2842
2843 If set, this generates bw/clat/iops log with per file private filenames. If
2844 not set, jobs with identical names will share the log filename. Default:
2845 true.
2846
2847.. option:: group_reporting
2848
2849 It may sometimes be interesting to display statistics for groups of jobs as
2850 a whole instead of for each individual job. This is especially true if
2851 :option:`numjobs` is used; looking at individual thread/process output
2852 quickly becomes unwieldy. To see the final report per-group instead of
2853 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
2854 same reporting group, unless if separated by a :option:`stonewall`, or by
2855 using :option:`new_group`.
2856
2857.. option:: new_group
2858
2859 Start a new reporting group. See: :option:`group_reporting`. If not given,
2860 all jobs in a file will be part of the same reporting group, unless
2861 separated by a :option:`stonewall`.
2862
589e88b7 2863.. option:: stats=bool
8243be59
JA
2864
2865 By default, fio collects and shows final output results for all jobs
2866 that run. If this option is set to 0, then fio will ignore it in
2867 the final stat output.
2868
f80dba8d
MT
2869.. option:: write_bw_log=str
2870
2871 If given, write a bandwidth log for this job. Can be used to store data of
074f0817 2872 the bandwidth of the jobs in their lifetime.
f80dba8d 2873
074f0817
SW
2874 If no str argument is given, the default filename of
2875 :file:`jobname_type.x.log` is used. Even when the argument is given, fio
2876 will still append the type of log. So if one specifies::
2877
2878 write_bw_log=foo
f80dba8d 2879
074f0817
SW
2880 The actual log name will be :file:`foo_bw.x.log` where `x` is the index
2881 of the job (`1..N`, where `N` is the number of jobs). If
2882 :option:`per_job_logs` is false, then the filename will not include the
2883 `.x` job index.
e3cedca7 2884
074f0817
SW
2885 The included :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
2886 text files into nice graphs. See `Log File Formats`_ for how data is
2887 structured within the file.
2888
2889.. option:: write_lat_log=str
e3cedca7 2890
074f0817 2891 Same as :option:`write_bw_log`, except this option creates I/O
77b7e675
SW
2892 submission (e.g., :file:`name_slat.x.log`), completion (e.g.,
2893 :file:`name_clat.x.log`), and total (e.g., :file:`name_lat.x.log`)
074f0817
SW
2894 latency files instead. See :option:`write_bw_log` for details about
2895 the filename format and `Log File Formats`_ for how data is structured
2896 within the files.
be4ecfdf 2897
f80dba8d 2898.. option:: write_hist_log=str
06842027 2899
074f0817 2900 Same as :option:`write_bw_log` but writes an I/O completion latency
77b7e675 2901 histogram file (e.g., :file:`name_hist.x.log`) instead. Note that this
074f0817
SW
2902 file will be empty unless :option:`log_hist_msec` has also been set.
2903 See :option:`write_bw_log` for details about the filename format and
2904 `Log File Formats`_ for how data is structured within the file.
06842027 2905
f80dba8d 2906.. option:: write_iops_log=str
06842027 2907
074f0817 2908 Same as :option:`write_bw_log`, but writes an IOPS file (e.g.
77b7e675 2909 :file:`name_iops.x.log`) instead. See :option:`write_bw_log` for
074f0817
SW
2910 details about the filename format and `Log File Formats`_ for how data
2911 is structured within the file.
06842027 2912
f80dba8d 2913.. option:: log_avg_msec=int
06842027 2914
f80dba8d
MT
2915 By default, fio will log an entry in the iops, latency, or bw log for every
2916 I/O that completes. When writing to the disk log, that can quickly grow to a
2917 very large size. Setting this option makes fio average the each log entry
2918 over the specified period of time, reducing the resolution of the log. See
2919 :option:`log_max_value` as well. Defaults to 0, logging all entries.
6fc82095 2920 Also see `Log File Formats`_.
06842027 2921
f80dba8d 2922.. option:: log_hist_msec=int
06842027 2923
f80dba8d
MT
2924 Same as :option:`log_avg_msec`, but logs entries for completion latency
2925 histograms. Computing latency percentiles from averages of intervals using
c60ebc45 2926 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
f80dba8d
MT
2927 histogram entries over the specified period of time, reducing log sizes for
2928 high IOPS devices while retaining percentile accuracy. See
074f0817
SW
2929 :option:`log_hist_coarseness` and :option:`write_hist_log` as well.
2930 Defaults to 0, meaning histogram logging is disabled.
06842027 2931
f80dba8d 2932.. option:: log_hist_coarseness=int
06842027 2933
f80dba8d
MT
2934 Integer ranging from 0 to 6, defining the coarseness of the resolution of
2935 the histogram logs enabled with :option:`log_hist_msec`. For each increment
2936 in coarseness, fio outputs half as many bins. Defaults to 0, for which
074f0817
SW
2937 histogram logs contain 1216 latency bins. See :option:`write_hist_log`
2938 and `Log File Formats`_.
