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 Limit runtime. The test will run until it completes the configured I/O
690 workload or until it has run for this specified amount of time, whichever
691 occurs first. It can be quite hard to determine for how long a specified
692 job will run, so this parameter is handy to cap the total runtime to a
693 given time. When the unit is omitted, the value is interpreted in
696 .. option:: time_based
698 If set, fio will run for the duration of the :option:`runtime` specified
699 even if the file(s) are completely read or written. It will simply loop over
700 the same workload as many times as the :option:`runtime` allows.
702 .. option:: startdelay=irange(time)
704 Delay the start of job for the specified amount of time. Can be a single
705 value or a range. When given as a range, each thread will choose a value
706 randomly from within the range. Value is in seconds if a unit is omitted.
708 .. option:: ramp_time=time
710 If set, fio will run the specified workload for this amount of time before
711 logging any performance numbers. Useful for letting performance settle
712 before logging results, thus minimizing the runtime required for stable
713 results. Note that the ``ramp_time`` is considered lead in time for a job,
714 thus it will increase the total runtime if a special timeout or
715 :option:`runtime` is specified. When the unit is omitted, the value is
718 .. option:: clocksource=str
720 Use the given clocksource as the base of timing. The supported options are:
723 :manpage:`gettimeofday(2)`
726 :manpage:`clock_gettime(2)`
729 Internal CPU clock source
731 cpu is the preferred clocksource if it is reliable, as it is very fast (and
732 fio is heavy on time calls). Fio will automatically use this clocksource if
733 it's supported and considered reliable on the system it is running on,
734 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
735 means supporting TSC Invariant.
737 .. option:: gtod_reduce=bool
739 Enable all of the :manpage:`gettimeofday(2)` reducing options
740 (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus
741 reduce precision of the timeout somewhat to really shrink the
742 :manpage:`gettimeofday(2)` call count. With this option enabled, we only do
743 about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all
744 time keeping was enabled.
746 .. option:: gtod_cpu=int
748 Sometimes it's cheaper to dedicate a single thread of execution to just
749 getting the current time. Fio (and databases, for instance) are very
750 intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set
751 one CPU aside for doing nothing but logging current time to a shared memory
752 location. Then the other threads/processes that run I/O workloads need only
753 copy that segment, instead of entering the kernel with a
754 :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time
755 calls will be excluded from other uses. Fio will manually clear it from the
756 CPU mask of other jobs.
762 .. option:: directory=str
764 Prefix filenames with this directory. Used to place files in a different
765 location than :file:`./`. You can specify a number of directories by
766 separating the names with a ':' character. These directories will be
767 assigned equally distributed to job clones created by :option:`numjobs` as
768 long as they are using generated filenames. If specific `filename(s)` are
769 set fio will use the first listed directory, and thereby matching the
770 `filename` semantic (which generates a file for each clone if not
771 specified, but lets all clones use the same file if set).
773 See the :option:`filename` option for information on how to escape "``:``"
774 characters within the directory path itself.
776 Note: To control the directory fio will use for internal state files
777 use :option:`--aux-path`.
779 .. option:: filename=str
781 Fio normally makes up a `filename` based on the job name, thread number, and
782 file number (see :option:`filename_format`). If you want to share files
783 between threads in a job or several
784 jobs with fixed file paths, specify a `filename` for each of them to override
785 the default. If the ioengine is file based, you can specify a number of files
786 by separating the names with a ':' colon. So if you wanted a job to open
787 :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use
788 ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is
789 specified, :option:`nrfiles` is ignored. The size of regular files specified
790 by this option will be :option:`size` divided by number of files unless an
791 explicit size is specified by :option:`filesize`.
793 Each colon in the wanted path must be escaped with a ``\``
794 character. For instance, if the path is :file:`/dev/dsk/foo@3,0:c` then you
795 would use ``filename=/dev/dsk/foo@3,0\:c`` and if the path is
796 :file:`F:\\filename` then you would use ``filename=F\:\filename``.
798 On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for
799 the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc.
800 Note: Windows and FreeBSD prevent write access to areas
801 of the disk containing in-use data (e.g. filesystems).
803 The filename "`-`" is a reserved name, meaning *stdin* or *stdout*. Which
804 of the two depends on the read/write direction set.
806 .. option:: filename_format=str
808 If sharing multiple files between jobs, it is usually necessary to have fio
809 generate the exact names that you want. By default, fio will name a file
810 based on the default file format specification of
811 :file:`jobname.jobnumber.filenumber`. With this option, that can be
812 customized. Fio will recognize and replace the following keywords in this
816 The name of the worker thread or process.
818 IP of the fio process when using client/server mode.
820 The incremental number of the worker thread or process.
822 The incremental number of the file for that worker thread or
825 To have dependent jobs share a set of files, this option can be set to have
826 fio generate filenames that are shared between the two. For instance, if
827 :file:`testfiles.$filenum` is specified, file number 4 for any job will be
828 named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum`
829 will be used if no other format specifier is given.
831 If you specify a path then the directories will be created up to the
832 main directory for the file. So for example if you specify
833 ``filename_format=a/b/c/$jobnum`` then the directories a/b/c will be
834 created before the file setup part of the job. If you specify
835 :option:`directory` then the path will be relative that directory,
836 otherwise it is treated as the absolute path.
838 .. option:: unique_filename=bool
840 To avoid collisions between networked clients, fio defaults to prefixing any
841 generated filenames (with a directory specified) with the source of the
842 client connecting. To disable this behavior, set this option to 0.
844 .. option:: opendir=str
846 Recursively open any files below directory `str`.
848 .. option:: lockfile=str
850 Fio defaults to not locking any files before it does I/O to them. If a file
851 or file descriptor is shared, fio can serialize I/O to that file to make the
852 end result consistent. This is usual for emulating real workloads that share
853 files. The lock modes are:
856 No locking. The default.
858 Only one thread or process may do I/O at a time, excluding all
861 Read-write locking on the file. Many readers may
862 access the file at the same time, but writes get exclusive access.
864 .. option:: nrfiles=int
866 Number of files to use for this job. Defaults to 1. The size of files
867 will be :option:`size` divided by this unless explicit size is specified by
868 :option:`filesize`. Files are created for each thread separately, and each
869 file will have a file number within its name by default, as explained in
870 :option:`filename` section.
873 .. option:: openfiles=int
875 Number of files to keep open at the same time. Defaults to the same as
876 :option:`nrfiles`, can be set smaller to limit the number simultaneous
879 .. option:: file_service_type=str
881 Defines how fio decides which file from a job to service next. The following
885 Choose a file at random.
888 Round robin over opened files. This is the default.
891 Finish one file before moving on to the next. Multiple files can
892 still be open depending on :option:`openfiles`.
895 Use a *Zipf* distribution to decide what file to access.
898 Use a *Pareto* distribution to decide what file to access.
901 Use a *Gaussian* (normal) distribution to decide what file to
907 For *random*, *roundrobin*, and *sequential*, a postfix can be appended to
908 tell fio how many I/Os to issue before switching to a new file. For example,
909 specifying ``file_service_type=random:8`` would cause fio to issue
910 8 I/Os before selecting a new file at random. For the non-uniform
911 distributions, a floating point postfix can be given to influence how the
912 distribution is skewed. See :option:`random_distribution` for a description
913 of how that would work.
915 .. option:: ioscheduler=str
917 Attempt to switch the device hosting the file to the specified I/O scheduler
920 .. option:: create_serialize=bool
922 If true, serialize the file creation for the jobs. This may be handy to
923 avoid interleaving of data files, which may greatly depend on the filesystem
924 used and even the number of processors in the system. Default: true.
926 .. option:: create_fsync=bool
928 :manpage:`fsync(2)` the data file after creation. This is the default.
930 .. option:: create_on_open=bool
932 If true, don't pre-create files but allow the job's open() to create a file
933 when it's time to do I/O. Default: false -- pre-create all necessary files
936 .. option:: create_only=bool
938 If true, fio will only run the setup phase of the job. If files need to be
939 laid out or updated on disk, only that will be done -- the actual job contents
940 are not executed. Default: false.
942 .. option:: allow_file_create=bool
944 If true, fio is permitted to create files as part of its workload. If this
945 option is false, then fio will error out if
946 the files it needs to use don't already exist. Default: true.
948 .. option:: allow_mounted_write=bool
950 If this isn't set, fio will abort jobs that are destructive (e.g. that write)
951 to what appears to be a mounted device or partition. This should help catch
952 creating inadvertently destructive tests, not realizing that the test will
953 destroy data on the mounted file system. Note that some platforms don't allow
954 writing against a mounted device regardless of this option. Default: false.
956 .. option:: pre_read=bool
958 If this is given, files will be pre-read into memory before starting the
959 given I/O operation. This will also clear the :option:`invalidate` flag,
960 since it is pointless to pre-read and then drop the cache. This will only
961 work for I/O engines that are seek-able, since they allow you to read the
962 same data multiple times. Thus it will not work on non-seekable I/O engines
963 (e.g. network, splice). Default: false.
965 .. option:: unlink=bool
967 Unlink the job files when done. Not the default, as repeated runs of that
968 job would then waste time recreating the file set again and again. Default:
971 .. option:: unlink_each_loop=bool
973 Unlink job files after each iteration or loop. Default: false.
975 .. option:: zonemode=str
980 The :option:`zonerange`, :option:`zonesize`,
981 :option `zonecapacity` and option:`zoneskip`
982 parameters are ignored.
984 I/O happens in a single zone until
985 :option:`zonesize` bytes have been transferred.
986 After that number of bytes has been
987 transferred processing of the next zone
988 starts. :option `zonecapacity` is ignored.
990 Zoned block device mode. I/O happens
991 sequentially in each zone, even if random I/O
992 has been selected. Random I/O happens across
993 all zones instead of being restricted to a
994 single zone. The :option:`zoneskip` parameter
995 is ignored. :option:`zonerange` and
996 :option:`zonesize` must be identical.
997 Trim is handled using a zone reset operation.
998 Trim only considers non-empty sequential write
999 required and sequential write preferred zones.
1001 .. option:: zonerange=int
1003 Size of a single zone. See also :option:`zonesize` and
1006 .. option:: zonesize=int
1008 For :option:`zonemode` =strided, this is the number of bytes to
1009 transfer before skipping :option:`zoneskip` bytes. If this parameter
1010 is smaller than :option:`zonerange` then only a fraction of each zone
1011 with :option:`zonerange` bytes will be accessed. If this parameter is
1012 larger than :option:`zonerange` then each zone will be accessed
1013 multiple times before skipping to the next zone.
1015 For :option:`zonemode` =zbd, this is the size of a single zone. The
1016 :option:`zonerange` parameter is ignored in this mode.
1019 .. option:: zonecapacity=int
1021 For :option:`zonemode` =zbd, this defines the capacity of a single zone,
1022 which is the accessible area starting from the zone start address.
1023 This parameter only applies when using :option:`zonemode` =zbd in
1024 combination with regular block devices. If not specified it defaults to
1025 the zone size. If the target device is a zoned block device, the zone
1026 capacity is obtained from the device information and this option is
1029 .. option:: zoneskip=int
1031 For :option:`zonemode` =strided, the number of bytes to skip after
1032 :option:`zonesize` bytes of data have been transferred. This parameter
1033 must be zero for :option:`zonemode` =zbd.
1035 .. option:: read_beyond_wp=bool
1037 This parameter applies to :option:`zonemode` =zbd only.
1039 Zoned block devices are block devices that consist of multiple zones.
1040 Each zone has a type, e.g. conventional or sequential. A conventional
1041 zone can be written at any offset that is a multiple of the block
1042 size. Sequential zones must be written sequentially. The position at
1043 which a write must occur is called the write pointer. A zoned block
1044 device can be either drive managed, host managed or host aware. For
1045 host managed devices the host must ensure that writes happen
1046 sequentially. Fio recognizes host managed devices and serializes
1047 writes to sequential zones for these devices.
1049 If a read occurs in a sequential zone beyond the write pointer then
1050 the zoned block device will complete the read without reading any data
1051 from the storage medium. Since such reads lead to unrealistically high
1052 bandwidth and IOPS numbers fio only reads beyond the write pointer if
1053 explicitly told to do so. Default: false.
1055 .. option:: max_open_zones=int
1057 A zone of a zoned block device is in the open state when it is partially
1058 written (i.e. not all sectors of the zone have been written). Zoned
1059 block devices may have a limit on the total number of zones that can
1060 be simultaneously in the open state, that is, the number of zones that
1061 can be written to simultaneously. The :option:`max_open_zones` parameter
1062 limits the number of zones to which write commands are issued by all fio
1063 jobs, that is, limits the number of zones that will be in the open
1064 state. This parameter is relevant only if the :option:`zonemode` =zbd is
1065 used. The default value is always equal to maximum number of open zones
1066 of the target zoned block device and a value higher than this limit
1067 cannot be specified by users unless the option
1068 :option:`ignore_zone_limits` is specified. When
1069 :option:`ignore_zone_limits` is specified or the target device has no
1070 limit on the number of zones that can be in an open state,
1071 :option:`max_open_zones` can specify 0 to disable any limit on the
1072 number of zones that can be simultaneously written to by all jobs.
1074 .. option:: job_max_open_zones=int
1076 In the same manner as :option:`max_open_zones`, limit the number of open
1077 zones per fio job, that is, the number of zones that a single job can
1078 simultaneously write to. A value of zero indicates no limit.
1081 .. option:: ignore_zone_limits=bool
1083 If this option is used, fio will ignore the maximum number of open
1084 zones limit of the zoned block device in use, thus allowing the
1085 option :option:`max_open_zones` value to be larger than the device
1086 reported limit. Default: false.
1088 .. option:: zone_reset_threshold=float
1090 A number between zero and one that indicates the ratio of written bytes
1091 in the zones with write pointers in the IO range to the size of the IO
1092 range. When current ratio is above this ratio, zones are reset
1093 periodically as :option:`zone_reset_frequency` specifies. If there are
1094 multiple jobs when using this option, the IO range for all write jobs
1097 .. option:: zone_reset_frequency=float
1099 A number between zero and one that indicates how often a zone reset
1100 should be issued if the zone reset threshold has been exceeded. A zone
1101 reset is submitted after each (1 / zone_reset_frequency) write
1102 requests. This and the previous parameter can be used to simulate
1103 garbage collection activity.
1109 .. option:: direct=bool
1111 If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that
1112 OpenBSD and ZFS on Solaris don't support direct I/O. On Windows the synchronous
1113 ioengines don't support direct I/O. Default: false.
1115 .. option:: buffered=bool
1117 If value is true, use buffered I/O. This is the opposite of the
1118 :option:`direct` option. Defaults to true.
1120 .. option:: readwrite=str, rw=str
1122 Type of I/O pattern. Accepted values are:
1129 Sequential trims (Linux block devices and SCSI
1130 character devices only).
1136 Random trims (Linux block devices and SCSI
1137 character devices only).
1139 Sequential mixed reads and writes.
1141 Random mixed reads and writes.
1143 Sequential trim+write sequences. Blocks will be trimmed first,
1144 then the same blocks will be written to. So if ``io_size=64K``
1145 is specified, Fio will trim a total of 64K bytes and also
1146 write 64K bytes on the same trimmed blocks. This behaviour
1147 will be consistent with ``number_ios`` or other Fio options
1148 limiting the total bytes or number of I/O's.
1150 Like trimwrite, but uses random offsets rather
1151 than sequential writes.
1153 Fio defaults to read if the option is not specified. For the mixed I/O
1154 types, the default is to split them 50/50. For certain types of I/O the
1155 result may still be skewed a bit, since the speed may be different.
1157 It is possible to specify the number of I/Os to do before getting a new
1158 offset by appending ``:<nr>`` to the end of the string given. For a
1159 random read, it would look like ``rw=randread:8`` for passing in an offset
1160 modifier with a value of 8. If the suffix is used with a sequential I/O
1161 pattern, then the *<nr>* value specified will be **added** to the generated
1162 offset for each I/O turning sequential I/O into sequential I/O with holes.
1163 For instance, using ``rw=write:4k`` will skip 4k for every write. Also see
1164 the :option:`rw_sequencer` option.
1166 .. option:: rw_sequencer=str
1168 If an offset modifier is given by appending a number to the ``rw=<str>``
1169 line, then this option controls how that number modifies the I/O offset
1170 being generated. Accepted values are:
1173 Generate sequential offset.
1175 Generate the same offset.
1177 ``sequential`` is only useful for random I/O, where fio would normally
1178 generate a new random offset for every I/O. If you append e.g. 8 to
1179 randread, i.e. ``rw=randread:8`` you would get a new random offset for
1180 every 8 I/Os. The result would be a sequence of 8 sequential offsets
1181 with a random starting point. However this behavior may change if a
1182 sequential I/O reaches end of the file. As sequential I/O is already
1183 sequential, setting ``sequential`` for that would not result in any
1184 difference. ``identical`` behaves in a similar fashion, except it sends
1185 the same offset 8 number of times before generating a new offset.
1190 rw_sequencer=sequential
1193 The generated sequence of offsets will look like this:
1194 4k, 8k, 12k, 16k, 20k, 24k, 28k, 32k, 92k, 96k, 100k, 104k, 108k,
1195 112k, 116k, 120k, 48k, 52k ...
1200 rw_sequencer=identical
1203 The generated sequence of offsets will look like this:
1204 4k, 4k, 4k, 4k, 4k, 4k, 4k, 4k, 92k, 92k, 92k, 92k, 92k, 92k, 92k, 92k,
1207 .. option:: unified_rw_reporting=str
1209 Fio normally reports statistics on a per data direction basis, meaning that
1210 reads, writes, and trims are accounted and reported separately. This option
1211 determines whether fio reports the results normally, summed together, or as
1213 Accepted values are:
1216 Normal statistics reporting.
1219 Statistics are summed per data direction and reported together.
1222 Statistics are reported normally, followed by the mixed statistics.
1225 Backward-compatible alias for **none**.
1228 Backward-compatible alias for **mixed**.
1233 .. option:: randrepeat=bool
1235 Seed all random number generators in a predictable way so the pattern
1236 is repeatable across runs. Default: true.
1238 .. option:: allrandrepeat=bool
1240 Alias for :option:`randrepeat`. Default: true.
1242 .. option:: randseed=int
1244 Seed the random number generators based on this seed value, to be able to
1245 control what sequence of output is being generated. If not set, the random
1246 sequence depends on the :option:`randrepeat` setting.
1248 .. option:: fallocate=str
1250 Whether pre-allocation is performed when laying down files.
1251 Accepted values are:
1254 Do not pre-allocate space.
1257 Use a platform's native pre-allocation call but fall back to
1258 **none** behavior if it fails/is not implemented.
