4 The first step in getting fio to simulate a desired I/O workload, is writing a
5 job file describing that specific setup. A job file may contain any number of
6 threads and/or files -- the typical contents of the job file is a *global*
7 section defining shared parameters, and one or more job sections describing the
8 jobs involved. When run, fio parses this file and sets everything up as
9 described. If we break down a job from top to bottom, it contains the following
14 Defines the I/O pattern issued to the file(s). We may only be reading
15 sequentially from this file(s), or we may be writing randomly. Or even
16 mixing reads and writes, sequentially or randomly.
17 Should we be doing buffered I/O, or direct/raw I/O?
21 In how large chunks are we issuing I/O? This may be a single value,
22 or it may describe a range of block sizes.
26 How much data are we going to be reading/writing.
30 How do we issue I/O? We could be memory mapping the file, we could be
31 using regular read/write, we could be using splice, async I/O, or even
36 If the I/O engine is async, how large a queuing depth do we want to
42 How many files are we spreading the workload over.
44 `Threads, processes and job synchronization`_
46 How many threads or processes should we spread this workload over.
48 The above are the basic parameters defined for a workload, in addition there's a
49 multitude of parameters that modify other aspects of how this job behaves.
55 .. option:: --debug=type
57 Enable verbose tracing `type` of various fio actions. May be ``all`` for all types
58 or individual types separated by a comma (e.g. ``--debug=file,mem`` will
59 enable file and memory debugging). Currently, additional logging is
63 Dump info related to processes.
65 Dump info related to file actions.
67 Dump info related to I/O queuing.
69 Dump info related to memory allocations.
71 Dump info related to blktrace setup.
73 Dump info related to I/O verification.
75 Enable all debug options.
77 Dump info related to random offset generation.
79 Dump info related to option matching and parsing.
81 Dump info related to disk utilization updates.
83 Dump info only related to job number x.
85 Dump info only related to mutex up/down ops.
87 Dump info related to profile extensions.
89 Dump info related to internal time keeping.
91 Dump info related to networking connections.
93 Dump info related to I/O rate switching.
95 Dump info related to log compress/decompress.
97 Show available debug options.
99 .. option:: --parse-only
101 Parse options only, don't start any I/O.
103 .. option:: --output=filename
105 Write output to file `filename`.
107 .. option:: --output-format=format
109 Set the reporting `format` to `normal`, `terse`, `json`, or `json+`. Multiple
110 formats can be selected, separated by a comma. `terse` is a CSV based
111 format. `json+` is like `json`, except it adds a full dump of the latency
114 .. option:: --bandwidth-log
116 Generate aggregate bandwidth logs.
118 .. option:: --minimal
120 Print statistics in a terse, semicolon-delimited format.
122 .. option:: --append-terse
124 Print statistics in selected mode AND terse, semicolon-delimited format.
125 **Deprecated**, use :option:`--output-format` instead to select multiple
128 .. option:: --terse-version=version
130 Set terse `version` output format (default 3, or 2 or 4 or 5).
132 .. option:: --version
134 Print version information and exit.
138 Print a summary of the command line options and exit.
140 .. option:: --cpuclock-test
142 Perform test and validation of internal CPU clock.
144 .. option:: --crctest=[test]
146 Test the speed of the built-in checksumming functions. If no argument is
147 given, all of them are tested. Alternatively, a comma separated list can
148 be passed, in which case the given ones are tested.
150 .. option:: --cmdhelp=command
152 Print help information for `command`. May be ``all`` for all commands.
154 .. option:: --enghelp=[ioengine[,command]]
156 List all commands defined by `ioengine`, or print help for `command`
157 defined by `ioengine`. If no `ioengine` is given, list all
160 .. option:: --showcmd=jobfile
162 Convert `jobfile` to a set of command-line options.
164 .. option:: --readonly
166 Turn on safety read-only checks, preventing writes. 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 Specifies when real-time ETA estimate should be printed. `when` may be
176 `always`, `never` or `auto`.
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 a full status dump of cumulative (from job start) values at `time`
186 intervals. This option does *not* provide per-period measurements. So
187 values such as bandwidth are running averages. When the time unit is omitted,
188 `time` is interpreted in seconds.
190 .. option:: --section=name
192 Only run specified section `name` in job file. Multiple sections can be specified.
193 The ``--section`` option allows one to combine related jobs into one file.
194 E.g. one job file could define light, moderate, and heavy sections. Tell
195 fio to run only the "heavy" section by giving ``--section=heavy``
196 command line option. One can also specify the "write" operations in one
197 section and "verify" operation in another section. The ``--section`` option
198 only applies to job sections. The reserved *global* section is always
201 .. option:: --alloc-size=kb
203 Set the internal smalloc pool size to `kb` in KiB. The
204 ``--alloc-size`` switch allows one to use a larger pool size for smalloc.
205 If running large jobs with randommap enabled, fio can run out of memory.
206 Smalloc is an internal allocator for shared structures from a fixed size
207 memory pool and can grow to 16 pools. The pool size defaults to 16MiB.
209 NOTE: While running :file:`.fio_smalloc.*` backing store files are visible
212 .. option:: --warnings-fatal
214 All fio parser warnings are fatal, causing fio to exit with an
217 .. option:: --max-jobs=nr
219 Set the maximum number of threads/processes to support to `nr`.
220 NOTE: On Linux, it may be necessary to increase the shared-memory
221 limit ('/proc/sys/kernel/shmmax') if fio runs into errors while
224 .. option:: --server=args
226 Start a backend server, with `args` specifying what to listen to.
227 See `Client/Server`_ section.
229 .. option:: --daemonize=pidfile
231 Background a fio server, writing the pid to the given `pidfile` file.
233 .. option:: --client=hostname
235 Instead of running the jobs locally, send and run them on the given `hostname`
236 or set of `hostname`s. See `Client/Server`_ section.
238 .. option:: --remote-config=file
240 Tell fio server to load this local `file`.
242 .. option:: --idle-prof=option
244 Report CPU idleness. `option` is one of the following:
247 Run unit work calibration only and exit.
250 Show aggregate system idleness and unit work.
253 As **system** but also show per CPU idleness.
255 .. option:: --inflate-log=log
257 Inflate and output compressed `log`.
259 .. option:: --trigger-file=file
261 Execute trigger command when `file` exists.
263 .. option:: --trigger-timeout=time
265 Execute trigger at this `time`.
267 .. option:: --trigger=command
269 Set this `command` as local trigger.
271 .. option:: --trigger-remote=command
273 Set this `command` as remote trigger.
275 .. option:: --aux-path=path
277 Use this `path` for fio state generated files.
279 Any parameters following the options will be assumed to be job files, unless
280 they match a job file parameter. Multiple job files can be listed and each job
281 file will be regarded as a separate group. Fio will :option:`stonewall`
282 execution between each group.
288 As previously described, fio accepts one or more job files describing what it is
289 supposed to do. The job file format is the classic ini file, where the names
290 enclosed in [] brackets define the job name. You are free to use any ASCII name
291 you want, except *global* which has special meaning. Following the job name is
292 a sequence of zero or more parameters, one per line, that define the behavior of
293 the job. If the first character in a line is a ';' or a '#', the entire line is
294 discarded as a comment.
296 A *global* section sets defaults for the jobs described in that file. A job may
297 override a *global* section parameter, and a job file may even have several
298 *global* sections if so desired. A job is only affected by a *global* section
301 The :option:`--cmdhelp` option also lists all options. If used with a `command`
302 argument, :option:`--cmdhelp` will detail the given `command`.
304 See the `examples/` directory for inspiration on how to write job files. Note
305 the copyright and license requirements currently apply to `examples/` files.
307 So let's look at a really simple job file that defines two processes, each
308 randomly reading from a 128MiB file:
312 ; -- start job file --
323 As you can see, the job file sections themselves are empty as all the described
324 parameters are shared. As no :option:`filename` option is given, fio makes up a
325 `filename` for each of the jobs as it sees fit. On the command line, this job
326 would look as follows::
328 $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
331 Let's look at an example that has a number of processes writing randomly to
336 ; -- start job file --
347 Here we have no *global* section, as we only have one job defined anyway. We
348 want to use async I/O here, with a depth of 4 for each file. We also increased
349 the buffer size used to 32KiB and define numjobs to 4 to fork 4 identical
350 jobs. The result is 4 processes each randomly writing to their own 64MiB
351 file. Instead of using the above job file, you could have given the parameters
352 on the command line. For this case, you would specify::
354 $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
356 When fio is utilized as a basis of any reasonably large test suite, it might be
357 desirable to share a set of standardized settings across multiple job files.
358 Instead of copy/pasting such settings, any section may pull in an external
359 :file:`filename.fio` file with *include filename* directive, as in the following
362 ; -- start job file including.fio --
366 include glob-include.fio
373 include test-include.fio
374 ; -- end job file including.fio --
378 ; -- start job file glob-include.fio --
381 ; -- end job file glob-include.fio --
385 ; -- start job file test-include.fio --
388 ; -- end job file test-include.fio --
390 Settings pulled into a section apply to that section only (except *global*
391 section). Include directives may be nested in that any included file may contain
392 further include directive(s). Include files may not contain [] sections.
395 Environment variables
396 ~~~~~~~~~~~~~~~~~~~~~
398 Fio also supports environment variable expansion in job files. Any sub-string of
399 the form ``${VARNAME}`` as part of an option value (in other words, on the right
400 of the '='), will be expanded to the value of the environment variable called
401 `VARNAME`. If no such environment variable is defined, or `VARNAME` is the
402 empty string, the empty string will be substituted.
404 As an example, let's look at a sample fio invocation and job file::
406 $ SIZE=64m NUMJOBS=4 fio jobfile.fio
410 ; -- start job file --
417 This will expand to the following equivalent job file at runtime:
421 ; -- start job file --
428 Fio ships with a few example job files, you can also look there for inspiration.
433 Additionally, fio has a set of reserved keywords that will be replaced
434 internally with the appropriate value. Those keywords are:
438 The architecture page size of the running system.
442 Megabytes of total memory in the system.
446 Number of online available CPUs.
448 These can be used on the command line or in the job file, and will be
449 automatically substituted with the current system values when the job is
450 run. Simple math is also supported on these keywords, so you can perform actions
455 and get that properly expanded to 8 times the size of memory in the machine.
461 This section describes in details each parameter associated with a job. Some
462 parameters take an option of a given type, such as an integer or a
463 string. Anywhere a numeric value is required, an arithmetic expression may be
464 used, provided it is surrounded by parentheses. Supported operators are:
473 For time values in expressions, units are microseconds by default. This is
474 different than for time values not in expressions (not enclosed in
475 parentheses). The following types are used:
482 String: A sequence of alphanumeric characters.
485 Integer with possible time suffix. Without a unit value is interpreted as
486 seconds unless otherwise specified. Accepts a suffix of 'd' for days, 'h' for
487 hours, 'm' for minutes, 's' for seconds, 'ms' (or 'msec') for milliseconds and
488 'us' (or 'usec') for microseconds. For example, use 10m for 10 minutes.
493 Integer. A whole number value, which may contain an integer prefix
494 and an integer suffix:
496 [*integer prefix*] **number** [*integer suffix*]
498 The optional *integer prefix* specifies the number's base. The default
499 is decimal. *0x* specifies hexadecimal.
501 The optional *integer suffix* specifies the number's units, and includes an
502 optional unit prefix and an optional unit. For quantities of data, the
503 default unit is bytes. For quantities of time, the default unit is seconds
504 unless otherwise specified.
506 With :option:`kb_base`\=1000, fio follows international standards for unit
507 prefixes. To specify power-of-10 decimal values defined in the
508 International System of Units (SI):
510 * *K* -- means kilo (K) or 1000
511 * *M* -- means mega (M) or 1000**2
512 * *G* -- means giga (G) or 1000**3
513 * *T* -- means tera (T) or 1000**4
514 * *P* -- means peta (P) or 1000**5
516 To specify power-of-2 binary values defined in IEC 80000-13:
518 * *Ki* -- means kibi (Ki) or 1024
519 * *Mi* -- means mebi (Mi) or 1024**2
520 * *Gi* -- means gibi (Gi) or 1024**3
521 * *Ti* -- means tebi (Ti) or 1024**4
522 * *Pi* -- means pebi (Pi) or 1024**5
524 With :option:`kb_base`\=1024 (the default), the unit prefixes are opposite
525 from those specified in the SI and IEC 80000-13 standards to provide
526 compatibility with old scripts. For example, 4k means 4096.
528 For quantities of data, an optional unit of 'B' may be included
529 (e.g., 'kB' is the same as 'k').
531 The *integer suffix* is not case sensitive (e.g., m/mi mean mebi/mega,
532 not milli). 'b' and 'B' both mean byte, not bit.
534 Examples with :option:`kb_base`\=1000:
536 * *4 KiB*: 4096, 4096b, 4096B, 4ki, 4kib, 4kiB, 4Ki, 4KiB
537 * *1 MiB*: 1048576, 1mi, 1024ki
538 * *1 MB*: 1000000, 1m, 1000k
539 * *1 TiB*: 1099511627776, 1ti, 1024gi, 1048576mi
540 * *1 TB*: 1000000000, 1t, 1000m, 1000000k
542 Examples with :option:`kb_base`\=1024 (default):
544 * *4 KiB*: 4096, 4096b, 4096B, 4k, 4kb, 4kB, 4K, 4KB
545 * *1 MiB*: 1048576, 1m, 1024k
546 * *1 MB*: 1000000, 1mi, 1000ki
547 * *1 TiB*: 1099511627776, 1t, 1024g, 1048576m
548 * *1 TB*: 1000000000, 1ti, 1000mi, 1000000ki
550 To specify times (units are not case sensitive):
554 * *M* -- means minutes
555 * *s* -- or sec means seconds (default)
556 * *ms* -- or *msec* means milliseconds
557 * *us* -- or *usec* means microseconds
559 If the option accepts an upper and lower range, use a colon ':' or
560 minus '-' to separate such values. See :ref:`irange <irange>`.
561 If the lower value specified happens to be larger than the upper value
562 the two values are swapped.
567 Boolean. Usually parsed as an integer, however only defined for
568 true and false (1 and 0).
573 Integer range with suffix. Allows value range to be given, such as
574 1024-4096. A colon may also be used as the separator, e.g. 1k:4k. If the
575 option allows two sets of ranges, they can be specified with a ',' or '/'
576 delimiter: 1k-4k/8k-32k. Also see :ref:`int <int>`.