8b28bd41 2939
f80dba8d 2940.. option:: log_max_value=bool
66c098b8 2941
f80dba8d
MT
2942 If :option:`log_avg_msec` is set, fio logs the average over that window. If
2943 you instead want to log the maximum value, set this option to 1. Defaults to
2944 0, meaning that averaged values are logged.
a696fa2a 2945
589e88b7 2946.. option:: log_offset=bool
a696fa2a 2947
f80dba8d 2948 If this is set, the iolog options will include the byte offset for the I/O
5a83478f
SW
2949 entry as well as the other data values. Defaults to 0 meaning that
2950 offsets are not present in logs. Also see `Log File Formats`_.
71bfa161 2951
f80dba8d 2952.. option:: log_compression=int
7de87099 2953
f80dba8d
MT
2954 If this is set, fio will compress the I/O logs as it goes, to keep the
2955 memory footprint lower. When a log reaches the specified size, that chunk is
2956 removed and compressed in the background. Given that I/O logs are fairly
2957 highly compressible, this yields a nice memory savings for longer runs. The
2958 downside is that the compression will consume some background CPU cycles, so
2959 it may impact the run. This, however, is also true if the logging ends up
2960 consuming most of the system memory. So pick your poison. The I/O logs are
2961 saved normally at the end of a run, by decompressing the chunks and storing
2962 them in the specified log file. This feature depends on the availability of
2963 zlib.
e0b0d892 2964
f80dba8d 2965.. option:: log_compression_cpus=str
e0b0d892 2966
f80dba8d
MT
2967 Define the set of CPUs that are allowed to handle online log compression for
2968 the I/O jobs. This can provide better isolation between performance
0cf90a62
SW
2969 sensitive jobs, and background compression work. See
2970 :option:`cpus_allowed` for the format used.
9e684a49 2971
f80dba8d 2972.. option:: log_store_compressed=bool
9e684a49 2973
f80dba8d
MT
2974 If set, fio will store the log files in a compressed format. They can be
2975 decompressed with fio, using the :option:`--inflate-log` command line
2976 parameter. The files will be stored with a :file:`.fz` suffix.
9e684a49 2977
f80dba8d 2978.. option:: log_unix_epoch=bool
9e684a49 2979
f80dba8d
MT
2980 If set, fio will log Unix timestamps to the log files produced by enabling
2981 write_type_log for each log type, instead of the default zero-based
2982 timestamps.
2983
2984.. option:: block_error_percentiles=bool
2985
2986 If set, record errors in trim block-sized units from writes and trims and
2987 output a histogram of how many trims it took to get to errors, and what kind
2988 of error was encountered.
2989
2990.. option:: bwavgtime=int
2991
2992 Average the calculated bandwidth over the given time. Value is specified in
2993 milliseconds. If the job also does bandwidth logging through
2994 :option:`write_bw_log`, then the minimum of this option and
2995 :option:`log_avg_msec` will be used. Default: 500ms.
2996
2997.. option:: iopsavgtime=int
2998
2999 Average the calculated IOPS over the given time. Value is specified in
3000 milliseconds. If the job also does IOPS logging through
3001 :option:`write_iops_log`, then the minimum of this option and
3002 :option:`log_avg_msec` will be used. Default: 500ms.
3003
3004.. option:: disk_util=bool
3005
3006 Generate disk utilization statistics, if the platform supports it.
3007 Default: true.
3008
3009.. option:: disable_lat=bool
3010
3011 Disable measurements of total latency numbers. Useful only for cutting back
3012 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
3013 performance at really high IOPS rates. Note that to really get rid of a
3014 large amount of these calls, this option must be used with
f75ede1d 3015 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
f80dba8d
MT
3016
3017.. option:: disable_clat=bool
3018
3019 Disable measurements of completion latency numbers. See
3020 :option:`disable_lat`.
3021
3022.. option:: disable_slat=bool
3023
3024 Disable measurements of submission latency numbers. See
f50fbdda 3025 :option:`disable_lat`.
f80dba8d 3026
f75ede1d 3027.. option:: disable_bw_measurement=bool, disable_bw=bool
f80dba8d
MT
3028
3029 Disable measurements of throughput/bandwidth numbers. See
3030 :option:`disable_lat`.
3031
3032.. option:: clat_percentiles=bool
3033
b599759b
JA
3034 Enable the reporting of percentiles of completion latencies. This
3035 option is mutually exclusive with :option:`lat_percentiles`.
3036
3037.. option:: lat_percentiles=bool
3038
b71968b1 3039 Enable the reporting of percentiles of I/O latencies. This is similar
b599759b
JA
3040 to :option:`clat_percentiles`, except that this includes the
3041 submission latency. This option is mutually exclusive with
3042 :option:`clat_percentiles`.
f80dba8d
MT
3043
3044.. option:: percentile_list=float_list
3045
c32ba107
JA
3046 Overwrite the default list of percentiles for completion latencies and
3047 the block error histogram. Each number is a floating number in the
3048 range (0,100], and the maximum length of the list is 20. Use ``:`` to
3049 separate the numbers, and list the numbers in ascending order. For
3050 example, ``--percentile_list=99.5:99.9`` will cause fio to report the
3051 values of completion latency below which 99.5% and 99.9% of the observed
3052 latencies fell, respectively.
f80dba8d 3053
e883cb35
JF
3054.. option:: significant_figures=int
3055
c32ba107
JA
3056 If using :option:`--output-format` of `normal`, set the significant
3057 figures to this value. Higher values will yield more precise IOPS and
3058 throughput units, while lower values will round. Requires a minimum
3059 value of 1 and a maximum value of 10. Defaults to 4.
e883cb35 3060
f80dba8d
MT
3061
3062Error handling
3063~~~~~~~~~~~~~~
3064
3065.. option:: exitall_on_error
3066
3067 When one job finishes in error, terminate the rest. The default is to wait
3068 for each job to finish.
