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 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:: --bandwidth-log
109 Generate aggregate bandwidth logs.
111 .. option:: --minimal
113 Print statistics in a terse, semicolon-delimited format.
115 .. option:: --append-terse
117 Print statistics in selected mode AND terse, semicolon-delimited format.
118 **deprecated**, use :option:`--output-format` instead to select multiple
121 .. option:: --output-format=type
123 Set the reporting format to `normal`, `terse`, `json`, or `json+`. Multiple
124 formats can be selected, separated by a comma. `terse` is a CSV based
125 format. `json+` is like `json`, except it adds a full dump of the latency
128 .. option:: --terse-version=type
130 Set terse version output format (default 3, or 2 or 4 or 5).
132 .. option:: --version
134 Print version info 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 be passed, in
148 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 :option:`ioengine`, or print help for `command`
157 defined by :option:`ioengine`. If no :option:`ioengine` is given, list all
160 .. option:: --showcmd=jobfile
162 Turn a job file into command line options.
164 .. option:: --readonly
166 Turn on safety read-only checks, preventing writes. The ``--readonly``
167 option is an extra safety guard to prevent users from accidentally starting
168 a write workload when that is not desired. Fio will only write if
169 `rw=write/randwrite/rw/randrw` is given. This extra safety net can be used
170 as an extra precaution as ``--readonly`` will also enable a write check in
171 the I/O engine core to prevent writes due to unknown user space bug(s).
173 .. option:: --eta=when
175 When real-time ETA estimate should be printed. May be `always`, `never` or
178 .. option:: --eta-newline=time
180 Force a new line for every `time` period passed. When the unit is omitted,
181 the value is interpreted in seconds.
183 .. option:: --status-interval=time
185 Force full status dump every `time` period passed. When the unit is
186 omitted, the value is interpreted in seconds.
188 .. option:: --section=name
190 Only run specified section in job file. Multiple sections can be specified.
191 The ``--section`` option allows one to combine related jobs into one file.
192 E.g. one job file could define light, moderate, and heavy sections. Tell
193 fio to run only the "heavy" section by giving ``--section=heavy``
194 command line option. One can also specify the "write" operations in one
195 section and "verify" operation in another section. The ``--section`` option
196 only applies to job sections. The reserved *global* section is always
199 .. option:: --alloc-size=kb
201 Set the internal smalloc pool to this size in KiB. The
202 ``--alloc-size`` switch allows one to use a larger pool size for smalloc.
203 If running large jobs with randommap enabled, fio can run out of memory.
204 Smalloc is an internal allocator for shared structures from a fixed size
205 memory pool and can grow to 16 pools. The pool size defaults to 16MiB.
207 NOTE: While running :file:`.fio_smalloc.*` backing store files are visible
210 .. option:: --warnings-fatal
212 All fio parser warnings are fatal, causing fio to exit with an
215 .. option:: --max-jobs=nr
217 Maximum number of threads/processes to support.
219 .. option:: --server=args
221 Start a backend server, with `args` specifying what to listen to.
222 See `Client/Server`_ section.
224 .. option:: --daemonize=pidfile
226 Background a fio server, writing the pid to the given `pidfile` file.
228 .. option:: --client=hostname
230 Instead of running the jobs locally, send and run them on the given host or
231 set of hosts. See `Client/Server`_ section.
233 .. option:: --remote-config=file
235 Tell fio server to load this local file.
237 .. option:: --idle-prof=option
239 Report CPU idleness. *option* is one of the following:
242 Run unit work calibration only and exit.
245 Show aggregate system idleness and unit work.
248 As **system** but also show per CPU idleness.
250 .. option:: --inflate-log=log
252 Inflate and output compressed log.
254 .. option:: --trigger-file=file
256 Execute trigger cmd when file exists.
258 .. option:: --trigger-timeout=t
260 Execute trigger at this time.
262 .. option:: --trigger=cmd
264 Set this command as local trigger.
266 .. option:: --trigger-remote=cmd
268 Set this command as remote trigger.
270 .. option:: --aux-path=path
272 Use this path for fio state generated files.
274 Any parameters following the options will be assumed to be job files, unless
275 they match a job file parameter. Multiple job files can be listed and each job
276 file will be regarded as a separate group. Fio will :option:`stonewall`
277 execution between each group.
283 As previously described, fio accepts one or more job files describing what it is
284 supposed to do. The job file format is the classic ini file, where the names
285 enclosed in [] brackets define the job name. You are free to use any ASCII name
286 you want, except *global* which has special meaning. Following the job name is
287 a sequence of zero or more parameters, one per line, that define the behavior of
288 the job. If the first character in a line is a ';' or a '#', the entire line is
289 discarded as a comment.
291 A *global* section sets defaults for the jobs described in that file. A job may
292 override a *global* section parameter, and a job file may even have several
293 *global* sections if so desired. A job is only affected by a *global* section
296 The :option:`--cmdhelp` option also lists all options. If used with an `option`
297 argument, :option:`--cmdhelp` will detail the given `option`.
299 See the `examples/` directory for inspiration on how to write job files. Note
300 the copyright and license requirements currently apply to `examples/` files.
302 So let's look at a really simple job file that defines two processes, each
303 randomly reading from a 128MiB file:
307 ; -- start job file --
318 As you can see, the job file sections themselves are empty as all the described
319 parameters are shared. As no :option:`filename` option is given, fio makes up a
320 `filename` for each of the jobs as it sees fit. On the command line, this job
321 would look as follows::
323 $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
326 Let's look at an example that has a number of processes writing randomly to
331 ; -- start job file --
342 Here we have no *global* section, as we only have one job defined anyway. We
343 want to use async I/O here, with a depth of 4 for each file. We also increased
344 the buffer size used to 32KiB and define numjobs to 4 to fork 4 identical
345 jobs. The result is 4 processes each randomly writing to their own 64MiB
346 file. Instead of using the above job file, you could have given the parameters
347 on the command line. For this case, you would specify::
349 $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
351 When fio is utilized as a basis of any reasonably large test suite, it might be
352 desirable to share a set of standardized settings across multiple job files.
353 Instead of copy/pasting such settings, any section may pull in an external
354 :file:`filename.fio` file with *include filename* directive, as in the following
357 ; -- start job file including.fio --
361 include glob-include.fio
368 include test-include.fio
369 ; -- end job file including.fio --
373 ; -- start job file glob-include.fio --
376 ; -- end job file glob-include.fio --
380 ; -- start job file test-include.fio --
383 ; -- end job file test-include.fio --
385 Settings pulled into a section apply to that section only (except *global*
386 section). Include directives may be nested in that any included file may contain
387 further include directive(s). Include files may not contain [] sections.
390 Environment variables
391 ~~~~~~~~~~~~~~~~~~~~~
393 Fio also supports environment variable expansion in job files. Any sub-string of
394 the form ``${VARNAME}`` as part of an option value (in other words, on the right
395 of the '='), will be expanded to the value of the environment variable called
396 `VARNAME`. If no such environment variable is defined, or `VARNAME` is the
397 empty string, the empty string will be substituted.
399 As an example, let's look at a sample fio invocation and job file::
401 $ SIZE=64m NUMJOBS=4 fio jobfile.fio
405 ; -- start job file --
412 This will expand to the following equivalent job file at runtime:
416 ; -- start job file --
423 Fio ships with a few example job files, you can also look there for inspiration.
428 Additionally, fio has a set of reserved keywords that will be replaced
429 internally with the appropriate value. Those keywords are:
433 The architecture page size of the running system.
437 Megabytes of total memory in the system.
441 Number of online available CPUs.
443 These can be used on the command line or in the job file, and will be
444 automatically substituted with the current system values when the job is
445 run. Simple math is also supported on these keywords, so you can perform actions
450 and get that properly expanded to 8 times the size of memory in the machine.
456 This section describes in details each parameter associated with a job. Some
457 parameters take an option of a given type, such as an integer or a
458 string. Anywhere a numeric value is required, an arithmetic expression may be
459 used, provided it is surrounded by parentheses. Supported operators are:
468 For time values in expressions, units are microseconds by default. This is
469 different than for time values not in expressions (not enclosed in
470 parentheses). The following types are used:
477 String. This is a sequence of alpha characters.
480 Integer with possible time suffix. Without a unit value is interpreted as
481 seconds unless otherwise specified. Accepts a suffix of 'd' for days, 'h' for
482 hours, 'm' for minutes, 's' for seconds, 'ms' (or 'msec') for milliseconds and
483 'us' (or 'usec') for microseconds. For example, use 10m for 10 minutes.
488 Integer. A whole number value, which may contain an integer prefix
489 and an integer suffix:
491 [*integer prefix*] **number** [*integer suffix*]
493 The optional *integer prefix* specifies the number's base. The default
494 is decimal. *0x* specifies hexadecimal.
496 The optional *integer suffix* specifies the number's units, and includes an
497 optional unit prefix and an optional unit. For quantities of data, the
498 default unit is bytes. For quantities of time, the default unit is seconds
499 unless otherwise specified.
501 With :option:`kb_base`\=1000, fio follows international standards for unit
502 prefixes. To specify power-of-10 decimal values defined in the
503 International System of Units (SI):
505 * *Ki* -- means kilo (K) or 1000
506 * *Mi* -- means mega (M) or 1000**2
507 * *Gi* -- means giga (G) or 1000**3
508 * *Ti* -- means tera (T) or 1000**4
509 * *Pi* -- means peta (P) or 1000**5
511 To specify power-of-2 binary values defined in IEC 80000-13:
513 * *k* -- means kibi (Ki) or 1024
514 * *M* -- means mebi (Mi) or 1024**2
515 * *G* -- means gibi (Gi) or 1024**3
516 * *T* -- means tebi (Ti) or 1024**4
517 * *P* -- means pebi (Pi) or 1024**5
519 With :option:`kb_base`\=1024 (the default), the unit prefixes are opposite
520 from those specified in the SI and IEC 80000-13 standards to provide
521 compatibility with old scripts. For example, 4k means 4096.
523 For quantities of data, an optional unit of 'B' may be included
524 (e.g., 'kB' is the same as 'k').
526 The *integer suffix* is not case sensitive (e.g., m/mi mean mebi/mega,
527 not milli). 'b' and 'B' both mean byte, not bit.
529 Examples with :option:`kb_base`\=1000:
531 * *4 KiB*: 4096, 4096b, 4096B, 4ki, 4kib, 4kiB, 4Ki, 4KiB
532 * *1 MiB*: 1048576, 1mi, 1024ki
533 * *1 MB*: 1000000, 1m, 1000k
534 * *1 TiB*: 1099511627776, 1ti, 1024gi, 1048576mi
535 * *1 TB*: 1000000000, 1t, 1000m, 1000000k
537 Examples with :option:`kb_base`\=1024 (default):
539 * *4 KiB*: 4096, 4096b, 4096B, 4k, 4kb, 4kB, 4K, 4KB
540 * *1 MiB*: 1048576, 1m, 1024k
541 * *1 MB*: 1000000, 1mi, 1000ki
542 * *1 TiB*: 1099511627776, 1t, 1024g, 1048576m
543 * *1 TB*: 1000000000, 1ti, 1000mi, 1000000ki
545 To specify times (units are not case sensitive):
549 * *M* -- means minutes
550 * *s* -- or sec means seconds (default)
551 * *ms* -- or *msec* means milliseconds
552 * *us* -- or *usec* means microseconds
554 If the option accepts an upper and lower range, use a colon ':' or
555 minus '-' to separate such values. See :ref:`irange <irange>`.
556 If the lower value specified happens to be larger than the upper value
557 the two values are swapped.
562 Boolean. Usually parsed as an integer, however only defined for
563 true and false (1 and 0).
568 Integer range with suffix. Allows value range to be given, such as
569 1024-4096. A colon may also be used as the separator, e.g. 1k:4k. If the
570 option allows two sets of ranges, they can be specified with a ',' or '/'
571 delimiter: 1k-4k/8k-32k. Also see :ref:`int <int>`.
574 A list of floating point numbers, separated by a ':' character.
580 .. option:: kb_base=int
582 Select the interpretation of unit prefixes in input parameters.
585 Inputs comply with IEC 80000-13 and the International
586 System of Units (SI). Use:
588 - power-of-2 values with IEC prefixes (e.g., KiB)
589 - power-of-10 values with SI prefixes (e.g., kB)
592 Compatibility mode (default). To avoid breaking old scripts:
594 - power-of-2 values with SI prefixes
595 - power-of-10 values with IEC prefixes
597 See :option:`bs` for more details on input parameters.
599 Outputs always use correct prefixes. Most outputs include both
602 bw=2383.3kB/s (2327.4KiB/s)
604 If only one value is reported, then kb_base selects the one to use:
606 **1000** -- SI prefixes
608 **1024** -- IEC prefixes
610 .. option:: unit_base=int
612 Base unit for reporting. Allowed values are:
615 Use auto-detection (default).
622 With the above in mind, here follows the complete list of fio job parameters.
630 ASCII name of the job. This may be used to override the name printed by fio
631 for this job. Otherwise the job name is used. On the command line this
632 parameter has the special purpose of also signaling the start of a new job.
634 .. option:: description=str
636 Text description of the job. Doesn't do anything except dump this text
637 description when this job is run. It's not parsed.
639 .. option:: loops=int
641 Run the specified number of iterations of this job. Used to repeat the same
642 workload a given number of times. Defaults to 1.
644 .. option:: numjobs=int
646 Create the specified number of clones of this job. Each clone of job
647 is spawned as an independent thread or process. May be used to setup a
648 larger number of threads/processes doing the same thing. Each thread is
649 reported separately; to see statistics for all clones as a whole, use
650 :option:`group_reporting` in conjunction with :option:`new_group`.
651 See :option:`--max-jobs`. Default: 1.
654 Time related parameters
655 ~~~~~~~~~~~~~~~~~~~~~~~
657 .. option:: runtime=time
659 Tell fio to terminate processing after the specified period of time. It
660 can be quite hard to determine for how long a specified job will run, so
661 this parameter is handy to cap the total runtime to a given time. When
662 the unit is omitted, the value is intepreted in seconds.
