4 The first step in getting fio to simulate a desired I/O workload, is writing a
5 job file describing that specific setup. A job file may contain any number of
6 threads and/or files -- the typical contents of the job file is a *global*
7 section defining shared parameters, and one or more job sections describing the
8 jobs involved. When run, fio parses this file and sets everything up as
9 described. If we break down a job from top to bottom, it contains the following
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?
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.
26 How much data are we going to be reading/writing.
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
36 If the I/O engine is async, how large a queuing depth do we want to
42 How many files are we spreading the workload over.
44 `Threads, processes and job synchronization`_
46 How many threads or processes should we spread this workload over.
48 The above are the basic parameters defined for a workload, in addition there's a
49 multitude of parameters that modify other aspects of how this job behaves.
55 .. option:: --debug=type
57 Enable verbose tracing `type` of various fio actions. May be ``all`` for all types
58 or individual types separated by a comma (e.g. ``--debug=file,mem`` will
59 enable file and memory debugging). Currently, additional logging is
63 Dump info related to processes.
65 Dump info related to file actions.
67 Dump info related to I/O queuing.
69 Dump info related to memory allocations.
71 Dump info related to blktrace setup.
73 Dump info related to I/O verification.
75 Enable all debug options.
77 Dump info related to random offset generation.
79 Dump info related to option matching and parsing.
81 Dump info related to disk utilization updates.
83 Dump info only related to job number x.
85 Dump info only related to mutex up/down ops.
87 Dump info related to profile extensions.
89 Dump info related to internal time keeping.
91 Dump info related to networking connections.
93 Dump info related to I/O rate switching.
95 Dump info related to log compress/decompress.
97 Dump info related to steadystate detection.
99 Dump info related to the helper thread.
101 Dump info related to support for zoned block devices.
103 Show available debug options.
105 .. option:: --parse-only
107 Parse options only, don't start any I/O.
109 .. option:: --merge-blktrace-only
111 Merge blktraces only, don't start any I/O.
113 .. option:: --output=filename
115 Write output to file `filename`.
117 .. option:: --output-format=format
119 Set the reporting `format` to `normal`, `terse`, `json`, or `json+`. Multiple
120 formats can be selected, separated by a comma. `terse` is a CSV based
121 format. `json+` is like `json`, except it adds a full dump of the latency
124 .. option:: --bandwidth-log
126 Generate aggregate bandwidth logs.
128 .. option:: --minimal
130 Print statistics in a terse, semicolon-delimited format.
132 .. option:: --append-terse
134 Print statistics in selected mode AND terse, semicolon-delimited format.
135 **Deprecated**, use :option:`--output-format` instead to select multiple
138 .. option:: --terse-version=version
140 Set terse `version` output format (default 3, or 2 or 4 or 5).
142 .. option:: --version
144 Print version information and exit.
148 Print a summary of the command line options and exit.
150 .. option:: --cpuclock-test
152 Perform test and validation of internal CPU clock.
154 .. option:: --crctest=[test]
156 Test the speed of the built-in checksumming functions. If no argument is
157 given, all of them are tested. Alternatively, a comma separated list can
158 be passed, in which case the given ones are tested.
160 .. option:: --cmdhelp=command
162 Print help information for `command`. May be ``all`` for all commands.
164 .. option:: --enghelp=[ioengine[,command]]
166 List all commands defined by `ioengine`, or print help for `command`
167 defined by `ioengine`. If no `ioengine` is given, list all
170 .. option:: --showcmd
172 Convert given job files to a set of command-line options.
174 .. option:: --readonly
176 Turn on safety read-only checks, preventing writes and trims. The
177 ``--readonly`` option is an extra safety guard to prevent users from
178 accidentally starting a write or trim workload when that is not desired.
179 Fio will only modify the device under test if
180 `rw=write/randwrite/rw/randrw/trim/randtrim/trimwrite` is given. This
181 safety net can be used as an extra precaution.
183 .. option:: --eta=when
185 Specifies when real-time ETA estimate should be printed. `when` may be
186 `always`, `never` or `auto`. `auto` is the default, it prints ETA
187 when requested if the output is a TTY. `always` disregards the output
188 type, and prints ETA when requested. `never` never prints ETA.
190 .. option:: --eta-interval=time
192 By default, fio requests client ETA status roughly every second. With
193 this option, the interval is configurable. Fio imposes a minimum
194 allowed time to avoid flooding the console, less than 250 msec is
197 .. option:: --eta-newline=time
199 Force a new line for every `time` period passed. When the unit is omitted,
200 the value is interpreted in seconds.
202 .. option:: --status-interval=time
204 Force a full status dump of cumulative (from job start) values at `time`
205 intervals. This option does *not* provide per-period measurements. So
206 values such as bandwidth are running averages. When the time unit is omitted,
207 `time` is interpreted in seconds. Note that using this option with
208 ``--output-format=json`` will yield output that technically isn't valid
209 json, since the output will be collated sets of valid json. It will need
210 to be split into valid sets of json after the run.
212 .. option:: --section=name
214 Only run specified section `name` in job file. Multiple sections can be specified.
215 The ``--section`` option allows one to combine related jobs into one file.
216 E.g. one job file could define light, moderate, and heavy sections. Tell
217 fio to run only the "heavy" section by giving ``--section=heavy``
218 command line option. One can also specify the "write" operations in one
219 section and "verify" operation in another section. The ``--section`` option
220 only applies to job sections. The reserved *global* section is always
223 .. option:: --alloc-size=kb
225 Allocate additional internal smalloc pools of size `kb` in KiB. The
226 ``--alloc-size`` option increases shared memory set aside for use by fio.
227 If running large jobs with randommap enabled, fio can run out of memory.
228 Smalloc is an internal allocator for shared structures from a fixed size
229 memory pool and can grow to 16 pools. The pool size defaults to 16MiB.
231 NOTE: While running :file:`.fio_smalloc.*` backing store files are visible
234 .. option:: --warnings-fatal
236 All fio parser warnings are fatal, causing fio to exit with an
239 .. option:: --max-jobs=nr
241 Set the maximum number of threads/processes to support to `nr`.
242 NOTE: On Linux, it may be necessary to increase the shared-memory
243 limit (:file:`/proc/sys/kernel/shmmax`) if fio runs into errors while
246 .. option:: --server=args
248 Start a backend server, with `args` specifying what to listen to.
249 See `Client/Server`_ section.
251 .. option:: --daemonize=pidfile
253 Background a fio server, writing the pid to the given `pidfile` file.
255 .. option:: --client=hostname
257 Instead of running the jobs locally, send and run them on the given `hostname`
258 or set of `hostname`\s. See `Client/Server`_ section.
260 .. option:: --remote-config=file
262 Tell fio server to load this local `file`.
264 .. option:: --idle-prof=option
266 Report CPU idleness. `option` is one of the following:
269 Run unit work calibration only and exit.
272 Show aggregate system idleness and unit work.
275 As **system** but also show per CPU idleness.
277 .. option:: --inflate-log=log
279 Inflate and output compressed `log`.
281 .. option:: --trigger-file=file
283 Execute trigger command when `file` exists.
285 .. option:: --trigger-timeout=time
287 Execute trigger at this `time`.
289 .. option:: --trigger=command
291 Set this `command` as local trigger.
293 .. option:: --trigger-remote=command
295 Set this `command` as remote trigger.
297 .. option:: --aux-path=path
299 Use the directory specified by `path` for generated state files instead
300 of the current working directory.
302 Any parameters following the options will be assumed to be job files, unless
303 they match a job file parameter. Multiple job files can be listed and each job
304 file will be regarded as a separate group. Fio will :option:`stonewall`
305 execution between each group.
311 As previously described, fio accepts one or more job files describing what it is
312 supposed to do. The job file format is the classic ini file, where the names
313 enclosed in [] brackets define the job name. You are free to use any ASCII name
314 you want, except *global* which has special meaning. Following the job name is
315 a sequence of zero or more parameters, one per line, that define the behavior of
316 the job. If the first character in a line is a ';' or a '#', the entire line is
317 discarded as a comment.
319 A *global* section sets defaults for the jobs described in that file. A job may
320 override a *global* section parameter, and a job file may even have several
321 *global* sections if so desired. A job is only affected by a *global* section
324 The :option:`--cmdhelp` option also lists all options. If used with a `command`
325 argument, :option:`--cmdhelp` will detail the given `command`.
327 See the `examples/` directory for inspiration on how to write job files. Note
328 the copyright and license requirements currently apply to `examples/` files.
330 So let's look at a really simple job file that defines two processes, each
331 randomly reading from a 128MiB file:
335 ; -- start job file --
346 As you can see, the job file sections themselves are empty as all the described
347 parameters are shared. As no :option:`filename` option is given, fio makes up a
348 `filename` for each of the jobs as it sees fit. On the command line, this job
349 would look as follows::
351 $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
354 Let's look at an example that has a number of processes writing randomly to
359 ; -- start job file --
370 Here we have no *global* section, as we only have one job defined anyway. We
371 want to use async I/O here, with a depth of 4 for each file. We also increased
372 the buffer size used to 32KiB and define numjobs to 4 to fork 4 identical
373 jobs. The result is 4 processes each randomly writing to their own 64MiB
374 file. Instead of using the above job file, you could have given the parameters
375 on the command line. For this case, you would specify::
377 $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
379 When fio is utilized as a basis of any reasonably large test suite, it might be
380 desirable to share a set of standardized settings across multiple job files.
381 Instead of copy/pasting such settings, any section may pull in an external
382 :file:`filename.fio` file with *include filename* directive, as in the following
385 ; -- start job file including.fio --
389 include glob-include.fio
396 include test-include.fio
397 ; -- end job file including.fio --
401 ; -- start job file glob-include.fio --
404 ; -- end job file glob-include.fio --
408 ; -- start job file test-include.fio --
411 ; -- end job file test-include.fio --
413 Settings pulled into a section apply to that section only (except *global*
414 section). Include directives may be nested in that any included file may contain
415 further include directive(s). Include files may not contain [] sections.
418 Environment variables
419 ~~~~~~~~~~~~~~~~~~~~~
421 Fio also supports environment variable expansion in job files. Any sub-string of
422 the form ``${VARNAME}`` as part of an option value (in other words, on the right
423 of the '='), will be expanded to the value of the environment variable called
424 `VARNAME`. If no such environment variable is defined, or `VARNAME` is the
425 empty string, the empty string will be substituted.
427 As an example, let's look at a sample fio invocation and job file::
429 $ SIZE=64m NUMJOBS=4 fio jobfile.fio
433 ; -- start job file --
440 This will expand to the following equivalent job file at runtime:
444 ; -- start job file --
451 Fio ships with a few example job files, you can also look there for inspiration.
456 Additionally, fio has a set of reserved keywords that will be replaced
457 internally with the appropriate value. Those keywords are:
461 The architecture page size of the running system.
465 Megabytes of total memory in the system.
469 Number of online available CPUs.
471 These can be used on the command line or in the job file, and will be
472 automatically substituted with the current system values when the job is
473 run. Simple math is also supported on these keywords, so you can perform actions
478 and get that properly expanded to 8 times the size of memory in the machine.
484 This section describes in details each parameter associated with a job. Some
485 parameters take an option of a given type, such as an integer or a
486 string. Anywhere a numeric value is required, an arithmetic expression may be
487 used, provided it is surrounded by parentheses. Supported operators are:
496 For time values in expressions, units are microseconds by default. This is
497 different than for time values not in expressions (not enclosed in
498 parentheses). The following types are used:
505 String: A sequence of alphanumeric characters.
508 Integer with possible time suffix. Without a unit value is interpreted as
509 seconds unless otherwise specified. Accepts a suffix of 'd' for days, 'h' for
510 hours, 'm' for minutes, 's' for seconds, 'ms' (or 'msec') for milliseconds and
511 'us' (or 'usec') for microseconds. For example, use 10m for 10 minutes.
516 Integer. A whole number value, which may contain an integer prefix
517 and an integer suffix:
519 [*integer prefix*] **number** [*integer suffix*]
521 The optional *integer prefix* specifies the number's base. The default
522 is decimal. *0x* specifies hexadecimal.
524 The optional *integer suffix* specifies the number's units, and includes an
525 optional unit prefix and an optional unit. For quantities of data, the
526 default unit is bytes. For quantities of time, the default unit is seconds
527 unless otherwise specified.
529 With :option:`kb_base`\=1000, fio follows international standards for unit
530 prefixes. To specify power-of-10 decimal values defined in the
531 International System of Units (SI):
533 * *K* -- means kilo (K) or 1000
534 * *M* -- means mega (M) or 1000**2
535 * *G* -- means giga (G) or 1000**3
536 * *T* -- means tera (T) or 1000**4
537 * *P* -- means peta (P) or 1000**5
539 To specify power-of-2 binary values defined in IEC 80000-13:
541 * *Ki* -- means kibi (Ki) or 1024
542 * *Mi* -- means mebi (Mi) or 1024**2
543 * *Gi* -- means gibi (Gi) or 1024**3
544 * *Ti* -- means tebi (Ti) or 1024**4
545 * *Pi* -- means pebi (Pi) or 1024**5
547 For Zone Block Device Mode:
550 With :option:`kb_base`\=1024 (the default), the unit prefixes are opposite
551 from those specified in the SI and IEC 80000-13 standards to provide
552 compatibility with old scripts. For example, 4k means 4096.
554 For quantities of data, an optional unit of 'B' may be included
555 (e.g., 'kB' is the same as 'k').
557 The *integer suffix* is not case sensitive (e.g., m/mi mean mebi/mega,
558 not milli). 'b' and 'B' both mean byte, not bit.
560 Examples with :option:`kb_base`\=1000:
562 * *4 KiB*: 4096, 4096b, 4096B, 4ki, 4kib, 4kiB, 4Ki, 4KiB
563 * *1 MiB*: 1048576, 1mi, 1024ki
564 * *1 MB*: 1000000, 1m, 1000k
565 * *1 TiB*: 1099511627776, 1ti, 1024gi, 1048576mi
566 * *1 TB*: 1000000000, 1t, 1000m, 1000000k
568 Examples with :option:`kb_base`\=1024 (default):
570 * *4 KiB*: 4096, 4096b, 4096B, 4k, 4kb, 4kB, 4K, 4KB
571 * *1 MiB*: 1048576, 1m, 1024k
572 * *1 MB*: 1000000, 1mi, 1000ki
573 * *1 TiB*: 1099511627776, 1t, 1024g, 1048576m
574 * *1 TB*: 1000000000, 1ti, 1000mi, 1000000ki
576 To specify times (units are not case sensitive):
580 * *M* -- means minutes
581 * *s* -- or sec means seconds (default)
582 * *ms* -- or *msec* means milliseconds
583 * *us* -- or *usec* means microseconds
585 If the option accepts an upper and lower range, use a colon ':' or
586 minus '-' to separate such values. See :ref:`irange <irange>`.
587 If the lower value specified happens to be larger than the upper value
588 the two values are swapped.
593 Boolean. Usually parsed as an integer, however only defined for
594 true and false (1 and 0).
599 Integer range with suffix. Allows value range to be given, such as
600 1024-4096. A colon may also be used as the separator, e.g. 1k:4k. If the
601 option allows two sets of ranges, they can be specified with a ',' or '/'
602 delimiter: 1k-4k/8k-32k. Also see :ref:`int <int>`.
605 A list of floating point numbers, separated by a ':' character.
607 With the above in mind, here follows the complete list of fio job parameters.
613 .. option:: kb_base=int
615 Select the interpretation of unit prefixes in input parameters.
618 Inputs comply with IEC 80000-13 and the International
619 System of Units (SI). Use:
621 - power-of-2 values with IEC prefixes (e.g., KiB)
622 - power-of-10 values with SI prefixes (e.g., kB)
625 Compatibility mode (default). To avoid breaking old scripts:
627 - power-of-2 values with SI prefixes
628 - power-of-10 values with IEC prefixes
630 See :option:`bs` for more details on input parameters.
632 Outputs always use correct prefixes. Most outputs include both
635 bw=2383.3kB/s (2327.4KiB/s)
637 If only one value is reported, then kb_base selects the one to use:
639 **1000** -- SI prefixes
641 **1024** -- IEC prefixes
643 .. option:: unit_base=int
645 Base unit for reporting. Allowed values are:
648 Use auto-detection (default).
660 ASCII name of the job. This may be used to override the name printed by fio
661 for this job. Otherwise the job name is used. On the command line this
662 parameter has the special purpose of also signaling the start of a new job.
664 .. option:: description=str
666 Text description of the job. Doesn't do anything except dump this text
667 description when this job is run. It's not parsed.
669 .. option:: loops=int
671 Run the specified number of iterations of this job. Used to repeat the same
672 workload a given number of times. Defaults to 1.
674 .. option:: numjobs=int
676 Create the specified number of clones of this job. Each clone of job
677 is spawned as an independent thread or process. May be used to setup a
678 larger number of threads/processes doing the same thing. Each thread is
679 reported separately; to see statistics for all clones as a whole, use
680 :option:`group_reporting` in conjunction with :option:`new_group`.
681 See :option:`--max-jobs`. Default: 1.
684 Time related parameters
685 ~~~~~~~~~~~~~~~~~~~~~~~
687 .. option:: runtime=time
689 Tell fio to terminate processing after the specified period of time. It
690 can be quite hard to determine for how long a specified job will run, so
691 this parameter is handy to cap the total runtime to a given time. When
692 the unit is omitted, the value is interpreted in seconds.
694 .. option:: time_based
696 If set, fio will run for the duration of the :option:`runtime` specified
697 even if the file(s) are completely read or written. It will simply loop over
698 the same workload as many times as the :option:`runtime` allows.
700 .. option:: startdelay=irange(time)
702 Delay the start of job for the specified amount of time. Can be a single
703 value or a range. When given as a range, each thread will choose a value
704 randomly from within the range. Value is in seconds if a unit is omitted.
706 .. option:: ramp_time=time
708 If set, fio will run the specified workload for this amount of time before
709 logging any performance numbers. Useful for letting performance settle
710 before logging results, thus minimizing the runtime required for stable
711 results. Note that the ``ramp_time`` is considered lead in time for a job,
712 thus it will increase the total runtime if a special timeout or
713 :option:`runtime` is specified. When the unit is omitted, the value is
716 .. option:: clocksource=str
718 Use the given clocksource as the base of timing. The supported options are:
721 :manpage:`gettimeofday(2)`
724 :manpage:`clock_gettime(2)`
727 Internal CPU clock source
729 cpu is the preferred clocksource if it is reliable, as it is very fast (and
730 fio is heavy on time calls). Fio will automatically use this clocksource if
731 it's supported and considered reliable on the system it is running on,
732 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
733 means supporting TSC Invariant.
735 .. option:: gtod_reduce=bool
737 Enable all of the :manpage:`gettimeofday(2)` reducing options
738 (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus
739 reduce precision of the timeout somewhat to really shrink the
740 :manpage:`gettimeofday(2)` call count. With this option enabled, we only do
741 about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all
742 time keeping was enabled.
744 .. option:: gtod_cpu=int
746 Sometimes it's cheaper to dedicate a single thread of execution to just
747 getting the current time. Fio (and databases, for instance) are very
748 intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set
749 one CPU aside for doing nothing but logging current time to a shared memory
750 location. Then the other threads/processes that run I/O workloads need only
751 copy that segment, instead of entering the kernel with a
752 :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time
753 calls will be excluded from other uses. Fio will manually clear it from the
754 CPU mask of other jobs.
760 .. option:: directory=str
762 Prefix filenames with this directory. Used to place files in a different
763 location than :file:`./`. You can specify a number of directories by
764 separating the names with a ':' character. These directories will be
765 assigned equally distributed to job clones created by :option:`numjobs` as
766 long as they are using generated filenames. If specific `filename(s)` are
767 set fio will use the first listed directory, and thereby matching the
768 `filename` semantic (which generates a file for each clone if not
769 specified, but lets all clones use the same file if set).
771 See the :option:`filename` option for information on how to escape "``:``"
772 characters within the directory path itself.
774 Note: To control the directory fio will use for internal state files
775 use :option:`--aux-path`.
777 .. option:: filename=str
779 Fio normally makes up a `filename` based on the job name, thread number, and
780 file number (see :option:`filename_format`). If you want to share files
781 between threads in a job or several
782 jobs with fixed file paths, specify a `filename` for each of them to override
783 the default. If the ioengine is file based, you can specify a number of files
784 by separating the names with a ':' colon. So if you wanted a job to open
785 :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use
786 ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is
787 specified, :option:`nrfiles` is ignored. The size of regular files specified
788 by this option will be :option:`size` divided by number of files unless an
789 explicit size is specified by :option:`filesize`.
791 Each colon in the wanted path must be escaped with a ``\``
792 character. For instance, if the path is :file:`/dev/dsk/foo@3,0:c` then you
793 would use ``filename=/dev/dsk/foo@3,0\:c`` and if the path is
794 :file:`F:\\filename` then you would use ``filename=F\:\filename``.
