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 Show available debug options.
99 .. option:: --parse-only
101 Parse options only, don't start any I/O.
103 .. option:: --output=filename
105 Write output to file `filename`.
107 .. option:: --output-format=format
109 Set the reporting `format` to `normal`, `terse`, `json`, or `json+`. Multiple
110 formats can be selected, separated by a comma. `terse` is a CSV based
111 format. `json+` is like `json`, except it adds a full dump of the latency
114 .. option:: --bandwidth-log
116 Generate aggregate bandwidth logs.
118 .. option:: --minimal
120 Print statistics in a terse, semicolon-delimited format.
122 .. option:: --append-terse
124 Print statistics in selected mode AND terse, semicolon-delimited format.
125 **Deprecated**, use :option:`--output-format` instead to select multiple
128 .. option:: --terse-version=version
130 Set terse `version` output format (default 3, or 2 or 4 or 5).
132 .. option:: --version
134 Print version information and exit.
138 Print a summary of the command line options and exit.
140 .. option:: --cpuclock-test
142 Perform test and validation of internal CPU clock.
144 .. option:: --crctest=[test]
146 Test the speed of the built-in checksumming functions. If no argument is
147 given, all of them are tested. Alternatively, a comma separated list can
148 be passed, in which case the given ones are tested.
150 .. option:: --cmdhelp=command
152 Print help information for `command`. May be ``all`` for all commands.
154 .. option:: --enghelp=[ioengine[,command]]
156 List all commands defined by `ioengine`, or print help for `command`
157 defined by `ioengine`. If no `ioengine` is given, list all
160 .. option:: --showcmd=jobfile
162 Convert `jobfile` to a set of command-line options.
164 .. option:: --readonly
166 Turn on safety read-only checks, preventing writes and trims. The
167 ``--readonly`` option is an extra safety guard to prevent users from
168 accidentally starting a write or trim workload when that is not desired.
169 Fio will only modify the device under test if
170 `rw=write/randwrite/rw/randrw/trim/randtrim/trimwrite` is given. This
171 safety net can be used as an extra precaution.
173 .. option:: --eta=when
175 Specifies when real-time ETA estimate should be printed. `when` may be
176 `always`, `never` or `auto`. `auto` is the default, it prints ETA
177 when requested if the output is a TTY. `always` disregards the output
178 type, and prints ETA when requested. `never` never prints ETA.
180 .. option:: --eta-interval=time
182 By default, fio requests client ETA status roughly every second. With
183 this option, the interval is configurable. Fio imposes a minimum
184 allowed time to avoid flooding the console, less than 250 msec is
187 .. option:: --eta-newline=time
189 Force a new line for every `time` period passed. When the unit is omitted,
190 the value is interpreted in seconds.
192 .. option:: --status-interval=time
194 Force a full status dump of cumulative (from job start) values at `time`
195 intervals. This option does *not* provide per-period measurements. So
196 values such as bandwidth are running averages. When the time unit is omitted,
197 `time` is interpreted in seconds.
199 .. option:: --section=name
201 Only run specified section `name` in job file. Multiple sections can be specified.
202 The ``--section`` option allows one to combine related jobs into one file.
203 E.g. one job file could define light, moderate, and heavy sections. Tell
204 fio to run only the "heavy" section by giving ``--section=heavy``
205 command line option. One can also specify the "write" operations in one
206 section and "verify" operation in another section. The ``--section`` option
207 only applies to job sections. The reserved *global* section is always
210 .. option:: --alloc-size=kb
212 Set the internal smalloc pool size to `kb` in KiB. The
213 ``--alloc-size`` switch allows one to use a larger pool size for smalloc.
214 If running large jobs with randommap enabled, fio can run out of memory.
215 Smalloc is an internal allocator for shared structures from a fixed size
216 memory pool and can grow to 16 pools. The pool size defaults to 16MiB.
218 NOTE: While running :file:`.fio_smalloc.*` backing store files are visible
221 .. option:: --warnings-fatal
223 All fio parser warnings are fatal, causing fio to exit with an
226 .. option:: --max-jobs=nr
228 Set the maximum number of threads/processes to support to `nr`.
229 NOTE: On Linux, it may be necessary to increase the shared-memory
230 limit (:file:`/proc/sys/kernel/shmmax`) if fio runs into errors while
233 .. option:: --server=args
235 Start a backend server, with `args` specifying what to listen to.
236 See `Client/Server`_ section.
238 .. option:: --daemonize=pidfile
240 Background a fio server, writing the pid to the given `pidfile` file.
242 .. option:: --client=hostname
244 Instead of running the jobs locally, send and run them on the given `hostname`
245 or set of `hostname`\s. See `Client/Server`_ section.
247 .. option:: --remote-config=file
249 Tell fio server to load this local `file`.
251 .. option:: --idle-prof=option
253 Report CPU idleness. `option` is one of the following:
256 Run unit work calibration only and exit.
259 Show aggregate system idleness and unit work.
262 As **system** but also show per CPU idleness.
264 .. option:: --inflate-log=log
266 Inflate and output compressed `log`.
268 .. option:: --trigger-file=file
270 Execute trigger command when `file` exists.
272 .. option:: --trigger-timeout=time
274 Execute trigger at this `time`.
276 .. option:: --trigger=command
278 Set this `command` as local trigger.
280 .. option:: --trigger-remote=command
282 Set this `command` as remote trigger.
284 .. option:: --aux-path=path
286 Use the directory specified by `path` for generated state files instead
287 of the current working directory.
289 Any parameters following the options will be assumed to be job files, unless
290 they match a job file parameter. Multiple job files can be listed and each job
291 file will be regarded as a separate group. Fio will :option:`stonewall`
292 execution between each group.
298 As previously described, fio accepts one or more job files describing what it is
299 supposed to do. The job file format is the classic ini file, where the names
300 enclosed in [] brackets define the job name. You are free to use any ASCII name
301 you want, except *global* which has special meaning. Following the job name is
302 a sequence of zero or more parameters, one per line, that define the behavior of
303 the job. If the first character in a line is a ';' or a '#', the entire line is
304 discarded as a comment.
306 A *global* section sets defaults for the jobs described in that file. A job may
307 override a *global* section parameter, and a job file may even have several
308 *global* sections if so desired. A job is only affected by a *global* section
311 The :option:`--cmdhelp` option also lists all options. If used with a `command`
312 argument, :option:`--cmdhelp` will detail the given `command`.
314 See the `examples/` directory for inspiration on how to write job files. Note
315 the copyright and license requirements currently apply to `examples/` files.
317 So let's look at a really simple job file that defines two processes, each
318 randomly reading from a 128MiB file:
322 ; -- start job file --
333 As you can see, the job file sections themselves are empty as all the described
334 parameters are shared. As no :option:`filename` option is given, fio makes up a
335 `filename` for each of the jobs as it sees fit. On the command line, this job
336 would look as follows::
338 $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
341 Let's look at an example that has a number of processes writing randomly to
346 ; -- start job file --
357 Here we have no *global* section, as we only have one job defined anyway. We
358 want to use async I/O here, with a depth of 4 for each file. We also increased
359 the buffer size used to 32KiB and define numjobs to 4 to fork 4 identical
360 jobs. The result is 4 processes each randomly writing to their own 64MiB
361 file. Instead of using the above job file, you could have given the parameters
362 on the command line. For this case, you would specify::
364 $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
366 When fio is utilized as a basis of any reasonably large test suite, it might be
367 desirable to share a set of standardized settings across multiple job files.
368 Instead of copy/pasting such settings, any section may pull in an external
369 :file:`filename.fio` file with *include filename* directive, as in the following
372 ; -- start job file including.fio --
376 include glob-include.fio
383 include test-include.fio
384 ; -- end job file including.fio --
388 ; -- start job file glob-include.fio --
391 ; -- end job file glob-include.fio --
395 ; -- start job file test-include.fio --
398 ; -- end job file test-include.fio --
400 Settings pulled into a section apply to that section only (except *global*
401 section). Include directives may be nested in that any included file may contain
402 further include directive(s). Include files may not contain [] sections.
405 Environment variables
406 ~~~~~~~~~~~~~~~~~~~~~
408 Fio also supports environment variable expansion in job files. Any sub-string of
409 the form ``${VARNAME}`` as part of an option value (in other words, on the right
410 of the '='), will be expanded to the value of the environment variable called
411 `VARNAME`. If no such environment variable is defined, or `VARNAME` is the
412 empty string, the empty string will be substituted.
414 As an example, let's look at a sample fio invocation and job file::
416 $ SIZE=64m NUMJOBS=4 fio jobfile.fio
420 ; -- start job file --
427 This will expand to the following equivalent job file at runtime:
431 ; -- start job file --
438 Fio ships with a few example job files, you can also look there for inspiration.
443 Additionally, fio has a set of reserved keywords that will be replaced
444 internally with the appropriate value. Those keywords are:
448 The architecture page size of the running system.
452 Megabytes of total memory in the system.
456 Number of online available CPUs.
458 These can be used on the command line or in the job file, and will be
459 automatically substituted with the current system values when the job is
460 run. Simple math is also supported on these keywords, so you can perform actions
465 and get that properly expanded to 8 times the size of memory in the machine.
471 This section describes in details each parameter associated with a job. Some
472 parameters take an option of a given type, such as an integer or a
473 string. Anywhere a numeric value is required, an arithmetic expression may be
474 used, provided it is surrounded by parentheses. Supported operators are:
483 For time values in expressions, units are microseconds by default. This is
484 different than for time values not in expressions (not enclosed in
485 parentheses). The following types are used:
492 String: A sequence of alphanumeric characters.
495 Integer with possible time suffix. Without a unit value is interpreted as
496 seconds unless otherwise specified. Accepts a suffix of 'd' for days, 'h' for
497 hours, 'm' for minutes, 's' for seconds, 'ms' (or 'msec') for milliseconds and
498 'us' (or 'usec') for microseconds. For example, use 10m for 10 minutes.
503 Integer. A whole number value, which may contain an integer prefix
504 and an integer suffix:
506 [*integer prefix*] **number** [*integer suffix*]
508 The optional *integer prefix* specifies the number's base. The default
509 is decimal. *0x* specifies hexadecimal.
511 The optional *integer suffix* specifies the number's units, and includes an
512 optional unit prefix and an optional unit. For quantities of data, the
513 default unit is bytes. For quantities of time, the default unit is seconds
514 unless otherwise specified.
516 With :option:`kb_base`\=1000, fio follows international standards for unit
517 prefixes. To specify power-of-10 decimal values defined in the
518 International System of Units (SI):
520 * *K* -- means kilo (K) or 1000
521 * *M* -- means mega (M) or 1000**2
522 * *G* -- means giga (G) or 1000**3
523 * *T* -- means tera (T) or 1000**4
524 * *P* -- means peta (P) or 1000**5
526 To specify power-of-2 binary values defined in IEC 80000-13:
528 * *Ki* -- means kibi (Ki) or 1024
529 * *Mi* -- means mebi (Mi) or 1024**2
530 * *Gi* -- means gibi (Gi) or 1024**3
531 * *Ti* -- means tebi (Ti) or 1024**4
532 * *Pi* -- means pebi (Pi) or 1024**5
534 With :option:`kb_base`\=1024 (the default), the unit prefixes are opposite
535 from those specified in the SI and IEC 80000-13 standards to provide
536 compatibility with old scripts. For example, 4k means 4096.
538 For quantities of data, an optional unit of 'B' may be included
539 (e.g., 'kB' is the same as 'k').
541 The *integer suffix* is not case sensitive (e.g., m/mi mean mebi/mega,
542 not milli). 'b' and 'B' both mean byte, not bit.
544 Examples with :option:`kb_base`\=1000:
546 * *4 KiB*: 4096, 4096b, 4096B, 4ki, 4kib, 4kiB, 4Ki, 4KiB
547 * *1 MiB*: 1048576, 1mi, 1024ki
548 * *1 MB*: 1000000, 1m, 1000k
549 * *1 TiB*: 1099511627776, 1ti, 1024gi, 1048576mi
550 * *1 TB*: 1000000000, 1t, 1000m, 1000000k
552 Examples with :option:`kb_base`\=1024 (default):
554 * *4 KiB*: 4096, 4096b, 4096B, 4k, 4kb, 4kB, 4K, 4KB
555 * *1 MiB*: 1048576, 1m, 1024k
556 * *1 MB*: 1000000, 1mi, 1000ki
557 * *1 TiB*: 1099511627776, 1t, 1024g, 1048576m
558 * *1 TB*: 1000000000, 1ti, 1000mi, 1000000ki
560 To specify times (units are not case sensitive):
564 * *M* -- means minutes
565 * *s* -- or sec means seconds (default)
566 * *ms* -- or *msec* means milliseconds
567 * *us* -- or *usec* means microseconds
569 If the option accepts an upper and lower range, use a colon ':' or
570 minus '-' to separate such values. See :ref:`irange <irange>`.
571 If the lower value specified happens to be larger than the upper value
572 the two values are swapped.
577 Boolean. Usually parsed as an integer, however only defined for
578 true and false (1 and 0).
583 Integer range with suffix. Allows value range to be given, such as
584 1024-4096. A colon may also be used as the separator, e.g. 1k:4k. If the
585 option allows two sets of ranges, they can be specified with a ',' or '/'
586 delimiter: 1k-4k/8k-32k. Also see :ref:`int <int>`.
589 A list of floating point numbers, separated by a ':' character.
591 With the above in mind, here follows the complete list of fio job parameters.
597 .. option:: kb_base=int
599 Select the interpretation of unit prefixes in input parameters.
602 Inputs comply with IEC 80000-13 and the International
603 System of Units (SI). Use:
605 - power-of-2 values with IEC prefixes (e.g., KiB)
606 - power-of-10 values with SI prefixes (e.g., kB)
609 Compatibility mode (default). To avoid breaking old scripts:
611 - power-of-2 values with SI prefixes
612 - power-of-10 values with IEC prefixes
614 See :option:`bs` for more details on input parameters.
616 Outputs always use correct prefixes. Most outputs include both
619 bw=2383.3kB/s (2327.4KiB/s)
621 If only one value is reported, then kb_base selects the one to use:
623 **1000** -- SI prefixes
625 **1024** -- IEC prefixes
627 .. option:: unit_base=int
629 Base unit for reporting. Allowed values are:
632 Use auto-detection (default).
644 ASCII name of the job. This may be used to override the name printed by fio
645 for this job. Otherwise the job name is used. On the command line this
646 parameter has the special purpose of also signaling the start of a new job.
648 .. option:: description=str
650 Text description of the job. Doesn't do anything except dump this text
651 description when this job is run. It's not parsed.
653 .. option:: loops=int
655 Run the specified number of iterations of this job. Used to repeat the same
656 workload a given number of times. Defaults to 1.
658 .. option:: numjobs=int
660 Create the specified number of clones of this job. Each clone of job
661 is spawned as an independent thread or process. May be used to setup a
662 larger number of threads/processes doing the same thing. Each thread is
663 reported separately; to see statistics for all clones as a whole, use
664 :option:`group_reporting` in conjunction with :option:`new_group`.
665 See :option:`--max-jobs`. Default: 1.
668 Time related parameters
669 ~~~~~~~~~~~~~~~~~~~~~~~
671 .. option:: runtime=time
673 Tell fio to terminate processing after the specified period of time. It
674 can be quite hard to determine for how long a specified job will run, so
675 this parameter is handy to cap the total runtime to a given time. When
676 the unit is omitted, the value is interpreted in seconds.
678 .. option:: time_based
680 If set, fio will run for the duration of the :option:`runtime` specified
681 even if the file(s) are completely read or written. It will simply loop over
682 the same workload as many times as the :option:`runtime` allows.
684 .. option:: startdelay=irange(time)
686 Delay the start of job for the specified amount of time. Can be a single
687 value or a range. When given as a range, each thread will choose a value
688 randomly from within the range. Value is in seconds if a unit is omitted.