579 A list of floating point numbers, separated by a ':' character.
581 With the above in mind, here follows the complete list of fio job parameters.
587 .. option:: kb_base=int
589 Select the interpretation of unit prefixes in input parameters.
592 Inputs comply with IEC 80000-13 and the International
593 System of Units (SI). Use:
595 - power-of-2 values with IEC prefixes (e.g., KiB)
596 - power-of-10 values with SI prefixes (e.g., kB)
599 Compatibility mode (default). To avoid breaking old scripts:
601 - power-of-2 values with SI prefixes
602 - power-of-10 values with IEC prefixes
604 See :option:`bs` for more details on input parameters.
606 Outputs always use correct prefixes. Most outputs include both
609 bw=2383.3kB/s (2327.4KiB/s)
611 If only one value is reported, then kb_base selects the one to use:
613 **1000** -- SI prefixes
615 **1024** -- IEC prefixes
617 .. option:: unit_base=int
619 Base unit for reporting. Allowed values are:
622 Use auto-detection (default).
634 ASCII name of the job. This may be used to override the name printed by fio
635 for this job. Otherwise the job name is used. On the command line this
636 parameter has the special purpose of also signaling the start of a new job.
638 .. option:: description=str
640 Text description of the job. Doesn't do anything except dump this text
641 description when this job is run. It's not parsed.
643 .. option:: loops=int
645 Run the specified number of iterations of this job. Used to repeat the same
646 workload a given number of times. Defaults to 1.
648 .. option:: numjobs=int
650 Create the specified number of clones of this job. Each clone of job
651 is spawned as an independent thread or process. May be used to setup a
652 larger number of threads/processes doing the same thing. Each thread is
653 reported separately; to see statistics for all clones as a whole, use
654 :option:`group_reporting` in conjunction with :option:`new_group`.
655 See :option:`--max-jobs`. Default: 1.
658 Time related parameters
659 ~~~~~~~~~~~~~~~~~~~~~~~
661 .. option:: runtime=time
663 Tell fio to terminate processing after the specified period of time. It
664 can be quite hard to determine for how long a specified job will run, so
665 this parameter is handy to cap the total runtime to a given time. When
666 the unit is omitted, the value is intepreted in seconds.
668 .. option:: time_based
670 If set, fio will run for the duration of the :option:`runtime` specified
671 even if the file(s) are completely read or written. It will simply loop over
672 the same workload as many times as the :option:`runtime` allows.
674 .. option:: startdelay=irange(time)
676 Delay the start of job for the specified amount of time. Can be a single
677 value or a range. When given as a range, each thread will choose a value
678 randomly from within the range. Value is in seconds if a unit is omitted.
680 .. option:: ramp_time=time
682 If set, fio will run the specified workload for this amount of time before
683 logging any performance numbers. Useful for letting performance settle
684 before logging results, thus minimizing the runtime required for stable
685 results. Note that the ``ramp_time`` is considered lead in time for a job,
686 thus it will increase the total runtime if a special timeout or
687 :option:`runtime` is specified. When the unit is omitted, the value is
690 .. option:: clocksource=str
692 Use the given clocksource as the base of timing. The supported options are:
695 :manpage:`gettimeofday(2)`
698 :manpage:`clock_gettime(2)`
701 Internal CPU clock source
703 cpu is the preferred clocksource if it is reliable, as it is very fast (and
704 fio is heavy on time calls). Fio will automatically use this clocksource if
705 it's supported and considered reliable on the system it is running on,
706 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
707 means supporting TSC Invariant.
709 .. option:: gtod_reduce=bool
711 Enable all of the :manpage:`gettimeofday(2)` reducing options
712 (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus
713 reduce precision of the timeout somewhat to really shrink the
714 :manpage:`gettimeofday(2)` call count. With this option enabled, we only do
715 about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all
716 time keeping was enabled.
718 .. option:: gtod_cpu=int
720 Sometimes it's cheaper to dedicate a single thread of execution to just
721 getting the current time. Fio (and databases, for instance) are very
722 intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set
723 one CPU aside for doing nothing but logging current time to a shared memory
724 location. Then the other threads/processes that run I/O workloads need only
725 copy that segment, instead of entering the kernel with a
726 :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time
727 calls will be excluded from other uses. Fio will manually clear it from the
728 CPU mask of other jobs.
734 .. option:: directory=str
736 Prefix filenames with this directory. Used to place files in a different
737 location than :file:`./`. You can specify a number of directories by
738 separating the names with a ':' character. These directories will be
739 assigned equally distributed to job clones created by :option:`numjobs` as
740 long as they are using generated filenames. If specific `filename(s)` are
741 set fio will use the first listed directory, and thereby matching the
742 `filename` semantic which generates a file each clone if not specified, but
743 let all clones use the same if set.
745 See the :option:`filename` option for information on how to escape "``:``" and
746 "``\``" characters within the directory path itself.
748 .. option:: filename=str
750 Fio normally makes up a `filename` based on the job name, thread number, and
751 file number (see :option:`filename_format`). If you want to share files
752 between threads in a job or several
753 jobs with fixed file paths, specify a `filename` for each of them to override
754 the default. If the ioengine is file based, you can specify a number of files
755 by separating the names with a ':' colon. So if you wanted a job to open
756 :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use
757 ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is
758 specified, :option:`nrfiles` is ignored. The size of regular files specified
759 by this option will be :option:`size` divided by number of files unless an
760 explicit size is specified by :option:`filesize`.
762 Each colon and backslash in the wanted path must be escaped with a ``\``
763 character. For instance, if the path is :file:`/dev/dsk/foo@3,0:c` then you
764 would use ``filename=/dev/dsk/foo@3,0\:c`` and if the path is
765 :file:`F:\\filename` then you would use ``filename=F\:\\filename``.
767 On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for
768 the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc.
769 Note: Windows and FreeBSD prevent write access to areas
770 of the disk containing in-use data (e.g. filesystems).
772 The filename "`-`" is a reserved name, meaning *stdin* or *stdout*. Which
773 of the two depends on the read/write direction set.
775 .. option:: filename_format=str
777 If sharing multiple files between jobs, it is usually necessary to have fio
778 generate the exact names that you want. By default, fio will name a file
779 based on the default file format specification of
780 :file:`jobname.jobnumber.filenumber`. With this option, that can be
781 customized. Fio will recognize and replace the following keywords in this
785 The name of the worker thread or process.
787 The incremental number of the worker thread or process.
789 The incremental number of the file for that worker thread or
792 To have dependent jobs share a set of files, this option can be set to have
793 fio generate filenames that are shared between the two. For instance, if
794 :file:`testfiles.$filenum` is specified, file number 4 for any job will be
795 named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum`
796 will be used if no other format specifier is given.
798 .. option:: unique_filename=bool
800 To avoid collisions between networked clients, fio defaults to prefixing any
801 generated filenames (with a directory specified) with the source of the
802 client connecting. To disable this behavior, set this option to 0.
804 .. option:: opendir=str
806 Recursively open any files below directory `str`.
808 .. option:: lockfile=str
810 Fio defaults to not locking any files before it does I/O to them. If a file
811 or file descriptor is shared, fio can serialize I/O to that file to make the
812 end result consistent. This is usual for emulating real workloads that share
813 files. The lock modes are:
816 No locking. The default.
818 Only one thread or process may do I/O at a time, excluding all
821 Read-write locking on the file. Many readers may
822 access the file at the same time, but writes get exclusive access.
824 .. option:: nrfiles=int
826 Number of files to use for this job. Defaults to 1. The size of files
827 will be :option:`size` divided by this unless explicit size is specified by
828 :option:`filesize`. Files are created for each thread separately, and each
829 file will have a file number within its name by default, as explained in
830 :option:`filename` section.
833 .. option:: openfiles=int
835 Number of files to keep open at the same time. Defaults to the same as
836 :option:`nrfiles`, can be set smaller to limit the number simultaneous
839 .. option:: file_service_type=str
841 Defines how fio decides which file from a job to service next. The following
845 Choose a file at random.
848 Round robin over opened files. This is the default.
851 Finish one file before moving on to the next. Multiple files can
852 still be open depending on :option:`openfiles`.
855 Use a *Zipf* distribution to decide what file to access.
858 Use a *Pareto* distribution to decide what file to access.
861 Use a *Gaussian* (normal) distribution to decide what file to
867 For *random*, *roundrobin*, and *sequential*, a postfix can be appended to
868 tell fio how many I/Os to issue before switching to a new file. For example,
869 specifying ``file_service_type=random:8`` would cause fio to issue
870 8 I/Os before selecting a new file at random. For the non-uniform
871 distributions, a floating point postfix can be given to influence how the
872 distribution is skewed. See :option:`random_distribution` for a description
873 of how that would work.
875 .. option:: ioscheduler=str
877 Attempt to switch the device hosting the file to the specified I/O scheduler
880 .. option:: create_serialize=bool
882 If true, serialize the file creation for the jobs. This may be handy to
883 avoid interleaving of data files, which may greatly depend on the filesystem
884 used and even the number of processors in the system. Default: true.
886 .. option:: create_fsync=bool
888 :manpage:`fsync(2)` the data file after creation. This is the default.
890 .. option:: create_on_open=bool
892 If true, don't pre-create files but allow the job's open() to create a file
893 when it's time to do I/O. Default: false -- pre-create all necessary files
896 .. option:: create_only=bool
898 If true, fio will only run the setup phase of the job. If files need to be
899 laid out or updated on disk, only that will be done -- the actual job contents
900 are not executed. Default: false.
902 .. option:: allow_file_create=bool
904 If true, fio is permitted to create files as part of its workload. If this
905 option is false, then fio will error out if
906 the files it needs to use don't already exist. Default: true.
908 .. option:: allow_mounted_write=bool
910 If this isn't set, fio will abort jobs that are destructive (e.g. that write)
911 to what appears to be a mounted device or partition. This should help catch
912 creating inadvertently destructive tests, not realizing that the test will
913 destroy data on the mounted file system. Note that some platforms don't allow
914 writing against a mounted device regardless of this option. Default: false.
916 .. option:: pre_read=bool
918 If this is given, files will be pre-read into memory before starting the
919 given I/O operation. This will also clear the :option:`invalidate` flag,
920 since it is pointless to pre-read and then drop the cache. This will only
921 work for I/O engines that are seek-able, since they allow you to read the
922 same data multiple times. Thus it will not work on non-seekable I/O engines
923 (e.g. network, splice). Default: false.
925 .. option:: unlink=bool
927 Unlink the job files when done. Not the default, as repeated runs of that
928 job would then waste time recreating the file set again and again. Default:
931 .. option:: unlink_each_loop=bool
933 Unlink job files after each iteration or loop. Default: false.
935 .. option:: zonesize=int
937 Divide a file into zones of the specified size. See :option:`zoneskip`.
939 .. option:: zonerange=int
941 Give size of an I/O zone. See :option:`zoneskip`.
943 .. option:: zoneskip=int
945 Skip the specified number of bytes when :option:`zonesize` data has been
946 read. The two zone options can be used to only do I/O on zones of a file.
952 .. option:: direct=bool
954 If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that
955 OpenBSD and ZFS on Solaris don't support direct I/O. On Windows the synchronous
956 ioengines don't support direct I/O. Default: false.
958 .. option:: atomic=bool
960 If value is true, attempt to use atomic direct I/O. Atomic writes are
961 guaranteed to be stable once acknowledged by the operating system. Only
962 Linux supports O_ATOMIC right now.
964 .. option:: buffered=bool
966 If value is true, use buffered I/O. This is the opposite of the
967 :option:`direct` option. Defaults to true.
969 .. option:: readwrite=str, rw=str
971 Type of I/O pattern. Accepted values are:
978 Sequential trims (Linux block devices only).
984 Random trims (Linux block devices only).
986 Sequential mixed reads and writes.
988 Random mixed reads and writes.
990 Sequential trim+write sequences. Blocks will be trimmed first,
991 then the same blocks will be written to.
993 Fio defaults to read if the option is not specified. For the mixed I/O
994 types, the default is to split them 50/50. For certain types of I/O the
995 result may still be skewed a bit, since the speed may be different.
997 It is possible to specify the number of I/Os to do before getting a new
998 offset by appending ``:<nr>`` to the end of the string given. For a
999 random read, it would look like ``rw=randread:8`` for passing in an offset
1000 modifier with a value of 8. If the suffix is used with a sequential I/O
1001 pattern, then the *<nr>* value specified will be **added** to the generated
1002 offset for each I/O turning sequential I/O into sequential I/O with holes.
1003 For instance, using ``rw=write:4k`` will skip 4k for every write. Also see
1004 the :option:`rw_sequencer` option.
1006 .. option:: rw_sequencer=str
1008 If an offset modifier is given by appending a number to the ``rw=<str>``
1009 line, then this option controls how that number modifies the I/O offset
1010 being generated. Accepted values are:
1013 Generate sequential offset.
1015 Generate the same offset.
1017 ``sequential`` is only useful for random I/O, where fio would normally
1018 generate a new random offset for every I/O. If you append e.g. 8 to randread,
1019 you would get a new random offset for every 8 I/Os. The result would be a
1020 seek for only every 8 I/Os, instead of for every I/O. Use ``rw=randread:8``
1021 to specify that. As sequential I/O is already sequential, setting
1022 ``sequential`` for that would not result in any differences. ``identical``
1023 behaves in a similar fashion, except it sends the same offset 8 number of
1024 times before generating a new offset.
1026 .. option:: unified_rw_reporting=bool
1028 Fio normally reports statistics on a per data direction basis, meaning that
1029 reads, writes, and trims are accounted and reported separately. If this
1030 option is set fio sums the results and report them as "mixed" instead.
1032 .. option:: randrepeat=bool
1034 Seed the random number generator used for random I/O patterns in a
1035 predictable way so the pattern is repeatable across runs. Default: true.
1037 .. option:: allrandrepeat=bool
1039 Seed all random number generators in a predictable way so results are
1040 repeatable across runs. Default: false.
1042 .. option:: randseed=int
1044 Seed the random number generators based on this seed value, to be able to
1045 control what sequence of output is being generated. If not set, the random
1046 sequence depends on the :option:`randrepeat` setting.
1048 .. option:: fallocate=str
1050 Whether pre-allocation is performed when laying down files.
1051 Accepted values are:
1054 Do not pre-allocate space.
1057 Use a platform's native pre-allocation call but fall back to
1058 **none** behavior if it fails/is not implemented.
1061 Pre-allocate via :manpage:`posix_fallocate(3)`.
1064 Pre-allocate via :manpage:`fallocate(2)` with
1065 FALLOC_FL_KEEP_SIZE set.
1068 Backward-compatible alias for **none**.
1071 Backward-compatible alias for **posix**.
1073 May not be available on all supported platforms. **keep** is only available
1074 on Linux. If using ZFS on Solaris this cannot be set to **posix**
1075 because ZFS doesn't support pre-allocation. Default: **native** if any
1076 pre-allocation methods are available, **none** if not.