1261 Pre-allocate via :manpage:`posix_fallocate(3)`.
1264 Pre-allocate via :manpage:`fallocate(2)` with
1265 FALLOC_FL_KEEP_SIZE set.
1268 Extend file to final size via :manpage:`ftruncate(2)`
1269 instead of allocating.
1272 Backward-compatible alias for **none**.
1275 Backward-compatible alias for **posix**.
1277 May not be available on all supported platforms. **keep** is only available
1278 on Linux. If using ZFS on Solaris this cannot be set to **posix**
1279 because ZFS doesn't support pre-allocation. Default: **native** if any
1280 pre-allocation methods except **truncate** are available, **none** if not.
1282 Note that using **truncate** on Windows will interact surprisingly
1283 with non-sequential write patterns. When writing to a file that has
1284 been extended by setting the end-of-file information, Windows will
1285 backfill the unwritten portion of the file up to that offset with
1286 zeroes before issuing the new write. This means that a single small
1287 write to the end of an extended file will stall until the entire
1288 file has been filled with zeroes.
1290 .. option:: fadvise_hint=str
1292 Use :manpage:`posix_fadvise(2)` or :manpage:`posix_fadvise(2)` to
1293 advise the kernel on what I/O patterns are likely to be issued.
1294 Accepted values are:
1297 Backwards-compatible hint for "no hint".
1300 Backwards compatible hint for "advise with fio workload type". This
1301 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1302 for a sequential workload.
1305 Advise using **FADV_SEQUENTIAL**.
1308 Advise using **FADV_RANDOM**.
1311 Advise using **FADV_NOREUSE**. This may be a no-op on older Linux
1312 kernels. Since Linux 6.3, it provides a hint to the LRU algorithm.
1313 See the :manpage:`posix_fadvise(2)` man page.
1315 .. option:: write_hint=str
1317 Use :manpage:`fcntl(2)` to advise the kernel what life time to expect
1318 from a write. Only supported on Linux, as of version 4.13. Accepted
1322 No particular life time associated with this file.
1325 Data written to this file has a short life time.
1328 Data written to this file has a medium life time.
1331 Data written to this file has a long life time.
1334 Data written to this file has a very long life time.
1336 The values are all relative to each other, and no absolute meaning
1337 should be associated with them.
1339 .. option:: offset=int
1341 Start I/O at the provided offset in the file, given as either a fixed size in
1342 bytes, zones or a percentage. If a percentage is given, the generated offset will be
1343 aligned to the minimum ``blocksize`` or to the value of ``offset_align`` if
1344 provided. Data before the given offset will not be touched. This
1345 effectively caps the file size at `real_size - offset`. Can be combined with
1346 :option:`size` to constrain the start and end range of the I/O workload.
1347 A percentage can be specified by a number between 1 and 100 followed by '%',
1348 for example, ``offset=20%`` to specify 20%. In ZBD mode, value can be set as
1349 number of zones using 'z'.
1351 .. option:: offset_align=int
1353 If set to non-zero value, the byte offset generated by a percentage ``offset``
1354 is aligned upwards to this value. Defaults to 0 meaning that a percentage
1355 offset is aligned to the minimum block size.
1357 .. option:: offset_increment=int
1359 If this is provided, then the real offset becomes `offset + offset_increment
1360 * thread_number`, where the thread number is a counter that starts at 0 and
1361 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1362 specified). This option is useful if there are several jobs which are
1363 intended to operate on a file in parallel disjoint segments, with even
1364 spacing between the starting points. Percentages can be used for this option.
1365 If a percentage is given, the generated offset will be aligned to the minimum
1366 ``blocksize`` or to the value of ``offset_align`` if provided. In ZBD mode, value can
1367 also be set as number of zones using 'z'.
1369 .. option:: number_ios=int
1371 Fio will normally perform I/Os until it has exhausted the size of the region
1372 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1373 condition). With this setting, the range/size can be set independently of
1374 the number of I/Os to perform. When fio reaches this number, it will exit
1375 normally and report status. Note that this does not extend the amount of I/O
1376 that will be done, it will only stop fio if this condition is met before
1377 other end-of-job criteria.
1379 .. option:: fsync=int
1381 If writing to a file, issue an :manpage:`fsync(2)` (or its equivalent) of
1382 the dirty data for every number of blocks given. For example, if you give 32
1383 as a parameter, fio will sync the file after every 32 writes issued. If fio is
1384 using non-buffered I/O, we may not sync the file. The exception is the sg
1385 I/O engine, which synchronizes the disk cache anyway. Defaults to 0, which
1386 means fio does not periodically issue and wait for a sync to complete. Also
1387 see :option:`end_fsync` and :option:`fsync_on_close`.
1389 .. option:: fdatasync=int
1391 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
1392 not metadata blocks. In Windows, DragonFlyBSD or OSX there is no
1393 :manpage:`fdatasync(2)` so this falls back to using :manpage:`fsync(2)`.
1394 Defaults to 0, which means fio does not periodically issue and wait for a
1395 data-only sync to complete.
1397 .. option:: write_barrier=int
1399 Make every `N-th` write a barrier write.
1401 .. option:: sync_file_range=str:int
1403 Use :manpage:`sync_file_range(2)` for every `int` number of write
1404 operations. Fio will track range of writes that have happened since the last
1405 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1408 SYNC_FILE_RANGE_WAIT_BEFORE
1410 SYNC_FILE_RANGE_WRITE
1412 SYNC_FILE_RANGE_WAIT_AFTER
1414 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1415 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1416 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1419 .. option:: overwrite=bool
1421 If true, writes to a file will always overwrite existing data. If the file
1422 doesn't already exist, it will be created before the write phase begins. If
1423 the file exists and is large enough for the specified write phase, nothing
1424 will be done. Default: false.
1426 .. option:: end_fsync=bool
1428 If true, :manpage:`fsync(2)` file contents when a write stage has completed.
1431 .. option:: fsync_on_close=bool
1433 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
1434 from :option:`end_fsync` in that it will happen on every file close, not
1435 just at the end of the job. Default: false.
1437 .. option:: rwmixread=int
1439 Percentage of a mixed workload that should be reads. Default: 50.
1441 .. option:: rwmixwrite=int
1443 Percentage of a mixed workload that should be writes. If both
1444 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1445 add up to 100%, the latter of the two will be used to override the
1446 first. This may interfere with a given rate setting, if fio is asked to
1447 limit reads or writes to a certain rate. If that is the case, then the
1448 distribution may be skewed. Default: 50.
1450 .. option:: random_distribution=str:float[:float][,str:float][,str:float]
1452 By default, fio will use a completely uniform random distribution when asked
1453 to perform random I/O. Sometimes it is useful to skew the distribution in
1454 specific ways, ensuring that some parts of the data is more hot than others.
1455 fio includes the following distribution models:
1458 Uniform random distribution
1467 Normal (Gaussian) distribution
1470 Zoned random distribution
1473 Zone absolute random distribution
1475 When using a **zipf** or **pareto** distribution, an input value is also
1476 needed to define the access pattern. For **zipf**, this is the `Zipf
1477 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
1478 program, :command:`fio-genzipf`, that can be used visualize what the given input
1479 values will yield in terms of hit rates. If you wanted to use **zipf** with
1480 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1481 option. If a non-uniform model is used, fio will disable use of the random
1482 map. For the **normal** distribution, a normal (Gaussian) deviation is
1483 supplied as a value between 0 and 100.
1485 The second, optional float is allowed for **pareto**, **zipf** and **normal** distributions.
1486 It allows one to set base of distribution in non-default place, giving more control
1487 over most probable outcome. This value is in range [0-1] which maps linearly to
1488 range of possible random values.
1489 Defaults are: random for **pareto** and **zipf**, and 0.5 for **normal**.
1490 If you wanted to use **zipf** with a `theta` of 1.2 centered on 1/4 of allowed value range,
1491 you would use ``random_distribution=zipf:1.2:0.25``.
1493 For a **zoned** distribution, fio supports specifying percentages of I/O
1494 access that should fall within what range of the file or device. For
1495 example, given a criteria of:
1497 * 60% of accesses should be to the first 10%
1498 * 30% of accesses should be to the next 20%
1499 * 8% of accesses should be to the next 30%
1500 * 2% of accesses should be to the next 40%
1502 we can define that through zoning of the random accesses. For the above
1503 example, the user would do::
1505 random_distribution=zoned:60/10:30/20:8/30:2/40
1507 A **zoned_abs** distribution works exactly like the **zoned**, except
1508 that it takes absolute sizes. For example, let's say you wanted to
1509 define access according to the following criteria:
1511 * 60% of accesses should be to the first 20G
1512 * 30% of accesses should be to the next 100G
1513 * 10% of accesses should be to the next 500G
1515 we can define an absolute zoning distribution with:
1517 random_distribution=zoned_abs=60/20G:30/100G:10/500g
1519 For both **zoned** and **zoned_abs**, fio supports defining up to
1522 Similarly to how :option:`bssplit` works for setting ranges and
1523 percentages of block sizes. Like :option:`bssplit`, it's possible to
1524 specify separate zones for reads, writes, and trims. If just one set
1525 is given, it'll apply to all of them. This goes for both **zoned**
1526 **zoned_abs** distributions.
1528 .. option:: percentage_random=int[,int][,int]
1530 For a random workload, set how big a percentage should be random. This
1531 defaults to 100%, in which case the workload is fully random. It can be set
1532 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1533 sequential. Any setting in between will result in a random mix of sequential
1534 and random I/O, at the given percentages. Comma-separated values may be
1535 specified for reads, writes, and trims as described in :option:`blocksize`.
1537 .. option:: norandommap
1539 Normally fio will cover every block of the file when doing random I/O. If
1540 this option is given, fio will just get a new random offset without looking
1541 at past I/O history. This means that some blocks may not be read or written,
1542 and that some blocks may be read/written more than once. If this option is
1543 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1544 only intact blocks are verified, i.e., partially-overwritten blocks are
1545 ignored. With an async I/O engine and an I/O depth > 1, it is possible for
1546 the same block to be overwritten, which can cause verification errors. Either
1547 do not use norandommap in this case, or also use the lfsr random generator.
1549 .. option:: softrandommap=bool
1551 See :option:`norandommap`. If fio runs with the random block map enabled and
1552 it fails to allocate the map, if this option is set it will continue without
1553 a random block map. As coverage will not be as complete as with random maps,
1554 this option is disabled by default.
1556 .. option:: random_generator=str
1558 Fio supports the following engines for generating I/O offsets for random I/O:
1561 Strong 2^88 cycle random number generator.
1563 Linear feedback shift register generator.
1565 Strong 64-bit 2^258 cycle random number generator.
1567 **tausworthe** is a strong random number generator, but it requires tracking
1568 on the side if we want to ensure that blocks are only read or written
1569 once. **lfsr** guarantees that we never generate the same offset twice, and
1570 it's also less computationally expensive. It's not a true random generator,
1571 however, though for I/O purposes it's typically good enough. **lfsr** only
1572 works with single block sizes, not with workloads that use multiple block
1573 sizes. If used with such a workload, fio may read or write some blocks
1574 multiple times. The default value is **tausworthe**, unless the required
1575 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1576 selected automatically.
1582 .. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1584 The block size in bytes used for I/O units. Default: 4096. A single value
1585 applies to reads, writes, and trims. Comma-separated values may be
1586 specified for reads, writes, and trims. A value not terminated in a comma
1587 applies to subsequent types.
1592 means 256k for reads, writes and trims.
1595 means 8k for reads, 32k for writes and trims.
1598 means 8k for reads, 32k for writes, and default for trims.
1601 means default for reads, 8k for writes and trims.
1604 means default for reads, 8k for writes, and default for trims.
1606 .. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1608 A range of block sizes in bytes for I/O units. The issued I/O unit will
1609 always be a multiple of the minimum size, unless
1610 :option:`blocksize_unaligned` is set.
1612 Comma-separated ranges may be specified for reads, writes, and trims as
1613 described in :option:`blocksize`.
1615 Example: ``bsrange=1k-4k,2k-8k``.
1617 .. option:: bssplit=str[,str][,str]
1619 Sometimes you want even finer grained control of the block sizes
1620 issued, not just an even split between them. This option allows you to
1621 weight various block sizes, so that you are able to define a specific
1622 amount of block sizes issued. The format for this option is::
1624 bssplit=blocksize/percentage:blocksize/percentage
1626 for as many block sizes as needed. So if you want to define a workload
1627 that has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would
1630 bssplit=4k/10:64k/50:32k/40
1632 Ordering does not matter. If the percentage is left blank, fio will
1633 fill in the remaining values evenly. So a bssplit option like this one::
1635 bssplit=4k/50:1k/:32k/
1637 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always
1638 add up to 100, if bssplit is given a range that adds up to more, it
1641 Comma-separated values may be specified for reads, writes, and trims as
1642 described in :option:`blocksize`.
1644 If you want a workload that has 50% 2k reads and 50% 4k reads, while
1645 having 90% 4k writes and 10% 8k writes, you would specify::
1647 bssplit=2k/50:4k/50,4k/90:8k/10
1649 Fio supports defining up to 64 different weights for each data
1652 .. option:: blocksize_unaligned, bs_unaligned
1654 If set, fio will issue I/O units with any size within
1655 :option:`blocksize_range`, not just multiples of the minimum size. This
1656 typically won't work with direct I/O, as that normally requires sector
1659 .. option:: bs_is_seq_rand=bool
1661 If this option is set, fio will use the normal read,write blocksize settings
1662 as sequential,random blocksize settings instead. Any random read or write
1663 will use the WRITE blocksize settings, and any sequential read or write will
1664 use the READ blocksize settings.
1666 .. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1668 Boundary to which fio will align random I/O units. Default:
1669 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1670 I/O, though it usually depends on the hardware block size. This option is
1671 mutually exclusive with using a random map for files, so it will turn off
1672 that option. Comma-separated values may be specified for reads, writes, and
1673 trims as described in :option:`blocksize`.
1679 .. option:: zero_buffers
1681 Initialize buffers with all zeros. Default: fill buffers with random data.
1683 .. option:: refill_buffers
1685 If this option is given, fio will refill the I/O buffers on every
1686 submit. Only makes sense if :option:`zero_buffers` isn't specified,
1687 naturally. Defaults to being unset i.e., the buffer is only filled at
1688 init time and the data in it is reused when possible but if any of
1689 :option:`verify`, :option:`buffer_compress_percentage` or
1690 :option:`dedupe_percentage` are enabled then `refill_buffers` is also
1691 automatically enabled.
1693 .. option:: scramble_buffers=bool
1695 If :option:`refill_buffers` is too costly and the target is using data
1696 deduplication, then setting this option will slightly modify the I/O buffer
1697 contents to defeat normal de-dupe attempts. This is not enough to defeat
1698 more clever block compression attempts, but it will stop naive dedupe of
1699 blocks. Default: true.
1701 .. option:: buffer_compress_percentage=int
1703 If this is set, then fio will attempt to provide I/O buffer content
1704 (on WRITEs) that compresses to the specified level. Fio does this by
1705 providing a mix of random data followed by fixed pattern data. The
1706 fixed pattern is either zeros, or the pattern specified by
1707 :option:`buffer_pattern`. If the `buffer_pattern` option is used, it
1708 might skew the compression ratio slightly. Setting
1709 `buffer_compress_percentage` to a value other than 100 will also
1710 enable :option:`refill_buffers` in order to reduce the likelihood that
1711 adjacent blocks are so similar that they over compress when seen
1712 together. See :option:`buffer_compress_chunk` for how to set a finer or
1713 coarser granularity for the random/fixed data region. Defaults to unset
1714 i.e., buffer data will not adhere to any compression level.
1716 .. option:: buffer_compress_chunk=int
1718 This setting allows fio to manage how big the random/fixed data region
1719 is when using :option:`buffer_compress_percentage`. When
1720 `buffer_compress_chunk` is set to some non-zero value smaller than the
1721 block size, fio can repeat the random/fixed region throughout the I/O
1722 buffer at the specified interval (which particularly useful when
1723 bigger block sizes are used for a job). When set to 0, fio will use a
1724 chunk size that matches the block size resulting in a single
1725 random/fixed region within the I/O buffer. Defaults to 512. When the
1726 unit is omitted, the value is interpreted in bytes.
1728 .. option:: buffer_pattern=str
1730 If set, fio will fill the I/O buffers with this pattern or with the contents
1731 of a file. If not set, the contents of I/O buffers are defined by the other
1732 options related to buffer contents. The setting can be any pattern of bytes,
1733 and can be prefixed with 0x for hex values. It may also be a string, where
1734 the string must then be wrapped with ``""``. Or it may also be a filename,
1735 where the filename must be wrapped with ``''`` in which case the file is
1736 opened and read. Note that not all the file contents will be read if that
1737 would cause the buffers to overflow. So, for example::
1739 buffer_pattern='filename'
1743 buffer_pattern="abcd"
1751 buffer_pattern=0xdeadface
1753 Also you can combine everything together in any order::
1755 buffer_pattern=0xdeadface"abcd"-12'filename'
1757 .. option:: dedupe_percentage=int
1759 If set, fio will generate this percentage of identical buffers when
1760 writing. These buffers will be naturally dedupable. The contents of the
1761 buffers depend on what other buffer compression settings have been set. It's
1762 possible to have the individual buffers either fully compressible, or not at
1763 all -- this option only controls the distribution of unique buffers. Setting
1764 this option will also enable :option:`refill_buffers` to prevent every buffer
1767 .. option:: dedupe_mode=str
1769 If ``dedupe_percentage=<int>`` is given, then this option controls how fio
1770 generates the dedupe buffers.
1773 Generate dedupe buffers by repeating previous writes
1775 Generate dedupe buffers from working set
1777 ``repeat`` is the default option for fio. Dedupe buffers are generated
1778 by repeating previous unique write.
1780 ``working_set`` is a more realistic workload.
1781 With ``working_set``, ``dedupe_working_set_percentage=<int>`` should be provided.
1782 Given that, fio will use the initial unique write buffers as its working set.
1783 Upon deciding to dedupe, fio will randomly choose a buffer from the working set.
1784 Note that by using ``working_set`` the dedupe percentage will converge
1785 to the desired over time while ``repeat`` maintains the desired percentage
1788 .. option:: dedupe_working_set_percentage=int
1790 If ``dedupe_mode=<str>`` is set to ``working_set``, then this controls
1791 the percentage of size of the file or device used as the buffers
1792 fio will choose to generate the dedupe buffers from
1794 Note that size needs to be explicitly provided and only 1 file per
1797 .. option:: dedupe_global=bool
1799 This controls whether the deduplication buffers will be shared amongst
1800 all jobs that have this option set. The buffers are spread evenly between
1803 .. option:: invalidate=bool
1805 Invalidate the buffer/page cache parts of the files to be used prior to
1806 starting I/O if the platform and file type support it. Defaults to true.