690 .. option:: ramp_time=time
692 If set, fio will run the specified workload for this amount of time before
693 logging any performance numbers. Useful for letting performance settle
694 before logging results, thus minimizing the runtime required for stable
695 results. Note that the ``ramp_time`` is considered lead in time for a job,
696 thus it will increase the total runtime if a special timeout or
697 :option:`runtime` is specified. When the unit is omitted, the value is
700 .. option:: clocksource=str
702 Use the given clocksource as the base of timing. The supported options are:
705 :manpage:`gettimeofday(2)`
708 :manpage:`clock_gettime(2)`
711 Internal CPU clock source
713 cpu is the preferred clocksource if it is reliable, as it is very fast (and
714 fio is heavy on time calls). Fio will automatically use this clocksource if
715 it's supported and considered reliable on the system it is running on,
716 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
717 means supporting TSC Invariant.
719 .. option:: gtod_reduce=bool
721 Enable all of the :manpage:`gettimeofday(2)` reducing options
722 (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus
723 reduce precision of the timeout somewhat to really shrink the
724 :manpage:`gettimeofday(2)` call count. With this option enabled, we only do
725 about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all
726 time keeping was enabled.
728 .. option:: gtod_cpu=int
730 Sometimes it's cheaper to dedicate a single thread of execution to just
731 getting the current time. Fio (and databases, for instance) are very
732 intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set
733 one CPU aside for doing nothing but logging current time to a shared memory
734 location. Then the other threads/processes that run I/O workloads need only
735 copy that segment, instead of entering the kernel with a
736 :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time
737 calls will be excluded from other uses. Fio will manually clear it from the
738 CPU mask of other jobs.
744 .. option:: directory=str
746 Prefix filenames with this directory. Used to place files in a different
747 location than :file:`./`. You can specify a number of directories by
748 separating the names with a ':' character. These directories will be
749 assigned equally distributed to job clones created by :option:`numjobs` as
750 long as they are using generated filenames. If specific `filename(s)` are
751 set fio will use the first listed directory, and thereby matching the
752 `filename` semantic (which generates a file for each clone if not
753 specified, but lets all clones use the same file if set).
755 See the :option:`filename` option for information on how to escape "``:``" and
756 "``\``" characters within the directory path itself.
758 Note: To control the directory fio will use for internal state files
759 use :option:`--aux-path`.
761 .. option:: filename=str
763 Fio normally makes up a `filename` based on the job name, thread number, and
764 file number (see :option:`filename_format`). If you want to share files
765 between threads in a job or several
766 jobs with fixed file paths, specify a `filename` for each of them to override
767 the default. If the ioengine is file based, you can specify a number of files
768 by separating the names with a ':' colon. So if you wanted a job to open
769 :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use
770 ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is
771 specified, :option:`nrfiles` is ignored. The size of regular files specified
772 by this option will be :option:`size` divided by number of files unless an
773 explicit size is specified by :option:`filesize`.
775 Each colon and backslash in the wanted path must be escaped with a ``\``
776 character. For instance, if the path is :file:`/dev/dsk/foo@3,0:c` then you
777 would use ``filename=/dev/dsk/foo@3,0\:c`` and if the path is
778 :file:`F:\\filename` then you would use ``filename=F\:\\filename``.
780 On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for
781 the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc.
782 Note: Windows and FreeBSD prevent write access to areas
783 of the disk containing in-use data (e.g. filesystems).
785 The filename "`-`" is a reserved name, meaning *stdin* or *stdout*. Which
786 of the two depends on the read/write direction set.
788 .. option:: filename_format=str
790 If sharing multiple files between jobs, it is usually necessary to have fio
791 generate the exact names that you want. By default, fio will name a file
792 based on the default file format specification of
793 :file:`jobname.jobnumber.filenumber`. With this option, that can be
794 customized. Fio will recognize and replace the following keywords in this
798 The name of the worker thread or process.
800 The incremental number of the worker thread or process.
802 The incremental number of the file for that worker thread or
805 To have dependent jobs share a set of files, this option can be set to have
806 fio generate filenames that are shared between the two. For instance, if
807 :file:`testfiles.$filenum` is specified, file number 4 for any job will be
808 named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum`
809 will be used if no other format specifier is given.
811 If you specify a path then the directories will be created up to the
812 main directory for the file. So for example if you specify
813 ``filename_format=a/b/c/$jobnum`` then the directories a/b/c will be
814 created before the file setup part of the job. If you specify
815 :option:`directory` then the path will be relative that directory,
816 otherwise it is treated as the absolute path.
818 .. option:: unique_filename=bool
820 To avoid collisions between networked clients, fio defaults to prefixing any
821 generated filenames (with a directory specified) with the source of the
822 client connecting. To disable this behavior, set this option to 0.
824 .. option:: opendir=str
826 Recursively open any files below directory `str`.
828 .. option:: lockfile=str
830 Fio defaults to not locking any files before it does I/O to them. If a file
831 or file descriptor is shared, fio can serialize I/O to that file to make the
832 end result consistent. This is usual for emulating real workloads that share
833 files. The lock modes are:
836 No locking. The default.
838 Only one thread or process may do I/O at a time, excluding all
841 Read-write locking on the file. Many readers may
842 access the file at the same time, but writes get exclusive access.
844 .. option:: nrfiles=int
846 Number of files to use for this job. Defaults to 1. The size of files
847 will be :option:`size` divided by this unless explicit size is specified by
848 :option:`filesize`. Files are created for each thread separately, and each
849 file will have a file number within its name by default, as explained in
850 :option:`filename` section.
853 .. option:: openfiles=int
855 Number of files to keep open at the same time. Defaults to the same as
856 :option:`nrfiles`, can be set smaller to limit the number simultaneous
859 .. option:: file_service_type=str
861 Defines how fio decides which file from a job to service next. The following
865 Choose a file at random.
868 Round robin over opened files. This is the default.
871 Finish one file before moving on to the next. Multiple files can
872 still be open depending on :option:`openfiles`.
875 Use a *Zipf* distribution to decide what file to access.
878 Use a *Pareto* distribution to decide what file to access.
881 Use a *Gaussian* (normal) distribution to decide what file to
887 For *random*, *roundrobin*, and *sequential*, a postfix can be appended to
888 tell fio how many I/Os to issue before switching to a new file. For example,
889 specifying ``file_service_type=random:8`` would cause fio to issue
890 8 I/Os before selecting a new file at random. For the non-uniform
891 distributions, a floating point postfix can be given to influence how the
892 distribution is skewed. See :option:`random_distribution` for a description
893 of how that would work.
895 .. option:: ioscheduler=str
897 Attempt to switch the device hosting the file to the specified I/O scheduler
900 .. option:: create_serialize=bool
902 If true, serialize the file creation for the jobs. This may be handy to
903 avoid interleaving of data files, which may greatly depend on the filesystem
904 used and even the number of processors in the system. Default: true.
906 .. option:: create_fsync=bool
908 :manpage:`fsync(2)` the data file after creation. This is the default.
910 .. option:: create_on_open=bool
912 If true, don't pre-create files but allow the job's open() to create a file
913 when it's time to do I/O. Default: false -- pre-create all necessary files
916 .. option:: create_only=bool
918 If true, fio will only run the setup phase of the job. If files need to be
919 laid out or updated on disk, only that will be done -- the actual job contents
920 are not executed. Default: false.
922 .. option:: allow_file_create=bool
924 If true, fio is permitted to create files as part of its workload. If this
925 option is false, then fio will error out if
926 the files it needs to use don't already exist. Default: true.
928 .. option:: allow_mounted_write=bool
930 If this isn't set, fio will abort jobs that are destructive (e.g. that write)
931 to what appears to be a mounted device or partition. This should help catch
932 creating inadvertently destructive tests, not realizing that the test will
933 destroy data on the mounted file system. Note that some platforms don't allow
934 writing against a mounted device regardless of this option. Default: false.
936 .. option:: pre_read=bool
938 If this is given, files will be pre-read into memory before starting the
939 given I/O operation. This will also clear the :option:`invalidate` flag,
940 since it is pointless to pre-read and then drop the cache. This will only
941 work for I/O engines that are seek-able, since they allow you to read the
942 same data multiple times. Thus it will not work on non-seekable I/O engines
943 (e.g. network, splice). Default: false.
945 .. option:: unlink=bool
947 Unlink the job files when done. Not the default, as repeated runs of that
948 job would then waste time recreating the file set again and again. Default:
951 .. option:: unlink_each_loop=bool
953 Unlink job files after each iteration or loop. Default: false.
955 .. option:: zonesize=int
957 Divide a file into zones of the specified size. See :option:`zoneskip`.
959 .. option:: zonerange=int
961 Give size of an I/O zone. See :option:`zoneskip`.
963 .. option:: zoneskip=int
965 Skip the specified number of bytes when :option:`zonesize` data has been
966 read. The two zone options can be used to only do I/O on zones of a file.
972 .. option:: direct=bool
974 If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that
975 OpenBSD and ZFS on Solaris don't support direct I/O. On Windows the synchronous
976 ioengines don't support direct I/O. Default: false.
978 .. option:: atomic=bool
980 If value is true, attempt to use atomic direct I/O. Atomic writes are
981 guaranteed to be stable once acknowledged by the operating system. Only
982 Linux supports O_ATOMIC right now.
984 .. option:: buffered=bool
986 If value is true, use buffered I/O. This is the opposite of the
987 :option:`direct` option. Defaults to true.
989 .. option:: readwrite=str, rw=str
991 Type of I/O pattern. Accepted values are:
998 Sequential trims (Linux block devices and SCSI
999 character devices only).
1005 Random trims (Linux block devices and SCSI
1006 character devices only).
1008 Sequential mixed reads and writes.
1010 Random mixed reads and writes.
1012 Sequential trim+write sequences. Blocks will be trimmed first,
1013 then the same blocks will be written to.
1015 Fio defaults to read if the option is not specified. For the mixed I/O
1016 types, the default is to split them 50/50. For certain types of I/O the
1017 result may still be skewed a bit, since the speed may be different.
1019 It is possible to specify the number of I/Os to do before getting a new
1020 offset by appending ``:<nr>`` to the end of the string given. For a
1021 random read, it would look like ``rw=randread:8`` for passing in an offset
1022 modifier with a value of 8. If the suffix is used with a sequential I/O
1023 pattern, then the *<nr>* value specified will be **added** to the generated
1024 offset for each I/O turning sequential I/O into sequential I/O with holes.
1025 For instance, using ``rw=write:4k`` will skip 4k for every write. Also see
1026 the :option:`rw_sequencer` option.
1028 .. option:: rw_sequencer=str
1030 If an offset modifier is given by appending a number to the ``rw=<str>``
1031 line, then this option controls how that number modifies the I/O offset
1032 being generated. Accepted values are:
1035 Generate sequential offset.
1037 Generate the same offset.
1039 ``sequential`` is only useful for random I/O, where fio would normally
1040 generate a new random offset for every I/O. If you append e.g. 8 to randread,
1041 you would get a new random offset for every 8 I/Os. The result would be a
1042 seek for only every 8 I/Os, instead of for every I/O. Use ``rw=randread:8``
1043 to specify that. As sequential I/O is already sequential, setting
1044 ``sequential`` for that would not result in any differences. ``identical``
1045 behaves in a similar fashion, except it sends the same offset 8 number of
1046 times before generating a new offset.
1048 .. option:: unified_rw_reporting=bool
1050 Fio normally reports statistics on a per data direction basis, meaning that
1051 reads, writes, and trims are accounted and reported separately. If this
1052 option is set fio sums the results and report them as "mixed" instead.
1054 .. option:: randrepeat=bool
1056 Seed the random number generator used for random I/O patterns in a
1057 predictable way so the pattern is repeatable across runs. Default: true.
1059 .. option:: allrandrepeat=bool
1061 Seed all random number generators in a predictable way so results are
1062 repeatable across runs. Default: false.
1064 .. option:: randseed=int
1066 Seed the random number generators based on this seed value, to be able to
1067 control what sequence of output is being generated. If not set, the random
1068 sequence depends on the :option:`randrepeat` setting.
1070 .. option:: fallocate=str
1072 Whether pre-allocation is performed when laying down files.
1073 Accepted values are:
1076 Do not pre-allocate space.
1079 Use a platform's native pre-allocation call but fall back to
1080 **none** behavior if it fails/is not implemented.
1083 Pre-allocate via :manpage:`posix_fallocate(3)`.
1086 Pre-allocate via :manpage:`fallocate(2)` with
1087 FALLOC_FL_KEEP_SIZE set.
1090 Backward-compatible alias for **none**.
1093 Backward-compatible alias for **posix**.
1095 May not be available on all supported platforms. **keep** is only available
1096 on Linux. If using ZFS on Solaris this cannot be set to **posix**
1097 because ZFS doesn't support pre-allocation. Default: **native** if any
1098 pre-allocation methods are available, **none** if not.
1100 .. option:: fadvise_hint=str
1102 Use :manpage:`posix_fadvise(2)` or :manpage:`posix_fadvise(2)` to
1103 advise the kernel on what I/O patterns are likely to be issued.
1104 Accepted values are:
1107 Backwards-compatible hint for "no hint".
1110 Backwards compatible hint for "advise with fio workload type". This
1111 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1112 for a sequential workload.
1115 Advise using **FADV_SEQUENTIAL**.
1118 Advise using **FADV_RANDOM**.
1120 .. option:: write_hint=str
1122 Use :manpage:`fcntl(2)` to advise the kernel what life time to expect
1123 from a write. Only supported on Linux, as of version 4.13. Accepted
1127 No particular life time associated with this file.
1130 Data written to this file has a short life time.
1133 Data written to this file has a medium life time.
1136 Data written to this file has a long life time.
1139 Data written to this file has a very long life time.
1141 The values are all relative to each other, and no absolute meaning
1142 should be associated with them.
1144 .. option:: offset=int
1146 Start I/O at the provided offset in the file, given as either a fixed size in
1147 bytes or a percentage. If a percentage is given, the generated offset will be
1148 aligned to the minimum ``blocksize`` or to the value of ``offset_align`` if
1149 provided. Data before the given offset will not be touched. This
1150 effectively caps the file size at `real_size - offset`. Can be combined with
1151 :option:`size` to constrain the start and end range of the I/O workload.
1152 A percentage can be specified by a number between 1 and 100 followed by '%',
1153 for example, ``offset=20%`` to specify 20%.
1155 .. option:: offset_align=int
1157 If set to non-zero value, the byte offset generated by a percentage ``offset``
1158 is aligned upwards to this value. Defaults to 0 meaning that a percentage
1159 offset is aligned to the minimum block size.
1161 .. option:: offset_increment=int
1163 If this is provided, then the real offset becomes `offset + offset_increment
1164 * thread_number`, where the thread number is a counter that starts at 0 and
1165 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1166 specified). This option is useful if there are several jobs which are
1167 intended to operate on a file in parallel disjoint segments, with even
1168 spacing between the starting points.
1170 .. option:: number_ios=int
1172 Fio will normally perform I/Os until it has exhausted the size of the region
1173 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1174 condition). With this setting, the range/size can be set independently of
1175 the number of I/Os to perform. When fio reaches this number, it will exit
1176 normally and report status. Note that this does not extend the amount of I/O
1177 that will be done, it will only stop fio if this condition is met before
1178 other end-of-job criteria.
1180 .. option:: fsync=int
1182 If writing to a file, issue an :manpage:`fsync(2)` (or its equivalent) of
1183 the dirty data for every number of blocks given. For example, if you give 32
1184 as a parameter, fio will sync the file after every 32 writes issued. If fio is
1185 using non-buffered I/O, we may not sync the file. The exception is the sg
1186 I/O engine, which synchronizes the disk cache anyway. Defaults to 0, which
1187 means fio does not periodically issue and wait for a sync to complete. Also
1188 see :option:`end_fsync` and :option:`fsync_on_close`.
1190 .. option:: fdatasync=int
1192 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
1193 not metadata blocks. In Windows, FreeBSD, and DragonFlyBSD there is no
1194 :manpage:`fdatasync(2)` so this falls back to using :manpage:`fsync(2)`.