3069
3070.. option:: continue_on_error=str
3071
3072 Normally fio will exit the job on the first observed failure. If this option
3073 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
3074 EILSEQ) until the runtime is exceeded or the I/O size specified is
3075 completed. If this option is used, there are two more stats that are
3076 appended, the total error count and the first error. The error field given
3077 in the stats is the first error that was hit during the run.
3078
3079 The allowed values are:
3080
3081 **none**
3082 Exit on any I/O or verify errors.
3083
3084 **read**
3085 Continue on read errors, exit on all others.
3086
3087 **write**
3088 Continue on write errors, exit on all others.
3089
3090 **io**
3091 Continue on any I/O error, exit on all others.
3092
3093 **verify**
3094 Continue on verify errors, exit on all others.
3095
3096 **all**
3097 Continue on all errors.
3098
3099 **0**
3100 Backward-compatible alias for 'none'.
3101
3102 **1**
3103 Backward-compatible alias for 'all'.
3104
3105.. option:: ignore_error=str
3106
3107 Sometimes you want to ignore some errors during test in that case you can
a35ef7cb
TK
3108 specify error list for each error type, instead of only being able to
3109 ignore the default 'non-fatal error' using :option:`continue_on_error`.
f80dba8d
MT
3110 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
3111 given error type is separated with ':'. Error may be symbol ('ENOSPC',
3112 'ENOMEM') or integer. Example::
3113
3114 ignore_error=EAGAIN,ENOSPC:122
3115
3116 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
a35ef7cb
TK
3117 WRITE. This option works by overriding :option:`continue_on_error` with
3118 the list of errors for each error type if any.
f80dba8d
MT
3119
3120.. option:: error_dump=bool
3121
3122 If set dump every error even if it is non fatal, true by default. If
3123 disabled only fatal error will be dumped.
3124
f75ede1d
SW
3125Running predefined workloads
3126----------------------------
3127
3128Fio includes predefined profiles that mimic the I/O workloads generated by
3129other tools.
3130
3131.. option:: profile=str
3132
3133 The predefined workload to run. Current profiles are:
3134
3135 **tiobench**
3136 Threaded I/O bench (tiotest/tiobench) like workload.
3137
3138 **act**
3139 Aerospike Certification Tool (ACT) like workload.
3140
3141To view a profile's additional options use :option:`--cmdhelp` after specifying
3142the profile. For example::
3143
f50fbdda 3144 $ fio --profile=act --cmdhelp
f75ede1d
SW
3145
3146Act profile options
3147~~~~~~~~~~~~~~~~~~~
3148
3149.. option:: device-names=str
3150 :noindex:
3151
3152 Devices to use.
3153
3154.. option:: load=int
3155 :noindex:
3156
3157 ACT load multiplier. Default: 1.
3158
3159.. option:: test-duration=time
3160 :noindex:
3161
947e0fe0
SW
3162 How long the entire test takes to run. When the unit is omitted, the value
3163 is given in seconds. Default: 24h.
f75ede1d
SW
3164
3165.. option:: threads-per-queue=int
3166 :noindex:
3167
f50fbdda 3168 Number of read I/O threads per device. Default: 8.
f75ede1d
SW
3169
3170.. option:: read-req-num-512-blocks=int
3171 :noindex:
3172
3173 Number of 512B blocks to read at the time. Default: 3.
3174
3175.. option:: large-block-op-kbytes=int
3176 :noindex:
3177
3178 Size of large block ops in KiB (writes). Default: 131072.
3179
3180.. option:: prep
3181 :noindex:
3182
3183 Set to run ACT prep phase.
3184
3185Tiobench profile options
3186~~~~~~~~~~~~~~~~~~~~~~~~
3187
3188.. option:: size=str
3189 :noindex:
3190
f50fbdda 3191 Size in MiB.
f75ede1d
SW
3192
3193.. option:: block=int
3194 :noindex:
3195
3196 Block size in bytes. Default: 4096.
3197
3198.. option:: numruns=int
3199 :noindex:
3200
3201 Number of runs.
3202
3203.. option:: dir=str
3204 :noindex:
3205
3206 Test directory.
3207
3208.. option:: threads=int
3209 :noindex:
3210
3211 Number of threads.
f80dba8d
MT
3212
3213Interpreting the output
3214-----------------------
3215
36214730
SW
3216..