1807 This will be ignored if :option:`pre_read` is also specified for the
1810 .. option:: sync=str
1812 Whether, and what type, of synchronous I/O to use for writes. The allowed
1816 Do not use synchronous IO, the default.
1822 Use synchronous file IO. For the majority of I/O engines,
1823 this means using O_SYNC.
1829 Use synchronous data IO. For the majority of I/O engines,
1830 this means using O_DSYNC.
1833 .. option:: iomem=str, mem=str
1835 Fio can use various types of memory as the I/O unit buffer. The allowed
1839 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1843 Use shared memory as the buffers. Allocated through
1844 :manpage:`shmget(2)`.
1847 Same as shm, but use huge pages as backing.
1850 Use :manpage:`mmap(2)` to allocate buffers. May either be anonymous memory, or can
1851 be file backed if a filename is given after the option. The format
1852 is `mem=mmap:/path/to/file`.
1855 Use a memory mapped huge file as the buffer backing. Append filename
1856 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1859 Same as mmap, but use a MMAP_SHARED mapping.
1862 Use GPU memory as the buffers for GPUDirect RDMA benchmark.
1863 The :option:`ioengine` must be `rdma`.
1865 The area allocated is a function of the maximum allowed bs size for the job,
1866 multiplied by the I/O depth given. Note that for **shmhuge** and
1867 **mmaphuge** to work, the system must have free huge pages allocated. This
1868 can normally be checked and set by reading/writing
1869 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1870 is 2 or 4MiB in size depending on the platform. So to calculate the
1871 number of huge pages you need for a given job file, add up the I/O
1872 depth of all jobs (normally one unless :option:`iodepth` is used) and
1873 multiply by the maximum bs set. Then divide that number by the huge
1874 page size. You can see the size of the huge pages in
1875 :file:`/proc/meminfo`. If no huge pages are allocated by having a
1876 non-zero number in `nr_hugepages`, using **mmaphuge** or **shmhuge**
1877 will fail. Also see :option:`hugepage-size`.
1879 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1880 should point there. So if it's mounted in :file:`/huge`, you would use
1881 `mem=mmaphuge:/huge/somefile`.
1883 .. option:: iomem_align=int, mem_align=int
1885 This indicates the memory alignment of the I/O memory buffers. Note that
1886 the given alignment is applied to the first I/O unit buffer, if using
1887 :option:`iodepth` the alignment of the following buffers are given by the
1888 :option:`bs` used. In other words, if using a :option:`bs` that is a
1889 multiple of the page sized in the system, all buffers will be aligned to
1890 this value. If using a :option:`bs` that is not page aligned, the alignment
1891 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1894 .. option:: hugepage-size=int
1896 Defines the size of a huge page. Must at least be equal to the system
1897 setting, see :file:`/proc/meminfo` and
1898 :file:`/sys/kernel/mm/hugepages/`. Defaults to 2 or 4MiB depending on
1899 the platform. Should probably always be a multiple of megabytes, so
1900 using ``hugepage-size=Xm`` is the preferred way to set this to avoid
1901 setting a non-pow-2 bad value.
1903 .. option:: lockmem=int
1905 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1906 simulate a smaller amount of memory. The amount specified is per worker.
1912 .. option:: size=int
1914 The total size of file I/O for each thread of this job. Fio will run until
1915 this many bytes has been transferred, unless runtime is altered by other means
1916 such as (1) :option:`runtime`, (2) :option:`io_size` (3) :option:`number_ios`,
1917 (4) gaps/holes while doing I/O's such as ``rw=read:16K``, or (5) sequential
1918 I/O reaching end of the file which is possible when :option:`percentage_random`
1920 Fio will divide this size between the available files determined by options
1921 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1922 specified by the job. If the result of division happens to be 0, the size is
1923 set to the physical size of the given files or devices if they exist.
1924 If this option is not specified, fio will use the full size of the given
1925 files or devices. If the files do not exist, size must be given. It is also
1926 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1927 given, fio will use 20% of the full size of the given files or devices.
1928 In ZBD mode, value can also be set as number of zones using 'z'.
1929 Can be combined with :option:`offset` to constrain the start and end range
1930 that I/O will be done within.
1932 .. option:: io_size=int, io_limit=int
1934 Normally fio operates within the region set by :option:`size`, which means
1935 that the :option:`size` option sets both the region and size of I/O to be
1936 performed. Sometimes that is not what you want. With this option, it is
1937 possible to define just the amount of I/O that fio should do. For instance,
1938 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1939 will perform I/O within the first 20GiB but exit when 5GiB have been
1940 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1941 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1942 the 0..20GiB region.
1944 .. option:: filesize=irange(int)
1946 Individual file sizes. May be a range, in which case fio will select sizes for
1947 files at random within the given range. If not given, each created file is the
1948 same size. This option overrides :option:`size` in terms of file size, i.e. if
1949 :option:`filesize` is specified then :option:`size` becomes merely the default
1950 for :option:`io_size` and has no effect at all if :option:`io_size` is set
1953 .. option:: file_append=bool
1955 Perform I/O after the end of the file. Normally fio will operate within the
1956 size of a file. If this option is set, then fio will append to the file
1957 instead. This has identical behavior to setting :option:`offset` to the size
1958 of a file. This option is ignored on non-regular files.
1960 .. option:: fill_device=bool, fill_fs=bool
1962 Sets size to something really large and waits for ENOSPC (no space left on
1963 device) or EDQUOT (disk quota exceeded)
1964 as the terminating condition. Only makes sense with sequential
1965 write. For a read workload, the mount point will be filled first then I/O
1966 started on the result. This option doesn't make sense if operating on a raw
1967 device node, since the size of that is already known by the file system.
1968 Additionally, writing beyond end-of-device will not return ENOSPC there.
1974 .. option:: ioengine=str
1976 Defines how the job issues I/O to the file. The following types are defined:
1979 Basic :manpage:`read(2)` or :manpage:`write(2)`
1980 I/O. :manpage:`lseek(2)` is used to position the I/O location.
1981 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
1984 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1985 all supported operating systems except for Windows.
1988 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
1989 queuing by coalescing adjacent I/Os into a single submission.
1992 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1995 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1998 Fast Linux native asynchronous I/O. Supports async IO
1999 for both direct and buffered IO.
2000 This engine defines engine specific options.
2003 Fast Linux native asynchronous I/O for pass through commands.
2004 This engine defines engine specific options.
2007 Linux native asynchronous I/O. Note that Linux may only support
2008 queued behavior with non-buffered I/O (set ``direct=1`` or
2010 This engine defines engine specific options.
2013 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
2014 :manpage:`aio_write(3)`.
2017 Solaris native asynchronous I/O.
2020 Windows native asynchronous I/O. Default on Windows.
2023 File is memory mapped with :manpage:`mmap(2)` and data copied
2024 to/from using :manpage:`memcpy(3)`.
2027 :manpage:`splice(2)` is used to transfer the data and
2028 :manpage:`vmsplice(2)` to transfer data from user space to the
2032 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
2033 ioctl, or if the target is an sg character device we use
2034 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
2035 I/O. Requires :option:`filename` option to specify either block or
2036 character devices. This engine supports trim operations.
2037 The sg engine includes engine specific options.
2040 Read, write, trim and ZBC/ZAC operations to a zoned
2041 block device using libzbc library. The target can be
2042 either an SG character device or a block device file.
2045 Doesn't transfer any data, just pretends to. This is mainly used to
2046 exercise fio itself and for debugging/testing purposes.
2049 Transfer over the network to given ``host:port``. Depending on the
2050 :option:`protocol` used, the :option:`hostname`, :option:`port`,
2051 :option:`listen` and :option:`filename` options are used to specify
2052 what sort of connection to make, while the :option:`protocol` option
2053 determines which protocol will be used. This engine defines engine
2057 Like **net**, but uses :manpage:`splice(2)` and
2058 :manpage:`vmsplice(2)` to map data and send/receive.
2059 This engine defines engine specific options.
2062 Doesn't transfer any data, but burns CPU cycles according to the
2063 :option:`cpuload`, :option:`cpuchunks` and :option:`cpumode` options.
2064 Setting :option:`cpuload`\=85 will cause that job to do nothing but burn 85%
2065 of the CPU. In case of SMP machines, use :option:`numjobs`\=<nr_of_cpu>
2066 to get desired CPU usage, as the cpuload only loads a
2067 single CPU at the desired rate. A job never finishes unless there is
2068 at least one non-cpuio job.
2069 Setting :option:`cpumode`\=qsort replace the default noop instructions loop
2070 by a qsort algorithm to consume more energy.
2073 The RDMA I/O engine supports both RDMA memory semantics
2074 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
2075 InfiniBand, RoCE and iWARP protocols. This engine defines engine
2079 I/O engine that does regular fallocate to simulate data transfer as
2083 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
2086 does fallocate(,mode = 0).
2089 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
2092 I/O engine that sends :manpage:`ftruncate(2)` operations in response
2093 to write (DDIR_WRITE) events. Each ftruncate issued sets the file's
2094 size to the current block offset. :option:`blocksize` is ignored.
2097 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
2098 defragment activity in request to DDIR_WRITE event.
2101 I/O engine supporting direct access to Ceph Reliable Autonomic
2102 Distributed Object Store (RADOS) via librados. This ioengine
2103 defines engine specific options.
2106 I/O engine supporting direct access to Ceph Rados Block Devices
2107 (RBD) via librbd without the need to use the kernel rbd driver. This
2108 ioengine defines engine specific options.
2111 I/O engine supporting GET/PUT requests over HTTP(S) with libcurl to
2112 a WebDAV or S3 endpoint. This ioengine defines engine specific options.
2114 This engine only supports direct IO of iodepth=1; you need to scale this
2115 via numjobs. blocksize defines the size of the objects to be created.
2117 TRIM is translated to object deletion.
2120 Using GlusterFS libgfapi sync interface to direct access to
2121 GlusterFS volumes without having to go through FUSE. This ioengine
2122 defines engine specific options.
2125 Using GlusterFS libgfapi async interface to direct access to
2126 GlusterFS volumes without having to go through FUSE. This ioengine
2127 defines engine specific options.
2130 Read and write through Hadoop (HDFS). The :option:`filename` option
2131 is used to specify host,port of the hdfs name-node to connect. This
2132 engine interprets offsets a little differently. In HDFS, files once
2133 created cannot be modified so random writes are not possible. To
2134 imitate this the libhdfs engine expects a bunch of small files to be
2135 created over HDFS and will randomly pick a file from them
2136 based on the offset generated by fio backend (see the example
2137 job file to create such files, use ``rw=write`` option). Please
2138 note, it may be necessary to set environment variables to work
2139 with HDFS/libhdfs properly. Each job uses its own connection to
2143 Read, write and erase an MTD character device (e.g.,
2144 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
2145 underlying device type, the I/O may have to go in a certain pattern,
2146 e.g., on NAND, writing sequentially to erase blocks and discarding
2147 before overwriting. The `trimwrite` mode works well for this
2151 Read and write using device DAX to a persistent memory device (e.g.,
2152 /dev/dax0.0) through the PMDK libpmem library.
2155 Prefix to specify loading an external I/O engine object file. Append
2156 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
2157 ioengine :file:`foo.o` in :file:`/tmp`. The path can be either
2158 absolute or relative. See :file:`engines/skeleton_external.c` for
2159 details of writing an external I/O engine.
2162 Simply create the files and do no I/O to them. You still need to
2163 set `filesize` so that all the accounting still occurs, but no
2164 actual I/O will be done other than creating the file.
2167 Simply do stat() and do no I/O to the file. You need to set 'filesize'
2168 and 'nrfiles', so that files will be created.
2169 This engine is to measure file lookup and meta data access.
2172 Simply delete the files by unlink() and do no I/O to them. You need to set 'filesize'
2173 and 'nrfiles', so that the files will be created.
2174 This engine is to measure file delete.
2177 Read and write using mmap I/O to a file on a filesystem
2178 mounted with DAX on a persistent memory device through the PMDK
2182 Synchronous read and write using DDN's Infinite Memory Engine (IME).
2183 This engine is very basic and issues calls to IME whenever an IO is
2187 Synchronous read and write using DDN's Infinite Memory Engine (IME).
2188 This engine uses iovecs and will try to stack as much IOs as possible
2189 (if the IOs are "contiguous" and the IO depth is not exceeded)
2190 before issuing a call to IME.
2193 Asynchronous read and write using DDN's Infinite Memory Engine (IME).
2194 This engine will try to stack as much IOs as possible by creating
2195 requests for IME. FIO will then decide when to commit these requests.
2198 Read and write iscsi lun with libiscsi.
2201 Read and write a Network Block Device (NBD).
2204 I/O engine supporting libcufile synchronous access to nvidia-fs and a
2205 GPUDirect Storage-supported filesystem. This engine performs
2206 I/O without transferring buffers between user-space and the kernel,
2207 unless :option:`verify` is set or :option:`cuda_io` is `posix`.
2208 :option:`iomem` must not be `cudamalloc`. This ioengine defines
2209 engine specific options.
2212 I/O engine supporting asynchronous read and write operations to the
2213 DAOS File System (DFS) via libdfs.
2216 I/O engine supporting asynchronous read and write operations to
2217 NFS filesystems from userspace via libnfs. This is useful for
2218 achieving higher concurrency and thus throughput than is possible
2222 Execute 3rd party tools. Could be used to perform monitoring during jobs runtime.
2225 I/O engine using the xNVMe C API, for NVMe devices. The xnvme engine provides
2226 flexibility to access GNU/Linux Kernel NVMe driver via libaio, IOCTLs, io_uring,
2227 the SPDK NVMe driver, or your own custom NVMe driver. The xnvme engine includes
2228 engine specific options. (See https://xnvme.io).
2231 Use the libblkio library
2232 (https://gitlab.com/libblkio/libblkio). The specific
2233 *driver* to use must be set using
2234 :option:`libblkio_driver`. If
2235 :option:`mem`/:option:`iomem` is not specified, memory
2236 allocation is delegated to libblkio (and so is
2237 guaranteed to work with the selected *driver*). One
2238 libblkio instance is used per process, so all jobs
2239 setting option :option:`thread` will share a single
2240 instance (with one queue per thread) and must specify
2241 compatible options. Note that some drivers don't allow
2242 several instances to access the same device or file
2243 simultaneously, but allow it for threads.
2245 I/O engine specific parameters
2246 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2248 In addition, there are some parameters which are only valid when a specific
2249 :option:`ioengine` is in use. These are used identically to normal parameters,
2250 with the caveat that when used on the command line, they must come after the
2251 :option:`ioengine` that defines them is selected.
2253 .. option:: cmdprio_percentage=int[,int] : [io_uring] [libaio]
2255 Set the percentage of I/O that will be issued with the highest priority.
2256 Default: 0. A single value applies to reads and writes. Comma-separated
2257 values may be specified for reads and writes. For this option to be
2258 effective, NCQ priority must be supported and enabled, and the :option:`direct`
2259 option must be set. fio must also be run as the root user. Unlike
2260 slat/clat/lat stats, which can be tracked and reported independently, per
2261 priority stats only track and report a single type of latency. By default,
2262 completion latency (clat) will be reported, if :option:`lat_percentiles` is
2263 set, total latency (lat) will be reported.
2265 .. option:: cmdprio_class=int[,int] : [io_uring] [libaio]
2267 Set the I/O priority class to use for I/Os that must be issued with
2268 a priority when :option:`cmdprio_percentage` or
2269 :option:`cmdprio_bssplit` is set. If not specified when
2270 :option:`cmdprio_percentage` or :option:`cmdprio_bssplit` is set,
2271 this defaults to the highest priority class. A single value applies
2272 to reads and writes. Comma-separated values may be specified for
2273 reads and writes. See :manpage:`ionice(1)`. See also the
2274 :option:`prioclass` option.
2276 .. option:: cmdprio=int[,int] : [io_uring] [libaio]
2278 Set the I/O priority value to use for I/Os that must be issued with
2279 a priority when :option:`cmdprio_percentage` or
2280 :option:`cmdprio_bssplit` is set. If not specified when
2281 :option:`cmdprio_percentage` or :option:`cmdprio_bssplit` is set,
2283 Linux limits us to a positive value between 0 and 7, with 0 being the
2284 highest. A single value applies to reads and writes. Comma-separated
2285 values may be specified for reads and writes. See :manpage:`ionice(1)`.
2286 Refer to an appropriate manpage for other operating systems since
2287 meaning of priority may differ. See also the :option:`prio` option.
2289 .. option:: cmdprio_bssplit=str[,str] : [io_uring] [libaio]
2291 To get a finer control over I/O priority, this option allows
2292 specifying the percentage of IOs that must have a priority set
2293 depending on the block size of the IO. This option is useful only
2294 when used together with the :option:`bssplit` option, that is,
2295 multiple different block sizes are used for reads and writes.
2297 The first accepted format for this option is the same as the format of
2298 the :option:`bssplit` option:
2300 cmdprio_bssplit=blocksize/percentage:blocksize/percentage
2302 In this case, each entry will use the priority class and priority
2303 level defined by the options :option:`cmdprio_class` and
2304 :option:`cmdprio` respectively.
2306 The second accepted format for this option is:
2308 cmdprio_bssplit=blocksize/percentage/class/level:blocksize/percentage/class/level
2310 In this case, the priority class and priority level is defined inside
2311 each entry. In comparison with the first accepted format, the second
2312 accepted format does not restrict all entries to have the same priority
2313 class and priority level.
2315 For both formats, only the read and write data directions are supported,
2316 values for trim IOs are ignored. This option is mutually exclusive with
2317 the :option:`cmdprio_percentage` option.
2319 .. option:: fixedbufs : [io_uring] [io_uring_cmd]
2321 If fio is asked to do direct IO, then Linux will map pages for each
2322 IO call, and release them when IO is done. If this option is set, the
2323 pages are pre-mapped before IO is started. This eliminates the need to
2324 map and release for each IO. This is more efficient, and reduces the
2327 .. option:: nonvectored=int : [io_uring] [io_uring_cmd]
2329 With this option, fio will use non-vectored read/write commands, where
2330 address must contain the address directly. Default is -1.
2332 .. option:: force_async=int : [io_uring] [io_uring_cmd]
2334 Normal operation for io_uring is to try and issue an sqe as
2335 non-blocking first, and if that fails, execute it in an async manner.
2336 With this option set to N, then every N request fio will ask sqe to
2337 be issued in an async manner. Default is 0.
2339 .. option:: registerfiles : [io_uring] [io_uring_cmd]
2341 With this option, fio registers the set of files being used with the
2342 kernel. This avoids the overhead of managing file counts in the kernel,
2343 making the submission and completion part more lightweight. Required
2344 for the below :option:`sqthread_poll` option.
2346 .. option:: sqthread_poll : [io_uring] [io_uring_cmd] [xnvme]
2348 Normally fio will submit IO by issuing a system call to notify the
2349 kernel of available items in the SQ ring. If this option is set, the
2350 act of submitting IO will be done by a polling thread in the kernel.