796 On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for
797 the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc.
798 Note: Windows and FreeBSD prevent write access to areas
799 of the disk containing in-use data (e.g. filesystems).
801 The filename "`-`" is a reserved name, meaning *stdin* or *stdout*. Which
802 of the two depends on the read/write direction set.
804 .. option:: filename_format=str
806 If sharing multiple files between jobs, it is usually necessary to have fio
807 generate the exact names that you want. By default, fio will name a file
808 based on the default file format specification of
809 :file:`jobname.jobnumber.filenumber`. With this option, that can be
810 customized. Fio will recognize and replace the following keywords in this
814 The name of the worker thread or process.
816 IP of the fio process when using client/server mode.
818 The incremental number of the worker thread or process.
820 The incremental number of the file for that worker thread or
823 To have dependent jobs share a set of files, this option can be set to have
824 fio generate filenames that are shared between the two. For instance, if
825 :file:`testfiles.$filenum` is specified, file number 4 for any job will be
826 named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum`
827 will be used if no other format specifier is given.
829 If you specify a path then the directories will be created up to the
830 main directory for the file. So for example if you specify
831 ``filename_format=a/b/c/$jobnum`` then the directories a/b/c will be
832 created before the file setup part of the job. If you specify
833 :option:`directory` then the path will be relative that directory,
834 otherwise it is treated as the absolute path.
836 .. option:: unique_filename=bool
838 To avoid collisions between networked clients, fio defaults to prefixing any
839 generated filenames (with a directory specified) with the source of the
840 client connecting. To disable this behavior, set this option to 0.
842 .. option:: opendir=str
844 Recursively open any files below directory `str`.
846 .. option:: lockfile=str
848 Fio defaults to not locking any files before it does I/O to them. If a file
849 or file descriptor is shared, fio can serialize I/O to that file to make the
850 end result consistent. This is usual for emulating real workloads that share
851 files. The lock modes are:
854 No locking. The default.
856 Only one thread or process may do I/O at a time, excluding all
859 Read-write locking on the file. Many readers may
860 access the file at the same time, but writes get exclusive access.
862 .. option:: nrfiles=int
864 Number of files to use for this job. Defaults to 1. The size of files
865 will be :option:`size` divided by this unless explicit size is specified by
866 :option:`filesize`. Files are created for each thread separately, and each
867 file will have a file number within its name by default, as explained in
868 :option:`filename` section.
871 .. option:: openfiles=int
873 Number of files to keep open at the same time. Defaults to the same as
874 :option:`nrfiles`, can be set smaller to limit the number simultaneous
877 .. option:: file_service_type=str
879 Defines how fio decides which file from a job to service next. The following
883 Choose a file at random.
886 Round robin over opened files. This is the default.
889 Finish one file before moving on to the next. Multiple files can
890 still be open depending on :option:`openfiles`.
893 Use a *Zipf* distribution to decide what file to access.
896 Use a *Pareto* distribution to decide what file to access.
899 Use a *Gaussian* (normal) distribution to decide what file to
905 For *random*, *roundrobin*, and *sequential*, a postfix can be appended to
906 tell fio how many I/Os to issue before switching to a new file. For example,
907 specifying ``file_service_type=random:8`` would cause fio to issue
908 8 I/Os before selecting a new file at random. For the non-uniform
909 distributions, a floating point postfix can be given to influence how the
910 distribution is skewed. See :option:`random_distribution` for a description
911 of how that would work.
913 .. option:: ioscheduler=str
915 Attempt to switch the device hosting the file to the specified I/O scheduler
918 .. option:: create_serialize=bool
920 If true, serialize the file creation for the jobs. This may be handy to
921 avoid interleaving of data files, which may greatly depend on the filesystem
922 used and even the number of processors in the system. Default: true.
924 .. option:: create_fsync=bool
926 :manpage:`fsync(2)` the data file after creation. This is the default.
928 .. option:: create_on_open=bool
930 If true, don't pre-create files but allow the job's open() to create a file
931 when it's time to do I/O. Default: false -- pre-create all necessary files
934 .. option:: create_only=bool
936 If true, fio will only run the setup phase of the job. If files need to be
937 laid out or updated on disk, only that will be done -- the actual job contents
938 are not executed. Default: false.
940 .. option:: allow_file_create=bool
942 If true, fio is permitted to create files as part of its workload. If this
943 option is false, then fio will error out if
944 the files it needs to use don't already exist. Default: true.
946 .. option:: allow_mounted_write=bool
948 If this isn't set, fio will abort jobs that are destructive (e.g. that write)
949 to what appears to be a mounted device or partition. This should help catch
950 creating inadvertently destructive tests, not realizing that the test will
951 destroy data on the mounted file system. Note that some platforms don't allow
952 writing against a mounted device regardless of this option. Default: false.
954 .. option:: pre_read=bool
956 If this is given, files will be pre-read into memory before starting the
957 given I/O operation. This will also clear the :option:`invalidate` flag,
958 since it is pointless to pre-read and then drop the cache. This will only
959 work for I/O engines that are seek-able, since they allow you to read the
960 same data multiple times. Thus it will not work on non-seekable I/O engines
961 (e.g. network, splice). Default: false.
963 .. option:: unlink=bool
965 Unlink the job files when done. Not the default, as repeated runs of that
966 job would then waste time recreating the file set again and again. Default:
969 .. option:: unlink_each_loop=bool
971 Unlink job files after each iteration or loop. Default: false.
973 .. option:: zonemode=str
978 The :option:`zonerange`, :option:`zonesize`,
979 :option `zonecapacity` and option:`zoneskip`
980 parameters are ignored.
982 I/O happens in a single zone until
983 :option:`zonesize` bytes have been transferred.
984 After that number of bytes has been
985 transferred processing of the next zone
986 starts. :option `zonecapacity` is ignored.
988 Zoned block device mode. I/O happens
989 sequentially in each zone, even if random I/O
990 has been selected. Random I/O happens across
991 all zones instead of being restricted to a
992 single zone. The :option:`zoneskip` parameter
993 is ignored. :option:`zonerange` and
994 :option:`zonesize` must be identical.
995 Trim is handled using a zone reset operation.
996 Trim only considers non-empty sequential write
997 required and sequential write preferred zones.
999 .. option:: zonerange=int
1001 Size of a single zone. See also :option:`zonesize` and
1004 .. option:: zonesize=int
1006 For :option:`zonemode` =strided, this is the number of bytes to
1007 transfer before skipping :option:`zoneskip` bytes. If this parameter
1008 is smaller than :option:`zonerange` then only a fraction of each zone
1009 with :option:`zonerange` bytes will be accessed. If this parameter is
1010 larger than :option:`zonerange` then each zone will be accessed
1011 multiple times before skipping to the next zone.
1013 For :option:`zonemode` =zbd, this is the size of a single zone. The
1014 :option:`zonerange` parameter is ignored in this mode.
1017 .. option:: zonecapacity=int
1019 For :option:`zonemode` =zbd, this defines the capacity of a single zone,
1020 which is the accessible area starting from the zone start address.
1021 This parameter only applies when using :option:`zonemode` =zbd in
1022 combination with regular block devices. If not specified it defaults to
1023 the zone size. If the target device is a zoned block device, the zone
1024 capacity is obtained from the device information and this option is
1027 .. option:: zoneskip=int
1029 For :option:`zonemode` =strided, the number of bytes to skip after
1030 :option:`zonesize` bytes of data have been transferred. This parameter
1031 must be zero for :option:`zonemode` =zbd.
1033 .. option:: read_beyond_wp=bool
1035 This parameter applies to :option:`zonemode` =zbd only.
1037 Zoned block devices are block devices that consist of multiple zones.
1038 Each zone has a type, e.g. conventional or sequential. A conventional
1039 zone can be written at any offset that is a multiple of the block
1040 size. Sequential zones must be written sequentially. The position at
1041 which a write must occur is called the write pointer. A zoned block
1042 device can be either drive managed, host managed or host aware. For
1043 host managed devices the host must ensure that writes happen
1044 sequentially. Fio recognizes host managed devices and serializes
1045 writes to sequential zones for these devices.
1047 If a read occurs in a sequential zone beyond the write pointer then
1048 the zoned block device will complete the read without reading any data
1049 from the storage medium. Since such reads lead to unrealistically high
1050 bandwidth and IOPS numbers fio only reads beyond the write pointer if
1051 explicitly told to do so. Default: false.
1053 .. option:: max_open_zones=int
1055 A zone of a zoned block device is in the open state when it is partially
1056 written (i.e. not all sectors of the zone have been written). Zoned
1057 block devices may have a limit on the total number of zones that can
1058 be simultaneously in the open state, that is, the number of zones that
1059 can be written to simultaneously. The :option:`max_open_zones` parameter
1060 limits the number of zones to which write commands are issued by all fio
1061 jobs, that is, limits the number of zones that will be in the open
1062 state. This parameter is relevant only if the :option:`zonemode` =zbd is
1063 used. The default value is always equal to maximum number of open zones
1064 of the target zoned block device and a value higher than this limit
1065 cannot be specified by users unless the option
1066 :option:`ignore_zone_limits` is specified. When
1067 :option:`ignore_zone_limits` is specified or the target device has no
1068 limit on the number of zones that can be in an open state,
1069 :option:`max_open_zones` can specify 0 to disable any limit on the
1070 number of zones that can be simultaneously written to by all jobs.
1072 .. option:: job_max_open_zones=int
1074 In the same manner as :option:`max_open_zones`, limit the number of open
1075 zones per fio job, that is, the number of zones that a single job can
1076 simultaneously write to. A value of zero indicates no limit.
1079 .. option:: ignore_zone_limits=bool
1081 If this option is used, fio will ignore the maximum number of open
1082 zones limit of the zoned block device in use, thus allowing the
1083 option :option:`max_open_zones` value to be larger than the device
1084 reported limit. Default: false.
1086 .. option:: zone_reset_threshold=float
1088 A number between zero and one that indicates the ratio of written bytes
1089 in the zones with write pointers in the IO range to the size of the IO
1090 range. When current ratio is above this ratio, zones are reset
1091 periodically as :option:`zone_reset_frequency` specifies.
1093 .. option:: zone_reset_frequency=float
1095 A number between zero and one that indicates how often a zone reset
1096 should be issued if the zone reset threshold has been exceeded. A zone
1097 reset is submitted after each (1 / zone_reset_frequency) write
1098 requests. This and the previous parameter can be used to simulate
1099 garbage collection activity.
1105 .. option:: direct=bool
1107 If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that
1108 OpenBSD and ZFS on Solaris don't support direct I/O. On Windows the synchronous
1109 ioengines don't support direct I/O. Default: false.
1111 .. option:: atomic=bool
1113 If value is true, attempt to use atomic direct I/O. Atomic writes are
1114 guaranteed to be stable once acknowledged by the operating system. Only
1115 Linux supports O_ATOMIC right now.
1117 .. option:: buffered=bool
1119 If value is true, use buffered I/O. This is the opposite of the
1120 :option:`direct` option. Defaults to true.
1122 .. option:: readwrite=str, rw=str
1124 Type of I/O pattern. Accepted values are:
1131 Sequential trims (Linux block devices and SCSI
1132 character devices only).
1138 Random trims (Linux block devices and SCSI
1139 character devices only).
1141 Sequential mixed reads and writes.
1143 Random mixed reads and writes.
1145 Sequential trim+write sequences. Blocks will be trimmed first,
1146 then the same blocks will be written to. So if ``io_size=64K``
1147 is specified, Fio will trim a total of 64K bytes and also
1148 write 64K bytes on the same trimmed blocks. This behaviour
1149 will be consistent with ``number_ios`` or other Fio options
1150 limiting the total bytes or number of I/O's.
1152 Like trimwrite, but uses random offsets rather
1153 than sequential writes.
1155 Fio defaults to read if the option is not specified. For the mixed I/O
1156 types, the default is to split them 50/50. For certain types of I/O the
1157 result may still be skewed a bit, since the speed may be different.
1159 It is possible to specify the number of I/Os to do before getting a new
1160 offset by appending ``:<nr>`` to the end of the string given. For a
1161 random read, it would look like ``rw=randread:8`` for passing in an offset
1162 modifier with a value of 8. If the suffix is used with a sequential I/O
1163 pattern, then the *<nr>* value specified will be **added** to the generated
1164 offset for each I/O turning sequential I/O into sequential I/O with holes.
1165 For instance, using ``rw=write:4k`` will skip 4k for every write. Also see
1166 the :option:`rw_sequencer` option.
1168 .. option:: rw_sequencer=str
1170 If an offset modifier is given by appending a number to the ``rw=<str>``
1171 line, then this option controls how that number modifies the I/O offset
1172 being generated. Accepted values are:
1175 Generate sequential offset.
1177 Generate the same offset.
1179 ``sequential`` is only useful for random I/O, where fio would normally
1180 generate a new random offset for every I/O. If you append e.g. 8 to
1181 randread, i.e. ``rw=randread:8`` you would get a new random offset for
1182 every 8 I/Os. The result would be a sequence of 8 sequential offsets
1183 with a random starting point. However this behavior may change if a
1184 sequential I/O reaches end of the file. As sequential I/O is already
1185 sequential, setting ``sequential`` for that would not result in any
1186 difference. ``identical`` behaves in a similar fashion, except it sends
1187 the same offset 8 number of times before generating a new offset.
1192 rw_sequencer=sequential
1195 The generated sequence of offsets will look like this:
1196 4k, 8k, 12k, 16k, 20k, 24k, 28k, 32k, 92k, 96k, 100k, 104k, 108k,
1197 112k, 116k, 120k, 48k, 52k ...
1202 rw_sequencer=identical
1205 The generated sequence of offsets will look like this:
1206 4k, 4k, 4k, 4k, 4k, 4k, 4k, 4k, 92k, 92k, 92k, 92k, 92k, 92k, 92k, 92k,
1209 .. option:: unified_rw_reporting=str
1211 Fio normally reports statistics on a per data direction basis, meaning that
1212 reads, writes, and trims are accounted and reported separately. This option
1213 determines whether fio reports the results normally, summed together, or as
1215 Accepted values are:
1218 Normal statistics reporting.
1221 Statistics are summed per data direction and reported together.
1224 Statistics are reported normally, followed by the mixed statistics.
1227 Backward-compatible alias for **none**.
1230 Backward-compatible alias for **mixed**.
1235 .. option:: randrepeat=bool
1237 Seed the random number generator used for random I/O patterns in a
1238 predictable way so the pattern is repeatable across runs. Default: true.
1240 .. option:: allrandrepeat=bool
1242 Seed all random number generators in a predictable way so results are
1243 repeatable across runs. Default: false.
1245 .. option:: randseed=int
1247 Seed the random number generators based on this seed value, to be able to
1248 control what sequence of output is being generated. If not set, the random
1249 sequence depends on the :option:`randrepeat` setting.
1251 .. option:: fallocate=str
1253 Whether pre-allocation is performed when laying down files.
1254 Accepted values are:
1257 Do not pre-allocate space.
1260 Use a platform's native pre-allocation call but fall back to
1261 **none** behavior if it fails/is not implemented.
1264 Pre-allocate via :manpage:`posix_fallocate(3)`.
1267 Pre-allocate via :manpage:`fallocate(2)` with
1268 FALLOC_FL_KEEP_SIZE set.
1271 Extend file to final size via :manpage:`ftruncate(2)`
1272 instead of allocating.
1275 Backward-compatible alias for **none**.
1278 Backward-compatible alias for **posix**.
1280 May not be available on all supported platforms. **keep** is only available
1281 on Linux. If using ZFS on Solaris this cannot be set to **posix**
1282 because ZFS doesn't support pre-allocation. Default: **native** if any
1283 pre-allocation methods except **truncate** are available, **none** if not.
1285 Note that using **truncate** on Windows will interact surprisingly
1286 with non-sequential write patterns. When writing to a file that has
1287 been extended by setting the end-of-file information, Windows will
1288 backfill the unwritten portion of the file up to that offset with
1289 zeroes before issuing the new write. This means that a single small
1290 write to the end of an extended file will stall until the entire
1291 file has been filled with zeroes.
1293 .. option:: fadvise_hint=str
1295 Use :manpage:`posix_fadvise(2)` or :manpage:`posix_fadvise(2)` to
1296 advise the kernel on what I/O patterns are likely to be issued.
1297 Accepted values are:
1300 Backwards-compatible hint for "no hint".
1303 Backwards compatible hint for "advise with fio workload type". This
1304 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1305 for a sequential workload.
1308 Advise using **FADV_SEQUENTIAL**.
1311 Advise using **FADV_RANDOM**.
1313 .. option:: write_hint=str
1315 Use :manpage:`fcntl(2)` to advise the kernel what life time to expect
1316 from a write. Only supported on Linux, as of version 4.13. Accepted
1320 No particular life time associated with this file.
1323 Data written to this file has a short life time.
1326 Data written to this file has a medium life time.
1329 Data written to this file has a long life time.
1332 Data written to this file has a very long life time.
1334 The values are all relative to each other, and no absolute meaning
1335 should be associated with them.
1337 .. option:: offset=int
1339 Start I/O at the provided offset in the file, given as either a fixed size in
1340 bytes, zones or a percentage. If a percentage is given, the generated offset will be
1341 aligned to the minimum ``blocksize`` or to the value of ``offset_align`` if
1342 provided. Data before the given offset will not be touched. This
1343 effectively caps the file size at `real_size - offset`. Can be combined with
1344 :option:`size` to constrain the start and end range of the I/O workload.
1345 A percentage can be specified by a number between 1 and 100 followed by '%',
1346 for example, ``offset=20%`` to specify 20%. In ZBD mode, value can be set as
1347 number of zones using 'z'.
1349 .. option:: offset_align=int
1351 If set to non-zero value, the byte offset generated by a percentage ``offset``
1352 is aligned upwards to this value. Defaults to 0 meaning that a percentage
1353 offset is aligned to the minimum block size.
1355 .. option:: offset_increment=int
1357 If this is provided, then the real offset becomes `offset + offset_increment
1358 * thread_number`, where the thread number is a counter that starts at 0 and
1359 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1360 specified). This option is useful if there are several jobs which are
1361 intended to operate on a file in parallel disjoint segments, with even
1362 spacing between the starting points. Percentages can be used for this option.
1363 If a percentage is given, the generated offset will be aligned to the minimum
1364 ``blocksize`` or to the value of ``offset_align`` if provided. In ZBD mode, value can
1365 also be set as number of zones using 'z'.
1367 .. option:: number_ios=int
1369 Fio will normally perform I/Os until it has exhausted the size of the region
1370 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1371 condition). With this setting, the range/size can be set independently of
1372 the number of I/Os to perform. When fio reaches this number, it will exit
1373 normally and report status. Note that this does not extend the amount of I/O
1374 that will be done, it will only stop fio if this condition is met before
1375 other end-of-job criteria.
1377 .. option:: fsync=int
1379 If writing to a file, issue an :manpage:`fsync(2)` (or its equivalent) of
1380 the dirty data for every number of blocks given. For example, if you give 32
1381 as a parameter, fio will sync the file after every 32 writes issued. If fio is
1382 using non-buffered I/O, we may not sync the file. The exception is the sg
1383 I/O engine, which synchronizes the disk cache anyway. Defaults to 0, which
1384 means fio does not periodically issue and wait for a sync to complete. Also
1385 see :option:`end_fsync` and :option:`fsync_on_close`.
1387 .. option:: fdatasync=int
1389 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
1390 not metadata blocks. In Windows, DragonFlyBSD or OSX there is no
1391 :manpage:`fdatasync(2)` so this falls back to using :manpage:`fsync(2)`.
1392 Defaults to 0, which means fio does not periodically issue and wait for a
1393 data-only sync to complete.
1395 .. option:: write_barrier=int
1397 Make every `N-th` write a barrier write.
1399 .. option:: sync_file_range=str:int
1401 Use :manpage:`sync_file_range(2)` for every `int` number of write
1402 operations. Fio will track range of writes that have happened since the last
1403 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1406 SYNC_FILE_RANGE_WAIT_BEFORE
1408 SYNC_FILE_RANGE_WRITE
1410 SYNC_FILE_RANGE_WAIT_AFTER
1412 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1413 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1414 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1417 .. option:: overwrite=bool
1419 If true, writes to a file will always overwrite existing data. If the file
1420 doesn't already exist, it will be created before the write phase begins. If
1421 the file exists and is large enough for the specified write phase, nothing
1422 will be done. Default: false.
1424 .. option:: end_fsync=bool
1426 If true, :manpage:`fsync(2)` file contents when a write stage has completed.
1429 .. option:: fsync_on_close=bool
1431 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
1432 from :option:`end_fsync` in that it will happen on every file close, not
1433 just at the end of the job. Default: false.
1435 .. option:: rwmixread=int
1437 Percentage of a mixed workload that should be reads. Default: 50.
1439 .. option:: rwmixwrite=int
1441 Percentage of a mixed workload that should be writes. If both
1442 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1443 add up to 100%, the latter of the two will be used to override the
1444 first. This may interfere with a given rate setting, if fio is asked to
1445 limit reads or writes to a certain rate. If that is the case, then the
1446 distribution may be skewed. Default: 50.