3217 Example output was based on the following:
3218 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --time_based \
3219 --rate=1256k --bs=14K --name=quick --runtime=1s --name=mixed \
3220 --runtime=2m --rw=rw
3221
f80dba8d
MT
3222Fio spits out a lot of output. While running, fio will display the status of the
3223jobs created. An example of that would be::
3224
9d25d068 3225 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 3226
36214730
SW
3227The characters inside the first set of square brackets denote the current status of
3228each thread. The first character is the first job defined in the job file, and so
3229forth. The possible values (in typical life cycle order) are:
f80dba8d
MT
3230
3231+------+-----+-----------------------------------------------------------+
3232| Idle | Run | |
3233+======+=====+===========================================================+
3234| P | | Thread setup, but not started. |
3235+------+-----+-----------------------------------------------------------+
3236| C | | Thread created. |
3237+------+-----+-----------------------------------------------------------+
3238| I | | Thread initialized, waiting or generating necessary data. |
3239+------+-----+-----------------------------------------------------------+
3240| | p | Thread running pre-reading file(s). |
3241+------+-----+-----------------------------------------------------------+
36214730
SW
3242| | / | Thread is in ramp period. |
3243+------+-----+-----------------------------------------------------------+
f80dba8d
MT
3244| | R | Running, doing sequential reads. |
3245+------+-----+-----------------------------------------------------------+
3246| | r | Running, doing random reads. |
3247+------+-----+-----------------------------------------------------------+
3248| | W | Running, doing sequential writes. |
3249+------+-----+-----------------------------------------------------------+
3250| | w | Running, doing random writes. |
3251+------+-----+-----------------------------------------------------------+
3252| | M | Running, doing mixed sequential reads/writes. |
3253+------+-----+-----------------------------------------------------------+
3254| | m | Running, doing mixed random reads/writes. |
3255+------+-----+-----------------------------------------------------------+
36214730
SW
3256| | D | Running, doing sequential trims. |
3257+------+-----+-----------------------------------------------------------+
3258| | d | Running, doing random trims. |
3259+------+-----+-----------------------------------------------------------+
3260| | F | Running, currently waiting for :manpage:`fsync(2)`. |
f80dba8d
MT
3261+------+-----+-----------------------------------------------------------+
3262| | V | Running, doing verification of written data. |
3263+------+-----+-----------------------------------------------------------+
36214730
SW
3264| f | | Thread finishing. |
3265+------+-----+-----------------------------------------------------------+
f80dba8d
MT
3266| E | | Thread exited, not reaped by main thread yet. |
3267+------+-----+-----------------------------------------------------------+
36214730 3268| _ | | Thread reaped. |
f80dba8d
MT
3269+------+-----+-----------------------------------------------------------+
3270| X | | Thread reaped, exited with an error. |
3271+------+-----+-----------------------------------------------------------+
3272| K | | Thread reaped, exited due to signal. |
3273+------+-----+-----------------------------------------------------------+
3274
36214730
SW
3275..
3276 Example output was based on the following:
3277 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --runtime=58m \
3278 --time_based --rate=2512k --bs=256K --numjobs=10 \
3279 --name=readers --rw=read --name=writers --rw=write
3280
f80dba8d 3281Fio will condense the thread string as not to take up more space on the command
36214730 3282line than needed. For instance, if you have 10 readers and 10 writers running,
f80dba8d
MT
3283the output would look like this::
3284
9d25d068 3285 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 3286
36214730
SW
3287Note that the status string is displayed in order, so it's possible to tell which of
3288the jobs are currently doing what. In the example above this means that jobs 1--10
3289are readers and 11--20 are writers.
f80dba8d
MT
3290
3291The other values are fairly self explanatory -- number of threads currently
36214730
SW
3292running and doing I/O, the number of currently open files (f=), the estimated
3293completion percentage, the rate of I/O since last check (read speed listed first,
f50fbdda
TK
3294then write speed and optionally trim speed) in terms of bandwidth and IOPS,
3295and time to completion for the current running group. It's impossible to estimate
3296runtime of the following groups (if any).
36214730
SW
3297
3298..
3299 Example output was based on the following:
3300 TZ=UTC fio --iodepth=16 --ioengine=posixaio --filename=/tmp/fiofile \
3301 --direct=1 --size=100M --time_based --runtime=50s --rate_iops=89 \
3302 --bs=7K --name=Client1 --rw=write
3303
3304When fio is done (or interrupted by :kbd:`Ctrl-C`), it will show the data for
3305each thread, group of threads, and disks in that order. For each overall thread (or
3306group) the output looks like::
3307
3308 Client1: (groupid=0, jobs=1): err= 0: pid=16109: Sat Jun 24 12:07:54 2017
3309 write: IOPS=88, BW=623KiB/s (638kB/s)(30.4MiB/50032msec)
3310 slat (nsec): min=500, max=145500, avg=8318.00, stdev=4781.50
3311 clat (usec): min=170, max=78367, avg=4019.02, stdev=8293.31
3312 lat (usec): min=174, max=78375, avg=4027.34, stdev=8291.79
3313 clat percentiles (usec):
3314 | 1.00th=[ 302], 5.00th=[ 326], 10.00th=[ 343], 20.00th=[ 363],