1195 Defaults to 0, which means fio does not periodically issue and wait for a
1196 data-only sync to complete.
1198 .. option:: write_barrier=int
1200 Make every `N-th` write a barrier write.
1202 .. option:: sync_file_range=str:int
1204 Use :manpage:`sync_file_range(2)` for every `int` number of write
1205 operations. Fio will track range of writes that have happened since the last
1206 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1209 SYNC_FILE_RANGE_WAIT_BEFORE
1211 SYNC_FILE_RANGE_WRITE
1213 SYNC_FILE_RANGE_WAIT_AFTER
1215 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1216 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1217 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1220 .. option:: overwrite=bool
1222 If true, writes to a file will always overwrite existing data. If the file
1223 doesn't already exist, it will be created before the write phase begins. If
1224 the file exists and is large enough for the specified write phase, nothing
1225 will be done. Default: false.
1227 .. option:: end_fsync=bool
1229 If true, :manpage:`fsync(2)` file contents when a write stage has completed.
1232 .. option:: fsync_on_close=bool
1234 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
1235 from :option:`end_fsync` in that it will happen on every file close, not
1236 just at the end of the job. Default: false.
1238 .. option:: rwmixread=int
1240 Percentage of a mixed workload that should be reads. Default: 50.
1242 .. option:: rwmixwrite=int
1244 Percentage of a mixed workload that should be writes. If both
1245 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1246 add up to 100%, the latter of the two will be used to override the
1247 first. This may interfere with a given rate setting, if fio is asked to
1248 limit reads or writes to a certain rate. If that is the case, then the
1249 distribution may be skewed. Default: 50.
1251 .. option:: random_distribution=str:float[,str:float][,str:float]
1253 By default, fio will use a completely uniform random distribution when asked
1254 to perform random I/O. Sometimes it is useful to skew the distribution in
1255 specific ways, ensuring that some parts of the data is more hot than others.
1256 fio includes the following distribution models:
1259 Uniform random distribution
1268 Normal (Gaussian) distribution
1271 Zoned random distribution
1274 Zone absolute random distribution
1276 When using a **zipf** or **pareto** distribution, an input value is also
1277 needed to define the access pattern. For **zipf**, this is the `Zipf
1278 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
1279 program, :command:`fio-genzipf`, that can be used visualize what the given input
1280 values will yield in terms of hit rates. If you wanted to use **zipf** with
1281 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1282 option. If a non-uniform model is used, fio will disable use of the random
1283 map. For the **normal** distribution, a normal (Gaussian) deviation is
1284 supplied as a value between 0 and 100.
1286 For a **zoned** distribution, fio supports specifying percentages of I/O
1287 access that should fall within what range of the file or device. For
1288 example, given a criteria of:
1290 * 60% of accesses should be to the first 10%
1291 * 30% of accesses should be to the next 20%
1292 * 8% of accesses should be to the next 30%
1293 * 2% of accesses should be to the next 40%
1295 we can define that through zoning of the random accesses. For the above
1296 example, the user would do::
1298 random_distribution=zoned:60/10:30/20:8/30:2/40
1300 A **zoned_abs** distribution works exactly like the **zoned**, except
1301 that it takes absolute sizes. For example, let's say you wanted to
1302 define access according to the following criteria:
1304 * 60% of accesses should be to the first 20G
1305 * 30% of accesses should be to the next 100G
1306 * 10% of accesses should be to the next 500G
1308 we can define an absolute zoning distribution with:
1310 random_distribution=zoned_abs=60/20G:30/100G:10/500g
1312 For both **zoned** and **zoned_abs**, fio supports defining up to
1315 Similarly to how :option:`bssplit` works for setting ranges and
1316 percentages of block sizes. Like :option:`bssplit`, it's possible to
1317 specify separate zones for reads, writes, and trims. If just one set
1318 is given, it'll apply to all of them. This goes for both **zoned**
1319 **zoned_abs** distributions.
1321 .. option:: percentage_random=int[,int][,int]
1323 For a random workload, set how big a percentage should be random. This
1324 defaults to 100%, in which case the workload is fully random. It can be set
1325 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1326 sequential. Any setting in between will result in a random mix of sequential
1327 and random I/O, at the given percentages. Comma-separated values may be
1328 specified for reads, writes, and trims as described in :option:`blocksize`.
1330 .. option:: norandommap
1332 Normally fio will cover every block of the file when doing random I/O. If
1333 this option is given, fio will just get a new random offset without looking
1334 at past I/O history. This means that some blocks may not be read or written,
1335 and that some blocks may be read/written more than once. If this option is
1336 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1337 only intact blocks are verified, i.e., partially-overwritten blocks are
1338 ignored. With an async I/O engine and an I/O depth > 1, it is possible for
1339 the same block to be overwritten, which can cause verification errors. Either
1340 do not use norandommap in this case, or also use the lfsr random generator.
1342 .. option:: softrandommap=bool
1344 See :option:`norandommap`. If fio runs with the random block map enabled and
1345 it fails to allocate the map, if this option is set it will continue without
1346 a random block map. As coverage will not be as complete as with random maps,
1347 this option is disabled by default.
1349 .. option:: random_generator=str
1351 Fio supports the following engines for generating I/O offsets for random I/O:
1354 Strong 2^88 cycle random number generator.
1356 Linear feedback shift register generator.
1358 Strong 64-bit 2^258 cycle random number generator.
1360 **tausworthe** is a strong random number generator, but it requires tracking
1361 on the side if we want to ensure that blocks are only read or written
1362 once. **lfsr** guarantees that we never generate the same offset twice, and
1363 it's also less computationally expensive. It's not a true random generator,
1364 however, though for I/O purposes it's typically good enough. **lfsr** only
1365 works with single block sizes, not with workloads that use multiple block
1366 sizes. If used with such a workload, fio may read or write some blocks
1367 multiple times. The default value is **tausworthe**, unless the required
1368 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1369 selected automatically.
1375 .. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1377 The block size in bytes used for I/O units. Default: 4096. A single value
1378 applies to reads, writes, and trims. Comma-separated values may be
1379 specified for reads, writes, and trims. A value not terminated in a comma
1380 applies to subsequent types.
1385 means 256k for reads, writes and trims.
1388 means 8k for reads, 32k for writes and trims.
1391 means 8k for reads, 32k for writes, and default for trims.
1394 means default for reads, 8k for writes and trims.
1397 means default for reads, 8k for writes, and default for trims.
1399 .. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1401 A range of block sizes in bytes for I/O units. The issued I/O unit will
1402 always be a multiple of the minimum size, unless
1403 :option:`blocksize_unaligned` is set.
1405 Comma-separated ranges may be specified for reads, writes, and trims as
1406 described in :option:`blocksize`.
1408 Example: ``bsrange=1k-4k,2k-8k``.
1410 .. option:: bssplit=str[,str][,str]
1412 Sometimes you want even finer grained control of the block sizes
1413 issued, not just an even split between them. This option allows you to
1414 weight various block sizes, so that you are able to define a specific
1415 amount of block sizes issued. The format for this option is::
1417 bssplit=blocksize/percentage:blocksize/percentage
1419 for as many block sizes as needed. So if you want to define a workload
1420 that has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would
1423 bssplit=4k/10:64k/50:32k/40
1425 Ordering does not matter. If the percentage is left blank, fio will
1426 fill in the remaining values evenly. So a bssplit option like this one::
1428 bssplit=4k/50:1k/:32k/
1430 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always
1431 add up to 100, if bssplit is given a range that adds up to more, it
1434 Comma-separated values may be specified for reads, writes, and trims as
1435 described in :option:`blocksize`.
1437 If you want a workload that has 50% 2k reads and 50% 4k reads, while
1438 having 90% 4k writes and 10% 8k writes, you would specify::
1440 bssplit=2k/50:4k/50,4k/90:8k/10
1442 Fio supports defining up to 64 different weights for each data
1445 .. option:: blocksize_unaligned, bs_unaligned
1447 If set, fio will issue I/O units with any size within
1448 :option:`blocksize_range`, not just multiples of the minimum size. This
1449 typically won't work with direct I/O, as that normally requires sector
1452 .. option:: bs_is_seq_rand=bool
1454 If this option is set, fio will use the normal read,write blocksize settings
1455 as sequential,random blocksize settings instead. Any random read or write
1456 will use the WRITE blocksize settings, and any sequential read or write will
1457 use the READ blocksize settings.
1459 .. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1461 Boundary to which fio will align random I/O units. Default:
1462 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1463 I/O, though it usually depends on the hardware block size. This option is
1464 mutually exclusive with using a random map for files, so it will turn off
1465 that option. Comma-separated values may be specified for reads, writes, and
1466 trims as described in :option:`blocksize`.
1472 .. option:: zero_buffers
1474 Initialize buffers with all zeros. Default: fill buffers with random data.
1476 .. option:: refill_buffers
1478 If this option is given, fio will refill the I/O buffers on every
1479 submit. Only makes sense if :option:`zero_buffers` isn't specified,
1480 naturally. Defaults to being unset i.e., the buffer is only filled at
1481 init time and the data in it is reused when possible but if any of
1482 :option:`verify`, :option:`buffer_compress_percentage` or
1483 :option:`dedupe_percentage` are enabled then `refill_buffers` is also
1484 automatically enabled.
1486 .. option:: scramble_buffers=bool
1488 If :option:`refill_buffers` is too costly and the target is using data
1489 deduplication, then setting this option will slightly modify the I/O buffer
1490 contents to defeat normal de-dupe attempts. This is not enough to defeat
1491 more clever block compression attempts, but it will stop naive dedupe of
1492 blocks. Default: true.
1494 .. option:: buffer_compress_percentage=int
1496 If this is set, then fio will attempt to provide I/O buffer content
1497 (on WRITEs) that compresses to the specified level. Fio does this by
1498 providing a mix of random data followed by fixed pattern data. The
1499 fixed pattern is either zeros, or the pattern specified by
1500 :option:`buffer_pattern`. If the `buffer_pattern` option is used, it
1501 might skew the compression ratio slightly. Setting
1502 `buffer_compress_percentage` to a value other than 100 will also
1503 enable :option:`refill_buffers` in order to reduce the likelihood that
1504 adjacent blocks are so similar that they over compress when seen
1505 together. See :option:`buffer_compress_chunk` for how to set a finer or
1506 coarser granularity for the random/fixed data region. Defaults to unset
1507 i.e., buffer data will not adhere to any compression level.
1509 .. option:: buffer_compress_chunk=int
1511 This setting allows fio to manage how big the random/fixed data region
1512 is when using :option:`buffer_compress_percentage`. When
1513 `buffer_compress_chunk` is set to some non-zero value smaller than the
1514 block size, fio can repeat the random/fixed region throughout the I/O
1515 buffer at the specified interval (which particularly useful when
1516 bigger block sizes are used for a job). When set to 0, fio will use a
1517 chunk size that matches the block size resulting in a single
1518 random/fixed region within the I/O buffer. Defaults to 512. When the
1519 unit is omitted, the value is interpreted in bytes.
1521 .. option:: buffer_pattern=str
1523 If set, fio will fill the I/O buffers with this pattern or with the contents
1524 of a file. If not set, the contents of I/O buffers are defined by the other
1525 options related to buffer contents. The setting can be any pattern of bytes,
1526 and can be prefixed with 0x for hex values. It may also be a string, where
1527 the string must then be wrapped with ``""``. Or it may also be a filename,
1528 where the filename must be wrapped with ``''`` in which case the file is
1529 opened and read. Note that not all the file contents will be read if that
1530 would cause the buffers to overflow. So, for example::
1532 buffer_pattern='filename'
1536 buffer_pattern="abcd"
1544 buffer_pattern=0xdeadface
1546 Also you can combine everything together in any order::
1548 buffer_pattern=0xdeadface"abcd"-12'filename'
1550 .. option:: dedupe_percentage=int
1552 If set, fio will generate this percentage of identical buffers when
1553 writing. These buffers will be naturally dedupable. The contents of the
1554 buffers depend on what other buffer compression settings have been set. It's
1555 possible to have the individual buffers either fully compressible, or not at
1556 all -- this option only controls the distribution of unique buffers. Setting
1557 this option will also enable :option:`refill_buffers` to prevent every buffer
1560 .. option:: invalidate=bool
1562 Invalidate the buffer/page cache parts of the files to be used prior to
1563 starting I/O if the platform and file type support it. Defaults to true.
1564 This will be ignored if :option:`pre_read` is also specified for the
1567 .. option:: sync=bool
1569 Use synchronous I/O for buffered writes. For the majority of I/O engines,
1570 this means using O_SYNC. Default: false.
1572 .. option:: iomem=str, mem=str
1574 Fio can use various types of memory as the I/O unit buffer. The allowed
1578 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1582 Use shared memory as the buffers. Allocated through
1583 :manpage:`shmget(2)`.
1586 Same as shm, but use huge pages as backing.
1589 Use :manpage:`mmap(2)` to allocate buffers. May either be anonymous memory, or can
1590 be file backed if a filename is given after the option. The format
1591 is `mem=mmap:/path/to/file`.
1594 Use a memory mapped huge file as the buffer backing. Append filename
1595 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1598 Same as mmap, but use a MMAP_SHARED mapping.
1601 Use GPU memory as the buffers for GPUDirect RDMA benchmark.
1602 The :option:`ioengine` must be `rdma`.
1604 The area allocated is a function of the maximum allowed bs size for the job,
1605 multiplied by the I/O depth given. Note that for **shmhuge** and
1606 **mmaphuge** to work, the system must have free huge pages allocated. This
1607 can normally be checked and set by reading/writing
1608 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1609 is 4MiB in size. So to calculate the number of huge pages you need for a
1610 given job file, add up the I/O depth of all jobs (normally one unless
1611 :option:`iodepth` is used) and multiply by the maximum bs set. Then divide
1612 that number by the huge page size. You can see the size of the huge pages in
1613 :file:`/proc/meminfo`. If no huge pages are allocated by having a non-zero
1614 number in `nr_hugepages`, using **mmaphuge** or **shmhuge** will fail. Also
1615 see :option:`hugepage-size`.
1617 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1618 should point there. So if it's mounted in :file:`/huge`, you would use
1619 `mem=mmaphuge:/huge/somefile`.
1621 .. option:: iomem_align=int, mem_align=int
1623 This indicates the memory alignment of the I/O memory buffers. Note that
1624 the given alignment is applied to the first I/O unit buffer, if using
1625 :option:`iodepth` the alignment of the following buffers are given by the
1626 :option:`bs` used. In other words, if using a :option:`bs` that is a
1627 multiple of the page sized in the system, all buffers will be aligned to
1628 this value. If using a :option:`bs` that is not page aligned, the alignment
1629 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1632 .. option:: hugepage-size=int
1634 Defines the size of a huge page. Must at least be equal to the system
1635 setting, see :file:`/proc/meminfo`. Defaults to 4MiB. Should probably
1636 always be a multiple of megabytes, so using ``hugepage-size=Xm`` is the
1637 preferred way to set this to avoid setting a non-pow-2 bad value.
1639 .. option:: lockmem=int
1641 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1642 simulate a smaller amount of memory. The amount specified is per worker.
1648 .. option:: size=int
1650 The total size of file I/O for each thread of this job. Fio will run until
1651 this many bytes has been transferred, unless runtime is limited by other options
1652 (such as :option:`runtime`, for instance, or increased/decreased by :option:`io_size`).
1653 Fio will divide this size between the available files determined by options
1654 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1655 specified by the job. If the result of division happens to be 0, the size is
1656 set to the physical size of the given files or devices if they exist.
1657 If this option is not specified, fio will use the full size of the given
1658 files or devices. If the files do not exist, size must be given. It is also
1659 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1660 given, fio will use 20% of the full size of the given files or devices.
1661 Can be combined with :option:`offset` to constrain the start and end range
1662 that I/O will be done within.
1664 .. option:: io_size=int, io_limit=int
1666 Normally fio operates within the region set by :option:`size`, which means
1667 that the :option:`size` option sets both the region and size of I/O to be
1668 performed. Sometimes that is not what you want. With this option, it is
1669 possible to define just the amount of I/O that fio should do. For instance,
1670 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1671 will perform I/O within the first 20GiB but exit when 5GiB have been
1672 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1673 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1674 the 0..20GiB region.
1676 .. option:: filesize=irange(int)
1678 Individual file sizes. May be a range, in which case fio will select sizes
1679 for files at random within the given range and limited to :option:`size` in
1680 total (if that is given). If not given, each created file is the same size.
1681 This option overrides :option:`size` in terms of file size, which means
1682 this value is used as a fixed size or possible range of each file.
1684 .. option:: file_append=bool
1686 Perform I/O after the end of the file. Normally fio will operate within the
1687 size of a file. If this option is set, then fio will append to the file
1688 instead. This has identical behavior to setting :option:`offset` to the size
1689 of a file. This option is ignored on non-regular files.
1691 .. option:: fill_device=bool, fill_fs=bool
1693 Sets size to something really large and waits for ENOSPC (no space left on
1694 device) as the terminating condition. Only makes sense with sequential
1695 write. For a read workload, the mount point will be filled first then I/O
1696 started on the result. This option doesn't make sense if operating on a raw
1697 device node, since the size of that is already known by the file system.
1698 Additionally, writing beyond end-of-device will not return ENOSPC there.
1704 .. option:: ioengine=str
1706 Defines how the job issues I/O to the file. The following types are defined:
1709 Basic :manpage:`read(2)` or :manpage:`write(2)`
1710 I/O. :manpage:`lseek(2)` is used to position the I/O location.
1711 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
1714 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1715 all supported operating systems except for Windows.
1718 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
1719 queuing by coalescing adjacent I/Os into a single submission.
1722 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1725 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1728 Linux native asynchronous I/O. Note that Linux may only support
1729 queued behavior with non-buffered I/O (set ``direct=1`` or
1731 This engine defines engine specific options.
1734 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
1735 :manpage:`aio_write(3)`.
1738 Solaris native asynchronous I/O.
1741 Windows native asynchronous I/O. Default on Windows.
1744 File is memory mapped with :manpage:`mmap(2)` and data copied
1745 to/from using :manpage:`memcpy(3)`.
1748 :manpage:`splice(2)` is used to transfer the data and
1749 :manpage:`vmsplice(2)` to transfer data from user space to the
1753 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
1754 ioctl, or if the target is an sg character device we use
1755 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
1756 I/O. Requires :option:`filename` option to specify either block or
1757 character devices. This engine supports trim operations.
1758 The sg engine includes engine specific options.
1761 Doesn't transfer any data, just pretends to. This is mainly used to
1762 exercise fio itself and for debugging/testing purposes.
1765 Transfer over the network to given ``host:port``. Depending on the
1766 :option:`protocol` used, the :option:`hostname`, :option:`port`,
1767 :option:`listen` and :option:`filename` options are used to specify
1768 what sort of connection to make, while the :option:`protocol` option
1769 determines which protocol will be used. This engine defines engine
1773 Like **net**, but uses :manpage:`splice(2)` and
1774 :manpage:`vmsplice(2)` to map data and send/receive.
1775 This engine defines engine specific options.
1778 Doesn't transfer any data, but burns CPU cycles according to the
1779 :option:`cpuload` and :option:`cpuchunks` options. Setting
1780 :option:`cpuload`\=85 will cause that job to do nothing but burn 85%
1781 of the CPU. In case of SMP machines, use :option:`numjobs`\=<nr_of_cpu>
1782 to get desired CPU usage, as the cpuload only loads a
1783 single CPU at the desired rate. A job never finishes unless there is
1784 at least one non-cpuio job.
1787 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall
1788 Interface approach to async I/O. See
1790 http://www.xmailserver.org/guasi-lib.html
1792 for more info on GUASI.
1795 The RDMA I/O engine supports both RDMA memory semantics
1796 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
1797 InfiniBand, RoCE and iWARP protocols. This engine defines engine
1801 I/O engine that does regular fallocate to simulate data transfer as
1805 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
1808 does fallocate(,mode = 0).
1811 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
1814 I/O engine that sends :manpage:`ftruncate(2)` operations in response
1815 to write (DDIR_WRITE) events. Each ftruncate issued sets the file's
1816 size to the current block offset. :option:`blocksize` is ignored.
1819 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
1820 defragment activity in request to DDIR_WRITE event.
1823 I/O engine supporting direct access to Ceph Reliable Autonomic
1824 Distributed Object Store (RADOS) via librados. This ioengine
1825 defines engine specific options.
1828 I/O engine supporting direct access to Ceph Rados Block Devices
1829 (RBD) via librbd without the need to use the kernel rbd driver. This
1830 ioengine defines engine specific options.
1833 I/O engine supporting GET/PUT requests over HTTP(S) with libcurl to
1834 a WebDAV or S3 endpoint. This ioengine defines engine specific options.
1836 This engine only supports direct IO of iodepth=1; you need to scale this
1837 via numjobs. blocksize defines the size of the objects to be created.
1839 TRIM is translated to object deletion.
1842 Using GlusterFS libgfapi sync interface to direct access to
1843 GlusterFS volumes without having to go through FUSE. This ioengine
1844 defines engine specific options.
1847 Using GlusterFS libgfapi async interface to direct access to
1848 GlusterFS volumes without having to go through FUSE. This ioengine
1849 defines engine specific options.
1852 Read and write through Hadoop (HDFS). The :option:`filename` option
1853 is used to specify host,port of the hdfs name-node to connect. This
1854 engine interprets offsets a little differently. In HDFS, files once
1855 created cannot be modified so random writes are not possible. To
1856 imitate this the libhdfs engine expects a bunch of small files to be
1857 created over HDFS and will randomly pick a file from them
1858 based on the offset generated by fio backend (see the example
1859 job file to create such files, use ``rw=write`` option). Please
1860 note, it may be necessary to set environment variables to work
1861 with HDFS/libhdfs properly. Each job uses its own connection to
1865 Read, write and erase an MTD character device (e.g.,
1866 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
1867 underlying device type, the I/O may have to go in a certain pattern,
1868 e.g., on NAND, writing sequentially to erase blocks and discarding
1869 before overwriting. The `trimwrite` mode works well for this
1873 Read and write using filesystem DAX to a file on a filesystem
1874 mounted with DAX on a persistent memory device through the PMDK
1878 Read and write using device DAX to a persistent memory device (e.g.,
1879 /dev/dax0.0) through the PMDK libpmem library.
1882 Prefix to specify loading an external I/O engine object file. Append
1883 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
1884 ioengine :file:`foo.o` in :file:`/tmp`. The path can be either
1885 absolute or relative. See :file:`engines/skeleton_external.c` for
1886 details of writing an external I/O engine.
1889 Simply create the files and do no I/O to them. You still need to
1890 set `filesize` so that all the accounting still occurs, but no
1891 actual I/O will be done other than creating the file.
1894 Read and write using mmap I/O to a file on a filesystem
1895 mounted with DAX on a persistent memory device through the PMDK
1898 I/O engine specific parameters
1899 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1901 In addition, there are some parameters which are only valid when a specific
1902 :option:`ioengine` is in use. These are used identically to normal parameters,
1903 with the caveat that when used on the command line, they must come after the
1904 :option:`ioengine` that defines them is selected.
1906 .. option:: userspace_reap : [libaio]
1908 Normally, with the libaio engine in use, fio will use the
1909 :manpage:`io_getevents(2)` system call to reap newly returned events. With
1910 this flag turned on, the AIO ring will be read directly from user-space to
1911 reap events. The reaping mode is only enabled when polling for a minimum of
1912 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
1914 .. option:: hipri : [pvsync2]
1916 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
1919 .. option:: hipri_percentage : [pvsync2]
1921 When hipri is set this determines the probability of a pvsync2 I/O being high
1922 priority. The default is 100%.
1924 .. option:: cpuload=int : [cpuio]
1926 Attempt to use the specified percentage of CPU cycles. This is a mandatory
1927 option when using cpuio I/O engine.
1929 .. option:: cpuchunks=int : [cpuio]
1931 Split the load into cycles of the given time. In microseconds.
1933 .. option:: exit_on_io_done=bool : [cpuio]
1935 Detect when I/O threads are done, then exit.
1937 .. option:: namenode=str : [libhdfs]
1939 The hostname or IP address of a HDFS cluster namenode to contact.
1941 .. option:: port=int
1945 The listening port of the HFDS cluster namenode.
1949 The TCP or UDP port to bind to or connect to. If this is used with
1950 :option:`numjobs` to spawn multiple instances of the same job type, then
1951 this will be the starting port number since fio will use a range of
1956 The port to use for RDMA-CM communication. This should be the same value
1957 on the client and the server side.
1959 .. option:: hostname=str : [netsplice] [net] [rdma]
1961 The hostname or IP address to use for TCP, UDP or RDMA-CM based I/O. If the job
1962 is a TCP listener or UDP reader, the hostname is not used and must be omitted
1963 unless it is a valid UDP multicast address.
1965 .. option:: interface=str : [netsplice] [net]
1967 The IP address of the network interface used to send or receive UDP
1970 .. option:: ttl=int : [netsplice] [net]
1972 Time-to-live value for outgoing UDP multicast packets. Default: 1.
1974 .. option:: nodelay=bool : [netsplice] [net]
1976 Set TCP_NODELAY on TCP connections.
1978 .. option:: protocol=str, proto=str : [netsplice] [net]
1980 The network protocol to use. Accepted values are:
1983 Transmission control protocol.
1985 Transmission control protocol V6.
1987 User datagram protocol.
1989 User datagram protocol V6.
1993 When the protocol is TCP or UDP, the port must also be given, as well as the
1994 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
1995 normal :option:`filename` option should be used and the port is invalid.
1997 .. option:: listen : [netsplice] [net]
1999 For TCP network connections, tell fio to listen for incoming connections
2000 rather than initiating an outgoing connection. The :option:`hostname` must
2001 be omitted if this option is used.
2003 .. option:: pingpong : [netsplice] [net]
2005 Normally a network writer will just continue writing data, and a network
2006 reader will just consume packages. If ``pingpong=1`` is set, a writer will
2007 send its normal payload to the reader, then wait for the reader to send the
2008 same payload back. This allows fio to measure network latencies. The
2009 submission and completion latencies then measure local time spent sending or
2010 receiving, and the completion latency measures how long it took for the
2011 other end to receive and send back. For UDP multicast traffic
2012 ``pingpong=1`` should only be set for a single reader when multiple readers
2013 are listening to the same address.
2015 .. option:: window_size : [netsplice] [net]
2017 Set the desired socket buffer size for the connection.
2019 .. option:: mss : [netsplice] [net]
2021 Set the TCP maximum segment size (TCP_MAXSEG).
2023 .. option:: donorname=str : [e4defrag]
2025 File will be used as a block donor (swap extents between files).
2027 .. option:: inplace=int : [e4defrag]
2029 Configure donor file blocks allocation strategy:
2032 Default. Preallocate donor's file on init.
2034 Allocate space immediately inside defragment event, and free right
2037 .. option:: clustername=str : [rbd,rados]
2039 Specifies the name of the Ceph cluster.
2041 .. option:: rbdname=str : [rbd]
2043 Specifies the name of the RBD.
2045 .. option:: pool=str : [rbd,rados]
2047 Specifies the name of the Ceph pool containing RBD or RADOS data.
2049 .. option:: clientname=str : [rbd,rados]
2051 Specifies the username (without the 'client.' prefix) used to access the
2052 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
2053 the full *type.id* string. If no type. prefix is given, fio will add
2054 'client.' by default.
2056 .. option:: busy_poll=bool : [rbd,rados]
2058 Poll store instead of waiting for completion. Usually this provides better
2059 throughput at cost of higher(up to 100%) CPU utilization.
2061 .. option:: skip_bad=bool : [mtd]
2063 Skip operations against known bad blocks.
2065 .. option:: hdfsdirectory : [libhdfs]
2067 libhdfs will create chunk in this HDFS directory.
2069 .. option:: chunk_size : [libhdfs]
2071 The size of the chunk to use for each file.
2073 .. option:: verb=str : [rdma]
2075 The RDMA verb to use on this side of the RDMA ioengine connection. Valid
2076 values are write, read, send and recv. These correspond to the equivalent
2077 RDMA verbs (e.g. write = rdma_write etc.). Note that this only needs to be
2078 specified on the client side of the connection. See the examples folder.
2080 .. option:: bindname=str : [rdma]
2082 The name to use to bind the local RDMA-CM connection to a local RDMA device.
2083 This could be a hostname or an IPv4 or IPv6 address. On the server side this
2084 will be passed into the rdma_bind_addr() function and on the client site it
2085 will be used in the rdma_resolve_add() function. This can be useful when
2086 multiple paths exist between the client and the server or in certain loopback
2089 .. option:: readfua=bool : [sg]
2091 With readfua option set to 1, read operations include
2092 the force unit access (fua) flag. Default is 0.
2094 .. option:: writefua=bool : [sg]
2096 With writefua option set to 1, write operations include
2097 the force unit access (fua) flag. Default is 0.
2099 .. option:: sg_write_mode=str : [sg]
2101 Specify the type of write commands to issue. This option can take three values:
2104 This is the default where write opcodes are issued as usual.
2106 Issue WRITE AND VERIFY commands. The BYTCHK bit is set to 0. This
2107 directs the device to carry out a medium verification with no data
2108 comparison. The writefua option is ignored with this selection.
2110 Issue WRITE SAME commands. This transfers a single block to the device
2111 and writes this same block of data to a contiguous sequence of LBAs
2112 beginning at the specified offset. fio's block size parameter specifies
2113 the amount of data written with each command. However, the amount of data
2114 actually transferred to the device is equal to the device's block
2115 (sector) size. For a device with 512 byte sectors, blocksize=8k will
2116 write 16 sectors with each command. fio will still generate 8k of data
2117 for each command but only the first 512 bytes will be used and
2118 transferred to the device. The writefua option is ignored with this
2121 .. option:: http_host=str : [http]
2123 Hostname to connect to. For S3, this could be the bucket hostname.
2124 Default is **localhost**
2126 .. option:: http_user=str : [http]
2128 Username for HTTP authentication.
2130 .. option:: http_pass=str : [http]
2132 Password for HTTP authentication.
2134 .. option:: https=bool : [http]
2136 Enable HTTPS instead of http. Default is **0**
2138 .. option:: http_s3=bool : [http]
2140 Enable S3 specific HTTP headers such as authenticating requests
2141 with AWS Signature Version 4. Default is **0**
2143 .. option:: http_s3_region=str : [http]
2145 The S3 region/zone string.
2146 Default is **us-east-1**
2148 .. option:: http_s3_key=str : [http]
2152 .. option:: http_s3_keyid=str : [http]
2154 The S3 key/access id.
2156 .. option:: http_verbose=int : [http]
2158 Enable verbose requests from libcurl. Useful for debugging. 1
2159 turns on verbose logging from libcurl, 2 additionally enables
2160 HTTP IO tracing. Default is **0**
2165 .. option:: iodepth=int
2167 Number of I/O units to keep in flight against the file. Note that
2168 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
2169 for small degrees when :option:`verify_async` is in use). Even async
2170 engines may impose OS restrictions causing the desired depth not to be
2171 achieved. This may happen on Linux when using libaio and not setting
2172 :option:`direct`\=1, since buffered I/O is not async on that OS. Keep an
2173 eye on the I/O depth distribution in the fio output to verify that the
2174 achieved depth is as expected. Default: 1.
2176 .. option:: iodepth_batch_submit=int, iodepth_batch=int
2178 This defines how many pieces of I/O to submit at once. It defaults to 1
2179 which means that we submit each I/O as soon as it is available, but can be
2180 raised to submit bigger batches of I/O at the time. If it is set to 0 the
2181 :option:`iodepth` value will be used.
2183 .. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
2185 This defines how many pieces of I/O to retrieve at once. It defaults to 1
2186 which means that we'll ask for a minimum of 1 I/O in the retrieval process
2187 from the kernel. The I/O retrieval will go on until we hit the limit set by
2188 :option:`iodepth_low`. If this variable is set to 0, then fio will always
2189 check for completed events before queuing more I/O. This helps reduce I/O
2190 latency, at the cost of more retrieval system calls.
2192 .. option:: iodepth_batch_complete_max=int
2194 This defines maximum pieces of I/O to retrieve at once. This variable should
2195 be used along with :option:`iodepth_batch_complete_min`\=int variable,
2196 specifying the range of min and max amount of I/O which should be
2197 retrieved. By default it is equal to the :option:`iodepth_batch_complete_min`
2202 iodepth_batch_complete_min=1
2203 iodepth_batch_complete_max=<iodepth>
2205 which means that we will retrieve at least 1 I/O and up to the whole
2206 submitted queue depth. If none of I/O has been completed yet, we will wait.
2210 iodepth_batch_complete_min=0
2211 iodepth_batch_complete_max=<iodepth>
2213 which means that we can retrieve up to the whole submitted queue depth, but
2214 if none of I/O has been completed yet, we will NOT wait and immediately exit
2215 the system call. In this example we simply do polling.
2217 .. option:: iodepth_low=int
2219 The low water mark indicating when to start filling the queue
2220 again. Defaults to the same as :option:`iodepth`, meaning that fio will
2221 attempt to keep the queue full at all times. If :option:`iodepth` is set to
2222 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
2223 16 requests, it will let the depth drain down to 4 before starting to fill
2226 .. option:: serialize_overlap=bool
2228 Serialize in-flight I/Os that might otherwise cause or suffer from data races.
2229 When two or more I/Os are submitted simultaneously, there is no guarantee that
2230 the I/Os will be processed or completed in the submitted order. Further, if
2231 two or more of those I/Os are writes, any overlapping region between them can
2232 become indeterminate/undefined on certain storage. These issues can cause
2233 verification to fail erratically when at least one of the racing I/Os is
2234 changing data and the overlapping region has a non-zero size. Setting
2235 ``serialize_overlap`` tells fio to avoid provoking this behavior by explicitly
2236 serializing in-flight I/Os that have a non-zero overlap. Note that setting
2237 this option can reduce both performance and the :option:`iodepth` achieved.
2238 Additionally this option does not work when :option:`io_submit_mode` is set to
2239 offload. Default: false.
2241 .. option:: io_submit_mode=str
2243 This option controls how fio submits the I/O to the I/O engine. The default
2244 is `inline`, which means that the fio job threads submit and reap I/O
2245 directly. If set to `offload`, the job threads will offload I/O submission
2246 to a dedicated pool of I/O threads. This requires some coordination and thus
2247 has a bit of extra overhead, especially for lower queue depth I/O where it
2248 can increase latencies. The benefit is that fio can manage submission rates
2249 independently of the device completion rates. This avoids skewed latency
2250 reporting if I/O gets backed up on the device side (the coordinated omission
2257 .. option:: thinktime=time
2259 Stall the job for the specified period of time after an I/O has completed before issuing the
2260 next. May be used to simulate processing being done by an application.
2261 When the unit is omitted, the value is interpreted in microseconds. See
2262 :option:`thinktime_blocks` and :option:`thinktime_spin`.
2264 .. option:: thinktime_spin=time
2266 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
2267 something with the data received, before falling back to sleeping for the
2268 rest of the period specified by :option:`thinktime`. When the unit is
2269 omitted, the value is interpreted in microseconds.
2271 .. option:: thinktime_blocks=int
2273 Only valid if :option:`thinktime` is set - control how many blocks to issue,
2274 before waiting :option:`thinktime` usecs. If not set, defaults to 1 which will make
2275 fio wait :option:`thinktime` usecs after every block. This effectively makes any
2276 queue depth setting redundant, since no more than 1 I/O will be queued
2277 before we have to complete it and do our :option:`thinktime`. In other words, this
2278 setting effectively caps the queue depth if the latter is larger.
2280 .. option:: rate=int[,int][,int]
2282 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
2283 suffix rules apply. Comma-separated values may be specified for reads,
2284 writes, and trims as described in :option:`blocksize`.
2286 For example, using `rate=1m,500k` would limit reads to 1MiB/sec and writes to
2287 500KiB/sec. Capping only reads or writes can be done with `rate=,500k` or
2288 `rate=500k,` where the former will only limit writes (to 500KiB/sec) and the
2289 latter will only limit reads.
2291 .. option:: rate_min=int[,int][,int]
2293 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
2294 to meet this requirement will cause the job to exit. Comma-separated values
2295 may be specified for reads, writes, and trims as described in
2296 :option:`blocksize`.
2298 .. option:: rate_iops=int[,int][,int]
2300 Cap the bandwidth to this number of IOPS. Basically the same as
2301 :option:`rate`, just specified independently of bandwidth. If the job is
2302 given a block size range instead of a fixed value, the smallest block size
2303 is used as the metric. Comma-separated values may be specified for reads,
2304 writes, and trims as described in :option:`blocksize`.
2306 .. option:: rate_iops_min=int[,int][,int]
2308 If fio doesn't meet this rate of I/O, it will cause the job to exit.
2309 Comma-separated values may be specified for reads, writes, and trims as
2310 described in :option:`blocksize`.
2312 .. option:: rate_process=str
2314 This option controls how fio manages rated I/O submissions. The default is
2315 `linear`, which submits I/O in a linear fashion with fixed delays between
2316 I/Os that gets adjusted based on I/O completion rates. If this is set to
2317 `poisson`, fio will submit I/O based on a more real world random request
2318 flow, known as the Poisson process
2319 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
2320 10^6 / IOPS for the given workload.
2322 .. option:: rate_ignore_thinktime=bool
2324 By default, fio will attempt to catch up to the specified rate setting,
2325 if any kind of thinktime setting was used. If this option is set, then
2326 fio will ignore the thinktime and continue doing IO at the specified
2327 rate, instead of entering a catch-up mode after thinktime is done.
2333 .. option:: latency_target=time
2335 If set, fio will attempt to find the max performance point that the given
2336 workload will run at while maintaining a latency below this target. When
2337 the unit is omitted, the value is interpreted in microseconds. See
2338 :option:`latency_window` and :option:`latency_percentile`.
2340 .. option:: latency_window=time
2342 Used with :option:`latency_target` to specify the sample window that the job
2343 is run at varying queue depths to test the performance. When the unit is
2344 omitted, the value is interpreted in microseconds.
2346 .. option:: latency_percentile=float
2348 The percentage of I/Os that must fall within the criteria specified by
2349 :option:`latency_target` and :option:`latency_window`. If not set, this
2350 defaults to 100.0, meaning that all I/Os must be equal or below to the value
2351 set by :option:`latency_target`.
2353 .. option:: max_latency=time
2355 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
2356 maximum latency. When the unit is omitted, the value is interpreted in
2359 .. option:: rate_cycle=int
2361 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
2362 of milliseconds. Defaults to 1000.
2368 .. option:: write_iolog=str
2370 Write the issued I/O patterns to the specified file. See
2371 :option:`read_iolog`. Specify a separate file for each job, otherwise the
2372 iologs will be interspersed and the file may be corrupt.
2374 .. option:: read_iolog=str
2376 Open an iolog with the specified filename and replay the I/O patterns it
2377 contains. This can be used to store a workload and replay it sometime
2378 later. The iolog given may also be a blktrace binary file, which allows fio
2379 to replay a workload captured by :command:`blktrace`. See
2380 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
2381 replay, the file needs to be turned into a blkparse binary data file first
2382 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
2384 .. option:: read_iolog_chunked=bool
2386 Determines how iolog is read. If false(default) entire :option:`read_iolog`
2387 will be read at once. If selected true, input from iolog will be read
2388 gradually. Useful when iolog is very large, or it is generated.
2390 .. option:: replay_no_stall=bool
2392 When replaying I/O with :option:`read_iolog` the default behavior is to
2393 attempt to respect the timestamps within the log and replay them with the
2394 appropriate delay between IOPS. By setting this variable fio will not
2395 respect the timestamps and attempt to replay them as fast as possible while
2396 still respecting ordering. The result is the same I/O pattern to a given
2397 device, but different timings.
2399 .. option:: replay_time_scale=int
2401 When replaying I/O with :option:`read_iolog`, fio will honor the
2402 original timing in the trace. With this option, it's possible to scale
2403 the time. It's a percentage option, if set to 50 it means run at 50%
2404 the original IO rate in the trace. If set to 200, run at twice the
2405 original IO rate. Defaults to 100.
2407 .. option:: replay_redirect=str
2409 While replaying I/O patterns using :option:`read_iolog` the default behavior
2410 is to replay the IOPS onto the major/minor device that each IOP was recorded
2411 from. This is sometimes undesirable because on a different machine those
2412 major/minor numbers can map to a different device. Changing hardware on the
2413 same system can also result in a different major/minor mapping.
2414 ``replay_redirect`` causes all I/Os to be replayed onto the single specified
2415 device regardless of the device it was recorded
2416 from. i.e. :option:`replay_redirect`\= :file:`/dev/sdc` would cause all I/O
2417 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
2418 multiple devices will be replayed onto a single device, if the trace
2419 contains multiple devices. If you want multiple devices to be replayed
2420 concurrently to multiple redirected devices you must blkparse your trace
2421 into separate traces and replay them with independent fio invocations.
2422 Unfortunately this also breaks the strict time ordering between multiple
2425 .. option:: replay_align=int
2427 Force alignment of I/O offsets and lengths in a trace to this power of 2
2430 .. option:: replay_scale=int
2432 Scale sector offsets down by this factor when replaying traces.
2434 .. option:: replay_skip=str
2436 Sometimes it's useful to skip certain IO types in a replay trace.
2437 This could be, for instance, eliminating the writes in the trace.
2438 Or not replaying the trims/discards, if you are redirecting to
2439 a device that doesn't support them. This option takes a comma
2440 separated list of read, write, trim, sync.
2443 Threads, processes and job synchronization
2444 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2448 Fio defaults to creating jobs by using fork, however if this option is
2449 given, fio will create jobs by using POSIX Threads' function
2450 :manpage:`pthread_create(3)` to create threads instead.
2452 .. option:: wait_for=str
2454 If set, the current job won't be started until all workers of the specified
2455 waitee job are done.
2457 ``wait_for`` operates on the job name basis, so there are a few
2458 limitations. First, the waitee must be defined prior to the waiter job
2459 (meaning no forward references). Second, if a job is being referenced as a
2460 waitee, it must have a unique name (no duplicate waitees).
2462 .. option:: nice=int
2464 Run the job with the given nice value. See man :manpage:`nice(2)`.
2466 On Windows, values less than -15 set the process class to "High"; -1 through
2467 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
2470 .. option:: prio=int
2472 Set the I/O priority value of this job. Linux limits us to a positive value
2473 between 0 and 7, with 0 being the highest. See man
2474 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
2475 systems since meaning of priority may differ.
2477 .. option:: prioclass=int
2479 Set the I/O priority class. See man :manpage:`ionice(1)`.
2481 .. option:: cpus_allowed=str
2483 Controls the same options as :option:`cpumask`, but accepts a textual
2484 specification of the permitted CPUs instead and CPUs are indexed from 0. So
2485 to use CPUs 0 and 5 you would specify ``cpus_allowed=0,5``. This option also
2486 allows a range of CPUs to be specified -- say you wanted a binding to CPUs
2487 0, 5, and 8 to 15, you would set ``cpus_allowed=0,5,8-15``.
2489 On Windows, when ``cpus_allowed`` is unset only CPUs from fio's current
2490 processor group will be used and affinity settings are inherited from the
2491 system. An fio build configured to target Windows 7 makes options that set
2492 CPUs processor group aware and values will set both the processor group
2493 and a CPU from within that group. For example, on a system where processor
2494 group 0 has 40 CPUs and processor group 1 has 32 CPUs, ``cpus_allowed``
2495 values between 0 and 39 will bind CPUs from processor group 0 and
2496 ``cpus_allowed`` values between 40 and 71 will bind CPUs from processor
2497 group 1. When using ``cpus_allowed_policy=shared`` all CPUs specified by a
2498 single ``cpus_allowed`` option must be from the same processor group. For
2499 Windows fio builds not built for Windows 7, CPUs will only be selected from
2500 (and be relative to) whatever processor group fio happens to be running in
2501 and CPUs from other processor groups cannot be used.
2503 .. option:: cpus_allowed_policy=str
2505 Set the policy of how fio distributes the CPUs specified by
2506 :option:`cpus_allowed` or :option:`cpumask`. Two policies are supported:
2509 All jobs will share the CPU set specified.
2511 Each job will get a unique CPU from the CPU set.
2513 **shared** is the default behavior, if the option isn't specified. If
2514 **split** is specified, then fio will will assign one cpu per job. If not
2515 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
2518 .. option:: cpumask=int
2520 Set the CPU affinity of this job. The parameter given is a bit mask of
2521 allowed CPUs the job may run on. So if you want the allowed CPUs to be 1
2522 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
2523 :manpage:`sched_setaffinity(2)`. This may not work on all supported
2524 operating systems or kernel versions. This option doesn't work well for a
2525 higher CPU count than what you can store in an integer mask, so it can only
2526 control cpus 1-32. For boxes with larger CPU counts, use
2527 :option:`cpus_allowed`.
2529 .. option:: numa_cpu_nodes=str
2531 Set this job running on specified NUMA nodes' CPUs. The arguments allow
2532 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
2533 NUMA options support, fio must be built on a system with libnuma-dev(el)
2536 .. option:: numa_mem_policy=str
2538 Set this job's memory policy and corresponding NUMA nodes. Format of the
2543 ``mode`` is one of the following memory policies: ``default``, ``prefer``,
2544 ``bind``, ``interleave`` or ``local``. For ``default`` and ``local`` memory
2545 policies, no node needs to be specified. For ``prefer``, only one node is
2546 allowed. For ``bind`` and ``interleave`` the ``nodelist`` may be as
2547 follows: a comma delimited list of numbers, A-B ranges, or `all`.
2549 .. option:: cgroup=str
2551 Add job to this control group. If it doesn't exist, it will be created. The
2552 system must have a mounted cgroup blkio mount point for this to work. If
2553 your system doesn't have it mounted, you can do so with::
2555 # mount -t cgroup -o blkio none /cgroup
2557 .. option:: cgroup_weight=int
2559 Set the weight of the cgroup to this value. See the documentation that comes
2560 with the kernel, allowed values are in the range of 100..1000.
2562 .. option:: cgroup_nodelete=bool
2564 Normally fio will delete the cgroups it has created after the job
2565 completion. To override this behavior and to leave cgroups around after the
2566 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
2567 to inspect various cgroup files after job completion. Default: false.
2569 .. option:: flow_id=int
2571 The ID of the flow. If not specified, it defaults to being a global
2572 flow. See :option:`flow`.
2574 .. option:: flow=int
2576 Weight in token-based flow control. If this value is used, then there is a
2577 'flow counter' which is used to regulate the proportion of activity between
2578 two or more jobs. Fio attempts to keep this flow counter near zero. The
2579 ``flow`` parameter stands for how much should be added or subtracted to the
2580 flow counter on each iteration of the main I/O loop. That is, if one job has
2581 ``flow=8`` and another job has ``flow=-1``, then there will be a roughly 1:8
2582 ratio in how much one runs vs the other.
2584 .. option:: flow_watermark=int
2586 The maximum value that the absolute value of the flow counter is allowed to
2587 reach before the job must wait for a lower value of the counter.
2589 .. option:: flow_sleep=int
2591 The period of time, in microseconds, to wait after the flow watermark has
2592 been exceeded before retrying operations.
2594 .. option:: stonewall, wait_for_previous
2596 Wait for preceding jobs in the job file to exit, before starting this
2597 one. Can be used to insert serialization points in the job file. A stone
2598 wall also implies starting a new reporting group, see
2599 :option:`group_reporting`.
2603 By default, fio will continue running all other jobs when one job finishes
2604 but sometimes this is not the desired action. Setting ``exitall`` will
2605 instead make fio terminate all other jobs when one job finishes.
2607 .. option:: exec_prerun=str
2609 Before running this job, issue the command specified through
2610 :manpage:`system(3)`. Output is redirected in a file called
2611 :file:`jobname.prerun.txt`.
2613 .. option:: exec_postrun=str
2615 After the job completes, issue the command specified though
2616 :manpage:`system(3)`. Output is redirected in a file called
2617 :file:`jobname.postrun.txt`.
2621 Instead of running as the invoking user, set the user ID to this value
2622 before the thread/process does any work.
2626 Set group ID, see :option:`uid`.
2632 .. option:: verify_only
2634 Do not perform specified workload, only verify data still matches previous
2635 invocation of this workload. This option allows one to check data multiple
2636 times at a later date without overwriting it. This option makes sense only
2637 for workloads that write data, and does not support workloads with the
2638 :option:`time_based` option set.
2640 .. option:: do_verify=bool
2642 Run the verify phase after a write phase. Only valid if :option:`verify` is
2645 .. option:: verify=str
2647 If writing to a file, fio can verify the file contents after each iteration
2648 of the job. Each verification method also implies verification of special
2649 header, which is written to the beginning of each block. This header also
2650 includes meta information, like offset of the block, block number, timestamp
2651 when block was written, etc. :option:`verify` can be combined with
2652 :option:`verify_pattern` option. The allowed values are:
2655 Use an md5 sum of the data area and store it in the header of
2659 Use an experimental crc64 sum of the data area and store it in the
2660 header of each block.
2663 Use a crc32c sum of the data area and store it in the header of
2664 each block. This will automatically use hardware acceleration
2665 (e.g. SSE4.2 on an x86 or CRC crypto extensions on ARM64) but will
2666 fall back to software crc32c if none is found. Generally the
2667 fastest checksum fio supports when hardware accelerated.
2673 Use a crc32 sum of the data area and store it in the header of each
2677 Use a crc16 sum of the data area and store it in the header of each
2681 Use a crc7 sum of the data area and store it in the header of each
2685 Use xxhash as the checksum function. Generally the fastest software
2686 checksum that fio supports.
2689 Use sha512 as the checksum function.
2692 Use sha256 as the checksum function.
2695 Use optimized sha1 as the checksum function.
2698 Use optimized sha3-224 as the checksum function.
2701 Use optimized sha3-256 as the checksum function.
2704 Use optimized sha3-384 as the checksum function.
2707 Use optimized sha3-512 as the checksum function.
2710 This option is deprecated, since now meta information is included in
2711 generic verification header and meta verification happens by
2712 default. For detailed information see the description of the
2713 :option:`verify` setting. This option is kept because of
2714 compatibility's sake with old configurations. Do not use it.
2717 Verify a strict pattern. Normally fio includes a header with some
2718 basic information and checksumming, but if this option is set, only
2719 the specific pattern set with :option:`verify_pattern` is verified.
2722 Only pretend to verify. Useful for testing internals with
2723 :option:`ioengine`\=null, not for much else.
2725 This option can be used for repeated burn-in tests of a system to make sure
2726 that the written data is also correctly read back. If the data direction
2727 given is a read or random read, fio will assume that it should verify a
2728 previously written file. If the data direction includes any form of write,
2729 the verify will be of the newly written data.
2731 To avoid false verification errors, do not use the norandommap option when
2732 verifying data with async I/O engines and I/O depths > 1. Or use the
2733 norandommap and the lfsr random generator together to avoid writing to the
2734 same offset with muliple outstanding I/Os.
2736 .. option:: verify_offset=int
2738 Swap the verification header with data somewhere else in the block before
2739 writing. It is swapped back before verifying.
2741 .. option:: verify_interval=int
2743 Write the verification header at a finer granularity than the
2744 :option:`blocksize`. It will be written for chunks the size of
2745 ``verify_interval``. :option:`blocksize` should divide this evenly.
2747 .. option:: verify_pattern=str
2749 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
2750 filling with totally random bytes, but sometimes it's interesting to fill
2751 with a known pattern for I/O verification purposes. Depending on the width
2752 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time (it can
2753 be either a decimal or a hex number). The ``verify_pattern`` if larger than
2754 a 32-bit quantity has to be a hex number that starts with either "0x" or
2755 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
2756 format, which means that for each block offset will be written and then
2757 verified back, e.g.::
2761 Or use combination of everything::
2763 verify_pattern=0xff%o"abcd"-12
2765 .. option:: verify_fatal=bool
2767 Normally fio will keep checking the entire contents before quitting on a
2768 block verification failure. If this option is set, fio will exit the job on
2769 the first observed failure. Default: false.
2771 .. option:: verify_dump=bool
2773 If set, dump the contents of both the original data block and the data block
2774 we read off disk to files. This allows later analysis to inspect just what
2775 kind of data corruption occurred. Off by default.
2777 .. option:: verify_async=int
2779 Fio will normally verify I/O inline from the submitting thread. This option
2780 takes an integer describing how many async offload threads to create for I/O
2781 verification instead, causing fio to offload the duty of verifying I/O
2782 contents to one or more separate threads. If using this offload option, even
2783 sync I/O engines can benefit from using an :option:`iodepth` setting higher
2784 than 1, as it allows them to have I/O in flight while verifies are running.
2785 Defaults to 0 async threads, i.e. verification is not asynchronous.
2787 .. option:: verify_async_cpus=str
2789 Tell fio to set the given CPU affinity on the async I/O verification
2790 threads. See :option:`cpus_allowed` for the format used.
2792 .. option:: verify_backlog=int
2794 Fio will normally verify the written contents of a job that utilizes verify
2795 once that job has completed. In other words, everything is written then
2796 everything is read back and verified. You may want to verify continually
2797 instead for a variety of reasons. Fio stores the meta data associated with
2798 an I/O block in memory, so for large verify workloads, quite a bit of memory
2799 would be used up holding this meta data. If this option is enabled, fio will
2800 write only N blocks before verifying these blocks.
2802 .. option:: verify_backlog_batch=int
2804 Control how many blocks fio will verify if :option:`verify_backlog` is
2805 set. If not set, will default to the value of :option:`verify_backlog`
2806 (meaning the entire queue is read back and verified). If
2807 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
2808 blocks will be verified, if ``verify_backlog_batch`` is larger than
2809 :option:`verify_backlog`, some blocks will be verified more than once.
2811 .. option:: verify_state_save=bool
2813 When a job exits during the write phase of a verify workload, save its
2814 current state. This allows fio to replay up until that point, if the verify
2815 state is loaded for the verify read phase. The format of the filename is,
2818 <type>-<jobname>-<jobindex>-verify.state.
2820 <type> is "local" for a local run, "sock" for a client/server socket
2821 connection, and "ip" (192.168.0.1, for instance) for a networked
2822 client/server connection. Defaults to true.
2824 .. option:: verify_state_load=bool
2826 If a verify termination trigger was used, fio stores the current write state
2827 of each thread. This can be used at verification time so that fio knows how
2828 far it should verify. Without this information, fio will run a full
2829 verification pass, according to the settings in the job file used. Default
2832 .. option:: trim_percentage=int
2834 Number of verify blocks to discard/trim.
2836 .. option:: trim_verify_zero=bool
2838 Verify that trim/discarded blocks are returned as zeros.
2840 .. option:: trim_backlog=int
2842 Trim after this number of blocks are written.
2844 .. option:: trim_backlog_batch=int
2846 Trim this number of I/O blocks.
2848 .. option:: experimental_verify=bool
2850 Enable experimental verification.
2855 .. option:: steadystate=str:float, ss=str:float
2857 Define the criterion and limit for assessing steady state performance. The
2858 first parameter designates the criterion whereas the second parameter sets
2859 the threshold. When the criterion falls below the threshold for the
2860 specified duration, the job will stop. For example, `iops_slope:0.1%` will
2861 direct fio to terminate the job when the least squares regression slope
2862 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
2863 this will apply to all jobs in the group. Below is the list of available
2864 steady state assessment criteria. All assessments are carried out using only
2865 data from the rolling collection window. Threshold limits can be expressed
2866 as a fixed value or as a percentage of the mean in the collection window.
2869 Collect IOPS data. Stop the job if all individual IOPS measurements
2870 are within the specified limit of the mean IOPS (e.g., ``iops:2``
2871 means that all individual IOPS values must be within 2 of the mean,
2872 whereas ``iops:0.2%`` means that all individual IOPS values must be
2873 within 0.2% of the mean IOPS to terminate the job).
2876 Collect IOPS data and calculate the least squares regression
2877 slope. Stop the job if the slope falls below the specified limit.
2880 Collect bandwidth data. Stop the job if all individual bandwidth
2881 measurements are within the specified limit of the mean bandwidth.
2884 Collect bandwidth data and calculate the least squares regression
2885 slope. Stop the job if the slope falls below the specified limit.
2887 .. option:: steadystate_duration=time, ss_dur=time
2889 A rolling window of this duration will be used to judge whether steady state
2890 has been reached. Data will be collected once per second. The default is 0
2891 which disables steady state detection. When the unit is omitted, the
2892 value is interpreted in seconds.
2894 .. option:: steadystate_ramp_time=time, ss_ramp=time
2896 Allow the job to run for the specified duration before beginning data
2897 collection for checking the steady state job termination criterion. The
2898 default is 0. When the unit is omitted, the value is interpreted in seconds.
2901 Measurements and reporting
2902 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2904 .. option:: per_job_logs=bool
2906 If set, this generates bw/clat/iops log with per file private filenames. If
2907 not set, jobs with identical names will share the log filename. Default:
2910 .. option:: group_reporting
2912 It may sometimes be interesting to display statistics for groups of jobs as
2913 a whole instead of for each individual job. This is especially true if
2914 :option:`numjobs` is used; looking at individual thread/process output
2915 quickly becomes unwieldy. To see the final report per-group instead of
2916 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
2917 same reporting group, unless if separated by a :option:`stonewall`, or by
2918 using :option:`new_group`.
2920 .. option:: new_group
2922 Start a new reporting group. See: :option:`group_reporting`. If not given,
2923 all jobs in a file will be part of the same reporting group, unless
2924 separated by a :option:`stonewall`.
2926 .. option:: stats=bool
2928 By default, fio collects and shows final output results for all jobs
2929 that run. If this option is set to 0, then fio will ignore it in
2930 the final stat output.
2932 .. option:: write_bw_log=str
2934 If given, write a bandwidth log for this job. Can be used to store data of
2935 the bandwidth of the jobs in their lifetime.
2937 If no str argument is given, the default filename of
2938 :file:`jobname_type.x.log` is used. Even when the argument is given, fio
2939 will still append the type of log. So if one specifies::
2943 The actual log name will be :file:`foo_bw.x.log` where `x` is the index
2944 of the job (`1..N`, where `N` is the number of jobs). If
2945 :option:`per_job_logs` is false, then the filename will not include the
2948 The included :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
2949 text files into nice graphs. See `Log File Formats`_ for how data is
2950 structured within the file.
2952 .. option:: write_lat_log=str
2954 Same as :option:`write_bw_log`, except this option creates I/O
2955 submission (e.g., :file:`name_slat.x.log`), completion (e.g.,
2956 :file:`name_clat.x.log`), and total (e.g., :file:`name_lat.x.log`)
2957 latency files instead. See :option:`write_bw_log` for details about
2958 the filename format and `Log File Formats`_ for how data is structured
2961 .. option:: write_hist_log=str
2963 Same as :option:`write_bw_log` but writes an I/O completion latency
2964 histogram file (e.g., :file:`name_hist.x.log`) instead. Note that this
2965 file will be empty unless :option:`log_hist_msec` has also been set.
2966 See :option:`write_bw_log` for details about the filename format and
2967 `Log File Formats`_ for how data is structured within the file.
2969 .. option:: write_iops_log=str
2971 Same as :option:`write_bw_log`, but writes an IOPS file (e.g.
2972 :file:`name_iops.x.log`) instead. Because fio defaults to individual
2973 I/O logging, the value entry in the IOPS log will be 1 unless windowed
2974 logging (see :option:`log_avg_msec`) has been enabled. See
2975 :option:`write_bw_log` for details about the filename format and `Log
2976 File Formats`_ for how data is structured within the file.
2978 .. option:: log_avg_msec=int
2980 By default, fio will log an entry in the iops, latency, or bw log for every
2981 I/O that completes. When writing to the disk log, that can quickly grow to a
2982 very large size. Setting this option makes fio average the each log entry
2983 over the specified period of time, reducing the resolution of the log. See
2984 :option:`log_max_value` as well. Defaults to 0, logging all entries.
2985 Also see `Log File Formats`_.
2987 .. option:: log_hist_msec=int
2989 Same as :option:`log_avg_msec`, but logs entries for completion latency
2990 histograms. Computing latency percentiles from averages of intervals using
2991 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
2992 histogram entries over the specified period of time, reducing log sizes for
2993 high IOPS devices while retaining percentile accuracy. See
2994 :option:`log_hist_coarseness` and :option:`write_hist_log` as well.
2995 Defaults to 0, meaning histogram logging is disabled.
2997 .. option:: log_hist_coarseness=int
2999 Integer ranging from 0 to 6, defining the coarseness of the resolution of
3000 the histogram logs enabled with :option:`log_hist_msec`. For each increment
3001 in coarseness, fio outputs half as many bins. Defaults to 0, for which
3002 histogram logs contain 1216 latency bins. See :option:`write_hist_log`
3003 and `Log File Formats`_.
3005 .. option:: log_max_value=bool
3007 If :option:`log_avg_msec` is set, fio logs the average over that window. If
3008 you instead want to log the maximum value, set this option to 1. Defaults to
3009 0, meaning that averaged values are logged.
3011 .. option:: log_offset=bool
3013 If this is set, the iolog options will include the byte offset for the I/O
3014 entry as well as the other data values. Defaults to 0 meaning that
3015 offsets are not present in logs. Also see `Log File Formats`_.
3017 .. option:: log_compression=int
3019 If this is set, fio will compress the I/O logs as it goes, to keep the
3020 memory footprint lower. When a log reaches the specified size, that chunk is
3021 removed and compressed in the background. Given that I/O logs are fairly
3022 highly compressible, this yields a nice memory savings for longer runs. The
3023 downside is that the compression will consume some background CPU cycles, so
3024 it may impact the run. This, however, is also true if the logging ends up
3025 consuming most of the system memory. So pick your poison. The I/O logs are
3026 saved normally at the end of a run, by decompressing the chunks and storing
3027 them in the specified log file. This feature depends on the availability of
3030 .. option:: log_compression_cpus=str
3032 Define the set of CPUs that are allowed to handle online log compression for
3033 the I/O jobs. This can provide better isolation between performance
3034 sensitive jobs, and background compression work. See
3035 :option:`cpus_allowed` for the format used.
3037 .. option:: log_store_compressed=bool
3039 If set, fio will store the log files in a compressed format. They can be
3040 decompressed with fio, using the :option:`--inflate-log` command line
3041 parameter. The files will be stored with a :file:`.fz` suffix.
3043 .. option:: log_unix_epoch=bool
3045 If set, fio will log Unix timestamps to the log files produced by enabling
3046 write_type_log for each log type, instead of the default zero-based
3049 .. option:: block_error_percentiles=bool
3051 If set, record errors in trim block-sized units from writes and trims and
3052 output a histogram of how many trims it took to get to errors, and what kind
3053 of error was encountered.
3055 .. option:: bwavgtime=int
3057 Average the calculated bandwidth over the given time. Value is specified in
3058 milliseconds. If the job also does bandwidth logging through
3059 :option:`write_bw_log`, then the minimum of this option and
3060 :option:`log_avg_msec` will be used. Default: 500ms.
3062 .. option:: iopsavgtime=int
3064 Average the calculated IOPS over the given time. Value is specified in
3065 milliseconds. If the job also does IOPS logging through
3066 :option:`write_iops_log`, then the minimum of this option and
3067 :option:`log_avg_msec` will be used. Default: 500ms.
3069 .. option:: disk_util=bool
3071 Generate disk utilization statistics, if the platform supports it.
3074 .. option:: disable_lat=bool
3076 Disable measurements of total latency numbers. Useful only for cutting back
3077 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
3078 performance at really high IOPS rates. Note that to really get rid of a
3079 large amount of these calls, this option must be used with
3080 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
3082 .. option:: disable_clat=bool
3084 Disable measurements of completion latency numbers. See
3085 :option:`disable_lat`.
3087 .. option:: disable_slat=bool
3089 Disable measurements of submission latency numbers. See
3090 :option:`disable_lat`.
3092 .. option:: disable_bw_measurement=bool, disable_bw=bool
3094 Disable measurements of throughput/bandwidth numbers. See
3095 :option:`disable_lat`.
3097 .. option:: clat_percentiles=bool
3099 Enable the reporting of percentiles of completion latencies. This
3100 option is mutually exclusive with :option:`lat_percentiles`.
3102 .. option:: lat_percentiles=bool
3104 Enable the reporting of percentiles of I/O latencies. This is similar
3105 to :option:`clat_percentiles`, except that this includes the
3106 submission latency. This option is mutually exclusive with
3107 :option:`clat_percentiles`.
3109 .. option:: percentile_list=float_list
3111 Overwrite the default list of percentiles for completion latencies and
3112 the block error histogram. Each number is a floating number in the
3113 range (0,100], and the maximum length of the list is 20. Use ``:`` to
3114 separate the numbers, and list the numbers in ascending order. For
3115 example, ``--percentile_list=99.5:99.9`` will cause fio to report the
3116 values of completion latency below which 99.5% and 99.9% of the observed
3117 latencies fell, respectively.
3119 .. option:: significant_figures=int
3121 If using :option:`--output-format` of `normal`, set the significant
3122 figures to this value. Higher values will yield more precise IOPS and
3123 throughput units, while lower values will round. Requires a minimum
3124 value of 1 and a maximum value of 10. Defaults to 4.
3130 .. option:: exitall_on_error
3132 When one job finishes in error, terminate the rest. The default is to wait
3133 for each job to finish.
3135 .. option:: continue_on_error=str
3137 Normally fio will exit the job on the first observed failure. If this option
3138 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
3139 EILSEQ) until the runtime is exceeded or the I/O size specified is
3140 completed. If this option is used, there are two more stats that are
3141 appended, the total error count and the first error. The error field given
3142 in the stats is the first error that was hit during the run.
3144 The allowed values are:
3147 Exit on any I/O or verify errors.
3150 Continue on read errors, exit on all others.
3153 Continue on write errors, exit on all others.
3156 Continue on any I/O error, exit on all others.
3159 Continue on verify errors, exit on all others.
3162 Continue on all errors.
3165 Backward-compatible alias for 'none'.
3168 Backward-compatible alias for 'all'.
3170 .. option:: ignore_error=str
3172 Sometimes you want to ignore some errors during test in that case you can
3173 specify error list for each error type, instead of only being able to
3174 ignore the default 'non-fatal error' using :option:`continue_on_error`.
3175 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
3176 given error type is separated with ':'. Error may be symbol ('ENOSPC',
3177 'ENOMEM') or integer. Example::
3179 ignore_error=EAGAIN,ENOSPC:122
3181 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
3182 WRITE. This option works by overriding :option:`continue_on_error` with
3183 the list of errors for each error type if any.
3185 .. option:: error_dump=bool
3187 If set dump every error even if it is non fatal, true by default. If
3188 disabled only fatal error will be dumped.
3190 Running predefined workloads
3191 ----------------------------
3193 Fio includes predefined profiles that mimic the I/O workloads generated by
3196 .. option:: profile=str
3198 The predefined workload to run. Current profiles are:
3201 Threaded I/O bench (tiotest/tiobench) like workload.
3204 Aerospike Certification Tool (ACT) like workload.
3206 To view a profile's additional options use :option:`--cmdhelp` after specifying
3207 the profile. For example::
3209 $ fio --profile=act --cmdhelp
3214 .. option:: device-names=str
3219 .. option:: load=int
3222 ACT load multiplier. Default: 1.
3224 .. option:: test-duration=time
3227 How long the entire test takes to run. When the unit is omitted, the value
3228 is given in seconds. Default: 24h.
3230 .. option:: threads-per-queue=int
3233 Number of read I/O threads per device. Default: 8.
3235 .. option:: read-req-num-512-blocks=int
3238 Number of 512B blocks to read at the time. Default: 3.
3240 .. option:: large-block-op-kbytes=int
3243 Size of large block ops in KiB (writes). Default: 131072.
3248 Set to run ACT prep phase.
3250 Tiobench profile options
3251 ~~~~~~~~~~~~~~~~~~~~~~~~
3253 .. option:: size=str
3258 .. option:: block=int
3261 Block size in bytes. Default: 4096.
3263 .. option:: numruns=int
3273 .. option:: threads=int
3278 Interpreting the output
3279 -----------------------
3282 Example output was based on the following:
3283 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --time_based \
3284 --rate=1256k --bs=14K --name=quick --runtime=1s --name=mixed \
3285 --runtime=2m --rw=rw
3287 Fio spits out a lot of output. While running, fio will display the status of the
3288 jobs created. An example of that would be::
3290 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]
3292 The characters inside the first set of square brackets denote the current status of
3293 each thread. The first character is the first job defined in the job file, and so
3294 forth. The possible values (in typical life cycle order) are:
3296 +------+-----+-----------------------------------------------------------+
3298 +======+=====+===========================================================+
3299 | P | | Thread setup, but not started. |
3300 +------+-----+-----------------------------------------------------------+
3301 | C | | Thread created. |
3302 +------+-----+-----------------------------------------------------------+
3303 | I | | Thread initialized, waiting or generating necessary data. |
3304 +------+-----+-----------------------------------------------------------+
3305 | | p | Thread running pre-reading file(s). |
3306 +------+-----+-----------------------------------------------------------+
3307 | | / | Thread is in ramp period. |
3308 +------+-----+-----------------------------------------------------------+
3309 | | R | Running, doing sequential reads. |
3310 +------+-----+-----------------------------------------------------------+
3311 | | r | Running, doing random reads. |
3312 +------+-----+-----------------------------------------------------------+
3313 | | W | Running, doing sequential writes. |
3314 +------+-----+-----------------------------------------------------------+
3315 | | w | Running, doing random writes. |
3316 +------+-----+-----------------------------------------------------------+
3317 | | M | Running, doing mixed sequential reads/writes. |
3318 +------+-----+-----------------------------------------------------------+
3319 | | m | Running, doing mixed random reads/writes. |
3320 +------+-----+-----------------------------------------------------------+
3321 | | D | Running, doing sequential trims. |
3322 +------+-----+-----------------------------------------------------------+
3323 | | d | Running, doing random trims. |
3324 +------+-----+-----------------------------------------------------------+
3325 | | F | Running, currently waiting for :manpage:`fsync(2)`. |
3326 +------+-----+-----------------------------------------------------------+
3327 | | V | Running, doing verification of written data. |
3328 +------+-----+-----------------------------------------------------------+
3329 | f | | Thread finishing. |
3330 +------+-----+-----------------------------------------------------------+
3331 | E | | Thread exited, not reaped by main thread yet. |
3332 +------+-----+-----------------------------------------------------------+
3333 | _ | | Thread reaped. |
3334 +------+-----+-----------------------------------------------------------+
3335 | X | | Thread reaped, exited with an error. |
3336 +------+-----+-----------------------------------------------------------+
3337 | K | | Thread reaped, exited due to signal. |
3338 +------+-----+-----------------------------------------------------------+
3341 Example output was based on the following:
3342 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --runtime=58m \
3343 --time_based --rate=2512k --bs=256K --numjobs=10 \
3344 --name=readers --rw=read --name=writers --rw=write
3346 Fio will condense the thread string as not to take up more space on the command
3347 line than needed. For instance, if you have 10 readers and 10 writers running,
3348 the output would look like this::
3350 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]
3352 Note that the status string is displayed in order, so it's possible to tell which of
3353 the jobs are currently doing what. In the example above this means that jobs 1--10
3354 are readers and 11--20 are writers.
3356 The other values are fairly self explanatory -- number of threads currently
3357 running and doing I/O, the number of currently open files (f=), the estimated
3358 completion percentage, the rate of I/O since last check (read speed listed first,
3359 then write speed and optionally trim speed) in terms of bandwidth and IOPS,
3360 and time to completion for the current running group. It's impossible to estimate
3361 runtime of the following groups (if any).
3364 Example output was based on the following:
3365 TZ=UTC fio --iodepth=16 --ioengine=posixaio --filename=/tmp/fiofile \
3366 --direct=1 --size=100M --time_based --runtime=50s --rate_iops=89 \
3367 --bs=7K --name=Client1 --rw=write
3369 When fio is done (or interrupted by :kbd:`Ctrl-C`), it will show the data for
3370 each thread, group of threads, and disks in that order. For each overall thread (or
3371 group) the output looks like::
3373 Client1: (groupid=0, jobs=1): err= 0: pid=16109: Sat Jun 24 12:07:54 2017
3374 write: IOPS=88, BW=623KiB/s (638kB/s)(30.4MiB/50032msec)
3375 slat (nsec): min=500, max=145500, avg=8318.00, stdev=4781.50
3376 clat (usec): min=170, max=78367, avg=4019.02, stdev=8293.31
3377 lat (usec): min=174, max=78375, avg=4027.34, stdev=8291.79
3378 clat percentiles (usec):
3379 | 1.00th=[ 302], 5.00th=[ 326], 10.00th=[ 343], 20.00th=[ 363],
3380 | 30.00th=[ 392], 40.00th=[ 404], 50.00th=[ 416], 60.00th=[ 445],
3381 | 70.00th=[ 816], 80.00th=[ 6718], 90.00th=[12911], 95.00th=[21627],
3382 | 99.00th=[43779], 99.50th=[51643], 99.90th=[68682], 99.95th=[72877],
3384 bw ( KiB/s): min= 532, max= 686, per=0.10%, avg=622.87, stdev=24.82, samples= 100
3385 iops : min= 76, max= 98, avg=88.98, stdev= 3.54, samples= 100
3386 lat (usec) : 250=0.04%, 500=64.11%, 750=4.81%, 1000=2.79%
3387 lat (msec) : 2=4.16%, 4=1.84%, 10=4.90%, 20=11.33%, 50=5.37%
3388 lat (msec) : 100=0.65%
3389 cpu : usr=0.27%, sys=0.18%, ctx=12072, majf=0, minf=21
3390 IO depths : 1=85.0%, 2=13.1%, 4=1.8%, 8=0.1%, 16=0.0%, 32=0.0%, >=64=0.0%
3391 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3392 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3393 issued rwt: total=0,4450,0, short=0,0,0, dropped=0,0,0
3394 latency : target=0, window=0, percentile=100.00%, depth=8
3396 The job name (or first job's name when using :option:`group_reporting`) is printed,
3397 along with the group id, count of jobs being aggregated, last error id seen (which
3398 is 0 when there are no errors), pid/tid of that thread and the time the job/group
3399 completed. Below are the I/O statistics for each data direction performed (showing
3400 writes in the example above). In the order listed, they denote:
3403 The string before the colon shows the I/O direction the statistics
3404 are for. **IOPS** is the average I/Os performed per second. **BW**
3405 is the average bandwidth rate shown as: value in power of 2 format
3406 (value in power of 10 format). The last two values show: (**total
3407 I/O performed** in power of 2 format / **runtime** of that thread).
3410 Submission latency (**min** being the minimum, **max** being the
3411 maximum, **avg** being the average, **stdev** being the standard
3412 deviation). This is the time it took to submit the I/O. For
3413 sync I/O this row is not displayed as the slat is really the
3414 completion latency (since queue/complete is one operation there).
3415 This value can be in nanoseconds, microseconds or milliseconds ---
3416 fio will choose the most appropriate base and print that (in the
3417 example above nanoseconds was the best scale). Note: in :option:`--minimal` mode
3418 latencies are always expressed in microseconds.
3421 Completion latency. Same names as slat, this denotes the time from
3422 submission to completion of the I/O pieces. For sync I/O, clat will
3423 usually be equal (or very close) to 0, as the time from submit to
3424 complete is basically just CPU time (I/O has already been done, see slat
3428 Total latency. Same names as slat and clat, this denotes the time from
3429 when fio created the I/O unit to completion of the I/O operation.
3432 Bandwidth statistics based on samples. Same names as the xlat stats,
3433 but also includes the number of samples taken (**samples**) and an
3434 approximate percentage of total aggregate bandwidth this thread
3435 received in its group (**per**). This last value is only really
3436 useful if the threads in this group are on the same disk, since they
3437 are then competing for disk access.
3440 IOPS statistics based on samples. Same names as bw.
3442 **lat (nsec/usec/msec)**
3443 The distribution of I/O completion latencies. This is the time from when
3444 I/O leaves fio and when it gets completed. Unlike the separate
3445 read/write/trim sections above, the data here and in the remaining
3446 sections apply to all I/Os for the reporting group. 250=0.04% means that
3447 0.04% of the I/Os completed in under 250us. 500=64.11% means that 64.11%
3448 of the I/Os required 250 to 499us for completion.
3451 CPU usage. User and system time, along with the number of context
3452 switches this thread went through, usage of system and user time, and
3453 finally the number of major and minor page faults. The CPU utilization
3454 numbers are averages for the jobs in that reporting group, while the
3455 context and fault counters are summed.
3458 The distribution of I/O depths over the job lifetime. The numbers are
3459 divided into powers of 2 and each entry covers depths from that value
3460 up to those that are lower than the next entry -- e.g., 16= covers
3461 depths from 16 to 31. Note that the range covered by a depth
3462 distribution entry can be different to the range covered by the
3463 equivalent submit/complete distribution entry.
3466 How many pieces of I/O were submitting in a single submit call. Each
3467 entry denotes that amount and below, until the previous entry -- e.g.,
3468 16=100% means that we submitted anywhere between 9 to 16 I/Os per submit
3469 call. Note that the range covered by a submit distribution entry can
3470 be different to the range covered by the equivalent depth distribution
3474 Like the above submit number, but for completions instead.
3477 The number of read/write/trim requests issued, and how many of them were
3481 These values are for :option:`latency_target` and related options. When
3482 these options are engaged, this section describes the I/O depth required
3483 to meet the specified latency target.
3486 Example output was based on the following:
3487 TZ=UTC fio --ioengine=null --iodepth=2 --size=100M --numjobs=2 \
3488 --rate_process=poisson --io_limit=32M --name=read --bs=128k \
3489 --rate=11M --name=write --rw=write --bs=2k --rate=700k
3491 After each client has been listed, the group statistics are printed. They
3492 will look like this::
3494 Run status group 0 (all jobs):
3495 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
3496 WRITE: bw=1231KiB/s (1261kB/s), 616KiB/s-621KiB/s (630kB/s-636kB/s), io=64.0MiB (67.1MB), run=52747-53223msec
3498 For each data direction it prints:
3501 Aggregate bandwidth of threads in this group followed by the
3502 minimum and maximum bandwidth of all the threads in this group.
3503 Values outside of brackets are power-of-2 format and those
3504 within are the equivalent value in a power-of-10 format.
3506 Aggregate I/O performed of all threads in this group. The
3507 format is the same as bw.
3509 The smallest and longest runtimes of the threads in this group.
3511 And finally, the disk statistics are printed. This is Linux specific. They will look like this::
3513 Disk stats (read/write):
3514 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
3516 Each value is printed for both reads and writes, with reads first. The
3520 Number of I/Os performed by all groups.
3522 Number of merges performed by the I/O scheduler.
3524 Number of ticks we kept the disk busy.
3526 Total time spent in the disk queue.
3528 The disk utilization. A value of 100% means we kept the disk
3529 busy constantly, 50% would be a disk idling half of the time.
3531 It is also possible to get fio to dump the current output while it is running,
3532 without terminating the job. To do that, send fio the **USR1** signal. You can
3533 also get regularly timed dumps by using the :option:`--status-interval`
3534 parameter, or by creating a file in :file:`/tmp` named
3535 :file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
3536 current output status.
3542 For scripted usage where you typically want to generate tables or graphs of the
3543 results, fio can output the results in a semicolon separated format. The format
3544 is one long line of values, such as::
3546 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%
3547 A description of this job goes here.
3549 The job description (if provided) follows on a second line.
3551 To enable terse output, use the :option:`--minimal` or
3552 :option:`--output-format`\=terse command line options. The
3553 first value is the version of the terse output format. If the output has to be
3554 changed for some reason, this number will be incremented by 1 to signify that
3557 Split up, the format is as follows (comments in brackets denote when a
3558 field was introduced or whether it's specific to some terse version):
3562 terse version, fio version [v3], jobname, groupid, error
3566 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3567 Submission latency: min, max, mean, stdev (usec)
3568 Completion latency: min, max, mean, stdev (usec)
3569 Completion latency percentiles: 20 fields (see below)
3570 Total latency: min, max, mean, stdev (usec)
3571 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3572 IOPS [v5]: min, max, mean, stdev, number of samples
3578 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3579 Submission latency: min, max, mean, stdev (usec)
3580 Completion latency: min, max, mean, stdev (usec)
3581 Completion latency percentiles: 20 fields (see below)
3582 Total latency: min, max, mean, stdev (usec)
3583 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3584 IOPS [v5]: min, max, mean, stdev, number of samples
3586 TRIM status [all but version 3]:
3588 Fields are similar to READ/WRITE status.
3592 user, system, context switches, major faults, minor faults
3596 <=1, 2, 4, 8, 16, 32, >=64
3598 I/O latencies microseconds::
3600 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
3602 I/O latencies milliseconds::
3604 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
3606 Disk utilization [v3]::
3608 disk name, read ios, write ios, read merges, write merges, read ticks, write ticks,
3609 time spent in queue, disk utilization percentage
3611 Additional Info (dependent on continue_on_error, default off)::
3613 total # errors, first error code
3615 Additional Info (dependent on description being set)::
3619 Completion latency percentiles can be a grouping of up to 20 sets, so for the
3620 terse output fio writes all of them. Each field will look like this::
3624 which is the Xth percentile, and the `usec` latency associated with it.
3626 For `Disk utilization`, all disks used by fio are shown. So for each disk there
3627 will be a disk utilization section.
3629 Below is a single line containing short names for each of the fields in the
3630 minimal output v3, separated by semicolons::
3632 terse_version_3;fio_version;jobname;groupid;error;read_kb;read_bandwidth;read_iops;read_runtime_ms;read_slat_min;read_slat_max;read_slat_mean;read_slat_dev;read_clat_min;read_clat_max;read_clat_mean;read_clat_dev;read_clat_pct01;read_clat_pct02;read_clat_pct03;read_clat_pct04;read_clat_pct05;read_clat_pct06;read_clat_pct07;read_clat_pct08;read_clat_pct09;read_clat_pct10;read_clat_pct11;read_clat_pct12;read_clat_pct13;read_clat_pct14;read_clat_pct15;read_clat_pct16;read_clat_pct17;read_clat_pct18;read_clat_pct19;read_clat_pct20;read_tlat_min;read_lat_max;read_lat_mean;read_lat_dev;read_bw_min;read_bw_max;read_bw_agg_pct;read_bw_mean;read_bw_dev;write_kb;write_bandwidth;write_iops;write_runtime_ms;write_slat_min;write_slat_max;write_slat_mean;write_slat_dev;write_clat_min;write_clat_max;write_clat_mean;write_clat_dev;write_clat_pct01;write_clat_pct02;write_clat_pct03;write_clat_pct04;write_clat_pct05;write_clat_pct06;write_clat_pct07;write_clat_pct08;write_clat_pct09;write_clat_pct10;write_clat_pct11;write_clat_pct12;write_clat_pct13;write_clat_pct14;write_clat_pct15;write_clat_pct16;write_clat_pct17;write_clat_pct18;write_clat_pct19;write_clat_pct20;write_tlat_min;write_lat_max;write_lat_mean;write_lat_dev;write_bw_min;write_bw_max;write_bw_agg_pct;write_bw_mean;write_bw_dev;cpu_user;cpu_sys;cpu_csw;cpu_mjf;cpu_minf;iodepth_1;iodepth_2;iodepth_4;iodepth_8;iodepth_16;iodepth_32;iodepth_64;lat_2us;lat_4us;lat_10us;lat_20us;lat_50us;lat_100us;lat_250us;lat_500us;lat_750us;lat_1000us;lat_2ms;lat_4ms;lat_10ms;lat_20ms;lat_50ms;lat_100ms;lat_250ms;lat_500ms;lat_750ms;lat_1000ms;lat_2000ms;lat_over_2000ms;disk_name;disk_read_iops;disk_write_iops;disk_read_merges;disk_write_merges;disk_read_ticks;write_ticks;disk_queue_time;disk_util
3638 The `json` output format is intended to be both human readable and convenient
3639 for automated parsing. For the most part its sections mirror those of the
3640 `normal` output. The `runtime` value is reported in msec and the `bw` value is
3641 reported in 1024 bytes per second units.
3647 The `json+` output format is identical to the `json` output format except that it
3648 adds a full dump of the completion latency bins. Each `bins` object contains a
3649 set of (key, value) pairs where keys are latency durations and values count how
3650 many I/Os had completion latencies of the corresponding duration. For example,
3653 "bins" : { "87552" : 1, "89600" : 1, "94720" : 1, "96768" : 1, "97792" : 1, "99840" : 1, "100864" : 2, "103936" : 6, "104960" : 534, "105984" : 5995, "107008" : 7529, ... }
3655 This data indicates that one I/O required 87,552ns to complete, two I/Os required
3656 100,864ns to complete, and 7529 I/Os required 107,008ns to complete.
3658 Also included with fio is a Python script `fio_jsonplus_clat2csv` that takes
3659 json+ output and generates CSV-formatted latency data suitable for plotting.
3661 The latency durations actually represent the midpoints of latency intervals.
3662 For details refer to :file:`stat.h`.
3668 There are two trace file format that you can encounter. The older (v1) format is
3669 unsupported since version 1.20-rc3 (March 2008). It will still be described
3670 below in case that you get an old trace and want to understand it.
3672 In any case the trace is a simple text file with a single action per line.
3675 Trace file format v1
3676 ~~~~~~~~~~~~~~~~~~~~
3678 Each line represents a single I/O action in the following format::
3682 where `rw=0/1` for read/write, and the `offset` and `length` entries being in bytes.
3684 This format is not supported in fio versions >= 1.20-rc3.
3687 Trace file format v2
3688 ~~~~~~~~~~~~~~~~~~~~
3690 The second version of the trace file format was added in fio version 1.17. It
3691 allows to access more then one file per trace and has a bigger set of possible
3694 The first line of the trace file has to be::
3698 Following this can be lines in two different formats, which are described below.
3700 The file management format::
3704 The `filename` is given as an absolute path. The `action` can be one of these:
3707 Add the given `filename` to the trace.
3709 Open the file with the given `filename`. The `filename` has to have
3710 been added with the **add** action before.
3712 Close the file with the given `filename`. The file has to have been
3716 The file I/O action format::
3718 filename action offset length
3720 The `filename` is given as an absolute path, and has to have been added and
3721 opened before it can be used with this format. The `offset` and `length` are
3722 given in bytes. The `action` can be one of these:
3725 Wait for `offset` microseconds. Everything below 100 is discarded.
3726 The time is relative to the previous `wait` statement.
3728 Read `length` bytes beginning from `offset`.
3730 Write `length` bytes beginning from `offset`.
3732 :manpage:`fsync(2)` the file.
3734 :manpage:`fdatasync(2)` the file.
3736 Trim the given file from the given `offset` for `length` bytes.
3738 CPU idleness profiling
3739 ----------------------
3741 In some cases, we want to understand CPU overhead in a test. For example, we
3742 test patches for the specific goodness of whether they reduce CPU usage.
3743 Fio implements a balloon approach to create a thread per CPU that runs at idle
3744 priority, meaning that it only runs when nobody else needs the cpu.
3745 By measuring the amount of work completed by the thread, idleness of each CPU
3746 can be derived accordingly.
3748 An unit work is defined as touching a full page of unsigned characters. Mean and
3749 standard deviation of time to complete an unit work is reported in "unit work"
3750 section. Options can be chosen to report detailed percpu idleness or overall
3751 system idleness by aggregating percpu stats.
3754 Verification and triggers
3755 -------------------------
3757 Fio is usually run in one of two ways, when data verification is done. The first
3758 is a normal write job of some sort with verify enabled. When the write phase has
3759 completed, fio switches to reads and verifies everything it wrote. The second
3760 model is running just the write phase, and then later on running the same job
3761 (but with reads instead of writes) to repeat the same I/O patterns and verify
3762 the contents. Both of these methods depend on the write phase being completed,
3763 as fio otherwise has no idea how much data was written.
3765 With verification triggers, fio supports dumping the current write state to
3766 local files. Then a subsequent read verify workload can load this state and know
3767 exactly where to stop. This is useful for testing cases where power is cut to a
3768 server in a managed fashion, for instance.
3770 A verification trigger consists of two things:
3772 1) Storing the write state of each job.
3773 2) Executing a trigger command.
3775 The write state is relatively small, on the order of hundreds of bytes to single
3776 kilobytes. It contains information on the number of completions done, the last X
3779 A trigger is invoked either through creation ('touch') of a specified file in
3780 the system, or through a timeout setting. If fio is run with
3781 :option:`--trigger-file`\= :file:`/tmp/trigger-file`, then it will continually
3782 check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
3783 will fire off the trigger (thus saving state, and executing the trigger
3786 For client/server runs, there's both a local and remote trigger. If fio is
3787 running as a server backend, it will send the job states back to the client for
3788 safe storage, then execute the remote trigger, if specified. If a local trigger
3789 is specified, the server will still send back the write state, but the client
3790 will then execute the trigger.
3792 Verification trigger example
3793 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3795 Let's say we want to run a powercut test on the remote Linux machine 'server'.
3796 Our write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
3797 some point during the run, and we'll run this test from the safety or our local
3798 machine, 'localbox'. On the server, we'll start the fio backend normally::
3800 server# fio --server
3802 and on the client, we'll fire off the workload::
3804 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
3806 We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
3808 echo b > /proc/sysrq-trigger
3810 on the server once it has received the trigger and sent us the write state. This
3811 will work, but it's not **really** cutting power to the server, it's merely
3812 abruptly rebooting it. If we have a remote way of cutting power to the server
3813 through IPMI or similar, we could do that through a local trigger command
3814 instead. Let's assume we have a script that does IPMI reboot of a given hostname,
3815 ipmi-reboot. On localbox, we could then have run fio with a local trigger
3818 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
3820 For this case, fio would wait for the server to send us the write state, then
3821 execute ``ipmi-reboot server`` when that happened.
3823 Loading verify state
3824 ~~~~~~~~~~~~~~~~~~~~
3826 To load stored write state, a read verification job file must contain the
3827 :option:`verify_state_load` option. If that is set, fio will load the previously
3828 stored state. For a local fio run this is done by loading the files directly,
3829 and on a client/server run, the server backend will ask the client to send the
3830 files over and load them from there.
3836 Fio supports a variety of log file formats, for logging latencies, bandwidth,
3837 and IOPS. The logs share a common format, which looks like this:
3839 *time* (`msec`), *value*, *data direction*, *block size* (`bytes`),
3842 *Time* for the log entry is always in milliseconds. The *value* logged depends
3843 on the type of log, it will be one of the following:
3846 Value is latency in nsecs
3852 *Data direction* is one of the following:
3861 The entry's *block size* is always in bytes. The *offset* is the position in bytes
3862 from the start of the file for that particular I/O. The logging of the offset can be
3863 toggled with :option:`log_offset`.
3865 Fio defaults to logging every individual I/O but when windowed logging is set
3866 through :option:`log_avg_msec`, either the average (by default) or the maximum
3867 (:option:`log_max_value` is set) *value* seen over the specified period of time
3868 is recorded. Each *data direction* seen within the window period will aggregate
3869 its values in a separate row. Further, when using windowed logging the *block
3870 size* and *offset* entries will always contain 0.
3875 Normally fio is invoked as a stand-alone application on the machine where the
3876 I/O workload should be generated. However, the backend and frontend of fio can
3877 be run separately i.e., the fio server can generate an I/O workload on the "Device
3878 Under Test" while being controlled by a client on another machine.
3880 Start the server on the machine which has access to the storage DUT::
3884 where `args` defines what fio listens to. The arguments are of the form
3885 ``type,hostname`` or ``IP,port``. *type* is either ``ip`` (or ip4) for TCP/IP
3886 v4, ``ip6`` for TCP/IP v6, or ``sock`` for a local unix domain socket.
3887 *hostname* is either a hostname or IP address, and *port* is the port to listen
3888 to (only valid for TCP/IP, not a local socket). Some examples:
3892 Start a fio server, listening on all interfaces on the default port (8765).
3894 2) ``fio --server=ip:hostname,4444``
3896 Start a fio server, listening on IP belonging to hostname and on port 4444.
3898 3) ``fio --server=ip6:::1,4444``
3900 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
3902 4) ``fio --server=,4444``
3904 Start a fio server, listening on all interfaces on port 4444.
3906 5) ``fio --server=1.2.3.4``
3908 Start a fio server, listening on IP 1.2.3.4 on the default port.
3910 6) ``fio --server=sock:/tmp/fio.sock``
3912 Start a fio server, listening on the local socket :file:`/tmp/fio.sock`.
3914 Once a server is running, a "client" can connect to the fio server with::
3916 fio <local-args> --client=<server> <remote-args> <job file(s)>
3918 where `local-args` are arguments for the client where it is running, `server`
3919 is the connect string, and `remote-args` and `job file(s)` are sent to the
3920 server. The `server` string follows the same format as it does on the server
3921 side, to allow IP/hostname/socket and port strings.
3923 Fio can connect to multiple servers this way::
3925 fio --client=<server1> <job file(s)> --client=<server2> <job file(s)>
3927 If the job file is located on the fio server, then you can tell the server to
3928 load a local file as well. This is done by using :option:`--remote-config` ::
3930 fio --client=server --remote-config /path/to/file.fio
3932 Then fio will open this local (to the server) job file instead of being passed
3933 one from the client.
3935 If you have many servers (example: 100 VMs/containers), you can input a pathname
3936 of a file containing host IPs/names as the parameter value for the
3937 :option:`--client` option. For example, here is an example :file:`host.list`
3938 file containing 2 hostnames::
3940 host1.your.dns.domain
3941 host2.your.dns.domain
3943 The fio command would then be::
3945 fio --client=host.list <job file(s)>
3947 In this mode, you cannot input server-specific parameters or job files -- all
3948 servers receive the same job file.
3950 In order to let ``fio --client`` runs use a shared filesystem from multiple
3951 hosts, ``fio --client`` now prepends the IP address of the server to the
3952 filename. For example, if fio is using the directory :file:`/mnt/nfs/fio` and is
3953 writing filename :file:`fileio.tmp`, with a :option:`--client` `hostfile`
3954 containing two hostnames ``h1`` and ``h2`` with IP addresses 192.168.10.120 and
3955 192.168.10.121, then fio will create two files::
3957 /mnt/nfs/fio/192.168.10.120.fileio.tmp
3958 /mnt/nfs/fio/192.168.10.121.fileio.tmp