664 .. option:: time_based
666 If set, fio will run for the duration of the :option:`runtime` specified
667 even if the file(s) are completely read or written. It will simply loop over
668 the same workload as many times as the :option:`runtime` allows.
670 .. option:: startdelay=irange(time)
672 Delay the start of job for the specified amount of time. Can be a single
673 value or a range. When given as a range, each thread will choose a value
674 randomly from within the range. Value is in seconds if a unit is omitted.
676 .. option:: ramp_time=time
678 If set, fio will run the specified workload for this amount of time before
679 logging any performance numbers. Useful for letting performance settle
680 before logging results, thus minimizing the runtime required for stable
681 results. Note that the ``ramp_time`` is considered lead in time for a job,
682 thus it will increase the total runtime if a special timeout or
683 :option:`runtime` is specified. When the unit is omitted, the value is
686 .. option:: clocksource=str
688 Use the given clocksource as the base of timing. The supported options are:
691 :manpage:`gettimeofday(2)`
694 :manpage:`clock_gettime(2)`
697 Internal CPU clock source
699 cpu is the preferred clocksource if it is reliable, as it is very fast (and
700 fio is heavy on time calls). Fio will automatically use this clocksource if
701 it's supported and considered reliable on the system it is running on,
702 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
703 means supporting TSC Invariant.
705 .. option:: gtod_reduce=bool
707 Enable all of the :manpage:`gettimeofday(2)` reducing options
708 (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus
709 reduce precision of the timeout somewhat to really shrink the
710 :manpage:`gettimeofday(2)` call count. With this option enabled, we only do
711 about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all
712 time keeping was enabled.
714 .. option:: gtod_cpu=int
716 Sometimes it's cheaper to dedicate a single thread of execution to just
717 getting the current time. Fio (and databases, for instance) are very
718 intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set
719 one CPU aside for doing nothing but logging current time to a shared memory
720 location. Then the other threads/processes that run I/O workloads need only
721 copy that segment, instead of entering the kernel with a
722 :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time
723 calls will be excluded from other uses. Fio will manually clear it from the
724 CPU mask of other jobs.
730 .. option:: directory=str
732 Prefix filenames with this directory. Used to place files in a different
733 location than :file:`./`. You can specify a number of directories by
734 separating the names with a ':' character. These directories will be
735 assigned equally distributed to job clones created by :option:`numjobs` as
736 long as they are using generated filenames. If specific `filename(s)` are
737 set fio will use the first listed directory, and thereby matching the
738 `filename` semantic which generates a file each clone if not specified, but
739 let all clones use the same if set.
741 See the :option:`filename` option for information on how to escape "``:``" and
742 "``\``" characters within the directory path itself.
744 .. option:: filename=str
746 Fio normally makes up a `filename` based on the job name, thread number, and
747 file number (see :option:`filename_format`). If you want to share files
748 between threads in a job or several
749 jobs with fixed file paths, specify a `filename` for each of them to override
750 the default. If the ioengine is file based, you can specify a number of files
751 by separating the names with a ':' colon. So if you wanted a job to open
752 :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use
753 ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is
754 specified, :option:`nrfiles` is ignored. The size of regular files specified
755 by this option will be :option:`size` divided by number of files unless an
756 explicit size is specified by :option:`filesize`.
758 Each colon and backslash in the wanted path must be escaped with a ``\``
759 character. For instance, if the path is :file:`/dev/dsk/foo@3,0:c` then you
760 would use ``filename=/dev/dsk/foo@3,0\:c`` and if the path is
761 :file:`F:\\filename` then you would use ``filename=F\:\\filename``.
763 On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for
764 the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc.
765 Note: Windows and FreeBSD prevent write access to areas
766 of the disk containing in-use data (e.g. filesystems).
768 The filename "`-`" is a reserved name, meaning *stdin* or *stdout*. Which
769 of the two depends on the read/write direction set.
771 .. option:: filename_format=str
773 If sharing multiple files between jobs, it is usually necessary to have fio
774 generate the exact names that you want. By default, fio will name a file
775 based on the default file format specification of
776 :file:`jobname.jobnumber.filenumber`. With this option, that can be
777 customized. Fio will recognize and replace the following keywords in this
781 The name of the worker thread or process.
783 The incremental number of the worker thread or process.
785 The incremental number of the file for that worker thread or
788 To have dependent jobs share a set of files, this option can be set to have
789 fio generate filenames that are shared between the two. For instance, if
790 :file:`testfiles.$filenum` is specified, file number 4 for any job will be
791 named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum`
792 will be used if no other format specifier is given.
794 .. option:: unique_filename=bool
796 To avoid collisions between networked clients, fio defaults to prefixing any
797 generated filenames (with a directory specified) with the source of the
798 client connecting. To disable this behavior, set this option to 0.
800 .. option:: opendir=str
802 Recursively open any files below directory `str`.
804 .. option:: lockfile=str
806 Fio defaults to not locking any files before it does I/O to them. If a file
807 or file descriptor is shared, fio can serialize I/O to that file to make the
808 end result consistent. This is usual for emulating real workloads that share
809 files. The lock modes are:
812 No locking. The default.
814 Only one thread or process may do I/O at a time, excluding all
817 Read-write locking on the file. Many readers may
818 access the file at the same time, but writes get exclusive access.
820 .. option:: nrfiles=int
822 Number of files to use for this job. Defaults to 1. The size of files
823 will be :option:`size` divided by this unless explicit size is specified by
824 :option:`filesize`. Files are created for each thread separately, and each
825 file will have a file number within its name by default, as explained in
826 :option:`filename` section.
829 .. option:: openfiles=int
831 Number of files to keep open at the same time. Defaults to the same as
832 :option:`nrfiles`, can be set smaller to limit the number simultaneous
835 .. option:: file_service_type=str
837 Defines how fio decides which file from a job to service next. The following
841 Choose a file at random.
844 Round robin over opened files. This is the default.
847 Finish one file before moving on to the next. Multiple files can
848 still be open depending on 'openfiles'.
851 Use a *Zipf* distribution to decide what file to access.
854 Use a *Pareto* distribution to decide what file to access.
857 Use a *Gaussian* (normal) distribution to decide what file to
863 For *random*, *roundrobin*, and *sequential*, a postfix can be appended to
864 tell fio how many I/Os to issue before switching to a new file. For example,
865 specifying ``file_service_type=random:8`` would cause fio to issue
866 8 I/Os before selecting a new file at random. For the non-uniform
867 distributions, a floating point postfix can be given to influence how the
868 distribution is skewed. See :option:`random_distribution` for a description
869 of how that would work.
871 .. option:: ioscheduler=str
873 Attempt to switch the device hosting the file to the specified I/O scheduler
876 .. option:: create_serialize=bool
878 If true, serialize the file creation for the jobs. This may be handy to
879 avoid interleaving of data files, which may greatly depend on the filesystem
880 used and even the number of processors in the system. Default: true.
882 .. option:: create_fsync=bool
884 :manpage:`fsync(2)` the data file after creation. This is the default.
886 .. option:: create_on_open=bool
888 If true, don't pre-create files but allow the job's open() to create a file
889 when it's time to do I/O. Default: false -- pre-create all necessary files
892 .. option:: create_only=bool
894 If true, fio will only run the setup phase of the job. If files need to be
895 laid out or updated on disk, only that will be done -- the actual job contents
896 are not executed. Default: false.
898 .. option:: allow_file_create=bool
900 If true, fio is permitted to create files as part of its workload. If this
901 option is false, then fio will error out if
902 the files it needs to use don't already exist. Default: true.
904 .. option:: allow_mounted_write=bool
906 If this isn't set, fio will abort jobs that are destructive (e.g. that write)
907 to what appears to be a mounted device or partition. This should help catch
908 creating inadvertently destructive tests, not realizing that the test will
909 destroy data on the mounted file system. Note that some platforms don't allow
910 writing against a mounted device regardless of this option. Default: false.
912 .. option:: pre_read=bool
914 If this is given, files will be pre-read into memory before starting the
915 given I/O operation. This will also clear the :option:`invalidate` flag,
916 since it is pointless to pre-read and then drop the cache. This will only
917 work for I/O engines that are seek-able, since they allow you to read the
918 same data multiple times. Thus it will not work on non-seekable I/O engines
919 (e.g. network, splice). Default: false.
921 .. option:: unlink=bool
923 Unlink the job files when done. Not the default, as repeated runs of that
924 job would then waste time recreating the file set again and again. Default:
927 .. option:: unlink_each_loop=bool
929 Unlink job files after each iteration or loop. Default: false.
931 .. option:: zonesize=int
933 Divide a file into zones of the specified size. See :option:`zoneskip`.
935 .. option:: zonerange=int
937 Give size of an I/O zone. See :option:`zoneskip`.
939 .. option:: zoneskip=int
941 Skip the specified number of bytes when :option:`zonesize` data has been
942 read. The two zone options can be used to only do I/O on zones of a file.
948 .. option:: direct=bool
950 If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that
951 ZFS on Solaris doesn't support direct I/O. On Windows the synchronous
952 ioengines don't support direct I/O. Default: false.
954 .. option:: atomic=bool
956 If value is true, attempt to use atomic direct I/O. Atomic writes are
957 guaranteed to be stable once acknowledged by the operating system. Only
958 Linux supports O_ATOMIC right now.
960 .. option:: buffered=bool
962 If value is true, use buffered I/O. This is the opposite of the
963 :option:`direct` option. Defaults to true.
965 .. option:: readwrite=str, rw=str
967 Type of I/O pattern. Accepted values are:
974 Sequential trims (Linux block devices only).
980 Random trims (Linux block devices only).
982 Sequential mixed reads and writes.
984 Random mixed reads and writes.
986 Sequential trim+write sequences. Blocks will be trimmed first,
987 then the same blocks will be written to.
989 Fio defaults to read if the option is not specified. For the mixed I/O
990 types, the default is to split them 50/50. For certain types of I/O the
991 result may still be skewed a bit, since the speed may be different.
993 It is possible to specify the number of I/Os to do before getting a new
994 offset by appending ``:<nr>`` to the end of the string given. For a
995 random read, it would look like ``rw=randread:8`` for passing in an offset
996 modifier with a value of 8. If the suffix is used with a sequential I/O
997 pattern, then the *<nr>* value specified will be **added** to the generated
998 offset for each I/O turning sequential I/O into sequential I/O with holes.
999 For instance, using ``rw=write:4k`` will skip 4k for every write. Also see
1000 the :option:`rw_sequencer` option.
1002 .. option:: rw_sequencer=str
1004 If an offset modifier is given by appending a number to the ``rw=<str>``
1005 line, then this option controls how that number modifies the I/O offset
1006 being generated. Accepted values are:
1009 Generate sequential offset.
1011 Generate the same offset.
1013 ``sequential`` is only useful for random I/O, where fio would normally
1014 generate a new random offset for every I/O. If you append e.g. 8 to randread,
1015 you would get a new random offset for every 8 I/O's. The result would be a
1016 seek for only every 8 I/O's, instead of for every I/O. Use ``rw=randread:8``
1017 to specify that. As sequential I/O is already sequential, setting
1018 ``sequential`` for that would not result in any differences. ``identical``
1019 behaves in a similar fashion, except it sends the same offset 8 number of
1020 times before generating a new offset.
1022 .. option:: unified_rw_reporting=bool
1024 Fio normally reports statistics on a per data direction basis, meaning that
1025 reads, writes, and trims are accounted and reported separately. If this
1026 option is set fio sums the results and report them as "mixed" instead.
1028 .. option:: randrepeat=bool
1030 Seed the random number generator used for random I/O patterns in a
1031 predictable way so the pattern is repeatable across runs. Default: true.
1033 .. option:: allrandrepeat=bool
1035 Seed all random number generators in a predictable way so results are
1036 repeatable across runs. Default: false.
1038 .. option:: randseed=int
1040 Seed the random number generators based on this seed value, to be able to
1041 control what sequence of output is being generated. If not set, the random
1042 sequence depends on the :option:`randrepeat` setting.
1044 .. option:: fallocate=str
1046 Whether pre-allocation is performed when laying down files.
1047 Accepted values are:
1050 Do not pre-allocate space.
1053 Use a platform's native pre-allocation call but fall back to
1054 **none** behavior if it fails/is not implemented.
1057 Pre-allocate via :manpage:`posix_fallocate(3)`.
1060 Pre-allocate via :manpage:`fallocate(2)` with
1061 FALLOC_FL_KEEP_SIZE set.
1064 Backward-compatible alias for **none**.
1067 Backward-compatible alias for **posix**.
1069 May not be available on all supported platforms. **keep** is only available
1070 on Linux. If using ZFS on Solaris this cannot be set to **posix**
1071 because ZFS doesn't support pre-allocation. Default: **native** if any
1072 pre-allocation methods are available, **none** if not.
1074 .. option:: fadvise_hint=str
1076 Use :manpage:`posix_fadvise(2)` to advise the kernel on what I/O patterns
1077 are likely to be issued. Accepted values are:
1080 Backwards-compatible hint for "no hint".
1083 Backwards compatible hint for "advise with fio workload type". This
1084 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1085 for a sequential workload.
1088 Advise using **FADV_SEQUENTIAL**.
1091 Advise using **FADV_RANDOM**.
1093 .. option:: fadvise_stream=int
1095 Use :manpage:`posix_fadvise(2)` to advise the kernel what stream ID the
1096 writes issued belong to. Only supported on Linux. Note, this option may
1097 change going forward.
1099 .. option:: offset=int
1101 Start I/O at the provided offset in the file, given as either a fixed size in
1102 bytes or a percentage. If a percentage is given, the next ``blockalign``-ed
1103 offset will be used. Data before the given offset will not be touched. This
1104 effectively caps the file size at `real_size - offset`. Can be combined with
1105 :option:`size` to constrain the start and end range of the I/O workload.