3315 | 30.00th=[ 392], 40.00th=[ 404], 50.00th=[ 416], 60.00th=[ 445],
3316 | 70.00th=[ 816], 80.00th=[ 6718], 90.00th=[12911], 95.00th=[21627],
3317 | 99.00th=[43779], 99.50th=[51643], 99.90th=[68682], 99.95th=[72877],
3318 | 99.99th=[78119]
3319 bw ( KiB/s): min= 532, max= 686, per=0.10%, avg=622.87, stdev=24.82, samples= 100
3320 iops : min= 76, max= 98, avg=88.98, stdev= 3.54, samples= 100
29092211
VF
3321 lat (usec) : 250=0.04%, 500=64.11%, 750=4.81%, 1000=2.79%
3322 lat (msec) : 2=4.16%, 4=1.84%, 10=4.90%, 20=11.33%, 50=5.37%
3323 lat (msec) : 100=0.65%
36214730
SW
3324 cpu : usr=0.27%, sys=0.18%, ctx=12072, majf=0, minf=21
3325 IO depths : 1=85.0%, 2=13.1%, 4=1.8%, 8=0.1%, 16=0.0%, 32=0.0%, >=64=0.0%
3326 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3327 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3328 issued rwt: total=0,4450,0, short=0,0,0, dropped=0,0,0
3329 latency : target=0, window=0, percentile=100.00%, depth=8
3330
3331The job name (or first job's name when using :option:`group_reporting`) is printed,
3332along with the group id, count of jobs being aggregated, last error id seen (which
3333is 0 when there are no errors), pid/tid of that thread and the time the job/group
3334completed. Below are the I/O statistics for each data direction performed (showing
3335writes in the example above). In the order listed, they denote:
3336
3337**read/write/trim**
3338 The string before the colon shows the I/O direction the statistics
3339 are for. **IOPS** is the average I/Os performed per second. **BW**
3340 is the average bandwidth rate shown as: value in power of 2 format
3341 (value in power of 10 format). The last two values show: (**total
3342 I/O performed** in power of 2 format / **runtime** of that thread).
f80dba8d
MT
3343
3344**slat**
36214730
SW
3345 Submission latency (**min** being the minimum, **max** being the
3346 maximum, **avg** being the average, **stdev** being the standard
3347 deviation). This is the time it took to submit the I/O. For
3348 sync I/O this row is not displayed as the slat is really the
3349 completion latency (since queue/complete is one operation there).
3350 This value can be in nanoseconds, microseconds or milliseconds ---
3351 fio will choose the most appropriate base and print that (in the
3352 example above nanoseconds was the best scale). Note: in :option:`--minimal` mode
0d237712 3353 latencies are always expressed in microseconds.
f80dba8d
MT
3354
3355**clat**
3356 Completion latency. Same names as slat, this denotes the time from
3357 submission to completion of the I/O pieces. For sync I/O, clat will
3358 usually be equal (or very close) to 0, as the time from submit to
3359 complete is basically just CPU time (I/O has already been done, see slat
3360 explanation).
3361
29092211
VF
3362**lat**
3363 Total latency. Same names as slat and clat, this denotes the time from
3364 when fio created the I/O unit to completion of the I/O operation.
3365
f80dba8d 3366**bw**
36214730
SW
3367 Bandwidth statistics based on samples. Same names as the xlat stats,
3368 but also includes the number of samples taken (**samples**) and an
3369 approximate percentage of total aggregate bandwidth this thread
3370 received in its group (**per**). This last value is only really
3371 useful if the threads in this group are on the same disk, since they
3372 are then competing for disk access.
3373
3374**iops**
3375 IOPS statistics based on samples. Same names as bw.
f80dba8d 3376
29092211
VF
3377**lat (nsec/usec/msec)**
3378 The distribution of I/O completion latencies. This is the time from when
3379 I/O leaves fio and when it gets completed. Unlike the separate
3380 read/write/trim sections above, the data here and in the remaining
3381 sections apply to all I/Os for the reporting group. 250=0.04% means that
3382 0.04% of the I/Os completed in under 250us. 500=64.11% means that 64.11%
3383 of the I/Os required 250 to 499us for completion.
3384
f80dba8d
MT
3385**cpu**
3386 CPU usage. User and system time, along with the number of context
3387 switches this thread went through, usage of system and user time, and
3388 finally the number of major and minor page faults. The CPU utilization
3389 numbers are averages for the jobs in that reporting group, while the
23a8e176 3390 context and fault counters are summed.
f80dba8d
MT
3391
3392**IO depths**
a2140525
SW
3393 The distribution of I/O depths over the job lifetime. The numbers are
3394 divided into powers of 2 and each entry covers depths from that value
3395 up to those that are lower than the next entry -- e.g., 16= covers
3396 depths from 16 to 31. Note that the range covered by a depth
3397 distribution entry can be different to the range covered by the
3398 equivalent submit/complete distribution entry.
f80dba8d
MT
3399
3400**IO submit**
3401 How many pieces of I/O were submitting in a single submit call. Each
c60ebc45 3402 entry denotes that amount and below, until the previous entry -- e.g.,
a2140525
SW
3403 16=100% means that we submitted anywhere between 9 to 16 I/Os per submit
3404 call. Note that the range covered by a submit distribution entry can
3405 be different to the range covered by the equivalent depth distribution
3406 entry.
f80dba8d
MT
3407
3408**IO complete**
3409 Like the above submit number, but for completions instead.
3410
36214730
SW
3411**IO issued rwt**
3412 The number of read/write/trim requests issued, and how many of them were
3413 short or dropped.
f80dba8d 3414
29092211 3415**IO latency**
ee21ebee 3416 These values are for :option:`latency_target` and related options. When
29092211
VF
3417 these options are engaged, this section describes the I/O depth required
3418 to meet the specified latency target.