2351 This frees up cycles for fio, at the cost of using more CPU in the
2352 system. As submission is just the time it takes to fill in the sqe
2353 entries and any syscall required to wake up the idle kernel thread,
2354 fio will not report submission latencies.
2356 .. option:: sqthread_poll_cpu=int : [io_uring] [io_uring_cmd]
2358 When :option:`sqthread_poll` is set, this option provides a way to
2359 define which CPU should be used for the polling thread.
2361 .. option:: cmd_type=str : [io_uring_cmd]
2363 Specifies the type of uring passthrough command to be used. Supported
2364 value is nvme. Default is nvme.
2368 [io_uring] [io_uring_cmd] [xnvme]
2370 If this option is set, fio will attempt to use polled IO completions.
2371 Normal IO completions generate interrupts to signal the completion of
2372 IO, polled completions do not. Hence they are require active reaping
2373 by the application. The benefits are more efficient IO for high IOPS
2374 scenarios, and lower latencies for low queue depth IO.
2378 Use poll queues. This is incompatible with
2379 :option:`libblkio_wait_mode=eventfd <libblkio_wait_mode>` and
2380 :option:`libblkio_force_enable_completion_eventfd`.
2384 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
2389 If this option is set, fio will attempt to use polled IO completions.
2390 This will have a similar effect as (io_uring)hipri. Only SCSI READ and
2391 WRITE commands will have the SGV4_FLAG_HIPRI set (not UNMAP (trim) nor
2392 VERIFY). Older versions of the Linux sg driver that do not support
2393 hipri will simply ignore this flag and do normal IO. The Linux SCSI
2394 Low Level Driver (LLD) that "owns" the device also needs to support
2395 hipri (also known as iopoll and mq_poll). The MegaRAID driver is an
2396 example of a SCSI LLD. Default: clear (0) which does normal
2397 (interrupted based) IO.
2399 .. option:: userspace_reap : [libaio]
2401 Normally, with the libaio engine in use, fio will use the
2402 :manpage:`io_getevents(2)` system call to reap newly returned events. With
2403 this flag turned on, the AIO ring will be read directly from user-space to
2404 reap events. The reaping mode is only enabled when polling for a minimum of
2405 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
2407 .. option:: hipri_percentage : [pvsync2]
2409 When hipri is set this determines the probability of a pvsync2 I/O being high
2410 priority. The default is 100%.
2412 .. option:: nowait=bool : [pvsync2] [libaio] [io_uring] [io_uring_cmd]
2414 By default if a request cannot be executed immediately (e.g. resource starvation,
2415 waiting on locks) it is queued and the initiating process will be blocked until
2416 the required resource becomes free.
2418 This option sets the RWF_NOWAIT flag (supported from the 4.14 Linux kernel) and
2419 the call will return instantly with EAGAIN or a partial result rather than waiting.
2421 It is useful to also use ignore_error=EAGAIN when using this option.
2423 Note: glibc 2.27, 2.28 have a bug in syscall wrappers preadv2, pwritev2.
2424 They return EOPNOTSUP instead of EAGAIN.
2426 For cached I/O, using this option usually means a request operates only with
2427 cached data. Currently the RWF_NOWAIT flag does not supported for cached write.
2429 For direct I/O, requests will only succeed if cache invalidation isn't required,
2430 file blocks are fully allocated and the disk request could be issued immediately.
2432 .. option:: fdp=bool : [io_uring_cmd]
2434 Enable Flexible Data Placement mode for write commands.
2436 .. option:: fdp_pli=str : [io_uring_cmd]
2438 Select which Placement ID Index/Indicies this job is allowed to use for
2439 writes. By default, the job will cycle through all available Placement
2440 IDs, so use this to isolate these identifiers to specific jobs. If you
2441 want fio to use placement identifier only at indices 0, 2 and 5 specify
2444 .. option:: cpuload=int : [cpuio]
2446 Attempt to use the specified percentage of CPU cycles. This is a mandatory
2447 option when using cpuio I/O engine.
2449 .. option:: cpuchunks=int : [cpuio]
2451 Split the load into cycles of the given time. In microseconds.
2453 .. option:: cpumode=str : [cpuio]
2455 Specify how to stress the CPU. It can take these two values:
2458 This is the default where the CPU executes noop instructions.
2460 Replace the default noop instructions loop with a qsort algorithm to
2461 consume more energy.
2463 .. option:: exit_on_io_done=bool : [cpuio]
2465 Detect when I/O threads are done, then exit.
2467 .. option:: namenode=str : [libhdfs]
2469 The hostname or IP address of a HDFS cluster namenode to contact.
2471 .. option:: port=int
2475 The listening port of the HFDS cluster namenode.
2479 The TCP or UDP port to bind to or connect to. If this is used with
2480 :option:`numjobs` to spawn multiple instances of the same job type, then
2481 this will be the starting port number since fio will use a range of
2486 The port to use for RDMA-CM communication. This should be the same value
2487 on the client and the server side.
2489 .. option:: hostname=str : [netsplice] [net] [rdma]
2491 The hostname or IP address to use for TCP, UDP or RDMA-CM based I/O. If the job
2492 is a TCP listener or UDP reader, the hostname is not used and must be omitted
2493 unless it is a valid UDP multicast address.
2495 .. option:: serverip=str : [librpma_*]
2497 The IP address to be used for RDMA-CM based I/O.
2499 .. option:: direct_write_to_pmem=bool : [librpma_*]
2501 Set to 1 only when Direct Write to PMem from the remote host is possible.
2502 Otherwise, set to 0.
2504 .. option:: busy_wait_polling=bool : [librpma_*_server]
2506 Set to 0 to wait for completion instead of busy-wait polling completion.
2509 .. option:: interface=str : [netsplice] [net]
2511 The IP address of the network interface used to send or receive UDP
2514 .. option:: ttl=int : [netsplice] [net]
2516 Time-to-live value for outgoing UDP multicast packets. Default: 1.
2518 .. option:: nodelay=bool : [netsplice] [net]
2520 Set TCP_NODELAY on TCP connections.
2522 .. option:: protocol=str, proto=str : [netsplice] [net]
2524 The network protocol to use. Accepted values are:
2527 Transmission control protocol.
2529 Transmission control protocol V6.
2531 User datagram protocol.
2533 User datagram protocol V6.
2537 When the protocol is TCP or UDP, the port must also be given, as well as the
2538 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
2539 normal :option:`filename` option should be used and the port is invalid.
2541 .. option:: listen : [netsplice] [net]
2543 For TCP network connections, tell fio to listen for incoming connections
2544 rather than initiating an outgoing connection. The :option:`hostname` must
2545 be omitted if this option is used.
2547 .. option:: pingpong : [netsplice] [net]
2549 Normally a network writer will just continue writing data, and a network
2550 reader will just consume packages. If ``pingpong=1`` is set, a writer will
2551 send its normal payload to the reader, then wait for the reader to send the
2552 same payload back. This allows fio to measure network latencies. The
2553 submission and completion latencies then measure local time spent sending or
2554 receiving, and the completion latency measures how long it took for the
2555 other end to receive and send back. For UDP multicast traffic
2556 ``pingpong=1`` should only be set for a single reader when multiple readers
2557 are listening to the same address.
2559 .. option:: window_size : [netsplice] [net]
2561 Set the desired socket buffer size for the connection.
2563 .. option:: mss : [netsplice] [net]
2565 Set the TCP maximum segment size (TCP_MAXSEG).
2567 .. option:: donorname=str : [e4defrag]
2569 File will be used as a block donor (swap extents between files).
2571 .. option:: inplace=int : [e4defrag]
2573 Configure donor file blocks allocation strategy:
2576 Default. Preallocate donor's file on init.
2578 Allocate space immediately inside defragment event, and free right
2581 .. option:: clustername=str : [rbd,rados]
2583 Specifies the name of the Ceph cluster.
2585 .. option:: rbdname=str : [rbd]
2587 Specifies the name of the RBD.
2589 .. option:: clientname=str : [rbd,rados]
2591 Specifies the username (without the 'client.' prefix) used to access the
2592 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
2593 the full *type.id* string. If no type. prefix is given, fio will add
2594 'client.' by default.
2596 .. option:: conf=str : [rados]
2598 Specifies the configuration path of ceph cluster, so conf file does not
2599 have to be /etc/ceph/ceph.conf.
2601 .. option:: busy_poll=bool : [rbd,rados]
2603 Poll store instead of waiting for completion. Usually this provides better
2604 throughput at cost of higher(up to 100%) CPU utilization.
2606 .. option:: touch_objects=bool : [rados]
2608 During initialization, touch (create if do not exist) all objects (files).
2609 Touching all objects affects ceph caches and likely impacts test results.
2612 .. option:: pool=str :
2616 Specifies the name of the Ceph pool containing RBD or RADOS data.
2620 Specify the label or UUID of the DAOS pool to connect to.
2622 .. option:: cont=str : [dfs]
2624 Specify the label or UUID of the DAOS container to open.
2626 .. option:: chunk_size=int
2630 Specify a different chunk size (in bytes) for the dfs file.
2631 Use DAOS container's chunk size by default.
2635 The size of the chunk to use for each file.
2637 .. option:: object_class=str : [dfs]
2639 Specify a different object class for the dfs file.
2640 Use DAOS container's object class by default.
2642 .. option:: skip_bad=bool : [mtd]
2644 Skip operations against known bad blocks.
2646 .. option:: hdfsdirectory : [libhdfs]
2648 libhdfs will create chunk in this HDFS directory.
2650 .. option:: verb=str : [rdma]
2652 The RDMA verb to use on this side of the RDMA ioengine connection. Valid
2653 values are write, read, send and recv. These correspond to the equivalent
2654 RDMA verbs (e.g. write = rdma_write etc.). Note that this only needs to be
2655 specified on the client side of the connection. See the examples folder.
2657 .. option:: bindname=str : [rdma]
2659 The name to use to bind the local RDMA-CM connection to a local RDMA device.
2660 This could be a hostname or an IPv4 or IPv6 address. On the server side this
2661 will be passed into the rdma_bind_addr() function and on the client site it
2662 will be used in the rdma_resolve_add() function. This can be useful when
2663 multiple paths exist between the client and the server or in certain loopback
2666 .. option:: stat_type=str : [filestat]
2668 Specify stat system call type to measure lookup/getattr performance.
2669 Default is **stat** for :manpage:`stat(2)`.
2671 .. option:: readfua=bool : [sg]
2673 With readfua option set to 1, read operations include
2674 the force unit access (fua) flag. Default is 0.
2676 .. option:: writefua=bool : [sg]
2678 With writefua option set to 1, write operations include
2679 the force unit access (fua) flag. Default is 0.
2681 .. option:: sg_write_mode=str : [sg]
2683 Specify the type of write commands to issue. This option can take three values:
2686 This is the default where write opcodes are issued as usual.
2687 **write_and_verify**
2688 Issue WRITE AND VERIFY commands. The BYTCHK bit is set to 0. This
2689 directs the device to carry out a medium verification with no data
2690 comparison. The writefua option is ignored with this selection.
2692 This option is deprecated. Use write_and_verify instead.
2694 Issue WRITE SAME commands. This transfers a single block to the device
2695 and writes this same block of data to a contiguous sequence of LBAs
2696 beginning at the specified offset. fio's block size parameter specifies
2697 the amount of data written with each command. However, the amount of data
2698 actually transferred to the device is equal to the device's block
2699 (sector) size. For a device with 512 byte sectors, blocksize=8k will
2700 write 16 sectors with each command. fio will still generate 8k of data
2701 for each command but only the first 512 bytes will be used and
2702 transferred to the device. The writefua option is ignored with this
2705 This option is deprecated. Use write_same instead.
2707 Issue WRITE SAME(16) commands as above but with the No Data Output
2708 Buffer (NDOB) bit set. No data will be transferred to the device with
2709 this bit set. Data written will be a pre-determined pattern such as
2712 Issue WRITE STREAM(16) commands. Use the **stream_id** option to specify
2713 the stream identifier.
2714 **verify_bytchk_00**
2715 Issue VERIFY commands with BYTCHK set to 00. This directs the
2716 device to carry out a medium verification with no data comparison.
2717 **verify_bytchk_01**
2718 Issue VERIFY commands with BYTCHK set to 01. This directs the device to
2719 compare the data on the device with the data transferred to the device.
2720 **verify_bytchk_11**
2721 Issue VERIFY commands with BYTCHK set to 11. This transfers a
2722 single block to the device and compares the contents of this block with the
2723 data on the device beginning at the specified offset. fio's block size
2724 parameter specifies the total amount of data compared with this command.
2725 However, only one block (sector) worth of data is transferred to the device.
2726 This is similar to the WRITE SAME command except that data is compared instead
2729 .. option:: stream_id=int : [sg]
2731 Set the stream identifier for WRITE STREAM commands. If this is set to 0 (which is not
2732 a valid stream identifier) fio will open a stream and then close it when done. Default
2735 .. option:: http_host=str : [http]
2737 Hostname to connect to. For S3, this could be the bucket hostname.
2738 Default is **localhost**
2740 .. option:: http_user=str : [http]
2742 Username for HTTP authentication.
2744 .. option:: http_pass=str : [http]
2746 Password for HTTP authentication.
2748 .. option:: https=str : [http]
2750 Enable HTTPS instead of http. *on* enables HTTPS; *insecure*
2751 will enable HTTPS, but disable SSL peer verification (use with
2752 caution!). Default is **off**
2754 .. option:: http_mode=str : [http]
2756 Which HTTP access mode to use: *webdav*, *swift*, or *s3*.
2757 Default is **webdav**
2759 .. option:: http_s3_region=str : [http]
2761 The S3 region/zone string.
2762 Default is **us-east-1**
2764 .. option:: http_s3_key=str : [http]
2768 .. option:: http_s3_keyid=str : [http]
2770 The S3 key/access id.
2772 .. option:: http_s3_sse_customer_key=str : [http]
2774 The encryption customer key in SSE server side.
2776 .. option:: http_s3_sse_customer_algorithm=str : [http]
2778 The encryption customer algorithm in SSE server side.
2779 Default is **AES256**
2781 .. option:: http_s3_storage_class=str : [http]
2783 Which storage class to access. User-customizable settings.
2784 Default is **STANDARD**
2786 .. option:: http_swift_auth_token=str : [http]
2788 The Swift auth token. See the example configuration file on how
2791 .. option:: http_verbose=int : [http]
2793 Enable verbose requests from libcurl. Useful for debugging. 1
2794 turns on verbose logging from libcurl, 2 additionally enables
2795 HTTP IO tracing. Default is **0**
2797 .. option:: uri=str : [nbd]
2799 Specify the NBD URI of the server to test. The string
2800 is a standard NBD URI
2801 (see https://github.com/NetworkBlockDevice/nbd/tree/master/doc).
2802 Example URIs: nbd://localhost:10809
2803 nbd+unix:///?socket=/tmp/socket
2804 nbds://tlshost/exportname
2806 .. option:: gpu_dev_ids=str : [libcufile]
2808 Specify the GPU IDs to use with CUDA. This is a colon-separated list of
2809 int. GPUs are assigned to workers roundrobin. Default is 0.
2811 .. option:: cuda_io=str : [libcufile]
2813 Specify the type of I/O to use with CUDA. Default is **cufile**.
2816 Use libcufile and nvidia-fs. This option performs I/O directly
2817 between a GPUDirect Storage filesystem and GPU buffers,
2818 avoiding use of a bounce buffer. If :option:`verify` is set,
2819 cudaMemcpy is used to copy verificaton data between RAM and GPU.
2820 Verification data is copied from RAM to GPU before a write
2821 and from GPU to RAM after a read. :option:`direct` must be 1.
2823 Use POSIX to perform I/O with a RAM buffer, and use cudaMemcpy
2824 to transfer data between RAM and the GPUs. Data is copied from
2825 GPU to RAM before a write and copied from RAM to GPU after a
2826 read. :option:`verify` does not affect use of cudaMemcpy.
2828 .. option:: nfs_url=str : [nfs]
2830 URL in libnfs format, eg nfs://<server|ipv4|ipv6>/path[?arg=val[&arg=val]*]
2831 Refer to the libnfs README for more details.
2833 .. option:: program=str : [exec]
2835 Specify the program to execute.
2837 .. option:: arguments=str : [exec]
2839 Specify arguments to pass to program.
2840 Some special variables can be expanded to pass fio's job details to the program.
2843 Replaced by the duration of the job in seconds.
2845 Replaced by the name of the job.
2847 .. option:: grace_time=int : [exec]
2849 Specify the time between the SIGTERM and SIGKILL signals. Default is 1 second.
2851 .. option:: std_redirect=bool : [exec]
2853 If set, stdout and stderr streams are redirected to files named from the job name. Default is true.
2855 .. option:: xnvme_async=str : [xnvme]
2857 Select the xnvme async command interface. This can take these values.
2860 This is default and use to emulate asynchronous I/O by using a
2861 single thread to create a queue pair on top of a synchronous
2862 I/O interface using the NVMe driver IOCTL.
2864 Emulate an asynchronous I/O interface with a pool of userspace
2865 threads on top of a synchronous I/O interface using the NVMe
2866 driver IOCTL. By default four threads are used.
2868 Linux native asynchronous I/O interface which supports both
2869 direct and buffered I/O.
2871 Fast Linux native asynchronous I/O interface for NVMe pass
2872 through commands. This only works with NVMe character device
2875 Use Linux aio for Asynchronous I/O.
2877 Use the posix asynchronous I/O interface to perform one or
2878 more I/O operations asynchronously.
2880 Use the user-space VFIO-based backend, implemented using
2881 libvfn instead of SPDK.
2883 Do not transfer any data; just pretend to. This is mainly used
2884 for introspective performance evaluation.
2886 .. option:: xnvme_sync=str : [xnvme]
2888 Select the xnvme synchronous command interface. This can take these values.
2891 This is default and uses Linux NVMe Driver ioctl() for
2894 This supports regular as well as vectored pread() and pwrite()
2897 This is the same as psync except that it also supports zone
2898 management commands using Linux block layer IOCTLs.
2900 .. option:: xnvme_admin=str : [xnvme]
2902 Select the xnvme admin command interface. This can take these values.
2905 This is default and uses linux NVMe Driver ioctl() for admin
2908 Use Linux Block Layer ioctl() and sysfs for admin commands.
2910 .. option:: xnvme_dev_nsid=int : [xnvme]
2912 xnvme namespace identifier for userspace NVMe driver, SPDK or vfio.
2914 .. option:: xnvme_dev_subnqn=str : [xnvme]
2916 Sets the subsystem NQN for fabrics. This is for xNVMe to utilize a
2917 fabrics target with multiple systems.