1106 A percentage can be specified by a number between 1 and 100 followed by '%',
1107 for example, ``offset=20%`` to specify 20%.
1109 .. option:: offset_increment=int
1111 If this is provided, then the real offset becomes `offset + offset_increment
1112 * thread_number`, where the thread number is a counter that starts at 0 and
1113 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1114 specified). This option is useful if there are several jobs which are
1115 intended to operate on a file in parallel disjoint segments, with even
1116 spacing between the starting points.
1118 .. option:: number_ios=int
1120 Fio will normally perform I/Os until it has exhausted the size of the region
1121 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1122 condition). With this setting, the range/size can be set independently of
1123 the number of I/Os to perform. When fio reaches this number, it will exit
1124 normally and report status. Note that this does not extend the amount of I/O
1125 that will be done, it will only stop fio if this condition is met before
1126 other end-of-job criteria.
1128 .. option:: fsync=int
1130 If writing to a file, issue an :manpage:`fsync(2)` (or its equivalent) of
1131 the dirty data for every number of blocks given. For example, if you give 32
1132 as a parameter, fio will sync the file after every 32 writes issued. If fio is
1133 using non-buffered I/O, we may not sync the file. The exception is the sg
1134 I/O engine, which synchronizes the disk cache anyway. Defaults to 0, which
1135 means fio does not periodically issue and wait for a sync to complete. Also
1136 see :option:`end_fsync` and :option:`fsync_on_close`.
1138 .. option:: fdatasync=int
1140 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
1141 not metadata blocks. In Windows, FreeBSD, and DragonFlyBSD there is no
1142 :manpage:`fdatasync(2)` so this falls back to using :manpage:`fsync(2)`.
1143 Defaults to 0, which means fio does not periodically issue and wait for a
1144 data-only sync to complete.
1146 .. option:: write_barrier=int
1148 Make every `N-th` write a barrier write.
1150 .. option:: sync_file_range=str:val
1152 Use :manpage:`sync_file_range(2)` for every `val` number of write
1153 operations. Fio will track range of writes that have happened since the last
1154 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1157 SYNC_FILE_RANGE_WAIT_BEFORE
1159 SYNC_FILE_RANGE_WRITE
1161 SYNC_FILE_RANGE_WAIT_AFTER
1163 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1164 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1165 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1168 .. option:: overwrite=bool
1170 If true, writes to a file will always overwrite existing data. If the file
1171 doesn't already exist, it will be created before the write phase begins. If
1172 the file exists and is large enough for the specified write phase, nothing
1173 will be done. Default: false.
1175 .. option:: end_fsync=bool
1177 If true, :manpage:`fsync(2)` file contents when a write stage has completed.
1180 .. option:: fsync_on_close=bool
1182 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
1183 from :option:`end_fsync` in that it will happen on every file close, not
1184 just at the end of the job. Default: false.
1186 .. option:: rwmixread=int
1188 Percentage of a mixed workload that should be reads. Default: 50.
1190 .. option:: rwmixwrite=int
1192 Percentage of a mixed workload that should be writes. If both
1193 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1194 add up to 100%, the latter of the two will be used to override the
1195 first. This may interfere with a given rate setting, if fio is asked to
1196 limit reads or writes to a certain rate. If that is the case, then the
1197 distribution may be skewed. Default: 50.
1199 .. option:: random_distribution=str:float[,str:float][,str:float]
1201 By default, fio will use a completely uniform random distribution when asked
1202 to perform random I/O. Sometimes it is useful to skew the distribution in
1203 specific ways, ensuring that some parts of the data is more hot than others.
1204 fio includes the following distribution models:
1207 Uniform random distribution
1216 Normal (Gaussian) distribution
1219 Zoned random distribution
1221 When using a **zipf** or **pareto** distribution, an input value is also
1222 needed to define the access pattern. For **zipf**, this is the `zipf
1223 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
1224 program, :command:`genzipf`, that can be used visualize what the given input
1225 values will yield in terms of hit rates. If you wanted to use **zipf** with
1226 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1227 option. If a non-uniform model is used, fio will disable use of the random
1228 map. For the **normal** distribution, a normal (Gaussian) deviation is
1229 supplied as a value between 0 and 100.
1231 For a **zoned** distribution, fio supports specifying percentages of I/O
1232 access that should fall within what range of the file or device. For
1233 example, given a criteria of:
1235 * 60% of accesses should be to the first 10%
1236 * 30% of accesses should be to the next 20%
1237 * 8% of accesses should be to to the next 30%
1238 * 2% of accesses should be to the next 40%
1240 we can define that through zoning of the random accesses. For the above
1241 example, the user would do::
1243 random_distribution=zoned:60/10:30/20:8/30:2/40
1245 similarly to how :option:`bssplit` works for setting ranges and percentages
1246 of block sizes. Like :option:`bssplit`, it's possible to specify separate
1247 zones for reads, writes, and trims. If just one set is given, it'll apply to
1250 .. option:: percentage_random=int[,int][,int]
1252 For a random workload, set how big a percentage should be random. This
1253 defaults to 100%, in which case the workload is fully random. It can be set
1254 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1255 sequential. Any setting in between will result in a random mix of sequential
1256 and random I/O, at the given percentages. Comma-separated values may be
1257 specified for reads, writes, and trims as described in :option:`blocksize`.
1259 .. option:: norandommap
1261 Normally fio will cover every block of the file when doing random I/O. If
1262 this option is given, fio will just get a new random offset without looking
1263 at past I/O history. This means that some blocks may not be read or written,
1264 and that some blocks may be read/written more than once. If this option is
1265 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1266 only intact blocks are verified, i.e., partially-overwritten blocks are
1269 .. option:: softrandommap=bool
1271 See :option:`norandommap`. If fio runs with the random block map enabled and
1272 it fails to allocate the map, if this option is set it will continue without
1273 a random block map. As coverage will not be as complete as with random maps,
1274 this option is disabled by default.
1276 .. option:: random_generator=str
1278 Fio supports the following engines for generating
1279 I/O offsets for random I/O:
1282 Strong 2^88 cycle random number generator
1284 Linear feedback shift register generator
1286 Strong 64-bit 2^258 cycle random number generator
1288 **tausworthe** is a strong random number generator, but it requires tracking
1289 on the side if we want to ensure that blocks are only read or written
1290 once. **LFSR** guarantees that we never generate the same offset twice, and
1291 it's also less computationally expensive. It's not a true random generator,
1292 however, though for I/O purposes it's typically good enough. **LFSR** only
1293 works with single block sizes, not with workloads that use multiple block
1294 sizes. If used with such a workload, fio may read or write some blocks
1295 multiple times. The default value is **tausworthe**, unless the required
1296 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1297 selected automatically.
1303 .. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1305 The block size in bytes used for I/O units. Default: 4096. A single value
1306 applies to reads, writes, and trims. Comma-separated values may be
1307 specified for reads, writes, and trims. A value not terminated in a comma
1308 applies to subsequent types.
1313 means 256k for reads, writes and trims.
1316 means 8k for reads, 32k for writes and trims.
1319 means 8k for reads, 32k for writes, and default for trims.
1322 means default for reads, 8k for writes and trims.
1325 means default for reads, 8k for writes, and default for trims.
1327 .. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1329 A range of block sizes in bytes for I/O units. The issued I/O unit will
1330 always be a multiple of the minimum size, unless
1331 :option:`blocksize_unaligned` is set.
1333 Comma-separated ranges may be specified for reads, writes, and trims as
1334 described in :option:`blocksize`.
1336 Example: ``bsrange=1k-4k,2k-8k``.
1338 .. option:: bssplit=str[,str][,str]
1340 Sometimes you want even finer grained control of the block sizes issued, not
1341 just an even split between them. This option allows you to weight various
1342 block sizes, so that you are able to define a specific amount of block sizes
1343 issued. The format for this option is::
1345 bssplit=blocksize/percentage:blocksize/percentage
1347 for as many block sizes as needed. So if you want to define a workload that
1348 has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would write::
1350 bssplit=4k/10:64k/50:32k/40
1352 Ordering does not matter. If the percentage is left blank, fio will fill in
1353 the remaining values evenly. So a bssplit option like this one::
1355 bssplit=4k/50:1k/:32k/
1357 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always add up
1358 to 100, if bssplit is given a range that adds up to more, it will error out.
1360 Comma-separated values may be specified for reads, writes, and trims as
1361 described in :option:`blocksize`.
1363 If you want a workload that has 50% 2k reads and 50% 4k reads, while having
1364 90% 4k writes and 10% 8k writes, you would specify::
1366 bssplit=2k/50:4k/50,4k/90,8k/10
1368 .. option:: blocksize_unaligned, bs_unaligned
1370 If set, fio will issue I/O units with any size within
1371 :option:`blocksize_range`, not just multiples of the minimum size. This
1372 typically won't work with direct I/O, as that normally requires sector
1375 .. option:: bs_is_seq_rand
1377 If this option is set, fio will use the normal read,write blocksize settings
1378 as sequential,random blocksize settings instead. Any random read or write
1379 will use the WRITE blocksize settings, and any sequential read or write will
1380 use the READ blocksize settings.
1382 .. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1384 Boundary to which fio will align random I/O units. Default:
1385 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1386 I/O, though it usually depends on the hardware block size. This option is
1387 mutually exclusive with using a random map for files, so it will turn off
1388 that option. Comma-separated values may be specified for reads, writes, and
1389 trims as described in :option:`blocksize`.
1395 .. option:: zero_buffers
1397 Initialize buffers with all zeros. Default: fill buffers with random data.
1399 .. option:: refill_buffers
1401 If this option is given, fio will refill the I/O buffers on every
1402 submit. The default is to only fill it at init time and reuse that
1403 data. Only makes sense if zero_buffers isn't specified, naturally. If data
1404 verification is enabled, `refill_buffers` is also automatically enabled.
1406 .. option:: scramble_buffers=bool
1408 If :option:`refill_buffers` is too costly and the target is using data
1409 deduplication, then setting this option will slightly modify the I/O buffer
1410 contents to defeat normal de-dupe attempts. This is not enough to defeat
1411 more clever block compression attempts, but it will stop naive dedupe of
1412 blocks. Default: true.
1414 .. option:: buffer_compress_percentage=int
1416 If this is set, then fio will attempt to provide I/O buffer content (on
1417 WRITEs) that compresses to the specified level. Fio does this by providing a
1418 mix of random data and a fixed pattern. The fixed pattern is either zeros,
1419 or the pattern specified by :option:`buffer_pattern`. If the pattern option
1420 is used, it might skew the compression ratio slightly. Note that this is per
1421 block size unit, for file/disk wide compression level that matches this
1422 setting, you'll also want to set :option:`refill_buffers`.
1424 .. option:: buffer_compress_chunk=int
1426 See :option:`buffer_compress_percentage`. This setting allows fio to manage
1427 how big the ranges of random data and zeroed data is. Without this set, fio
1428 will provide :option:`buffer_compress_percentage` of blocksize random data,
1429 followed by the remaining zeroed. With this set to some chunk size smaller
1430 than the block size, fio can alternate random and zeroed data throughout the
1433 .. option:: buffer_pattern=str
1435 If set, fio will fill the I/O buffers with this pattern or with the contents
1436 of a file. If not set, the contents of I/O buffers are defined by the other
1437 options related to buffer contents. The setting can be any pattern of bytes,
1438 and can be prefixed with 0x for hex values. It may also be a string, where
1439 the string must then be wrapped with ``""``. Or it may also be a filename,
1440 where the filename must be wrapped with ``''`` in which case the file is
1441 opened and read. Note that not all the file contents will be read if that
1442 would cause the buffers to overflow. So, for example::
1444 buffer_pattern='filename'
1448 buffer_pattern="abcd"
1456 buffer_pattern=0xdeadface
1458 Also you can combine everything together in any order::
1460 buffer_pattern=0xdeadface"abcd"-12'filename'
1462 .. option:: dedupe_percentage=int
1464 If set, fio will generate this percentage of identical buffers when
1465 writing. These buffers will be naturally dedupable. The contents of the
1466 buffers depend on what other buffer compression settings have been set. It's
1467 possible to have the individual buffers either fully compressible, or not at
1468 all. This option only controls the distribution of unique buffers.
1470 .. option:: invalidate=bool
1472 Invalidate the buffer/page cache parts of the files to be used prior to
1473 starting I/O if the platform and file type support it. Defaults to true.
1474 This will be ignored if :option:`pre_read` is also specified for the
1477 .. option:: sync=bool
1479 Use synchronous I/O for buffered writes. For the majority of I/O engines,
1480 this means using O_SYNC. Default: false.
1482 .. option:: iomem=str, mem=str
1484 Fio can use various types of memory as the I/O unit buffer. The allowed
1488 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1492 Use shared memory as the buffers. Allocated through
1493 :manpage:`shmget(2)`.
1496 Same as shm, but use huge pages as backing.
1499 Use :manpage:`mmap(2)` to allocate buffers. May either be anonymous memory, or can
1500 be file backed if a filename is given after the option. The format
1501 is `mem=mmap:/path/to/file`.
1504 Use a memory mapped huge file as the buffer backing. Append filename
1505 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1508 Same as mmap, but use a MMAP_SHARED mapping.
1511 Use GPU memory as the buffers for GPUDirect RDMA benchmark.
1513 The area allocated is a function of the maximum allowed bs size for the job,
1514 multiplied by the I/O depth given. Note that for **shmhuge** and
1515 **mmaphuge** to work, the system must have free huge pages allocated. This
1516 can normally be checked and set by reading/writing
1517 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1518 is 4MiB in size. So to calculate the number of huge pages you need for a
1519 given job file, add up the I/O depth of all jobs (normally one unless
1520 :option:`iodepth` is used) and multiply by the maximum bs set. Then divide
1521 that number by the huge page size. You can see the size of the huge pages in
1522 :file:`/proc/meminfo`. If no huge pages are allocated by having a non-zero
1523 number in `nr_hugepages`, using **mmaphuge** or **shmhuge** will fail. Also
1524 see :option:`hugepage-size`.
1526 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1527 should point there. So if it's mounted in :file:`/huge`, you would use
1528 `mem=mmaphuge:/huge/somefile`.
1530 .. option:: iomem_align=int
1532 This indicates the memory alignment of the I/O memory buffers. Note that
1533 the given alignment is applied to the first I/O unit buffer, if using
1534 :option:`iodepth` the alignment of the following buffers are given by the
1535 :option:`bs` used. In other words, if using a :option:`bs` that is a
1536 multiple of the page sized in the system, all buffers will be aligned to
1537 this value. If using a :option:`bs` that is not page aligned, the alignment
1538 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1541 .. option:: hugepage-size=int
1543 Defines the size of a huge page. Must at least be equal to the system
1544 setting, see :file:`/proc/meminfo`. Defaults to 4MiB. Should probably
1545 always be a multiple of megabytes, so using ``hugepage-size=Xm`` is the
1546 preferred way to set this to avoid setting a non-pow-2 bad value.
1548 .. option:: lockmem=int
1550 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1551 simulate a smaller amount of memory. The amount specified is per worker.
1557 .. option:: size=int
1559 The total size of file I/O for each thread of this job. Fio will run until
1560 this many bytes has been transferred, unless runtime is limited by other options
1561 (such as :option:`runtime`, for instance, or increased/decreased by :option:`io_size`).
1562 Fio will divide this size between the available files determined by options
1563 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1564 specified by the job. If the result of division happens to be 0, the size is
1565 set to the physical size of the given files or devices if they exist.
1566 If this option is not specified, fio will use the full size of the given
1567 files or devices. If the files do not exist, size must be given. It is also
1568 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1569 given, fio will use 20% of the full size of the given files or devices.
1570 Can be combined with :option:`offset` to constrain the start and end range
1571 that I/O will be done within.
1573 .. option:: io_size=int, io_limit=int
1575 Normally fio operates within the region set by :option:`size`, which means
1576 that the :option:`size` option sets both the region and size of I/O to be
1577 performed. Sometimes that is not what you want. With this option, it is
1578 possible to define just the amount of I/O that fio should do. For instance,
1579 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1580 will perform I/O within the first 20GiB but exit when 5GiB have been
1581 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1582 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1583 the 0..20GiB region.
1585 .. option:: filesize=irange(int)
1587 Individual file sizes. May be a range, in which case fio will select sizes
1588 for files at random within the given range and limited to :option:`size` in
1589 total (if that is given). If not given, each created file is the same size.
1590 This option overrides :option:`size` in terms of file size, which means
1591 this value is used as a fixed size or possible range of each file.
1593 .. option:: file_append=bool
1595 Perform I/O after the end of the file. Normally fio will operate within the
1596 size of a file. If this option is set, then fio will append to the file
1597 instead. This has identical behavior to setting :option:`offset` to the size
1598 of a file. This option is ignored on non-regular files.
1600 .. option:: fill_device=bool, fill_fs=bool
1602 Sets size to something really large and waits for ENOSPC (no space left on
1603 device) as the terminating condition. Only makes sense with sequential
1604 write. For a read workload, the mount point will be filled first then I/O
1605 started on the result. This option doesn't make sense if operating on a raw
1606 device node, since the size of that is already known by the file system.
1607 Additionally, writing beyond end-of-device will not return ENOSPC there.
1613 .. option:: ioengine=str
1615 Defines how the job issues I/O to the file. The following types are defined:
1618 Basic :manpage:`read(2)` or :manpage:`write(2)`
1619 I/O. :manpage:`lseek(2)` is used to position the I/O location.
1620 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
1623 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1624 all supported operating systems except for Windows.
1627 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
1628 queuing by coalescing adjacent I/Os into a single submission.
1631 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1634 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1637 Linux native asynchronous I/O. Note that Linux may only support
1638 queued behavior with non-buffered I/O (set ``direct=1`` or
1640 This engine defines engine specific options.
1643 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
1644 :manpage:`aio_write(3)`.
1647 Solaris native asynchronous I/O.
1650 Windows native asynchronous I/O. Default on Windows.
1653 File is memory mapped with :manpage:`mmap(2)` and data copied
1654 to/from using :manpage:`memcpy(3)`.
1657 :manpage:`splice(2)` is used to transfer the data and
1658 :manpage:`vmsplice(2)` to transfer data from user space to the
1662 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
1663 ioctl, or if the target is an sg character device we use
1664 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
1665 I/O. Requires filename option to specify either block or character
1669 Doesn't transfer any data, just pretends to. This is mainly used to
1670 exercise fio itself and for debugging/testing purposes.
1673 Transfer over the network to given ``host:port``. Depending on the
1674 :option:`protocol` used, the :option:`hostname`, :option:`port`,
1675 :option:`listen` and :option:`filename` options are used to specify
1676 what sort of connection to make, while the :option:`protocol` option
1677 determines which protocol will be used. This engine defines engine
1681 Like **net**, but uses :manpage:`splice(2)` and
1682 :manpage:`vmsplice(2)` to map data and send/receive.
1683 This engine defines engine specific options.
1686 Doesn't transfer any data, but burns CPU cycles according to the
1687 :option:`cpuload` and :option:`cpuchunks` options. Setting
1688 :option:`cpuload`\=85 will cause that job to do nothing but burn 85%
1689 of the CPU. In case of SMP machines, use :option:`numjobs`
1690 =<no_of_cpu> to get desired CPU usage, as the cpuload only loads a
1691 single CPU at the desired rate. A job never finishes unless there is
1692 at least one non-cpuio job.
1695 The GUASI I/O engine is the Generic Userspace Asyncronous Syscall
1696 Interface approach to async I/O. See
1698 http://www.xmailserver.org/guasi-lib.html
1700 for more info on GUASI.
1703 The RDMA I/O engine supports both RDMA memory semantics
1704 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
1705 InfiniBand, RoCE and iWARP protocols.
1708 I/O engine that does regular fallocate to simulate data transfer as
1712 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
1715 does fallocate(,mode = 0).
1718 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
1721 I/O engine that sends :manpage:`ftruncate(2)` operations in response
1722 to write (DDIR_WRITE) events. Each ftruncate issued sets the file's
1723 size to the current block offset. Block size is ignored.
1726 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
1727 defragment activity in request to DDIR_WRITE event.
1730 I/O engine supporting direct access to Ceph Rados Block Devices
1731 (RBD) via librbd without the need to use the kernel rbd driver. This
1732 ioengine defines engine specific options.
1735 Using GlusterFS libgfapi sync interface to direct access to
1736 GlusterFS volumes without having to go through FUSE. This ioengine
1737 defines engine specific options.
1740 Using GlusterFS libgfapi async interface to direct access to
1741 GlusterFS volumes without having to go through FUSE. This ioengine
1742 defines engine specific options.
1745 Read and write through Hadoop (HDFS). The :file:`filename` option
1746 is used to specify host,port of the hdfs name-node to connect. This
1747 engine interprets offsets a little differently. In HDFS, files once
1748 created cannot be modified so random writes are not possible. To
1749 imitate this the libhdfs engine expects a bunch of small files to be
1750 created over HDFS and will randomly pick a file from them
1751 based on the offset generated by fio backend (see the example
1752 job file to create such files, use ``rw=write`` option). Please
1753 note, it may be necessary to set environment variables to work
1754 with HDFS/libhdfs properly. Each job uses its own connection to
1758 Read, write and erase an MTD character device (e.g.,
1759 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
1760 underlying device type, the I/O may have to go in a certain pattern,
1761 e.g., on NAND, writing sequentially to erase blocks and discarding
1762 before overwriting. The `trimwrite` mode works well for this
1766 Read and write using filesystem DAX to a file on a filesystem
1767 mounted with DAX on a persistent memory device through the NVML
1771 Read and write using device DAX to a persistent memory device (e.g.,
1772 /dev/dax0.0) through the NVML libpmem library.
1775 Prefix to specify loading an external I/O engine object file. Append
1776 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
1777 ioengine :file:`foo.o` in :file:`/tmp`.
1780 I/O engine specific parameters
1781 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1783 In addition, there are some parameters which are only valid when a specific
1784 ioengine is in use. These are used identically to normal parameters, with the
1785 caveat that when used on the command line, they must come after the
1786 :option:`ioengine` that defines them is selected.
1788 .. option:: userspace_reap : [libaio]
1790 Normally, with the libaio engine in use, fio will use the
1791 :manpage:`io_getevents(2)` system call to reap newly returned events. With
1792 this flag turned on, the AIO ring will be read directly from user-space to
1793 reap events. The reaping mode is only enabled when polling for a minimum of
1794 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
1796 .. option:: hipri : [pvsync2]
1798 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
1801 .. option:: cpuload=int : [cpuio]
1803 Attempt to use the specified percentage of CPU cycles. This is a mandatory
1804 option when using cpuio I/O engine.
1806 .. option:: cpuchunks=int : [cpuio]
1808 Split the load into cycles of the given time. In microseconds.
1810 .. option:: exit_on_io_done=bool : [cpuio]
1812 Detect when I/O threads are done, then exit.
1814 .. option:: hostname=str : [netsplice] [net]
1816 The hostname or IP address to use for TCP or UDP based I/O. If the job is
1817 a TCP listener or UDP reader, the hostname is not used and must be omitted
1818 unless it is a valid UDP multicast address.
1820 .. option:: namenode=str : [libhdfs]
1822 The hostname or IP address of a HDFS cluster namenode to contact.
1824 .. option:: port=int
1828 The TCP or UDP port to bind to or connect to. If this is used with
1829 :option:`numjobs` to spawn multiple instances of the same job type, then
1830 this will be the starting port number since fio will use a range of
1835 the listening port of the HFDS cluster namenode.
1837 .. option:: interface=str : [netsplice] [net]
1839 The IP address of the network interface used to send or receive UDP
1842 .. option:: ttl=int : [netsplice] [net]
1844 Time-to-live value for outgoing UDP multicast packets. Default: 1.
1846 .. option:: nodelay=bool : [netsplice] [net]
1848 Set TCP_NODELAY on TCP connections.
1850 .. option:: protocol=str : [netsplice] [net]
1852 .. option:: proto=str : [netsplice] [net]
1854 The network protocol to use. Accepted values are:
1857 Transmission control protocol.
1859 Transmission control protocol V6.
1861 User datagram protocol.
1863 User datagram protocol V6.
1867 When the protocol is TCP or UDP, the port must also be given, as well as the
1868 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
1869 normal filename option should be used and the port is invalid.
1871 .. option:: listen : [net]
1873 For TCP network connections, tell fio to listen for incoming connections
1874 rather than initiating an outgoing connection. The :option:`hostname` must
1875 be omitted if this option is used.
1877 .. option:: pingpong : [net]
1879 Normally a network writer will just continue writing data, and a network
1880 reader will just consume packages. If ``pingpong=1`` is set, a writer will
1881 send its normal payload to the reader, then wait for the reader to send the
1882 same payload back. This allows fio to measure network latencies. The
1883 submission and completion latencies then measure local time spent sending or
1884 receiving, and the completion latency measures how long it took for the
1885 other end to receive and send back. For UDP multicast traffic
1886 ``pingpong=1`` should only be set for a single reader when multiple readers
1887 are listening to the same address.
1889 .. option:: window_size : [net]
1891 Set the desired socket buffer size for the connection.
1893 .. option:: mss : [net]
1895 Set the TCP maximum segment size (TCP_MAXSEG).
1897 .. option:: donorname=str : [e4defrag]
1899 File will be used as a block donor (swap extents between files).
1901 .. option:: inplace=int : [e4defrag]
1903 Configure donor file blocks allocation strategy:
1906 Default. Preallocate donor's file on init.
1908 Allocate space immediately inside defragment event, and free right
1911 .. option:: clustername=str : [rbd]
1913 Specifies the name of the Ceph cluster.
1915 .. option:: rbdname=str : [rbd]
1917 Specifies the name of the RBD.
1919 .. option:: pool=str : [rbd]
1921 Specifies the name of the Ceph pool containing RBD.
1923 .. option:: clientname=str : [rbd]
1925 Specifies the username (without the 'client.' prefix) used to access the
1926 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
1927 the full *type.id* string. If no type. prefix is given, fio will add
1928 'client.' by default.
1930 .. option:: skip_bad=bool : [mtd]
1932 Skip operations against known bad blocks.
1934 .. option:: hdfsdirectory : [libhdfs]
1936 libhdfs will create chunk in this HDFS directory.
1938 .. option:: chunk_size : [libhdfs]
1940 the size of the chunk to use for each file.
1946 .. option:: iodepth=int
1948 Number of I/O units to keep in flight against the file. Note that
1949 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
1950 for small degrees when :option:`verify_async` is in use). Even async
1951 engines may impose OS restrictions causing the desired depth not to be
1952 achieved. This may happen on Linux when using libaio and not setting
1953 :option:`direct`\=1, since buffered I/O is not async on that OS. Keep an
1954 eye on the I/O depth distribution in the fio output to verify that the
1955 achieved depth is as expected. Default: 1.
1957 .. option:: iodepth_batch_submit=int, iodepth_batch=int
1959 This defines how many pieces of I/O to submit at once. It defaults to 1
1960 which means that we submit each I/O as soon as it is available, but can be
1961 raised to submit bigger batches of I/O at the time. If it is set to 0 the
1962 :option:`iodepth` value will be used.
1964 .. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
1966 This defines how many pieces of I/O to retrieve at once. It defaults to 1
1967 which means that we'll ask for a minimum of 1 I/O in the retrieval process
1968 from the kernel. The I/O retrieval will go on until we hit the limit set by
1969 :option:`iodepth_low`. If this variable is set to 0, then fio will always
1970 check for completed events before queuing more I/O. This helps reduce I/O
1971 latency, at the cost of more retrieval system calls.
1973 .. option:: iodepth_batch_complete_max=int
1975 This defines maximum pieces of I/O to retrieve at once. This variable should
1976 be used along with :option:`iodepth_batch_complete_min`\=int variable,
1977 specifying the range of min and max amount of I/O which should be
1978 retrieved. By default it is equal to the :option:`iodepth_batch_complete_min`
1983 iodepth_batch_complete_min=1
1984 iodepth_batch_complete_max=<iodepth>
1986 which means that we will retrieve at least 1 I/O and up to the whole
1987 submitted queue depth. If none of I/O has been completed yet, we will wait.
1991 iodepth_batch_complete_min=0
1992 iodepth_batch_complete_max=<iodepth>
1994 which means that we can retrieve up to the whole submitted queue depth, but
1995 if none of I/O has been completed yet, we will NOT wait and immediately exit
1996 the system call. In this example we simply do polling.
1998 .. option:: iodepth_low=int
2000 The low water mark indicating when to start filling the queue
2001 again. Defaults to the same as :option:`iodepth`, meaning that fio will
2002 attempt to keep the queue full at all times. If :option:`iodepth` is set to
2003 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
2004 16 requests, it will let the depth drain down to 4 before starting to fill
2007 .. option:: io_submit_mode=str
2009 This option controls how fio submits the I/O to the I/O engine. The default
2010 is `inline`, which means that the fio job threads submit and reap I/O
2011 directly. If set to `offload`, the job threads will offload I/O submission
2012 to a dedicated pool of I/O threads. This requires some coordination and thus
2013 has a bit of extra overhead, especially for lower queue depth I/O where it
2014 can increase latencies. The benefit is that fio can manage submission rates
2015 independently of the device completion rates. This avoids skewed latency
2016 reporting if I/O gets backed up on the device side (the coordinated omission
2023 .. option:: thinktime=time
2025 Stall the job for the specified period of time after an I/O has completed before issuing the
2026 next. May be used to simulate processing being done by an application.
2027 When the unit is omitted, the value is interpreted in microseconds. See
2028 :option:`thinktime_blocks` and :option:`thinktime_spin`.
2030 .. option:: thinktime_spin=time
2032 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
2033 something with the data received, before falling back to sleeping for the
2034 rest of the period specified by :option:`thinktime`. When the unit is
2035 omitted, the value is interpreted in microseconds.
2037 .. option:: thinktime_blocks=int
2039 Only valid if :option:`thinktime` is set - control how many blocks to issue,
2040 before waiting `thinktime` usecs. If not set, defaults to 1 which will make
2041 fio wait `thinktime` usecs after every block. This effectively makes any
2042 queue depth setting redundant, since no more than 1 I/O will be queued
2043 before we have to complete it and do our thinktime. In other words, this
2044 setting effectively caps the queue depth if the latter is larger.
2046 .. option:: rate=int[,int][,int]
2048 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
2049 suffix rules apply. Comma-separated values may be specified for reads,
2050 writes, and trims as described in :option:`blocksize`.
2052 For example, using `rate=1m,500k` would limit reads to 1MiB/sec and writes to
2053 500KiB/sec. Capping only reads or writes can be done with `rate=,500k` or
2054 `rate=500k,` where the former will only limit writes (to 500KiB/sec) and the
2055 latter will only limit reads.
2057 .. option:: rate_min=int[,int][,int]
2059 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
2060 to meet this requirement will cause the job to exit. Comma-separated values
2061 may be specified for reads, writes, and trims as described in
2062 :option:`blocksize`.
2064 .. option:: rate_iops=int[,int][,int]
2066 Cap the bandwidth to this number of IOPS. Basically the same as
2067 :option:`rate`, just specified independently of bandwidth. If the job is
2068 given a block size range instead of a fixed value, the smallest block size
2069 is used as the metric. Comma-separated values may be specified for reads,
2070 writes, and trims as described in :option:`blocksize`.
2072 .. option:: rate_iops_min=int[,int][,int]
2074 If fio doesn't meet this rate of I/O, it will cause the job to exit.
2075 Comma-separated values may be specified for reads, writes, and trims as
2076 described in :option:`blocksize`.
2078 .. option:: rate_process=str
2080 This option controls how fio manages rated I/O submissions. The default is
2081 `linear`, which submits I/O in a linear fashion with fixed delays between
2082 I/Os that gets adjusted based on I/O completion rates. If this is set to
2083 `poisson`, fio will submit I/O based on a more real world random request
2084 flow, known as the Poisson process
2085 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
2086 10^6 / IOPS for the given workload.
2092 .. option:: latency_target=time
2094 If set, fio will attempt to find the max performance point that the given
2095 workload will run at while maintaining a latency below this target. When
2096 the unit is omitted, the value is interpreted in microseconds. See
2097 :option:`latency_window` and :option:`latency_percentile`.
2099 .. option:: latency_window=time
2101 Used with :option:`latency_target` to specify the sample window that the job
2102 is run at varying queue depths to test the performance. When the unit is
2103 omitted, the value is interpreted in microseconds.
2105 .. option:: latency_percentile=float
2107 The percentage of I/Os that must fall within the criteria specified by
2108 :option:`latency_target` and :option:`latency_window`. If not set, this
2109 defaults to 100.0, meaning that all I/Os must be equal or below to the value
2110 set by :option:`latency_target`.
2112 .. option:: max_latency=time
2114 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
2115 maximum latency. When the unit is omitted, the value is interpreted in
2118 .. option:: rate_cycle=int
2120 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
2121 of milliseconds. Defaults to 1000.
2127 .. option:: write_iolog=str
2129 Write the issued I/O patterns to the specified file. See
2130 :option:`read_iolog`. Specify a separate file for each job, otherwise the
2131 iologs will be interspersed and the file may be corrupt.
2133 .. option:: read_iolog=str
2135 Open an iolog with the specified filename and replay the I/O patterns it
2136 contains. This can be used to store a workload and replay it sometime
2137 later. The iolog given may also be a blktrace binary file, which allows fio
2138 to replay a workload captured by :command:`blktrace`. See
2139 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
2140 replay, the file needs to be turned into a blkparse binary data file first
2141 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
2143 .. option:: replay_no_stall=int
2145 When replaying I/O with :option:`read_iolog` the default behavior is to
2146 attempt to respect the timestamps within the log and replay them with the
2147 appropriate delay between IOPS. By setting this variable fio will not
2148 respect the timestamps and attempt to replay them as fast as possible while
2149 still respecting ordering. The result is the same I/O pattern to a given
2150 device, but different timings.
2152 .. option:: replay_redirect=str
2154 While replaying I/O patterns using :option:`read_iolog` the default behavior
2155 is to replay the IOPS onto the major/minor device that each IOP was recorded
2156 from. This is sometimes undesirable because on a different machine those
2157 major/minor numbers can map to a different device. Changing hardware on the
2158 same system can also result in a different major/minor mapping.
2159 ``replay_redirect`` causes all I/Os to be replayed onto the single specified
2160 device regardless of the device it was recorded
2161 from. i.e. :option:`replay_redirect`\= :file:`/dev/sdc` would cause all I/O
2162 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
2163 multiple devices will be replayed onto a single device, if the trace
2164 contains multiple devices. If you want multiple devices to be replayed
2165 concurrently to multiple redirected devices you must blkparse your trace
2166 into separate traces and replay them with independent fio invocations.
2167 Unfortunately this also breaks the strict time ordering between multiple
2170 .. option:: replay_align=int
2172 Force alignment of I/O offsets and lengths in a trace to this power of 2
2175 .. option:: replay_scale=int
2177 Scale sector offsets down by this factor when replaying traces.
2180 Threads, processes and job synchronization
2181 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2185 Fio defaults to creating jobs by using fork, however if this option is
2186 given, fio will create jobs by using POSIX Threads' function
2187 :manpage:`pthread_create(3)` to create threads instead.
2189 .. option:: wait_for=str
2191 If set, the current job won't be started until all workers of the specified
2192 waitee job are done.
2194 ``wait_for`` operates on the job name basis, so there are a few
2195 limitations. First, the waitee must be defined prior to the waiter job
2196 (meaning no forward references). Second, if a job is being referenced as a
2197 waitee, it must have a unique name (no duplicate waitees).
2199 .. option:: nice=int
2201 Run the job with the given nice value. See man :manpage:`nice(2)`.
2203 On Windows, values less than -15 set the process class to "High"; -1 through
2204 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
2207 .. option:: prio=int
2209 Set the I/O priority value of this job. Linux limits us to a positive value
2210 between 0 and 7, with 0 being the highest. See man
2211 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
2212 systems since meaning of priority may differ.
2214 .. option:: prioclass=int
2216 Set the I/O priority class. See man :manpage:`ionice(1)`.
2218 .. option:: cpumask=int
2220 Set the CPU affinity of this job. The parameter given is a bit mask of
2221 allowed CPUs the job may run on. So if you want the allowed CPUs to be 1
2222 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
2223 :manpage:`sched_setaffinity(2)`. This may not work on all supported
2224 operating systems or kernel versions. This option doesn't work well for a
2225 higher CPU count than what you can store in an integer mask, so it can only
2226 control cpus 1-32. For boxes with larger CPU counts, use
2227 :option:`cpus_allowed`.
2229 .. option:: cpus_allowed=str
2231 Controls the same options as :option:`cpumask`, but accepts a textual
2232 specification of the permitted CPUs instead. So to use CPUs 1 and 5 you
2233 would specify ``cpus_allowed=1,5``. This option also allows a range of CPUs
2234 to be specified -- say you wanted a binding to CPUs 1, 5, and 8 to 15, you
2235 would set ``cpus_allowed=1,5,8-15``.
2237 .. option:: cpus_allowed_policy=str
2239 Set the policy of how fio distributes the CPUs specified by
2240 :option:`cpus_allowed` or :option:`cpumask`. Two policies are supported:
2243 All jobs will share the CPU set specified.
2245 Each job will get a unique CPU from the CPU set.
2247 **shared** is the default behavior, if the option isn't specified. If
2248 **split** is specified, then fio will will assign one cpu per job. If not
2249 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
2252 .. option:: numa_cpu_nodes=str
2254 Set this job running on specified NUMA nodes' CPUs. The arguments allow
2255 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
2256 NUMA options support, fio must be built on a system with libnuma-dev(el)
2259 .. option:: numa_mem_policy=str
2261 Set this job's memory policy and corresponding NUMA nodes. Format of the
2266 ``mode`` is one of the following memory poicies: ``default``, ``prefer``,
2267 ``bind``, ``interleave`` or ``local``. For ``default`` and ``local`` memory
2268 policies, no node needs to be specified. For ``prefer``, only one node is
2269 allowed. For ``bind`` and ``interleave`` the ``nodelist`` may be as
2270 follows: a comma delimited list of numbers, A-B ranges, or `all`.
2272 .. option:: cgroup=str
2274 Add job to this control group. If it doesn't exist, it will be created. The
2275 system must have a mounted cgroup blkio mount point for this to work. If
2276 your system doesn't have it mounted, you can do so with::
2278 # mount -t cgroup -o blkio none /cgroup
2280 .. option:: cgroup_weight=int
2282 Set the weight of the cgroup to this value. See the documentation that comes
2283 with the kernel, allowed values are in the range of 100..1000.
2285 .. option:: cgroup_nodelete=bool
2287 Normally fio will delete the cgroups it has created after the job
2288 completion. To override this behavior and to leave cgroups around after the
2289 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
2290 to inspect various cgroup files after job completion. Default: false.
2292 .. option:: flow_id=int
2294 The ID of the flow. If not specified, it defaults to being a global
2295 flow. See :option:`flow`.
2297 .. option:: flow=int
2299 Weight in token-based flow control. If this value is used, then there is a
2300 'flow counter' which is used to regulate the proportion of activity between
2301 two or more jobs. Fio attempts to keep this flow counter near zero. The
2302 ``flow`` parameter stands for how much should be added or subtracted to the
2303 flow counter on each iteration of the main I/O loop. That is, if one job has
2304 ``flow=8`` and another job has ``flow=-1``, then there will be a roughly 1:8
2305 ratio in how much one runs vs the other.
2307 .. option:: flow_watermark=int
2309 The maximum value that the absolute value of the flow counter is allowed to
2310 reach before the job must wait for a lower value of the counter.
2312 .. option:: flow_sleep=int
2314 The period of time, in microseconds, to wait after the flow watermark has
2315 been exceeded before retrying operations.
2317 .. option:: stonewall, wait_for_previous
2319 Wait for preceding jobs in the job file to exit, before starting this
2320 one. Can be used to insert serialization points in the job file. A stone
2321 wall also implies starting a new reporting group, see
2322 :option:`group_reporting`.
2326 By default, fio will continue running all other jobs when one job finishes
2327 but sometimes this is not the desired action. Setting ``exitall`` will
2328 instead make fio terminate all other jobs when one job finishes.
2330 .. option:: exec_prerun=str
2332 Before running this job, issue the command specified through
2333 :manpage:`system(3)`. Output is redirected in a file called
2334 :file:`jobname.prerun.txt`.
2336 .. option:: exec_postrun=str
2338 After the job completes, issue the command specified though
2339 :manpage:`system(3)`. Output is redirected in a file called
2340 :file:`jobname.postrun.txt`.
2344 Instead of running as the invoking user, set the user ID to this value
2345 before the thread/process does any work.
2349 Set group ID, see :option:`uid`.
2355 .. option:: verify_only
2357 Do not perform specified workload, only verify data still matches previous
2358 invocation of this workload. This option allows one to check data multiple
2359 times at a later date without overwriting it. This option makes sense only
2360 for workloads that write data, and does not support workloads with the
2361 :option:`time_based` option set.
2363 .. option:: do_verify=bool
2365 Run the verify phase after a write phase. Only valid if :option:`verify` is
2368 .. option:: verify=str
2370 If writing to a file, fio can verify the file contents after each iteration
2371 of the job. Each verification method also implies verification of special
2372 header, which is written to the beginning of each block. This header also
2373 includes meta information, like offset of the block, block number, timestamp
2374 when block was written, etc. :option:`verify` can be combined with
2375 :option:`verify_pattern` option. The allowed values are:
2378 Use an md5 sum of the data area and store it in the header of
2382 Use an experimental crc64 sum of the data area and store it in the
2383 header of each block.
2386 Use a crc32c sum of the data area and store it in the header of
2387 each block. This will automatically use hardware acceleration
2388 (e.g. SSE4.2 on an x86 or CRC crypto extensions on ARM64) but will
2389 fall back to software crc32c if none is found. Generally the
2390 fatest checksum fio supports when hardware accelerated.
2396 Use a crc32 sum of the data area and store it in the header of each
2400 Use a crc16 sum of the data area and store it in the header of each
2404 Use a crc7 sum of the data area and store it in the header of each
2408 Use xxhash as the checksum function. Generally the fastest software
2409 checksum that fio supports.
2412 Use sha512 as the checksum function.
2415 Use sha256 as the checksum function.
2418 Use optimized sha1 as the checksum function.
2421 Use optimized sha3-224 as the checksum function.
2424 Use optimized sha3-256 as the checksum function.
2427 Use optimized sha3-384 as the checksum function.
2430 Use optimized sha3-512 as the checksum function.
2433 This option is deprecated, since now meta information is included in
2434 generic verification header and meta verification happens by
2435 default. For detailed information see the description of the
2436 :option:`verify` setting. This option is kept because of
2437 compatibility's sake with old configurations. Do not use it.
2440 Verify a strict pattern. Normally fio includes a header with some
2441 basic information and checksumming, but if this option is set, only
2442 the specific pattern set with :option:`verify_pattern` is verified.
2445 Only pretend to verify. Useful for testing internals with
2446 :option:`ioengine`\=null, not for much else.
2448 This option can be used for repeated burn-in tests of a system to make sure
2449 that the written data is also correctly read back. If the data direction
2450 given is a read or random read, fio will assume that it should verify a
2451 previously written file. If the data direction includes any form of write,
2452 the verify will be of the newly written data.
2454 .. option:: verifysort=bool
2456 If true, fio will sort written verify blocks when it deems it faster to read
2457 them back in a sorted manner. This is often the case when overwriting an
2458 existing file, since the blocks are already laid out in the file system. You
2459 can ignore this option unless doing huge amounts of really fast I/O where
2460 the red-black tree sorting CPU time becomes significant. Default: true.
2462 .. option:: verifysort_nr=int
2464 Pre-load and sort verify blocks for a read workload.
2466 .. option:: verify_offset=int
2468 Swap the verification header with data somewhere else in the block before
2469 writing. It is swapped back before verifying.
2471 .. option:: verify_interval=int
2473 Write the verification header at a finer granularity than the
2474 :option:`blocksize`. It will be written for chunks the size of
2475 ``verify_interval``. :option:`blocksize` should divide this evenly.
2477 .. option:: verify_pattern=str
2479 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
2480 filling with totally random bytes, but sometimes it's interesting to fill
2481 with a known pattern for I/O verification purposes. Depending on the width
2482 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time (it can
2483 be either a decimal or a hex number). The ``verify_pattern`` if larger than
2484 a 32-bit quantity has to be a hex number that starts with either "0x" or
2485 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
2486 format, which means that for each block offset will be written and then
2487 verified back, e.g.::
2491 Or use combination of everything::
2493 verify_pattern=0xff%o"abcd"-12
2495 .. option:: verify_fatal=bool
2497 Normally fio will keep checking the entire contents before quitting on a
2498 block verification failure. If this option is set, fio will exit the job on
2499 the first observed failure. Default: false.
2501 .. option:: verify_dump=bool
2503 If set, dump the contents of both the original data block and the data block
2504 we read off disk to files. This allows later analysis to inspect just what
2505 kind of data corruption occurred. Off by default.
2507 .. option:: verify_async=int
2509 Fio will normally verify I/O inline from the submitting thread. This option
2510 takes an integer describing how many async offload threads to create for I/O
2511 verification instead, causing fio to offload the duty of verifying I/O
2512 contents to one or more separate threads. If using this offload option, even
2513 sync I/O engines can benefit from using an :option:`iodepth` setting higher
2514 than 1, as it allows them to have I/O in flight while verifies are running.
2515 Defaults to 0 async threads, i.e. verification is not asynchronous.
2517 .. option:: verify_async_cpus=str
2519 Tell fio to set the given CPU affinity on the async I/O verification
2520 threads. See :option:`cpus_allowed` for the format used.
2522 .. option:: verify_backlog=int
2524 Fio will normally verify the written contents of a job that utilizes verify
2525 once that job has completed. In other words, everything is written then
2526 everything is read back and verified. You may want to verify continually
2527 instead for a variety of reasons. Fio stores the meta data associated with
2528 an I/O block in memory, so for large verify workloads, quite a bit of memory
2529 would be used up holding this meta data. If this option is enabled, fio will
2530 write only N blocks before verifying these blocks.
2532 .. option:: verify_backlog_batch=int
2534 Control how many blocks fio will verify if :option:`verify_backlog` is
2535 set. If not set, will default to the value of :option:`verify_backlog`
2536 (meaning the entire queue is read back and verified). If
2537 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
2538 blocks will be verified, if ``verify_backlog_batch`` is larger than
2539 :option:`verify_backlog`, some blocks will be verified more than once.
2541 .. option:: verify_state_save=bool
2543 When a job exits during the write phase of a verify workload, save its
2544 current state. This allows fio to replay up until that point, if the verify
2545 state is loaded for the verify read phase. The format of the filename is,
2548 <type>-<jobname>-<jobindex>-verify.state.
2550 <type> is "local" for a local run, "sock" for a client/server socket
2551 connection, and "ip" (192.168.0.1, for instance) for a networked
2552 client/server connection. Defaults to true.
2554 .. option:: verify_state_load=bool
2556 If a verify termination trigger was used, fio stores the current write state
2557 of each thread. This can be used at verification time so that fio knows how
2558 far it should verify. Without this information, fio will run a full
2559 verification pass, according to the settings in the job file used. Default
2562 .. option:: trim_percentage=int
2564 Number of verify blocks to discard/trim.
2566 .. option:: trim_verify_zero=bool
2568 Verify that trim/discarded blocks are returned as zeros.
2570 .. option:: trim_backlog=int
2572 Verify that trim/discarded blocks are returned as zeros.
2574 .. option:: trim_backlog_batch=int
2576 Trim this number of I/O blocks.
2578 .. option:: experimental_verify=bool
2580 Enable experimental verification.
2586 .. option:: steadystate=str:float, ss=str:float
2588 Define the criterion and limit for assessing steady state performance. The
2589 first parameter designates the criterion whereas the second parameter sets
2590 the threshold. When the criterion falls below the threshold for the
2591 specified duration, the job will stop. For example, `iops_slope:0.1%` will
2592 direct fio to terminate the job when the least squares regression slope
2593 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
2594 this will apply to all jobs in the group. Below is the list of available
2595 steady state assessment criteria. All assessments are carried out using only
2596 data from the rolling collection window. Threshold limits can be expressed
2597 as a fixed value or as a percentage of the mean in the collection window.
2600 Collect IOPS data. Stop the job if all individual IOPS measurements
2601 are within the specified limit of the mean IOPS (e.g., ``iops:2``
2602 means that all individual IOPS values must be within 2 of the mean,
2603 whereas ``iops:0.2%`` means that all individual IOPS values must be
2604 within 0.2% of the mean IOPS to terminate the job).
2607 Collect IOPS data and calculate the least squares regression
2608 slope. Stop the job if the slope falls below the specified limit.
2611 Collect bandwidth data. Stop the job if all individual bandwidth
2612 measurements are within the specified limit of the mean bandwidth.
2615 Collect bandwidth data and calculate the least squares regression
2616 slope. Stop the job if the slope falls below the specified limit.
2618 .. option:: steadystate_duration=time, ss_dur=time
2620 A rolling window of this duration will be used to judge whether steady state
2621 has been reached. Data will be collected once per second. The default is 0
2622 which disables steady state detection. When the unit is omitted, the
2623 value is interpreted in seconds.
2625 .. option:: steadystate_ramp_time=time, ss_ramp=time
2627 Allow the job to run for the specified duration before beginning data
2628 collection for checking the steady state job termination criterion. The
2629 default is 0. When the unit is omitted, the value is interpreted in seconds.
2632 Measurements and reporting
2633 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2635 .. option:: per_job_logs=bool
2637 If set, this generates bw/clat/iops log with per file private filenames. If
2638 not set, jobs with identical names will share the log filename. Default:
2641 .. option:: group_reporting
2643 It may sometimes be interesting to display statistics for groups of jobs as
2644 a whole instead of for each individual job. This is especially true if
2645 :option:`numjobs` is used; looking at individual thread/process output
2646 quickly becomes unwieldy. To see the final report per-group instead of
2647 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
2648 same reporting group, unless if separated by a :option:`stonewall`, or by
2649 using :option:`new_group`.
2651 .. option:: new_group
2653 Start a new reporting group. See: :option:`group_reporting`. If not given,
2654 all jobs in a file will be part of the same reporting group, unless
2655 separated by a :option:`stonewall`.
2659 By default, fio collects and shows final output results for all jobs
2660 that run. If this option is set to 0, then fio will ignore it in
2661 the final stat output.
2663 .. option:: write_bw_log=str
2665 If given, write a bandwidth log for this job. Can be used to store data of
2666 the bandwidth of the jobs in their lifetime. The included
2667 :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
2668 text files into nice graphs. See :option:`write_lat_log` for behavior of
2669 given filename. For this option, the postfix is :file:`_bw.x.log`, where `x`
2670 is the index of the job (`1..N`, where `N` is the number of jobs). If
2671 :option:`per_job_logs` is false, then the filename will not include the job
2672 index. See `Log File Formats`_.
2674 .. option:: write_lat_log=str
2676 Same as :option:`write_bw_log`, except that this option stores I/O
2677 submission, completion, and total latencies instead. If no filename is given
2678 with this option, the default filename of :file:`jobname_type.log` is
2679 used. Even if the filename is given, fio will still append the type of
2680 log. So if one specifies::
2684 The actual log names will be :file:`foo_slat.x.log`, :file:`foo_clat.x.log`,
2685 and :file:`foo_lat.x.log`, where `x` is the index of the job (1..N, where N
2686 is the number of jobs). This helps :command:`fio_generate_plot` find the
2687 logs automatically. If :option:`per_job_logs` is false, then the filename
2688 will not include the job index. See `Log File Formats`_.
2690 .. option:: write_hist_log=str
2692 Same as :option:`write_lat_log`, but writes I/O completion latency
2693 histograms. If no filename is given with this option, the default filename
2694 of :file:`jobname_clat_hist.x.log` is used, where `x` is the index of the
2695 job (1..N, where `N` is the number of jobs). Even if the filename is given,
2696 fio will still append the type of log. If :option:`per_job_logs` is false,
2697 then the filename will not include the job index. See `Log File Formats`_.
2699 .. option:: write_iops_log=str
2701 Same as :option:`write_bw_log`, but writes IOPS. If no filename is given
2702 with this option, the default filename of :file:`jobname_type.x.log` is
2703 used,where `x` is the index of the job (1..N, where `N` is the number of
2704 jobs). Even if the filename is given, fio will still append the type of
2705 log. If :option:`per_job_logs` is false, then the filename will not include
2706 the job index. See `Log File Formats`_.
2708 .. option:: log_avg_msec=int
2710 By default, fio will log an entry in the iops, latency, or bw log for every
2711 I/O that completes. When writing to the disk log, that can quickly grow to a
2712 very large size. Setting this option makes fio average the each log entry
2713 over the specified period of time, reducing the resolution of the log. See
2714 :option:`log_max_value` as well. Defaults to 0, logging all entries.
2715 Also see `Log File Formats`_.
2717 .. option:: log_hist_msec=int
2719 Same as :option:`log_avg_msec`, but logs entries for completion latency
2720 histograms. Computing latency percentiles from averages of intervals using
2721 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
2722 histogram entries over the specified period of time, reducing log sizes for
2723 high IOPS devices while retaining percentile accuracy. See
2724 :option:`log_hist_coarseness` as well. Defaults to 0, meaning histogram
2725 logging is disabled.
2727 .. option:: log_hist_coarseness=int
2729 Integer ranging from 0 to 6, defining the coarseness of the resolution of
2730 the histogram logs enabled with :option:`log_hist_msec`. For each increment
2731 in coarseness, fio outputs half as many bins. Defaults to 0, for which
2732 histogram logs contain 1216 latency bins. See `Log File Formats`_.
2734 .. option:: log_max_value=bool
2736 If :option:`log_avg_msec` is set, fio logs the average over that window. If
2737 you instead want to log the maximum value, set this option to 1. Defaults to
2738 0, meaning that averaged values are logged.
2740 .. option:: log_offset=int
2742 If this is set, the iolog options will include the byte offset for the I/O
2743 entry as well as the other data values.
2745 .. option:: log_compression=int
2747 If this is set, fio will compress the I/O logs as it goes, to keep the
2748 memory footprint lower. When a log reaches the specified size, that chunk is
2749 removed and compressed in the background. Given that I/O logs are fairly
2750 highly compressible, this yields a nice memory savings for longer runs. The
2751 downside is that the compression will consume some background CPU cycles, so
2752 it may impact the run. This, however, is also true if the logging ends up
2753 consuming most of the system memory. So pick your poison. The I/O logs are
2754 saved normally at the end of a run, by decompressing the chunks and storing
2755 them in the specified log file. This feature depends on the availability of
2758 .. option:: log_compression_cpus=str
2760 Define the set of CPUs that are allowed to handle online log compression for
2761 the I/O jobs. This can provide better isolation between performance
2762 sensitive jobs, and background compression work.
2764 .. option:: log_store_compressed=bool
2766 If set, fio will store the log files in a compressed format. They can be
2767 decompressed with fio, using the :option:`--inflate-log` command line
2768 parameter. The files will be stored with a :file:`.fz` suffix.
2770 .. option:: log_unix_epoch=bool
2772 If set, fio will log Unix timestamps to the log files produced by enabling
2773 write_type_log for each log type, instead of the default zero-based
2776 .. option:: block_error_percentiles=bool
2778 If set, record errors in trim block-sized units from writes and trims and
2779 output a histogram of how many trims it took to get to errors, and what kind
2780 of error was encountered.
2782 .. option:: bwavgtime=int
2784 Average the calculated bandwidth over the given time. Value is specified in
2785 milliseconds. If the job also does bandwidth logging through
2786 :option:`write_bw_log`, then the minimum of this option and
2787 :option:`log_avg_msec` will be used. Default: 500ms.
2789 .. option:: iopsavgtime=int
2791 Average the calculated IOPS over the given time. Value is specified in
2792 milliseconds. If the job also does IOPS logging through
2793 :option:`write_iops_log`, then the minimum of this option and
2794 :option:`log_avg_msec` will be used. Default: 500ms.
2796 .. option:: disk_util=bool
2798 Generate disk utilization statistics, if the platform supports it.
2801 .. option:: disable_lat=bool
2803 Disable measurements of total latency numbers. Useful only for cutting back
2804 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
2805 performance at really high IOPS rates. Note that to really get rid of a
2806 large amount of these calls, this option must be used with
2807 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
2809 .. option:: disable_clat=bool
2811 Disable measurements of completion latency numbers. See
2812 :option:`disable_lat`.
2814 .. option:: disable_slat=bool
2816 Disable measurements of submission latency numbers. See
2817 :option:`disable_slat`.
2819 .. option:: disable_bw_measurement=bool, disable_bw=bool
2821 Disable measurements of throughput/bandwidth numbers. See
2822 :option:`disable_lat`.
2824 .. option:: clat_percentiles=bool
2826 Enable the reporting of percentiles of completion latencies.
2828 .. option:: percentile_list=float_list
2830 Overwrite the default list of percentiles for completion latencies and the
2831 block error histogram. Each number is a floating number in the range
2832 (0,100], and the maximum length of the list is 20. Use ``:`` to separate the
2833 numbers, and list the numbers in ascending order. For example,
2834 ``--percentile_list=99.5:99.9`` will cause fio to report the values of
2835 completion latency below which 99.5% and 99.9% of the observed latencies
2842 .. option:: exitall_on_error
2844 When one job finishes in error, terminate the rest. The default is to wait
2845 for each job to finish.
2847 .. option:: continue_on_error=str
2849 Normally fio will exit the job on the first observed failure. If this option
2850 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
2851 EILSEQ) until the runtime is exceeded or the I/O size specified is
2852 completed. If this option is used, there are two more stats that are
2853 appended, the total error count and the first error. The error field given
2854 in the stats is the first error that was hit during the run.
2856 The allowed values are:
2859 Exit on any I/O or verify errors.
2862 Continue on read errors, exit on all others.
2865 Continue on write errors, exit on all others.
2868 Continue on any I/O error, exit on all others.
2871 Continue on verify errors, exit on all others.
2874 Continue on all errors.
2877 Backward-compatible alias for 'none'.
2880 Backward-compatible alias for 'all'.
2882 .. option:: ignore_error=str
2884 Sometimes you want to ignore some errors during test in that case you can
2885 specify error list for each error type, instead of only being able to
2886 ignore the default 'non-fatal error' using :option:`continue_on_error`.
2887 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
2888 given error type is separated with ':'. Error may be symbol ('ENOSPC',
2889 'ENOMEM') or integer. Example::
2891 ignore_error=EAGAIN,ENOSPC:122
2893 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
2894 WRITE. This option works by overriding :option:`continue_on_error` with
2895 the list of errors for each error type if any.
2897 .. option:: error_dump=bool
2899 If set dump every error even if it is non fatal, true by default. If
2900 disabled only fatal error will be dumped.
2902 Running predefined workloads
2903 ----------------------------
2905 Fio includes predefined profiles that mimic the I/O workloads generated by
2908 .. option:: profile=str
2910 The predefined workload to run. Current profiles are:
2913 Threaded I/O bench (tiotest/tiobench) like workload.
2916 Aerospike Certification Tool (ACT) like workload.
2918 To view a profile's additional options use :option:`--cmdhelp` after specifying
2919 the profile. For example::
2921 $ fio --profile=act --cmdhelp
2926 .. option:: device-names=str
2931 .. option:: load=int
2934 ACT load multiplier. Default: 1.
2936 .. option:: test-duration=time
2939 How long the entire test takes to run. When the unit is omitted, the value
2940 is given in seconds. Default: 24h.
2942 .. option:: threads-per-queue=int
2945 Number of read IO threads per device. Default: 8.
2947 .. option:: read-req-num-512-blocks=int
2950 Number of 512B blocks to read at the time. Default: 3.
2952 .. option:: large-block-op-kbytes=int
2955 Size of large block ops in KiB (writes). Default: 131072.
2960 Set to run ACT prep phase.
2962 Tiobench profile options
2963 ~~~~~~~~~~~~~~~~~~~~~~~~
2965 .. option:: size=str
2970 .. option:: block=int
2973 Block size in bytes. Default: 4096.
2975 .. option:: numruns=int
2985 .. option:: threads=int
2990 Interpreting the output
2991 -----------------------
2994 Example output was based on the following:
2995 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --time_based \
2996 --rate=1256k --bs=14K --name=quick --runtime=1s --name=mixed \
2997 --runtime=2m --rw=rw
2999 Fio spits out a lot of output. While running, fio will display the status of the
3000 jobs created. An example of that would be::
3002 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]
3004 The characters inside the first set of square brackets denote the current status of
3005 each thread. The first character is the first job defined in the job file, and so
3006 forth. The possible values (in typical life cycle order) are:
3008 +------+-----+-----------------------------------------------------------+
3010 +======+=====+===========================================================+
3011 | P | | Thread setup, but not started. |
3012 +------+-----+-----------------------------------------------------------+
3013 | C | | Thread created. |
3014 +------+-----+-----------------------------------------------------------+
3015 | I | | Thread initialized, waiting or generating necessary data. |
3016 +------+-----+-----------------------------------------------------------+
3017 | | p | Thread running pre-reading file(s). |
3018 +------+-----+-----------------------------------------------------------+
3019 | | / | Thread is in ramp period. |
3020 +------+-----+-----------------------------------------------------------+
3021 | | R | Running, doing sequential reads. |
3022 +------+-----+-----------------------------------------------------------+
3023 | | r | Running, doing random reads. |
3024 +------+-----+-----------------------------------------------------------+
3025 | | W | Running, doing sequential writes. |
3026 +------+-----+-----------------------------------------------------------+
3027 | | w | Running, doing random writes. |
3028 +------+-----+-----------------------------------------------------------+
3029 | | M | Running, doing mixed sequential reads/writes. |
3030 +------+-----+-----------------------------------------------------------+
3031 | | m | Running, doing mixed random reads/writes. |
3032 +------+-----+-----------------------------------------------------------+
3033 | | D | Running, doing sequential trims. |
3034 +------+-----+-----------------------------------------------------------+
3035 | | d | Running, doing random trims. |
3036 +------+-----+-----------------------------------------------------------+
3037 | | F | Running, currently waiting for :manpage:`fsync(2)`. |
3038 +------+-----+-----------------------------------------------------------+
3039 | | V | Running, doing verification of written data. |
3040 +------+-----+-----------------------------------------------------------+
3041 | f | | Thread finishing. |
3042 +------+-----+-----------------------------------------------------------+
3043 | E | | Thread exited, not reaped by main thread yet. |
3044 +------+-----+-----------------------------------------------------------+
3045 | _ | | Thread reaped. |
3046 +------+-----+-----------------------------------------------------------+
3047 | X | | Thread reaped, exited with an error. |
3048 +------+-----+-----------------------------------------------------------+
3049 | K | | Thread reaped, exited due to signal. |
3050 +------+-----+-----------------------------------------------------------+
3053 Example output was based on the following:
3054 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --runtime=58m \
3055 --time_based --rate=2512k --bs=256K --numjobs=10 \
3056 --name=readers --rw=read --name=writers --rw=write
3058 Fio will condense the thread string as not to take up more space on the command
3059 line than needed. For instance, if you have 10 readers and 10 writers running,
3060 the output would look like this::
3062 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]
3064 Note that the status string is displayed in order, so it's possible to tell which of
3065 the jobs are currently doing what. In the example above this means that jobs 1--10
3066 are readers and 11--20 are writers.
3068 The other values are fairly self explanatory -- number of threads currently
3069 running and doing I/O, the number of currently open files (f=), the estimated
3070 completion percentage, the rate of I/O since last check (read speed listed first,
3071 then write speed and optionally trim speed) in terms of bandwidth and IOPS, and time to completion for the current
3072 running group. It's impossible to estimate runtime of the following groups (if
3076 Example output was based on the following:
3077 TZ=UTC fio --iodepth=16 --ioengine=posixaio --filename=/tmp/fiofile \
3078 --direct=1 --size=100M --time_based --runtime=50s --rate_iops=89 \
3079 --bs=7K --name=Client1 --rw=write
3081 When fio is done (or interrupted by :kbd:`Ctrl-C`), it will show the data for
3082 each thread, group of threads, and disks in that order. For each overall thread (or
3083 group) the output looks like::
3085 Client1: (groupid=0, jobs=1): err= 0: pid=16109: Sat Jun 24 12:07:54 2017
3086 write: IOPS=88, BW=623KiB/s (638kB/s)(30.4MiB/50032msec)
3087 slat (nsec): min=500, max=145500, avg=8318.00, stdev=4781.50
3088 clat (usec): min=170, max=78367, avg=4019.02, stdev=8293.31
3089 lat (usec): min=174, max=78375, avg=4027.34, stdev=8291.79
3090 clat percentiles (usec):
3091 | 1.00th=[ 302], 5.00th=[ 326], 10.00th=[ 343], 20.00th=[ 363],
3092 | 30.00th=[ 392], 40.00th=[ 404], 50.00th=[ 416], 60.00th=[ 445],
3093 | 70.00th=[ 816], 80.00th=[ 6718], 90.00th=[12911], 95.00th=[21627],
3094 | 99.00th=[43779], 99.50th=[51643], 99.90th=[68682], 99.95th=[72877],
3096 bw ( KiB/s): min= 532, max= 686, per=0.10%, avg=622.87, stdev=24.82, samples= 100
3097 iops : min= 76, max= 98, avg=88.98, stdev= 3.54, samples= 100
3098 lat (usec) : 250=0.04%, 500=64.11%, 750=4.81%, 1000=2.79%
3099 lat (msec) : 2=4.16%, 4=1.84%, 10=4.90%, 20=11.33%, 50=5.37%
3100 lat (msec) : 100=0.65%
3101 cpu : usr=0.27%, sys=0.18%, ctx=12072, majf=0, minf=21
3102 IO depths : 1=85.0%, 2=13.1%, 4=1.8%, 8=0.1%, 16=0.0%, 32=0.0%, >=64=0.0%
3103 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3104 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3105 issued rwt: total=0,4450,0, short=0,0,0, dropped=0,0,0
3106 latency : target=0, window=0, percentile=100.00%, depth=8
3108 The job name (or first job's name when using :option:`group_reporting`) is printed,
3109 along with the group id, count of jobs being aggregated, last error id seen (which
3110 is 0 when there are no errors), pid/tid of that thread and the time the job/group
3111 completed. Below are the I/O statistics for each data direction performed (showing
3112 writes in the example above). In the order listed, they denote:
3115 The string before the colon shows the I/O direction the statistics
3116 are for. **IOPS** is the average I/Os performed per second. **BW**
3117 is the average bandwidth rate shown as: value in power of 2 format
3118 (value in power of 10 format). The last two values show: (**total
3119 I/O performed** in power of 2 format / **runtime** of that thread).
3122 Submission latency (**min** being the minimum, **max** being the
3123 maximum, **avg** being the average, **stdev** being the standard
3124 deviation). This is the time it took to submit the I/O. For
3125 sync I/O this row is not displayed as the slat is really the
3126 completion latency (since queue/complete is one operation there).
3127 This value can be in nanoseconds, microseconds or milliseconds ---
3128 fio will choose the most appropriate base and print that (in the
3129 example above nanoseconds was the best scale). Note: in :option:`--minimal` mode
3130 latencies are always expressed in microseconds.
3133 Completion latency. Same names as slat, this denotes the time from
3134 submission to completion of the I/O pieces. For sync I/O, clat will
3135 usually be equal (or very close) to 0, as the time from submit to
3136 complete is basically just CPU time (I/O has already been done, see slat
3140 Bandwidth statistics based on samples. Same names as the xlat stats,
3141 but also includes the number of samples taken (**samples**) and an
3142 approximate percentage of total aggregate bandwidth this thread
3143 received in its group (**per**). This last value is only really
3144 useful if the threads in this group are on the same disk, since they
3145 are then competing for disk access.
3148 IOPS statistics based on samples. Same names as bw.
3151 CPU usage. User and system time, along with the number of context
3152 switches this thread went through, usage of system and user time, and
3153 finally the number of major and minor page faults. The CPU utilization
3154 numbers are averages for the jobs in that reporting group, while the
3155 context and fault counters are summed.
3158 The distribution of I/O depths over the job lifetime. The numbers are
3159 divided into powers of 2 and each entry covers depths from that value
3160 up to those that are lower than the next entry -- e.g., 16= covers
3161 depths from 16 to 31. Note that the range covered by a depth
3162 distribution entry can be different to the range covered by the
3163 equivalent submit/complete distribution entry.
3166 How many pieces of I/O were submitting in a single submit call. Each
3167 entry denotes that amount and below, until the previous entry -- e.g.,
3168 16=100% means that we submitted anywhere between 9 to 16 I/Os per submit
3169 call. Note that the range covered by a submit distribution entry can
3170 be different to the range covered by the equivalent depth distribution
3174 Like the above submit number, but for completions instead.
3177 The number of read/write/trim requests issued, and how many of them were
3181 The distribution of I/O completion latencies. This is the time from when
3182 I/O leaves fio and when it gets completed. The numbers follow the same
3183 pattern as the I/O depths, meaning that 2=1.6% means that 1.6% of the
3184 I/O completed within 2 msecs, 20=12.8% means that 12.8% of the I/O took
3185 more than 10 msecs, but less than (or equal to) 20 msecs.
3188 Example output was based on the following:
3189 TZ=UTC fio --ioengine=null --iodepth=2 --size=100M --numjobs=2 \
3190 --rate_process=poisson --io_limit=32M --name=read --bs=128k \
3191 --rate=11M --name=write --rw=write --bs=2k --rate=700k
3193 After each client has been listed, the group statistics are printed. They
3194 will look like this::
3196 Run status group 0 (all jobs):
3197 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
3198 WRITE: bw=1231KiB/s (1261kB/s), 616KiB/s-621KiB/s (630kB/s-636kB/s), io=64.0MiB (67.1MB), run=52747-53223msec
3200 For each data direction it prints:
3203 Aggregate bandwidth of threads in this group followed by the
3204 minimum and maximum bandwidth of all the threads in this group.
3205 Values outside of brackets are power-of-2 format and those
3206 within are the equivalent value in a power-of-10 format.
3208 Aggregate I/O performed of all threads in this group. The
3209 format is the same as bw.
3211 The smallest and longest runtimes of the threads in this group.
3213 And finally, the disk statistics are printed. They will look like this::
3215 Disk stats (read/write):
3216 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
3218 Each value is printed for both reads and writes, with reads first. The
3222 Number of I/Os performed by all groups.
3224 Number of merges I/O the I/O scheduler.
3226 Number of ticks we kept the disk busy.
3228 Total time spent in the disk queue.
3230 The disk utilization. A value of 100% means we kept the disk
3231 busy constantly, 50% would be a disk idling half of the time.
3233 It is also possible to get fio to dump the current output while it is running,
3234 without terminating the job. To do that, send fio the **USR1** signal. You can
3235 also get regularly timed dumps by using the :option:`--status-interval`
3236 parameter, or by creating a file in :file:`/tmp` named
3237 :file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
3238 current output status.
3244 For scripted usage where you typically want to generate tables or graphs of the
3245 results, fio can output the results in a semicolon separated format. The format
3246 is one long line of values, such as::
3248 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%
3249 A description of this job goes here.
3251 The job description (if provided) follows on a second line.
3253 To enable terse output, use the :option:`--minimal` or
3254 :option:`--output-format`\=terse command line options. The
3255 first value is the version of the terse output format. If the output has to be
3256 changed for some reason, this number will be incremented by 1 to signify that
3259 Split up, the format is as follows (comments in brackets denote when a
3260 field was introduced or whether its specific to some terse version):
3264 terse version, fio version [v3], jobname, groupid, error
3268 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3269 Submission latency: min, max, mean, stdev (usec)
3270 Completion latency: min, max, mean, stdev (usec)
3271 Completion latency percentiles: 20 fields (see below)
3272 Total latency: min, max, mean, stdev (usec)
3273 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3274 IOPS [v5]: min, max, mean, stdev, number of samples
3280 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3281 Submission latency: min, max, mean, stdev (usec)
3282 Completion latency: min, max, mean, stdev (usec)
3283 Completion latency percentiles: 20 fields (see below)
3284 Total latency: min, max, mean, stdev (usec)
3285 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3286 IOPS [v5]: min, max, mean, stdev, number of samples
3288 TRIM status [all but version 3]:
3290 Fields are similar to READ/WRITE status.
3294 user, system, context switches, major faults, minor faults
3298 <=1, 2, 4, 8, 16, 32, >=64
3300 I/O latencies microseconds::
3302 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
3304 I/O latencies milliseconds::
3306 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
3308 Disk utilization [v3]::
3310 Disk name, Read ios, write ios,
3311 Read merges, write merges,
3312 Read ticks, write ticks,
3313 Time spent in queue, disk utilization percentage
3315 Additional Info (dependent on continue_on_error, default off)::
3317 total # errors, first error code
3319 Additional Info (dependent on description being set)::
3323 Completion latency percentiles can be a grouping of up to 20 sets, so for the
3324 terse output fio writes all of them. Each field will look like this::
3328 which is the Xth percentile, and the `usec` latency associated with it.
3330 For disk utilization, all disks used by fio are shown. So for each disk there
3331 will be a disk utilization section.
3333 Below is a single line containing short names for each of the fields in the
3334 minimal output v3, separated by semicolons::
3336 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
3342 There are two trace file format that you can encounter. The older (v1) format is
3343 unsupported since version 1.20-rc3 (March 2008). It will still be described
3344 below in case that you get an old trace and want to understand it.
3346 In any case the trace is a simple text file with a single action per line.
3349 Trace file format v1
3350 ~~~~~~~~~~~~~~~~~~~~
3352 Each line represents a single I/O action in the following format::
3356 where `rw=0/1` for read/write, and the offset and length entries being in bytes.
3358 This format is not supported in fio versions >= 1.20-rc3.
3361 Trace file format v2
3362 ~~~~~~~~~~~~~~~~~~~~
3364 The second version of the trace file format was added in fio version 1.17. It
3365 allows to access more then one file per trace and has a bigger set of possible
3368 The first line of the trace file has to be::
3372 Following this can be lines in two different formats, which are described below.
3374 The file management format::
3378 The filename is given as an absolute path. The action can be one of these:
3381 Add the given filename to the trace.
3383 Open the file with the given filename. The filename has to have
3384 been added with the **add** action before.
3386 Close the file with the given filename. The file has to have been
3390 The file I/O action format::
3392 filename action offset length
3394 The `filename` is given as an absolute path, and has to have been added and
3395 opened before it can be used with this format. The `offset` and `length` are
3396 given in bytes. The `action` can be one of these:
3399 Wait for `offset` microseconds. Everything below 100 is discarded.
3400 The time is relative to the previous `wait` statement.
3402 Read `length` bytes beginning from `offset`.
3404 Write `length` bytes beginning from `offset`.
3406 :manpage:`fsync(2)` the file.
3408 :manpage:`fdatasync(2)` the file.
3410 Trim the given file from the given `offset` for `length` bytes.
3412 CPU idleness profiling
3413 ----------------------
3415 In some cases, we want to understand CPU overhead in a test. For example, we
3416 test patches for the specific goodness of whether they reduce CPU usage.
3417 Fio implements a balloon approach to create a thread per CPU that runs at idle
3418 priority, meaning that it only runs when nobody else needs the cpu.
3419 By measuring the amount of work completed by the thread, idleness of each CPU
3420 can be derived accordingly.
3422 An unit work is defined as touching a full page of unsigned characters. Mean and
3423 standard deviation of time to complete an unit work is reported in "unit work"
3424 section. Options can be chosen to report detailed percpu idleness or overall
3425 system idleness by aggregating percpu stats.
3428 Verification and triggers
3429 -------------------------
3431 Fio is usually run in one of two ways, when data verification is done. The first
3432 is a normal write job of some sort with verify enabled. When the write phase has
3433 completed, fio switches to reads and verifies everything it wrote. The second
3434 model is running just the write phase, and then later on running the same job
3435 (but with reads instead of writes) to repeat the same I/O patterns and verify
3436 the contents. Both of these methods depend on the write phase being completed,
3437 as fio otherwise has no idea how much data was written.
3439 With verification triggers, fio supports dumping the current write state to
3440 local files. Then a subsequent read verify workload can load this state and know
3441 exactly where to stop. This is useful for testing cases where power is cut to a
3442 server in a managed fashion, for instance.
3444 A verification trigger consists of two things:
3446 1) Storing the write state of each job.
3447 2) Executing a trigger command.
3449 The write state is relatively small, on the order of hundreds of bytes to single
3450 kilobytes. It contains information on the number of completions done, the last X
3453 A trigger is invoked either through creation ('touch') of a specified file in
3454 the system, or through a timeout setting. If fio is run with
3455 :option:`--trigger-file`\= :file:`/tmp/trigger-file`, then it will continually
3456 check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
3457 will fire off the trigger (thus saving state, and executing the trigger
3460 For client/server runs, there's both a local and remote trigger. If fio is
3461 running as a server backend, it will send the job states back to the client for
3462 safe storage, then execute the remote trigger, if specified. If a local trigger
3463 is specified, the server will still send back the write state, but the client
3464 will then execute the trigger.
3466 Verification trigger example
3467 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3469 Let's say we want to run a powercut test on the remote machine 'server'. Our
3470 write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
3471 some point during the run, and we'll run this test from the safety or our local
3472 machine, 'localbox'. On the server, we'll start the fio backend normally::
3474 server# fio --server
3476 and on the client, we'll fire off the workload::
3478 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
3480 We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
3482 echo b > /proc/sysrq-trigger
3484 on the server once it has received the trigger and sent us the write state. This
3485 will work, but it's not **really** cutting power to the server, it's merely
3486 abruptly rebooting it. If we have a remote way of cutting power to the server
3487 through IPMI or similar, we could do that through a local trigger command
3488 instead. Let's assume we have a script that does IPMI reboot of a given hostname,
3489 ipmi-reboot. On localbox, we could then have run fio with a local trigger
3492 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
3494 For this case, fio would wait for the server to send us the write state, then
3495 execute ``ipmi-reboot server`` when that happened.
3497 Loading verify state
3498 ~~~~~~~~~~~~~~~~~~~~
3500 To load stored write state, a read verification job file must contain the
3501 :option:`verify_state_load` option. If that is set, fio will load the previously
3502 stored state. For a local fio run this is done by loading the files directly,
3503 and on a client/server run, the server backend will ask the client to send the
3504 files over and load them from there.
3510 Fio supports a variety of log file formats, for logging latencies, bandwidth,
3511 and IOPS. The logs share a common format, which looks like this:
3513 *time* (`msec`), *value*, *data direction*, *offset*
3515 Time for the log entry is always in milliseconds. The *value* logged depends
3516 on the type of log, it will be one of the following:
3519 Value is latency in usecs
3525 *Data direction* is one of the following:
3534 The *offset* is the offset, in bytes, from the start of the file, for that
3535 particular I/O. The logging of the offset can be toggled with
3536 :option:`log_offset`.
3538 Fio defaults to logging every individual I/O. When IOPS are logged for individual
3539 I/Os the value entry will always be 1. If windowed logging is enabled through
3540 :option:`log_avg_msec`, fio logs the average values over the specified period of time.
3541 If windowed logging is enabled and :option:`log_max_value` is set, then fio logs
3542 maximum values in that window instead of averages. Since 'data direction' and
3543 'offset' are per-I/O values, they aren't applicable if windowed logging is enabled.
3548 Normally fio is invoked as a stand-alone application on the machine where the
3549 I/O workload should be generated. However, the backend and frontend of fio can
3550 be run separately i.e., the fio server can generate an I/O workload on the "Device
3551 Under Test" while being controlled by a client on another machine.
3553 Start the server on the machine which has access to the storage DUT::
3557 where `args` defines what fio listens to. The arguments are of the form
3558 ``type,hostname`` or ``IP,port``. *type* is either ``ip`` (or ip4) for TCP/IP
3559 v4, ``ip6`` for TCP/IP v6, or ``sock`` for a local unix domain socket.
3560 *hostname* is either a hostname or IP address, and *port* is the port to listen
3561 to (only valid for TCP/IP, not a local socket). Some examples:
3565 Start a fio server, listening on all interfaces on the default port (8765).
3567 2) ``fio --server=ip:hostname,4444``
3569 Start a fio server, listening on IP belonging to hostname and on port 4444.
3571 3) ``fio --server=ip6:::1,4444``
3573 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
3575 4) ``fio --server=,4444``
3577 Start a fio server, listening on all interfaces on port 4444.
3579 5) ``fio --server=1.2.3.4``
3581 Start a fio server, listening on IP 1.2.3.4 on the default port.
3583 6) ``fio --server=sock:/tmp/fio.sock``
3585 Start a fio server, listening on the local socket :file:`/tmp/fio.sock`.
3587 Once a server is running, a "client" can connect to the fio server with::
3589 fio <local-args> --client=<server> <remote-args> <job file(s)>
3591 where `local-args` are arguments for the client where it is running, `server`
3592 is the connect string, and `remote-args` and `job file(s)` are sent to the
3593 server. The `server` string follows the same format as it does on the server
3594 side, to allow IP/hostname/socket and port strings.
3596 Fio can connect to multiple servers this way::
3598 fio --client=<server1> <job file(s)> --client=<server2> <job file(s)>
3600 If the job file is located on the fio server, then you can tell the server to
3601 load a local file as well. This is done by using :option:`--remote-config` ::
3603 fio --client=server --remote-config /path/to/file.fio
3605 Then fio will open this local (to the server) job file instead of being passed
3606 one from the client.
3608 If you have many servers (example: 100 VMs/containers), you can input a pathname
3609 of a file containing host IPs/names as the parameter value for the
3610 :option:`--client` option. For example, here is an example :file:`host.list`
3611 file containing 2 hostnames::
3613 host1.your.dns.domain
3614 host2.your.dns.domain
3616 The fio command would then be::
3618 fio --client=host.list <job file(s)>
3620 In this mode, you cannot input server-specific parameters or job files -- all
3621 servers receive the same job file.
3623 In order to let ``fio --client`` runs use a shared filesystem from multiple
3624 hosts, ``fio --client`` now prepends the IP address of the server to the
3625 filename. For example, if fio is using the directory :file:`/mnt/nfs/fio` and is
3626 writing filename :file:`fileio.tmp`, with a :option:`--client` `hostfile`
3627 containing two hostnames ``h1`` and ``h2`` with IP addresses 192.168.10.120 and
3628 192.168.10.121, then fio will create two files::
3630 /mnt/nfs/fio/192.168.10.120.fileio.tmp
3631 /mnt/nfs/fio/192.168.10.121.fileio.tmp