71bfa161 3419
36214730
SW
3420..
3421 Example output was based on the following:
3422 TZ=UTC fio --ioengine=null --iodepth=2 --size=100M --numjobs=2 \
3423 --rate_process=poisson --io_limit=32M --name=read --bs=128k \
3424 --rate=11M --name=write --rw=write --bs=2k --rate=700k
3425
71bfa161 3426After each client has been listed, the group statistics are printed. They
f80dba8d 3427will look like this::
71bfa161 3428
f80dba8d 3429 Run status group 0 (all jobs):
36214730
SW
3430 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
3431 WRITE: bw=1231KiB/s (1261kB/s), 616KiB/s-621KiB/s (630kB/s-636kB/s), io=64.0MiB (67.1MB), run=52747-53223msec
71bfa161 3432
36214730 3433For each data direction it prints:
71bfa161 3434
36214730
SW
3435**bw**
3436 Aggregate bandwidth of threads in this group followed by the
3437 minimum and maximum bandwidth of all the threads in this group.
3438 Values outside of brackets are power-of-2 format and those
3439 within are the equivalent value in a power-of-10 format.
f80dba8d 3440**io**
36214730
SW
3441 Aggregate I/O performed of all threads in this group. The
3442 format is the same as bw.
3443**run**
3444 The smallest and longest runtimes of the threads in this group.
71bfa161 3445
f50fbdda 3446And finally, the disk statistics are printed. This is Linux specific. They will look like this::
71bfa161 3447
f80dba8d
MT
3448 Disk stats (read/write):
3449 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
71bfa161
JA
3450
3451Each value is printed for both reads and writes, with reads first. The
3452numbers denote:
3453
f80dba8d 3454**ios**
c60ebc45 3455 Number of I/Os performed by all groups.
f80dba8d 3456**merge**
007c7be9 3457 Number of merges performed by the I/O scheduler.
f80dba8d
MT
3458**ticks**
3459 Number of ticks we kept the disk busy.
36214730 3460**in_queue**
f80dba8d
MT
3461 Total time spent in the disk queue.
3462**util**
3463 The disk utilization. A value of 100% means we kept the disk
71bfa161
JA
3464 busy constantly, 50% would be a disk idling half of the time.
3465
f80dba8d
MT
3466It is also possible to get fio to dump the current output while it is running,
3467without terminating the job. To do that, send fio the **USR1** signal. You can
3468also get regularly timed dumps by using the :option:`--status-interval`
3469parameter, or by creating a file in :file:`/tmp` named
3470:file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
3471current output status.
8423bd11 3472
71bfa161 3473
f80dba8d
MT
3474Terse output
3475------------
71bfa161 3476
f80dba8d
MT
3477For scripted usage where you typically want to generate tables or graphs of the
3478results, fio can output the results in a semicolon separated format. The format
3479is one long line of values, such as::
71bfa161 3480
f80dba8d
MT
3481 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%
3482 A description of this job goes here.
562c2d2f
DN
3483
3484The job description (if provided) follows on a second line.
71bfa161 3485
a7f77fa6
SW
3486To enable terse output, use the :option:`--minimal` or
3487:option:`--output-format`\=terse command line options. The
f80dba8d
MT
3488first value is the version of the terse output format. If the output has to be
3489changed for some reason, this number will be incremented by 1 to signify that
3490change.
6820cb3b 3491
a2c95580 3492Split up, the format is as follows (comments in brackets denote when a
007c7be9 3493field was introduced or whether it's specific to some terse version):
71bfa161 3494
f80dba8d
MT
3495 ::
3496
f50fbdda 3497 terse version, fio version [v3], jobname, groupid, error
f80dba8d
MT
3498
3499 READ status::
3500
3501 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3502 Submission latency: min, max, mean, stdev (usec)
3503 Completion latency: min, max, mean, stdev (usec)
3504 Completion latency percentiles: 20 fields (see below)
3505 Total latency: min, max, mean, stdev (usec)
f50fbdda
TK
3506 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3507 IOPS [v5]: min, max, mean, stdev, number of samples
f80dba8d
MT
3508
3509 WRITE status:
3510
3511 ::
3512
3513 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3514 Submission latency: min, max, mean, stdev (usec)
247823cc 3515 Completion latency: min, max, mean, stdev (usec)
f80dba8d
MT
3516 Completion latency percentiles: 20 fields (see below)
3517 Total latency: min, max, mean, stdev (usec)
f50fbdda
TK
3518 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3519 IOPS [v5]: min, max, mean, stdev, number of samples
a2c95580
AH
3520
3521 TRIM status [all but version 3]:
3522
f50fbdda 3523 Fields are similar to READ/WRITE status.
f80dba8d
MT
3524
3525 CPU usage::
3526
3527 user, system, context switches, major faults, minor faults
3528
3529 I/O depths::
3530
3531 <=1, 2, 4, 8, 16, 32, >=64
3532
3533 I/O latencies microseconds::
3534
3535 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
3536
3537 I/O latencies milliseconds::
3538
3539 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
3540
a2c95580 3541 Disk utilization [v3]::
f80dba8d 3542
f50fbdda
TK
3543 disk name, read ios, write ios, read merges, write merges, read ticks, write ticks,
3544 time spent in queue, disk utilization percentage
f80dba8d
MT
3545
3546 Additional Info (dependent on continue_on_error, default off)::
3547
3548 total # errors, first error code
3549
3550 Additional Info (dependent on description being set)::
3551
3552 Text description
3553
3554Completion latency percentiles can be a grouping of up to 20 sets, so for the
3555terse output fio writes all of them. Each field will look like this::
1db92cb6 3556
f50fbdda 3557 1.00%=6112
1db92cb6 3558
f80dba8d 3559which is the Xth percentile, and the `usec` latency associated with it.