1448 .. option:: random_distribution=str:float[:float][,str:float][,str:float]
1450 By default, fio will use a completely uniform random distribution when asked
1451 to perform random I/O. Sometimes it is useful to skew the distribution in
1452 specific ways, ensuring that some parts of the data is more hot than others.
1453 fio includes the following distribution models:
1456 Uniform random distribution
1465 Normal (Gaussian) distribution
1468 Zoned random distribution
1471 Zone absolute random distribution
1473 When using a **zipf** or **pareto** distribution, an input value is also
1474 needed to define the access pattern. For **zipf**, this is the `Zipf
1475 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
1476 program, :command:`fio-genzipf`, that can be used visualize what the given input
1477 values will yield in terms of hit rates. If you wanted to use **zipf** with
1478 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1479 option. If a non-uniform model is used, fio will disable use of the random
1480 map. For the **normal** distribution, a normal (Gaussian) deviation is
1481 supplied as a value between 0 and 100.
1483 The second, optional float is allowed for **pareto**, **zipf** and **normal** distributions.
1484 It allows one to set base of distribution in non-default place, giving more control
1485 over most probable outcome. This value is in range [0-1] which maps linearly to
1486 range of possible random values.
1487 Defaults are: random for **pareto** and **zipf**, and 0.5 for **normal**.
1488 If you wanted to use **zipf** with a `theta` of 1.2 centered on 1/4 of allowed value range,
1489 you would use ``random_distribution=zipf:1.2:0.25``.
1491 For a **zoned** distribution, fio supports specifying percentages of I/O
1492 access that should fall within what range of the file or device. For
1493 example, given a criteria of:
1495 * 60% of accesses should be to the first 10%
1496 * 30% of accesses should be to the next 20%
1497 * 8% of accesses should be to the next 30%
1498 * 2% of accesses should be to the next 40%
1500 we can define that through zoning of the random accesses. For the above
1501 example, the user would do::
1503 random_distribution=zoned:60/10:30/20:8/30:2/40
1505 A **zoned_abs** distribution works exactly like the **zoned**, except
1506 that it takes absolute sizes. For example, let's say you wanted to
1507 define access according to the following criteria:
1509 * 60% of accesses should be to the first 20G
1510 * 30% of accesses should be to the next 100G
1511 * 10% of accesses should be to the next 500G
1513 we can define an absolute zoning distribution with:
1515 random_distribution=zoned_abs=60/20G:30/100G:10/500g
1517 For both **zoned** and **zoned_abs**, fio supports defining up to
1520 Similarly to how :option:`bssplit` works for setting ranges and
1521 percentages of block sizes. Like :option:`bssplit`, it's possible to
1522 specify separate zones for reads, writes, and trims. If just one set
1523 is given, it'll apply to all of them. This goes for both **zoned**
1524 **zoned_abs** distributions.
1526 .. option:: percentage_random=int[,int][,int]
1528 For a random workload, set how big a percentage should be random. This
1529 defaults to 100%, in which case the workload is fully random. It can be set
1530 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1531 sequential. Any setting in between will result in a random mix of sequential
1532 and random I/O, at the given percentages. Comma-separated values may be
1533 specified for reads, writes, and trims as described in :option:`blocksize`.
1535 .. option:: norandommap
1537 Normally fio will cover every block of the file when doing random I/O. If
1538 this option is given, fio will just get a new random offset without looking
1539 at past I/O history. This means that some blocks may not be read or written,
1540 and that some blocks may be read/written more than once. If this option is
1541 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1542 only intact blocks are verified, i.e., partially-overwritten blocks are
1543 ignored. With an async I/O engine and an I/O depth > 1, it is possible for
1544 the same block to be overwritten, which can cause verification errors. Either
1545 do not use norandommap in this case, or also use the lfsr random generator.
1547 .. option:: softrandommap=bool
1549 See :option:`norandommap`. If fio runs with the random block map enabled and
1550 it fails to allocate the map, if this option is set it will continue without
1551 a random block map. As coverage will not be as complete as with random maps,
1552 this option is disabled by default.
1554 .. option:: random_generator=str
1556 Fio supports the following engines for generating I/O offsets for random I/O:
1559 Strong 2^88 cycle random number generator.
1561 Linear feedback shift register generator.
1563 Strong 64-bit 2^258 cycle random number generator.
1565 **tausworthe** is a strong random number generator, but it requires tracking
1566 on the side if we want to ensure that blocks are only read or written
1567 once. **lfsr** guarantees that we never generate the same offset twice, and
1568 it's also less computationally expensive. It's not a true random generator,
1569 however, though for I/O purposes it's typically good enough. **lfsr** only
1570 works with single block sizes, not with workloads that use multiple block
1571 sizes. If used with such a workload, fio may read or write some blocks
1572 multiple times. The default value is **tausworthe**, unless the required
1573 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1574 selected automatically.
1580 .. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1582 The block size in bytes used for I/O units. Default: 4096. A single value
1583 applies to reads, writes, and trims. Comma-separated values may be
1584 specified for reads, writes, and trims. A value not terminated in a comma
1585 applies to subsequent types.
1590 means 256k for reads, writes and trims.
1593 means 8k for reads, 32k for writes and trims.
1596 means 8k for reads, 32k for writes, and default for trims.
1599 means default for reads, 8k for writes and trims.
1602 means default for reads, 8k for writes, and default for trims.
1604 .. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1606 A range of block sizes in bytes for I/O units. The issued I/O unit will
1607 always be a multiple of the minimum size, unless
1608 :option:`blocksize_unaligned` is set.
1610 Comma-separated ranges may be specified for reads, writes, and trims as
1611 described in :option:`blocksize`.
1613 Example: ``bsrange=1k-4k,2k-8k``.
1615 .. option:: bssplit=str[,str][,str]
1617 Sometimes you want even finer grained control of the block sizes
1618 issued, not just an even split between them. This option allows you to
1619 weight various block sizes, so that you are able to define a specific
1620 amount of block sizes issued. The format for this option is::
1622 bssplit=blocksize/percentage:blocksize/percentage
1624 for as many block sizes as needed. So if you want to define a workload
1625 that has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would
1628 bssplit=4k/10:64k/50:32k/40
1630 Ordering does not matter. If the percentage is left blank, fio will
1631 fill in the remaining values evenly. So a bssplit option like this one::
1633 bssplit=4k/50:1k/:32k/
1635 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always
1636 add up to 100, if bssplit is given a range that adds up to more, it
1639 Comma-separated values may be specified for reads, writes, and trims as
1640 described in :option:`blocksize`.
1642 If you want a workload that has 50% 2k reads and 50% 4k reads, while
1643 having 90% 4k writes and 10% 8k writes, you would specify::
1645 bssplit=2k/50:4k/50,4k/90:8k/10
1647 Fio supports defining up to 64 different weights for each data
1650 .. option:: blocksize_unaligned, bs_unaligned
1652 If set, fio will issue I/O units with any size within
1653 :option:`blocksize_range`, not just multiples of the minimum size. This
1654 typically won't work with direct I/O, as that normally requires sector
1657 .. option:: bs_is_seq_rand=bool
1659 If this option is set, fio will use the normal read,write blocksize settings
1660 as sequential,random blocksize settings instead. Any random read or write
1661 will use the WRITE blocksize settings, and any sequential read or write will
1662 use the READ blocksize settings.
1664 .. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1666 Boundary to which fio will align random I/O units. Default:
1667 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1668 I/O, though it usually depends on the hardware block size. This option is
1669 mutually exclusive with using a random map for files, so it will turn off
1670 that option. Comma-separated values may be specified for reads, writes, and
1671 trims as described in :option:`blocksize`.
1677 .. option:: zero_buffers
1679 Initialize buffers with all zeros. Default: fill buffers with random data.
1681 .. option:: refill_buffers
1683 If this option is given, fio will refill the I/O buffers on every
1684 submit. Only makes sense if :option:`zero_buffers` isn't specified,
1685 naturally. Defaults to being unset i.e., the buffer is only filled at
1686 init time and the data in it is reused when possible but if any of
1687 :option:`verify`, :option:`buffer_compress_percentage` or
1688 :option:`dedupe_percentage` are enabled then `refill_buffers` is also
1689 automatically enabled.
1691 .. option:: scramble_buffers=bool
1693 If :option:`refill_buffers` is too costly and the target is using data
1694 deduplication, then setting this option will slightly modify the I/O buffer
1695 contents to defeat normal de-dupe attempts. This is not enough to defeat
1696 more clever block compression attempts, but it will stop naive dedupe of
1697 blocks. Default: true.
1699 .. option:: buffer_compress_percentage=int
1701 If this is set, then fio will attempt to provide I/O buffer content
1702 (on WRITEs) that compresses to the specified level. Fio does this by
1703 providing a mix of random data followed by fixed pattern data. The
1704 fixed pattern is either zeros, or the pattern specified by
1705 :option:`buffer_pattern`. If the `buffer_pattern` option is used, it
1706 might skew the compression ratio slightly. Setting
1707 `buffer_compress_percentage` to a value other than 100 will also
1708 enable :option:`refill_buffers` in order to reduce the likelihood that
1709 adjacent blocks are so similar that they over compress when seen
1710 together. See :option:`buffer_compress_chunk` for how to set a finer or
1711 coarser granularity for the random/fixed data region. Defaults to unset
1712 i.e., buffer data will not adhere to any compression level.
1714 .. option:: buffer_compress_chunk=int
1716 This setting allows fio to manage how big the random/fixed data region
1717 is when using :option:`buffer_compress_percentage`. When
1718 `buffer_compress_chunk` is set to some non-zero value smaller than the
1719 block size, fio can repeat the random/fixed region throughout the I/O
1720 buffer at the specified interval (which particularly useful when
1721 bigger block sizes are used for a job). When set to 0, fio will use a
1722 chunk size that matches the block size resulting in a single
1723 random/fixed region within the I/O buffer. Defaults to 512. When the
1724 unit is omitted, the value is interpreted in bytes.
1726 .. option:: buffer_pattern=str
1728 If set, fio will fill the I/O buffers with this pattern or with the contents
1729 of a file. If not set, the contents of I/O buffers are defined by the other
1730 options related to buffer contents. The setting can be any pattern of bytes,
1731 and can be prefixed with 0x for hex values. It may also be a string, where
1732 the string must then be wrapped with ``""``. Or it may also be a filename,
1733 where the filename must be wrapped with ``''`` in which case the file is
1734 opened and read. Note that not all the file contents will be read if that
1735 would cause the buffers to overflow. So, for example::
1737 buffer_pattern='filename'
1741 buffer_pattern="abcd"
1749 buffer_pattern=0xdeadface
1751 Also you can combine everything together in any order::
1753 buffer_pattern=0xdeadface"abcd"-12'filename'
1755 .. option:: dedupe_percentage=int
1757 If set, fio will generate this percentage of identical buffers when
1758 writing. These buffers will be naturally dedupable. The contents of the
1759 buffers depend on what other buffer compression settings have been set. It's
1760 possible to have the individual buffers either fully compressible, or not at
1761 all -- this option only controls the distribution of unique buffers. Setting
1762 this option will also enable :option:`refill_buffers` to prevent every buffer
1765 .. option:: dedupe_mode=str
1767 If ``dedupe_percentage=<int>`` is given, then this option controls how fio
1768 generates the dedupe buffers.
1771 Generate dedupe buffers by repeating previous writes
1773 Generate dedupe buffers from working set
1775 ``repeat`` is the default option for fio. Dedupe buffers are generated
1776 by repeating previous unique write.
1778 ``working_set`` is a more realistic workload.
1779 With ``working_set``, ``dedupe_working_set_percentage=<int>`` should be provided.
1780 Given that, fio will use the initial unique write buffers as its working set.
1781 Upon deciding to dedupe, fio will randomly choose a buffer from the working set.
1782 Note that by using ``working_set`` the dedupe percentage will converge
1783 to the desired over time while ``repeat`` maintains the desired percentage
1786 .. option:: dedupe_working_set_percentage=int
1788 If ``dedupe_mode=<str>`` is set to ``working_set``, then this controls
1789 the percentage of size of the file or device used as the buffers
1790 fio will choose to generate the dedupe buffers from
1792 Note that size needs to be explicitly provided and only 1 file per
1795 .. option:: dedupe_global=bool
1797 This controls whether the deduplication buffers will be shared amongst
1798 all jobs that have this option set. The buffers are spread evenly between
1801 .. option:: invalidate=bool
1803 Invalidate the buffer/page cache parts of the files to be used prior to
1804 starting I/O if the platform and file type support it. Defaults to true.
1805 This will be ignored if :option:`pre_read` is also specified for the
1808 .. option:: sync=str
1810 Whether, and what type, of synchronous I/O to use for writes. The allowed
1814 Do not use synchronous IO, the default.
1820 Use synchronous file IO. For the majority of I/O engines,
1821 this means using O_SYNC.
1827 Use synchronous data IO. For the majority of I/O engines,
1828 this means using O_DSYNC.
1831 .. option:: iomem=str, mem=str
1833 Fio can use various types of memory as the I/O unit buffer. The allowed
1837 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1841 Use shared memory as the buffers. Allocated through
1842 :manpage:`shmget(2)`.
1845 Same as shm, but use huge pages as backing.
1848 Use :manpage:`mmap(2)` to allocate buffers. May either be anonymous memory, or can
1849 be file backed if a filename is given after the option. The format
1850 is `mem=mmap:/path/to/file`.
1853 Use a memory mapped huge file as the buffer backing. Append filename
1854 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1857 Same as mmap, but use a MMAP_SHARED mapping.
1860 Use GPU memory as the buffers for GPUDirect RDMA benchmark.
1861 The :option:`ioengine` must be `rdma`.
1863 The area allocated is a function of the maximum allowed bs size for the job,
1864 multiplied by the I/O depth given. Note that for **shmhuge** and
1865 **mmaphuge** to work, the system must have free huge pages allocated. This
1866 can normally be checked and set by reading/writing
1867 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1868 is 2 or 4MiB in size depending on the platform. So to calculate the
1869 number of huge pages you need for a given job file, add up the I/O
1870 depth of all jobs (normally one unless :option:`iodepth` is used) and
1871 multiply by the maximum bs set. Then divide that number by the huge
1872 page size. You can see the size of the huge pages in
1873 :file:`/proc/meminfo`. If no huge pages are allocated by having a
1874 non-zero number in `nr_hugepages`, using **mmaphuge** or **shmhuge**
1875 will fail. Also see :option:`hugepage-size`.
1877 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1878 should point there. So if it's mounted in :file:`/huge`, you would use
1879 `mem=mmaphuge:/huge/somefile`.
1881 .. option:: iomem_align=int, mem_align=int
1883 This indicates the memory alignment of the I/O memory buffers. Note that
1884 the given alignment is applied to the first I/O unit buffer, if using
1885 :option:`iodepth` the alignment of the following buffers are given by the
1886 :option:`bs` used. In other words, if using a :option:`bs` that is a
1887 multiple of the page sized in the system, all buffers will be aligned to
1888 this value. If using a :option:`bs` that is not page aligned, the alignment
1889 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1892 .. option:: hugepage-size=int
1894 Defines the size of a huge page. Must at least be equal to the system
1895 setting, see :file:`/proc/meminfo` and
1896 :file:`/sys/kernel/mm/hugepages/`. Defaults to 2 or 4MiB depending on
1897 the platform. Should probably always be a multiple of megabytes, so
1898 using ``hugepage-size=Xm`` is the preferred way to set this to avoid
1899 setting a non-pow-2 bad value.
1901 .. option:: lockmem=int
1903 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1904 simulate a smaller amount of memory. The amount specified is per worker.
1910 .. option:: size=int
1912 The total size of file I/O for each thread of this job. Fio will run until
1913 this many bytes has been transferred, unless runtime is altered by other means
1914 such as (1) :option:`runtime`, (2) :option:`io_size` (3) :option:`number_ios`,
1915 (4) gaps/holes while doing I/O's such as ``rw=read:16K``, or (5) sequential
1916 I/O reaching end of the file which is possible when :option:`percentage_random`
1918 Fio will divide this size between the available files determined by options
1919 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1920 specified by the job. If the result of division happens to be 0, the size is
1921 set to the physical size of the given files or devices if they exist.
1922 If this option is not specified, fio will use the full size of the given
1923 files or devices. If the files do not exist, size must be given. It is also
1924 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1925 given, fio will use 20% of the full size of the given files or devices.
1926 In ZBD mode, value can also be set as number of zones using 'z'.
1927 Can be combined with :option:`offset` to constrain the start and end range
1928 that I/O will be done within.
1930 .. option:: io_size=int, io_limit=int
1932 Normally fio operates within the region set by :option:`size`, which means
1933 that the :option:`size` option sets both the region and size of I/O to be
1934 performed. Sometimes that is not what you want. With this option, it is
1935 possible to define just the amount of I/O that fio should do. For instance,
1936 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1937 will perform I/O within the first 20GiB but exit when 5GiB have been
1938 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1939 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1940 the 0..20GiB region.
1942 .. option:: filesize=irange(int)
1944 Individual file sizes. May be a range, in which case fio will select sizes for
1945 files at random within the given range. If not given, each created file is the
1946 same size. This option overrides :option:`size` in terms of file size, i.e. if
1947 :option:`filesize` is specified then :option:`size` becomes merely the default
1948 for :option:`io_size` and has no effect at all if :option:`io_size` is set
1951 .. option:: file_append=bool
1953 Perform I/O after the end of the file. Normally fio will operate within the
1954 size of a file. If this option is set, then fio will append to the file
1955 instead. This has identical behavior to setting :option:`offset` to the size
1956 of a file. This option is ignored on non-regular files.
1958 .. option:: fill_device=bool, fill_fs=bool
1960 Sets size to something really large and waits for ENOSPC (no space left on
1961 device) or EDQUOT (disk quota exceeded)
1962 as the terminating condition. Only makes sense with sequential
1963 write. For a read workload, the mount point will be filled first then I/O
1964 started on the result. This option doesn't make sense if operating on a raw
1965 device node, since the size of that is already known by the file system.
1966 Additionally, writing beyond end-of-device will not return ENOSPC there.
1972 .. option:: ioengine=str
1974 Defines how the job issues I/O to the file. The following types are defined:
1977 Basic :manpage:`read(2)` or :manpage:`write(2)`
1978 I/O. :manpage:`lseek(2)` is used to position the I/O location.
1979 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
1982 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1983 all supported operating systems except for Windows.
1986 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
1987 queuing by coalescing adjacent I/Os into a single submission.
1990 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1993 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1996 Fast Linux native asynchronous I/O. Supports async IO
1997 for both direct and buffered IO.
1998 This engine defines engine specific options.
2001 Fast Linux native asynchronous I/O for pass through commands.
2002 This engine defines engine specific options.
2005 Linux native asynchronous I/O. Note that Linux may only support
2006 queued behavior with non-buffered I/O (set ``direct=1`` or
2008 This engine defines engine specific options.
2011 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
2012 :manpage:`aio_write(3)`.
2015 Solaris native asynchronous I/O.
2018 Windows native asynchronous I/O. Default on Windows.
2021 File is memory mapped with :manpage:`mmap(2)` and data copied
2022 to/from using :manpage:`memcpy(3)`.
2025 :manpage:`splice(2)` is used to transfer the data and
2026 :manpage:`vmsplice(2)` to transfer data from user space to the
2030 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
2031 ioctl, or if the target is an sg character device we use
2032 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
2033 I/O. Requires :option:`filename` option to specify either block or
2034 character devices. This engine supports trim operations.
2035 The sg engine includes engine specific options.
2038 Read, write, trim and ZBC/ZAC operations to a zoned
2039 block device using libzbc library. The target can be
2040 either an SG character device or a block device file.
2043 Doesn't transfer any data, just pretends to. This is mainly used to
2044 exercise fio itself and for debugging/testing purposes.
2047 Transfer over the network to given ``host:port``. Depending on the
2048 :option:`protocol` used, the :option:`hostname`, :option:`port`,
2049 :option:`listen` and :option:`filename` options are used to specify
2050 what sort of connection to make, while the :option:`protocol` option
2051 determines which protocol will be used. This engine defines engine
2055 Like **net**, but uses :manpage:`splice(2)` and
2056 :manpage:`vmsplice(2)` to map data and send/receive.
2057 This engine defines engine specific options.
2060 Doesn't transfer any data, but burns CPU cycles according to the
2061 :option:`cpuload`, :option:`cpuchunks` and :option:`cpumode` options.
2062 Setting :option:`cpuload`\=85 will cause that job to do nothing but burn 85%
2063 of the CPU. In case of SMP machines, use :option:`numjobs`\=<nr_of_cpu>
2064 to get desired CPU usage, as the cpuload only loads a
2065 single CPU at the desired rate. A job never finishes unless there is
2066 at least one non-cpuio job.