1078 .. option:: fadvise_hint=str
1080 Use :manpage:`posix_fadvise(2)` to advise the kernel on what I/O patterns
1081 are likely to be issued. Accepted values are:
1084 Backwards-compatible hint for "no hint".
1087 Backwards compatible hint for "advise with fio workload type". This
1088 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1089 for a sequential workload.
1092 Advise using **FADV_SEQUENTIAL**.
1095 Advise using **FADV_RANDOM**.
1097 .. option:: write_hint=str
1099 Use :manpage:`fcntl(2)` to advise the kernel what life time to expect
1100 from a write. Only supported on Linux, as of version 4.13. Accepted
1104 No particular life time associated with this file.
1107 Data written to this file has a short life time.
1110 Data written to this file has a medium life time.
1113 Data written to this file has a long life time.
1116 Data written to this file has a very long life time.
1118 The values are all relative to each other, and no absolute meaning
1119 should be associated with them.
1121 .. option:: offset=int
1123 Start I/O at the provided offset in the file, given as either a fixed size in
1124 bytes or a percentage. If a percentage is given, the next ``blockalign``-ed
1125 offset will be used. Data before the given offset will not be touched. This
1126 effectively caps the file size at `real_size - offset`. Can be combined with
1127 :option:`size` to constrain the start and end range of the I/O workload.
1128 A percentage can be specified by a number between 1 and 100 followed by '%',
1129 for example, ``offset=20%`` to specify 20%.
1131 .. option:: offset_increment=int
1133 If this is provided, then the real offset becomes `offset + offset_increment
1134 * thread_number`, where the thread number is a counter that starts at 0 and
1135 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1136 specified). This option is useful if there are several jobs which are
1137 intended to operate on a file in parallel disjoint segments, with even
1138 spacing between the starting points.
1140 .. option:: number_ios=int
1142 Fio will normally perform I/Os until it has exhausted the size of the region
1143 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1144 condition). With this setting, the range/size can be set independently of
1145 the number of I/Os to perform. When fio reaches this number, it will exit
1146 normally and report status. Note that this does not extend the amount of I/O
1147 that will be done, it will only stop fio if this condition is met before
1148 other end-of-job criteria.
1150 .. option:: fsync=int
1152 If writing to a file, issue an :manpage:`fsync(2)` (or its equivalent) of
1153 the dirty data for every number of blocks given. For example, if you give 32
1154 as a parameter, fio will sync the file after every 32 writes issued. If fio is
1155 using non-buffered I/O, we may not sync the file. The exception is the sg
1156 I/O engine, which synchronizes the disk cache anyway. Defaults to 0, which
1157 means fio does not periodically issue and wait for a sync to complete. Also
1158 see :option:`end_fsync` and :option:`fsync_on_close`.
1160 .. option:: fdatasync=int
1162 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
1163 not metadata blocks. In Windows, FreeBSD, and DragonFlyBSD there is no
1164 :manpage:`fdatasync(2)` so this falls back to using :manpage:`fsync(2)`.
1165 Defaults to 0, which means fio does not periodically issue and wait for a
1166 data-only sync to complete.
1168 .. option:: write_barrier=int
1170 Make every `N-th` write a barrier write.
1172 .. option:: sync_file_range=str:int
1174 Use :manpage:`sync_file_range(2)` for every `int` number of write
1175 operations. Fio will track range of writes that have happened since the last
1176 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1179 SYNC_FILE_RANGE_WAIT_BEFORE
1181 SYNC_FILE_RANGE_WRITE
1183 SYNC_FILE_RANGE_WAIT_AFTER
1185 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1186 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1187 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1190 .. option:: overwrite=bool
1192 If true, writes to a file will always overwrite existing data. If the file
1193 doesn't already exist, it will be created before the write phase begins. If
1194 the file exists and is large enough for the specified write phase, nothing
1195 will be done. Default: false.
1197 .. option:: end_fsync=bool
1199 If true, :manpage:`fsync(2)` file contents when a write stage has completed.
1202 .. option:: fsync_on_close=bool
1204 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
1205 from :option:`end_fsync` in that it will happen on every file close, not
1206 just at the end of the job. Default: false.
1208 .. option:: rwmixread=int
1210 Percentage of a mixed workload that should be reads. Default: 50.
1212 .. option:: rwmixwrite=int
1214 Percentage of a mixed workload that should be writes. If both
1215 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1216 add up to 100%, the latter of the two will be used to override the
1217 first. This may interfere with a given rate setting, if fio is asked to
1218 limit reads or writes to a certain rate. If that is the case, then the
1219 distribution may be skewed. Default: 50.
1221 .. option:: random_distribution=str:float[,str:float][,str:float]
1223 By default, fio will use a completely uniform random distribution when asked
1224 to perform random I/O. Sometimes it is useful to skew the distribution in
1225 specific ways, ensuring that some parts of the data is more hot than others.
1226 fio includes the following distribution models:
1229 Uniform random distribution
1238 Normal (Gaussian) distribution
1241 Zoned random distribution
1243 When using a **zipf** or **pareto** distribution, an input value is also
1244 needed to define the access pattern. For **zipf**, this is the `Zipf
1245 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
1246 program, :command:`fio-genzipf`, that can be used visualize what the given input
1247 values will yield in terms of hit rates. If you wanted to use **zipf** with
1248 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1249 option. If a non-uniform model is used, fio will disable use of the random
1250 map. For the **normal** distribution, a normal (Gaussian) deviation is
1251 supplied as a value between 0 and 100.
1253 For a **zoned** distribution, fio supports specifying percentages of I/O
1254 access that should fall within what range of the file or device. For
1255 example, given a criteria of:
1257 * 60% of accesses should be to the first 10%
1258 * 30% of accesses should be to the next 20%
1259 * 8% of accesses should be to the next 30%
1260 * 2% of accesses should be to the next 40%
1262 we can define that through zoning of the random accesses. For the above
1263 example, the user would do::
1265 random_distribution=zoned:60/10:30/20:8/30:2/40
1267 similarly to how :option:`bssplit` works for setting ranges and percentages
1268 of block sizes. Like :option:`bssplit`, it's possible to specify separate
1269 zones for reads, writes, and trims. If just one set is given, it'll apply to
1272 .. option:: percentage_random=int[,int][,int]
1274 For a random workload, set how big a percentage should be random. This
1275 defaults to 100%, in which case the workload is fully random. It can be set
1276 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1277 sequential. Any setting in between will result in a random mix of sequential
1278 and random I/O, at the given percentages. Comma-separated values may be
1279 specified for reads, writes, and trims as described in :option:`blocksize`.
1281 .. option:: norandommap
1283 Normally fio will cover every block of the file when doing random I/O. If
1284 this option is given, fio will just get a new random offset without looking
1285 at past I/O history. This means that some blocks may not be read or written,
1286 and that some blocks may be read/written more than once. If this option is
1287 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1288 only intact blocks are verified, i.e., partially-overwritten blocks are
1291 .. option:: softrandommap=bool
1293 See :option:`norandommap`. If fio runs with the random block map enabled and
1294 it fails to allocate the map, if this option is set it will continue without
1295 a random block map. As coverage will not be as complete as with random maps,
1296 this option is disabled by default.
1298 .. option:: random_generator=str
1300 Fio supports the following engines for generating I/O offsets for random I/O:
1303 Strong 2^88 cycle random number generator.
1305 Linear feedback shift register generator.
1307 Strong 64-bit 2^258 cycle random number generator.
1309 **tausworthe** is a strong random number generator, but it requires tracking
1310 on the side if we want to ensure that blocks are only read or written
1311 once. **lfsr** guarantees that we never generate the same offset twice, and
1312 it's also less computationally expensive. It's not a true random generator,
1313 however, though for I/O purposes it's typically good enough. **lfsr** only
1314 works with single block sizes, not with workloads that use multiple block
1315 sizes. If used with such a workload, fio may read or write some blocks
1316 multiple times. The default value is **tausworthe**, unless the required
1317 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1318 selected automatically.
1324 .. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1326 The block size in bytes used for I/O units. Default: 4096. A single value
1327 applies to reads, writes, and trims. Comma-separated values may be
1328 specified for reads, writes, and trims. A value not terminated in a comma
1329 applies to subsequent types.
1334 means 256k for reads, writes and trims.
1337 means 8k for reads, 32k for writes and trims.
1340 means 8k for reads, 32k for writes, and default for trims.
1343 means default for reads, 8k for writes and trims.
1346 means default for reads, 8k for writes, and default for trims.
1348 .. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1350 A range of block sizes in bytes for I/O units. The issued I/O unit will
1351 always be a multiple of the minimum size, unless
1352 :option:`blocksize_unaligned` is set.
1354 Comma-separated ranges may be specified for reads, writes, and trims as
1355 described in :option:`blocksize`.
1357 Example: ``bsrange=1k-4k,2k-8k``.
1359 .. option:: bssplit=str[,str][,str]
1361 Sometimes you want even finer grained control of the block sizes issued, not
1362 just an even split between them. This option allows you to weight various
1363 block sizes, so that you are able to define a specific amount of block sizes
1364 issued. The format for this option is::
1366 bssplit=blocksize/percentage:blocksize/percentage
1368 for as many block sizes as needed. So if you want to define a workload that
1369 has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would write::
1371 bssplit=4k/10:64k/50:32k/40
1373 Ordering does not matter. If the percentage is left blank, fio will fill in
1374 the remaining values evenly. So a bssplit option like this one::
1376 bssplit=4k/50:1k/:32k/
1378 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always add up
1379 to 100, if bssplit is given a range that adds up to more, it will error out.
1381 Comma-separated values may be specified for reads, writes, and trims as
1382 described in :option:`blocksize`.
1384 If you want a workload that has 50% 2k reads and 50% 4k reads, while having
1385 90% 4k writes and 10% 8k writes, you would specify::
1387 bssplit=2k/50:4k/50,4k/90,8k/10
1389 .. option:: blocksize_unaligned, bs_unaligned
1391 If set, fio will issue I/O units with any size within
1392 :option:`blocksize_range`, not just multiples of the minimum size. This
1393 typically won't work with direct I/O, as that normally requires sector
1396 .. option:: bs_is_seq_rand=bool
1398 If this option is set, fio will use the normal read,write blocksize settings
1399 as sequential,random blocksize settings instead. Any random read or write
1400 will use the WRITE blocksize settings, and any sequential read or write will
1401 use the READ blocksize settings.
1403 .. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1405 Boundary to which fio will align random I/O units. Default:
1406 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1407 I/O, though it usually depends on the hardware block size. This option is
1408 mutually exclusive with using a random map for files, so it will turn off
1409 that option. Comma-separated values may be specified for reads, writes, and
1410 trims as described in :option:`blocksize`.
1416 .. option:: zero_buffers
1418 Initialize buffers with all zeros. Default: fill buffers with random data.
1420 .. option:: refill_buffers
1422 If this option is given, fio will refill the I/O buffers on every
1423 submit. The default is to only fill it at init time and reuse that
1424 data. Only makes sense if zero_buffers isn't specified, naturally. If data
1425 verification is enabled, `refill_buffers` is also automatically enabled.
1427 .. option:: scramble_buffers=bool
1429 If :option:`refill_buffers` is too costly and the target is using data
1430 deduplication, then setting this option will slightly modify the I/O buffer
1431 contents to defeat normal de-dupe attempts. This is not enough to defeat
1432 more clever block compression attempts, but it will stop naive dedupe of
1433 blocks. Default: true.
1435 .. option:: buffer_compress_percentage=int
1437 If this is set, then fio will attempt to provide I/O buffer content (on
1438 WRITEs) that compresses to the specified level. Fio does this by providing a
1439 mix of random data and a fixed pattern. The fixed pattern is either zeros,
1440 or the pattern specified by :option:`buffer_pattern`. If the pattern option
1441 is used, it might skew the compression ratio slightly. Note that this is per
1442 block size unit, for file/disk wide compression level that matches this
1443 setting, you'll also want to set :option:`refill_buffers`.
1445 .. option:: buffer_compress_chunk=int
1447 See :option:`buffer_compress_percentage`. This setting allows fio to manage
1448 how big the ranges of random data and zeroed data is. Without this set, fio
1449 will provide :option:`buffer_compress_percentage` of blocksize random data,
1450 followed by the remaining zeroed. With this set to some chunk size smaller
1451 than the block size, fio can alternate random and zeroed data throughout the
1454 .. option:: buffer_pattern=str
1456 If set, fio will fill the I/O buffers with this pattern or with the contents
1457 of a file. If not set, the contents of I/O buffers are defined by the other
1458 options related to buffer contents. The setting can be any pattern of bytes,
1459 and can be prefixed with 0x for hex values. It may also be a string, where
1460 the string must then be wrapped with ``""``. Or it may also be a filename,
1461 where the filename must be wrapped with ``''`` in which case the file is
1462 opened and read. Note that not all the file contents will be read if that
1463 would cause the buffers to overflow. So, for example::
1465 buffer_pattern='filename'
1469 buffer_pattern="abcd"
1477 buffer_pattern=0xdeadface
1479 Also you can combine everything together in any order::
1481 buffer_pattern=0xdeadface"abcd"-12'filename'
1483 .. option:: dedupe_percentage=int
1485 If set, fio will generate this percentage of identical buffers when
1486 writing. These buffers will be naturally dedupable. The contents of the
1487 buffers depend on what other buffer compression settings have been set. It's
1488 possible to have the individual buffers either fully compressible, or not at
1489 all. This option only controls the distribution of unique buffers.
1491 .. option:: invalidate=bool
1493 Invalidate the buffer/page cache parts of the files to be used prior to
1494 starting I/O if the platform and file type support it. Defaults to true.
1495 This will be ignored if :option:`pre_read` is also specified for the
1498 .. option:: sync=bool
1500 Use synchronous I/O for buffered writes. For the majority of I/O engines,
1501 this means using O_SYNC. Default: false.
1503 .. option:: iomem=str, mem=str
1505 Fio can use various types of memory as the I/O unit buffer. The allowed
1509 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1513 Use shared memory as the buffers. Allocated through
1514 :manpage:`shmget(2)`.
1517 Same as shm, but use huge pages as backing.
1520 Use :manpage:`mmap(2)` to allocate buffers. May either be anonymous memory, or can
1521 be file backed if a filename is given after the option. The format
1522 is `mem=mmap:/path/to/file`.
1525 Use a memory mapped huge file as the buffer backing. Append filename
1526 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1529 Same as mmap, but use a MMAP_SHARED mapping.
1532 Use GPU memory as the buffers for GPUDirect RDMA benchmark.
1533 The :option:`ioengine` must be `rdma`.
1535 The area allocated is a function of the maximum allowed bs size for the job,
1536 multiplied by the I/O depth given. Note that for **shmhuge** and
1537 **mmaphuge** to work, the system must have free huge pages allocated. This
1538 can normally be checked and set by reading/writing
1539 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1540 is 4MiB in size. So to calculate the number of huge pages you need for a
1541 given job file, add up the I/O depth of all jobs (normally one unless
1542 :option:`iodepth` is used) and multiply by the maximum bs set. Then divide
1543 that number by the huge page size. You can see the size of the huge pages in
1544 :file:`/proc/meminfo`. If no huge pages are allocated by having a non-zero
1545 number in `nr_hugepages`, using **mmaphuge** or **shmhuge** will fail. Also
1546 see :option:`hugepage-size`.
1548 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1549 should point there. So if it's mounted in :file:`/huge`, you would use
1550 `mem=mmaphuge:/huge/somefile`.
1552 .. option:: iomem_align=int, mem_align=int
1554 This indicates the memory alignment of the I/O memory buffers. Note that
1555 the given alignment is applied to the first I/O unit buffer, if using
1556 :option:`iodepth` the alignment of the following buffers are given by the
1557 :option:`bs` used. In other words, if using a :option:`bs` that is a
1558 multiple of the page sized in the system, all buffers will be aligned to
1559 this value. If using a :option:`bs` that is not page aligned, the alignment
1560 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1563 .. option:: hugepage-size=int
1565 Defines the size of a huge page. Must at least be equal to the system
1566 setting, see :file:`/proc/meminfo`. Defaults to 4MiB. Should probably
1567 always be a multiple of megabytes, so using ``hugepage-size=Xm`` is the
1568 preferred way to set this to avoid setting a non-pow-2 bad value.
1570 .. option:: lockmem=int
1572 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1573 simulate a smaller amount of memory. The amount specified is per worker.
1579 .. option:: size=int
1581 The total size of file I/O for each thread of this job. Fio will run until
1582 this many bytes has been transferred, unless runtime is limited by other options
1583 (such as :option:`runtime`, for instance, or increased/decreased by :option:`io_size`).
1584 Fio will divide this size between the available files determined by options
1585 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1586 specified by the job. If the result of division happens to be 0, the size is
1587 set to the physical size of the given files or devices if they exist.
1588 If this option is not specified, fio will use the full size of the given
1589 files or devices. If the files do not exist, size must be given. It is also
1590 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1591 given, fio will use 20% of the full size of the given files or devices.
1592 Can be combined with :option:`offset` to constrain the start and end range
1593 that I/O will be done within.
1595 .. option:: io_size=int, io_limit=int
1597 Normally fio operates within the region set by :option:`size`, which means
1598 that the :option:`size` option sets both the region and size of I/O to be
1599 performed. Sometimes that is not what you want. With this option, it is
1600 possible to define just the amount of I/O that fio should do. For instance,
1601 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1602 will perform I/O within the first 20GiB but exit when 5GiB have been
1603 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1604 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1605 the 0..20GiB region.
1607 .. option:: filesize=irange(int)
1609 Individual file sizes. May be a range, in which case fio will select sizes
1610 for files at random within the given range and limited to :option:`size` in
1611 total (if that is given). If not given, each created file is the same size.
1612 This option overrides :option:`size` in terms of file size, which means
1613 this value is used as a fixed size or possible range of each file.
1615 .. option:: file_append=bool
1617 Perform I/O after the end of the file. Normally fio will operate within the
1618 size of a file. If this option is set, then fio will append to the file
1619 instead. This has identical behavior to setting :option:`offset` to the size
1620 of a file. This option is ignored on non-regular files.
1622 .. option:: fill_device=bool, fill_fs=bool
1624 Sets size to something really large and waits for ENOSPC (no space left on
1625 device) as the terminating condition. Only makes sense with sequential
1626 write. For a read workload, the mount point will be filled first then I/O
1627 started on the result. This option doesn't make sense if operating on a raw
1628 device node, since the size of that is already known by the file system.
1629 Additionally, writing beyond end-of-device will not return ENOSPC there.
1635 .. option:: ioengine=str
1637 Defines how the job issues I/O to the file. The following types are defined:
1640 Basic :manpage:`read(2)` or :manpage:`write(2)`
1641 I/O. :manpage:`lseek(2)` is used to position the I/O location.
1642 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
1645 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1646 all supported operating systems except for Windows.
1649 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
1650 queuing by coalescing adjacent I/Os into a single submission.
1653 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1656 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1659 Linux native asynchronous I/O. Note that Linux may only support
1660 queued behavior with non-buffered I/O (set ``direct=1`` or
1662 This engine defines engine specific options.
1665 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
1666 :manpage:`aio_write(3)`.
1669 Solaris native asynchronous I/O.
1672 Windows native asynchronous I/O. Default on Windows.
1675 File is memory mapped with :manpage:`mmap(2)` and data copied
1676 to/from using :manpage:`memcpy(3)`.
1679 :manpage:`splice(2)` is used to transfer the data and
1680 :manpage:`vmsplice(2)` to transfer data from user space to the
1684 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
1685 ioctl, or if the target is an sg character device we use
1686 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
1687 I/O. Requires :option:`filename` option to specify either block or
1691 Doesn't transfer any data, just pretends to. This is mainly used to
1692 exercise fio itself and for debugging/testing purposes.
1695 Transfer over the network to given ``host:port``. Depending on the
1696 :option:`protocol` used, the :option:`hostname`, :option:`port`,
1697 :option:`listen` and :option:`filename` options are used to specify
1698 what sort of connection to make, while the :option:`protocol` option
1699 determines which protocol will be used. This engine defines engine
1703 Like **net**, but uses :manpage:`splice(2)` and
1704 :manpage:`vmsplice(2)` to map data and send/receive.
1705 This engine defines engine specific options.
1708 Doesn't transfer any data, but burns CPU cycles according to the
1709 :option:`cpuload` and :option:`cpuchunks` options. Setting
1710 :option:`cpuload`\=85 will cause that job to do nothing but burn 85%
1711 of the CPU. In case of SMP machines, use :option:`numjobs`=<nr_of_cpu>
1712 to get desired CPU usage, as the cpuload only loads a
1713 single CPU at the desired rate. A job never finishes unless there is
1714 at least one non-cpuio job.
1717 The GUASI I/O engine is the Generic Userspace Asyncronous Syscall
1718 Interface approach to async I/O. See
1720 http://www.xmailserver.org/guasi-lib.html
1722 for more info on GUASI.
1725 The RDMA I/O engine supports both RDMA memory semantics
1726 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
1727 InfiniBand, RoCE and iWARP protocols.
1730 I/O engine that does regular fallocate to simulate data transfer as
1734 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
1737 does fallocate(,mode = 0).
1740 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
1743 I/O engine that sends :manpage:`ftruncate(2)` operations in response
1744 to write (DDIR_WRITE) events. Each ftruncate issued sets the file's
1745 size to the current block offset. :option:`blocksize` is ignored.
1748 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
1749 defragment activity in request to DDIR_WRITE event.
1752 I/O engine supporting direct access to Ceph Rados Block Devices
1753 (RBD) via librbd without the need to use the kernel rbd driver. This
1754 ioengine defines engine specific options.
1757 Using GlusterFS libgfapi sync interface to direct access to
1758 GlusterFS volumes without having to go through FUSE. This ioengine
1759 defines engine specific options.
1762 Using GlusterFS libgfapi async interface to direct access to
1763 GlusterFS volumes without having to go through FUSE. This ioengine
1764 defines engine specific options.
1767 Read and write through Hadoop (HDFS). The :option:`filename` option
1768 is used to specify host,port of the hdfs name-node to connect. This
1769 engine interprets offsets a little differently. In HDFS, files once
1770 created cannot be modified so random writes are not possible. To
1771 imitate this the libhdfs engine expects a bunch of small files to be
1772 created over HDFS and will randomly pick a file from them
1773 based on the offset generated by fio backend (see the example
1774 job file to create such files, use ``rw=write`` option). Please
1775 note, it may be necessary to set environment variables to work
1776 with HDFS/libhdfs properly. Each job uses its own connection to
1780 Read, write and erase an MTD character device (e.g.,
1781 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
1782 underlying device type, the I/O may have to go in a certain pattern,
1783 e.g., on NAND, writing sequentially to erase blocks and discarding
1784 before overwriting. The `trimwrite` mode works well for this
1788 Read and write using filesystem DAX to a file on a filesystem
1789 mounted with DAX on a persistent memory device through the NVML
1793 Read and write using device DAX to a persistent memory device (e.g.,
1794 /dev/dax0.0) through the NVML libpmem library.
1797 Prefix to specify loading an external I/O engine object file. Append
1798 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
1799 ioengine :file:`foo.o` in :file:`/tmp`. The path can be either
1800 absolute or relative. See :file:`engines/skeleton_external.c` for
1801 details of writing an external I/O engine.
1804 Simply create the files and do no IO to them. You still need to
1805 set `filesize` so that all the accounting still occurs, but no
1806 actual IO will be done other than creating the file.
1808 I/O engine specific parameters
1809 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1811 In addition, there are some parameters which are only valid when a specific
1812 :option:`ioengine` is in use. These are used identically to normal parameters,
1813 with the caveat that when used on the command line, they must come after the
1814 :option:`ioengine` that defines them is selected.
1816 .. option:: userspace_reap : [libaio]
1818 Normally, with the libaio engine in use, fio will use the
1819 :manpage:`io_getevents(2)` system call to reap newly returned events. With
1820 this flag turned on, the AIO ring will be read directly from user-space to
1821 reap events. The reaping mode is only enabled when polling for a minimum of
1822 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
1824 .. option:: hipri : [pvsync2]
1826 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
1829 .. option:: hipri_percentage : [pvsync2]
1831 When hipri is set this determines the probability of a pvsync2 I/O being high
1832 priority. The default is 100%.
1834 .. option:: cpuload=int : [cpuio]
1836 Attempt to use the specified percentage of CPU cycles. This is a mandatory
1837 option when using cpuio I/O engine.
1839 .. option:: cpuchunks=int : [cpuio]
1841 Split the load into cycles of the given time. In microseconds.
1843 .. option:: exit_on_io_done=bool : [cpuio]
1845 Detect when I/O threads are done, then exit.
1847 .. option:: namenode=str : [libhdfs]
1849 The hostname or IP address of a HDFS cluster namenode to contact.
1851 .. option:: port=int
1855 The listening port of the HFDS cluster namenode.
1859 The TCP or UDP port to bind to or connect to. If this is used with
1860 :option:`numjobs` to spawn multiple instances of the same job type, then
1861 this will be the starting port number since fio will use a range of
1864 .. option:: hostname=str : [netsplice] [net]
1866 The hostname or IP address to use for TCP or UDP based I/O. If the job is
1867 a TCP listener or UDP reader, the hostname is not used and must be omitted
1868 unless it is a valid UDP multicast address.
1870 .. option:: interface=str : [netsplice] [net]
1872 The IP address of the network interface used to send or receive UDP
1875 .. option:: ttl=int : [netsplice] [net]
1877 Time-to-live value for outgoing UDP multicast packets. Default: 1.
1879 .. option:: nodelay=bool : [netsplice] [net]
1881 Set TCP_NODELAY on TCP connections.
1883 .. option:: protocol=str, proto=str : [netsplice] [net]
1885 The network protocol to use. Accepted values are:
1888 Transmission control protocol.
1890 Transmission control protocol V6.
1892 User datagram protocol.
1894 User datagram protocol V6.
1898 When the protocol is TCP or UDP, the port must also be given, as well as the
1899 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
1900 normal :option:`filename` option should be used and the port is invalid.
1902 .. option:: listen : [netsplice] [net]
1904 For TCP network connections, tell fio to listen for incoming connections
1905 rather than initiating an outgoing connection. The :option:`hostname` must
1906 be omitted if this option is used.
1908 .. option:: pingpong : [netsplice] [net]
1910 Normally a network writer will just continue writing data, and a network
1911 reader will just consume packages. If ``pingpong=1`` is set, a writer will
1912 send its normal payload to the reader, then wait for the reader to send the
1913 same payload back. This allows fio to measure network latencies. The
1914 submission and completion latencies then measure local time spent sending or
1915 receiving, and the completion latency measures how long it took for the
1916 other end to receive and send back. For UDP multicast traffic
1917 ``pingpong=1`` should only be set for a single reader when multiple readers
1918 are listening to the same address.
1920 .. option:: window_size : [netsplice] [net]
1922 Set the desired socket buffer size for the connection.
1924 .. option:: mss : [netsplice] [net]
1926 Set the TCP maximum segment size (TCP_MAXSEG).
1928 .. option:: donorname=str : [e4defrag]
1930 File will be used as a block donor (swap extents between files).
1932 .. option:: inplace=int : [e4defrag]
1934 Configure donor file blocks allocation strategy:
1937 Default. Preallocate donor's file on init.
1939 Allocate space immediately inside defragment event, and free right
1942 .. option:: clustername=str : [rbd]
1944 Specifies the name of the Ceph cluster.
1946 .. option:: rbdname=str : [rbd]
1948 Specifies the name of the RBD.
1950 .. option:: pool=str : [rbd]
1952 Specifies the name of the Ceph pool containing RBD.
1954 .. option:: clientname=str : [rbd]
1956 Specifies the username (without the 'client.' prefix) used to access the
1957 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
1958 the full *type.id* string. If no type. prefix is given, fio will add
1959 'client.' by default.
1961 .. option:: skip_bad=bool : [mtd]
1963 Skip operations against known bad blocks.
1965 .. option:: hdfsdirectory : [libhdfs]
1967 libhdfs will create chunk in this HDFS directory.
1969 .. option:: chunk_size : [libhdfs]
1971 The size of the chunk to use for each file.
1977 .. option:: iodepth=int
1979 Number of I/O units to keep in flight against the file. Note that
1980 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
1981 for small degrees when :option:`verify_async` is in use). Even async
1982 engines may impose OS restrictions causing the desired depth not to be
1983 achieved. This may happen on Linux when using libaio and not setting
1984 :option:`direct`\=1, since buffered I/O is not async on that OS. Keep an
1985 eye on the I/O depth distribution in the fio output to verify that the
1986 achieved depth is as expected. Default: 1.
1988 .. option:: iodepth_batch_submit=int, iodepth_batch=int
1990 This defines how many pieces of I/O to submit at once. It defaults to 1
1991 which means that we submit each I/O as soon as it is available, but can be
1992 raised to submit bigger batches of I/O at the time. If it is set to 0 the
1993 :option:`iodepth` value will be used.
1995 .. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
1997 This defines how many pieces of I/O to retrieve at once. It defaults to 1
1998 which means that we'll ask for a minimum of 1 I/O in the retrieval process
1999 from the kernel. The I/O retrieval will go on until we hit the limit set by
2000 :option:`iodepth_low`. If this variable is set to 0, then fio will always
2001 check for completed events before queuing more I/O. This helps reduce I/O
2002 latency, at the cost of more retrieval system calls.
2004 .. option:: iodepth_batch_complete_max=int
2006 This defines maximum pieces of I/O to retrieve at once. This variable should
2007 be used along with :option:`iodepth_batch_complete_min`\=int variable,
2008 specifying the range of min and max amount of I/O which should be
2009 retrieved. By default it is equal to the :option:`iodepth_batch_complete_min`
2014 iodepth_batch_complete_min=1
2015 iodepth_batch_complete_max=<iodepth>
2017 which means that we will retrieve at least 1 I/O and up to the whole
2018 submitted queue depth. If none of I/O has been completed yet, we will wait.
2022 iodepth_batch_complete_min=0
2023 iodepth_batch_complete_max=<iodepth>
2025 which means that we can retrieve up to the whole submitted queue depth, but
2026 if none of I/O has been completed yet, we will NOT wait and immediately exit
2027 the system call. In this example we simply do polling.
2029 .. option:: iodepth_low=int
2031 The low water mark indicating when to start filling the queue
2032 again. Defaults to the same as :option:`iodepth`, meaning that fio will
2033 attempt to keep the queue full at all times. If :option:`iodepth` is set to
2034 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
2035 16 requests, it will let the depth drain down to 4 before starting to fill
2038 .. option:: serialize_overlap=bool
2040 Serialize in-flight I/Os that might otherwise cause or suffer from data races.
2041 When two or more I/Os are submitted simultaneously, there is no guarantee that
2042 the I/Os will be processed or completed in the submitted order. Further, if
2043 two or more of those I/Os are writes, any overlapping region between them can
2044 become indeterminate/undefined on certain storage. These issues can cause
2045 verification to fail erratically when at least one of the racing I/Os is
2046 changing data and the overlapping region has a non-zero size. Setting
2047 ``serialize_overlap`` tells fio to avoid provoking this behavior by explicitly
2048 serializing in-flight I/Os that have a non-zero overlap. Note that setting
2049 this option can reduce both performance and the `:option:iodepth` achieved.
2050 Additionally this option does not work when :option:`io_submit_mode` is set to
2051 offload. Default: false.
2053 .. option:: io_submit_mode=str
2055 This option controls how fio submits the I/O to the I/O engine. The default
2056 is `inline`, which means that the fio job threads submit and reap I/O
2057 directly. If set to `offload`, the job threads will offload I/O submission
2058 to a dedicated pool of I/O threads. This requires some coordination and thus
2059 has a bit of extra overhead, especially for lower queue depth I/O where it
2060 can increase latencies. The benefit is that fio can manage submission rates
2061 independently of the device completion rates. This avoids skewed latency
2062 reporting if I/O gets backed up on the device side (the coordinated omission
2069 .. option:: thinktime=time
2071 Stall the job for the specified period of time after an I/O has completed before issuing the
2072 next. May be used to simulate processing being done by an application.
2073 When the unit is omitted, the value is interpreted in microseconds. See
2074 :option:`thinktime_blocks` and :option:`thinktime_spin`.
2076 .. option:: thinktime_spin=time
2078 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
2079 something with the data received, before falling back to sleeping for the
2080 rest of the period specified by :option:`thinktime`. When the unit is
2081 omitted, the value is interpreted in microseconds.
2083 .. option:: thinktime_blocks=int
2085 Only valid if :option:`thinktime` is set - control how many blocks to issue,
2086 before waiting :option:`thinktime` usecs. If not set, defaults to 1 which will make
2087 fio wait :option:`thinktime` usecs after every block. This effectively makes any
2088 queue depth setting redundant, since no more than 1 I/O will be queued
2089 before we have to complete it and do our :option:`thinktime`. In other words, this
2090 setting effectively caps the queue depth if the latter is larger.
2092 .. option:: rate=int[,int][,int]
2094 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
2095 suffix rules apply. Comma-separated values may be specified for reads,
2096 writes, and trims as described in :option:`blocksize`.
2098 For example, using `rate=1m,500k` would limit reads to 1MiB/sec and writes to
2099 500KiB/sec. Capping only reads or writes can be done with `rate=,500k` or
2100 `rate=500k,` where the former will only limit writes (to 500KiB/sec) and the
2101 latter will only limit reads.
2103 .. option:: rate_min=int[,int][,int]
2105 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
2106 to meet this requirement will cause the job to exit. Comma-separated values
2107 may be specified for reads, writes, and trims as described in
2108 :option:`blocksize`.
2110 .. option:: rate_iops=int[,int][,int]
2112 Cap the bandwidth to this number of IOPS. Basically the same as
2113 :option:`rate`, just specified independently of bandwidth. If the job is
2114 given a block size range instead of a fixed value, the smallest block size
2115 is used as the metric. Comma-separated values may be specified for reads,
2116 writes, and trims as described in :option:`blocksize`.
2118 .. option:: rate_iops_min=int[,int][,int]
2120 If fio doesn't meet this rate of I/O, it will cause the job to exit.
2121 Comma-separated values may be specified for reads, writes, and trims as
2122 described in :option:`blocksize`.
2124 .. option:: rate_process=str
2126 This option controls how fio manages rated I/O submissions. The default is
2127 `linear`, which submits I/O in a linear fashion with fixed delays between
2128 I/Os that gets adjusted based on I/O completion rates. If this is set to
2129 `poisson`, fio will submit I/O based on a more real world random request
2130 flow, known as the Poisson process
2131 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
2132 10^6 / IOPS for the given workload.
2138 .. option:: latency_target=time
2140 If set, fio will attempt to find the max performance point that the given
2141 workload will run at while maintaining a latency below this target. When
2142 the unit is omitted, the value is interpreted in microseconds. See
2143 :option:`latency_window` and :option:`latency_percentile`.
2145 .. option:: latency_window=time
2147 Used with :option:`latency_target` to specify the sample window that the job
2148 is run at varying queue depths to test the performance. When the unit is
2149 omitted, the value is interpreted in microseconds.
2151 .. option:: latency_percentile=float
2153 The percentage of I/Os that must fall within the criteria specified by
2154 :option:`latency_target` and :option:`latency_window`. If not set, this
2155 defaults to 100.0, meaning that all I/Os must be equal or below to the value
2156 set by :option:`latency_target`.
2158 .. option:: max_latency=time
2160 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
2161 maximum latency. When the unit is omitted, the value is interpreted in
2164 .. option:: rate_cycle=int
2166 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
2167 of milliseconds. Defaults to 1000.
2173 .. option:: write_iolog=str
2175 Write the issued I/O patterns to the specified file. See
2176 :option:`read_iolog`. Specify a separate file for each job, otherwise the
2177 iologs will be interspersed and the file may be corrupt.
2179 .. option:: read_iolog=str
2181 Open an iolog with the specified filename and replay the I/O patterns it
2182 contains. This can be used to store a workload and replay it sometime
2183 later. The iolog given may also be a blktrace binary file, which allows fio
2184 to replay a workload captured by :command:`blktrace`. See
2185 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
2186 replay, the file needs to be turned into a blkparse binary data file first
2187 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
2189 .. option:: replay_no_stall=bool
2191 When replaying I/O with :option:`read_iolog` the default behavior is to
2192 attempt to respect the timestamps within the log and replay them with the
2193 appropriate delay between IOPS. By setting this variable fio will not
2194 respect the timestamps and attempt to replay them as fast as possible while
2195 still respecting ordering. The result is the same I/O pattern to a given
2196 device, but different timings.
2198 .. option:: replay_redirect=str
2200 While replaying I/O patterns using :option:`read_iolog` the default behavior
2201 is to replay the IOPS onto the major/minor device that each IOP was recorded
2202 from. This is sometimes undesirable because on a different machine those
2203 major/minor numbers can map to a different device. Changing hardware on the
2204 same system can also result in a different major/minor mapping.
2205 ``replay_redirect`` causes all I/Os to be replayed onto the single specified
2206 device regardless of the device it was recorded
2207 from. i.e. :option:`replay_redirect`\= :file:`/dev/sdc` would cause all I/O
2208 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
2209 multiple devices will be replayed onto a single device, if the trace
2210 contains multiple devices. If you want multiple devices to be replayed
2211 concurrently to multiple redirected devices you must blkparse your trace
2212 into separate traces and replay them with independent fio invocations.
2213 Unfortunately this also breaks the strict time ordering between multiple
2216 .. option:: replay_align=int
2218 Force alignment of I/O offsets and lengths in a trace to this power of 2
2221 .. option:: replay_scale=int
2223 Scale sector offsets down by this factor when replaying traces.
2226 Threads, processes and job synchronization
2227 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2231 Fio defaults to creating jobs by using fork, however if this option is
2232 given, fio will create jobs by using POSIX Threads' function
2233 :manpage:`pthread_create(3)` to create threads instead.
2235 .. option:: wait_for=str
2237 If set, the current job won't be started until all workers of the specified
2238 waitee job are done.
2240 ``wait_for`` operates on the job name basis, so there are a few
2241 limitations. First, the waitee must be defined prior to the waiter job
2242 (meaning no forward references). Second, if a job is being referenced as a
2243 waitee, it must have a unique name (no duplicate waitees).
2245 .. option:: nice=int
2247 Run the job with the given nice value. See man :manpage:`nice(2)`.
2249 On Windows, values less than -15 set the process class to "High"; -1 through
2250 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
2253 .. option:: prio=int
2255 Set the I/O priority value of this job. Linux limits us to a positive value
2256 between 0 and 7, with 0 being the highest. See man
2257 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
2258 systems since meaning of priority may differ.
2260 .. option:: prioclass=int
2262 Set the I/O priority class. See man :manpage:`ionice(1)`.
2264 .. option:: cpumask=int
2266 Set the CPU affinity of this job. The parameter given is a bit mask of
2267 allowed CPUs the job may run on. So if you want the allowed CPUs to be 1
2268 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
2269 :manpage:`sched_setaffinity(2)`. This may not work on all supported
2270 operating systems or kernel versions. This option doesn't work well for a
2271 higher CPU count than what you can store in an integer mask, so it can only
2272 control cpus 1-32. For boxes with larger CPU counts, use
2273 :option:`cpus_allowed`.
2275 .. option:: cpus_allowed=str
2277 Controls the same options as :option:`cpumask`, but accepts a textual
2278 specification of the permitted CPUs instead. So to use CPUs 1 and 5 you
2279 would specify ``cpus_allowed=1,5``. This option also allows a range of CPUs
2280 to be specified -- say you wanted a binding to CPUs 1, 5, and 8 to 15, you
2281 would set ``cpus_allowed=1,5,8-15``.
2283 .. option:: cpus_allowed_policy=str
2285 Set the policy of how fio distributes the CPUs specified by
2286 :option:`cpus_allowed` or :option:`cpumask`. Two policies are supported:
2289 All jobs will share the CPU set specified.
2291 Each job will get a unique CPU from the CPU set.
2293 **shared** is the default behavior, if the option isn't specified. If
2294 **split** is specified, then fio will will assign one cpu per job. If not
2295 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
2298 .. option:: numa_cpu_nodes=str
2300 Set this job running on specified NUMA nodes' CPUs. The arguments allow
2301 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
2302 NUMA options support, fio must be built on a system with libnuma-dev(el)
2305 .. option:: numa_mem_policy=str
2307 Set this job's memory policy and corresponding NUMA nodes. Format of the
2312 ``mode`` is one of the following memory poicies: ``default``, ``prefer``,
2313 ``bind``, ``interleave`` or ``local``. For ``default`` and ``local`` memory
2314 policies, no node needs to be specified. For ``prefer``, only one node is
2315 allowed. For ``bind`` and ``interleave`` the ``nodelist`` may be as
2316 follows: a comma delimited list of numbers, A-B ranges, or `all`.
2318 .. option:: cgroup=str
2320 Add job to this control group. If it doesn't exist, it will be created. The
2321 system must have a mounted cgroup blkio mount point for this to work. If
2322 your system doesn't have it mounted, you can do so with::
2324 # mount -t cgroup -o blkio none /cgroup
2326 .. option:: cgroup_weight=int
2328 Set the weight of the cgroup to this value. See the documentation that comes
2329 with the kernel, allowed values are in the range of 100..1000.
2331 .. option:: cgroup_nodelete=bool
2333 Normally fio will delete the cgroups it has created after the job
2334 completion. To override this behavior and to leave cgroups around after the
2335 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
2336 to inspect various cgroup files after job completion. Default: false.
2338 .. option:: flow_id=int
2340 The ID of the flow. If not specified, it defaults to being a global
2341 flow. See :option:`flow`.
2343 .. option:: flow=int
2345 Weight in token-based flow control. If this value is used, then there is a
2346 'flow counter' which is used to regulate the proportion of activity between
2347 two or more jobs. Fio attempts to keep this flow counter near zero. The
2348 ``flow`` parameter stands for how much should be added or subtracted to the
2349 flow counter on each iteration of the main I/O loop. That is, if one job has
2350 ``flow=8`` and another job has ``flow=-1``, then there will be a roughly 1:8
2351 ratio in how much one runs vs the other.
2353 .. option:: flow_watermark=int
2355 The maximum value that the absolute value of the flow counter is allowed to
2356 reach before the job must wait for a lower value of the counter.
2358 .. option:: flow_sleep=int
2360 The period of time, in microseconds, to wait after the flow watermark has
2361 been exceeded before retrying operations.
2363 .. option:: stonewall, wait_for_previous
2365 Wait for preceding jobs in the job file to exit, before starting this
2366 one. Can be used to insert serialization points in the job file. A stone
2367 wall also implies starting a new reporting group, see
2368 :option:`group_reporting`.
2372 By default, fio will continue running all other jobs when one job finishes
2373 but sometimes this is not the desired action. Setting ``exitall`` will
2374 instead make fio terminate all other jobs when one job finishes.
2376 .. option:: exec_prerun=str
2378 Before running this job, issue the command specified through
2379 :manpage:`system(3)`. Output is redirected in a file called
2380 :file:`jobname.prerun.txt`.
2382 .. option:: exec_postrun=str
2384 After the job completes, issue the command specified though
2385 :manpage:`system(3)`. Output is redirected in a file called
2386 :file:`jobname.postrun.txt`.
2390 Instead of running as the invoking user, set the user ID to this value
2391 before the thread/process does any work.
2395 Set group ID, see :option:`uid`.
2401 .. option:: verify_only
2403 Do not perform specified workload, only verify data still matches previous
2404 invocation of this workload. This option allows one to check data multiple
2405 times at a later date without overwriting it. This option makes sense only
2406 for workloads that write data, and does not support workloads with the
2407 :option:`time_based` option set.
2409 .. option:: do_verify=bool
2411 Run the verify phase after a write phase. Only valid if :option:`verify` is
2414 .. option:: verify=str
2416 If writing to a file, fio can verify the file contents after each iteration
2417 of the job. Each verification method also implies verification of special
2418 header, which is written to the beginning of each block. This header also
2419 includes meta information, like offset of the block, block number, timestamp
2420 when block was written, etc. :option:`verify` can be combined with
2421 :option:`verify_pattern` option. The allowed values are:
2424 Use an md5 sum of the data area and store it in the header of
2428 Use an experimental crc64 sum of the data area and store it in the
2429 header of each block.
2432 Use a crc32c sum of the data area and store it in the header of
2433 each block. This will automatically use hardware acceleration
2434 (e.g. SSE4.2 on an x86 or CRC crypto extensions on ARM64) but will
2435 fall back to software crc32c if none is found. Generally the
2436 fatest checksum fio supports when hardware accelerated.
2442 Use a crc32 sum of the data area and store it in the header of each
2446 Use a crc16 sum of the data area and store it in the header of each
2450 Use a crc7 sum of the data area and store it in the header of each
2454 Use xxhash as the checksum function. Generally the fastest software
2455 checksum that fio supports.
2458 Use sha512 as the checksum function.
2461 Use sha256 as the checksum function.
2464 Use optimized sha1 as the checksum function.
2467 Use optimized sha3-224 as the checksum function.
2470 Use optimized sha3-256 as the checksum function.
2473 Use optimized sha3-384 as the checksum function.
2476 Use optimized sha3-512 as the checksum function.
2479 This option is deprecated, since now meta information is included in
2480 generic verification header and meta verification happens by
2481 default. For detailed information see the description of the
2482 :option:`verify` setting. This option is kept because of
2483 compatibility's sake with old configurations. Do not use it.
2486 Verify a strict pattern. Normally fio includes a header with some
2487 basic information and checksumming, but if this option is set, only
2488 the specific pattern set with :option:`verify_pattern` is verified.
2491 Only pretend to verify. Useful for testing internals with
2492 :option:`ioengine`\=null, not for much else.
2494 This option can be used for repeated burn-in tests of a system to make sure
2495 that the written data is also correctly read back. If the data direction
2496 given is a read or random read, fio will assume that it should verify a
2497 previously written file. If the data direction includes any form of write,
2498 the verify will be of the newly written data.
2500 .. option:: verifysort=bool
2502 If true, fio will sort written verify blocks when it deems it faster to read
2503 them back in a sorted manner. This is often the case when overwriting an
2504 existing file, since the blocks are already laid out in the file system. You
2505 can ignore this option unless doing huge amounts of really fast I/O where
2506 the red-black tree sorting CPU time becomes significant. Default: true.
2508 .. option:: verifysort_nr=int
2510 Pre-load and sort verify blocks for a read workload.
2512 .. option:: verify_offset=int
2514 Swap the verification header with data somewhere else in the block before
2515 writing. It is swapped back before verifying.
2517 .. option:: verify_interval=int
2519 Write the verification header at a finer granularity than the
2520 :option:`blocksize`. It will be written for chunks the size of
2521 ``verify_interval``. :option:`blocksize` should divide this evenly.
2523 .. option:: verify_pattern=str
2525 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
2526 filling with totally random bytes, but sometimes it's interesting to fill
2527 with a known pattern for I/O verification purposes. Depending on the width
2528 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time (it can
2529 be either a decimal or a hex number). The ``verify_pattern`` if larger than
2530 a 32-bit quantity has to be a hex number that starts with either "0x" or
2531 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
2532 format, which means that for each block offset will be written and then
2533 verified back, e.g.::
2537 Or use combination of everything::
2539 verify_pattern=0xff%o"abcd"-12
2541 .. option:: verify_fatal=bool
2543 Normally fio will keep checking the entire contents before quitting on a
2544 block verification failure. If this option is set, fio will exit the job on
2545 the first observed failure. Default: false.
2547 .. option:: verify_dump=bool
2549 If set, dump the contents of both the original data block and the data block
2550 we read off disk to files. This allows later analysis to inspect just what
2551 kind of data corruption occurred. Off by default.
2553 .. option:: verify_async=int
2555 Fio will normally verify I/O inline from the submitting thread. This option
2556 takes an integer describing how many async offload threads to create for I/O
2557 verification instead, causing fio to offload the duty of verifying I/O
2558 contents to one or more separate threads. If using this offload option, even
2559 sync I/O engines can benefit from using an :option:`iodepth` setting higher
2560 than 1, as it allows them to have I/O in flight while verifies are running.
2561 Defaults to 0 async threads, i.e. verification is not asynchronous.
2563 .. option:: verify_async_cpus=str
2565 Tell fio to set the given CPU affinity on the async I/O verification
2566 threads. See :option:`cpus_allowed` for the format used.
2568 .. option:: verify_backlog=int
2570 Fio will normally verify the written contents of a job that utilizes verify
2571 once that job has completed. In other words, everything is written then
2572 everything is read back and verified. You may want to verify continually
2573 instead for a variety of reasons. Fio stores the meta data associated with
2574 an I/O block in memory, so for large verify workloads, quite a bit of memory
2575 would be used up holding this meta data. If this option is enabled, fio will
2576 write only N blocks before verifying these blocks.
2578 .. option:: verify_backlog_batch=int
2580 Control how many blocks fio will verify if :option:`verify_backlog` is
2581 set. If not set, will default to the value of :option:`verify_backlog`
2582 (meaning the entire queue is read back and verified). If
2583 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
2584 blocks will be verified, if ``verify_backlog_batch`` is larger than
2585 :option:`verify_backlog`, some blocks will be verified more than once.
2587 .. option:: verify_state_save=bool
2589 When a job exits during the write phase of a verify workload, save its
2590 current state. This allows fio to replay up until that point, if the verify
2591 state is loaded for the verify read phase. The format of the filename is,
2594 <type>-<jobname>-<jobindex>-verify.state.
2596 <type> is "local" for a local run, "sock" for a client/server socket
2597 connection, and "ip" (192.168.0.1, for instance) for a networked
2598 client/server connection. Defaults to true.
2600 .. option:: verify_state_load=bool
2602 If a verify termination trigger was used, fio stores the current write state
2603 of each thread. This can be used at verification time so that fio knows how
2604 far it should verify. Without this information, fio will run a full
2605 verification pass, according to the settings in the job file used. Default
2608 .. option:: trim_percentage=int
2610 Number of verify blocks to discard/trim.
2612 .. option:: trim_verify_zero=bool
2614 Verify that trim/discarded blocks are returned as zeros.
2616 .. option:: trim_backlog=int
2618 Trim after this number of blocks are written.
2620 .. option:: trim_backlog_batch=int
2622 Trim this number of I/O blocks.
2624 .. option:: experimental_verify=bool
2626 Enable experimental verification.
2631 .. option:: steadystate=str:float, ss=str:float
2633 Define the criterion and limit for assessing steady state performance. The
2634 first parameter designates the criterion whereas the second parameter sets
2635 the threshold. When the criterion falls below the threshold for the
2636 specified duration, the job will stop. For example, `iops_slope:0.1%` will
2637 direct fio to terminate the job when the least squares regression slope
2638 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
2639 this will apply to all jobs in the group. Below is the list of available
2640 steady state assessment criteria. All assessments are carried out using only
2641 data from the rolling collection window. Threshold limits can be expressed
2642 as a fixed value or as a percentage of the mean in the collection window.
2645 Collect IOPS data. Stop the job if all individual IOPS measurements
2646 are within the specified limit of the mean IOPS (e.g., ``iops:2``
2647 means that all individual IOPS values must be within 2 of the mean,
2648 whereas ``iops:0.2%`` means that all individual IOPS values must be
2649 within 0.2% of the mean IOPS to terminate the job).
2652 Collect IOPS data and calculate the least squares regression
2653 slope. Stop the job if the slope falls below the specified limit.
2656 Collect bandwidth data. Stop the job if all individual bandwidth
2657 measurements are within the specified limit of the mean bandwidth.
2660 Collect bandwidth data and calculate the least squares regression
2661 slope. Stop the job if the slope falls below the specified limit.
2663 .. option:: steadystate_duration=time, ss_dur=time
2665 A rolling window of this duration will be used to judge whether steady state
2666 has been reached. Data will be collected once per second. The default is 0
2667 which disables steady state detection. When the unit is omitted, the
2668 value is interpreted in seconds.
2670 .. option:: steadystate_ramp_time=time, ss_ramp=time
2672 Allow the job to run for the specified duration before beginning data
2673 collection for checking the steady state job termination criterion. The
2674 default is 0. When the unit is omitted, the value is interpreted in seconds.
2677 Measurements and reporting
2678 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2680 .. option:: per_job_logs=bool
2682 If set, this generates bw/clat/iops log with per file private filenames. If
2683 not set, jobs with identical names will share the log filename. Default:
2686 .. option:: group_reporting
2688 It may sometimes be interesting to display statistics for groups of jobs as
2689 a whole instead of for each individual job. This is especially true if
2690 :option:`numjobs` is used; looking at individual thread/process output
2691 quickly becomes unwieldy. To see the final report per-group instead of
2692 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
2693 same reporting group, unless if separated by a :option:`stonewall`, or by
2694 using :option:`new_group`.
2696 .. option:: new_group
2698 Start a new reporting group. See: :option:`group_reporting`. If not given,
2699 all jobs in a file will be part of the same reporting group, unless
2700 separated by a :option:`stonewall`.
2702 .. option:: stats=bool
2704 By default, fio collects and shows final output results for all jobs
2705 that run. If this option is set to 0, then fio will ignore it in
2706 the final stat output.
2708 .. option:: write_bw_log=str
2710 If given, write a bandwidth log for this job. Can be used to store data of
2711 the bandwidth of the jobs in their lifetime. The included
2712 :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
2713 text files into nice graphs. See :option:`write_lat_log` for behavior of
2714 given filename. For this option, the postfix is :file:`_bw.x.log`, where `x`
2715 is the index of the job (`1..N`, where `N` is the number of jobs). If
2716 :option:`per_job_logs` is false, then the filename will not include the job
2717 index. See `Log File Formats`_.
2719 .. option:: write_lat_log=str
2721 Same as :option:`write_bw_log`, except that this option stores I/O
2722 submission, completion, and total latencies instead. If no filename is given
2723 with this option, the default filename of :file:`jobname_type.log` is
2724 used. Even if the filename is given, fio will still append the type of
2725 log. So if one specifies::
2729 The actual log names will be :file:`foo_slat.x.log`, :file:`foo_clat.x.log`,
2730 and :file:`foo_lat.x.log`, where `x` is the index of the job (`1..N`, where `N`
2731 is the number of jobs). This helps :command:`fio_generate_plots` find the
2732 logs automatically. If :option:`per_job_logs` is false, then the filename
2733 will not include the job index. See `Log File Formats`_.
2735 .. option:: write_hist_log=str
2737 Same as :option:`write_lat_log`, but writes I/O completion latency
2738 histograms. If no filename is given with this option, the default filename
2739 of :file:`jobname_clat_hist.x.log` is used, where `x` is the index of the
2740 job (`1..N`, where `N` is the number of jobs). Even if the filename is given,
2741 fio will still append the type of log. If :option:`per_job_logs` is false,
2742 then the filename will not include the job index. See `Log File Formats`_.
2744 .. option:: write_iops_log=str
2746 Same as :option:`write_bw_log`, but writes IOPS. If no filename is given
2747 with this option, the default filename of :file:`jobname_type.x.log` is
2748 used, where `x` is the index of the job (`1..N`, where `N` is the number of
2749 jobs). Even if the filename is given, fio will still append the type of
2750 log. If :option:`per_job_logs` is false, then the filename will not include
2751 the job index. See `Log File Formats`_.
2753 .. option:: log_avg_msec=int
2755 By default, fio will log an entry in the iops, latency, or bw log for every
2756 I/O that completes. When writing to the disk log, that can quickly grow to a
2757 very large size. Setting this option makes fio average the each log entry
2758 over the specified period of time, reducing the resolution of the log. See
2759 :option:`log_max_value` as well. Defaults to 0, logging all entries.
2760 Also see `Log File Formats`_.
2762 .. option:: log_hist_msec=int
2764 Same as :option:`log_avg_msec`, but logs entries for completion latency
2765 histograms. Computing latency percentiles from averages of intervals using
2766 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
2767 histogram entries over the specified period of time, reducing log sizes for
2768 high IOPS devices while retaining percentile accuracy. See
2769 :option:`log_hist_coarseness` as well. Defaults to 0, meaning histogram
2770 logging is disabled.
2772 .. option:: log_hist_coarseness=int
2774 Integer ranging from 0 to 6, defining the coarseness of the resolution of
2775 the histogram logs enabled with :option:`log_hist_msec`. For each increment
2776 in coarseness, fio outputs half as many bins. Defaults to 0, for which
2777 histogram logs contain 1216 latency bins. See `Log File Formats`_.
2779 .. option:: log_max_value=bool
2781 If :option:`log_avg_msec` is set, fio logs the average over that window. If
2782 you instead want to log the maximum value, set this option to 1. Defaults to
2783 0, meaning that averaged values are logged.
2785 .. option:: log_offset=bool
2787 If this is set, the iolog options will include the byte offset for the I/O
2788 entry as well as the other data values. Defaults to 0 meaning that
2789 offsets are not present in logs. Also see `Log File Formats`_.
2791 .. option:: log_compression=int
2793 If this is set, fio will compress the I/O logs as it goes, to keep the
2794 memory footprint lower. When a log reaches the specified size, that chunk is
2795 removed and compressed in the background. Given that I/O logs are fairly
2796 highly compressible, this yields a nice memory savings for longer runs. The
2797 downside is that the compression will consume some background CPU cycles, so
2798 it may impact the run. This, however, is also true if the logging ends up
2799 consuming most of the system memory. So pick your poison. The I/O logs are
2800 saved normally at the end of a run, by decompressing the chunks and storing
2801 them in the specified log file. This feature depends on the availability of
2804 .. option:: log_compression_cpus=str
2806 Define the set of CPUs that are allowed to handle online log compression for
2807 the I/O jobs. This can provide better isolation between performance
2808 sensitive jobs, and background compression work.
2810 .. option:: log_store_compressed=bool
2812 If set, fio will store the log files in a compressed format. They can be
2813 decompressed with fio, using the :option:`--inflate-log` command line
2814 parameter. The files will be stored with a :file:`.fz` suffix.
2816 .. option:: log_unix_epoch=bool
2818 If set, fio will log Unix timestamps to the log files produced by enabling
2819 write_type_log for each log type, instead of the default zero-based
2822 .. option:: block_error_percentiles=bool
2824 If set, record errors in trim block-sized units from writes and trims and
2825 output a histogram of how many trims it took to get to errors, and what kind
2826 of error was encountered.
2828 .. option:: bwavgtime=int
2830 Average the calculated bandwidth over the given time. Value is specified in
2831 milliseconds. If the job also does bandwidth logging through
2832 :option:`write_bw_log`, then the minimum of this option and
2833 :option:`log_avg_msec` will be used. Default: 500ms.
2835 .. option:: iopsavgtime=int
2837 Average the calculated IOPS over the given time. Value is specified in
2838 milliseconds. If the job also does IOPS logging through
2839 :option:`write_iops_log`, then the minimum of this option and
2840 :option:`log_avg_msec` will be used. Default: 500ms.
2842 .. option:: disk_util=bool
2844 Generate disk utilization statistics, if the platform supports it.
2847 .. option:: disable_lat=bool
2849 Disable measurements of total latency numbers. Useful only for cutting back
2850 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
2851 performance at really high IOPS rates. Note that to really get rid of a
2852 large amount of these calls, this option must be used with
2853 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
2855 .. option:: disable_clat=bool
2857 Disable measurements of completion latency numbers. See
2858 :option:`disable_lat`.
2860 .. option:: disable_slat=bool
2862 Disable measurements of submission latency numbers. See
2863 :option:`disable_lat`.
2865 .. option:: disable_bw_measurement=bool, disable_bw=bool
2867 Disable measurements of throughput/bandwidth numbers. See
2868 :option:`disable_lat`.
2870 .. option:: clat_percentiles=bool
2872 Enable the reporting of percentiles of completion latencies. This
2873 option is mutually exclusive with :option:`lat_percentiles`.
2875 .. option:: lat_percentiles=bool
2877 Enable the reporting of percentiles of IO latencies. This is similar
2878 to :option:`clat_percentiles`, except that this includes the
2879 submission latency. This option is mutually exclusive with
2880 :option:`clat_percentiles`.
2882 .. option:: percentile_list=float_list
2884 Overwrite the default list of percentiles for completion latencies and the
2885 block error histogram. Each number is a floating number in the range
2886 (0,100], and the maximum length of the list is 20. Use ``:`` to separate the
2887 numbers, and list the numbers in ascending order. For example,
2888 ``--percentile_list=99.5:99.9`` will cause fio to report the values of
2889 completion latency below which 99.5% and 99.9% of the observed latencies
2896 .. option:: exitall_on_error
2898 When one job finishes in error, terminate the rest. The default is to wait
2899 for each job to finish.
2901 .. option:: continue_on_error=str
2903 Normally fio will exit the job on the first observed failure. If this option
2904 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
2905 EILSEQ) until the runtime is exceeded or the I/O size specified is
2906 completed. If this option is used, there are two more stats that are
2907 appended, the total error count and the first error. The error field given
2908 in the stats is the first error that was hit during the run.
2910 The allowed values are:
2913 Exit on any I/O or verify errors.
2916 Continue on read errors, exit on all others.
2919 Continue on write errors, exit on all others.
2922 Continue on any I/O error, exit on all others.
2925 Continue on verify errors, exit on all others.
2928 Continue on all errors.
2931 Backward-compatible alias for 'none'.
2934 Backward-compatible alias for 'all'.
2936 .. option:: ignore_error=str
2938 Sometimes you want to ignore some errors during test in that case you can
2939 specify error list for each error type, instead of only being able to
2940 ignore the default 'non-fatal error' using :option:`continue_on_error`.
2941 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
2942 given error type is separated with ':'. Error may be symbol ('ENOSPC',
2943 'ENOMEM') or integer. Example::
2945 ignore_error=EAGAIN,ENOSPC:122
2947 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
2948 WRITE. This option works by overriding :option:`continue_on_error` with
2949 the list of errors for each error type if any.
2951 .. option:: error_dump=bool
2953 If set dump every error even if it is non fatal, true by default. If
2954 disabled only fatal error will be dumped.
2956 Running predefined workloads
2957 ----------------------------
2959 Fio includes predefined profiles that mimic the I/O workloads generated by
2962 .. option:: profile=str
2964 The predefined workload to run. Current profiles are:
2967 Threaded I/O bench (tiotest/tiobench) like workload.
2970 Aerospike Certification Tool (ACT) like workload.
2972 To view a profile's additional options use :option:`--cmdhelp` after specifying
2973 the profile. For example::
2975 $ fio --profile=act --cmdhelp
2980 .. option:: device-names=str
2985 .. option:: load=int
2988 ACT load multiplier. Default: 1.
2990 .. option:: test-duration=time
2993 How long the entire test takes to run. When the unit is omitted, the value
2994 is given in seconds. Default: 24h.
2996 .. option:: threads-per-queue=int
2999 Number of read I/O threads per device. Default: 8.
3001 .. option:: read-req-num-512-blocks=int
3004 Number of 512B blocks to read at the time. Default: 3.
3006 .. option:: large-block-op-kbytes=int
3009 Size of large block ops in KiB (writes). Default: 131072.
3014 Set to run ACT prep phase.
3016 Tiobench profile options
3017 ~~~~~~~~~~~~~~~~~~~~~~~~
3019 .. option:: size=str
3024 .. option:: block=int
3027 Block size in bytes. Default: 4096.
3029 .. option:: numruns=int
3039 .. option:: threads=int
3044 Interpreting the output
3045 -----------------------
3048 Example output was based on the following:
3049 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --time_based \
3050 --rate=1256k --bs=14K --name=quick --runtime=1s --name=mixed \
3051 --runtime=2m --rw=rw
3053 Fio spits out a lot of output. While running, fio will display the status of the
3054 jobs created. An example of that would be::
3056 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]
3058 The characters inside the first set of square brackets denote the current status of
3059 each thread. The first character is the first job defined in the job file, and so
3060 forth. The possible values (in typical life cycle order) are:
3062 +------+-----+-----------------------------------------------------------+
3064 +======+=====+===========================================================+
3065 | P | | Thread setup, but not started. |
3066 +------+-----+-----------------------------------------------------------+
3067 | C | | Thread created. |
3068 +------+-----+-----------------------------------------------------------+
3069 | I | | Thread initialized, waiting or generating necessary data. |
3070 +------+-----+-----------------------------------------------------------+
3071 | | p | Thread running pre-reading file(s). |
3072 +------+-----+-----------------------------------------------------------+
3073 | | / | Thread is in ramp period. |
3074 +------+-----+-----------------------------------------------------------+
3075 | | R | Running, doing sequential reads. |
3076 +------+-----+-----------------------------------------------------------+
3077 | | r | Running, doing random reads. |
3078 +------+-----+-----------------------------------------------------------+
3079 | | W | Running, doing sequential writes. |
3080 +------+-----+-----------------------------------------------------------+
3081 | | w | Running, doing random writes. |
3082 +------+-----+-----------------------------------------------------------+
3083 | | M | Running, doing mixed sequential reads/writes. |
3084 +------+-----+-----------------------------------------------------------+
3085 | | m | Running, doing mixed random reads/writes. |
3086 +------+-----+-----------------------------------------------------------+
3087 | | D | Running, doing sequential trims. |
3088 +------+-----+-----------------------------------------------------------+
3089 | | d | Running, doing random trims. |
3090 +------+-----+-----------------------------------------------------------+
3091 | | F | Running, currently waiting for :manpage:`fsync(2)`. |
3092 +------+-----+-----------------------------------------------------------+
3093 | | V | Running, doing verification of written data. |
3094 +------+-----+-----------------------------------------------------------+
3095 | f | | Thread finishing. |
3096 +------+-----+-----------------------------------------------------------+
3097 | E | | Thread exited, not reaped by main thread yet. |
3098 +------+-----+-----------------------------------------------------------+
3099 | _ | | Thread reaped. |
3100 +------+-----+-----------------------------------------------------------+
3101 | X | | Thread reaped, exited with an error. |
3102 +------+-----+-----------------------------------------------------------+
3103 | K | | Thread reaped, exited due to signal. |
3104 +------+-----+-----------------------------------------------------------+
3107 Example output was based on the following:
3108 TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --runtime=58m \
3109 --time_based --rate=2512k --bs=256K --numjobs=10 \
3110 --name=readers --rw=read --name=writers --rw=write
3112 Fio will condense the thread string as not to take up more space on the command
3113 line than needed. For instance, if you have 10 readers and 10 writers running,
3114 the output would look like this::
3116 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]
3118 Note that the status string is displayed in order, so it's possible to tell which of
3119 the jobs are currently doing what. In the example above this means that jobs 1--10
3120 are readers and 11--20 are writers.
3122 The other values are fairly self explanatory -- number of threads currently
3123 running and doing I/O, the number of currently open files (f=), the estimated
3124 completion percentage, the rate of I/O since last check (read speed listed first,
3125 then write speed and optionally trim speed) in terms of bandwidth and IOPS,
3126 and time to completion for the current running group. It's impossible to estimate
3127 runtime of the following groups (if any).
3130 Example output was based on the following:
3131 TZ=UTC fio --iodepth=16 --ioengine=posixaio --filename=/tmp/fiofile \
3132 --direct=1 --size=100M --time_based --runtime=50s --rate_iops=89 \
3133 --bs=7K --name=Client1 --rw=write
3135 When fio is done (or interrupted by :kbd:`Ctrl-C`), it will show the data for
3136 each thread, group of threads, and disks in that order. For each overall thread (or
3137 group) the output looks like::
3139 Client1: (groupid=0, jobs=1): err= 0: pid=16109: Sat Jun 24 12:07:54 2017
3140 write: IOPS=88, BW=623KiB/s (638kB/s)(30.4MiB/50032msec)
3141 slat (nsec): min=500, max=145500, avg=8318.00, stdev=4781.50
3142 clat (usec): min=170, max=78367, avg=4019.02, stdev=8293.31
3143 lat (usec): min=174, max=78375, avg=4027.34, stdev=8291.79
3144 clat percentiles (usec):
3145 | 1.00th=[ 302], 5.00th=[ 326], 10.00th=[ 343], 20.00th=[ 363],
3146 | 30.00th=[ 392], 40.00th=[ 404], 50.00th=[ 416], 60.00th=[ 445],
3147 | 70.00th=[ 816], 80.00th=[ 6718], 90.00th=[12911], 95.00th=[21627],
3148 | 99.00th=[43779], 99.50th=[51643], 99.90th=[68682], 99.95th=[72877],
3150 bw ( KiB/s): min= 532, max= 686, per=0.10%, avg=622.87, stdev=24.82, samples= 100
3151 iops : min= 76, max= 98, avg=88.98, stdev= 3.54, samples= 100
3152 lat (usec) : 250=0.04%, 500=64.11%, 750=4.81%, 1000=2.79%
3153 lat (msec) : 2=4.16%, 4=1.84%, 10=4.90%, 20=11.33%, 50=5.37%
3154 lat (msec) : 100=0.65%
3155 cpu : usr=0.27%, sys=0.18%, ctx=12072, majf=0, minf=21
3156 IO depths : 1=85.0%, 2=13.1%, 4=1.8%, 8=0.1%, 16=0.0%, 32=0.0%, >=64=0.0%
3157 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3158 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3159 issued rwt: total=0,4450,0, short=0,0,0, dropped=0,0,0
3160 latency : target=0, window=0, percentile=100.00%, depth=8
3162 The job name (or first job's name when using :option:`group_reporting`) is printed,
3163 along with the group id, count of jobs being aggregated, last error id seen (which
3164 is 0 when there are no errors), pid/tid of that thread and the time the job/group
3165 completed. Below are the I/O statistics for each data direction performed (showing
3166 writes in the example above). In the order listed, they denote:
3169 The string before the colon shows the I/O direction the statistics
3170 are for. **IOPS** is the average I/Os performed per second. **BW**
3171 is the average bandwidth rate shown as: value in power of 2 format
3172 (value in power of 10 format). The last two values show: (**total
3173 I/O performed** in power of 2 format / **runtime** of that thread).
3176 Submission latency (**min** being the minimum, **max** being the
3177 maximum, **avg** being the average, **stdev** being the standard
3178 deviation). This is the time it took to submit the I/O. For
3179 sync I/O this row is not displayed as the slat is really the
3180 completion latency (since queue/complete is one operation there).
3181 This value can be in nanoseconds, microseconds or milliseconds ---
3182 fio will choose the most appropriate base and print that (in the
3183 example above nanoseconds was the best scale). Note: in :option:`--minimal` mode
3184 latencies are always expressed in microseconds.
3187 Completion latency. Same names as slat, this denotes the time from
3188 submission to completion of the I/O pieces. For sync I/O, clat will
3189 usually be equal (or very close) to 0, as the time from submit to
3190 complete is basically just CPU time (I/O has already been done, see slat
3194 Total latency. Same names as slat and clat, this denotes the time from
3195 when fio created the I/O unit to completion of the I/O operation.
3198 Bandwidth statistics based on samples. Same names as the xlat stats,
3199 but also includes the number of samples taken (**samples**) and an
3200 approximate percentage of total aggregate bandwidth this thread
3201 received in its group (**per**). This last value is only really
3202 useful if the threads in this group are on the same disk, since they
3203 are then competing for disk access.
3206 IOPS statistics based on samples. Same names as bw.
3208 **lat (nsec/usec/msec)**
3209 The distribution of I/O completion latencies. This is the time from when
3210 I/O leaves fio and when it gets completed. Unlike the separate
3211 read/write/trim sections above, the data here and in the remaining
3212 sections apply to all I/Os for the reporting group. 250=0.04% means that
3213 0.04% of the I/Os completed in under 250us. 500=64.11% means that 64.11%
3214 of the I/Os required 250 to 499us for completion.
3217 CPU usage. User and system time, along with the number of context
3218 switches this thread went through, usage of system and user time, and
3219 finally the number of major and minor page faults. The CPU utilization
3220 numbers are averages for the jobs in that reporting group, while the
3221 context and fault counters are summed.
3224 The distribution of I/O depths over the job lifetime. The numbers are
3225 divided into powers of 2 and each entry covers depths from that value
3226 up to those that are lower than the next entry -- e.g., 16= covers
3227 depths from 16 to 31. Note that the range covered by a depth
3228 distribution entry can be different to the range covered by the
3229 equivalent submit/complete distribution entry.
3232 How many pieces of I/O were submitting in a single submit call. Each
3233 entry denotes that amount and below, until the previous entry -- e.g.,
3234 16=100% means that we submitted anywhere between 9 to 16 I/Os per submit
3235 call. Note that the range covered by a submit distribution entry can
3236 be different to the range covered by the equivalent depth distribution
3240 Like the above submit number, but for completions instead.
3243 The number of read/write/trim requests issued, and how many of them were
3247 These values are for `--latency-target` and related options. When
3248 these options are engaged, this section describes the I/O depth required
3249 to meet the specified latency target.
3252 Example output was based on the following:
3253 TZ=UTC fio --ioengine=null --iodepth=2 --size=100M --numjobs=2 \
3254 --rate_process=poisson --io_limit=32M --name=read --bs=128k \
3255 --rate=11M --name=write --rw=write --bs=2k --rate=700k
3257 After each client has been listed, the group statistics are printed. They
3258 will look like this::
3260 Run status group 0 (all jobs):
3261 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
3262 WRITE: bw=1231KiB/s (1261kB/s), 616KiB/s-621KiB/s (630kB/s-636kB/s), io=64.0MiB (67.1MB), run=52747-53223msec
3264 For each data direction it prints:
3267 Aggregate bandwidth of threads in this group followed by the
3268 minimum and maximum bandwidth of all the threads in this group.
3269 Values outside of brackets are power-of-2 format and those
3270 within are the equivalent value in a power-of-10 format.
3272 Aggregate I/O performed of all threads in this group. The
3273 format is the same as bw.
3275 The smallest and longest runtimes of the threads in this group.
3277 And finally, the disk statistics are printed. This is Linux specific. They will look like this::
3279 Disk stats (read/write):
3280 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
3282 Each value is printed for both reads and writes, with reads first. The
3286 Number of I/Os performed by all groups.
3288 Number of merges performed by the I/O scheduler.
3290 Number of ticks we kept the disk busy.
3292 Total time spent in the disk queue.
3294 The disk utilization. A value of 100% means we kept the disk
3295 busy constantly, 50% would be a disk idling half of the time.
3297 It is also possible to get fio to dump the current output while it is running,
3298 without terminating the job. To do that, send fio the **USR1** signal. You can
3299 also get regularly timed dumps by using the :option:`--status-interval`
3300 parameter, or by creating a file in :file:`/tmp` named
3301 :file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
3302 current output status.
3308 For scripted usage where you typically want to generate tables or graphs of the
3309 results, fio can output the results in a semicolon separated format. The format
3310 is one long line of values, such as::
3312 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%
3313 A description of this job goes here.
3315 The job description (if provided) follows on a second line.
3317 To enable terse output, use the :option:`--minimal` or
3318 :option:`--output-format`\=terse command line options. The
3319 first value is the version of the terse output format. If the output has to be
3320 changed for some reason, this number will be incremented by 1 to signify that
3323 Split up, the format is as follows (comments in brackets denote when a
3324 field was introduced or whether it's specific to some terse version):
3328 terse version, fio version [v3], jobname, groupid, error
3332 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3333 Submission latency: min, max, mean, stdev (usec)
3334 Completion latency: min, max, mean, stdev (usec)
3335 Completion latency percentiles: 20 fields (see below)
3336 Total latency: min, max, mean, stdev (usec)
3337 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3338 IOPS [v5]: min, max, mean, stdev, number of samples
3344 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3345 Submission latency: min, max, mean, stdev (usec)
3346 Completion latency: min, max, mean, stdev (usec)
3347 Completion latency percentiles: 20 fields (see below)
3348 Total latency: min, max, mean, stdev (usec)
3349 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3350 IOPS [v5]: min, max, mean, stdev, number of samples
3352 TRIM status [all but version 3]:
3354 Fields are similar to READ/WRITE status.
3358 user, system, context switches, major faults, minor faults
3362 <=1, 2, 4, 8, 16, 32, >=64
3364 I/O latencies microseconds::
3366 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
3368 I/O latencies milliseconds::
3370 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
3372 Disk utilization [v3]::
3374 disk name, read ios, write ios, read merges, write merges, read ticks, write ticks,
3375 time spent in queue, disk utilization percentage
3377 Additional Info (dependent on continue_on_error, default off)::
3379 total # errors, first error code
3381 Additional Info (dependent on description being set)::
3385 Completion latency percentiles can be a grouping of up to 20 sets, so for the
3386 terse output fio writes all of them. Each field will look like this::
3390 which is the Xth percentile, and the `usec` latency associated with it.
3392 For `Disk utilization`, all disks used by fio are shown. So for each disk there
3393 will be a disk utilization section.
3395 Below is a single line containing short names for each of the fields in the
3396 minimal output v3, separated by semicolons::
3398 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
3404 The `json` output format is intended to be both human readable and convenient
3405 for automated parsing. For the most part its sections mirror those of the
3406 `normal` output. The `runtime` value is reported in msec and the `bw` value is
3407 reported in 1024 bytes per second units.
3413 The `json+` output format is identical to the `json` output format except that it
3414 adds a full dump of the completion latency bins. Each `bins` object contains a
3415 set of (key, value) pairs where keys are latency durations and values count how
3416 many I/Os had completion latencies of the corresponding duration. For example,
3419 "bins" : { "87552" : 1, "89600" : 1, "94720" : 1, "96768" : 1, "97792" : 1, "99840" : 1, "100864" : 2, "103936" : 6, "104960" : 534, "105984" : 5995, "107008" : 7529, ... }
3421 This data indicates that one I/O required 87,552ns to complete, two I/Os required
3422 100,864ns to complete, and 7529 I/Os required 107,008ns to complete.
3424 Also included with fio is a Python script `fio_jsonplus_clat2csv` that takes
3425 json+ output and generates CSV-formatted latency data suitable for plotting.
3427 The latency durations actually represent the midpoints of latency intervals.
3428 For details refer to :file:`stat.h`.
3434 There are two trace file format that you can encounter. The older (v1) format is
3435 unsupported since version 1.20-rc3 (March 2008). It will still be described
3436 below in case that you get an old trace and want to understand it.
3438 In any case the trace is a simple text file with a single action per line.
3441 Trace file format v1
3442 ~~~~~~~~~~~~~~~~~~~~
3444 Each line represents a single I/O action in the following format::
3448 where `rw=0/1` for read/write, and the `offset` and `length` entries being in bytes.
3450 This format is not supported in fio versions >= 1.20-rc3.
3453 Trace file format v2
3454 ~~~~~~~~~~~~~~~~~~~~
3456 The second version of the trace file format was added in fio version 1.17. It
3457 allows to access more then one file per trace and has a bigger set of possible
3460 The first line of the trace file has to be::
3464 Following this can be lines in two different formats, which are described below.
3466 The file management format::
3470 The `filename` is given as an absolute path. The `action` can be one of these:
3473 Add the given `filename` to the trace.
3475 Open the file with the given `filename`. The `filename` has to have
3476 been added with the **add** action before.
3478 Close the file with the given `filename`. The file has to have been
3482 The file I/O action format::
3484 filename action offset length
3486 The `filename` is given as an absolute path, and has to have been added and
3487 opened before it can be used with this format. The `offset` and `length` are
3488 given in bytes. The `action` can be one of these:
3491 Wait for `offset` microseconds. Everything below 100 is discarded.
3492 The time is relative to the previous `wait` statement.
3494 Read `length` bytes beginning from `offset`.
3496 Write `length` bytes beginning from `offset`.
3498 :manpage:`fsync(2)` the file.
3500 :manpage:`fdatasync(2)` the file.
3502 Trim the given file from the given `offset` for `length` bytes.
3504 CPU idleness profiling
3505 ----------------------
3507 In some cases, we want to understand CPU overhead in a test. For example, we
3508 test patches for the specific goodness of whether they reduce CPU usage.
3509 Fio implements a balloon approach to create a thread per CPU that runs at idle
3510 priority, meaning that it only runs when nobody else needs the cpu.
3511 By measuring the amount of work completed by the thread, idleness of each CPU
3512 can be derived accordingly.
3514 An unit work is defined as touching a full page of unsigned characters. Mean and
3515 standard deviation of time to complete an unit work is reported in "unit work"
3516 section. Options can be chosen to report detailed percpu idleness or overall
3517 system idleness by aggregating percpu stats.
3520 Verification and triggers
3521 -------------------------
3523 Fio is usually run in one of two ways, when data verification is done. The first
3524 is a normal write job of some sort with verify enabled. When the write phase has
3525 completed, fio switches to reads and verifies everything it wrote. The second
3526 model is running just the write phase, and then later on running the same job
3527 (but with reads instead of writes) to repeat the same I/O patterns and verify
3528 the contents. Both of these methods depend on the write phase being completed,
3529 as fio otherwise has no idea how much data was written.
3531 With verification triggers, fio supports dumping the current write state to
3532 local files. Then a subsequent read verify workload can load this state and know
3533 exactly where to stop. This is useful for testing cases where power is cut to a
3534 server in a managed fashion, for instance.
3536 A verification trigger consists of two things:
3538 1) Storing the write state of each job.
3539 2) Executing a trigger command.
3541 The write state is relatively small, on the order of hundreds of bytes to single
3542 kilobytes. It contains information on the number of completions done, the last X
3545 A trigger is invoked either through creation ('touch') of a specified file in
3546 the system, or through a timeout setting. If fio is run with
3547 :option:`--trigger-file`\= :file:`/tmp/trigger-file`, then it will continually
3548 check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
3549 will fire off the trigger (thus saving state, and executing the trigger
3552 For client/server runs, there's both a local and remote trigger. If fio is
3553 running as a server backend, it will send the job states back to the client for
3554 safe storage, then execute the remote trigger, if specified. If a local trigger
3555 is specified, the server will still send back the write state, but the client
3556 will then execute the trigger.
3558 Verification trigger example
3559 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3561 Let's say we want to run a powercut test on the remote Linux machine 'server'.
3562 Our write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
3563 some point during the run, and we'll run this test from the safety or our local
3564 machine, 'localbox'. On the server, we'll start the fio backend normally::
3566 server# fio --server
3568 and on the client, we'll fire off the workload::
3570 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
3572 We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
3574 echo b > /proc/sysrq-trigger
3576 on the server once it has received the trigger and sent us the write state. This
3577 will work, but it's not **really** cutting power to the server, it's merely
3578 abruptly rebooting it. If we have a remote way of cutting power to the server
3579 through IPMI or similar, we could do that through a local trigger command
3580 instead. Let's assume we have a script that does IPMI reboot of a given hostname,
3581 ipmi-reboot. On localbox, we could then have run fio with a local trigger
3584 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
3586 For this case, fio would wait for the server to send us the write state, then
3587 execute ``ipmi-reboot server`` when that happened.
3589 Loading verify state
3590 ~~~~~~~~~~~~~~~~~~~~
3592 To load stored write state, a read verification job file must contain the
3593 :option:`verify_state_load` option. If that is set, fio will load the previously
3594 stored state. For a local fio run this is done by loading the files directly,
3595 and on a client/server run, the server backend will ask the client to send the
3596 files over and load them from there.
3602 Fio supports a variety of log file formats, for logging latencies, bandwidth,
3603 and IOPS. The logs share a common format, which looks like this:
3605 *time* (`msec`), *value*, *data direction*, *block size* (`bytes`),
3608 *Time* for the log entry is always in milliseconds. The *value* logged depends
3609 on the type of log, it will be one of the following:
3612 Value is latency in nsecs
3618 *Data direction* is one of the following:
3627 The entry's *block size* is always in bytes. The *offset* is the offset, in bytes,
3628 from the start of the file, for that particular I/O. The logging of the offset can be
3629 toggled with :option:`log_offset`.
3631 Fio defaults to logging every individual I/O. When IOPS are logged for individual
3632 I/Os the *value* entry will always be 1. If windowed logging is enabled through
3633 :option:`log_avg_msec`, fio logs the average values over the specified period of time.
3634 If windowed logging is enabled and :option:`log_max_value` is set, then fio logs
3635 maximum values in that window instead of averages. Since *data direction*, *block
3636 size* and *offset* are per-I/O values, if windowed logging is enabled they
3637 aren't applicable and will be 0.
3642 Normally fio is invoked as a stand-alone application on the machine where the
3643 I/O workload should be generated. However, the backend and frontend of fio can
3644 be run separately i.e., the fio server can generate an I/O workload on the "Device
3645 Under Test" while being controlled by a client on another machine.
3647 Start the server on the machine which has access to the storage DUT::
3651 where `args` defines what fio listens to. The arguments are of the form
3652 ``type,hostname`` or ``IP,port``. *type* is either ``ip`` (or ip4) for TCP/IP
3653 v4, ``ip6`` for TCP/IP v6, or ``sock`` for a local unix domain socket.
3654 *hostname* is either a hostname or IP address, and *port* is the port to listen
3655 to (only valid for TCP/IP, not a local socket). Some examples:
3659 Start a fio server, listening on all interfaces on the default port (8765).
3661 2) ``fio --server=ip:hostname,4444``
3663 Start a fio server, listening on IP belonging to hostname and on port 4444.
3665 3) ``fio --server=ip6:::1,4444``
3667 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
3669 4) ``fio --server=,4444``
3671 Start a fio server, listening on all interfaces on port 4444.
3673 5) ``fio --server=1.2.3.4``
3675 Start a fio server, listening on IP 1.2.3.4 on the default port.
3677 6) ``fio --server=sock:/tmp/fio.sock``
3679 Start a fio server, listening on the local socket :file:`/tmp/fio.sock`.
3681 Once a server is running, a "client" can connect to the fio server with::
3683 fio <local-args> --client=<server> <remote-args> <job file(s)>
3685 where `local-args` are arguments for the client where it is running, `server`
3686 is the connect string, and `remote-args` and `job file(s)` are sent to the
3687 server. The `server` string follows the same format as it does on the server
3688 side, to allow IP/hostname/socket and port strings.
3690 Fio can connect to multiple servers this way::
3692 fio --client=<server1> <job file(s)> --client=<server2> <job file(s)>
3694 If the job file is located on the fio server, then you can tell the server to
3695 load a local file as well. This is done by using :option:`--remote-config` ::
3697 fio --client=server --remote-config /path/to/file.fio
3699 Then fio will open this local (to the server) job file instead of being passed
3700 one from the client.
3702 If you have many servers (example: 100 VMs/containers), you can input a pathname
3703 of a file containing host IPs/names as the parameter value for the
3704 :option:`--client` option. For example, here is an example :file:`host.list`
3705 file containing 2 hostnames::
3707 host1.your.dns.domain
3708 host2.your.dns.domain
3710 The fio command would then be::
3712 fio --client=host.list <job file(s)>
3714 In this mode, you cannot input server-specific parameters or job files -- all
3715 servers receive the same job file.
3717 In order to let ``fio --client`` runs use a shared filesystem from multiple
3718 hosts, ``fio --client`` now prepends the IP address of the server to the
3719 filename. For example, if fio is using the directory :file:`/mnt/nfs/fio` and is
3720 writing filename :file:`fileio.tmp`, with a :option:`--client` `hostfile`
3721 containing two hostnames ``h1`` and ``h2`` with IP addresses 192.168.10.120 and
3722 192.168.10.121, then fio will create two files::
3724 /mnt/nfs/fio/192.168.10.120.fileio.tmp
3725 /mnt/nfs/fio/192.168.10.121.fileio.tmp