1db92cb6 3560
f50fbdda 3561For `Disk utilization`, all disks used by fio are shown. So for each disk there
f80dba8d 3562will be a disk utilization section.
f2f788dd 3563
2fc26c3d 3564Below is a single line containing short names for each of the fields in the
2831be97 3565minimal output v3, separated by semicolons::
2fc26c3d 3566
f50fbdda 3567 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 3568
25c8b9d7 3569
44c82dba
VF
3570JSON output
3571------------
3572
3573The `json` output format is intended to be both human readable and convenient
3574for automated parsing. For the most part its sections mirror those of the
3575`normal` output. The `runtime` value is reported in msec and the `bw` value is
3576reported in 1024 bytes per second units.
3577
3578
d29c4a91
VF
3579JSON+ output
3580------------
3581
3582The `json+` output format is identical to the `json` output format except that it
3583adds a full dump of the completion latency bins. Each `bins` object contains a
3584set of (key, value) pairs where keys are latency durations and values count how
3585many I/Os had completion latencies of the corresponding duration. For example,
3586consider:
3587
3588 "bins" : { "87552" : 1, "89600" : 1, "94720" : 1, "96768" : 1, "97792" : 1, "99840" : 1, "100864" : 2, "103936" : 6, "104960" : 534, "105984" : 5995, "107008" : 7529, ... }
3589
3590This data indicates that one I/O required 87,552ns to complete, two I/Os required
3591100,864ns to complete, and 7529 I/Os required 107,008ns to complete.
3592
3593Also included with fio is a Python script `fio_jsonplus_clat2csv` that takes
3594json+ output and generates CSV-formatted latency data suitable for plotting.
3595
3596The latency durations actually represent the midpoints of latency intervals.
f50fbdda 3597For details refer to :file:`stat.h`.
d29c4a91
VF
3598
3599
f80dba8d
MT
3600Trace file format
3601-----------------
3602
3603There are two trace file format that you can encounter. The older (v1) format is
3604unsupported since version 1.20-rc3 (March 2008). It will still be described
25c8b9d7
PD
3605below in case that you get an old trace and want to understand it.
3606
3607In any case the trace is a simple text file with a single action per line.
3608
3609
f80dba8d
MT
3610Trace file format v1
3611~~~~~~~~~~~~~~~~~~~~
3612
3613Each line represents a single I/O action in the following format::
3614
3615 rw, offset, length
25c8b9d7 3616
f50fbdda 3617where `rw=0/1` for read/write, and the `offset` and `length` entries being in bytes.
25c8b9d7 3618
22413915 3619This format is not supported in fio versions >= 1.20-rc3.
25c8b9d7 3620
25c8b9d7 3621
f80dba8d
MT
3622Trace file format v2
3623~~~~~~~~~~~~~~~~~~~~
25c8b9d7 3624
f80dba8d
MT
3625The second version of the trace file format was added in fio version 1.17. It
3626allows to access more then one file per trace and has a bigger set of possible
3627file actions.
25c8b9d7 3628
f80dba8d 3629The first line of the trace file has to be::
25c8b9d7 3630
f80dba8d 3631 fio version 2 iolog
25c8b9d7
PD
3632
3633Following this can be lines in two different formats, which are described below.
3634
f80dba8d 3635The file management format::
25c8b9d7 3636
f80dba8d 3637 filename action
25c8b9d7 3638
f50fbdda 3639The `filename` is given as an absolute path. The `action` can be one of these:
25c8b9d7 3640
f80dba8d 3641**add**
f50fbdda 3642 Add the given `filename` to the trace.
f80dba8d 3643**open**
f50fbdda 3644 Open the file with the given `filename`. The `filename` has to have
f80dba8d
MT
3645 been added with the **add** action before.
3646**close**
f50fbdda 3647 Close the file with the given `filename`. The file has to have been
f80dba8d
MT
3648 opened before.
3649
3650
3651The file I/O action format::
3652
3653 filename action offset length
3654
3655The `filename` is given as an absolute path, and has to have been added and
3656opened before it can be used with this format. The `offset` and `length` are
3657given in bytes. The `action` can be one of these:
3658
3659**wait**
3660 Wait for `offset` microseconds. Everything below 100 is discarded.
3661 The time is relative to the previous `wait` statement.
3662**read**
3663 Read `length` bytes beginning from `offset`.
3664**write**
3665 Write `length` bytes beginning from `offset`.
3666**sync**
3667 :manpage:`fsync(2)` the file.