2067 Setting :option:`cpumode`\=qsort replace the default noop instructions loop
2068 by a qsort algorithm to consume more energy.
2071 The RDMA I/O engine supports both RDMA memory semantics
2072 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
2073 InfiniBand, RoCE and iWARP protocols. This engine defines engine
2077 I/O engine that does regular fallocate to simulate data transfer as
2081 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
2084 does fallocate(,mode = 0).
2087 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
2090 I/O engine that sends :manpage:`ftruncate(2)` operations in response
2091 to write (DDIR_WRITE) events. Each ftruncate issued sets the file's
2092 size to the current block offset. :option:`blocksize` is ignored.
2095 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
2096 defragment activity in request to DDIR_WRITE event.
2099 I/O engine supporting direct access to Ceph Reliable Autonomic
2100 Distributed Object Store (RADOS) via librados. This ioengine
2101 defines engine specific options.
2104 I/O engine supporting direct access to Ceph Rados Block Devices
2105 (RBD) via librbd without the need to use the kernel rbd driver. This
2106 ioengine defines engine specific options.
2109 I/O engine supporting GET/PUT requests over HTTP(S) with libcurl to
2110 a WebDAV or S3 endpoint. This ioengine defines engine specific options.
2112 This engine only supports direct IO of iodepth=1; you need to scale this
2113 via numjobs. blocksize defines the size of the objects to be created.
2115 TRIM is translated to object deletion.
2118 Using GlusterFS libgfapi sync interface to direct access to
2119 GlusterFS volumes without having to go through FUSE. This ioengine
2120 defines engine specific options.
2123 Using GlusterFS libgfapi async interface to direct access to
2124 GlusterFS volumes without having to go through FUSE. This ioengine
2125 defines engine specific options.
2128 Read and write through Hadoop (HDFS). The :option:`filename` option
2129 is used to specify host,port of the hdfs name-node to connect. This
2130 engine interprets offsets a little differently. In HDFS, files once
2131 created cannot be modified so random writes are not possible. To
2132 imitate this the libhdfs engine expects a bunch of small files to be
2133 created over HDFS and will randomly pick a file from them
2134 based on the offset generated by fio backend (see the example
2135 job file to create such files, use ``rw=write`` option). Please
2136 note, it may be necessary to set environment variables to work
2137 with HDFS/libhdfs properly. Each job uses its own connection to
2141 Read, write and erase an MTD character device (e.g.,
2142 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
2143 underlying device type, the I/O may have to go in a certain pattern,
2144 e.g., on NAND, writing sequentially to erase blocks and discarding
2145 before overwriting. The `trimwrite` mode works well for this
2149 Read and write using filesystem DAX to a file on a filesystem
2150 mounted with DAX on a persistent memory device through the PMDK
2154 Read and write using device DAX to a persistent memory device (e.g.,
2155 /dev/dax0.0) through the PMDK libpmem library.
2158 Prefix to specify loading an external I/O engine object file. Append
2159 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
2160 ioengine :file:`foo.o` in :file:`/tmp`. The path can be either
2161 absolute or relative. See :file:`engines/skeleton_external.c` for
2162 details of writing an external I/O engine.
2165 Simply create the files and do no I/O to them. You still need to
2166 set `filesize` so that all the accounting still occurs, but no
2167 actual I/O will be done other than creating the file.
2170 Simply do stat() and do no I/O to the file. You need to set 'filesize'
2171 and 'nrfiles', so that files will be created.
2172 This engine is to measure file lookup and meta data access.
2175 Simply delete the files by unlink() and do no I/O to them. You need to set 'filesize'
2176 and 'nrfiles', so that the files will be created.
2177 This engine is to measure file delete.
2180 Read and write using mmap I/O to a file on a filesystem
2181 mounted with DAX on a persistent memory device through the PMDK
2185 Synchronous read and write using DDN's Infinite Memory Engine (IME).
2186 This engine is very basic and issues calls to IME whenever an IO is
2190 Synchronous read and write using DDN's Infinite Memory Engine (IME).
2191 This engine uses iovecs and will try to stack as much IOs as possible
2192 (if the IOs are "contiguous" and the IO depth is not exceeded)
2193 before issuing a call to IME.
2196 Asynchronous read and write using DDN's Infinite Memory Engine (IME).
2197 This engine will try to stack as much IOs as possible by creating
2198 requests for IME. FIO will then decide when to commit these requests.
2201 Read and write iscsi lun with libiscsi.
2204 Read and write a Network Block Device (NBD).
2207 I/O engine supporting libcufile synchronous access to nvidia-fs and a
2208 GPUDirect Storage-supported filesystem. This engine performs
2209 I/O without transferring buffers between user-space and the kernel,
2210 unless :option:`verify` is set or :option:`cuda_io` is `posix`.
2211 :option:`iomem` must not be `cudamalloc`. This ioengine defines
2212 engine specific options.
2215 I/O engine supporting asynchronous read and write operations to the
2216 DAOS File System (DFS) via libdfs.
2219 I/O engine supporting asynchronous read and write operations to
2220 NFS filesystems from userspace via libnfs. This is useful for
2221 achieving higher concurrency and thus throughput than is possible
2225 Execute 3rd party tools. Could be used to perform monitoring during jobs runtime.
2228 I/O engine using the xNVMe C API, for NVMe devices. The xnvme engine provides
2229 flexibility to access GNU/Linux Kernel NVMe driver via libaio, IOCTLs, io_uring,
2230 the SPDK NVMe driver, or your own custom NVMe driver. The xnvme engine includes
2231 engine specific options. (See https://xnvme.io).
2234 Use the libblkio library
2235 (https://gitlab.com/libblkio/libblkio). The specific
2236 *driver* to use must be set using
2237 :option:`libblkio_driver`. If
2238 :option:`mem`/:option:`iomem` is not specified, memory
2239 allocation is delegated to libblkio (and so is
2240 guaranteed to work with the selected *driver*). One
2241 libblkio instance is used per process, so all jobs
2242 setting option :option:`thread` will share a single
2243 instance (with one queue per thread) and must specify
2244 compatible options. Note that some drivers don't allow
2245 several instances to access the same device or file
2246 simultaneously, but allow it for threads.
2248 I/O engine specific parameters
2249 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2251 In addition, there are some parameters which are only valid when a specific
2252 :option:`ioengine` is in use. These are used identically to normal parameters,
2253 with the caveat that when used on the command line, they must come after the
2254 :option:`ioengine` that defines them is selected.
2256 .. option:: cmdprio_percentage=int[,int] : [io_uring] [libaio]
2258 Set the percentage of I/O that will be issued with the highest priority.
2259 Default: 0. A single value applies to reads and writes. Comma-separated
2260 values may be specified for reads and writes. For this option to be
2261 effective, NCQ priority must be supported and enabled, and the :option:`direct`
2262 option must be set. fio must also be run as the root user. Unlike
2263 slat/clat/lat stats, which can be tracked and reported independently, per
2264 priority stats only track and report a single type of latency. By default,
2265 completion latency (clat) will be reported, if :option:`lat_percentiles` is
2266 set, total latency (lat) will be reported.
2268 .. option:: cmdprio_class=int[,int] : [io_uring] [libaio]
2270 Set the I/O priority class to use for I/Os that must be issued with
2271 a priority when :option:`cmdprio_percentage` or
2272 :option:`cmdprio_bssplit` is set. If not specified when
2273 :option:`cmdprio_percentage` or :option:`cmdprio_bssplit` is set,
2274 this defaults to the highest priority class. A single value applies
2275 to reads and writes. Comma-separated values may be specified for
2276 reads and writes. See :manpage:`ionice(1)`. See also the
2277 :option:`prioclass` option.
2279 .. option:: cmdprio=int[,int] : [io_uring] [libaio]
2281 Set the I/O priority value to use for I/Os that must be issued with
2282 a priority when :option:`cmdprio_percentage` or
2283 :option:`cmdprio_bssplit` is set. If not specified when
2284 :option:`cmdprio_percentage` or :option:`cmdprio_bssplit` is set,
2286 Linux limits us to a positive value between 0 and 7, with 0 being the
2287 highest. A single value applies to reads and writes. Comma-separated
2288 values may be specified for reads and writes. See :manpage:`ionice(1)`.
2289 Refer to an appropriate manpage for other operating systems since
2290 meaning of priority may differ. See also the :option:`prio` option.
2292 .. option:: cmdprio_bssplit=str[,str] : [io_uring] [libaio]
2294 To get a finer control over I/O priority, this option allows
2295 specifying the percentage of IOs that must have a priority set
2296 depending on the block size of the IO. This option is useful only
2297 when used together with the :option:`bssplit` option, that is,
2298 multiple different block sizes are used for reads and writes.
2300 The first accepted format for this option is the same as the format of
2301 the :option:`bssplit` option:
2303 cmdprio_bssplit=blocksize/percentage:blocksize/percentage
2305 In this case, each entry will use the priority class and priority
2306 level defined by the options :option:`cmdprio_class` and
2307 :option:`cmdprio` respectively.
2309 The second accepted format for this option is:
2311 cmdprio_bssplit=blocksize/percentage/class/level:blocksize/percentage/class/level
2313 In this case, the priority class and priority level is defined inside
2314 each entry. In comparison with the first accepted format, the second
2315 accepted format does not restrict all entries to have the same priority
2316 class and priority level.
2318 For both formats, only the read and write data directions are supported,
2319 values for trim IOs are ignored. This option is mutually exclusive with
2320 the :option:`cmdprio_percentage` option.
2322 .. option:: fixedbufs : [io_uring] [io_uring_cmd]
2324 If fio is asked to do direct IO, then Linux will map pages for each
2325 IO call, and release them when IO is done. If this option is set, the
2326 pages are pre-mapped before IO is started. This eliminates the need to
2327 map and release for each IO. This is more efficient, and reduces the
2330 .. option:: nonvectored=int : [io_uring] [io_uring_cmd]
2332 With this option, fio will use non-vectored read/write commands, where
2333 address must contain the address directly. Default is -1.
2335 .. option:: force_async=int : [io_uring] [io_uring_cmd]
2337 Normal operation for io_uring is to try and issue an sqe as
2338 non-blocking first, and if that fails, execute it in an async manner.
2339 With this option set to N, then every N request fio will ask sqe to
2340 be issued in an async manner. Default is 0.
2342 .. option:: registerfiles : [io_uring] [io_uring_cmd]
2344 With this option, fio registers the set of files being used with the
2345 kernel. This avoids the overhead of managing file counts in the kernel,
2346 making the submission and completion part more lightweight. Required
2347 for the below :option:`sqthread_poll` option.
2349 .. option:: sqthread_poll : [io_uring] [io_uring_cmd] [xnvme]
2351 Normally fio will submit IO by issuing a system call to notify the
2352 kernel of available items in the SQ ring. If this option is set, the
2353 act of submitting IO will be done by a polling thread in the kernel.
2354 This frees up cycles for fio, at the cost of using more CPU in the
2355 system. As submission is just the time it takes to fill in the sqe
2356 entries and any syscall required to wake up the idle kernel thread,
2357 fio will not report submission latencies.
2359 .. option:: sqthread_poll_cpu=int : [io_uring] [io_uring_cmd]
2361 When :option:`sqthread_poll` is set, this option provides a way to
2362 define which CPU should be used for the polling thread.
2364 .. option:: cmd_type=str : [io_uring_cmd]
2366 Specifies the type of uring passthrough command to be used. Supported
2367 value is nvme. Default is nvme.
2371 [io_uring] [io_uring_cmd] [xnvme]
2373 If this option is set, fio will attempt to use polled IO completions.
2374 Normal IO completions generate interrupts to signal the completion of
2375 IO, polled completions do not. Hence they are require active reaping
2376 by the application. The benefits are more efficient IO for high IOPS
2377 scenarios, and lower latencies for low queue depth IO.
2381 Use poll queues. This is incompatible with
2382 :option:`libblkio_wait_mode=eventfd <libblkio_wait_mode>` and
2383 :option:`libblkio_force_enable_completion_eventfd`.
2387 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
2392 If this option is set, fio will attempt to use polled IO completions.
2393 This will have a similar effect as (io_uring)hipri. Only SCSI READ and
2394 WRITE commands will have the SGV4_FLAG_HIPRI set (not UNMAP (trim) nor
2395 VERIFY). Older versions of the Linux sg driver that do not support
2396 hipri will simply ignore this flag and do normal IO. The Linux SCSI
2397 Low Level Driver (LLD) that "owns" the device also needs to support
2398 hipri (also known as iopoll and mq_poll). The MegaRAID driver is an
2399 example of a SCSI LLD. Default: clear (0) which does normal
2400 (interrupted based) IO.
2402 .. option:: userspace_reap : [libaio]
2404 Normally, with the libaio engine in use, fio will use the
2405 :manpage:`io_getevents(2)` system call to reap newly returned events. With
2406 this flag turned on, the AIO ring will be read directly from user-space to
2407 reap events. The reaping mode is only enabled when polling for a minimum of
2408 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
2410 .. option:: hipri_percentage : [pvsync2]
2412 When hipri is set this determines the probability of a pvsync2 I/O being high
2413 priority. The default is 100%.
2415 .. option:: nowait=bool : [pvsync2] [libaio] [io_uring] [io_uring_cmd]
2417 By default if a request cannot be executed immediately (e.g. resource starvation,
2418 waiting on locks) it is queued and the initiating process will be blocked until
2419 the required resource becomes free.
2421 This option sets the RWF_NOWAIT flag (supported from the 4.14 Linux kernel) and
2422 the call will return instantly with EAGAIN or a partial result rather than waiting.
2424 It is useful to also use ignore_error=EAGAIN when using this option.
2426 Note: glibc 2.27, 2.28 have a bug in syscall wrappers preadv2, pwritev2.
2427 They return EOPNOTSUP instead of EAGAIN.
2429 For cached I/O, using this option usually means a request operates only with
2430 cached data. Currently the RWF_NOWAIT flag does not supported for cached write.
2432 For direct I/O, requests will only succeed if cache invalidation isn't required,
2433 file blocks are fully allocated and the disk request could be issued immediately.
2435 .. option:: cpuload=int : [cpuio]
2437 Attempt to use the specified percentage of CPU cycles. This is a mandatory
2438 option when using cpuio I/O engine.
2440 .. option:: cpuchunks=int : [cpuio]
2442 Split the load into cycles of the given time. In microseconds.
2444 .. option:: cpumode=str : [cpuio]
2446 Specify how to stress the CPU. It can take these two values:
2449 This is the default where the CPU executes noop instructions.
2451 Replace the default noop instructions loop with a qsort algorithm to
2452 consume more energy.
2454 .. option:: exit_on_io_done=bool : [cpuio]
2456 Detect when I/O threads are done, then exit.
2458 .. option:: namenode=str : [libhdfs]
2460 The hostname or IP address of a HDFS cluster namenode to contact.
2462 .. option:: port=int
2466 The listening port of the HFDS cluster namenode.
2470 The TCP or UDP port to bind to or connect to. If this is used with
2471 :option:`numjobs` to spawn multiple instances of the same job type, then
2472 this will be the starting port number since fio will use a range of
2477 The port to use for RDMA-CM communication. This should be the same value
2478 on the client and the server side.
2480 .. option:: hostname=str : [netsplice] [net] [rdma]
2482 The hostname or IP address to use for TCP, UDP or RDMA-CM based I/O. If the job
2483 is a TCP listener or UDP reader, the hostname is not used and must be omitted
2484 unless it is a valid UDP multicast address.
2486 .. option:: serverip=str : [librpma_*]
2488 The IP address to be used for RDMA-CM based I/O.
2490 .. option:: direct_write_to_pmem=bool : [librpma_*]
2492 Set to 1 only when Direct Write to PMem from the remote host is possible.
2493 Otherwise, set to 0.
2495 .. option:: busy_wait_polling=bool : [librpma_*_server]
2497 Set to 0 to wait for completion instead of busy-wait polling completion.
2500 .. option:: interface=str : [netsplice] [net]
2502 The IP address of the network interface used to send or receive UDP
2505 .. option:: ttl=int : [netsplice] [net]
2507 Time-to-live value for outgoing UDP multicast packets. Default: 1.
2509 .. option:: nodelay=bool : [netsplice] [net]
2511 Set TCP_NODELAY on TCP connections.
2513 .. option:: protocol=str, proto=str : [netsplice] [net]
2515 The network protocol to use. Accepted values are:
2518 Transmission control protocol.
2520 Transmission control protocol V6.
2522 User datagram protocol.
2524 User datagram protocol V6.
2528 When the protocol is TCP or UDP, the port must also be given, as well as the
2529 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
2530 normal :option:`filename` option should be used and the port is invalid.
2532 .. option:: listen : [netsplice] [net]
2534 For TCP network connections, tell fio to listen for incoming connections
2535 rather than initiating an outgoing connection. The :option:`hostname` must
2536 be omitted if this option is used.
2538 .. option:: pingpong : [netsplice] [net]
2540 Normally a network writer will just continue writing data, and a network
2541 reader will just consume packages. If ``pingpong=1`` is set, a writer will
2542 send its normal payload to the reader, then wait for the reader to send the
2543 same payload back. This allows fio to measure network latencies. The
2544 submission and completion latencies then measure local time spent sending or
2545 receiving, and the completion latency measures how long it took for the
2546 other end to receive and send back. For UDP multicast traffic
2547 ``pingpong=1`` should only be set for a single reader when multiple readers
2548 are listening to the same address.
2550 .. option:: window_size : [netsplice] [net]
2552 Set the desired socket buffer size for the connection.
2554 .. option:: mss : [netsplice] [net]
2556 Set the TCP maximum segment size (TCP_MAXSEG).
2558 .. option:: donorname=str : [e4defrag]
2560 File will be used as a block donor (swap extents between files).
2562 .. option:: inplace=int : [e4defrag]
2564 Configure donor file blocks allocation strategy:
2567 Default. Preallocate donor's file on init.
2569 Allocate space immediately inside defragment event, and free right
2572 .. option:: clustername=str : [rbd,rados]
2574 Specifies the name of the Ceph cluster.
2576 .. option:: rbdname=str : [rbd]
2578 Specifies the name of the RBD.
2580 .. option:: clientname=str : [rbd,rados]
2582 Specifies the username (without the 'client.' prefix) used to access the
2583 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
2584 the full *type.id* string. If no type. prefix is given, fio will add
2585 'client.' by default.
2587 .. option:: conf=str : [rados]
2589 Specifies the configuration path of ceph cluster, so conf file does not
2590 have to be /etc/ceph/ceph.conf.
2592 .. option:: busy_poll=bool : [rbd,rados]
2594 Poll store instead of waiting for completion. Usually this provides better
2595 throughput at cost of higher(up to 100%) CPU utilization.
2597 .. option:: touch_objects=bool : [rados]
2599 During initialization, touch (create if do not exist) all objects (files).
2600 Touching all objects affects ceph caches and likely impacts test results.
2603 .. option:: pool=str :
2607 Specifies the name of the Ceph pool containing RBD or RADOS data.
2611 Specify the label or UUID of the DAOS pool to connect to.
2613 .. option:: cont=str : [dfs]
2615 Specify the label or UUID of the DAOS container to open.
2617 .. option:: chunk_size=int
2621 Specify a different chunk size (in bytes) for the dfs file.
2622 Use DAOS container's chunk size by default.
2626 The size of the chunk to use for each file.
2628 .. option:: object_class=str : [dfs]
2630 Specify a different object class for the dfs file.
2631 Use DAOS container's object class by default.
2633 .. option:: skip_bad=bool : [mtd]
2635 Skip operations against known bad blocks.
2637 .. option:: hdfsdirectory : [libhdfs]
2639 libhdfs will create chunk in this HDFS directory.
2641 .. option:: verb=str : [rdma]
2643 The RDMA verb to use on this side of the RDMA ioengine connection. Valid
2644 values are write, read, send and recv. These correspond to the equivalent
2645 RDMA verbs (e.g. write = rdma_write etc.). Note that this only needs to be
2646 specified on the client side of the connection. See the examples folder.
2648 .. option:: bindname=str : [rdma]
2650 The name to use to bind the local RDMA-CM connection to a local RDMA device.
2651 This could be a hostname or an IPv4 or IPv6 address. On the server side this
2652 will be passed into the rdma_bind_addr() function and on the client site it
2653 will be used in the rdma_resolve_add() function. This can be useful when
2654 multiple paths exist between the client and the server or in certain loopback
2657 .. option:: stat_type=str : [filestat]
2659 Specify stat system call type to measure lookup/getattr performance.
2660 Default is **stat** for :manpage:`stat(2)`.
2662 .. option:: readfua=bool : [sg]
2664 With readfua option set to 1, read operations include
2665 the force unit access (fua) flag. Default is 0.
2667 .. option:: writefua=bool : [sg]
2669 With writefua option set to 1, write operations include
2670 the force unit access (fua) flag. Default is 0.
2672 .. option:: sg_write_mode=str : [sg]
2674 Specify the type of write commands to issue. This option can take three values:
2677 This is the default where write opcodes are issued as usual.
2678 **write_and_verify**
2679 Issue WRITE AND VERIFY commands. The BYTCHK bit is set to 0. This
2680 directs the device to carry out a medium verification with no data
2681 comparison. The writefua option is ignored with this selection.
2683 This option is deprecated. Use write_and_verify instead.
2685 Issue WRITE SAME commands. This transfers a single block to the device
2686 and writes this same block of data to a contiguous sequence of LBAs
2687 beginning at the specified offset. fio's block size parameter specifies
2688 the amount of data written with each command. However, the amount of data
2689 actually transferred to the device is equal to the device's block
2690 (sector) size. For a device with 512 byte sectors, blocksize=8k will
2691 write 16 sectors with each command. fio will still generate 8k of data
2692 for each command but only the first 512 bytes will be used and
2693 transferred to the device. The writefua option is ignored with this
2696 This option is deprecated. Use write_same instead.
2698 Issue WRITE SAME(16) commands as above but with the No Data Output
2699 Buffer (NDOB) bit set. No data will be transferred to the device with
2700 this bit set. Data written will be a pre-determined pattern such as
2703 Issue WRITE STREAM(16) commands. Use the **stream_id** option to specify
2704 the stream identifier.
2705 **verify_bytchk_00**
2706 Issue VERIFY commands with BYTCHK set to 00. This directs the
2707 device to carry out a medium verification with no data comparison.
2708 **verify_bytchk_01**
2709 Issue VERIFY commands with BYTCHK set to 01. This directs the device to
2710 compare the data on the device with the data transferred to the device.
2711 **verify_bytchk_11**
2712 Issue VERIFY commands with BYTCHK set to 11. This transfers a
2713 single block to the device and compares the contents of this block with the
2714 data on the device beginning at the specified offset. fio's block size
2715 parameter specifies the total amount of data compared with this command.
2716 However, only one block (sector) worth of data is transferred to the device.
2717 This is similar to the WRITE SAME command except that data is compared instead
2720 .. option:: stream_id=int : [sg]
2722 Set the stream identifier for WRITE STREAM commands. If this is set to 0 (which is not
2723 a valid stream identifier) fio will open a stream and then close it when done. Default
2726 .. option:: http_host=str : [http]
2728 Hostname to connect to. For S3, this could be the bucket hostname.
2729 Default is **localhost**
2731 .. option:: http_user=str : [http]
2733 Username for HTTP authentication.
2735 .. option:: http_pass=str : [http]
2737 Password for HTTP authentication.
2739 .. option:: https=str : [http]
2741 Enable HTTPS instead of http. *on* enables HTTPS; *insecure*
2742 will enable HTTPS, but disable SSL peer verification (use with
2743 caution!). Default is **off**
2745 .. option:: http_mode=str : [http]
2747 Which HTTP access mode to use: *webdav*, *swift*, or *s3*.
2748 Default is **webdav**
2750 .. option:: http_s3_region=str : [http]
2752 The S3 region/zone string.
2753 Default is **us-east-1**
2755 .. option:: http_s3_key=str : [http]
2759 .. option:: http_s3_keyid=str : [http]
2761 The S3 key/access id.
2763 .. option:: http_s3_sse_customer_key=str : [http]
2765 The encryption customer key in SSE server side.
2767 .. option:: http_s3_sse_customer_algorithm=str : [http]
2769 The encryption customer algorithm in SSE server side.
2770 Default is **AES256**
2772 .. option:: http_s3_storage_class=str : [http]
2774 Which storage class to access. User-customizable settings.
2775 Default is **STANDARD**
2777 .. option:: http_swift_auth_token=str : [http]
2779 The Swift auth token. See the example configuration file on how
2782 .. option:: http_verbose=int : [http]
2784 Enable verbose requests from libcurl. Useful for debugging. 1
2785 turns on verbose logging from libcurl, 2 additionally enables
2786 HTTP IO tracing. Default is **0**
2788 .. option:: uri=str : [nbd]
2790 Specify the NBD URI of the server to test. The string
2791 is a standard NBD URI
2792 (see https://github.com/NetworkBlockDevice/nbd/tree/master/doc).
2793 Example URIs: nbd://localhost:10809
2794 nbd+unix:///?socket=/tmp/socket
2795 nbds://tlshost/exportname
2797 .. option:: gpu_dev_ids=str : [libcufile]
2799 Specify the GPU IDs to use with CUDA. This is a colon-separated list of
2800 int. GPUs are assigned to workers roundrobin. Default is 0.
2802 .. option:: cuda_io=str : [libcufile]
2804 Specify the type of I/O to use with CUDA. Default is **cufile**.
2807 Use libcufile and nvidia-fs. This option performs I/O directly
2808 between a GPUDirect Storage filesystem and GPU buffers,
2809 avoiding use of a bounce buffer. If :option:`verify` is set,
2810 cudaMemcpy is used to copy verificaton data between RAM and GPU.
2811 Verification data is copied from RAM to GPU before a write
2812 and from GPU to RAM after a read. :option:`direct` must be 1.
2814 Use POSIX to perform I/O with a RAM buffer, and use cudaMemcpy
2815 to transfer data between RAM and the GPUs. Data is copied from
2816 GPU to RAM before a write and copied from RAM to GPU after a
2817 read. :option:`verify` does not affect use of cudaMemcpy.
2819 .. option:: nfs_url=str : [nfs]
2821 URL in libnfs format, eg nfs://<server|ipv4|ipv6>/path[?arg=val[&arg=val]*]
2822 Refer to the libnfs README for more details.
2824 .. option:: program=str : [exec]
2826 Specify the program to execute.
2828 .. option:: arguments=str : [exec]
2830 Specify arguments to pass to program.
2831 Some special variables can be expanded to pass fio's job details to the program.
2834 Replaced by the duration of the job in seconds.
2836 Replaced by the name of the job.
2838 .. option:: grace_time=int : [exec]
2840 Specify the time between the SIGTERM and SIGKILL signals. Default is 1 second.
2842 .. option:: std_redirect=bool : [exec]
2844 If set, stdout and stderr streams are redirected to files named from the job name. Default is true.
2846 .. option:: xnvme_async=str : [xnvme]
2848 Select the xnvme async command interface. This can take these values.
2851 This is default and use to emulate asynchronous I/O by using a
2852 single thread to create a queue pair on top of a synchronous
2853 I/O interface using the NVMe driver IOCTL.
2855 Emulate an asynchronous I/O interface with a pool of userspace
2856 threads on top of a synchronous I/O interface using the NVMe
2857 driver IOCTL. By default four threads are used.
2859 Linux native asynchronous I/O interface which supports both
2860 direct and buffered I/O.
2862 Fast Linux native asynchronous I/O interface for NVMe pass
2863 through commands. This only works with NVMe character device
2866 Use Linux aio for Asynchronous I/O.
2868 Use the posix asynchronous I/O interface to perform one or
2869 more I/O operations asynchronously.
2871 Use the user-space VFIO-based backend, implemented using
2872 libvfn instead of SPDK.
2874 Do not transfer any data; just pretend to. This is mainly used
2875 for introspective performance evaluation.
2877 .. option:: xnvme_sync=str : [xnvme]
2879 Select the xnvme synchronous command interface. This can take these values.
2882 This is default and uses Linux NVMe Driver ioctl() for
2885 This supports regular as well as vectored pread() and pwrite()
2888 This is the same as psync except that it also supports zone
2889 management commands using Linux block layer IOCTLs.
2891 .. option:: xnvme_admin=str : [xnvme]
2893 Select the xnvme admin command interface. This can take these values.
2896 This is default and uses linux NVMe Driver ioctl() for admin
2899 Use Linux Block Layer ioctl() and sysfs for admin commands.
2901 .. option:: xnvme_dev_nsid=int : [xnvme]
2903 xnvme namespace identifier for userspace NVMe driver, SPDK or vfio.
2905 .. option:: xnvme_dev_subnqn=str : [xnvme]
2907 Sets the subsystem NQN for fabrics. This is for xNVMe to utilize a
2908 fabrics target with multiple systems.
2910 .. option:: xnvme_mem=str : [xnvme]
2912 Select the xnvme memory backend. This can take these values.
2915 This is the default posix memory backend for linux NVMe driver.
2917 Use hugepages, instead of existing posix memory backend. The
2918 memory backend uses hugetlbfs. This require users to allocate
2919 hugepages, mount hugetlbfs and set an enviornment variable for
2922 Uses SPDK's memory allocator.
2924 Uses libvfn's memory allocator. This also specifies the use
2925 of libvfn backend instead of SPDK.
2927 .. option:: xnvme_iovec=int : [xnvme]
2929 If this option is set. xnvme will use vectored read/write commands.
2931 .. option:: libblkio_driver=str : [libblkio]
2933 The libblkio *driver* to use. Different drivers access devices through
2934 different underlying interfaces. Available drivers depend on the
2935 libblkio version in use and are listed at
2936 https://libblkio.gitlab.io/libblkio/blkio.html#drivers
2938 .. option:: libblkio_path=str : [libblkio]
2940 Sets the value of the driver-specific "path" property before connecting
2941 the libblkio instance, which identifies the target device or file on
2942 which to perform I/O. Its exact semantics are driver-dependent and not
2943 all drivers may support it; see
2944 https://libblkio.gitlab.io/libblkio/blkio.html#drivers
2946 .. option:: libblkio_pre_connect_props=str : [libblkio]
2948 A colon-separated list of additional libblkio properties to be set after
2949 creating but before connecting the libblkio instance. Each property must
2950 have the format ``<name>=<value>``. Colons can be escaped as ``\:``.
2951 These are set after the engine sets any other properties, so those can
2952 be overriden. Available properties depend on the libblkio version in use
2954 https://libblkio.gitlab.io/libblkio/blkio.html#properties
2956 .. option:: libblkio_num_entries=int : [libblkio]
2958 Sets the value of the driver-specific "num-entries" property before
2959 starting the libblkio instance. Its exact semantics are driver-dependent
2960 and not all drivers may support it; see
2961 https://libblkio.gitlab.io/libblkio/blkio.html#drivers
2963 .. option:: libblkio_queue_size=int : [libblkio]
2965 Sets the value of the driver-specific "queue-size" property before
2966 starting the libblkio instance. Its exact semantics are driver-dependent
2967 and not all drivers may support it; see
2968 https://libblkio.gitlab.io/libblkio/blkio.html#drivers
2970 .. option:: libblkio_pre_start_props=str : [libblkio]
2972 A colon-separated list of additional libblkio properties to be set after
2973 connecting but before starting the libblkio instance. Each property must
2974 have the format ``<name>=<value>``. Colons can be escaped as ``\:``.
2975 These are set after the engine sets any other properties, so those can
2976 be overriden. Available properties depend on the libblkio version in use
2978 https://libblkio.gitlab.io/libblkio/blkio.html#properties
2980 .. option:: libblkio_vectored : [libblkio]
2982 Submit vectored read and write requests.
2984 .. option:: libblkio_write_zeroes_on_trim : [libblkio]
2986 Submit trims as "write zeroes" requests instead of discard requests.
2988 .. option:: libblkio_wait_mode=str : [libblkio]
2990 How to wait for completions:
2993 Use a blocking call to ``blkioq_do_io()``.
2995 Use a blocking call to ``read()`` on the completion eventfd.
2997 Use a busy loop with a non-blocking call to ``blkioq_do_io()``.
2999 .. option:: libblkio_force_enable_completion_eventfd : [libblkio]
3001 Enable the queue's completion eventfd even when unused. This may impact
3002 performance. The default is to enable it only if
3003 :option:`libblkio_wait_mode=eventfd <libblkio_wait_mode>`.
3008 .. option:: iodepth=int
3010 Number of I/O units to keep in flight against the file. Note that
3011 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
3012 for small degrees when :option:`verify_async` is in use). Even async
3013 engines may impose OS restrictions causing the desired depth not to be
3014 achieved. This may happen on Linux when using libaio and not setting
3015 :option:`direct`\=1, since buffered I/O is not async on that OS. Keep an
3016 eye on the I/O depth distribution in the fio output to verify that the
3017 achieved depth is as expected. Default: 1.
3019 .. option:: iodepth_batch_submit=int, iodepth_batch=int
3021 This defines how many pieces of I/O to submit at once. It defaults to 1
3022 which means that we submit each I/O as soon as it is available, but can be
3023 raised to submit bigger batches of I/O at the time. If it is set to 0 the
3024 :option:`iodepth` value will be used.
3026 .. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
3028 This defines how many pieces of I/O to retrieve at once. It defaults to 1
3029 which means that we'll ask for a minimum of 1 I/O in the retrieval process
3030 from the kernel. The I/O retrieval will go on until we hit the limit set by
3031 :option:`iodepth_low`. If this variable is set to 0, then fio will always
3032 check for completed events before queuing more I/O. This helps reduce I/O
3033 latency, at the cost of more retrieval system calls.
3035 .. option:: iodepth_batch_complete_max=int
3037 This defines maximum pieces of I/O to retrieve at once. This variable should
3038 be used along with :option:`iodepth_batch_complete_min`\=int variable,
3039 specifying the range of min and max amount of I/O which should be
3040 retrieved. By default it is equal to the :option:`iodepth_batch_complete_min`
3045 iodepth_batch_complete_min=1
3046 iodepth_batch_complete_max=<iodepth>
3048 which means that we will retrieve at least 1 I/O and up to the whole
3049 submitted queue depth. If none of I/O has been completed yet, we will wait.
3053 iodepth_batch_complete_min=0
3054 iodepth_batch_complete_max=<iodepth>
3056 which means that we can retrieve up to the whole submitted queue depth, but
3057 if none of I/O has been completed yet, we will NOT wait and immediately exit
3058 the system call. In this example we simply do polling.
3060 .. option:: iodepth_low=int
3062 The low water mark indicating when to start filling the queue
3063 again. Defaults to the same as :option:`iodepth`, meaning that fio will
3064 attempt to keep the queue full at all times. If :option:`iodepth` is set to
3065 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
3066 16 requests, it will let the depth drain down to 4 before starting to fill
3069 .. option:: serialize_overlap=bool
3071 Serialize in-flight I/Os that might otherwise cause or suffer from data races.
3072 When two or more I/Os are submitted simultaneously, there is no guarantee that
3073 the I/Os will be processed or completed in the submitted order. Further, if
3074 two or more of those I/Os are writes, any overlapping region between them can
3075 become indeterminate/undefined on certain storage. These issues can cause
3076 verification to fail erratically when at least one of the racing I/Os is
3077 changing data and the overlapping region has a non-zero size. Setting
3078 ``serialize_overlap`` tells fio to avoid provoking this behavior by explicitly
3079 serializing in-flight I/Os that have a non-zero overlap. Note that setting
3080 this option can reduce both performance and the :option:`iodepth` achieved.
3082 This option only applies to I/Os issued for a single job except when it is
3083 enabled along with :option:`io_submit_mode`\=offload. In offload mode, fio
3084 will check for overlap among all I/Os submitted by offload jobs with :option:`serialize_overlap`
3089 .. option:: io_submit_mode=str
3091 This option controls how fio submits the I/O to the I/O engine. The default
3092 is `inline`, which means that the fio job threads submit and reap I/O
3093 directly. If set to `offload`, the job threads will offload I/O submission
3094 to a dedicated pool of I/O threads. This requires some coordination and thus
3095 has a bit of extra overhead, especially for lower queue depth I/O where it
3096 can increase latencies. The benefit is that fio can manage submission rates
3097 independently of the device completion rates. This avoids skewed latency
3098 reporting if I/O gets backed up on the device side (the coordinated omission
3099 problem). Note that this option cannot reliably be used with async IO
3106 .. option:: thinktime=time
3108 Stall the job for the specified period of time after an I/O has completed before issuing the
3109 next. May be used to simulate processing being done by an application.
3110 When the unit is omitted, the value is interpreted in microseconds. See
3111 :option:`thinktime_blocks`, :option:`thinktime_iotime` and :option:`thinktime_spin`.
3113 .. option:: thinktime_spin=time
3115 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
3116 something with the data received, before falling back to sleeping for the
3117 rest of the period specified by :option:`thinktime`. When the unit is
3118 omitted, the value is interpreted in microseconds.
3120 .. option:: thinktime_blocks=int
3122 Only valid if :option:`thinktime` is set - control how many blocks to issue,
3123 before waiting :option:`thinktime` usecs. If not set, defaults to 1 which will make
3124 fio wait :option:`thinktime` usecs after every block. This effectively makes any
3125 queue depth setting redundant, since no more than 1 I/O will be queued
3126 before we have to complete it and do our :option:`thinktime`. In other words, this
3127 setting effectively caps the queue depth if the latter is larger.
3129 .. option:: thinktime_blocks_type=str
3131 Only valid if :option:`thinktime` is set - control how :option:`thinktime_blocks`
3132 triggers. The default is `complete`, which triggers thinktime when fio completes
3133 :option:`thinktime_blocks` blocks. If this is set to `issue`, then the trigger happens
3136 .. option:: thinktime_iotime=time
3138 Only valid if :option:`thinktime` is set - control :option:`thinktime`
3139 interval by time. The :option:`thinktime` stall is repeated after IOs
3140 are executed for :option:`thinktime_iotime`. For example,
3141 ``--thinktime_iotime=9s --thinktime=1s`` repeat 10-second cycle with IOs
3142 for 9 seconds and stall for 1 second. When the unit is omitted,
3143 :option:`thinktime_iotime` is interpreted as a number of seconds. If
3144 this option is used together with :option:`thinktime_blocks`, the
3145 :option:`thinktime` stall is repeated after :option:`thinktime_iotime`
3146 or after :option:`thinktime_blocks` IOs, whichever happens first.
3148 .. option:: rate=int[,int][,int]
3150 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
3151 suffix rules apply. Comma-separated values may be specified for reads,
3152 writes, and trims as described in :option:`blocksize`.
3154 For example, using `rate=1m,500k` would limit reads to 1MiB/sec and writes to
3155 500KiB/sec. Capping only reads or writes can be done with `rate=,500k` or
3156 `rate=500k,` where the former will only limit writes (to 500KiB/sec) and the
3157 latter will only limit reads.
3159 .. option:: rate_min=int[,int][,int]
3161 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
3162 to meet this requirement will cause the job to exit. Comma-separated values
3163 may be specified for reads, writes, and trims as described in
3164 :option:`blocksize`.
3166 .. option:: rate_iops=int[,int][,int]
3168 Cap the bandwidth to this number of IOPS. Basically the same as
3169 :option:`rate`, just specified independently of bandwidth. If the job is
3170 given a block size range instead of a fixed value, the smallest block size
3171 is used as the metric. Comma-separated values may be specified for reads,
3172 writes, and trims as described in :option:`blocksize`.
3174 .. option:: rate_iops_min=int[,int][,int]
3176 If fio doesn't meet this rate of I/O, it will cause the job to exit.
3177 Comma-separated values may be specified for reads, writes, and trims as
3178 described in :option:`blocksize`.
3180 .. option:: rate_process=str
3182 This option controls how fio manages rated I/O submissions. The default is
3183 `linear`, which submits I/O in a linear fashion with fixed delays between
3184 I/Os that gets adjusted based on I/O completion rates. If this is set to
3185 `poisson`, fio will submit I/O based on a more real world random request
3186 flow, known as the Poisson process
3187 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
3188 10^6 / IOPS for the given workload.
3190 .. option:: rate_ignore_thinktime=bool
3192 By default, fio will attempt to catch up to the specified rate setting,
3193 if any kind of thinktime setting was used. If this option is set, then
3194 fio will ignore the thinktime and continue doing IO at the specified
3195 rate, instead of entering a catch-up mode after thinktime is done.
3201 .. option:: latency_target=time
3203 If set, fio will attempt to find the max performance point that the given
3204 workload will run at while maintaining a latency below this target. When
3205 the unit is omitted, the value is interpreted in microseconds. See
3206 :option:`latency_window` and :option:`latency_percentile`.
3208 .. option:: latency_window=time
3210 Used with :option:`latency_target` to specify the sample window that the job
3211 is run at varying queue depths to test the performance. When the unit is
3212 omitted, the value is interpreted in microseconds.
3214 .. option:: latency_percentile=float
3216 The percentage of I/Os that must fall within the criteria specified by
3217 :option:`latency_target` and :option:`latency_window`. If not set, this
3218 defaults to 100.0, meaning that all I/Os must be equal or below to the value
3219 set by :option:`latency_target`.
3221 .. option:: latency_run=bool
3223 Used with :option:`latency_target`. If false (default), fio will find
3224 the highest queue depth that meets :option:`latency_target` and exit. If
3225 true, fio will continue running and try to meet :option:`latency_target`
3226 by adjusting queue depth.
3228 .. option:: max_latency=time[,time][,time]
3230 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
3231 maximum latency. When the unit is omitted, the value is interpreted in
3232 microseconds. Comma-separated values may be specified for reads, writes,
3233 and trims as described in :option:`blocksize`.
3235 .. option:: rate_cycle=int
3237 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
3238 of milliseconds. Defaults to 1000.
3244 .. option:: write_iolog=str
3246 Write the issued I/O patterns to the specified file. See
3247 :option:`read_iolog`. Specify a separate file for each job, otherwise the
3248 iologs will be interspersed and the file may be corrupt. This file will
3249 be opened in append mode.
3251 .. option:: read_iolog=str
3253 Open an iolog with the specified filename and replay the I/O patterns it
3254 contains. This can be used to store a workload and replay it sometime
3255 later. The iolog given may also be a blktrace binary file, which allows fio
3256 to replay a workload captured by :command:`blktrace`. See
3257 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
3258 replay, the file needs to be turned into a blkparse binary data file first
3259 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
3260 You can specify a number of files by separating the names with a ':'
3261 character. See the :option:`filename` option for information on how to
3262 escape ':' characters within the file names. These files will
3263 be sequentially assigned to job clones created by :option:`numjobs`.
3264 '-' is a reserved name, meaning read from stdin, notably if
3265 :option:`filename` is set to '-' which means stdin as well, then
3266 this flag can't be set to '-'.
3268 .. option:: read_iolog_chunked=bool
3270 Determines how iolog is read. If false(default) entire :option:`read_iolog`
3271 will be read at once. If selected true, input from iolog will be read
3272 gradually. Useful when iolog is very large, or it is generated.
3274 .. option:: merge_blktrace_file=str
3276 When specified, rather than replaying the logs passed to :option:`read_iolog`,
3277 the logs go through a merge phase which aggregates them into a single
3278 blktrace. The resulting file is then passed on as the :option:`read_iolog`
3279 parameter. The intention here is to make the order of events consistent.
3280 This limits the influence of the scheduler compared to replaying multiple
3281 blktraces via concurrent jobs.
3283 .. option:: merge_blktrace_scalars=float_list
3285 This is a percentage based option that is index paired with the list of
3286 files passed to :option:`read_iolog`. When merging is performed, scale
3287 the time of each event by the corresponding amount. For example,
3288 ``--merge_blktrace_scalars="50:100"`` runs the first trace in halftime
3289 and the second trace in realtime. This knob is separately tunable from
3290 :option:`replay_time_scale` which scales the trace during runtime and
3291 does not change the output of the merge unlike this option.
3293 .. option:: merge_blktrace_iters=float_list
3295 This is a whole number option that is index paired with the list of files
3296 passed to :option:`read_iolog`. When merging is performed, run each trace
3297 for the specified number of iterations. For example,
3298 ``--merge_blktrace_iters="2:1"`` runs the first trace for two iterations
3299 and the second trace for one iteration.
3301 .. option:: replay_no_stall=bool
3303 When replaying I/O with :option:`read_iolog` the default behavior is to
3304 attempt to respect the timestamps within the log and replay them with the
3305 appropriate delay between IOPS. By setting this variable fio will not
3306 respect the timestamps and attempt to replay them as fast as possible while
3307 still respecting ordering. The result is the same I/O pattern to a given
3308 device, but different timings.
3310 .. option:: replay_time_scale=int
3312 When replaying I/O with :option:`read_iolog`, fio will honor the
3313 original timing in the trace. With this option, it's possible to scale
3314 the time. It's a percentage option, if set to 50 it means run at 50%
3315 the original IO rate in the trace. If set to 200, run at twice the
3316 original IO rate. Defaults to 100.
3318 .. option:: replay_redirect=str
3320 While replaying I/O patterns using :option:`read_iolog` the default behavior
3321 is to replay the IOPS onto the major/minor device that each IOP was recorded
3322 from. This is sometimes undesirable because on a different machine those
3323 major/minor numbers can map to a different device. Changing hardware on the
3324 same system can also result in a different major/minor mapping.
3325 ``replay_redirect`` causes all I/Os to be replayed onto the single specified
3326 device regardless of the device it was recorded
3327 from. i.e. :option:`replay_redirect`\= :file:`/dev/sdc` would cause all I/O
3328 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
3329 multiple devices will be replayed onto a single device, if the trace
3330 contains multiple devices. If you want multiple devices to be replayed
3331 concurrently to multiple redirected devices you must blkparse your trace
3332 into separate traces and replay them with independent fio invocations.
3333 Unfortunately this also breaks the strict time ordering between multiple
3336 .. option:: replay_align=int
3338 Force alignment of the byte offsets in a trace to this value. The value
3339 must be a power of 2.
3341 .. option:: replay_scale=int
3343 Scale byte offsets down by this factor when replaying traces. Should most
3344 likely use :option:`replay_align` as well.
3346 .. option:: replay_skip=str
3348 Sometimes it's useful to skip certain IO types in a replay trace.
3349 This could be, for instance, eliminating the writes in the trace.
3350 Or not replaying the trims/discards, if you are redirecting to
3351 a device that doesn't support them. This option takes a comma
3352 separated list of read, write, trim, sync.
3355 Threads, processes and job synchronization
3356 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3360 Fio defaults to creating jobs by using fork, however if this option is
3361 given, fio will create jobs by using POSIX Threads' function
3362 :manpage:`pthread_create(3)` to create threads instead.
3364 .. option:: wait_for=str
3366 If set, the current job won't be started until all workers of the specified
3367 waitee job are done.
3369 ``wait_for`` operates on the job name basis, so there are a few
3370 limitations. First, the waitee must be defined prior to the waiter job
3371 (meaning no forward references). Second, if a job is being referenced as a
3372 waitee, it must have a unique name (no duplicate waitees).
3374 .. option:: nice=int
3376 Run the job with the given nice value. See man :manpage:`nice(2)`.
3378 On Windows, values less than -15 set the process class to "High"; -1 through
3379 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
3382 .. option:: prio=int
3384 Set the I/O priority value of this job. Linux limits us to a positive value
3385 between 0 and 7, with 0 being the highest. See man
3386 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
3387 systems since meaning of priority may differ. For per-command priority
3388 setting, see I/O engine specific :option:`cmdprio_percentage` and
3389 :option:`cmdprio` options.
3391 .. option:: prioclass=int
3393 Set the I/O priority class. See man :manpage:`ionice(1)`. For per-command
3394 priority setting, see I/O engine specific :option:`cmdprio_percentage`
3395 and :option:`cmdprio_class` options.
3397 .. option:: cpus_allowed=str
3399 Controls the same options as :option:`cpumask`, but accepts a textual
3400 specification of the permitted CPUs instead and CPUs are indexed from 0. So
3401 to use CPUs 0 and 5 you would specify ``cpus_allowed=0,5``. This option also
3402 allows a range of CPUs to be specified -- say you wanted a binding to CPUs
3403 0, 5, and 8 to 15, you would set ``cpus_allowed=0,5,8-15``.
3405 On Windows, when ``cpus_allowed`` is unset only CPUs from fio's current
3406 processor group will be used and affinity settings are inherited from the
3407 system. An fio build configured to target Windows 7 makes options that set
3408 CPUs processor group aware and values will set both the processor group
3409 and a CPU from within that group. For example, on a system where processor
3410 group 0 has 40 CPUs and processor group 1 has 32 CPUs, ``cpus_allowed``
3411 values between 0 and 39 will bind CPUs from processor group 0 and
3412 ``cpus_allowed`` values between 40 and 71 will bind CPUs from processor
3413 group 1. When using ``cpus_allowed_policy=shared`` all CPUs specified by a
3414 single ``cpus_allowed`` option must be from the same processor group. For
3415 Windows fio builds not built for Windows 7, CPUs will only be selected from
3416 (and be relative to) whatever processor group fio happens to be running in
3417 and CPUs from other processor groups cannot be used.
3419 .. option:: cpus_allowed_policy=str
3421 Set the policy of how fio distributes the CPUs specified by
3422 :option:`cpus_allowed` or :option:`cpumask`. Two policies are supported:
3425 All jobs will share the CPU set specified.
3427 Each job will get a unique CPU from the CPU set.
3429 **shared** is the default behavior, if the option isn't specified. If
3430 **split** is specified, then fio will assign one cpu per job. If not
3431 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
3434 .. option:: cpumask=int
3436 Set the CPU affinity of this job. The parameter given is a bit mask of
3437 allowed CPUs the job may run on. So if you want the allowed CPUs to be 1
3438 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
3439 :manpage:`sched_setaffinity(2)`. This may not work on all supported
3440 operating systems or kernel versions. This option doesn't work well for a
3441 higher CPU count than what you can store in an integer mask, so it can only
3442 control cpus 1-32. For boxes with larger CPU counts, use
3443 :option:`cpus_allowed`.
3445 .. option:: numa_cpu_nodes=str
3447 Set this job running on specified NUMA nodes' CPUs. The arguments allow
3448 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
3449 NUMA options support, fio must be built on a system with libnuma-dev(el)
3452 .. option:: numa_mem_policy=str
3454 Set this job's memory policy and corresponding NUMA nodes. Format of the
3459 ``mode`` is one of the following memory policies: ``default``, ``prefer``,
3460 ``bind``, ``interleave`` or ``local``. For ``default`` and ``local`` memory
3461 policies, no node needs to be specified. For ``prefer``, only one node is
3462 allowed. For ``bind`` and ``interleave`` the ``nodelist`` may be as
3463 follows: a comma delimited list of numbers, A-B ranges, or `all`.
3465 .. option:: cgroup=str
3467 Add job to this control group. If it doesn't exist, it will be created. The
3468 system must have a mounted cgroup blkio mount point for this to work. If
3469 your system doesn't have it mounted, you can do so with::
3471 # mount -t cgroup -o blkio none /cgroup
3473 .. option:: cgroup_weight=int
3475 Set the weight of the cgroup to this value. See the documentation that comes
3476 with the kernel, allowed values are in the range of 100..1000.
3478 .. option:: cgroup_nodelete=bool
3480 Normally fio will delete the cgroups it has created after the job
3481 completion. To override this behavior and to leave cgroups around after the
3482 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
3483 to inspect various cgroup files after job completion. Default: false.
3485 .. option:: flow_id=int
3487 The ID of the flow. If not specified, it defaults to being a global
3488 flow. See :option:`flow`.
3490 .. option:: flow=int
3492 Weight in token-based flow control. If this value is used, then fio
3493 regulates the activity between two or more jobs sharing the same
3494 flow_id. Fio attempts to keep each job activity proportional to other
3495 jobs' activities in the same flow_id group, with respect to requested
3496 weight per job. That is, if one job has `flow=3', another job has
3497 `flow=2' and another with `flow=1`, then there will be a roughly 3:2:1
3498 ratio in how much one runs vs the others.
3500 .. option:: flow_sleep=int
3502 The period of time, in microseconds, to wait after the flow counter
3503 has exceeded its proportion before retrying operations.
3505 .. option:: stonewall, wait_for_previous
3507 Wait for preceding jobs in the job file to exit, before starting this
3508 one. Can be used to insert serialization points in the job file. A stone
3509 wall also implies starting a new reporting group, see
3510 :option:`group_reporting`.
3514 By default, fio will continue running all other jobs when one job finishes.
3515 Sometimes this is not the desired action. Setting ``exitall`` will instead
3516 make fio terminate all jobs in the same group, as soon as one job of that
3519 .. option:: exit_what=str
3521 By default, fio will continue running all other jobs when one job finishes.
3522 Sometimes this is not the desired action. Setting ``exitall`` will
3523 instead make fio terminate all jobs in the same group. The option
3524 ``exit_what`` allows to control which jobs get terminated when ``exitall`` is
3525 enabled. The default is ``group`` and does not change the behaviour of
3526 ``exitall``. The setting ``all`` terminates all jobs. The setting ``stonewall``
3527 terminates all currently running jobs across all groups and continues execution
3528 with the next stonewalled group.
3530 .. option:: exec_prerun=str
3532 Before running this job, issue the command specified through
3533 :manpage:`system(3)`. Output is redirected in a file called
3534 :file:`jobname.prerun.txt`.
3536 .. option:: exec_postrun=str
3538 After the job completes, issue the command specified though
3539 :manpage:`system(3)`. Output is redirected in a file called
3540 :file:`jobname.postrun.txt`.
3544 Instead of running as the invoking user, set the user ID to this value
3545 before the thread/process does any work.
3549 Set group ID, see :option:`uid`.
3555 .. option:: verify_only
3557 Do not perform specified workload, only verify data still matches previous
3558 invocation of this workload. This option allows one to check data multiple
3559 times at a later date without overwriting it. This option makes sense only
3560 for workloads that write data, and does not support workloads with the
3561 :option:`time_based` option set.
3563 .. option:: do_verify=bool
3565 Run the verify phase after a write phase. Only valid if :option:`verify` is
3568 .. option:: verify=str
3570 If writing to a file, fio can verify the file contents after each iteration
3571 of the job. Each verification method also implies verification of special
3572 header, which is written to the beginning of each block. This header also
3573 includes meta information, like offset of the block, block number, timestamp
3574 when block was written, etc. :option:`verify` can be combined with
3575 :option:`verify_pattern` option. The allowed values are:
3578 Use an md5 sum of the data area and store it in the header of
3582 Use an experimental crc64 sum of the data area and store it in the
3583 header of each block.
3586 Use a crc32c sum of the data area and store it in the header of
3587 each block. This will automatically use hardware acceleration
3588 (e.g. SSE4.2 on an x86 or CRC crypto extensions on ARM64) but will
3589 fall back to software crc32c if none is found. Generally the
3590 fastest checksum fio supports when hardware accelerated.
3596 Use a crc32 sum of the data area and store it in the header of each
3600 Use a crc16 sum of the data area and store it in the header of each
3604 Use a crc7 sum of the data area and store it in the header of each
3608 Use xxhash as the checksum function. Generally the fastest software
3609 checksum that fio supports.
3612 Use sha512 as the checksum function.
3615 Use sha256 as the checksum function.
3618 Use optimized sha1 as the checksum function.
3621 Use optimized sha3-224 as the checksum function.
3624 Use optimized sha3-256 as the checksum function.
3627 Use optimized sha3-384 as the checksum function.
3630 Use optimized sha3-512 as the checksum function.
3633 This option is deprecated, since now meta information is included in
3634 generic verification header and meta verification happens by
3635 default. For detailed information see the description of the
3636 :option:`verify` setting. This option is kept because of
3637 compatibility's sake with old configurations. Do not use it.
3640 Verify a strict pattern. Normally fio includes a header with some
3641 basic information and checksumming, but if this option is set, only
3642 the specific pattern set with :option:`verify_pattern` is verified.
3645 Only pretend to verify. Useful for testing internals with
3646 :option:`ioengine`\=null, not for much else.
3648 This option can be used for repeated burn-in tests of a system to make sure
3649 that the written data is also correctly read back. If the data direction
3650 given is a read or random read, fio will assume that it should verify a
3651 previously written file. If the data direction includes any form of write,
3652 the verify will be of the newly written data.
3654 To avoid false verification errors, do not use the norandommap option when
3655 verifying data with async I/O engines and I/O depths > 1. Or use the
3656 norandommap and the lfsr random generator together to avoid writing to the
3657 same offset with multiple outstanding I/Os.
3659 .. option:: verify_offset=int
3661 Swap the verification header with data somewhere else in the block before
3662 writing. It is swapped back before verifying.
3664 .. option:: verify_interval=int
3666 Write the verification header at a finer granularity than the
3667 :option:`blocksize`. It will be written for chunks the size of
3668 ``verify_interval``. :option:`blocksize` should divide this evenly.
3670 .. option:: verify_pattern=str
3672 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
3673 filling with totally random bytes, but sometimes it's interesting to fill
3674 with a known pattern for I/O verification purposes. Depending on the width
3675 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time (it can
3676 be either a decimal or a hex number). The ``verify_pattern`` if larger than
3677 a 32-bit quantity has to be a hex number that starts with either "0x" or
3678 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
3679 format, which means that for each block offset will be written and then
3680 verified back, e.g.::
3684 Or use combination of everything::
3686 verify_pattern=0xff%o"abcd"-12
3688 .. option:: verify_fatal=bool
3690 Normally fio will keep checking the entire contents before quitting on a
3691 block verification failure. If this option is set, fio will exit the job on
3692 the first observed failure. Default: false.
3694 .. option:: verify_dump=bool
3696 If set, dump the contents of both the original data block and the data block
3697 we read off disk to files. This allows later analysis to inspect just what
3698 kind of data corruption occurred. Off by default.
3700 .. option:: verify_async=int
3702 Fio will normally verify I/O inline from the submitting thread. This option
3703 takes an integer describing how many async offload threads to create for I/O
3704 verification instead, causing fio to offload the duty of verifying I/O
3705 contents to one or more separate threads. If using this offload option, even
3706 sync I/O engines can benefit from using an :option:`iodepth` setting higher
3707 than 1, as it allows them to have I/O in flight while verifies are running.
3708 Defaults to 0 async threads, i.e. verification is not asynchronous.
3710 .. option:: verify_async_cpus=str
3712 Tell fio to set the given CPU affinity on the async I/O verification
3713 threads. See :option:`cpus_allowed` for the format used.
3715 .. option:: verify_backlog=int
3717 Fio will normally verify the written contents of a job that utilizes verify
3718 once that job has completed. In other words, everything is written then
3719 everything is read back and verified. You may want to verify continually
3720 instead for a variety of reasons. Fio stores the meta data associated with
3721 an I/O block in memory, so for large verify workloads, quite a bit of memory
3722 would be used up holding this meta data. If this option is enabled, fio will
3723 write only N blocks before verifying these blocks.
3725 .. option:: verify_backlog_batch=int
3727 Control how many blocks fio will verify if :option:`verify_backlog` is
3728 set. If not set, will default to the value of :option:`verify_backlog`
3729 (meaning the entire queue is read back and verified). If
3730 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
3731 blocks will be verified, if ``verify_backlog_batch`` is larger than
3732 :option:`verify_backlog`, some blocks will be verified more than once.
3734 .. option:: verify_state_save=bool
3736 When a job exits during the write phase of a verify workload, save its
3737 current state. This allows fio to replay up until that point, if the verify
3738 state is loaded for the verify read phase. The format of the filename is,
3741 <type>-<jobname>-<jobindex>-verify.state.
3743 <type> is "local" for a local run, "sock" for a client/server socket
3744 connection, and "ip" (192.168.0.1, for instance) for a networked
3745 client/server connection. Defaults to true.
3747 .. option:: verify_state_load=bool
3749 If a verify termination trigger was used, fio stores the current write state
3750 of each thread. This can be used at verification time so that fio knows how
3751 far it should verify. Without this information, fio will run a full
3752 verification pass, according to the settings in the job file used. Default
3755 .. option:: trim_percentage=int
3757 Number of verify blocks to discard/trim.
3759 .. option:: trim_verify_zero=bool
3761 Verify that trim/discarded blocks are returned as zeros.
3763 .. option:: trim_backlog=int
3765 Trim after this number of blocks are written.
3767 .. option:: trim_backlog_batch=int
3769 Trim this number of I/O blocks.
3771 .. option:: experimental_verify=bool
3773 Enable experimental verification. Standard verify records I/O metadata
3774 for later use during the verification phase. Experimental verify
3775 instead resets the file after the write phase and then replays I/Os for
3776 the verification phase.
3781 .. option:: steadystate=str:float, ss=str:float
3783 Define the criterion and limit for assessing steady state performance. The
3784 first parameter designates the criterion whereas the second parameter sets
3785 the threshold. When the criterion falls below the threshold for the
3786 specified duration, the job will stop. For example, `iops_slope:0.1%` will
3787 direct fio to terminate the job when the least squares regression slope
3788 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
3789 this will apply to all jobs in the group. Below is the list of available
3790 steady state assessment criteria. All assessments are carried out using only
3791 data from the rolling collection window. Threshold limits can be expressed
3792 as a fixed value or as a percentage of the mean in the collection window.
3794 When using this feature, most jobs should include the :option:`time_based`
3795 and :option:`runtime` options or the :option:`loops` option so that fio does not
3796 stop running after it has covered the full size of the specified file(s) or device(s).
3799 Collect IOPS data. Stop the job if all individual IOPS measurements
3800 are within the specified limit of the mean IOPS (e.g., ``iops:2``
3801 means that all individual IOPS values must be within 2 of the mean,
3802 whereas ``iops:0.2%`` means that all individual IOPS values must be
3803 within 0.2% of the mean IOPS to terminate the job).
3806 Collect IOPS data and calculate the least squares regression
3807 slope. Stop the job if the slope falls below the specified limit.
3810 Collect bandwidth data. Stop the job if all individual bandwidth
3811 measurements are within the specified limit of the mean bandwidth.
3814 Collect bandwidth data and calculate the least squares regression
3815 slope. Stop the job if the slope falls below the specified limit.
3817 .. option:: steadystate_duration=time, ss_dur=time
3819 A rolling window of this duration will be used to judge whether steady state
3820 has been reached. Data will be collected once per second. The default is 0
3821 which disables steady state detection. When the unit is omitted, the
3822 value is interpreted in seconds.
3824 .. option:: steadystate_ramp_time=time, ss_ramp=time
3826 Allow the job to run for the specified duration before beginning data
3827 collection for checking the steady state job termination criterion. The
3828 default is 0. When the unit is omitted, the value is interpreted in seconds.
3831 Measurements and reporting
3832 ~~~~~~~~~~~~~~~~~~~~~~~~~~
3834 .. option:: per_job_logs=bool
3836 If set, this generates bw/clat/iops log with per file private filenames. If
3837 not set, jobs with identical names will share the log filename. Default:
3840 .. option:: group_reporting
3842 It may sometimes be interesting to display statistics for groups of jobs as
3843 a whole instead of for each individual job. This is especially true if
3844 :option:`numjobs` is used; looking at individual thread/process output
3845 quickly becomes unwieldy. To see the final report per-group instead of
3846 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
3847 same reporting group, unless if separated by a :option:`stonewall`, or by
3848 using :option:`new_group`.
3850 .. option:: new_group
3852 Start a new reporting group. See: :option:`group_reporting`. If not given,
3853 all jobs in a file will be part of the same reporting group, unless
3854 separated by a :option:`stonewall`.
3856 .. option:: stats=bool
3858 By default, fio collects and shows final output results for all jobs
3859 that run. If this option is set to 0, then fio will ignore it in
3860 the final stat output.
3862 .. option:: write_bw_log=str
3864 If given, write a bandwidth log for this job. Can be used to store data of
3865 the bandwidth of the jobs in their lifetime.
3867 If no str argument is given, the default filename of
3868 :file:`jobname_type.x.log` is used. Even when the argument is given, fio
3869 will still append the type of log. So if one specifies::
3873 The actual log name will be :file:`foo_bw.x.log` where `x` is the index
3874 of the job (`1..N`, where `N` is the number of jobs). If
3875 :option:`per_job_logs` is false, then the filename will not include the
3878 The included :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
3879 text files into nice graphs. See `Log File Formats`_ for how data is
3880 structured within the file.
3882 .. option:: write_lat_log=str
3884 Same as :option:`write_bw_log`, except this option creates I/O
3885 submission (e.g., :file:`name_slat.x.log`), completion (e.g.,
3886 :file:`name_clat.x.log`), and total (e.g., :file:`name_lat.x.log`)
3887 latency files instead. See :option:`write_bw_log` for details about
3888 the filename format and `Log File Formats`_ for how data is structured
3891 .. option:: write_hist_log=str
3893 Same as :option:`write_bw_log` but writes an I/O completion latency
3894 histogram file (e.g., :file:`name_hist.x.log`) instead. Note that this
3895 file will be empty unless :option:`log_hist_msec` has also been set.
3896 See :option:`write_bw_log` for details about the filename format and
3897 `Log File Formats`_ for how data is structured within the file.
3899 .. option:: write_iops_log=str
3901 Same as :option:`write_bw_log`, but writes an IOPS file (e.g.
3902 :file:`name_iops.x.log`) instead. Because fio defaults to individual
3903 I/O logging, the value entry in the IOPS log will be 1 unless windowed
3904 logging (see :option:`log_avg_msec`) has been enabled. See
3905 :option:`write_bw_log` for details about the filename format and `Log
3906 File Formats`_ for how data is structured within the file.
3908 .. option:: log_entries=int
3910 By default, fio will log an entry in the iops, latency, or bw log for
3911 every I/O that completes. The initial number of I/O log entries is 1024.
3912 When the log entries are all used, new log entries are dynamically
3913 allocated. This dynamic log entry allocation may negatively impact
3914 time-related statistics such as I/O tail latencies (e.g. 99.9th percentile
3915 completion latency). This option allows specifying a larger initial
3916 number of log entries to avoid run-time allocations of new log entries,
3917 resulting in more precise time-related I/O statistics.
3918 Also see :option:`log_avg_msec`. Defaults to 1024.
3920 .. option:: log_avg_msec=int
3922 By default, fio will log an entry in the iops, latency, or bw log for every
3923 I/O that completes. When writing to the disk log, that can quickly grow to a
3924 very large size. Setting this option makes fio average the each log entry
3925 over the specified period of time, reducing the resolution of the log. See
3926 :option:`log_max_value` as well. Defaults to 0, logging all entries.
3927 Also see `Log File Formats`_.
3929 .. option:: log_hist_msec=int
3931 Same as :option:`log_avg_msec`, but logs entries for completion latency
3932 histograms. Computing latency percentiles from averages of intervals using
3933 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
3934 histogram entries over the specified period of time, reducing log sizes for
3935 high IOPS devices while retaining percentile accuracy. See
3936 :option:`log_hist_coarseness` and :option:`write_hist_log` as well.
3937 Defaults to 0, meaning histogram logging is disabled.
3939 .. option:: log_hist_coarseness=int
3941 Integer ranging from 0 to 6, defining the coarseness of the resolution of
3942 the histogram logs enabled with :option:`log_hist_msec`. For each increment
3943 in coarseness, fio outputs half as many bins. Defaults to 0, for which
3944 histogram logs contain 1216 latency bins. See :option:`write_hist_log`
3945 and `Log File Formats`_.
3947 .. option:: log_max_value=bool
3949 If :option:`log_avg_msec` is set, fio logs the average over that window. If
3950 you instead want to log the maximum value, set this option to 1. Defaults to
3951 0, meaning that averaged values are logged.
3953 .. option:: log_offset=bool
3955 If this is set, the iolog options will include the byte offset for the I/O
3956 entry as well as the other data values. Defaults to 0 meaning that
3957 offsets are not present in logs. Also see `Log File Formats`_.
3959 .. option:: log_compression=int
3961 If this is set, fio will compress the I/O logs as it goes, to keep the
3962 memory footprint lower. When a log reaches the specified size, that chunk is
3963 removed and compressed in the background. Given that I/O logs are fairly
3964 highly compressible, this yields a nice memory savings for longer runs. The
3965 downside is that the compression will consume some background CPU cycles, so
3966 it may impact the run. This, however, is also true if the logging ends up
3967 consuming most of the system memory. So pick your poison. The I/O logs are
3968 saved normally at the end of a run, by decompressing the chunks and storing
3969 them in the specified log file. This feature depends on the availability of
3972 .. option:: log_compression_cpus=str
3974 Define the set of CPUs that are allowed to handle online log compression for
3975 the I/O jobs. This can provide better isolation between performance
3976 sensitive jobs, and background compression work. See
3977 :option:`cpus_allowed` for the format used.
3979 .. option:: log_store_compressed=bool
3981 If set, fio will store the log files in a compressed format. They can be
3982 decompressed with fio, using the :option:`--inflate-log` command line
3983 parameter. The files will be stored with a :file:`.fz` suffix.
3985 .. option:: log_unix_epoch=bool
3987 If set, fio will log Unix timestamps to the log files produced by enabling
3988 write_type_log for each log type, instead of the default zero-based
3991 .. option:: log_alternate_epoch=bool
3993 If set, fio will log timestamps based on the epoch used by the clock specified
3994 in the log_alternate_epoch_clock_id option, to the log files produced by
3995 enabling write_type_log for each log type, instead of the default zero-based
3998 .. option:: log_alternate_epoch_clock_id=int
4000 Specifies the clock_id to be used by clock_gettime to obtain the alternate epoch
4001 if either log_unix_epoch or log_alternate_epoch are true. Otherwise has no
4002 effect. Default value is 0, or CLOCK_REALTIME.
4004 .. option:: block_error_percentiles=bool
4006 If set, record errors in trim block-sized units from writes and trims and
4007 output a histogram of how many trims it took to get to errors, and what kind
4008 of error was encountered.
4010 .. option:: bwavgtime=int
4012 Average the calculated bandwidth over the given time. Value is specified in
4013 milliseconds. If the job also does bandwidth logging through
4014 :option:`write_bw_log`, then the minimum of this option and
4015 :option:`log_avg_msec` will be used. Default: 500ms.
4017 .. option:: iopsavgtime=int
4019 Average the calculated IOPS over the given time. Value is specified in
4020 milliseconds. If the job also does IOPS logging through
4021 :option:`write_iops_log`, then the minimum of this option and
4022 :option:`log_avg_msec` will be used. Default: 500ms.
4024 .. option:: disk_util=bool
4026 Generate disk utilization statistics, if the platform supports it.
4029 .. option:: disable_lat=bool
4031 Disable measurements of total latency numbers. Useful only for cutting back
4032 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
4033 performance at really high IOPS rates. Note that to really get rid of a
4034 large amount of these calls, this option must be used with
4035 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
4037 .. option:: disable_clat=bool
4039 Disable measurements of completion latency numbers. See
4040 :option:`disable_lat`.
4042 .. option:: disable_slat=bool
4044 Disable measurements of submission latency numbers. See
4045 :option:`disable_lat`.
4047 .. option:: disable_bw_measurement=bool, disable_bw=bool
4049 Disable measurements of throughput/bandwidth numbers. See
4050 :option:`disable_lat`.
4052 .. option:: slat_percentiles=bool
4054 Report submission latency percentiles. Submission latency is not recorded
4055 for synchronous ioengines.
4057 .. option:: clat_percentiles=bool
4059 Report completion latency percentiles.
4061 .. option:: lat_percentiles=bool
4063 Report total latency percentiles. Total latency is the sum of submission
4064 latency and completion latency.
4066 .. option:: percentile_list=float_list
4068 Overwrite the default list of percentiles for latencies and the block error
4069 histogram. Each number is a floating point number in the range (0,100], and
4070 the maximum length of the list is 20. Use ``:`` to separate the numbers. For
4071 example, ``--percentile_list=99.5:99.9`` will cause fio to report the
4072 latency durations below which 99.5% and 99.9% of the observed latencies fell,
4075 .. option:: significant_figures=int
4077 If using :option:`--output-format` of `normal`, set the significant
4078 figures to this value. Higher values will yield more precise IOPS and
4079 throughput units, while lower values will round. Requires a minimum
4080 value of 1 and a maximum value of 10. Defaults to 4.
4086 .. option:: exitall_on_error
4088 When one job finishes in error, terminate the rest. The default is to wait
4089 for each job to finish.
4091 .. option:: continue_on_error=str
4093 Normally fio will exit the job on the first observed failure. If this option
4094 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
4095 EILSEQ) until the runtime is exceeded or the I/O size specified is
4096 completed. If this option is used, there are two more stats that are
4097 appended, the total error count and the first error. The error field given
4098 in the stats is the first error that was hit during the run.
4100 Note: a write error from the device may go unnoticed by fio when using
4101 buffered IO, as the write() (or similar) system call merely dirties the
4102 kernel pages, unless :option:`sync` or :option:`direct` is used. Device IO
4103 errors occur when the dirty data is actually written out to disk. If fully
4104 sync writes aren't desirable, :option:`fsync` or :option:`fdatasync` can be
4105 used as well. This is specific to writes, as reads are always synchronous.
4107 The allowed values are:
4110 Exit on any I/O or verify errors.
4113 Continue on read errors, exit on all others.
4116 Continue on write errors, exit on all others.
4119 Continue on any I/O error, exit on all others.
4122 Continue on verify errors, exit on all others.
4125 Continue on all errors.
4128 Backward-compatible alias for 'none'.
4131 Backward-compatible alias for 'all'.
4133 .. option:: ignore_error=str
4135 Sometimes you want to ignore some errors during test in that case you can
4136 specify error list for each error type, instead of only being able to
4137 ignore the default 'non-fatal error' using :option:`continue_on_error`.
4138 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
4139 given error type is separated with ':'. Error may be symbol ('ENOSPC',
4140 'ENOMEM') or integer. Example::
4142 ignore_error=EAGAIN,ENOSPC:122
4144 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
4145 WRITE. This option works by overriding :option:`continue_on_error` with
4146 the list of errors for each error type if any.
4148 .. option:: error_dump=bool
4150 If set dump every error even if it is non fatal, true by default. If
4151 disabled only fatal error will be dumped.
4153 Running predefined workloads
4154 ----------------------------
4156 Fio includes predefined profiles that mimic the I/O workloads generated by
4159 .. option:: profile=str
4161 The predefined workload to run. Current profiles are:
4164 Threaded I/O bench (tiotest/tiobench) like workload.
4167 Aerospike Certification Tool (ACT) like workload.
4169 To view a profile's additional options use :option:`--cmdhelp` after specifying
4170 the profile. For example::
4172 $ fio --profile=act --cmdhelp
4177 .. option:: device-names=str
4182 .. option:: load=int
4185 ACT load multiplier. Default: 1.
4187 .. option:: test-duration=time
4190 How long the entire test takes to run. When the unit is omitted, the value
4191 is given in seconds. Default: 24h.
4193 .. option:: threads-per-queue=int
4196 Number of read I/O threads per device. Default: 8.
4198 .. option:: read-req-num-512-blocks=int
4201 Number of 512B blocks to read at the time. Default: 3.
4203 .. option:: large-block-op-kbytes=int
4206 Size of large block ops in KiB (writes). Default: 131072.
4211 Set to run ACT prep phase.
4213 Tiobench profile options
4214 ~~~~~~~~~~~~~~~~~~~~~~~~
4216 .. option:: size=str
4221 .. option:: block=int
4224 Block size in bytes. Default: 4096.
4226 .. option:: numruns=int
4236 .. option:: threads=int
4241 Interpreting the output
4242 -----------------------
4245 Example output was based on the following:
4246 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --time_based \
4247 --rate=1256k --bs=14K --name=quick --runtime=1s --name=mixed \
4248 --runtime=2m --rw=rw
4250 Fio spits out a lot of output. While running, fio will display the status of the
4251 jobs created. An example of that would be::
4253 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]
4255 The characters inside the first set of square brackets denote the current status of
4256 each thread. The first character is the first job defined in the job file, and so
4257 forth. The possible values (in typical life cycle order) are:
4259 +------+-----+-----------------------------------------------------------+
4261 +======+=====+===========================================================+
4262 | P | | Thread setup, but not started. |
4263 +------+-----+-----------------------------------------------------------+
4264 | C | | Thread created. |
4265 +------+-----+-----------------------------------------------------------+
4266 | I | | Thread initialized, waiting or generating necessary data. |
4267 +------+-----+-----------------------------------------------------------+
4268 | | p | Thread running pre-reading file(s). |
4269 +------+-----+-----------------------------------------------------------+
4270 | | / | Thread is in ramp period. |
4271 +------+-----+-----------------------------------------------------------+
4272 | | R | Running, doing sequential reads. |
4273 +------+-----+-----------------------------------------------------------+
4274 | | r | Running, doing random reads. |
4275 +------+-----+-----------------------------------------------------------+
4276 | | W | Running, doing sequential writes. |
4277 +------+-----+-----------------------------------------------------------+
4278 | | w | Running, doing random writes. |
4279 +------+-----+-----------------------------------------------------------+
4280 | | M | Running, doing mixed sequential reads/writes. |
4281 +------+-----+-----------------------------------------------------------+
4282 | | m | Running, doing mixed random reads/writes. |
4283 +------+-----+-----------------------------------------------------------+
4284 | | D | Running, doing sequential trims. |
4285 +------+-----+-----------------------------------------------------------+
4286 | | d | Running, doing random trims. |
4287 +------+-----+-----------------------------------------------------------+
4288 | | F | Running, currently waiting for :manpage:`fsync(2)`. |
4289 +------+-----+-----------------------------------------------------------+
4290 | | V | Running, doing verification of written data. |
4291 +------+-----+-----------------------------------------------------------+
4292 | f | | Thread finishing. |
4293 +------+-----+-----------------------------------------------------------+
4294 | E | | Thread exited, not reaped by main thread yet. |
4295 +------+-----+-----------------------------------------------------------+
4296 | _ | | Thread reaped. |
4297 +------+-----+-----------------------------------------------------------+
4298 | X | | Thread reaped, exited with an error. |
4299 +------+-----+-----------------------------------------------------------+
4300 | K | | Thread reaped, exited due to signal. |
4301 +------+-----+-----------------------------------------------------------+
4304 Example output was based on the following:
4305 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --runtime=58m \
4306 --time_based --rate=2512k --bs=256K --numjobs=10 \
4307 --name=readers --rw=read --name=writers --rw=write
4309 Fio will condense the thread string as not to take up more space on the command
4310 line than needed. For instance, if you have 10 readers and 10 writers running,
4311 the output would look like this::
4313 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]
4315 Note that the status string is displayed in order, so it's possible to tell which of
4316 the jobs are currently doing what. In the example above this means that jobs 1--10
4317 are readers and 11--20 are writers.
4319 The other values are fairly self explanatory -- number of threads currently
4320 running and doing I/O, the number of currently open files (f=), the estimated
4321 completion percentage, the rate of I/O since last check (read speed listed first,
4322 then write speed and optionally trim speed) in terms of bandwidth and IOPS,
4323 and time to completion for the current running group. It's impossible to estimate
4324 runtime of the following groups (if any).
4327 Example output was based on the following:
4328 TZ=UTC fio --iodepth=16 --ioengine=posixaio --filename=/tmp/fiofile \
4329 --direct=1 --size=100M --time_based --runtime=50s --rate_iops=89 \
4330 --bs=7K --name=Client1 --rw=write
4332 When fio is done (or interrupted by :kbd:`Ctrl-C`), it will show the data for
4333 each thread, group of threads, and disks in that order. For each overall thread (or
4334 group) the output looks like::
4336 Client1: (groupid=0, jobs=1): err= 0: pid=16109: Sat Jun 24 12:07:54 2017
4337 write: IOPS=88, BW=623KiB/s (638kB/s)(30.4MiB/50032msec)
4338 slat (nsec): min=500, max=145500, avg=8318.00, stdev=4781.50
4339 clat (usec): min=170, max=78367, avg=4019.02, stdev=8293.31
4340 lat (usec): min=174, max=78375, avg=4027.34, stdev=8291.79
4341 clat percentiles (usec):
4342 | 1.00th=[ 302], 5.00th=[ 326], 10.00th=[ 343], 20.00th=[ 363],
4343 | 30.00th=[ 392], 40.00th=[ 404], 50.00th=[ 416], 60.00th=[ 445],
4344 | 70.00th=[ 816], 80.00th=[ 6718], 90.00th=[12911], 95.00th=[21627],
4345 | 99.00th=[43779], 99.50th=[51643], 99.90th=[68682], 99.95th=[72877],
4347 bw ( KiB/s): min= 532, max= 686, per=0.10%, avg=622.87, stdev=24.82, samples= 100
4348 iops : min= 76, max= 98, avg=88.98, stdev= 3.54, samples= 100
4349 lat (usec) : 250=0.04%, 500=64.11%, 750=4.81%, 1000=2.79%
4350 lat (msec) : 2=4.16%, 4=1.84%, 10=4.90%, 20=11.33%, 50=5.37%
4351 lat (msec) : 100=0.65%
4352 cpu : usr=0.27%, sys=0.18%, ctx=12072, majf=0, minf=21
4353 IO depths : 1=85.0%, 2=13.1%, 4=1.8%, 8=0.1%, 16=0.0%, 32=0.0%, >=64=0.0%
4354 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
4355 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
4356 issued rwt: total=0,4450,0, short=0,0,0, dropped=0,0,0
4357 latency : target=0, window=0, percentile=100.00%, depth=8
4359 The job name (or first job's name when using :option:`group_reporting`) is printed,
4360 along with the group id, count of jobs being aggregated, last error id seen (which
4361 is 0 when there are no errors), pid/tid of that thread and the time the job/group
4362 completed. Below are the I/O statistics for each data direction performed (showing
4363 writes in the example above). In the order listed, they denote:
4366 The string before the colon shows the I/O direction the statistics
4367 are for. **IOPS** is the average I/Os performed per second. **BW**
4368 is the average bandwidth rate shown as: value in power of 2 format
4369 (value in power of 10 format). The last two values show: (**total
4370 I/O performed** in power of 2 format / **runtime** of that thread).
4373 Submission latency (**min** being the minimum, **max** being the
4374 maximum, **avg** being the average, **stdev** being the standard
4375 deviation). This is the time from when fio initialized the I/O
4376 to submission. For synchronous ioengines this includes the time
4377 up until just before the ioengine's queue function is called.
4378 For asynchronous ioengines this includes the time up through the
4379 completion of the ioengine's queue function (and commit function
4380 if it is defined). For sync I/O this row is not displayed as the
4381 slat is negligible. This value can be in nanoseconds,
4382 microseconds or milliseconds --- fio will choose the most
4383 appropriate base and print that (in the example above
4384 nanoseconds was the best scale). Note: in :option:`--minimal`
4385 mode latencies are always expressed in microseconds.
4388 Completion latency. Same names as slat, this denotes the time from
4389 submission to completion of the I/O pieces. For sync I/O, this
4390 represents the time from when the I/O was submitted to the
4391 operating system to when it was completed. For asynchronous
4392 ioengines this is the time from when the ioengine's queue (and
4393 commit if available) functions were completed to when the I/O's
4394 completion was reaped by fio.
4397 Total latency. Same names as slat and clat, this denotes the time from
4398 when fio created the I/O unit to completion of the I/O operation.
4399 It is the sum of submission and completion latency.
4402 Bandwidth statistics based on samples. Same names as the xlat stats,
4403 but also includes the number of samples taken (**samples**) and an
4404 approximate percentage of total aggregate bandwidth this thread
4405 received in its group (**per**). This last value is only really
4406 useful if the threads in this group are on the same disk, since they
4407 are then competing for disk access.
4410 IOPS statistics based on samples. Same names as bw.
4412 **lat (nsec/usec/msec)**
4413 The distribution of I/O completion latencies. This is the time from when
4414 I/O leaves fio and when it gets completed. Unlike the separate
4415 read/write/trim sections above, the data here and in the remaining
4416 sections apply to all I/Os for the reporting group. 250=0.04% means that
4417 0.04% of the I/Os completed in under 250us. 500=64.11% means that 64.11%
4418 of the I/Os required 250 to 499us for completion.
4421 CPU usage. User and system time, along with the number of context
4422 switches this thread went through, usage of system and user time, and
4423 finally the number of major and minor page faults. The CPU utilization
4424 numbers are averages for the jobs in that reporting group, while the
4425 context and fault counters are summed.
4428 The distribution of I/O depths over the job lifetime. The numbers are
4429 divided into powers of 2 and each entry covers depths from that value
4430 up to those that are lower than the next entry -- e.g., 16= covers
4431 depths from 16 to 31. Note that the range covered by a depth
4432 distribution entry can be different to the range covered by the
4433 equivalent submit/complete distribution entry.
4436 How many pieces of I/O were submitting in a single submit call. Each
4437 entry denotes that amount and below, until the previous entry -- e.g.,
4438 16=100% means that we submitted anywhere between 9 to 16 I/Os per submit
4439 call. Note that the range covered by a submit distribution entry can
4440 be different to the range covered by the equivalent depth distribution
4444 Like the above submit number, but for completions instead.
4447 The number of read/write/trim requests issued, and how many of them were
4451 These values are for :option:`latency_target` and related options. When
4452 these options are engaged, this section describes the I/O depth required
4453 to meet the specified latency target.
4456 Example output was based on the following:
4457 TZ=UTC fio --ioengine=null --iodepth=2 --size=100M --numjobs=2 \
4458 --rate_process=poisson --io_limit=32M --name=read --bs=128k \
4459 --rate=11M --name=write --rw=write --bs=2k --rate=700k
4461 After each client has been listed, the group statistics are printed. They
4462 will look like this::
4464 Run status group 0 (all jobs):
4465 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
4466 WRITE: bw=1231KiB/s (1261kB/s), 616KiB/s-621KiB/s (630kB/s-636kB/s), io=64.0MiB (67.1MB), run=52747-53223msec
4468 For each data direction it prints:
4471 Aggregate bandwidth of threads in this group followed by the
4472 minimum and maximum bandwidth of all the threads in this group.
4473 Values outside of brackets are power-of-2 format and those
4474 within are the equivalent value in a power-of-10 format.
4476 Aggregate I/O performed of all threads in this group. The
4477 format is the same as bw.
4479 The smallest and longest runtimes of the threads in this group.
4481 And finally, the disk statistics are printed. This is Linux specific. They will look like this::
4483 Disk stats (read/write):
4484 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
4486 Each value is printed for both reads and writes, with reads first. The
4490 Number of I/Os performed by all groups.
4492 Number of merges performed by the I/O scheduler.
4494 Number of ticks we kept the disk busy.
4496 Total time spent in the disk queue.
4498 The disk utilization. A value of 100% means we kept the disk
4499 busy constantly, 50% would be a disk idling half of the time.
4501 It is also possible to get fio to dump the current output while it is running,
4502 without terminating the job. To do that, send fio the **USR1** signal. You can
4503 also get regularly timed dumps by using the :option:`--status-interval`
4504 parameter, or by creating a file in :file:`/tmp` named
4505 :file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
4506 current output status.
4512 For scripted usage where you typically want to generate tables or graphs of the
4513 results, fio can output the results in a semicolon separated format. The format
4514 is one long line of values, such as::
4516 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%
4517 A description of this job goes here.
4519 The job description (if provided) follows on a second line for terse v2.
4520 It appears on the same line for other terse versions.
4522 To enable terse output, use the :option:`--minimal` or
4523 :option:`--output-format`\=terse command line options. The
4524 first value is the version of the terse output format. If the output has to be
4525 changed for some reason, this number will be incremented by 1 to signify that
4528 Split up, the format is as follows (comments in brackets denote when a
4529 field was introduced or whether it's specific to some terse version):
4533 terse version, fio version [v3], jobname, groupid, error
4537 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
4538 Submission latency: min, max, mean, stdev (usec)
4539 Completion latency: min, max, mean, stdev (usec)
4540 Completion latency percentiles: 20 fields (see below)
4541 Total latency: min, max, mean, stdev (usec)
4542 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
4543 IOPS [v5]: min, max, mean, stdev, number of samples
4549 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
4550 Submission latency: min, max, mean, stdev (usec)
4551 Completion latency: min, max, mean, stdev (usec)
4552 Completion latency percentiles: 20 fields (see below)
4553 Total latency: min, max, mean, stdev (usec)
4554 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
4555 IOPS [v5]: min, max, mean, stdev, number of samples
4557 TRIM status [all but version 3]:
4559 Fields are similar to READ/WRITE status.
4563 user, system, context switches, major faults, minor faults
4567 <=1, 2, 4, 8, 16, 32, >=64
4569 I/O latencies microseconds::
4571 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
4573 I/O latencies milliseconds::
4575 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
4577 Disk utilization [v3]::
4579 disk name, read ios, write ios, read merges, write merges, read ticks, write ticks,
4580 time spent in queue, disk utilization percentage
4582 Additional Info (dependent on continue_on_error, default off)::
4584 total # errors, first error code
4586 Additional Info (dependent on description being set)::
4590 Completion latency percentiles can be a grouping of up to 20 sets, so for the
4591 terse output fio writes all of them. Each field will look like this::
4595 which is the Xth percentile, and the `usec` latency associated with it.
4597 For `Disk utilization`, all disks used by fio are shown. So for each disk there
4598 will be a disk utilization section.
4600 Below is a single line containing short names for each of the fields in the
4601 minimal output v3, separated by semicolons::
4603 terse_version_3;fio_version;jobname;groupid;error;read_kb;read_bandwidth_kb;read_iops;read_runtime_ms;read_slat_min_us;read_slat_max_us;read_slat_mean_us;read_slat_dev_us;read_clat_min_us;read_clat_max_us;read_clat_mean_us;read_clat_dev_us;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_us;read_lat_max_us;read_lat_mean_us;read_lat_dev_us;read_bw_min_kb;read_bw_max_kb;read_bw_agg_pct;read_bw_mean_kb;read_bw_dev_kb;write_kb;write_bandwidth_kb;write_iops;write_runtime_ms;write_slat_min_us;write_slat_max_us;write_slat_mean_us;write_slat_dev_us;write_clat_min_us;write_clat_max_us;write_clat_mean_us;write_clat_dev_us;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_us;write_lat_max_us;write_lat_mean_us;write_lat_dev_us;write_bw_min_kb;write_bw_max_kb;write_bw_agg_pct;write_bw_mean_kb;write_bw_dev_kb;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
4605 In client/server mode terse output differs from what appears when jobs are run
4606 locally. Disk utilization data is omitted from the standard terse output and
4607 for v3 and later appears on its own separate line at the end of each terse
4614 The `json` output format is intended to be both human readable and convenient
4615 for automated parsing. For the most part its sections mirror those of the
4616 `normal` output. The `runtime` value is reported in msec and the `bw` value is
4617 reported in 1024 bytes per second units.
4623 The `json+` output format is identical to the `json` output format except that it
4624 adds a full dump of the completion latency bins. Each `bins` object contains a
4625 set of (key, value) pairs where keys are latency durations and values count how
4626 many I/Os had completion latencies of the corresponding duration. For example,
4629 "bins" : { "87552" : 1, "89600" : 1, "94720" : 1, "96768" : 1, "97792" : 1, "99840" : 1, "100864" : 2, "103936" : 6, "104960" : 534, "105984" : 5995, "107008" : 7529, ... }
4631 This data indicates that one I/O required 87,552ns to complete, two I/Os required
4632 100,864ns to complete, and 7529 I/Os required 107,008ns to complete.
4634 Also included with fio is a Python script `fio_jsonplus_clat2csv` that takes
4635 json+ output and generates CSV-formatted latency data suitable for plotting.
4637 The latency durations actually represent the midpoints of latency intervals.
4638 For details refer to :file:`stat.h`.
4644 There are two trace file format that you can encounter. The older (v1) format is
4645 unsupported since version 1.20-rc3 (March 2008). It will still be described
4646 below in case that you get an old trace and want to understand it.
4648 In any case the trace is a simple text file with a single action per line.
4651 Trace file format v1
4652 ~~~~~~~~~~~~~~~~~~~~
4654 Each line represents a single I/O action in the following format::
4658 where `rw=0/1` for read/write, and the `offset` and `length` entries being in bytes.
4660 This format is not supported in fio versions >= 1.20-rc3.
4663 Trace file format v2
4664 ~~~~~~~~~~~~~~~~~~~~
4666 The second version of the trace file format was added in fio version 1.17. It
4667 allows one to access more than one file per trace and has a bigger set of possible
4670 The first line of the trace file has to be::
4674 Following this can be lines in two different formats, which are described below.
4676 The file management format::
4680 The `filename` is given as an absolute path. The `action` can be one of these:
4683 Add the given `filename` to the trace.
4685 Open the file with the given `filename`. The `filename` has to have
4686 been added with the **add** action before.
4688 Close the file with the given `filename`. The file has to have been
4692 The file I/O action format::
4694 filename action offset length
4696 The `filename` is given as an absolute path, and has to have been added and
4697 opened before it can be used with this format. The `offset` and `length` are
4698 given in bytes. The `action` can be one of these:
4701 Wait for `offset` microseconds. Everything below 100 is discarded.
4702 The time is relative to the previous `wait` statement. Note that
4703 action `wait` is not allowed as of version 3, as the same behavior
4704 can be achieved using timestamps.
4706 Read `length` bytes beginning from `offset`.
4708 Write `length` bytes beginning from `offset`.
4710 :manpage:`fsync(2)` the file.
4712 :manpage:`fdatasync(2)` the file.
4714 Trim the given file from the given `offset` for `length` bytes.
4717 Trace file format v3
4718 ~~~~~~~~~~~~~~~~~~~~
4720 The third version of the trace file format was added in fio version 3.31. It
4721 forces each action to have a timestamp associated with it.
4723 The first line of the trace file has to be::
4727 Following this can be lines in two different formats, which are described below.
4729 The file management format::
4731 timestamp filename action
4733 The file I/O action format::
4735 timestamp filename action offset length
4737 The `timestamp` is relative to the beginning of the run (ie starts at 0). The
4738 `filename`, `action`, `offset` and `length` are identical to version 2, except
4739 that version 3 does not allow the `wait` action.
4742 I/O Replay - Merging Traces
4743 ---------------------------
4745 Colocation is a common practice used to get the most out of a machine.
4746 Knowing which workloads play nicely with each other and which ones don't is
4747 a much harder task. While fio can replay workloads concurrently via multiple
4748 jobs, it leaves some variability up to the scheduler making results harder to
4749 reproduce. Merging is a way to make the order of events consistent.
4751 Merging is integrated into I/O replay and done when a
4752 :option:`merge_blktrace_file` is specified. The list of files passed to
4753 :option:`read_iolog` go through the merge process and output a single file
4754 stored to the specified file. The output file is passed on as if it were the
4755 only file passed to :option:`read_iolog`. An example would look like::
4757 $ fio --read_iolog="<file1>:<file2>" --merge_blktrace_file="<output_file>"
4759 Creating only the merged file can be done by passing the command line argument
4760 :option:`--merge-blktrace-only`.
4762 Scaling traces can be done to see the relative impact of any particular trace
4763 being slowed down or sped up. :option:`merge_blktrace_scalars` takes in a colon
4764 separated list of percentage scalars. It is index paired with the files passed
4765 to :option:`read_iolog`.
4767 With scaling, it may be desirable to match the running time of all traces.
4768 This can be done with :option:`merge_blktrace_iters`. It is index paired with
4769 :option:`read_iolog` just like :option:`merge_blktrace_scalars`.
4771 In an example, given two traces, A and B, each 60s long. If we want to see
4772 the impact of trace A issuing IOs twice as fast and repeat trace A over the
4773 runtime of trace B, the following can be done::
4775 $ fio --read_iolog="<trace_a>:"<trace_b>" --merge_blktrace_file"<output_file>" --merge_blktrace_scalars="50:100" --merge_blktrace_iters="2:1"
4777 This runs trace A at 2x the speed twice for approximately the same runtime as
4778 a single run of trace B.
4781 CPU idleness profiling
4782 ----------------------
4784 In some cases, we want to understand CPU overhead in a test. For example, we
4785 test patches for the specific goodness of whether they reduce CPU usage.
4786 Fio implements a balloon approach to create a thread per CPU that runs at idle
4787 priority, meaning that it only runs when nobody else needs the cpu.
4788 By measuring the amount of work completed by the thread, idleness of each CPU
4789 can be derived accordingly.
4791 An unit work is defined as touching a full page of unsigned characters. Mean and
4792 standard deviation of time to complete an unit work is reported in "unit work"
4793 section. Options can be chosen to report detailed percpu idleness or overall
4794 system idleness by aggregating percpu stats.
4797 Verification and triggers
4798 -------------------------
4800 Fio is usually run in one of two ways, when data verification is done. The first
4801 is a normal write job of some sort with verify enabled. When the write phase has
4802 completed, fio switches to reads and verifies everything it wrote. The second
4803 model is running just the write phase, and then later on running the same job
4804 (but with reads instead of writes) to repeat the same I/O patterns and verify
4805 the contents. Both of these methods depend on the write phase being completed,
4806 as fio otherwise has no idea how much data was written.
4808 With verification triggers, fio supports dumping the current write state to
4809 local files. Then a subsequent read verify workload can load this state and know
4810 exactly where to stop. This is useful for testing cases where power is cut to a
4811 server in a managed fashion, for instance.
4813 A verification trigger consists of two things:
4815 1) Storing the write state of each job.
4816 2) Executing a trigger command.
4818 The write state is relatively small, on the order of hundreds of bytes to single
4819 kilobytes. It contains information on the number of completions done, the last X
4822 A trigger is invoked either through creation ('touch') of a specified file in
4823 the system, or through a timeout setting. If fio is run with
4824 :option:`--trigger-file`\= :file:`/tmp/trigger-file`, then it will continually
4825 check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
4826 will fire off the trigger (thus saving state, and executing the trigger
4829 For client/server runs, there's both a local and remote trigger. If fio is
4830 running as a server backend, it will send the job states back to the client for
4831 safe storage, then execute the remote trigger, if specified. If a local trigger
4832 is specified, the server will still send back the write state, but the client
4833 will then execute the trigger.
4835 Verification trigger example
4836 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4838 Let's say we want to run a powercut test on the remote Linux machine 'server'.
4839 Our write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
4840 some point during the run, and we'll run this test from the safety or our local
4841 machine, 'localbox'. On the server, we'll start the fio backend normally::
4843 server# fio --server
4845 and on the client, we'll fire off the workload::
4847 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
4849 We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
4851 echo b > /proc/sysrq-trigger
4853 on the server once it has received the trigger and sent us the write state. This
4854 will work, but it's not **really** cutting power to the server, it's merely
4855 abruptly rebooting it. If we have a remote way of cutting power to the server
4856 through IPMI or similar, we could do that through a local trigger command
4857 instead. Let's assume we have a script that does IPMI reboot of a given hostname,
4858 ipmi-reboot. On localbox, we could then have run fio with a local trigger
4861 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
4863 For this case, fio would wait for the server to send us the write state, then
4864 execute ``ipmi-reboot server`` when that happened.
4866 Loading verify state
4867 ~~~~~~~~~~~~~~~~~~~~
4869 To load stored write state, a read verification job file must contain the
4870 :option:`verify_state_load` option. If that is set, fio will load the previously
4871 stored state. For a local fio run this is done by loading the files directly,
4872 and on a client/server run, the server backend will ask the client to send the
4873 files over and load them from there.
4879 Fio supports a variety of log file formats, for logging latencies, bandwidth,
4880 and IOPS. The logs share a common format, which looks like this:
4882 *time* (`msec`), *value*, *data direction*, *block size* (`bytes`),
4883 *offset* (`bytes`), *command priority*
4885 *Time* for the log entry is always in milliseconds. The *value* logged depends
4886 on the type of log, it will be one of the following:
4889 Value is latency in nsecs
4895 *Data direction* is one of the following:
4904 The entry's *block size* is always in bytes. The *offset* is the position in bytes
4905 from the start of the file for that particular I/O. The logging of the offset can be
4906 toggled with :option:`log_offset`.
4908 *Command priority* is 0 for normal priority and 1 for high priority. This is controlled
4909 by the ioengine specific :option:`cmdprio_percentage`.
4911 Fio defaults to logging every individual I/O but when windowed logging is set
4912 through :option:`log_avg_msec`, either the average (by default) or the maximum
4913 (:option:`log_max_value` is set) *value* seen over the specified period of time
4914 is recorded. Each *data direction* seen within the window period will aggregate
4915 its values in a separate row. Further, when using windowed logging the *block
4916 size* and *offset* entries will always contain 0.
4922 Normally fio is invoked as a stand-alone application on the machine where the
4923 I/O workload should be generated. However, the backend and frontend of fio can
4924 be run separately i.e., the fio server can generate an I/O workload on the "Device
4925 Under Test" while being controlled by a client on another machine.
4927 Start the server on the machine which has access to the storage DUT::
4931 where `args` defines what fio listens to. The arguments are of the form
4932 ``type,hostname`` or ``IP,port``. *type* is either ``ip`` (or ip4) for TCP/IP
4933 v4, ``ip6`` for TCP/IP v6, or ``sock`` for a local unix domain socket.
4934 *hostname* is either a hostname or IP address, and *port* is the port to listen
4935 to (only valid for TCP/IP, not a local socket). Some examples:
4939 Start a fio server, listening on all interfaces on the default port (8765).
4941 2) ``fio --server=ip:hostname,4444``
4943 Start a fio server, listening on IP belonging to hostname and on port 4444.
4945 3) ``fio --server=ip6:::1,4444``
4947 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
4949 4) ``fio --server=,4444``
4951 Start a fio server, listening on all interfaces on port 4444.
4953 5) ``fio --server=1.2.3.4``
4955 Start a fio server, listening on IP 1.2.3.4 on the default port.
4957 6) ``fio --server=sock:/tmp/fio.sock``
4959 Start a fio server, listening on the local socket :file:`/tmp/fio.sock`.
4961 Once a server is running, a "client" can connect to the fio server with::
4963 fio <local-args> --client=<server> <remote-args> <job file(s)>
4965 where `local-args` are arguments for the client where it is running, `server`
4966 is the connect string, and `remote-args` and `job file(s)` are sent to the
4967 server. The `server` string follows the same format as it does on the server
4968 side, to allow IP/hostname/socket and port strings.
4970 Fio can connect to multiple servers this way::
4972 fio --client=<server1> <job file(s)> --client=<server2> <job file(s)>
4974 If the job file is located on the fio server, then you can tell the server to
4975 load a local file as well. This is done by using :option:`--remote-config` ::
4977 fio --client=server --remote-config /path/to/file.fio
4979 Then fio will open this local (to the server) job file instead of being passed
4980 one from the client.
4982 If you have many servers (example: 100 VMs/containers), you can input a pathname
4983 of a file containing host IPs/names as the parameter value for the
4984 :option:`--client` option. For example, here is an example :file:`host.list`
4985 file containing 2 hostnames::
4987 host1.your.dns.domain
4988 host2.your.dns.domain
4990 The fio command would then be::
4992 fio --client=host.list <job file(s)>
4994 In this mode, you cannot input server-specific parameters or job files -- all
4995 servers receive the same job file.
4997 In order to let ``fio --client`` runs use a shared filesystem from multiple
4998 hosts, ``fio --client`` now prepends the IP address of the server to the
4999 filename. For example, if fio is using the directory :file:`/mnt/nfs/fio` and is
5000 writing filename :file:`fileio.tmp`, with a :option:`--client` `hostfile`
5001 containing two hostnames ``h1`` and ``h2`` with IP addresses 192.168.10.120 and
5002 192.168.10.121, then fio will create two files::
5004 /mnt/nfs/fio/192.168.10.120.fileio.tmp
5005 /mnt/nfs/fio/192.168.10.121.fileio.tmp
5007 Terse output in client/server mode will differ slightly from what is produced
5008 when fio is run in stand-alone mode. See the terse output section for details.