2919 .. option:: xnvme_mem=str : [xnvme]
2921 Select the xnvme memory backend. This can take these values.
2924 This is the default posix memory backend for linux NVMe driver.
2926 Use hugepages, instead of existing posix memory backend. The
2927 memory backend uses hugetlbfs. This require users to allocate
2928 hugepages, mount hugetlbfs and set an enviornment variable for
2931 Uses SPDK's memory allocator.
2933 Uses libvfn's memory allocator. This also specifies the use
2934 of libvfn backend instead of SPDK.
2936 .. option:: xnvme_iovec=int : [xnvme]
2938 If this option is set. xnvme will use vectored read/write commands.
2940 .. option:: libblkio_driver=str : [libblkio]
2942 The libblkio *driver* to use. Different drivers access devices through
2943 different underlying interfaces. Available drivers depend on the
2944 libblkio version in use and are listed at
2945 https://libblkio.gitlab.io/libblkio/blkio.html#drivers
2947 .. option:: libblkio_path=str : [libblkio]
2949 Sets the value of the driver-specific "path" property before connecting
2950 the libblkio instance, which identifies the target device or file on
2951 which to perform I/O. Its exact semantics are driver-dependent and not
2952 all drivers may support it; see
2953 https://libblkio.gitlab.io/libblkio/blkio.html#drivers
2955 .. option:: libblkio_pre_connect_props=str : [libblkio]
2957 A colon-separated list of additional libblkio properties to be set after
2958 creating but before connecting the libblkio instance. Each property must
2959 have the format ``<name>=<value>``. Colons can be escaped as ``\:``.
2960 These are set after the engine sets any other properties, so those can
2961 be overriden. Available properties depend on the libblkio version in use
2963 https://libblkio.gitlab.io/libblkio/blkio.html#properties
2965 .. option:: libblkio_num_entries=int : [libblkio]
2967 Sets the value of the driver-specific "num-entries" property before
2968 starting the libblkio instance. Its exact semantics are driver-dependent
2969 and not all drivers may support it; see
2970 https://libblkio.gitlab.io/libblkio/blkio.html#drivers
2972 .. option:: libblkio_queue_size=int : [libblkio]
2974 Sets the value of the driver-specific "queue-size" property before
2975 starting the libblkio instance. Its exact semantics are driver-dependent
2976 and not all drivers may support it; see
2977 https://libblkio.gitlab.io/libblkio/blkio.html#drivers
2979 .. option:: libblkio_pre_start_props=str : [libblkio]
2981 A colon-separated list of additional libblkio properties to be set after
2982 connecting but before starting the libblkio instance. Each property must
2983 have the format ``<name>=<value>``. Colons can be escaped as ``\:``.
2984 These are set after the engine sets any other properties, so those can
2985 be overriden. Available properties depend on the libblkio version in use
2987 https://libblkio.gitlab.io/libblkio/blkio.html#properties
2989 .. option:: libblkio_vectored : [libblkio]
2991 Submit vectored read and write requests.
2993 .. option:: libblkio_write_zeroes_on_trim : [libblkio]
2995 Submit trims as "write zeroes" requests instead of discard requests.
2997 .. option:: libblkio_wait_mode=str : [libblkio]
2999 How to wait for completions:
3002 Use a blocking call to ``blkioq_do_io()``.
3004 Use a blocking call to ``read()`` on the completion eventfd.
3006 Use a busy loop with a non-blocking call to ``blkioq_do_io()``.
3008 .. option:: libblkio_force_enable_completion_eventfd : [libblkio]
3010 Enable the queue's completion eventfd even when unused. This may impact
3011 performance. The default is to enable it only if
3012 :option:`libblkio_wait_mode=eventfd <libblkio_wait_mode>`.
3017 .. option:: iodepth=int
3019 Number of I/O units to keep in flight against the file. Note that
3020 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
3021 for small degrees when :option:`verify_async` is in use). Even async
3022 engines may impose OS restrictions causing the desired depth not to be
3023 achieved. This may happen on Linux when using libaio and not setting
3024 :option:`direct`\=1, since buffered I/O is not async on that OS. Keep an
3025 eye on the I/O depth distribution in the fio output to verify that the
3026 achieved depth is as expected. Default: 1.
3028 .. option:: iodepth_batch_submit=int, iodepth_batch=int
3030 This defines how many pieces of I/O to submit at once. It defaults to 1
3031 which means that we submit each I/O as soon as it is available, but can be
3032 raised to submit bigger batches of I/O at the time. If it is set to 0 the
3033 :option:`iodepth` value will be used.
3035 .. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
3037 This defines how many pieces of I/O to retrieve at once. It defaults to 1
3038 which means that we'll ask for a minimum of 1 I/O in the retrieval process
3039 from the kernel. The I/O retrieval will go on until we hit the limit set by
3040 :option:`iodepth_low`. If this variable is set to 0, then fio will always
3041 check for completed events before queuing more I/O. This helps reduce I/O
3042 latency, at the cost of more retrieval system calls.
3044 .. option:: iodepth_batch_complete_max=int
3046 This defines maximum pieces of I/O to retrieve at once. This variable should
3047 be used along with :option:`iodepth_batch_complete_min`\=int variable,
3048 specifying the range of min and max amount of I/O which should be
3049 retrieved. By default it is equal to the :option:`iodepth_batch_complete_min`
3054 iodepth_batch_complete_min=1
3055 iodepth_batch_complete_max=<iodepth>
3057 which means that we will retrieve at least 1 I/O and up to the whole
3058 submitted queue depth. If none of I/O has been completed yet, we will wait.
3062 iodepth_batch_complete_min=0
3063 iodepth_batch_complete_max=<iodepth>
3065 which means that we can retrieve up to the whole submitted queue depth, but
3066 if none of I/O has been completed yet, we will NOT wait and immediately exit
3067 the system call. In this example we simply do polling.
3069 .. option:: iodepth_low=int
3071 The low water mark indicating when to start filling the queue
3072 again. Defaults to the same as :option:`iodepth`, meaning that fio will
3073 attempt to keep the queue full at all times. If :option:`iodepth` is set to
3074 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
3075 16 requests, it will let the depth drain down to 4 before starting to fill
3078 .. option:: serialize_overlap=bool
3080 Serialize in-flight I/Os that might otherwise cause or suffer from data races.
3081 When two or more I/Os are submitted simultaneously, there is no guarantee that
3082 the I/Os will be processed or completed in the submitted order. Further, if
3083 two or more of those I/Os are writes, any overlapping region between them can
3084 become indeterminate/undefined on certain storage. These issues can cause
3085 verification to fail erratically when at least one of the racing I/Os is
3086 changing data and the overlapping region has a non-zero size. Setting
3087 ``serialize_overlap`` tells fio to avoid provoking this behavior by explicitly
3088 serializing in-flight I/Os that have a non-zero overlap. Note that setting
3089 this option can reduce both performance and the :option:`iodepth` achieved.
3091 This option only applies to I/Os issued for a single job except when it is
3092 enabled along with :option:`io_submit_mode`\=offload. In offload mode, fio
3093 will check for overlap among all I/Os submitted by offload jobs with :option:`serialize_overlap`
3098 .. option:: io_submit_mode=str
3100 This option controls how fio submits the I/O to the I/O engine. The default
3101 is `inline`, which means that the fio job threads submit and reap I/O
3102 directly. If set to `offload`, the job threads will offload I/O submission
3103 to a dedicated pool of I/O threads. This requires some coordination and thus
3104 has a bit of extra overhead, especially for lower queue depth I/O where it
3105 can increase latencies. The benefit is that fio can manage submission rates
3106 independently of the device completion rates. This avoids skewed latency
3107 reporting if I/O gets backed up on the device side (the coordinated omission
3108 problem). Note that this option cannot reliably be used with async IO
3115 .. option:: thinktime=time
3117 Stall the job for the specified period of time after an I/O has completed before issuing the
3118 next. May be used to simulate processing being done by an application.
3119 When the unit is omitted, the value is interpreted in microseconds. See
3120 :option:`thinktime_blocks`, :option:`thinktime_iotime` and :option:`thinktime_spin`.
3122 .. option:: thinktime_spin=time
3124 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
3125 something with the data received, before falling back to sleeping for the
3126 rest of the period specified by :option:`thinktime`. When the unit is
3127 omitted, the value is interpreted in microseconds.
3129 .. option:: thinktime_blocks=int
3131 Only valid if :option:`thinktime` is set - control how many blocks to issue,
3132 before waiting :option:`thinktime` usecs. If not set, defaults to 1 which will make
3133 fio wait :option:`thinktime` usecs after every block. This effectively makes any
3134 queue depth setting redundant, since no more than 1 I/O will be queued
3135 before we have to complete it and do our :option:`thinktime`. In other words, this
3136 setting effectively caps the queue depth if the latter is larger.
3138 .. option:: thinktime_blocks_type=str
3140 Only valid if :option:`thinktime` is set - control how :option:`thinktime_blocks`
3141 triggers. The default is `complete`, which triggers thinktime when fio completes
3142 :option:`thinktime_blocks` blocks. If this is set to `issue`, then the trigger happens
3145 .. option:: thinktime_iotime=time
3147 Only valid if :option:`thinktime` is set - control :option:`thinktime`
3148 interval by time. The :option:`thinktime` stall is repeated after IOs
3149 are executed for :option:`thinktime_iotime`. For example,
3150 ``--thinktime_iotime=9s --thinktime=1s`` repeat 10-second cycle with IOs
3151 for 9 seconds and stall for 1 second. When the unit is omitted,
3152 :option:`thinktime_iotime` is interpreted as a number of seconds. If
3153 this option is used together with :option:`thinktime_blocks`, the
3154 :option:`thinktime` stall is repeated after :option:`thinktime_iotime`
3155 or after :option:`thinktime_blocks` IOs, whichever happens first.
3157 .. option:: rate=int[,int][,int]
3159 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
3160 suffix rules apply. Comma-separated values may be specified for reads,
3161 writes, and trims as described in :option:`blocksize`.
3163 For example, using `rate=1m,500k` would limit reads to 1MiB/sec and writes to
3164 500KiB/sec. Capping only reads or writes can be done with `rate=,500k` or
3165 `rate=500k,` where the former will only limit writes (to 500KiB/sec) and the
3166 latter will only limit reads.
3168 .. option:: rate_min=int[,int][,int]
3170 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
3171 to meet this requirement will cause the job to exit. Comma-separated values
3172 may be specified for reads, writes, and trims as described in
3173 :option:`blocksize`.
3175 .. option:: rate_iops=int[,int][,int]
3177 Cap the bandwidth to this number of IOPS. Basically the same as
3178 :option:`rate`, just specified independently of bandwidth. If the job is
3179 given a block size range instead of a fixed value, the smallest block size
3180 is used as the metric. Comma-separated values may be specified for reads,
3181 writes, and trims as described in :option:`blocksize`.
3183 .. option:: rate_iops_min=int[,int][,int]
3185 If fio doesn't meet this rate of I/O, it will cause the job to exit.
3186 Comma-separated values may be specified for reads, writes, and trims as
3187 described in :option:`blocksize`.
3189 .. option:: rate_process=str
3191 This option controls how fio manages rated I/O submissions. The default is
3192 `linear`, which submits I/O in a linear fashion with fixed delays between
3193 I/Os that gets adjusted based on I/O completion rates. If this is set to
3194 `poisson`, fio will submit I/O based on a more real world random request
3195 flow, known as the Poisson process
3196 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
3197 10^6 / IOPS for the given workload.
3199 .. option:: rate_ignore_thinktime=bool
3201 By default, fio will attempt to catch up to the specified rate setting,
3202 if any kind of thinktime setting was used. If this option is set, then
3203 fio will ignore the thinktime and continue doing IO at the specified
3204 rate, instead of entering a catch-up mode after thinktime is done.
3210 .. option:: latency_target=time
3212 If set, fio will attempt to find the max performance point that the given
3213 workload will run at while maintaining a latency below this target. When
3214 the unit is omitted, the value is interpreted in microseconds. See
3215 :option:`latency_window` and :option:`latency_percentile`.
3217 .. option:: latency_window=time
3219 Used with :option:`latency_target` to specify the sample window that the job
3220 is run at varying queue depths to test the performance. When the unit is
3221 omitted, the value is interpreted in microseconds.
3223 .. option:: latency_percentile=float
3225 The percentage of I/Os that must fall within the criteria specified by
3226 :option:`latency_target` and :option:`latency_window`. If not set, this
3227 defaults to 100.0, meaning that all I/Os must be equal or below to the value
3228 set by :option:`latency_target`.
3230 .. option:: latency_run=bool
3232 Used with :option:`latency_target`. If false (default), fio will find
3233 the highest queue depth that meets :option:`latency_target` and exit. If
3234 true, fio will continue running and try to meet :option:`latency_target`
3235 by adjusting queue depth.
3237 .. option:: max_latency=time[,time][,time]
3239 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
3240 maximum latency. When the unit is omitted, the value is interpreted in
3241 microseconds. Comma-separated values may be specified for reads, writes,
3242 and trims as described in :option:`blocksize`.
3244 .. option:: rate_cycle=int
3246 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
3247 of milliseconds. Defaults to 1000.
3253 .. option:: write_iolog=str
3255 Write the issued I/O patterns to the specified file. See
3256 :option:`read_iolog`. Specify a separate file for each job, otherwise the
3257 iologs will be interspersed and the file may be corrupt. This file will
3258 be opened in append mode.
3260 .. option:: read_iolog=str
3262 Open an iolog with the specified filename and replay the I/O patterns it
3263 contains. This can be used to store a workload and replay it sometime
3264 later. The iolog given may also be a blktrace binary file, which allows fio
3265 to replay a workload captured by :command:`blktrace`. See
3266 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
3267 replay, the file needs to be turned into a blkparse binary data file first
3268 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
3269 You can specify a number of files by separating the names with a ':'
3270 character. See the :option:`filename` option for information on how to
3271 escape ':' characters within the file names. These files will
3272 be sequentially assigned to job clones created by :option:`numjobs`.
3273 '-' is a reserved name, meaning read from stdin, notably if
3274 :option:`filename` is set to '-' which means stdin as well, then
3275 this flag can't be set to '-'.
3277 .. option:: read_iolog_chunked=bool
3279 Determines how iolog is read. If false(default) entire :option:`read_iolog`
3280 will be read at once. If selected true, input from iolog will be read
3281 gradually. Useful when iolog is very large, or it is generated.
3283 .. option:: merge_blktrace_file=str
3285 When specified, rather than replaying the logs passed to :option:`read_iolog`,
3286 the logs go through a merge phase which aggregates them into a single
3287 blktrace. The resulting file is then passed on as the :option:`read_iolog`
3288 parameter. The intention here is to make the order of events consistent.
3289 This limits the influence of the scheduler compared to replaying multiple
3290 blktraces via concurrent jobs.
3292 .. option:: merge_blktrace_scalars=float_list
3294 This is a percentage based option that is index paired with the list of
3295 files passed to :option:`read_iolog`. When merging is performed, scale
3296 the time of each event by the corresponding amount. For example,
3297 ``--merge_blktrace_scalars="50:100"`` runs the first trace in halftime
3298 and the second trace in realtime. This knob is separately tunable from
3299 :option:`replay_time_scale` which scales the trace during runtime and
3300 does not change the output of the merge unlike this option.
3302 .. option:: merge_blktrace_iters=float_list
3304 This is a whole number option that is index paired with the list of files
3305 passed to :option:`read_iolog`. When merging is performed, run each trace
3306 for the specified number of iterations. For example,
3307 ``--merge_blktrace_iters="2:1"`` runs the first trace for two iterations
3308 and the second trace for one iteration.
3310 .. option:: replay_no_stall=bool
3312 When replaying I/O with :option:`read_iolog` the default behavior is to
3313 attempt to respect the timestamps within the log and replay them with the
3314 appropriate delay between IOPS. By setting this variable fio will not
3315 respect the timestamps and attempt to replay them as fast as possible while
3316 still respecting ordering. The result is the same I/O pattern to a given
3317 device, but different timings.
3319 .. option:: replay_time_scale=int
3321 When replaying I/O with :option:`read_iolog`, fio will honor the
3322 original timing in the trace. With this option, it's possible to scale
3323 the time. It's a percentage option, if set to 50 it means run at 50%
3324 the original IO rate in the trace. If set to 200, run at twice the
3325 original IO rate. Defaults to 100.
3327 .. option:: replay_redirect=str
3329 While replaying I/O patterns using :option:`read_iolog` the default behavior
3330 is to replay the IOPS onto the major/minor device that each IOP was recorded
3331 from. This is sometimes undesirable because on a different machine those
3332 major/minor numbers can map to a different device. Changing hardware on the
3333 same system can also result in a different major/minor mapping.
3334 ``replay_redirect`` causes all I/Os to be replayed onto the single specified
3335 device regardless of the device it was recorded
3336 from. i.e. :option:`replay_redirect`\= :file:`/dev/sdc` would cause all I/O
3337 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
3338 multiple devices will be replayed onto a single device, if the trace
3339 contains multiple devices. If you want multiple devices to be replayed
3340 concurrently to multiple redirected devices you must blkparse your trace
3341 into separate traces and replay them with independent fio invocations.
3342 Unfortunately this also breaks the strict time ordering between multiple
3345 .. option:: replay_align=int
3347 Force alignment of the byte offsets in a trace to this value. The value
3348 must be a power of 2.
3350 .. option:: replay_scale=int
3352 Scale byte offsets down by this factor when replaying traces. Should most
3353 likely use :option:`replay_align` as well.
3355 .. option:: replay_skip=str
3357 Sometimes it's useful to skip certain IO types in a replay trace.
3358 This could be, for instance, eliminating the writes in the trace.
3359 Or not replaying the trims/discards, if you are redirecting to
3360 a device that doesn't support them. This option takes a comma
3361 separated list of read, write, trim, sync.
3364 Threads, processes and job synchronization
3365 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3369 Fio defaults to creating jobs by using fork, however if this option is
3370 given, fio will create jobs by using POSIX Threads' function
3371 :manpage:`pthread_create(3)` to create threads instead.
3373 .. option:: wait_for=str
3375 If set, the current job won't be started until all workers of the specified
3376 waitee job are done.
3378 ``wait_for`` operates on the job name basis, so there are a few
3379 limitations. First, the waitee must be defined prior to the waiter job
3380 (meaning no forward references). Second, if a job is being referenced as a
3381 waitee, it must have a unique name (no duplicate waitees).
3383 .. option:: nice=int
3385 Run the job with the given nice value. See man :manpage:`nice(2)`.
3387 On Windows, values less than -15 set the process class to "High"; -1 through
3388 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
3391 .. option:: prio=int
3393 Set the I/O priority value of this job. Linux limits us to a positive value
3394 between 0 and 7, with 0 being the highest. See man
3395 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
3396 systems since meaning of priority may differ. For per-command priority
3397 setting, see I/O engine specific :option:`cmdprio_percentage` and
3398 :option:`cmdprio` options.
3400 .. option:: prioclass=int
3402 Set the I/O priority class. See man :manpage:`ionice(1)`. For per-command
3403 priority setting, see I/O engine specific :option:`cmdprio_percentage`
3404 and :option:`cmdprio_class` options.
3406 .. option:: cpus_allowed=str
3408 Controls the same options as :option:`cpumask`, but accepts a textual
3409 specification of the permitted CPUs instead and CPUs are indexed from 0. So
3410 to use CPUs 0 and 5 you would specify ``cpus_allowed=0,5``. This option also
3411 allows a range of CPUs to be specified -- say you wanted a binding to CPUs
3412 0, 5, and 8 to 15, you would set ``cpus_allowed=0,5,8-15``.
3414 On Windows, when ``cpus_allowed`` is unset only CPUs from fio's current
3415 processor group will be used and affinity settings are inherited from the
3416 system. An fio build configured to target Windows 7 makes options that set
3417 CPUs processor group aware and values will set both the processor group
3418 and a CPU from within that group. For example, on a system where processor
3419 group 0 has 40 CPUs and processor group 1 has 32 CPUs, ``cpus_allowed``
3420 values between 0 and 39 will bind CPUs from processor group 0 and
3421 ``cpus_allowed`` values between 40 and 71 will bind CPUs from processor
3422 group 1. When using ``cpus_allowed_policy=shared`` all CPUs specified by a
3423 single ``cpus_allowed`` option must be from the same processor group. For
3424 Windows fio builds not built for Windows 7, CPUs will only be selected from
3425 (and be relative to) whatever processor group fio happens to be running in
3426 and CPUs from other processor groups cannot be used.
3428 .. option:: cpus_allowed_policy=str
3430 Set the policy of how fio distributes the CPUs specified by
3431 :option:`cpus_allowed` or :option:`cpumask`. Two policies are supported:
3434 All jobs will share the CPU set specified.
3436 Each job will get a unique CPU from the CPU set.
3438 **shared** is the default behavior, if the option isn't specified. If
3439 **split** is specified, then fio will assign one cpu per job. If not
3440 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
3443 .. option:: cpumask=int
3445 Set the CPU affinity of this job. The parameter given is a bit mask of
3446 allowed CPUs the job may run on. So if you want the allowed CPUs to be 1
3447 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
3448 :manpage:`sched_setaffinity(2)`. This may not work on all supported
3449 operating systems or kernel versions. This option doesn't work well for a
3450 higher CPU count than what you can store in an integer mask, so it can only
3451 control cpus 1-32. For boxes with larger CPU counts, use
3452 :option:`cpus_allowed`.
3454 .. option:: numa_cpu_nodes=str
3456 Set this job running on specified NUMA nodes' CPUs. The arguments allow
3457 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
3458 NUMA options support, fio must be built on a system with libnuma-dev(el)
3461 .. option:: numa_mem_policy=str
3463 Set this job's memory policy and corresponding NUMA nodes. Format of the
3468 ``mode`` is one of the following memory policies: ``default``, ``prefer``,
3469 ``bind``, ``interleave`` or ``local``. For ``default`` and ``local`` memory
3470 policies, no node needs to be specified. For ``prefer``, only one node is
3471 allowed. For ``bind`` and ``interleave`` the ``nodelist`` may be as
3472 follows: a comma delimited list of numbers, A-B ranges, or `all`.
3474 .. option:: cgroup=str
3476 Add job to this control group. If it doesn't exist, it will be created. The
3477 system must have a mounted cgroup blkio mount point for this to work. If
3478 your system doesn't have it mounted, you can do so with::
3480 # mount -t cgroup -o blkio none /cgroup
3482 .. option:: cgroup_weight=int
3484 Set the weight of the cgroup to this value. See the documentation that comes
3485 with the kernel, allowed values are in the range of 100..1000.
3487 .. option:: cgroup_nodelete=bool
3489 Normally fio will delete the cgroups it has created after the job
3490 completion. To override this behavior and to leave cgroups around after the
3491 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
3492 to inspect various cgroup files after job completion. Default: false.
3494 .. option:: flow_id=int
3496 The ID of the flow. If not specified, it defaults to being a global
3497 flow. See :option:`flow`.
3499 .. option:: flow=int
3501 Weight in token-based flow control. If this value is used, then fio
3502 regulates the activity between two or more jobs sharing the same
3503 flow_id. Fio attempts to keep each job activity proportional to other
3504 jobs' activities in the same flow_id group, with respect to requested
3505 weight per job. That is, if one job has `flow=3', another job has
3506 `flow=2' and another with `flow=1`, then there will be a roughly 3:2:1
3507 ratio in how much one runs vs the others.
3509 .. option:: flow_sleep=int
3511 The period of time, in microseconds, to wait after the flow counter
3512 has exceeded its proportion before retrying operations.
3514 .. option:: stonewall, wait_for_previous
3516 Wait for preceding jobs in the job file to exit, before starting this
3517 one. Can be used to insert serialization points in the job file. A stone
3518 wall also implies starting a new reporting group, see
3519 :option:`group_reporting`.
3523 By default, fio will continue running all other jobs when one job finishes.
3524 Sometimes this is not the desired action. Setting ``exitall`` will instead
3525 make fio terminate all jobs in the same group, as soon as one job of that
3528 .. option:: exit_what=str
3530 By default, fio will continue running all other jobs when one job finishes.
3531 Sometimes this is not the desired action. Setting ``exitall`` will
3532 instead make fio terminate all jobs in the same group. The option
3533 ``exit_what`` allows to control which jobs get terminated when ``exitall`` is
3534 enabled. The default is ``group`` and does not change the behaviour of
3535 ``exitall``. The setting ``all`` terminates all jobs. The setting ``stonewall``
3536 terminates all currently running jobs across all groups and continues execution
3537 with the next stonewalled group.
3539 .. option:: exec_prerun=str
3541 Before running this job, issue the command specified through
3542 :manpage:`system(3)`. Output is redirected in a file called
3543 :file:`jobname.prerun.txt`.
3545 .. option:: exec_postrun=str
3547 After the job completes, issue the command specified though
3548 :manpage:`system(3)`. Output is redirected in a file called
3549 :file:`jobname.postrun.txt`.
3553 Instead of running as the invoking user, set the user ID to this value
3554 before the thread/process does any work.
3558 Set group ID, see :option:`uid`.
3564 .. option:: verify_only
3566 Do not perform specified workload, only verify data still matches previous
3567 invocation of this workload. This option allows one to check data multiple
3568 times at a later date without overwriting it. This option makes sense only
3569 for workloads that write data, and does not support workloads with the
3570 :option:`time_based` option set.
3572 .. option:: do_verify=bool
3574 Run the verify phase after a write phase. Only valid if :option:`verify` is
3577 .. option:: verify=str
3579 If writing to a file, fio can verify the file contents after each iteration
3580 of the job. Each verification method also implies verification of special
3581 header, which is written to the beginning of each block. This header also
3582 includes meta information, like offset of the block, block number, timestamp
3583 when block was written, etc. :option:`verify` can be combined with
3584 :option:`verify_pattern` option. The allowed values are:
3587 Use an md5 sum of the data area and store it in the header of
3591 Use an experimental crc64 sum of the data area and store it in the
3592 header of each block.
3595 Use a crc32c sum of the data area and store it in the header of
3596 each block. This will automatically use hardware acceleration
3597 (e.g. SSE4.2 on an x86 or CRC crypto extensions on ARM64) but will
3598 fall back to software crc32c if none is found. Generally the
3599 fastest checksum fio supports when hardware accelerated.
3605 Use a crc32 sum of the data area and store it in the header of each
3609 Use a crc16 sum of the data area and store it in the header of each
3613 Use a crc7 sum of the data area and store it in the header of each
3617 Use xxhash as the checksum function. Generally the fastest software
3618 checksum that fio supports.
3621 Use sha512 as the checksum function.
3624 Use sha256 as the checksum function.
3627 Use optimized sha1 as the checksum function.
3630 Use optimized sha3-224 as the checksum function.
3633 Use optimized sha3-256 as the checksum function.
3636 Use optimized sha3-384 as the checksum function.
3639 Use optimized sha3-512 as the checksum function.
3642 This option is deprecated, since now meta information is included in
3643 generic verification header and meta verification happens by
3644 default. For detailed information see the description of the
3645 :option:`verify` setting. This option is kept because of
3646 compatibility's sake with old configurations. Do not use it.
3649 Verify a strict pattern. Normally fio includes a header with some
3650 basic information and checksumming, but if this option is set, only
3651 the specific pattern set with :option:`verify_pattern` is verified.
3654 Only pretend to verify. Useful for testing internals with
3655 :option:`ioengine`\=null, not for much else.
3657 This option can be used for repeated burn-in tests of a system to make sure
3658 that the written data is also correctly read back. If the data direction
3659 given is a read or random read, fio will assume that it should verify a
3660 previously written file. If the data direction includes any form of write,
3661 the verify will be of the newly written data.
3663 To avoid false verification errors, do not use the norandommap option when
3664 verifying data with async I/O engines and I/O depths > 1. Or use the
3665 norandommap and the lfsr random generator together to avoid writing to the
3666 same offset with multiple outstanding I/Os.
3668 .. option:: verify_offset=int
3670 Swap the verification header with data somewhere else in the block before
3671 writing. It is swapped back before verifying.
3673 .. option:: verify_interval=int
3675 Write the verification header at a finer granularity than the
3676 :option:`blocksize`. It will be written for chunks the size of
3677 ``verify_interval``. :option:`blocksize` should divide this evenly.
3679 .. option:: verify_pattern=str
3681 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
3682 filling with totally random bytes, but sometimes it's interesting to fill
3683 with a known pattern for I/O verification purposes. Depending on the width
3684 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time (it can
3685 be either a decimal or a hex number). The ``verify_pattern`` if larger than
3686 a 32-bit quantity has to be a hex number that starts with either "0x" or
3687 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
3688 format, which means that for each block offset will be written and then
3689 verified back, e.g.::
3693 Or use combination of everything::
3695 verify_pattern=0xff%o"abcd"-12
3697 .. option:: verify_fatal=bool
3699 Normally fio will keep checking the entire contents before quitting on a
3700 block verification failure. If this option is set, fio will exit the job on
3701 the first observed failure. Default: false.
3703 .. option:: verify_dump=bool
3705 If set, dump the contents of both the original data block and the data block
3706 we read off disk to files. This allows later analysis to inspect just what
3707 kind of data corruption occurred. Off by default.
3709 .. option:: verify_async=int
3711 Fio will normally verify I/O inline from the submitting thread. This option
3712 takes an integer describing how many async offload threads to create for I/O
3713 verification instead, causing fio to offload the duty of verifying I/O
3714 contents to one or more separate threads. If using this offload option, even
3715 sync I/O engines can benefit from using an :option:`iodepth` setting higher
3716 than 1, as it allows them to have I/O in flight while verifies are running.
3717 Defaults to 0 async threads, i.e. verification is not asynchronous.
3719 .. option:: verify_async_cpus=str
3721 Tell fio to set the given CPU affinity on the async I/O verification
3722 threads. See :option:`cpus_allowed` for the format used.
3724 .. option:: verify_backlog=int
3726 Fio will normally verify the written contents of a job that utilizes verify
3727 once that job has completed. In other words, everything is written then
3728 everything is read back and verified. You may want to verify continually
3729 instead for a variety of reasons. Fio stores the meta data associated with
3730 an I/O block in memory, so for large verify workloads, quite a bit of memory
3731 would be used up holding this meta data. If this option is enabled, fio will
3732 write only N blocks before verifying these blocks.
3734 .. option:: verify_backlog_batch=int
3736 Control how many blocks fio will verify if :option:`verify_backlog` is
3737 set. If not set, will default to the value of :option:`verify_backlog`
3738 (meaning the entire queue is read back and verified). If
3739 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
3740 blocks will be verified, if ``verify_backlog_batch`` is larger than
3741 :option:`verify_backlog`, some blocks will be verified more than once.
3743 .. option:: verify_state_save=bool
3745 When a job exits during the write phase of a verify workload, save its
3746 current state. This allows fio to replay up until that point, if the verify
3747 state is loaded for the verify read phase. The format of the filename is,
3750 <type>-<jobname>-<jobindex>-verify.state.
3752 <type> is "local" for a local run, "sock" for a client/server socket
3753 connection, and "ip" (192.168.0.1, for instance) for a networked
3754 client/server connection. Defaults to true.
3756 .. option:: verify_state_load=bool
3758 If a verify termination trigger was used, fio stores the current write state
3759 of each thread. This can be used at verification time so that fio knows how
3760 far it should verify. Without this information, fio will run a full
3761 verification pass, according to the settings in the job file used. Default
3764 .. option:: trim_percentage=int
3766 Number of verify blocks to discard/trim.
3768 .. option:: trim_verify_zero=bool
3770 Verify that trim/discarded blocks are returned as zeros.
3772 .. option:: trim_backlog=int
3774 Trim after this number of blocks are written.
3776 .. option:: trim_backlog_batch=int
3778 Trim this number of I/O blocks.
3780 .. option:: experimental_verify=bool
3782 Enable experimental verification. Standard verify records I/O metadata
3783 for later use during the verification phase. Experimental verify
3784 instead resets the file after the write phase and then replays I/Os for
3785 the verification phase.
3790 .. option:: steadystate=str:float, ss=str:float
3792 Define the criterion and limit for assessing steady state performance. The
3793 first parameter designates the criterion whereas the second parameter sets
3794 the threshold. When the criterion falls below the threshold for the
3795 specified duration, the job will stop. For example, `iops_slope:0.1%` will
3796 direct fio to terminate the job when the least squares regression slope
3797 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
3798 this will apply to all jobs in the group. Below is the list of available
3799 steady state assessment criteria. All assessments are carried out using only
3800 data from the rolling collection window. Threshold limits can be expressed
3801 as a fixed value or as a percentage of the mean in the collection window.
3803 When using this feature, most jobs should include the :option:`time_based`
3804 and :option:`runtime` options or the :option:`loops` option so that fio does not
3805 stop running after it has covered the full size of the specified file(s) or device(s).
3808 Collect IOPS data. Stop the job if all individual IOPS measurements
3809 are within the specified limit of the mean IOPS (e.g., ``iops:2``
3810 means that all individual IOPS values must be within 2 of the mean,
3811 whereas ``iops:0.2%`` means that all individual IOPS values must be
3812 within 0.2% of the mean IOPS to terminate the job).
3815 Collect IOPS data and calculate the least squares regression
3816 slope. Stop the job if the slope falls below the specified limit.
3819 Collect bandwidth data. Stop the job if all individual bandwidth
3820 measurements are within the specified limit of the mean bandwidth.
3823 Collect bandwidth data and calculate the least squares regression
3824 slope. Stop the job if the slope falls below the specified limit.
3826 .. option:: steadystate_duration=time, ss_dur=time
3828 A rolling window of this duration will be used to judge whether steady
3829 state has been reached. Data will be collected every
3830 :option:`ss_interval`. The default is 0 which disables steady state
3831 detection. When the unit is omitted, the value is interpreted in
3834 .. option:: steadystate_ramp_time=time, ss_ramp=time
3836 Allow the job to run for the specified duration before beginning data
3837 collection for checking the steady state job termination criterion. The
3838 default is 0. When the unit is omitted, the value is interpreted in seconds.
3840 .. option:: steadystate_check_interval=time, ss_interval=time
3842 The values during the rolling window will be collected with a period of
3843 this value. If :option:`ss_interval` is 30s and :option:`ss_dur` is
3844 300s, 10 measurements will be taken. Default is 1s but that might not
3845 converge, especially for slower devices, so set this accordingly. When
3846 the unit is omitted, the value is interpreted in seconds.
3849 Measurements and reporting
3850 ~~~~~~~~~~~~~~~~~~~~~~~~~~
3852 .. option:: per_job_logs=bool
3854 If set, this generates bw/clat/iops log with per file private filenames. If
3855 not set, jobs with identical names will share the log filename. Default:
3858 .. option:: group_reporting
3860 It may sometimes be interesting to display statistics for groups of jobs as
3861 a whole instead of for each individual job. This is especially true if
3862 :option:`numjobs` is used; looking at individual thread/process output
3863 quickly becomes unwieldy. To see the final report per-group instead of
3864 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
3865 same reporting group, unless if separated by a :option:`stonewall`, or by
3866 using :option:`new_group`.
3868 .. option:: new_group
3870 Start a new reporting group. See: :option:`group_reporting`. If not given,
3871 all jobs in a file will be part of the same reporting group, unless
3872 separated by a :option:`stonewall`.
3874 .. option:: stats=bool
3876 By default, fio collects and shows final output results for all jobs
3877 that run. If this option is set to 0, then fio will ignore it in
3878 the final stat output.
3880 .. option:: write_bw_log=str
3882 If given, write a bandwidth log for this job. Can be used to store data of
3883 the bandwidth of the jobs in their lifetime.
3885 If no str argument is given, the default filename of
3886 :file:`jobname_type.x.log` is used. Even when the argument is given, fio
3887 will still append the type of log. So if one specifies::
3891 The actual log name will be :file:`foo_bw.x.log` where `x` is the index
3892 of the job (`1..N`, where `N` is the number of jobs). If
3893 :option:`per_job_logs` is false, then the filename will not include the
3896 The included :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
3897 text files into nice graphs. See `Log File Formats`_ for how data is
3898 structured within the file.
3900 .. option:: write_lat_log=str
3902 Same as :option:`write_bw_log`, except this option creates I/O
3903 submission (e.g., :file:`name_slat.x.log`), completion (e.g.,
3904 :file:`name_clat.x.log`), and total (e.g., :file:`name_lat.x.log`)
3905 latency files instead. See :option:`write_bw_log` for details about
3906 the filename format and `Log File Formats`_ for how data is structured
3909 .. option:: write_hist_log=str
3911 Same as :option:`write_bw_log` but writes an I/O completion latency
3912 histogram file (e.g., :file:`name_hist.x.log`) instead. Note that this
3913 file will be empty unless :option:`log_hist_msec` has also been set.
3914 See :option:`write_bw_log` for details about the filename format and
3915 `Log File Formats`_ for how data is structured within the file.
3917 .. option:: write_iops_log=str
3919 Same as :option:`write_bw_log`, but writes an IOPS file (e.g.
3920 :file:`name_iops.x.log`) instead. Because fio defaults to individual
3921 I/O logging, the value entry in the IOPS log will be 1 unless windowed
3922 logging (see :option:`log_avg_msec`) has been enabled. See
3923 :option:`write_bw_log` for details about the filename format and `Log
3924 File Formats`_ for how data is structured within the file.
3926 .. option:: log_entries=int
3928 By default, fio will log an entry in the iops, latency, or bw log for
3929 every I/O that completes. The initial number of I/O log entries is 1024.
3930 When the log entries are all used, new log entries are dynamically
3931 allocated. This dynamic log entry allocation may negatively impact
3932 time-related statistics such as I/O tail latencies (e.g. 99.9th percentile
3933 completion latency). This option allows specifying a larger initial
3934 number of log entries to avoid run-time allocations of new log entries,
3935 resulting in more precise time-related I/O statistics.
3936 Also see :option:`log_avg_msec`. Defaults to 1024.
3938 .. option:: log_avg_msec=int
3940 By default, fio will log an entry in the iops, latency, or bw log for every
3941 I/O that completes. When writing to the disk log, that can quickly grow to a
3942 very large size. Setting this option makes fio average the each log entry
3943 over the specified period of time, reducing the resolution of the log. See
3944 :option:`log_max_value` as well. Defaults to 0, logging all entries.
3945 Also see `Log File Formats`_.
3947 .. option:: log_hist_msec=int
3949 Same as :option:`log_avg_msec`, but logs entries for completion latency
3950 histograms. Computing latency percentiles from averages of intervals using
3951 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
3952 histogram entries over the specified period of time, reducing log sizes for
3953 high IOPS devices while retaining percentile accuracy. See
3954 :option:`log_hist_coarseness` and :option:`write_hist_log` as well.
3955 Defaults to 0, meaning histogram logging is disabled.
3957 .. option:: log_hist_coarseness=int
3959 Integer ranging from 0 to 6, defining the coarseness of the resolution of
3960 the histogram logs enabled with :option:`log_hist_msec`. For each increment
3961 in coarseness, fio outputs half as many bins. Defaults to 0, for which
3962 histogram logs contain 1216 latency bins. See :option:`write_hist_log`
3963 and `Log File Formats`_.
3965 .. option:: log_max_value=bool
3967 If :option:`log_avg_msec` is set, fio logs the average over that window. If
3968 you instead want to log the maximum value, set this option to 1. Defaults to
3969 0, meaning that averaged values are logged.
3971 .. option:: log_offset=bool
3973 If this is set, the iolog options will include the byte offset for the I/O
3974 entry as well as the other data values. Defaults to 0 meaning that
3975 offsets are not present in logs. Also see `Log File Formats`_.
3977 .. option:: log_compression=int
3979 If this is set, fio will compress the I/O logs as it goes, to keep the
3980 memory footprint lower. When a log reaches the specified size, that chunk is
3981 removed and compressed in the background. Given that I/O logs are fairly
3982 highly compressible, this yields a nice memory savings for longer runs. The
3983 downside is that the compression will consume some background CPU cycles, so
3984 it may impact the run. This, however, is also true if the logging ends up
3985 consuming most of the system memory. So pick your poison. The I/O logs are
3986 saved normally at the end of a run, by decompressing the chunks and storing
3987 them in the specified log file. This feature depends on the availability of
3990 .. option:: log_compression_cpus=str
3992 Define the set of CPUs that are allowed to handle online log compression for
3993 the I/O jobs. This can provide better isolation between performance
3994 sensitive jobs, and background compression work. See
3995 :option:`cpus_allowed` for the format used.
3997 .. option:: log_store_compressed=bool
3999 If set, fio will store the log files in a compressed format. They can be
4000 decompressed with fio, using the :option:`--inflate-log` command line
4001 parameter. The files will be stored with a :file:`.fz` suffix.
4003 .. option:: log_unix_epoch=bool
4005 If set, fio will log Unix timestamps to the log files produced by enabling
4006 write_type_log for each log type, instead of the default zero-based
4009 .. option:: log_alternate_epoch=bool
4011 If set, fio will log timestamps based on the epoch used by the clock specified
4012 in the log_alternate_epoch_clock_id option, to the log files produced by
4013 enabling write_type_log for each log type, instead of the default zero-based
4016 .. option:: log_alternate_epoch_clock_id=int
4018 Specifies the clock_id to be used by clock_gettime to obtain the alternate epoch
4019 if either log_unix_epoch or log_alternate_epoch are true. Otherwise has no
4020 effect. Default value is 0, or CLOCK_REALTIME.
4022 .. option:: block_error_percentiles=bool
4024 If set, record errors in trim block-sized units from writes and trims and
4025 output a histogram of how many trims it took to get to errors, and what kind
4026 of error was encountered.
4028 .. option:: bwavgtime=int
4030 Average the calculated bandwidth over the given time. Value is specified in
4031 milliseconds. If the job also does bandwidth logging through
4032 :option:`write_bw_log`, then the minimum of this option and
4033 :option:`log_avg_msec` will be used. Default: 500ms.
4035 .. option:: iopsavgtime=int
4037 Average the calculated IOPS over the given time. Value is specified in
4038 milliseconds. If the job also does IOPS logging through
4039 :option:`write_iops_log`, then the minimum of this option and
4040 :option:`log_avg_msec` will be used. Default: 500ms.
4042 .. option:: disk_util=bool
4044 Generate disk utilization statistics, if the platform supports it.
4047 .. option:: disable_lat=bool
4049 Disable measurements of total latency numbers. Useful only for cutting back
4050 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
4051 performance at really high IOPS rates. Note that to really get rid of a
4052 large amount of these calls, this option must be used with
4053 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
4055 .. option:: disable_clat=bool
4057 Disable measurements of completion latency numbers. See
4058 :option:`disable_lat`.
4060 .. option:: disable_slat=bool
4062 Disable measurements of submission latency numbers. See
4063 :option:`disable_lat`.
4065 .. option:: disable_bw_measurement=bool, disable_bw=bool
4067 Disable measurements of throughput/bandwidth numbers. See
4068 :option:`disable_lat`.
4070 .. option:: slat_percentiles=bool
4072 Report submission latency percentiles. Submission latency is not recorded
4073 for synchronous ioengines.
4075 .. option:: clat_percentiles=bool
4077 Report completion latency percentiles.
4079 .. option:: lat_percentiles=bool
4081 Report total latency percentiles. Total latency is the sum of submission
4082 latency and completion latency.
4084 .. option:: percentile_list=float_list
4086 Overwrite the default list of percentiles for latencies and the block error
4087 histogram. Each number is a floating point number in the range (0,100], and
4088 the maximum length of the list is 20. Use ``:`` to separate the numbers. For
4089 example, ``--percentile_list=99.5:99.9`` will cause fio to report the
4090 latency durations below which 99.5% and 99.9% of the observed latencies fell,
4093 .. option:: significant_figures=int
4095 If using :option:`--output-format` of `normal`, set the significant
4096 figures to this value. Higher values will yield more precise IOPS and
4097 throughput units, while lower values will round. Requires a minimum
4098 value of 1 and a maximum value of 10. Defaults to 4.
4104 .. option:: exitall_on_error
4106 When one job finishes in error, terminate the rest. The default is to wait
4107 for each job to finish.
4109 .. option:: continue_on_error=str
4111 Normally fio will exit the job on the first observed failure. If this option
4112 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
4113 EILSEQ) until the runtime is exceeded or the I/O size specified is
4114 completed. If this option is used, there are two more stats that are
4115 appended, the total error count and the first error. The error field given
4116 in the stats is the first error that was hit during the run.
4118 Note: a write error from the device may go unnoticed by fio when using
4119 buffered IO, as the write() (or similar) system call merely dirties the
4120 kernel pages, unless :option:`sync` or :option:`direct` is used. Device IO
4121 errors occur when the dirty data is actually written out to disk. If fully
4122 sync writes aren't desirable, :option:`fsync` or :option:`fdatasync` can be
4123 used as well. This is specific to writes, as reads are always synchronous.
4125 The allowed values are:
4128 Exit on any I/O or verify errors.
4131 Continue on read errors, exit on all others.
4134 Continue on write errors, exit on all others.
4137 Continue on any I/O error, exit on all others.
4140 Continue on verify errors, exit on all others.
4143 Continue on all errors.
4146 Backward-compatible alias for 'none'.
4149 Backward-compatible alias for 'all'.
4151 .. option:: ignore_error=str
4153 Sometimes you want to ignore some errors during test in that case you can
4154 specify error list for each error type, instead of only being able to
4155 ignore the default 'non-fatal error' using :option:`continue_on_error`.
4156 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
4157 given error type is separated with ':'. Error may be symbol ('ENOSPC',
4158 'ENOMEM') or integer. Example::
4160 ignore_error=EAGAIN,ENOSPC:122
4162 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
4163 WRITE. This option works by overriding :option:`continue_on_error` with
4164 the list of errors for each error type if any.
4166 .. option:: error_dump=bool
4168 If set dump every error even if it is non fatal, true by default. If
4169 disabled only fatal error will be dumped.
4171 Running predefined workloads
4172 ----------------------------
4174 Fio includes predefined profiles that mimic the I/O workloads generated by
4177 .. option:: profile=str
4179 The predefined workload to run. Current profiles are:
4182 Threaded I/O bench (tiotest/tiobench) like workload.
4185 Aerospike Certification Tool (ACT) like workload.
4187 To view a profile's additional options use :option:`--cmdhelp` after specifying
4188 the profile. For example::
4190 $ fio --profile=act --cmdhelp
4195 .. option:: device-names=str
4200 .. option:: load=int
4203 ACT load multiplier. Default: 1.
4205 .. option:: test-duration=time
4208 How long the entire test takes to run. When the unit is omitted, the value
4209 is given in seconds. Default: 24h.
4211 .. option:: threads-per-queue=int
4214 Number of read I/O threads per device. Default: 8.
4216 .. option:: read-req-num-512-blocks=int
4219 Number of 512B blocks to read at the time. Default: 3.
4221 .. option:: large-block-op-kbytes=int
4224 Size of large block ops in KiB (writes). Default: 131072.
4229 Set to run ACT prep phase.
4231 Tiobench profile options
4232 ~~~~~~~~~~~~~~~~~~~~~~~~
4234 .. option:: size=str
4239 .. option:: block=int
4242 Block size in bytes. Default: 4096.
4244 .. option:: numruns=int
4254 .. option:: threads=int
4259 Interpreting the output
4260 -----------------------
4263 Example output was based on the following:
4264 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --time_based \
4265 --rate=1256k --bs=14K --name=quick --runtime=1s --name=mixed \
4266 --runtime=2m --rw=rw
4268 Fio spits out a lot of output. While running, fio will display the status of the
4269 jobs created. An example of that would be::
4271 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]
4273 The characters inside the first set of square brackets denote the current status of
4274 each thread. The first character is the first job defined in the job file, and so
4275 forth. The possible values (in typical life cycle order) are:
4277 +------+-----+-----------------------------------------------------------+
4279 +======+=====+===========================================================+
4280 | P | | Thread setup, but not started. |
4281 +------+-----+-----------------------------------------------------------+
4282 | C | | Thread created. |
4283 +------+-----+-----------------------------------------------------------+
4284 | I | | Thread initialized, waiting or generating necessary data. |
4285 +------+-----+-----------------------------------------------------------+
4286 | | p | Thread running pre-reading file(s). |
4287 +------+-----+-----------------------------------------------------------+
4288 | | / | Thread is in ramp period. |
4289 +------+-----+-----------------------------------------------------------+
4290 | | R | Running, doing sequential reads. |
4291 +------+-----+-----------------------------------------------------------+
4292 | | r | Running, doing random reads. |
4293 +------+-----+-----------------------------------------------------------+
4294 | | W | Running, doing sequential writes. |
4295 +------+-----+-----------------------------------------------------------+
4296 | | w | Running, doing random writes. |
4297 +------+-----+-----------------------------------------------------------+
4298 | | M | Running, doing mixed sequential reads/writes. |
4299 +------+-----+-----------------------------------------------------------+
4300 | | m | Running, doing mixed random reads/writes. |
4301 +------+-----+-----------------------------------------------------------+
4302 | | D | Running, doing sequential trims. |
4303 +------+-----+-----------------------------------------------------------+
4304 | | d | Running, doing random trims. |
4305 +------+-----+-----------------------------------------------------------+
4306 | | F | Running, currently waiting for :manpage:`fsync(2)`. |
4307 +------+-----+-----------------------------------------------------------+
4308 | | V | Running, doing verification of written data. |
4309 +------+-----+-----------------------------------------------------------+
4310 | f | | Thread finishing. |
4311 +------+-----+-----------------------------------------------------------+
4312 | E | | Thread exited, not reaped by main thread yet. |
4313 +------+-----+-----------------------------------------------------------+
4314 | _ | | Thread reaped. |
4315 +------+-----+-----------------------------------------------------------+
4316 | X | | Thread reaped, exited with an error. |
4317 +------+-----+-----------------------------------------------------------+
4318 | K | | Thread reaped, exited due to signal. |
4319 +------+-----+-----------------------------------------------------------+
4322 Example output was based on the following:
4323 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --runtime=58m \
4324 --time_based --rate=2512k --bs=256K --numjobs=10 \
4325 --name=readers --rw=read --name=writers --rw=write
4327 Fio will condense the thread string as not to take up more space on the command
4328 line than needed. For instance, if you have 10 readers and 10 writers running,
4329 the output would look like this::
4331 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]
4333 Note that the status string is displayed in order, so it's possible to tell which of
4334 the jobs are currently doing what. In the example above this means that jobs 1--10
4335 are readers and 11--20 are writers.
4337 The other values are fairly self explanatory -- number of threads currently
4338 running and doing I/O, the number of currently open files (f=), the estimated
4339 completion percentage, the rate of I/O since last check (read speed listed first,
4340 then write speed and optionally trim speed) in terms of bandwidth and IOPS,
4341 and time to completion for the current running group. It's impossible to estimate
4342 runtime of the following groups (if any).
4345 Example output was based on the following:
4346 TZ=UTC fio --iodepth=16 --ioengine=posixaio --filename=/tmp/fiofile \
4347 --direct=1 --size=100M --time_based --runtime=50s --rate_iops=89 \
4348 --bs=7K --name=Client1 --rw=write
4350 When fio is done (or interrupted by :kbd:`Ctrl-C`), it will show the data for
4351 each thread, group of threads, and disks in that order. For each overall thread (or
4352 group) the output looks like::
4354 Client1: (groupid=0, jobs=1): err= 0: pid=16109: Sat Jun 24 12:07:54 2017
4355 write: IOPS=88, BW=623KiB/s (638kB/s)(30.4MiB/50032msec)
4356 slat (nsec): min=500, max=145500, avg=8318.00, stdev=4781.50
4357 clat (usec): min=170, max=78367, avg=4019.02, stdev=8293.31
4358 lat (usec): min=174, max=78375, avg=4027.34, stdev=8291.79
4359 clat percentiles (usec):
4360 | 1.00th=[ 302], 5.00th=[ 326], 10.00th=[ 343], 20.00th=[ 363],
4361 | 30.00th=[ 392], 40.00th=[ 404], 50.00th=[ 416], 60.00th=[ 445],
4362 | 70.00th=[ 816], 80.00th=[ 6718], 90.00th=[12911], 95.00th=[21627],
4363 | 99.00th=[43779], 99.50th=[51643], 99.90th=[68682], 99.95th=[72877],
4365 bw ( KiB/s): min= 532, max= 686, per=0.10%, avg=622.87, stdev=24.82, samples= 100
4366 iops : min= 76, max= 98, avg=88.98, stdev= 3.54, samples= 100
4367 lat (usec) : 250=0.04%, 500=64.11%, 750=4.81%, 1000=2.79%
4368 lat (msec) : 2=4.16%, 4=1.84%, 10=4.90%, 20=11.33%, 50=5.37%
4369 lat (msec) : 100=0.65%
4370 cpu : usr=0.27%, sys=0.18%, ctx=12072, majf=0, minf=21
4371 IO depths : 1=85.0%, 2=13.1%, 4=1.8%, 8=0.1%, 16=0.0%, 32=0.0%, >=64=0.0%
4372 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
4373 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
4374 issued rwt: total=0,4450,0, short=0,0,0, dropped=0,0,0
4375 latency : target=0, window=0, percentile=100.00%, depth=8
4377 The job name (or first job's name when using :option:`group_reporting`) is printed,
4378 along with the group id, count of jobs being aggregated, last error id seen (which
4379 is 0 when there are no errors), pid/tid of that thread and the time the job/group
4380 completed. Below are the I/O statistics for each data direction performed (showing
4381 writes in the example above). In the order listed, they denote:
4384 The string before the colon shows the I/O direction the statistics
4385 are for. **IOPS** is the average I/Os performed per second. **BW**
4386 is the average bandwidth rate shown as: value in power of 2 format
4387 (value in power of 10 format). The last two values show: (**total
4388 I/O performed** in power of 2 format / **runtime** of that thread).
4391 Submission latency (**min** being the minimum, **max** being the
4392 maximum, **avg** being the average, **stdev** being the standard
4393 deviation). This is the time from when fio initialized the I/O
4394 to submission. For synchronous ioengines this includes the time
4395 up until just before the ioengine's queue function is called.
4396 For asynchronous ioengines this includes the time up through the
4397 completion of the ioengine's queue function (and commit function
4398 if it is defined). For sync I/O this row is not displayed as the
4399 slat is negligible. This value can be in nanoseconds,
4400 microseconds or milliseconds --- fio will choose the most
4401 appropriate base and print that (in the example above
4402 nanoseconds was the best scale). Note: in :option:`--minimal`
4403 mode latencies are always expressed in microseconds.
4406 Completion latency. Same names as slat, this denotes the time from
4407 submission to completion of the I/O pieces. For sync I/O, this
4408 represents the time from when the I/O was submitted to the
4409 operating system to when it was completed. For asynchronous
4410 ioengines this is the time from when the ioengine's queue (and
4411 commit if available) functions were completed to when the I/O's
4412 completion was reaped by fio.
4415 Total latency. Same names as slat and clat, this denotes the time from
4416 when fio created the I/O unit to completion of the I/O operation.
4417 It is the sum of submission and completion latency.
4420 Bandwidth statistics based on samples. Same names as the xlat stats,
4421 but also includes the number of samples taken (**samples**) and an
4422 approximate percentage of total aggregate bandwidth this thread
4423 received in its group (**per**). This last value is only really
4424 useful if the threads in this group are on the same disk, since they
4425 are then competing for disk access.
4428 IOPS statistics based on samples. Same names as bw.
4430 **lat (nsec/usec/msec)**
4431 The distribution of I/O completion latencies. This is the time from when
4432 I/O leaves fio and when it gets completed. Unlike the separate
4433 read/write/trim sections above, the data here and in the remaining
4434 sections apply to all I/Os for the reporting group. 250=0.04% means that
4435 0.04% of the I/Os completed in under 250us. 500=64.11% means that 64.11%
4436 of the I/Os required 250 to 499us for completion.
4439 CPU usage. User and system time, along with the number of context
4440 switches this thread went through, usage of system and user time, and
4441 finally the number of major and minor page faults. The CPU utilization
4442 numbers are averages for the jobs in that reporting group, while the
4443 context and fault counters are summed.
4446 The distribution of I/O depths over the job lifetime. The numbers are
4447 divided into powers of 2 and each entry covers depths from that value
4448 up to those that are lower than the next entry -- e.g., 16= covers
4449 depths from 16 to 31. Note that the range covered by a depth
4450 distribution entry can be different to the range covered by the
4451 equivalent submit/complete distribution entry.
4454 How many pieces of I/O were submitting in a single submit call. Each
4455 entry denotes that amount and below, until the previous entry -- e.g.,
4456 16=100% means that we submitted anywhere between 9 to 16 I/Os per submit
4457 call. Note that the range covered by a submit distribution entry can
4458 be different to the range covered by the equivalent depth distribution
4462 Like the above submit number, but for completions instead.
4465 The number of read/write/trim requests issued, and how many of them were
4469 These values are for :option:`latency_target` and related options. When
4470 these options are engaged, this section describes the I/O depth required
4471 to meet the specified latency target.
4474 Example output was based on the following:
4475 TZ=UTC fio --ioengine=null --iodepth=2 --size=100M --numjobs=2 \
4476 --rate_process=poisson --io_limit=32M --name=read --bs=128k \
4477 --rate=11M --name=write --rw=write --bs=2k --rate=700k
4479 After each client has been listed, the group statistics are printed. They
4480 will look like this::
4482 Run status group 0 (all jobs):
4483 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
4484 WRITE: bw=1231KiB/s (1261kB/s), 616KiB/s-621KiB/s (630kB/s-636kB/s), io=64.0MiB (67.1MB), run=52747-53223msec
4486 For each data direction it prints:
4489 Aggregate bandwidth of threads in this group followed by the
4490 minimum and maximum bandwidth of all the threads in this group.
4491 Values outside of brackets are power-of-2 format and those
4492 within are the equivalent value in a power-of-10 format.
4494 Aggregate I/O performed of all threads in this group. The
4495 format is the same as bw.
4497 The smallest and longest runtimes of the threads in this group.
4499 And finally, the disk statistics are printed. This is Linux specific. They will look like this::
4501 Disk stats (read/write):
4502 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
4504 Each value is printed for both reads and writes, with reads first. The
4508 Number of I/Os performed by all groups.
4510 Number of merges performed by the I/O scheduler.
4512 Number of ticks we kept the disk busy.
4514 Total time spent in the disk queue.
4516 The disk utilization. A value of 100% means we kept the disk
4517 busy constantly, 50% would be a disk idling half of the time.
4519 It is also possible to get fio to dump the current output while it is running,
4520 without terminating the job. To do that, send fio the **USR1** signal. You can
4521 also get regularly timed dumps by using the :option:`--status-interval`
4522 parameter, or by creating a file in :file:`/tmp` named
4523 :file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
4524 current output status.
4530 For scripted usage where you typically want to generate tables or graphs of the
4531 results, fio can output the results in a semicolon separated format. The format
4532 is one long line of values, such as::
4534 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%
4535 A description of this job goes here.
4537 The job description (if provided) follows on a second line for terse v2.
4538 It appears on the same line for other terse versions.
4540 To enable terse output, use the :option:`--minimal` or
4541 :option:`--output-format`\=terse command line options. The
4542 first value is the version of the terse output format. If the output has to be
4543 changed for some reason, this number will be incremented by 1 to signify that
4546 Split up, the format is as follows (comments in brackets denote when a
4547 field was introduced or whether it's specific to some terse version):
4551 terse version, fio version [v3], jobname, groupid, error
4555 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
4556 Submission latency: min, max, mean, stdev (usec)
4557 Completion latency: min, max, mean, stdev (usec)
4558 Completion latency percentiles: 20 fields (see below)
4559 Total latency: min, max, mean, stdev (usec)
4560 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
4561 IOPS [v5]: min, max, mean, stdev, number of samples
4567 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
4568 Submission latency: min, max, mean, stdev (usec)
4569 Completion latency: min, max, mean, stdev (usec)
4570 Completion latency percentiles: 20 fields (see below)
4571 Total latency: min, max, mean, stdev (usec)
4572 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
4573 IOPS [v5]: min, max, mean, stdev, number of samples
4575 TRIM status [all but version 3]:
4577 Fields are similar to READ/WRITE status.
4581 user, system, context switches, major faults, minor faults
4585 <=1, 2, 4, 8, 16, 32, >=64
4587 I/O latencies microseconds::
4589 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
4591 I/O latencies milliseconds::
4593 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
4595 Disk utilization [v3]::
4597 disk name, read ios, write ios, read merges, write merges, read ticks, write ticks,
4598 time spent in queue, disk utilization percentage
4600 Additional Info (dependent on continue_on_error, default off)::
4602 total # errors, first error code
4604 Additional Info (dependent on description being set)::
4608 Completion latency percentiles can be a grouping of up to 20 sets, so for the
4609 terse output fio writes all of them. Each field will look like this::
4613 which is the Xth percentile, and the `usec` latency associated with it.
4615 For `Disk utilization`, all disks used by fio are shown. So for each disk there
4616 will be a disk utilization section.
4618 Below is a single line containing short names for each of the fields in the
4619 minimal output v3, separated by semicolons::
4621 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
4623 In client/server mode terse output differs from what appears when jobs are run
4624 locally. Disk utilization data is omitted from the standard terse output and
4625 for v3 and later appears on its own separate line at the end of each terse
4632 The `json` output format is intended to be both human readable and convenient
4633 for automated parsing. For the most part its sections mirror those of the
4634 `normal` output. The `runtime` value is reported in msec and the `bw` value is
4635 reported in 1024 bytes per second units.
4641 The `json+` output format is identical to the `json` output format except that it
4642 adds a full dump of the completion latency bins. Each `bins` object contains a
4643 set of (key, value) pairs where keys are latency durations and values count how
4644 many I/Os had completion latencies of the corresponding duration. For example,
4647 "bins" : { "87552" : 1, "89600" : 1, "94720" : 1, "96768" : 1, "97792" : 1, "99840" : 1, "100864" : 2, "103936" : 6, "104960" : 534, "105984" : 5995, "107008" : 7529, ... }
4649 This data indicates that one I/O required 87,552ns to complete, two I/Os required
4650 100,864ns to complete, and 7529 I/Os required 107,008ns to complete.
4652 Also included with fio is a Python script `fio_jsonplus_clat2csv` that takes
4653 json+ output and generates CSV-formatted latency data suitable for plotting.
4655 The latency durations actually represent the midpoints of latency intervals.
4656 For details refer to :file:`stat.h`.
4662 There are two trace file format that you can encounter. The older (v1) format is
4663 unsupported since version 1.20-rc3 (March 2008). It will still be described
4664 below in case that you get an old trace and want to understand it.
4666 In any case the trace is a simple text file with a single action per line.
4669 Trace file format v1
4670 ~~~~~~~~~~~~~~~~~~~~
4672 Each line represents a single I/O action in the following format::
4676 where `rw=0/1` for read/write, and the `offset` and `length` entries being in bytes.
4678 This format is not supported in fio versions >= 1.20-rc3.
4681 Trace file format v2
4682 ~~~~~~~~~~~~~~~~~~~~
4684 The second version of the trace file format was added in fio version 1.17. It
4685 allows one to access more than one file per trace and has a bigger set of possible
4688 The first line of the trace file has to be::
4692 Following this can be lines in two different formats, which are described below.
4694 The file management format::
4698 The `filename` is given as an absolute path. The `action` can be one of these:
4701 Add the given `filename` to the trace.
4703 Open the file with the given `filename`. The `filename` has to have
4704 been added with the **add** action before.
4706 Close the file with the given `filename`. The file has to have been
4710 The file I/O action format::
4712 filename action offset length
4714 The `filename` is given as an absolute path, and has to have been added and
4715 opened before it can be used with this format. The `offset` and `length` are
4716 given in bytes. The `action` can be one of these:
4719 Wait for `offset` microseconds. Everything below 100 is discarded.
4720 The time is relative to the previous `wait` statement. Note that
4721 action `wait` is not allowed as of version 3, as the same behavior
4722 can be achieved using timestamps.
4724 Read `length` bytes beginning from `offset`.
4726 Write `length` bytes beginning from `offset`.
4728 :manpage:`fsync(2)` the file.
4730 :manpage:`fdatasync(2)` the file.
4732 Trim the given file from the given `offset` for `length` bytes.
4735 Trace file format v3
4736 ~~~~~~~~~~~~~~~~~~~~
4738 The third version of the trace file format was added in fio version 3.31. It
4739 forces each action to have a timestamp associated with it.
4741 The first line of the trace file has to be::
4745 Following this can be lines in two different formats, which are described below.
4747 The file management format::
4749 timestamp filename action
4751 The file I/O action format::
4753 timestamp filename action offset length
4755 The `timestamp` is relative to the beginning of the run (ie starts at 0). The
4756 `filename`, `action`, `offset` and `length` are identical to version 2, except
4757 that version 3 does not allow the `wait` action.
4760 I/O Replay - Merging Traces
4761 ---------------------------
4763 Colocation is a common practice used to get the most out of a machine.
4764 Knowing which workloads play nicely with each other and which ones don't is
4765 a much harder task. While fio can replay workloads concurrently via multiple
4766 jobs, it leaves some variability up to the scheduler making results harder to
4767 reproduce. Merging is a way to make the order of events consistent.
4769 Merging is integrated into I/O replay and done when a
4770 :option:`merge_blktrace_file` is specified. The list of files passed to
4771 :option:`read_iolog` go through the merge process and output a single file
4772 stored to the specified file. The output file is passed on as if it were the
4773 only file passed to :option:`read_iolog`. An example would look like::
4775 $ fio --read_iolog="<file1>:<file2>" --merge_blktrace_file="<output_file>"
4777 Creating only the merged file can be done by passing the command line argument
4778 :option:`--merge-blktrace-only`.
4780 Scaling traces can be done to see the relative impact of any particular trace
4781 being slowed down or sped up. :option:`merge_blktrace_scalars` takes in a colon
4782 separated list of percentage scalars. It is index paired with the files passed
4783 to :option:`read_iolog`.
4785 With scaling, it may be desirable to match the running time of all traces.
4786 This can be done with :option:`merge_blktrace_iters`. It is index paired with
4787 :option:`read_iolog` just like :option:`merge_blktrace_scalars`.
4789 In an example, given two traces, A and B, each 60s long. If we want to see
4790 the impact of trace A issuing IOs twice as fast and repeat trace A over the
4791 runtime of trace B, the following can be done::
4793 $ fio --read_iolog="<trace_a>:"<trace_b>" --merge_blktrace_file"<output_file>" --merge_blktrace_scalars="50:100" --merge_blktrace_iters="2:1"
4795 This runs trace A at 2x the speed twice for approximately the same runtime as
4796 a single run of trace B.
4799 CPU idleness profiling
4800 ----------------------
4802 In some cases, we want to understand CPU overhead in a test. For example, we
4803 test patches for the specific goodness of whether they reduce CPU usage.
4804 Fio implements a balloon approach to create a thread per CPU that runs at idle
4805 priority, meaning that it only runs when nobody else needs the cpu.
4806 By measuring the amount of work completed by the thread, idleness of each CPU
4807 can be derived accordingly.
4809 An unit work is defined as touching a full page of unsigned characters. Mean and
4810 standard deviation of time to complete an unit work is reported in "unit work"
4811 section. Options can be chosen to report detailed percpu idleness or overall
4812 system idleness by aggregating percpu stats.
4815 Verification and triggers
4816 -------------------------
4818 Fio is usually run in one of two ways, when data verification is done. The first
4819 is a normal write job of some sort with verify enabled. When the write phase has
4820 completed, fio switches to reads and verifies everything it wrote. The second
4821 model is running just the write phase, and then later on running the same job
4822 (but with reads instead of writes) to repeat the same I/O patterns and verify
4823 the contents. Both of these methods depend on the write phase being completed,
4824 as fio otherwise has no idea how much data was written.
4826 With verification triggers, fio supports dumping the current write state to
4827 local files. Then a subsequent read verify workload can load this state and know
4828 exactly where to stop. This is useful for testing cases where power is cut to a
4829 server in a managed fashion, for instance.
4831 A verification trigger consists of two things:
4833 1) Storing the write state of each job.
4834 2) Executing a trigger command.
4836 The write state is relatively small, on the order of hundreds of bytes to single
4837 kilobytes. It contains information on the number of completions done, the last X
4840 A trigger is invoked either through creation ('touch') of a specified file in
4841 the system, or through a timeout setting. If fio is run with
4842 :option:`--trigger-file`\= :file:`/tmp/trigger-file`, then it will continually
4843 check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
4844 will fire off the trigger (thus saving state, and executing the trigger
4847 For client/server runs, there's both a local and remote trigger. If fio is
4848 running as a server backend, it will send the job states back to the client for
4849 safe storage, then execute the remote trigger, if specified. If a local trigger
4850 is specified, the server will still send back the write state, but the client
4851 will then execute the trigger.
4853 Verification trigger example
4854 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
4856 Let's say we want to run a powercut test on the remote Linux machine 'server'.
4857 Our write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
4858 some point during the run, and we'll run this test from the safety or our local
4859 machine, 'localbox'. On the server, we'll start the fio backend normally::
4861 server# fio --server
4863 and on the client, we'll fire off the workload::
4865 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
4867 We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
4869 echo b > /proc/sysrq-trigger
4871 on the server once it has received the trigger and sent us the write state. This
4872 will work, but it's not **really** cutting power to the server, it's merely
4873 abruptly rebooting it. If we have a remote way of cutting power to the server
4874 through IPMI or similar, we could do that through a local trigger command
4875 instead. Let's assume we have a script that does IPMI reboot of a given hostname,
4876 ipmi-reboot. On localbox, we could then have run fio with a local trigger
4879 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
4881 For this case, fio would wait for the server to send us the write state, then
4882 execute ``ipmi-reboot server`` when that happened.
4884 Loading verify state
4885 ~~~~~~~~~~~~~~~~~~~~
4887 To load stored write state, a read verification job file must contain the
4888 :option:`verify_state_load` option. If that is set, fio will load the previously
4889 stored state. For a local fio run this is done by loading the files directly,
4890 and on a client/server run, the server backend will ask the client to send the
4891 files over and load them from there.
4897 Fio supports a variety of log file formats, for logging latencies, bandwidth,
4898 and IOPS. The logs share a common format, which looks like this:
4900 *time* (`msec`), *value*, *data direction*, *block size* (`bytes`),
4901 *offset* (`bytes`), *command priority*
4903 *Time* for the log entry is always in milliseconds. The *value* logged depends
4904 on the type of log, it will be one of the following:
4907 Value is latency in nsecs
4913 *Data direction* is one of the following:
4922 The entry's *block size* is always in bytes. The *offset* is the position in bytes
4923 from the start of the file for that particular I/O. The logging of the offset can be
4924 toggled with :option:`log_offset`.
4926 *Command priority* is 0 for normal priority and 1 for high priority. This is controlled
4927 by the ioengine specific :option:`cmdprio_percentage`.
4929 Fio defaults to logging every individual I/O but when windowed logging is set
4930 through :option:`log_avg_msec`, either the average (by default) or the maximum
4931 (:option:`log_max_value` is set) *value* seen over the specified period of time
4932 is recorded. Each *data direction* seen within the window period will aggregate
4933 its values in a separate row. Further, when using windowed logging the *block
4934 size* and *offset* entries will always contain 0.
4940 Normally fio is invoked as a stand-alone application on the machine where the
4941 I/O workload should be generated. However, the backend and frontend of fio can
4942 be run separately i.e., the fio server can generate an I/O workload on the "Device
4943 Under Test" while being controlled by a client on another machine.
4945 Start the server on the machine which has access to the storage DUT::
4949 where `args` defines what fio listens to. The arguments are of the form
4950 ``type,hostname`` or ``IP,port``. *type* is either ``ip`` (or ip4) for TCP/IP
4951 v4, ``ip6`` for TCP/IP v6, or ``sock`` for a local unix domain socket.
4952 *hostname* is either a hostname or IP address, and *port* is the port to listen
4953 to (only valid for TCP/IP, not a local socket). Some examples:
4957 Start a fio server, listening on all interfaces on the default port (8765).
4959 2) ``fio --server=ip:hostname,4444``
4961 Start a fio server, listening on IP belonging to hostname and on port 4444.
4963 3) ``fio --server=ip6:::1,4444``
4965 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
4967 4) ``fio --server=,4444``
4969 Start a fio server, listening on all interfaces on port 4444.
4971 5) ``fio --server=1.2.3.4``
4973 Start a fio server, listening on IP 1.2.3.4 on the default port.
4975 6) ``fio --server=sock:/tmp/fio.sock``
4977 Start a fio server, listening on the local socket :file:`/tmp/fio.sock`.
4979 Once a server is running, a "client" can connect to the fio server with::
4981 fio <local-args> --client=<server> <remote-args> <job file(s)>
4983 where `local-args` are arguments for the client where it is running, `server`
4984 is the connect string, and `remote-args` and `job file(s)` are sent to the
4985 server. The `server` string follows the same format as it does on the server
4986 side, to allow IP/hostname/socket and port strings.
4988 Fio can connect to multiple servers this way::
4990 fio --client=<server1> <job file(s)> --client=<server2> <job file(s)>
4992 If the job file is located on the fio server, then you can tell the server to
4993 load a local file as well. This is done by using :option:`--remote-config` ::
4995 fio --client=server --remote-config /path/to/file.fio
4997 Then fio will open this local (to the server) job file instead of being passed
4998 one from the client.
5000 If you have many servers (example: 100 VMs/containers), you can input a pathname
5001 of a file containing host IPs/names as the parameter value for the
5002 :option:`--client` option. For example, here is an example :file:`host.list`
5003 file containing 2 hostnames::
5005 host1.your.dns.domain
5006 host2.your.dns.domain
5008 The fio command would then be::
5010 fio --client=host.list <job file(s)>
5012 In this mode, you cannot input server-specific parameters or job files -- all
5013 servers receive the same job file.
5015 In order to let ``fio --client`` runs use a shared filesystem from multiple
5016 hosts, ``fio --client`` now prepends the IP address of the server to the
5017 filename. For example, if fio is using the directory :file:`/mnt/nfs/fio` and is
5018 writing filename :file:`fileio.tmp`, with a :option:`--client` `hostfile`
5019 containing two hostnames ``h1`` and ``h2`` with IP addresses 192.168.10.120 and
5020 192.168.10.121, then fio will create two files::
5022 /mnt/nfs/fio/192.168.10.120.fileio.tmp
5023 /mnt/nfs/fio/192.168.10.121.fileio.tmp
5025 Terse output in client/server mode will differ slightly from what is produced
5026 when fio is run in stand-alone mode. See the terse output section for details.