3668**datasync**
3669 :manpage:`fdatasync(2)` the file.
3670**trim**
3671 Trim the given file from the given `offset` for `length` bytes.
3672
3673CPU idleness profiling
3674----------------------
3675
3676In some cases, we want to understand CPU overhead in a test. For example, we
3677test patches for the specific goodness of whether they reduce CPU usage.
3678Fio implements a balloon approach to create a thread per CPU that runs at idle
3679priority, meaning that it only runs when nobody else needs the cpu.
3680By measuring the amount of work completed by the thread, idleness of each CPU
3681can be derived accordingly.
3682
3683An unit work is defined as touching a full page of unsigned characters. Mean and
3684standard deviation of time to complete an unit work is reported in "unit work"
3685section. Options can be chosen to report detailed percpu idleness or overall
3686system idleness by aggregating percpu stats.
3687
3688
3689Verification and triggers
3690-------------------------
3691
3692Fio is usually run in one of two ways, when data verification is done. The first
3693is a normal write job of some sort with verify enabled. When the write phase has
3694completed, fio switches to reads and verifies everything it wrote. The second
3695model is running just the write phase, and then later on running the same job
3696(but with reads instead of writes) to repeat the same I/O patterns and verify
3697the contents. Both of these methods depend on the write phase being completed,
3698as fio otherwise has no idea how much data was written.
3699
3700With verification triggers, fio supports dumping the current write state to
3701local files. Then a subsequent read verify workload can load this state and know
3702exactly where to stop. This is useful for testing cases where power is cut to a
3703server in a managed fashion, for instance.
99b9a85a
JA
3704
3705A verification trigger consists of two things:
3706
f80dba8d
MT
37071) Storing the write state of each job.
37082) Executing a trigger command.
99b9a85a 3709
f80dba8d
MT
3710The write state is relatively small, on the order of hundreds of bytes to single
3711kilobytes. It contains information on the number of completions done, the last X
3712completions, etc.
99b9a85a 3713
f80dba8d
MT
3714A trigger is invoked either through creation ('touch') of a specified file in
3715the system, or through a timeout setting. If fio is run with
9207a0cb 3716:option:`--trigger-file`\= :file:`/tmp/trigger-file`, then it will continually
f80dba8d
MT
3717check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
3718will fire off the trigger (thus saving state, and executing the trigger
99b9a85a
JA
3719command).
3720
f80dba8d
MT
3721For client/server runs, there's both a local and remote trigger. If fio is
3722running as a server backend, it will send the job states back to the client for
3723safe storage, then execute the remote trigger, if specified. If a local trigger
3724is specified, the server will still send back the write state, but the client
3725will then execute the trigger.
99b9a85a 3726
f80dba8d
MT
3727Verification trigger example
3728~~~~~~~~~~~~~~~~~~~~~~~~~~~~
99b9a85a 3729
f50fbdda
TK
3730Let's say we want to run a powercut test on the remote Linux machine 'server'.
3731Our write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
f80dba8d
MT
3732some point during the run, and we'll run this test from the safety or our local
3733machine, 'localbox'. On the server, we'll start the fio backend normally::
99b9a85a 3734
f80dba8d 3735 server# fio --server
99b9a85a 3736
f80dba8d 3737and on the client, we'll fire off the workload::
99b9a85a 3738
f80dba8d 3739 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
99b9a85a 3740
f80dba8d 3741We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
99b9a85a 3742
f80dba8d 3743 echo b > /proc/sysrq-trigger
99b9a85a 3744
f80dba8d
MT
3745on the server once it has received the trigger and sent us the write state. This
3746will work, but it's not **really** cutting power to the server, it's merely
3747abruptly rebooting it. If we have a remote way of cutting power to the server
3748through IPMI or similar, we could do that through a local trigger command
4502cb42 3749instead. Let's assume we have a script that does IPMI reboot of a given hostname,
f80dba8d
MT
3750ipmi-reboot. On localbox, we could then have run fio with a local trigger
3751instead::
99b9a85a 3752
f80dba8d 3753 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
99b9a85a 3754
f80dba8d
MT
3755For this case, fio would wait for the server to send us the write state, then
3756execute ``ipmi-reboot server`` when that happened.
3757
3758Loading verify state
3759~~~~~~~~~~~~~~~~~~~~
3760
4502cb42 3761To load stored write state, a read verification job file must contain the
f80dba8d 3762:option:`verify_state_load` option. If that is set, fio will load the previously
99b9a85a 3763stored state. For a local fio run this is done by loading the files directly,
f80dba8d
MT
3764and on a client/server run, the server backend will ask the client to send the
3765files over and load them from there.
a3ae5b05
JA
3766
3767
f80dba8d
MT
3768Log File Formats
3769----------------
a3ae5b05
JA
3770
3771Fio supports a variety of log file formats, for logging latencies, bandwidth,
3772and IOPS. The logs share a common format, which looks like this:
3773
5a83478f
SW
3774 *time* (`msec`), *value*, *data direction*, *block size* (`bytes`),
3775 *offset* (`bytes`)
a3ae5b05 3776
5a83478f 3777*Time* for the log entry is always in milliseconds. The *value* logged depends
a3ae5b05
JA
3778on the type of log, it will be one of the following: