4 The first step in getting fio to simulate a desired I/O workload, is writing a
5 job file describing that specific setup. A job file may contain any number of
6 threads and/or files -- the typical contents of the job file is a *global*
7 section defining shared parameters, and one or more job sections describing the
8 jobs involved. When run, fio parses this file and sets everything up as
9 described. If we break down a job from top to bottom, it contains the following
14 Defines the I/O pattern issued to the file(s). We may only be reading
15 sequentially from this file(s), or we may be writing randomly. Or even
16 mixing reads and writes, sequentially or randomly.
17 Should we be doing buffered I/O, or direct/raw I/O?
21 In how large chunks are we issuing I/O? This may be a single value,
22 or it may describe a range of block sizes.
26 How much data are we going to be reading/writing.
30 How do we issue I/O? We could be memory mapping the file, we could be
31 using regular read/write, we could be using splice, async I/O, or even
36 If the I/O engine is async, how large a queuing depth do we want to
42 How many files are we spreading the workload over.
44 `Threads, processes and job synchronization`_
46 How many threads or processes should we spread this workload over.
48 The above are the basic parameters defined for a workload, in addition there's a
49 multitude of parameters that modify other aspects of how this job behaves.
55 .. option:: --debug=type
57 Enable verbose tracing of various fio actions. May be ``all`` for all types
58 or individual types separated by a comma (e.g. ``--debug=file,mem`` will
59 enable file and memory debugging). Currently, additional logging is
63 Dump info related to processes.
65 Dump info related to file actions.
67 Dump info related to I/O queuing.
69 Dump info related to memory allocations.
71 Dump info related to blktrace setup.
73 Dump info related to I/O verification.
75 Enable all debug options.
77 Dump info related to random offset generation.
79 Dump info related to option matching and parsing.
81 Dump info related to disk utilization updates.
83 Dump info only related to job number x.
85 Dump info only related to mutex up/down ops.
87 Dump info related to profile extensions.
89 Dump info related to internal time keeping.
91 Dump info related to networking connections.
93 Dump info related to I/O rate switching.
95 Dump info related to log compress/decompress.
97 Show available debug options.
99 .. option:: --parse-only
101 Parse options only, don\'t start any I/O.
103 .. option:: --output=filename
105 Write output to file `filename`.
107 .. option:: --bandwidth-log
109 Generate aggregate bandwidth logs.
111 .. option:: --minimal
113 Print statistics in a terse, semicolon-delimited format.
115 .. option:: --append-terse
117 Print statistics in selected mode AND terse, semicolon-delimited format.
118 **deprecated**, use :option:`--output-format` instead to select multiple
121 .. option:: --output-format=type
123 Set the reporting format to `normal`, `terse`, `json`, or `json+`. Multiple
124 formats can be selected, separate by a comma. `terse` is a CSV based
125 format. `json+` is like `json`, except it adds a full dump of the latency
128 .. option:: --terse-version=type
130 Set terse version output format (default 3, or 2 or 4).
132 .. option:: --version
134 Print version info 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 builtin checksumming functions. If no argument is
147 given, all of them are tested. Or a comma separated list can be passed, in
148 which case the given ones are tested.
150 .. option:: --cmdhelp=command
152 Print help information for `command`. May be ``all`` for all commands.
154 .. option:: --enghelp=[ioengine[,command]]
156 List all commands defined by :option:`ioengine`, or print help for `command`
157 defined by :option:`ioengine`. If no :option:`ioengine` is given, list all
160 .. option:: --showcmd=jobfile
162 Turn a job file into command line options.
164 .. option:: --readonly
166 Turn on safety read-only checks, preventing writes. The ``--readonly``
167 option is an extra safety guard to prevent users from accidentally starting
168 a write workload when that is not desired. Fio will only write if
169 `rw=write/randwrite/rw/randrw` is given. This extra safety net can be used
170 as an extra precaution as ``--readonly`` will also enable a write check in
171 the I/O engine core to prevent writes due to unknown user space bug(s).
173 .. option:: --eta=when
175 When real-time ETA estimate should be printed. May be `always`, `never` or
178 .. option:: --eta-newline=time
180 Force a new line for every `time` period passed.
182 .. option:: --status-interval=time
184 Force full status dump every `time` period passed.
186 .. option:: --section=name
188 Only run specified section in job file. Multiple sections can be specified.
189 The ``--section`` option allows one to combine related jobs into one file.
190 E.g. one job file could define light, moderate, and heavy sections. Tell
191 fio to run only the "heavy" section by giving ``--section=heavy``
192 command line option. One can also specify the "write" operations in one
193 section and "verify" operation in another section. The ``--section`` option
194 only applies to job sections. The reserved *global* section is always
197 .. option:: --alloc-size=kb
199 Set the internal smalloc pool to this size in kb (def 1024). The
200 ``--alloc-size`` switch allows one to use a larger pool size for smalloc.
201 If running large jobs with randommap enabled, fio can run out of memory.
202 Smalloc is an internal allocator for shared structures from a fixed size
203 memory pool. The pool size defaults to 16M and can grow to 8 pools.
205 NOTE: While running :file:`.fio_smalloc.*` backing store files are visible
208 .. option:: --warnings-fatal
210 All fio parser warnings are fatal, causing fio to exit with an
213 .. option:: --max-jobs=nr
215 Maximum number of threads/processes to support.
217 .. option:: --server=args
219 Start a backend server, with `args` specifying what to listen to.
220 See `Client/Server`_ section.
222 .. option:: --daemonize=pidfile
224 Background a fio server, writing the pid to the given `pidfile` file.
226 .. option:: --client=hostname
228 Instead of running the jobs locally, send and run them on the given host or
229 set of hosts. See `Client/Server`_ section.
231 .. option:: --remote-config=file
233 Tell fio server to load this local file.
235 .. option:: --idle-prof=option
237 Report cpu idleness on a system or percpu basis
238 ``--idle-prof=system,percpu`` or
239 run unit work calibration only ``--idle-prof=calibrate``.
241 .. option:: --inflate-log=log
243 Inflate and output compressed log.
245 .. option:: --trigger-file=file
247 Execute trigger cmd when file exists.
249 .. option:: --trigger-timeout=t
251 Execute trigger at this time.
253 .. option:: --trigger=cmd
255 Set this command as local trigger.
257 .. option:: --trigger-remote=cmd
259 Set this command as remote trigger.
261 .. option:: --aux-path=path
263 Use this path for fio state generated files.
265 Any parameters following the options will be assumed to be job files, unless
266 they match a job file parameter. Multiple job files can be listed and each job
267 file will be regarded as a separate group. Fio will :option:`stonewall`
268 execution between each group.
274 As previously described, fio accepts one or more job files describing what it is
275 supposed to do. The job file format is the classic ini file, where the names
276 enclosed in [] brackets define the job name. You are free to use any ASCII name
277 you want, except *global* which has special meaning. Following the job name is
278 a sequence of zero or more parameters, one per line, that define the behavior of
279 the job. If the first character in a line is a ';' or a '#', the entire line is
280 discarded as a comment.
282 A *global* section sets defaults for the jobs described in that file. A job may
283 override a *global* section parameter, and a job file may even have several
284 *global* sections if so desired. A job is only affected by a *global* section
287 The :option:`--cmdhelp` option also lists all options. If used with an `option`
288 argument, :option:`--cmdhelp` will detail the given `option`.
290 See the `examples/` directory for inspiration on how to write job files. Note
291 the copyright and license requirements currently apply to `examples/` files.
293 So let's look at a really simple job file that defines two processes, each
294 randomly reading from a 128MiB file:
298 ; -- start job file --
309 As you can see, the job file sections themselves are empty as all the described
310 parameters are shared. As no :option:`filename` option is given, fio makes up a
311 `filename` for each of the jobs as it sees fit. On the command line, this job
312 would look as follows::
314 $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
317 Let's look at an example that has a number of processes writing randomly to
322 ; -- start job file --
333 Here we have no *global* section, as we only have one job defined anyway. We
334 want to use async I/O here, with a depth of 4 for each file. We also increased
335 the buffer size used to 32KiB and define numjobs to 4 to fork 4 identical
336 jobs. The result is 4 processes each randomly writing to their own 64MiB
337 file. Instead of using the above job file, you could have given the parameters
338 on the command line. For this case, you would specify::
340 $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
342 When fio is utilized as a basis of any reasonably large test suite, it might be
343 desirable to share a set of standardized settings across multiple job files.
344 Instead of copy/pasting such settings, any section may pull in an external
345 :file:`filename.fio` file with *include filename* directive, as in the following
348 ; -- start job file including.fio --
352 include glob-include.fio
359 include test-include.fio
360 ; -- end job file including.fio --
364 ; -- start job file glob-include.fio --
367 ; -- end job file glob-include.fio --
371 ; -- start job file test-include.fio --
374 ; -- end job file test-include.fio --
376 Settings pulled into a section apply to that section only (except *global*
377 section). Include directives may be nested in that any included file may contain
378 further include directive(s). Include files may not contain [] sections.
381 Environment variables
382 ~~~~~~~~~~~~~~~~~~~~~
384 Fio also supports environment variable expansion in job files. Any sub-string of
385 the form ``${VARNAME}`` as part of an option value (in other words, on the right
386 of the '='), will be expanded to the value of the environment variable called
387 `VARNAME`. If no such environment variable is defined, or `VARNAME` is the
388 empty string, the empty string will be substituted.
390 As an example, let's look at a sample fio invocation and job file::
392 $ SIZE=64m NUMJOBS=4 fio jobfile.fio
396 ; -- start job file --
403 This will expand to the following equivalent job file at runtime:
407 ; -- start job file --
414 Fio ships with a few example job files, you can also look there for inspiration.
419 Additionally, fio has a set of reserved keywords that will be replaced
420 internally with the appropriate value. Those keywords are:
424 The architecture page size of the running system.
428 Megabytes of total memory in the system.
432 Number of online available CPUs.
434 These can be used on the command line or in the job file, and will be
435 automatically substituted with the current system values when the job is
436 run. Simple math is also supported on these keywords, so you can perform actions
441 and get that properly expanded to 8 times the size of memory in the machine.
447 This section describes in details each parameter associated with a job. Some
448 parameters take an option of a given type, such as an integer or a
449 string. Anywhere a numeric value is required, an arithmetic expression may be
450 used, provided it is surrounded by parentheses. Supported operators are:
459 For time values in expressions, units are microseconds by default. This is
460 different than for time values not in expressions (not enclosed in
461 parentheses). The following types are used:
468 String. This is a sequence of alpha characters.
471 Integer with possible time suffix. In seconds unless otherwise
472 specified, use e.g. 10m for 10 minutes. Accepts s/m/h for seconds, minutes,
473 and hours, and accepts 'ms' (or 'msec') for milliseconds, and 'us' (or
474 'usec') for microseconds.
479 Integer. A whole number value, which may contain an integer prefix
480 and an integer suffix:
482 [*integer prefix*] **number** [*integer suffix*]
484 The optional *integer prefix* specifies the number's base. The default
485 is decimal. *0x* specifies hexadecimal.
487 The optional *integer suffix* specifies the number's units, and includes an
488 optional unit prefix and an optional unit. For quantities of data, the
489 default unit is bytes. For quantities of time, the default unit is seconds.
491 With :option:`kb_base` =1000, fio follows international standards for unit
492 prefixes. To specify power-of-10 decimal values defined in the
493 International System of Units (SI):
495 * *Ki* -- means kilo (K) or 1000
496 * *Mi* -- means mega (M) or 1000**2
497 * *Gi* -- means giga (G) or 1000**3
498 * *Ti* -- means tera (T) or 1000**4
499 * *Pi* -- means peta (P) or 1000**5
501 To specify power-of-2 binary values defined in IEC 80000-13:
503 * *k* -- means kibi (Ki) or 1024
504 * *M* -- means mebi (Mi) or 1024**2
505 * *G* -- means gibi (Gi) or 1024**3
506 * *T* -- means tebi (Ti) or 1024**4
507 * *P* -- means pebi (Pi) or 1024**5
509 With :option:`kb_base` =1024 (the default), the unit prefixes are opposite
510 from those specified in the SI and IEC 80000-13 standards to provide
511 compatibility with old scripts. For example, 4k means 4096.
513 For quantities of data, an optional unit of 'B' may be included
514 (e.g., 'kB' is the same as 'k').
516 The *integer suffix* is not case sensitive (e.g., m/mi mean mebi/mega,
517 not milli). 'b' and 'B' both mean byte, not bit.
519 Examples with :option:`kb_base` =1000:
521 * *4 KiB*: 4096, 4096b, 4096B, 4ki, 4kib, 4kiB, 4Ki, 4KiB
522 * *1 MiB*: 1048576, 1mi, 1024ki
523 * *1 MB*: 1000000, 1m, 1000k
524 * *1 TiB*: 1099511627776, 1ti, 1024gi, 1048576mi
525 * *1 TB*: 1000000000, 1t, 1000m, 1000000k
527 Examples with :option:`kb_base` =1024 (default):
529 * *4 KiB*: 4096, 4096b, 4096B, 4k, 4kb, 4kB, 4K, 4KB
530 * *1 MiB*: 1048576, 1m, 1024k
531 * *1 MB*: 1000000, 1mi, 1000ki
532 * *1 TiB*: 1099511627776, 1t, 1024g, 1048576m
533 * *1 TB*: 1000000000, 1ti, 1000mi, 1000000ki
535 To specify times (units are not case sensitive):
539 * *M* -- mean minutes
540 * *s* -- or sec means seconds (default)
541 * *ms* -- or *msec* means milliseconds
542 * *us* -- or *usec* means microseconds
544 If the option accepts an upper and lower range, use a colon ':' or
545 minus '-' to separate such values. See :ref:`irange <irange>`.
550 Boolean. Usually parsed as an integer, however only defined for
551 true and false (1 and 0).
556 Integer range with suffix. Allows value range to be given, such as
557 1024-4096. A colon may also be used as the separator, e.g. 1k:4k. If the
558 option allows two sets of ranges, they can be specified with a ',' or '/'
559 delimiter: 1k-4k/8k-32k. Also see :ref:`int <int>`.
562 A list of floating point numbers, separated by a ':' character.
568 .. option:: kb_base=int
570 Select the interpretation of unit prefixes in input parameters.
573 Inputs comply with IEC 80000-13 and the International
574 System of Units (SI). Use:
576 - power-of-2 values with IEC prefixes (e.g., KiB)
577 - power-of-10 values with SI prefixes (e.g., kB)
580 Compatibility mode (default). To avoid breaking old scripts:
582 - power-of-2 values with SI prefixes
583 - power-of-10 values with IEC prefixes
585 See :option:`bs` for more details on input parameters.
587 Outputs always use correct prefixes. Most outputs include both
590 bw=2383.3kB/s (2327.4KiB/s)
592 If only one value is reported, then kb_base selects the one to use:
594 **1000** -- SI prefixes
596 **1024** -- IEC prefixes
598 .. option:: unit_base=int
600 Base unit for reporting. Allowed values are:
603 Use auto-detection (default).
610 With the above in mind, here follows the complete list of fio job parameters.
618 ASCII name of the job. This may be used to override the name printed by fio
619 for this job. Otherwise the job name is used. On the command line this
620 parameter has the special purpose of also signaling the start of a new job.
622 .. option:: description=str
624 Text description of the job. Doesn't do anything except dump this text
625 description when this job is run. It's not parsed.
627 .. option:: loops=int
629 Run the specified number of iterations of this job. Used to repeat the same
630 workload a given number of times. Defaults to 1.
632 .. option:: numjobs=int
634 Create the specified number of clones of this job. Each clone of job
635 is spawned as an independent thread or process. May be used to setup a
636 larger number of threads/processes doing the same thing. Each thread is
637 reported separately; to see statistics for all clones as a whole, use
638 :option:`group_reporting` in conjunction with :option:`new_group`.
639 See :option:`--max-jobs`.
642 Time related parameters
643 ~~~~~~~~~~~~~~~~~~~~~~~
645 .. option:: runtime=time
647 Tell fio to terminate processing after the specified period of time. It
648 can be quite hard to determine for how long a specified job will run, so
649 this parameter is handy to cap the total runtime to a given time. When
650 the unit is omitted, the value is given in seconds.
652 .. option:: time_based
654 If set, fio will run for the duration of the :option:`runtime` specified
655 even if the file(s) are completely read or written. It will simply loop over
656 the same workload as many times as the :option:`runtime` allows.
658 .. option:: startdelay=irange(time)
660 Delay start of job for the specified number of seconds. Supports all time
661 suffixes to allow specification of hours, minutes, seconds and milliseconds
662 -- seconds are the default if a unit is omitted. Can be given as a range
663 which causes each thread to choose randomly out of the range.
665 .. option:: ramp_time=time
667 If set, fio will run the specified workload for this amount of time before
668 logging any performance numbers. Useful for letting performance settle
669 before logging results, thus minimizing the runtime required for stable
670 results. Note that the ``ramp_time`` is considered lead in time for a job,
671 thus it will increase the total runtime if a special timeout or
672 :option:`runtime` is specified. When the unit is omitted, the value is
675 .. option:: clocksource=str
677 Use the given clocksource as the base of timing. The supported options are:
680 :manpage:`gettimeofday(2)`
683 :manpage:`clock_gettime(2)`
686 Internal CPU clock source
688 cpu is the preferred clocksource if it is reliable, as it is very fast (and
689 fio is heavy on time calls). Fio will automatically use this clocksource if
690 it's supported and considered reliable on the system it is running on,
691 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
692 means supporting TSC Invariant.
694 .. option:: gtod_reduce=bool
696 Enable all of the :manpage:`gettimeofday(2)` reducing options
697 (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus
698 reduce precision of the timeout somewhat to really shrink the
699 :manpage:`gettimeofday(2)` call count. With this option enabled, we only do
700 about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all
701 time keeping was enabled.
703 .. option:: gtod_cpu=int
705 Sometimes it's cheaper to dedicate a single thread of execution to just
706 getting the current time. Fio (and databases, for instance) are very
707 intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set
708 one CPU aside for doing nothing but logging current time to a shared memory
709 location. Then the other threads/processes that run I/O workloads need only
710 copy that segment, instead of entering the kernel with a
711 :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time
712 calls will be excluded from other uses. Fio will manually clear it from the
713 CPU mask of other jobs.
719 .. option:: directory=str
721 Prefix filenames with this directory. Used to place files in a different
722 location than :file:`./`. You can specify a number of directories by
723 separating the names with a ':' character. These directories will be
724 assigned equally distributed to job clones creates with :option:`numjobs` as
725 long as they are using generated filenames. If specific `filename(s)` are
726 set fio will use the first listed directory, and thereby matching the
727 `filename` semantic which generates a file each clone if not specified, but
728 let all clones use the same if set.
730 See the :option:`filename` option for escaping certain characters.
732 .. option:: filename=str
734 Fio normally makes up a `filename` based on the job name, thread number, and
735 file number. If you want to share files between threads in a job or several
736 jobs with fixed file paths, specify a `filename` for each of them to override
737 the default. If the ioengine is file based, you can specify a number of files
738 by separating the names with a ':' colon. So if you wanted a job to open
739 :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use
740 ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is
741 specified, :option:`nrfiles` is ignored. The size of regular files specified
742 by this option will be :option:`size` divided by number of files unless
743 explicit size is specified by :option:`filesize`.
745 On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for
746 the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc.
747 Note: Windows and FreeBSD prevent write access to areas
748 of the disk containing in-use data (e.g. filesystems). If the wanted
749 `filename` does need to include a colon, then escape that with a ``\``
750 character. For instance, if the `filename` is :file:`/dev/dsk/foo@3,0:c`,
751 then you would use ``filename="/dev/dsk/foo@3,0\:c"``. The
752 :file:`-` is a reserved name, meaning stdin or stdout. Which of the two
753 depends on the read/write direction set.
755 .. option:: filename_format=str
757 If sharing multiple files between jobs, it is usually necessary to have fio
758 generate the exact names that you want. By default, fio will name a file
759 based on the default file format specification of
760 :file:`jobname.jobnumber.filenumber`. With this option, that can be
761 customized. Fio will recognize and replace the following keywords in this
765 The name of the worker thread or process.
767 The incremental number of the worker thread or process.
769 The incremental number of the file for that worker thread or
772 To have dependent jobs share a set of files, this option can be set to have
773 fio generate filenames that are shared between the two. For instance, if
774 :file:`testfiles.$filenum` is specified, file number 4 for any job will be
775 named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum`
776 will be used if no other format specifier is given.
778 .. option:: unique_filename=bool
780 To avoid collisions between networked clients, fio defaults to prefixing any
781 generated filenames (with a directory specified) with the source of the
782 client connecting. To disable this behavior, set this option to 0.
784 .. option:: opendir=str
786 Recursively open any files below directory `str`.
788 .. option:: lockfile=str
790 Fio defaults to not locking any files before it does I/O to them. If a file
791 or file descriptor is shared, fio can serialize I/O to that file to make the
792 end result consistent. This is usual for emulating real workloads that share
793 files. The lock modes are:
796 No locking. The default.
798 Only one thread or process may do I/O at a time, excluding all
801 Read-write locking on the file. Many readers may
802 access the file at the same time, but writes get exclusive access.
804 .. option:: nrfiles=int
806 Number of files to use for this job. Defaults to 1. The size of files
807 will be :option:`size` divided by this unless explicit size is specified by
808 :option:`filesize`. Files are created for each thread separately, and each
809 file will have a file number within its name by default, as explained in
810 :option:`filename` section.
813 .. option:: openfiles=int
815 Number of files to keep open at the same time. Defaults to the same as
816 :option:`nrfiles`, can be set smaller to limit the number simultaneous
819 .. option:: file_service_type=str
821 Defines how fio decides which file from a job to service next. The following
825 Choose a file at random.
828 Round robin over opened files. This is the default.
831 Finish one file before moving on to the next. Multiple files can
832 still be open depending on 'openfiles'.
835 Use a *Zipf* distribution to decide what file to access.
838 Use a *Pareto* distribution to decide what file to access.
841 Use a *Gaussian* (normal) distribution to decide what file to
844 For *random*, *roundrobin*, and *sequential*, a postfix can be appended to
845 tell fio how many I/Os to issue before switching to a new file. For example,
846 specifying ``file_service_type=random:8`` would cause fio to issue
847 8 I/Os before selecting a new file at random. For the non-uniform
848 distributions, a floating point postfix can be given to influence how the
849 distribution is skewed. See :option:`random_distribution` for a description
850 of how that would work.
852 .. option:: ioscheduler=str
854 Attempt to switch the device hosting the file to the specified I/O scheduler
857 .. option:: create_serialize=bool
859 If true, serialize the file creation for the jobs. This may be handy to
860 avoid interleaving of data files, which may greatly depend on the filesystem
861 used and even the number of processors in the system.
863 .. option:: create_fsync=bool
865 fsync the data file after creation. This is the default.
867 .. option:: create_on_open=bool
869 Don't pre-setup the files for I/O, just create open() when it's time to do
872 .. option:: create_only=bool
874 If true, fio will only run the setup phase of the job. If files need to be
875 laid out or updated on disk, only that will be done. The actual job contents
878 .. option:: allow_file_create=bool
880 If true, fio is permitted to create files as part of its workload. This is
881 the default behavior. If this option is false, then fio will error out if
882 the files it needs to use don't already exist. Default: true.
884 .. option:: allow_mounted_write=bool
886 If this isn't set, fio will abort jobs that are destructive (e.g. that write)
887 to what appears to be a mounted device or partition. This should help catch
888 creating inadvertently destructive tests, not realizing that the test will
889 destroy data on the mounted file system. Note that some platforms don't allow
890 writing against a mounted device regardless of this option. Default: false.
892 .. option:: pre_read=bool
894 If this is given, files will be pre-read into memory before starting the
895 given I/O operation. This will also clear the :option:`invalidate` flag,
896 since it is pointless to pre-read and then drop the cache. This will only
897 work for I/O engines that are seek-able, since they allow you to read the
898 same data multiple times. Thus it will not work on e.g. network or splice I/O.
900 .. option:: unlink=bool
902 Unlink the job files when done. Not the default, as repeated runs of that
903 job would then waste time recreating the file set again and again.
905 .. option:: unlink_each_loop=bool
907 Unlink job files after each iteration or loop.
909 .. option:: zonesize=int
911 Divide a file into zones of the specified size. See :option:`zoneskip`.
913 .. option:: zonerange=int
915 Give size of an I/O zone. See :option:`zoneskip`.
917 .. option:: zoneskip=int
919 Skip the specified number of bytes when :option:`zonesize` data has been
920 read. The two zone options can be used to only do I/O on zones of a file.
926 .. option:: direct=bool
928 If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that
929 ZFS on Solaris doesn't support direct I/O. On Windows the synchronous
930 ioengines don't support direct I/O. Default: false.
932 .. option:: atomic=bool
934 If value is true, attempt to use atomic direct I/O. Atomic writes are
935 guaranteed to be stable once acknowledged by the operating system. Only
936 Linux supports O_ATOMIC right now.
938 .. option:: buffered=bool
940 If value is true, use buffered I/O. This is the opposite of the
941 :option:`direct` option. Defaults to true.
943 .. option:: readwrite=str, rw=str
945 Type of I/O pattern. Accepted values are:
952 Sequential trims (Linux block devices only).
958 Random trims (Linux block devices only).
960 Sequential mixed reads and writes.
962 Random mixed reads and writes.
964 Sequential trim+write sequences. Blocks will be trimmed first,
965 then the same blocks will be written to.
967 Fio defaults to read if the option is not specified. For the mixed I/O
968 types, the default is to split them 50/50. For certain types of I/O the
969 result may still be skewed a bit, since the speed may be different. It is
970 possible to specify a number of I/O's to do before getting a new offset,
971 this is done by appending a ``:<nr>`` to the end of the string given. For a
972 random read, it would look like ``rw=randread:8`` for passing in an offset
973 modifier with a value of 8. If the suffix is used with a sequential I/O
974 pattern, then the value specified will be added to the generated offset for
975 each I/O. For instance, using ``rw=write:4k`` will skip 4k for every
976 write. It turns sequential I/O into sequential I/O with holes. See the
977 :option:`rw_sequencer` option.
979 .. option:: rw_sequencer=str
981 If an offset modifier is given by appending a number to the ``rw=<str>``
982 line, then this option controls how that number modifies the I/O offset
983 being generated. Accepted values are:
986 Generate sequential offset.
988 Generate the same offset.
990 ``sequential`` is only useful for random I/O, where fio would normally
991 generate a new random offset for every I/O. If you append e.g. 8 to randread,
992 you would get a new random offset for every 8 I/O's. The result would be a
993 seek for only every 8 I/O's, instead of for every I/O. Use ``rw=randread:8``
994 to specify that. As sequential I/O is already sequential, setting
995 ``sequential`` for that would not result in any differences. ``identical``
996 behaves in a similar fashion, except it sends the same offset 8 number of
997 times before generating a new offset.
999 .. option:: unified_rw_reporting=bool
1001 Fio normally reports statistics on a per data direction basis, meaning that
1002 reads, writes, and trims are accounted and reported separately. If this
1003 option is set fio sums the results and report them as "mixed" instead.
1005 .. option:: randrepeat=bool
1007 Seed the random number generator used for random I/O patterns in a
1008 predictable way so the pattern is repeatable across runs. Default: true.
1010 .. option:: allrandrepeat=bool
1012 Seed all random number generators in a predictable way so results are
1013 repeatable across runs. Default: false.
1015 .. option:: randseed=int
1017 Seed the random number generators based on this seed value, to be able to
1018 control what sequence of output is being generated. If not set, the random
1019 sequence depends on the :option:`randrepeat` setting.
1021 .. option:: fallocate=str
1023 Whether pre-allocation is performed when laying down files.
1024 Accepted values are:
1027 Do not pre-allocate space.
1030 Pre-allocate via :manpage:`posix_fallocate(3)`.
1033 Pre-allocate via :manpage:`fallocate(2)` with
1034 FALLOC_FL_KEEP_SIZE set.
1037 Backward-compatible alias for **none**.
1040 Backward-compatible alias for **posix**.
1042 May not be available on all supported platforms. **keep** is only available
1043 on Linux. If using ZFS on Solaris this must be set to **none** because ZFS
1044 doesn't support it. Default: **posix**.
1046 .. option:: fadvise_hint=str
1048 Use :manpage:`posix_fadvise(2)` to advise the kernel on what I/O patterns
1049 are likely to be issued. Accepted values are:
1052 Backwards-compatible hint for "no hint".
1055 Backwards compatible hint for "advise with fio workload type". This
1056 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1057 for a sequential workload.
1060 Advise using **FADV_SEQUENTIAL**.
1063 Advise using **FADV_RANDOM**.
1065 .. option:: fadvise_stream=int
1067 Use :manpage:`posix_fadvise(2)` to advise the kernel what stream ID the
1068 writes issued belong to. Only supported on Linux. Note, this option may
1069 change going forward.
1071 .. option:: offset=int
1073 Start I/O at the given offset in the file. The data before the given offset
1074 will not be touched. This effectively caps the file size at `real_size -
1075 offset`. Can be combined with :option:`size` to constrain the start and
1076 end range that I/O will be done within.
1078 .. option:: offset_increment=int
1080 If this is provided, then the real offset becomes `offset + offset_increment
1081 * thread_number`, where the thread number is a counter that starts at 0 and
1082 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1083 specified). This option is useful if there are several jobs which are
1084 intended to operate on a file in parallel disjoint segments, with even
1085 spacing between the starting points.
1087 .. option:: number_ios=int
1089 Fio will normally perform I/Os until it has exhausted the size of the region
1090 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1091 condition). With this setting, the range/size can be set independently of
1092 the number of I/Os to perform. When fio reaches this number, it will exit
1093 normally and report status. Note that this does not extend the amount of I/O
1094 that will be done, it will only stop fio if this condition is met before
1095 other end-of-job criteria.
1097 .. option:: fsync=int
1099 If writing to a file, issue a sync of the dirty data for every number of
1100 blocks given. For example, if you give 32 as a parameter, fio will sync the
1101 file for every 32 writes issued. If fio is using non-buffered I/O, we may
1102 not sync the file. The exception is the sg I/O engine, which synchronizes
1103 the disk cache anyway. Defaults to 0, which means no sync every certain
1106 .. option:: fdatasync=int
1108 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
1109 not metadata blocks. In Windows, FreeBSD, and DragonFlyBSD there is no
1110 :manpage:`fdatasync(2)`, this falls back to using :manpage:`fsync(2)`.
1111 Defaults to 0, which means no sync data every certain number of writes.
1113 .. option:: write_barrier=int
1115 Make every `N-th` write a barrier write.
1117 .. option:: sync_file_range=str:val
1119 Use :manpage:`sync_file_range(2)` for every `val` number of write
1120 operations. Fio will track range of writes that have happened since the last
1121 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1124 SYNC_FILE_RANGE_WAIT_BEFORE
1126 SYNC_FILE_RANGE_WRITE
1128 SYNC_FILE_RANGE_WAIT_AFTER
1130 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1131 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1132 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1135 .. option:: overwrite=bool
1137 If true, writes to a file will always overwrite existing data. If the file
1138 doesn't already exist, it will be created before the write phase begins. If
1139 the file exists and is large enough for the specified write phase, nothing
1142 .. option:: end_fsync=bool
1144 If true, fsync file contents when a write stage has completed.
1146 .. option:: fsync_on_close=bool
1148 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
1149 from end_fsync in that it will happen on every file close, not just at the
1152 .. option:: rwmixread=int
1154 Percentage of a mixed workload that should be reads. Default: 50.
1156 .. option:: rwmixwrite=int
1158 Percentage of a mixed workload that should be writes. If both
1159 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1160 add up to 100%, the latter of the two will be used to override the
1161 first. This may interfere with a given rate setting, if fio is asked to
1162 limit reads or writes to a certain rate. If that is the case, then the
1163 distribution may be skewed. Default: 50.
1165 .. option:: random_distribution=str:float[,str:float][,str:float]
1167 By default, fio will use a completely uniform random distribution when asked
1168 to perform random I/O. Sometimes it is useful to skew the distribution in
1169 specific ways, ensuring that some parts of the data is more hot than others.
1170 fio includes the following distribution models:
1173 Uniform random distribution
1182 Normal (Gaussian) distribution
1185 Zoned random distribution
1187 When using a **zipf** or **pareto** distribution, an input value is also
1188 needed to define the access pattern. For **zipf**, this is the `zipf
1189 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
1190 program, :command:`genzipf`, that can be used visualize what the given input
1191 values will yield in terms of hit rates. If you wanted to use **zipf** with
1192 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1193 option. If a non-uniform model is used, fio will disable use of the random
1194 map. For the **gauss** distribution, a normal deviation is supplied as a
1195 value between 0 and 100.
1197 For a **zoned** distribution, fio supports specifying percentages of I/O
1198 access that should fall within what range of the file or device. For
1199 example, given a criteria of:
1201 * 60% of accesses should be to the first 10%
1202 * 30% of accesses should be to the next 20%
1203 * 8% of accesses should be to to the next 30%
1204 * 2% of accesses should be to the next 40%
1206 we can define that through zoning of the random accesses. For the above
1207 example, the user would do::
1209 random_distribution=zoned:60/10:30/20:8/30:2/40
1211 similarly to how :option:`bssplit` works for setting ranges and percentages
1212 of block sizes. Like :option:`bssplit`, it's possible to specify separate
1213 zones for reads, writes, and trims. If just one set is given, it'll apply to
1216 .. option:: percentage_random=int[,int][,int]
1218 For a random workload, set how big a percentage should be random. This
1219 defaults to 100%, in which case the workload is fully random. It can be set
1220 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1221 sequential. Any setting in between will result in a random mix of sequential
1222 and random I/O, at the given percentages. Comma-separated values may be
1223 specified for reads, writes, and trims as described in :option:`blocksize`.
1225 .. option:: norandommap
1227 Normally fio will cover every block of the file when doing random I/O. If
1228 this option is given, fio will just get a new random offset without looking
1229 at past I/O history. This means that some blocks may not be read or written,
1230 and that some blocks may be read/written more than once. If this option is
1231 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1232 only intact blocks are verified, i.e., partially-overwritten blocks are
1235 .. option:: softrandommap=bool
1237 See :option:`norandommap`. If fio runs with the random block map enabled and
1238 it fails to allocate the map, if this option is set it will continue without
1239 a random block map. As coverage will not be as complete as with random maps,
1240 this option is disabled by default.
1242 .. option:: random_generator=str
1244 Fio supports the following engines for generating
1245 I/O offsets for random I/O:
1248 Strong 2^88 cycle random number generator
1250 Linear feedback shift register generator
1252 Strong 64-bit 2^258 cycle random number generator
1254 **tausworthe** is a strong random number generator, but it requires tracking
1255 on the side if we want to ensure that blocks are only read or written
1256 once. **LFSR** guarantees that we never generate the same offset twice, and
1257 it's also less computationally expensive. It's not a true random generator,
1258 however, though for I/O purposes it's typically good enough. **LFSR** only
1259 works with single block sizes, not with workloads that use multiple block
1260 sizes. If used with such a workload, fio may read or write some blocks
1261 multiple times. The default value is **tausworthe**, unless the required
1262 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1263 selected automatically.
1269 .. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1271 The block size in bytes used for I/O units. Default: 4096. A single value
1272 applies to reads, writes, and trims. Comma-separated values may be
1273 specified for reads, writes, and trims. A value not terminated in a comma
1274 applies to subsequent types.
1279 means 256k for reads, writes and trims.
1282 means 8k for reads, 32k for writes and trims.
1285 means 8k for reads, 32k for writes, and default for trims.
1288 means default for reads, 8k for writes and trims.
1291 means default for reads, 8k for writes, and default for writes.
1293 .. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1295 A range of block sizes in bytes for I/O units. The issued I/O unit will
1296 always be a multiple of the minimum size, unless
1297 :option:`blocksize_unaligned` is set.
1299 Comma-separated ranges may be specified for reads, writes, and trims as
1300 described in :option:`blocksize`.
1302 Example: ``bsrange=1k-4k,2k-8k``.
1304 .. option:: bssplit=str[,str][,str]
1306 Sometimes you want even finer grained control of the block sizes issued, not
1307 just an even split between them. This option allows you to weight various
1308 block sizes, so that you are able to define a specific amount of block sizes
1309 issued. The format for this option is::
1311 bssplit=blocksize/percentage:blocksize/percentage
1313 for as many block sizes as needed. So if you want to define a workload that
1314 has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would write::
1316 bssplit=4k/10:64k/50:32k/40
1318 Ordering does not matter. If the percentage is left blank, fio will fill in
1319 the remaining values evenly. So a bssplit option like this one::
1321 bssplit=4k/50:1k/:32k/
1323 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always add up
1324 to 100, if bssplit is given a range that adds up to more, it will error out.
1326 Comma-separated values may be specified for reads, writes, and trims as
1327 described in :option:`blocksize`.
1329 If you want a workload that has 50% 2k reads and 50% 4k reads, while having
1330 90% 4k writes and 10% 8k writes, you would specify::
1332 bssplit=2k/50:4k/50,4k/90,8k/10
1334 .. option:: blocksize_unaligned, bs_unaligned
1336 If set, fio will issue I/O units with any size within
1337 :option:`blocksize_range`, not just multiples of the minimum size. This
1338 typically won't work with direct I/O, as that normally requires sector
1341 .. option:: bs_is_seq_rand
1343 If this option is set, fio will use the normal read,write blocksize settings
1344 as sequential,random blocksize settings instead. Any random read or write
1345 will use the WRITE blocksize settings, and any sequential read or write will
1346 use the READ blocksize settings.
1348 .. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1350 Boundary to which fio will align random I/O units. Default:
1351 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1352 I/O, though it usually depends on the hardware block size. This option is
1353 mutually exclusive with using a random map for files, so it will turn off
1354 that option. Comma-separated values may be specified for reads, writes, and
1355 trims as described in :option:`blocksize`.
1361 .. option:: zero_buffers
1363 Initialize buffers with all zeros. Default: fill buffers with random data.
1365 .. option:: refill_buffers
1367 If this option is given, fio will refill the I/O buffers on every
1368 submit. The default is to only fill it at init time and reuse that
1369 data. Only makes sense if zero_buffers isn't specified, naturally. If data
1370 verification is enabled, `refill_buffers` is also automatically enabled.
1372 .. option:: scramble_buffers=bool
1374 If :option:`refill_buffers` is too costly and the target is using data
1375 deduplication, then setting this option will slightly modify the I/O buffer
1376 contents to defeat normal de-dupe attempts. This is not enough to defeat
1377 more clever block compression attempts, but it will stop naive dedupe of
1378 blocks. Default: true.
1380 .. option:: buffer_compress_percentage=int
1382 If this is set, then fio will attempt to provide I/O buffer content (on
1383 WRITEs) that compress to the specified level. Fio does this by providing a
1384 mix of random data and a fixed pattern. The fixed pattern is either zeroes,
1385 or the pattern specified by :option:`buffer_pattern`. If the pattern option
1386 is used, it might skew the compression ratio slightly. Note that this is per
1387 block size unit, for file/disk wide compression level that matches this
1388 setting, you'll also want to set :option:`refill_buffers`.
1390 .. option:: buffer_compress_chunk=int
1392 See :option:`buffer_compress_percentage`. This setting allows fio to manage
1393 how big the ranges of random data and zeroed data is. Without this set, fio
1394 will provide :option:`buffer_compress_percentage` of blocksize random data,
1395 followed by the remaining zeroed. With this set to some chunk size smaller
1396 than the block size, fio can alternate random and zeroed data throughout the
1399 .. option:: buffer_pattern=str
1401 If set, fio will fill the I/O buffers with this pattern. If not set, the
1402 contents of I/O buffers is defined by the other options related to buffer
1403 contents. The setting can be any pattern of bytes, and can be prefixed with
1404 0x for hex values. It may also be a string, where the string must then be
1405 wrapped with ``""``, e.g.::
1407 buffer_pattern="abcd"
1415 buffer_pattern=0xdeadface
1417 Also you can combine everything together in any order::
1419 buffer_pattern=0xdeadface"abcd"-12
1421 .. option:: dedupe_percentage=int
1423 If set, fio will generate this percentage of identical buffers when
1424 writing. These buffers will be naturally dedupable. The contents of the
1425 buffers depend on what other buffer compression settings have been set. It's
1426 possible to have the individual buffers either fully compressible, or not at
1427 all. This option only controls the distribution of unique buffers.
1429 .. option:: invalidate=bool
1431 Invalidate the buffer/page cache parts for this file prior to starting
1432 I/O if the platform and file type support it. Defaults to true.
1433 This will be ignored if :option:`pre_read` is also specified for the
1436 .. option:: sync=bool
1438 Use synchronous I/O for buffered writes. For the majority of I/O engines,
1439 this means using O_SYNC. Default: false.
1441 .. option:: iomem=str, mem=str
1443 Fio can use various types of memory as the I/O unit buffer. The allowed
1447 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1451 Use shared memory as the buffers. Allocated through
1452 :manpage:`shmget(2)`.
1455 Same as shm, but use huge pages as backing.
1458 Use mmap to allocate buffers. May either be anonymous memory, or can
1459 be file backed if a filename is given after the option. The format
1460 is `mem=mmap:/path/to/file`.
1463 Use a memory mapped huge file as the buffer backing. Append filename
1464 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1467 Same as mmap, but use a MMAP_SHARED mapping.
1469 The area allocated is a function of the maximum allowed bs size for the job,
1470 multiplied by the I/O depth given. Note that for **shmhuge** and
1471 **mmaphuge** to work, the system must have free huge pages allocated. This
1472 can normally be checked and set by reading/writing
1473 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1474 is 4MiB in size. So to calculate the number of huge pages you need for a
1475 given job file, add up the I/O depth of all jobs (normally one unless
1476 :option:`iodepth` is used) and multiply by the maximum bs set. Then divide
1477 that number by the huge page size. You can see the size of the huge pages in
1478 :file:`/proc/meminfo`. If no huge pages are allocated by having a non-zero
1479 number in `nr_hugepages`, using **mmaphuge** or **shmhuge** will fail. Also
1480 see :option:`hugepage-size`.
1482 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1483 should point there. So if it's mounted in :file:`/huge`, you would use
1484 `mem=mmaphuge:/huge/somefile`.
1486 .. option:: iomem_align=int
1488 This indicates the memory alignment of the I/O memory buffers. Note that
1489 the given alignment is applied to the first I/O unit buffer, if using
1490 :option:`iodepth` the alignment of the following buffers are given by the
1491 :option:`bs` used. In other words, if using a :option:`bs` that is a
1492 multiple of the page sized in the system, all buffers will be aligned to
1493 this value. If using a :option:`bs` that is not page aligned, the alignment
1494 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1497 .. option:: hugepage-size=int
1499 Defines the size of a huge page. Must at least be equal to the system
1500 setting, see :file:`/proc/meminfo`. Defaults to 4MiB. Should probably
1501 always be a multiple of megabytes, so using ``hugepage-size=Xm`` is the
1502 preferred way to set this to avoid setting a non-pow-2 bad value.
1504 .. option:: lockmem=int
1506 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1507 simulate a smaller amount of memory. The amount specified is per worker.
1513 .. option:: size=int
1515 The total size of file I/O for each thread of this job. Fio will run until
1516 this many bytes has been transferred, unless runtime is limited by other options
1517 (such as :option:`runtime`, for instance, or increased/decreased by :option:`io_size`).
1518 Fio will divide this size between the available files determined by options
1519 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1520 specified by the job. If the result of division happens to be 0, the size is
1521 set to the physical size of the given files or devices if they exist.
1522 If this option is not specified, fio will use the full size of the given
1523 files or devices. If the files do not exist, size must be given. It is also
1524 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1525 given, fio will use 20% of the full size of the given files or devices.
1526 Can be combined with :option:`offset` to constrain the start and end range
1527 that I/O will be done within.
1529 .. option:: io_size=int, io_limit=int
1531 Normally fio operates within the region set by :option:`size`, which means
1532 that the :option:`size` option sets both the region and size of I/O to be
1533 performed. Sometimes that is not what you want. With this option, it is
1534 possible to define just the amount of I/O that fio should do. For instance,
1535 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1536 will perform I/O within the first 20GiB but exit when 5GiB have been
1537 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1538 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1539 the 0..20GiB region.
1541 .. option:: filesize=int
1543 Individual file sizes. May be a range, in which case fio will select sizes
1544 for files at random within the given range and limited to :option:`size` in
1545 total (if that is given). If not given, each created file is the same size.
1546 This option overrides :option:`size` in terms of file size, which means
1547 this value is used as a fixed size or possible range of each file.
1549 .. option:: file_append=bool
1551 Perform I/O after the end of the file. Normally fio will operate within the
1552 size of a file. If this option is set, then fio will append to the file
1553 instead. This has identical behavior to setting :option:`offset` to the size
1554 of a file. This option is ignored on non-regular files.
1556 .. option:: fill_device=bool, fill_fs=bool
1558 Sets size to something really large and waits for ENOSPC (no space left on
1559 device) as the terminating condition. Only makes sense with sequential
1560 write. For a read workload, the mount point will be filled first then I/O
1561 started on the result. This option doesn't make sense if operating on a raw
1562 device node, since the size of that is already known by the file system.
1563 Additionally, writing beyond end-of-device will not return ENOSPC there.
1569 .. option:: ioengine=str
1571 Defines how the job issues I/O to the file. The following types are defined:
1574 Basic :manpage:`read(2)` or :manpage:`write(2)`
1575 I/O. :manpage:`lseek(2)` is used to position the I/O location.
1576 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
1579 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1580 all supported operating systems except for Windows.
1583 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
1584 queuing by coalescing adjacent I/Os into a single submission.
1587 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1590 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1593 Linux native asynchronous I/O. Note that Linux may only support
1594 queued behaviour with non-buffered I/O (set ``direct=1`` or
1596 This engine defines engine specific options.
1599 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
1600 :manpage:`aio_write(3)`.
1603 Solaris native asynchronous I/O.
1606 Windows native asynchronous I/O. Default on Windows.
1609 File is memory mapped with :manpage:`mmap(2)` and data copied
1610 to/from using :manpage:`memcpy(3)`.
1613 :manpage:`splice(2)` is used to transfer the data and
1614 :manpage:`vmsplice(2)` to transfer data from user space to the
1618 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
1619 ioctl, or if the target is an sg character device we use
1620 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
1621 I/O. Requires filename option to specify either block or character
1625 Doesn't transfer any data, just pretends to. This is mainly used to
1626 exercise fio itself and for debugging/testing purposes.
1629 Transfer over the network to given ``host:port``. Depending on the
1630 :option:`protocol` used, the :option:`hostname`, :option:`port`,
1631 :option:`listen` and :option:`filename` options are used to specify
1632 what sort of connection to make, while the :option:`protocol` option
1633 determines which protocol will be used. This engine defines engine
1637 Like **net**, but uses :manpage:`splice(2)` and
1638 :manpage:`vmsplice(2)` to map data and send/receive.
1639 This engine defines engine specific options.
1642 Doesn't transfer any data, but burns CPU cycles according to the
1643 :option:`cpuload` and :option:`cpuchunks` options. Setting
1644 :option:`cpuload` =85 will cause that job to do nothing but burn 85%
1645 of the CPU. In case of SMP machines, use :option:`numjobs`
1646 =<no_of_cpu> to get desired CPU usage, as the cpuload only loads a
1647 single CPU at the desired rate. A job never finishes unless there is
1648 at least one non-cpuio job.
1651 The GUASI I/O engine is the Generic Userspace Asyncronous Syscall
1652 Interface approach to async I/O. See
1654 http://www.xmailserver.org/guasi-lib.html
1656 for more info on GUASI.
1659 The RDMA I/O engine supports both RDMA memory semantics
1660 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
1661 InfiniBand, RoCE and iWARP protocols.
1664 I/O engine that does regular fallocate to simulate data transfer as
1668 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
1671 does fallocate(,mode = 0).
1674 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
1677 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
1678 defragment activity in request to DDIR_WRITE event.
1681 I/O engine supporting direct access to Ceph Rados Block Devices
1682 (RBD) via librbd without the need to use the kernel rbd driver. This
1683 ioengine defines engine specific options.
1686 Using Glusterfs libgfapi sync interface to direct access to
1687 Glusterfs volumes without having to go through FUSE. This ioengine
1688 defines engine specific options.
1691 Using Glusterfs libgfapi async interface to direct access to
1692 Glusterfs volumes without having to go through FUSE. This ioengine
1693 defines engine specific options.
1696 Read and write through Hadoop (HDFS). The :file:`filename` option
1697 is used to specify host,port of the hdfs name-node to connect. This
1698 engine interprets offsets a little differently. In HDFS, files once
1699 created cannot be modified. So random writes are not possible. To
1700 imitate this, libhdfs engine expects bunch of small files to be
1701 created over HDFS, and engine will randomly pick a file out of those
1702 files based on the offset generated by fio backend. (see the example
1703 job file to create such files, use ``rw=write`` option). Please
1704 note, you might want to set necessary environment variables to work
1705 with hdfs/libhdfs properly. Each job uses its own connection to
1709 Read, write and erase an MTD character device (e.g.,
1710 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
1711 underlying device type, the I/O may have to go in a certain pattern,
1712 e.g., on NAND, writing sequentially to erase blocks and discarding
1713 before overwriting. The writetrim mode works well for this
1717 Read and write using filesystem DAX to a file on a filesystem
1718 mounted with DAX on a persistent memory device through the NVML
1722 Read and write using device DAX to a persistent memory device (e.g.,
1723 /dev/dax0.0) through the NVML libpmem library.
1726 Prefix to specify loading an external I/O engine object file. Append
1727 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
1728 ioengine :file:`foo.o` in :file:`/tmp`.
1731 I/O engine specific parameters
1732 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1734 In addition, there are some parameters which are only valid when a specific
1735 ioengine is in use. These are used identically to normal parameters, with the
1736 caveat that when used on the command line, they must come after the
1737 :option:`ioengine` that defines them is selected.
1739 .. option:: userspace_reap : [libaio]
1741 Normally, with the libaio engine in use, fio will use the
1742 :manpage:`io_getevents(2)` system call to reap newly returned events. With
1743 this flag turned on, the AIO ring will be read directly from user-space to
1744 reap events. The reaping mode is only enabled when polling for a minimum of
1745 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
1747 .. option:: hipri : [pvsync2]
1749 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
1752 .. option:: cpuload=int : [cpuio]
1754 Attempt to use the specified percentage of CPU cycles. This is a mandatory
1755 option when using cpuio I/O engine.
1757 .. option:: cpuchunks=int : [cpuio]
1759 Split the load into cycles of the given time. In microseconds.
1761 .. option:: exit_on_io_done=bool : [cpuio]
1763 Detect when I/O threads are done, then exit.
1765 .. option:: hostname=str : [netsplice] [net]
1767 The host name or IP address to use for TCP or UDP based I/O. If the job is
1768 a TCP listener or UDP reader, the host name is not used and must be omitted
1769 unless it is a valid UDP multicast address.
1771 .. option:: namenode=str : [libhdfs]
1773 The host name or IP address of a HDFS cluster namenode to contact.
1775 .. option:: port=int
1779 The TCP or UDP port to bind to or connect to. If this is used with
1780 :option:`numjobs` to spawn multiple instances of the same job type, then
1781 this will be the starting port number since fio will use a range of
1786 the listening port of the HFDS cluster namenode.
1788 .. option:: interface=str : [netsplice] [net]
1790 The IP address of the network interface used to send or receive UDP
1793 .. option:: ttl=int : [netsplice] [net]
1795 Time-to-live value for outgoing UDP multicast packets. Default: 1.
1797 .. option:: nodelay=bool : [netsplice] [net]
1799 Set TCP_NODELAY on TCP connections.
1801 .. option:: protocol=str : [netsplice] [net]
1803 .. option:: proto=str : [netsplice] [net]
1805 The network protocol to use. Accepted values are:
1808 Transmission control protocol.
1810 Transmission control protocol V6.
1812 User datagram protocol.
1814 User datagram protocol V6.
1818 When the protocol is TCP or UDP, the port must also be given, as well as the
1819 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
1820 normal filename option should be used and the port is invalid.
1822 .. option:: listen : [net]
1824 For TCP network connections, tell fio to listen for incoming connections
1825 rather than initiating an outgoing connection. The :option:`hostname` must
1826 be omitted if this option is used.
1828 .. option:: pingpong : [net]
1830 Normally a network writer will just continue writing data, and a network
1831 reader will just consume packages. If ``pingpong=1`` is set, a writer will
1832 send its normal payload to the reader, then wait for the reader to send the
1833 same payload back. This allows fio to measure network latencies. The
1834 submission and completion latencies then measure local time spent sending or
1835 receiving, and the completion latency measures how long it took for the
1836 other end to receive and send back. For UDP multicast traffic
1837 ``pingpong=1`` should only be set for a single reader when multiple readers
1838 are listening to the same address.
1840 .. option:: window_size : [net]
1842 Set the desired socket buffer size for the connection.
1844 .. option:: mss : [net]
1846 Set the TCP maximum segment size (TCP_MAXSEG).
1848 .. option:: donorname=str : [e4defrag]
1850 File will be used as a block donor(swap extents between files).
1852 .. option:: inplace=int : [e4defrag]
1854 Configure donor file blocks allocation strategy:
1857 Default. Preallocate donor's file on init.
1859 Allocate space immediately inside defragment event, and free right
1862 .. option:: clustername=str : [rbd]
1864 Specifies the name of the Ceph cluster.
1866 .. option:: rbdname=str : [rbd]
1868 Specifies the name of the RBD.
1870 .. option:: pool=str : [rbd]
1872 Specifies the name of the Ceph pool containing RBD.
1874 .. option:: clientname=str : [rbd]
1876 Specifies the username (without the 'client.' prefix) used to access the
1877 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
1878 the full *type.id* string. If no type. prefix is given, fio will add
1879 'client.' by default.
1881 .. option:: skip_bad=bool : [mtd]
1883 Skip operations against known bad blocks.
1885 .. option:: hdfsdirectory : [libhdfs]
1887 libhdfs will create chunk in this HDFS directory.
1889 .. option:: chunk_size : [libhdfs]
1891 the size of the chunk to use for each file.
1897 .. option:: iodepth=int
1899 Number of I/O units to keep in flight against the file. Note that
1900 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
1901 for small degrees when :option:`verify_async` is in use). Even async
1902 engines may impose OS restrictions causing the desired depth not to be
1903 achieved. This may happen on Linux when using libaio and not setting
1904 :option:`direct` =1, since buffered I/O is not async on that OS. Keep an
1905 eye on the I/O depth distribution in the fio output to verify that the
1906 achieved depth is as expected. Default: 1.
1908 .. option:: iodepth_batch_submit=int, iodepth_batch=int
1910 This defines how many pieces of I/O to submit at once. It defaults to 1
1911 which means that we submit each I/O as soon as it is available, but can be
1912 raised to submit bigger batches of I/O at the time. If it is set to 0 the
1913 :option:`iodepth` value will be used.
1915 .. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
1917 This defines how many pieces of I/O to retrieve at once. It defaults to 1
1918 which means that we'll ask for a minimum of 1 I/O in the retrieval process
1919 from the kernel. The I/O retrieval will go on until we hit the limit set by
1920 :option:`iodepth_low`. If this variable is set to 0, then fio will always
1921 check for completed events before queuing more I/O. This helps reduce I/O
1922 latency, at the cost of more retrieval system calls.
1924 .. option:: iodepth_batch_complete_max=int
1926 This defines maximum pieces of I/O to retrieve at once. This variable should
1927 be used along with :option:`iodepth_batch_complete_min` =int variable,
1928 specifying the range of min and max amount of I/O which should be
1929 retrieved. By default it is equal to :option:`iodepth_batch_complete_min`
1934 iodepth_batch_complete_min=1
1935 iodepth_batch_complete_max=<iodepth>
1937 which means that we will retrieve at least 1 I/O and up to the whole
1938 submitted queue depth. If none of I/O has been completed yet, we will wait.
1942 iodepth_batch_complete_min=0
1943 iodepth_batch_complete_max=<iodepth>
1945 which means that we can retrieve up to the whole submitted queue depth, but
1946 if none of I/O has been completed yet, we will NOT wait and immediately exit
1947 the system call. In this example we simply do polling.
1949 .. option:: iodepth_low=int
1951 The low water mark indicating when to start filling the queue
1952 again. Defaults to the same as :option:`iodepth`, meaning that fio will
1953 attempt to keep the queue full at all times. If :option:`iodepth` is set to
1954 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
1955 16 requests, it will let the depth drain down to 4 before starting to fill
1958 .. option:: io_submit_mode=str
1960 This option controls how fio submits the I/O to the I/O engine. The default
1961 is `inline`, which means that the fio job threads submit and reap I/O
1962 directly. If set to `offload`, the job threads will offload I/O submission
1963 to a dedicated pool of I/O threads. This requires some coordination and thus
1964 has a bit of extra overhead, especially for lower queue depth I/O where it
1965 can increase latencies. The benefit is that fio can manage submission rates
1966 independently of the device completion rates. This avoids skewed latency
1967 reporting if I/O gets back up on the device side (the coordinated omission
1974 .. option:: thinktime=time
1976 Stall the job for the specified period of time after an I/O has completed before issuing the
1977 next. May be used to simulate processing being done by an application.
1978 When the unit is omitted, the value is given in microseconds. See
1979 :option:`thinktime_blocks` and :option:`thinktime_spin`.
1981 .. option:: thinktime_spin=time
1983 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
1984 something with the data received, before falling back to sleeping for the
1985 rest of the period specified by :option:`thinktime`. When the unit is
1986 omitted, the value is given in microseconds.
1988 .. option:: thinktime_blocks=int
1990 Only valid if :option:`thinktime` is set - control how many blocks to issue,
1991 before waiting `thinktime` usecs. If not set, defaults to 1 which will make
1992 fio wait `thinktime` usecs after every block. This effectively makes any
1993 queue depth setting redundant, since no more than 1 I/O will be queued
1994 before we have to complete it and do our thinktime. In other words, this
1995 setting effectively caps the queue depth if the latter is larger.
1997 .. option:: rate=int[,int][,int]
1999 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
2000 suffix rules apply. Comma-separated values may be specified for reads,
2001 writes, and trims as described in :option:`blocksize`.
2003 .. option:: rate_min=int[,int][,int]
2005 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
2006 to meet this requirement will cause the job to exit. Comma-separated values
2007 may be specified for reads, writes, and trims as described in
2008 :option:`blocksize`.
2010 .. option:: rate_iops=int[,int][,int]
2012 Cap the bandwidth to this number of IOPS. Basically the same as
2013 :option:`rate`, just specified independently of bandwidth. If the job is
2014 given a block size range instead of a fixed value, the smallest block size
2015 is used as the metric. Comma-separated values may be specified for reads,
2016 writes, and trims as described in :option:`blocksize`.
2018 .. option:: rate_iops_min=int[,int][,int]
2020 If fio doesn't meet this rate of I/O, it will cause the job to exit.
2021 Comma-separated values may be specified for reads, writes, and trims as
2022 described in :option:`blocksize`.
2024 .. option:: rate_process=str
2026 This option controls how fio manages rated I/O submissions. The default is
2027 `linear`, which submits I/O in a linear fashion with fixed delays between
2028 I/Os that gets adjusted based on I/O completion rates. If this is set to
2029 `poisson`, fio will submit I/O based on a more real world random request
2030 flow, known as the Poisson process
2031 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
2032 10^6 / IOPS for the given workload.
2038 .. option:: latency_target=time
2040 If set, fio will attempt to find the max performance point that the given
2041 workload will run at while maintaining a latency below this target. When
2042 the unit is omitted, the value is given in microseconds. See
2043 :option:`latency_window` and :option:`latency_percentile`.
2045 .. option:: latency_window=time
2047 Used with :option:`latency_target` to specify the sample window that the job
2048 is run at varying queue depths to test the performance. When the unit is
2049 omitted, the value is given in microseconds.
2051 .. option:: latency_percentile=float
2053 The percentage of I/Os that must fall within the criteria specified by
2054 :option:`latency_target` and :option:`latency_window`. If not set, this
2055 defaults to 100.0, meaning that all I/Os must be equal or below to the value
2056 set by :option:`latency_target`.
2058 .. option:: max_latency=time
2060 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
2061 maximum latency. When the unit is omitted, the value is given in
2064 .. option:: rate_cycle=int
2066 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
2073 .. option:: write_iolog=str
2075 Write the issued I/O patterns to the specified file. See
2076 :option:`read_iolog`. Specify a separate file for each job, otherwise the
2077 iologs will be interspersed and the file may be corrupt.
2079 .. option:: read_iolog=str
2081 Open an iolog with the specified file name and replay the I/O patterns it
2082 contains. This can be used to store a workload and replay it sometime
2083 later. The iolog given may also be a blktrace binary file, which allows fio
2084 to replay a workload captured by :command:`blktrace`. See
2085 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
2086 replay, the file needs to be turned into a blkparse binary data file first
2087 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
2089 .. option:: replay_no_stall=int
2091 When replaying I/O with :option:`read_iolog` the default behavior is to
2092 attempt to respect the time stamps within the log and replay them with the
2093 appropriate delay between IOPS. By setting this variable fio will not
2094 respect the timestamps and attempt to replay them as fast as possible while
2095 still respecting ordering. The result is the same I/O pattern to a given
2096 device, but different timings.
2098 .. option:: replay_redirect=str
2100 While replaying I/O patterns using :option:`read_iolog` the default behavior
2101 is to replay the IOPS onto the major/minor device that each IOP was recorded
2102 from. This is sometimes undesirable because on a different machine those
2103 major/minor numbers can map to a different device. Changing hardware on the
2104 same system can also result in a different major/minor mapping.
2105 ``replay_redirect`` causes all IOPS to be replayed onto the single specified
2106 device regardless of the device it was recorded
2107 from. i.e. :option:`replay_redirect` = :file:`/dev/sdc` would cause all I/O
2108 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
2109 multiple devices will be replayed onto a single device, if the trace
2110 contains multiple devices. If you want multiple devices to be replayed
2111 concurrently to multiple redirected devices you must blkparse your trace
2112 into separate traces and replay them with independent fio invocations.
2113 Unfortunately this also breaks the strict time ordering between multiple
2116 .. option:: replay_align=int
2118 Force alignment of I/O offsets and lengths in a trace to this power of 2
2121 .. option:: replay_scale=int
2123 Scale sector offsets down by this factor when replaying traces.
2126 Threads, processes and job synchronization
2127 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2131 Fio defaults to forking jobs, however if this option is given, fio will use
2132 POSIX Threads function :manpage:`pthread_create(3)` to create threads instead
2133 of forking processes.
2135 .. option:: wait_for=str
2137 Specifies the name of the already defined job to wait for. Single waitee
2138 name only may be specified. If set, the job won't be started until all
2139 workers of the waitee job are done.
2141 ``wait_for`` operates on the job name basis, so there are a few
2142 limitations. First, the waitee must be defined prior to the waiter job
2143 (meaning no forward references). Second, if a job is being referenced as a
2144 waitee, it must have a unique name (no duplicate waitees).
2146 .. option:: nice=int
2148 Run the job with the given nice value. See man :manpage:`nice(2)`.
2150 On Windows, values less than -15 set the process class to "High"; -1 through
2151 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
2154 .. option:: prio=int
2156 Set the I/O priority value of this job. Linux limits us to a positive value
2157 between 0 and 7, with 0 being the highest. See man
2158 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
2159 systems since meaning of priority may differ.
2161 .. option:: prioclass=int
2163 Set the I/O priority class. See man :manpage:`ionice(1)`.
2165 .. option:: cpumask=int
2167 Set the CPU affinity of this job. The parameter given is a bitmask of
2168 allowed CPU's the job may run on. So if you want the allowed CPUs to be 1
2169 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
2170 :manpage:`sched_setaffinity(2)`. This may not work on all supported
2171 operating systems or kernel versions. This option doesn't work well for a
2172 higher CPU count than what you can store in an integer mask, so it can only
2173 control cpus 1-32. For boxes with larger CPU counts, use
2174 :option:`cpus_allowed`.
2176 .. option:: cpus_allowed=str
2178 Controls the same options as :option:`cpumask`, but it allows a text setting
2179 of the permitted CPUs instead. So to use CPUs 1 and 5, you would specify
2180 ``cpus_allowed=1,5``. This options also allows a range of CPUs. Say you
2181 wanted a binding to CPUs 1, 5, and 8-15, you would set
2182 ``cpus_allowed=1,5,8-15``.
2184 .. option:: cpus_allowed_policy=str
2186 Set the policy of how fio distributes the CPUs specified by
2187 :option:`cpus_allowed` or cpumask. Two policies are supported:
2190 All jobs will share the CPU set specified.
2192 Each job will get a unique CPU from the CPU set.
2194 **shared** is the default behaviour, if the option isn't specified. If
2195 **split** is specified, then fio will will assign one cpu per job. If not
2196 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
2199 .. option:: numa_cpu_nodes=str
2201 Set this job running on specified NUMA nodes' CPUs. The arguments allow
2202 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
2203 numa options support, fio must be built on a system with libnuma-dev(el)
2206 .. option:: numa_mem_policy=str
2208 Set this job's memory policy and corresponding NUMA nodes. Format of the
2213 ``mode`` is one of the following memory policy: ``default``, ``prefer``,
2214 ``bind``, ``interleave``, ``local`` For ``default`` and ``local`` memory
2215 policy, no node is needed to be specified. For ``prefer``, only one node is
2216 allowed. For ``bind`` and ``interleave``, it allow comma delimited list of
2217 numbers, A-B ranges, or `all`.
2219 .. option:: cgroup=str
2221 Add job to this control group. If it doesn't exist, it will be created. The
2222 system must have a mounted cgroup blkio mount point for this to work. If
2223 your system doesn't have it mounted, you can do so with::
2225 # mount -t cgroup -o blkio none /cgroup
2227 .. option:: cgroup_weight=int
2229 Set the weight of the cgroup to this value. See the documentation that comes
2230 with the kernel, allowed values are in the range of 100..1000.
2232 .. option:: cgroup_nodelete=bool
2234 Normally fio will delete the cgroups it has created after the job
2235 completion. To override this behavior and to leave cgroups around after the
2236 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
2237 to inspect various cgroup files after job completion. Default: false.
2239 .. option:: flow_id=int
2241 The ID of the flow. If not specified, it defaults to being a global
2242 flow. See :option:`flow`.
2244 .. option:: flow=int
2246 Weight in token-based flow control. If this value is used, then there is a
2247 'flow counter' which is used to regulate the proportion of activity between
2248 two or more jobs. Fio attempts to keep this flow counter near zero. The
2249 ``flow`` parameter stands for how much should be added or subtracted to the
2250 flow counter on each iteration of the main I/O loop. That is, if one job has
2251 ``flow=8`` and another job has ``flow=-1``, then there will be a roughly 1:8
2252 ratio in how much one runs vs the other.
2254 .. option:: flow_watermark=int
2256 The maximum value that the absolute value of the flow counter is allowed to
2257 reach before the job must wait for a lower value of the counter.
2259 .. option:: flow_sleep=int
2261 The period of time, in microseconds, to wait after the flow watermark has
2262 been exceeded before retrying operations.
2264 .. option:: stonewall, wait_for_previous
2266 Wait for preceding jobs in the job file to exit, before starting this
2267 one. Can be used to insert serialization points in the job file. A stone
2268 wall also implies starting a new reporting group, see
2269 :option:`group_reporting`.
2273 When one job finishes, terminate the rest. The default is to wait for each
2274 job to finish, sometimes that is not the desired action.
2276 .. option:: exec_prerun=str
2278 Before running this job, issue the command specified through
2279 :manpage:`system(3)`. Output is redirected in a file called
2280 :file:`jobname.prerun.txt`.
2282 .. option:: exec_postrun=str
2284 After the job completes, issue the command specified though
2285 :manpage:`system(3)`. Output is redirected in a file called
2286 :file:`jobname.postrun.txt`.
2290 Instead of running as the invoking user, set the user ID to this value
2291 before the thread/process does any work.
2295 Set group ID, see :option:`uid`.
2301 .. option:: verify_only
2303 Do not perform specified workload, only verify data still matches previous
2304 invocation of this workload. This option allows one to check data multiple
2305 times at a later date without overwriting it. This option makes sense only
2306 for workloads that write data, and does not support workloads with the
2307 :option:`time_based` option set.
2309 .. option:: do_verify=bool
2311 Run the verify phase after a write phase. Only valid if :option:`verify` is
2314 .. option:: verify=str
2316 If writing to a file, fio can verify the file contents after each iteration
2317 of the job. Each verification method also implies verification of special
2318 header, which is written to the beginning of each block. This header also
2319 includes meta information, like offset of the block, block number, timestamp
2320 when block was written, etc. :option:`verify` can be combined with
2321 :option:`verify_pattern` option. The allowed values are:
2324 Use an md5 sum of the data area and store it in the header of
2328 Use an experimental crc64 sum of the data area and store it in the
2329 header of each block.
2332 Use a crc32c sum of the data area and store it in the header of each
2336 Use hardware assisted crc32c calculation provided on SSE4.2 enabled
2337 processors. Falls back to regular software crc32c, if not supported
2341 Use a crc32 sum of the data area and store it in the header of each
2345 Use a crc16 sum of the data area and store it in the header of each
2349 Use a crc7 sum of the data area and store it in the header of each
2353 Use xxhash as the checksum function. Generally the fastest software
2354 checksum that fio supports.
2357 Use sha512 as the checksum function.
2360 Use sha256 as the checksum function.
2363 Use optimized sha1 as the checksum function.
2366 Use optimized sha3-224 as the checksum function.
2369 Use optimized sha3-256 as the checksum function.
2372 Use optimized sha3-384 as the checksum function.
2375 Use optimized sha3-512 as the checksum function.
2378 This option is deprecated, since now meta information is included in
2379 generic verification header and meta verification happens by
2380 default. For detailed information see the description of the
2381 :option:`verify` setting. This option is kept because of
2382 compatibility's sake with old configurations. Do not use it.
2385 Verify a strict pattern. Normally fio includes a header with some
2386 basic information and checksumming, but if this option is set, only
2387 the specific pattern set with :option:`verify_pattern` is verified.
2390 Only pretend to verify. Useful for testing internals with
2391 :option:`ioengine` `=null`, not for much else.
2393 This option can be used for repeated burn-in tests of a system to make sure
2394 that the written data is also correctly read back. If the data direction
2395 given is a read or random read, fio will assume that it should verify a
2396 previously written file. If the data direction includes any form of write,
2397 the verify will be of the newly written data.
2399 .. option:: verifysort=bool
2401 If true, fio will sort written verify blocks when it deems it faster to read
2402 them back in a sorted manner. This is often the case when overwriting an
2403 existing file, since the blocks are already laid out in the file system. You
2404 can ignore this option unless doing huge amounts of really fast I/O where
2405 the red-black tree sorting CPU time becomes significant. Default: true.
2407 .. option:: verifysort_nr=int
2409 Pre-load and sort verify blocks for a read workload.
2411 .. option:: verify_offset=int
2413 Swap the verification header with data somewhere else in the block before
2414 writing. It is swapped back before verifying.
2416 .. option:: verify_interval=int
2418 Write the verification header at a finer granularity than the
2419 :option:`blocksize`. It will be written for chunks the size of
2420 ``verify_interval``. :option:`blocksize` should divide this evenly.
2422 .. option:: verify_pattern=str
2424 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
2425 filling with totally random bytes, but sometimes it's interesting to fill
2426 with a known pattern for I/O verification purposes. Depending on the width
2427 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time(it can
2428 be either a decimal or a hex number). The ``verify_pattern`` if larger than
2429 a 32-bit quantity has to be a hex number that starts with either "0x" or
2430 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
2431 format, which means that for each block offset will be written and then
2432 verified back, e.g.::
2436 Or use combination of everything::
2438 verify_pattern=0xff%o"abcd"-12
2440 .. option:: verify_fatal=bool
2442 Normally fio will keep checking the entire contents before quitting on a
2443 block verification failure. If this option is set, fio will exit the job on
2444 the first observed failure. Default: false.
2446 .. option:: verify_dump=bool
2448 If set, dump the contents of both the original data block and the data block
2449 we read off disk to files. This allows later analysis to inspect just what
2450 kind of data corruption occurred. Off by default.
2452 .. option:: verify_async=int
2454 Fio will normally verify I/O inline from the submitting thread. This option
2455 takes an integer describing how many async offload threads to create for I/O
2456 verification instead, causing fio to offload the duty of verifying I/O
2457 contents to one or more separate threads. If using this offload option, even
2458 sync I/O engines can benefit from using an :option:`iodepth` setting higher
2459 than 1, as it allows them to have I/O in flight while verifies are running.
2461 .. option:: verify_async_cpus=str
2463 Tell fio to set the given CPU affinity on the async I/O verification
2464 threads. See :option:`cpus_allowed` for the format used.
2466 .. option:: verify_backlog=int
2468 Fio will normally verify the written contents of a job that utilizes verify
2469 once that job has completed. In other words, everything is written then
2470 everything is read back and verified. You may want to verify continually
2471 instead for a variety of reasons. Fio stores the meta data associated with
2472 an I/O block in memory, so for large verify workloads, quite a bit of memory
2473 would be used up holding this meta data. If this option is enabled, fio will
2474 write only N blocks before verifying these blocks.
2476 .. option:: verify_backlog_batch=int
2478 Control how many blocks fio will verify if :option:`verify_backlog` is
2479 set. If not set, will default to the value of :option:`verify_backlog`
2480 (meaning the entire queue is read back and verified). If
2481 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
2482 blocks will be verified, if ``verify_backlog_batch`` is larger than
2483 :option:`verify_backlog`, some blocks will be verified more than once.
2485 .. option:: verify_state_save=bool
2487 When a job exits during the write phase of a verify workload, save its
2488 current state. This allows fio to replay up until that point, if the verify
2489 state is loaded for the verify read phase. The format of the filename is,
2492 <type>-<jobname>-<jobindex>-verify.state.
2494 <type> is "local" for a local run, "sock" for a client/server socket
2495 connection, and "ip" (192.168.0.1, for instance) for a networked
2496 client/server connection.
2498 .. option:: verify_state_load=bool
2500 If a verify termination trigger was used, fio stores the current write state
2501 of each thread. This can be used at verification time so that fio knows how
2502 far it should verify. Without this information, fio will run a full
2503 verification pass, according to the settings in the job file used.
2505 .. option:: trim_percentage=int
2507 Number of verify blocks to discard/trim.
2509 .. option:: trim_verify_zero=bool
2511 Verify that trim/discarded blocks are returned as zeroes.
2513 .. option:: trim_backlog=int
2515 Verify that trim/discarded blocks are returned as zeroes.
2517 .. option:: trim_backlog_batch=int
2519 Trim this number of I/O blocks.
2521 .. option:: experimental_verify=bool
2523 Enable experimental verification.
2529 .. option:: steadystate=str:float, ss=str:float
2531 Define the criterion and limit for assessing steady state performance. The
2532 first parameter designates the criterion whereas the second parameter sets
2533 the threshold. When the criterion falls below the threshold for the
2534 specified duration, the job will stop. For example, `iops_slope:0.1%` will
2535 direct fio to terminate the job when the least squares regression slope
2536 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
2537 this will apply to all jobs in the group. Below is the list of available
2538 steady state assessment criteria. All assessments are carried out using only
2539 data from the rolling collection window. Threshold limits can be expressed
2540 as a fixed value or as a percentage of the mean in the collection window.
2543 Collect IOPS data. Stop the job if all individual IOPS measurements
2544 are within the specified limit of the mean IOPS (e.g., ``iops:2``
2545 means that all individual IOPS values must be within 2 of the mean,
2546 whereas ``iops:0.2%`` means that all individual IOPS values must be
2547 within 0.2% of the mean IOPS to terminate the job).
2550 Collect IOPS data and calculate the least squares regression
2551 slope. Stop the job if the slope falls below the specified limit.
2554 Collect bandwidth data. Stop the job if all individual bandwidth
2555 measurements are within the specified limit of the mean bandwidth.
2558 Collect bandwidth data and calculate the least squares regression
2559 slope. Stop the job if the slope falls below the specified limit.
2561 .. option:: steadystate_duration=time, ss_dur=time
2563 A rolling window of this duration will be used to judge whether steady state
2564 has been reached. Data will be collected once per second. The default is 0
2565 which disables steady state detection. When the unit is omitted, the
2566 value is given in seconds.
2568 .. option:: steadystate_ramp_time=time, ss_ramp=time
2570 Allow the job to run for the specified duration before beginning data
2571 collection for checking the steady state job termination criterion. The
2572 default is 0. When the unit is omitted, the value is given in seconds.
2575 Measurements and reporting
2576 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2578 .. option:: per_job_logs=bool
2580 If set, this generates bw/clat/iops log with per file private filenames. If
2581 not set, jobs with identical names will share the log filename. Default:
2584 .. option:: group_reporting
2586 It may sometimes be interesting to display statistics for groups of jobs as
2587 a whole instead of for each individual job. This is especially true if
2588 :option:`numjobs` is used; looking at individual thread/process output
2589 quickly becomes unwieldy. To see the final report per-group instead of
2590 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
2591 same reporting group, unless if separated by a :option:`stonewall`, or by
2592 using :option:`new_group`.
2594 .. option:: new_group
2596 Start a new reporting group. See: :option:`group_reporting`. If not given,
2597 all jobs in a file will be part of the same reporting group, unless
2598 separated by a :option:`stonewall`.
2602 By default, fio collects and shows final output results for all jobs
2603 that run. If this option is set to 0, then fio will ignore it in
2604 the final stat output.
2606 .. option:: write_bw_log=str
2608 If given, write a bandwidth log for this job. Can be used to store data of
2609 the bandwidth of the jobs in their lifetime. The included
2610 :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
2611 text files into nice graphs. See :option:`write_lat_log` for behaviour of
2612 given filename. For this option, the postfix is :file:`_bw.x.log`, where `x`
2613 is the index of the job (`1..N`, where `N` is the number of jobs). If
2614 :option:`per_job_logs` is false, then the filename will not include the job
2615 index. See `Log File Formats`_.
2617 .. option:: write_lat_log=str
2619 Same as :option:`write_bw_log`, except that this option stores I/O
2620 submission, completion, and total latencies instead. If no filename is given
2621 with this option, the default filename of :file:`jobname_type.log` is
2622 used. Even if the filename is given, fio will still append the type of
2623 log. So if one specifies::
2627 The actual log names will be :file:`foo_slat.x.log`, :file:`foo_clat.x.log`,
2628 and :file:`foo_lat.x.log`, where `x` is the index of the job (1..N, where N
2629 is the number of jobs). This helps :command:`fio_generate_plot` find the
2630 logs automatically. If :option:`per_job_logs` is false, then the filename
2631 will not include the job index. See `Log File Formats`_.
2633 .. option:: write_hist_log=str
2635 Same as :option:`write_lat_log`, but writes I/O completion latency
2636 histograms. If no filename is given with this option, the default filename
2637 of :file:`jobname_clat_hist.x.log` is used, where `x` is the index of the
2638 job (1..N, where `N` is the number of jobs). Even if the filename is given,
2639 fio will still append the type of log. If :option:`per_job_logs` is false,
2640 then the filename will not include the job index. See `Log File Formats`_.
2642 .. option:: write_iops_log=str
2644 Same as :option:`write_bw_log`, but writes IOPS. If no filename is given
2645 with this option, the default filename of :file:`jobname_type.x.log` is
2646 used,where `x` is the index of the job (1..N, where `N` is the number of
2647 jobs). Even if the filename is given, fio will still append the type of
2648 log. If :option:`per_job_logs` is false, then the filename will not include
2649 the job index. See `Log File Formats`_.
2651 .. option:: log_avg_msec=int
2653 By default, fio will log an entry in the iops, latency, or bw log for every
2654 I/O that completes. When writing to the disk log, that can quickly grow to a
2655 very large size. Setting this option makes fio average the each log entry
2656 over the specified period of time, reducing the resolution of the log. See
2657 :option:`log_max_value` as well. Defaults to 0, logging all entries.
2659 .. option:: log_hist_msec=int
2661 Same as :option:`log_avg_msec`, but logs entries for completion latency
2662 histograms. Computing latency percentiles from averages of intervals using
2663 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
2664 histogram entries over the specified period of time, reducing log sizes for
2665 high IOPS devices while retaining percentile accuracy. See
2666 :option:`log_hist_coarseness` as well. Defaults to 0, meaning histogram
2667 logging is disabled.
2669 .. option:: log_hist_coarseness=int
2671 Integer ranging from 0 to 6, defining the coarseness of the resolution of
2672 the histogram logs enabled with :option:`log_hist_msec`. For each increment
2673 in coarseness, fio outputs half as many bins. Defaults to 0, for which
2674 histogram logs contain 1216 latency bins. See `Log File Formats`_.
2676 .. option:: log_max_value=bool
2678 If :option:`log_avg_msec` is set, fio logs the average over that window. If
2679 you instead want to log the maximum value, set this option to 1. Defaults to
2680 0, meaning that averaged values are logged.
2682 .. option:: log_offset=int
2684 If this is set, the iolog options will include the byte offset for the I/O
2685 entry as well as the other data values.
2687 .. option:: log_compression=int
2689 If this is set, fio will compress the I/O logs as it goes, to keep the
2690 memory footprint lower. When a log reaches the specified size, that chunk is
2691 removed and compressed in the background. Given that I/O logs are fairly
2692 highly compressible, this yields a nice memory savings for longer runs. The
2693 downside is that the compression will consume some background CPU cycles, so
2694 it may impact the run. This, however, is also true if the logging ends up
2695 consuming most of the system memory. So pick your poison. The I/O logs are
2696 saved normally at the end of a run, by decompressing the chunks and storing
2697 them in the specified log file. This feature depends on the availability of
2700 .. option:: log_compression_cpus=str
2702 Define the set of CPUs that are allowed to handle online log compression for
2703 the I/O jobs. This can provide better isolation between performance
2704 sensitive jobs, and background compression work.
2706 .. option:: log_store_compressed=bool
2708 If set, fio will store the log files in a compressed format. They can be
2709 decompressed with fio, using the :option:`--inflate-log` command line
2710 parameter. The files will be stored with a :file:`.fz` suffix.
2712 .. option:: log_unix_epoch=bool
2714 If set, fio will log Unix timestamps to the log files produced by enabling
2715 write_type_log for each log type, instead of the default zero-based
2718 .. option:: block_error_percentiles=bool
2720 If set, record errors in trim block-sized units from writes and trims and
2721 output a histogram of how many trims it took to get to errors, and what kind
2722 of error was encountered.
2724 .. option:: bwavgtime=int
2726 Average the calculated bandwidth over the given time. Value is specified in
2727 milliseconds. If the job also does bandwidth logging through
2728 :option:`write_bw_log`, then the minimum of this option and
2729 :option:`log_avg_msec` will be used. Default: 500ms.
2731 .. option:: iopsavgtime=int
2733 Average the calculated IOPS over the given time. Value is specified in
2734 milliseconds. If the job also does IOPS logging through
2735 :option:`write_iops_log`, then the minimum of this option and
2736 :option:`log_avg_msec` will be used. Default: 500ms.
2738 .. option:: disk_util=bool
2740 Generate disk utilization statistics, if the platform supports it.
2743 .. option:: disable_lat=bool
2745 Disable measurements of total latency numbers. Useful only for cutting back
2746 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
2747 performance at really high IOPS rates. Note that to really get rid of a
2748 large amount of these calls, this option must be used with
2749 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
2751 .. option:: disable_clat=bool
2753 Disable measurements of completion latency numbers. See
2754 :option:`disable_lat`.
2756 .. option:: disable_slat=bool
2758 Disable measurements of submission latency numbers. See
2759 :option:`disable_slat`.
2761 .. option:: disable_bw_measurement=bool, disable_bw=bool
2763 Disable measurements of throughput/bandwidth numbers. See
2764 :option:`disable_lat`.
2766 .. option:: clat_percentiles=bool
2768 Enable the reporting of percentiles of completion latencies.
2770 .. option:: percentile_list=float_list
2772 Overwrite the default list of percentiles for completion latencies and the
2773 block error histogram. Each number is a floating number in the range
2774 (0,100], and the maximum length of the list is 20. Use ``:`` to separate the
2775 numbers, and list the numbers in ascending order. For example,
2776 ``--percentile_list=99.5:99.9`` will cause fio to report the values of
2777 completion latency below which 99.5% and 99.9% of the observed latencies
2784 .. option:: exitall_on_error
2786 When one job finishes in error, terminate the rest. The default is to wait
2787 for each job to finish.
2789 .. option:: continue_on_error=str
2791 Normally fio will exit the job on the first observed failure. If this option
2792 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
2793 EILSEQ) until the runtime is exceeded or the I/O size specified is
2794 completed. If this option is used, there are two more stats that are
2795 appended, the total error count and the first error. The error field given
2796 in the stats is the first error that was hit during the run.
2798 The allowed values are:
2801 Exit on any I/O or verify errors.
2804 Continue on read errors, exit on all others.
2807 Continue on write errors, exit on all others.
2810 Continue on any I/O error, exit on all others.
2813 Continue on verify errors, exit on all others.
2816 Continue on all errors.
2819 Backward-compatible alias for 'none'.
2822 Backward-compatible alias for 'all'.
2824 .. option:: ignore_error=str
2826 Sometimes you want to ignore some errors during test in that case you can
2827 specify error list for each error type.
2828 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
2829 given error type is separated with ':'. Error may be symbol ('ENOSPC',
2830 'ENOMEM') or integer. Example::
2832 ignore_error=EAGAIN,ENOSPC:122
2834 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
2837 .. option:: error_dump=bool
2839 If set dump every error even if it is non fatal, true by default. If
2840 disabled only fatal error will be dumped.
2842 Running predefined workloads
2843 ----------------------------
2845 Fio includes predefined profiles that mimic the I/O workloads generated by
2848 .. option:: profile=str
2850 The predefined workload to run. Current profiles are:
2853 Threaded I/O bench (tiotest/tiobench) like workload.
2856 Aerospike Certification Tool (ACT) like workload.
2858 To view a profile's additional options use :option:`--cmdhelp` after specifying
2859 the profile. For example::
2861 $ fio --profile=act --cmdhelp
2866 .. option:: device-names=str
2871 .. option:: load=int
2874 ACT load multiplier. Default: 1.
2876 .. option:: test-duration=time
2879 How long the entire test takes to run. Default: 24h.
2881 .. option:: threads-per-queue=int
2884 Number of read IO threads per device. Default: 8.
2886 .. option:: read-req-num-512-blocks=int
2889 Number of 512B blocks to read at the time. Default: 3.
2891 .. option:: large-block-op-kbytes=int
2894 Size of large block ops in KiB (writes). Default: 131072.
2899 Set to run ACT prep phase.
2901 Tiobench profile options
2902 ~~~~~~~~~~~~~~~~~~~~~~~~
2904 .. option:: size=str
2909 .. option:: block=int
2912 Block size in bytes. Default: 4096.
2914 .. option:: numruns=int
2924 .. option:: threads=int
2929 Interpreting the output
2930 -----------------------
2932 Fio spits out a lot of output. While running, fio will display the status of the
2933 jobs created. An example of that would be::
2935 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]
2937 The characters inside the square brackets denote the current status of each
2938 thread. The possible values (in typical life cycle order) are:
2940 +------+-----+-----------------------------------------------------------+
2942 +======+=====+===========================================================+
2943 | P | | Thread setup, but not started. |
2944 +------+-----+-----------------------------------------------------------+
2945 | C | | Thread created. |
2946 +------+-----+-----------------------------------------------------------+
2947 | I | | Thread initialized, waiting or generating necessary data. |
2948 +------+-----+-----------------------------------------------------------+
2949 | | p | Thread running pre-reading file(s). |
2950 +------+-----+-----------------------------------------------------------+
2951 | | R | Running, doing sequential reads. |
2952 +------+-----+-----------------------------------------------------------+
2953 | | r | Running, doing random reads. |
2954 +------+-----+-----------------------------------------------------------+
2955 | | W | Running, doing sequential writes. |
2956 +------+-----+-----------------------------------------------------------+
2957 | | w | Running, doing random writes. |
2958 +------+-----+-----------------------------------------------------------+
2959 | | M | Running, doing mixed sequential reads/writes. |
2960 +------+-----+-----------------------------------------------------------+
2961 | | m | Running, doing mixed random reads/writes. |
2962 +------+-----+-----------------------------------------------------------+
2963 | | F | Running, currently waiting for :manpage:`fsync(2)` |
2964 +------+-----+-----------------------------------------------------------+
2965 | | V | Running, doing verification of written data. |
2966 +------+-----+-----------------------------------------------------------+
2967 | E | | Thread exited, not reaped by main thread yet. |
2968 +------+-----+-----------------------------------------------------------+
2969 | _ | | Thread reaped, or |
2970 +------+-----+-----------------------------------------------------------+
2971 | X | | Thread reaped, exited with an error. |
2972 +------+-----+-----------------------------------------------------------+
2973 | K | | Thread reaped, exited due to signal. |
2974 +------+-----+-----------------------------------------------------------+
2976 Fio will condense the thread string as not to take up more space on the command
2977 line as is needed. For instance, if you have 10 readers and 10 writers running,
2978 the output would look like this::
2980 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]
2982 Fio will still maintain the ordering, though. So the above means that jobs 1..10
2983 are readers, and 11..20 are writers.
2985 The other values are fairly self explanatory -- number of threads currently
2986 running and doing I/O, the number of currently open files (f=), the rate of I/O
2987 since last check (read speed listed first, then write speed and optionally trim
2988 speed), and the estimated completion percentage and time for the current
2989 running group. It's impossible to estimate runtime of the following groups (if
2990 any). Note that the string is displayed in order, so it's possible to tell which
2991 of the jobs are currently doing what. The first character is the first job
2992 defined in the job file, and so forth.
2994 When fio is done (or interrupted by :kbd:`ctrl-c`), it will show the data for
2995 each thread, group of threads, and disks in that order. For each data direction,
2996 the output looks like::
2998 Client1 (g=0): err= 0:
2999 write: io= 32MiB, bw= 666KiB/s, iops=89 , runt= 50320msec
3000 slat (msec): min= 0, max= 136, avg= 0.03, stdev= 1.92
3001 clat (msec): min= 0, max= 631, avg=48.50, stdev=86.82
3002 bw (KiB/s) : min= 0, max= 1196, per=51.00%, avg=664.02, stdev=681.68
3003 cpu : usr=1.49%, sys=0.25%, ctx=7969, majf=0, minf=17
3004 IO depths : 1=0.1%, 2=0.3%, 4=0.5%, 8=99.0%, 16=0.0%, 32=0.0%, >32=0.0%
3005 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3006 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3007 issued r/w: total=0/32768, short=0/0
3008 lat (msec): 2=1.6%, 4=0.0%, 10=3.2%, 20=12.8%, 50=38.4%, 100=24.8%,
3009 lat (msec): 250=15.2%, 500=0.0%, 750=0.0%, 1000=0.0%, >=2048=0.0%
3011 The client number is printed, along with the group id and error of that
3012 thread. Below is the I/O statistics, here for writes. In the order listed, they
3016 Number of megabytes I/O performed.
3019 Average bandwidth rate.
3022 Average I/Os performed per second.
3025 The runtime of that thread.
3028 Submission latency (avg being the average, stdev being the standard
3029 deviation). This is the time it took to submit the I/O. For sync I/O,
3030 the slat is really the completion latency, since queue/complete is one
3031 operation there. This value can be in milliseconds or microseconds, fio
3032 will choose the most appropriate base and print that. In the example
3033 above, milliseconds is the best scale. Note: in :option:`--minimal` mode
3034 latencies are always expressed in microseconds.
3037 Completion latency. Same names as slat, this denotes the time from
3038 submission to completion of the I/O pieces. For sync I/O, clat will
3039 usually be equal (or very close) to 0, as the time from submit to
3040 complete is basically just CPU time (I/O has already been done, see slat
3044 Bandwidth. Same names as the xlat stats, but also includes an
3045 approximate percentage of total aggregate bandwidth this thread received
3046 in this group. This last value is only really useful if the threads in
3047 this group are on the same disk, since they are then competing for disk
3051 CPU usage. User and system time, along with the number of context
3052 switches this thread went through, usage of system and user time, and
3053 finally the number of major and minor page faults. The CPU utilization
3054 numbers are averages for the jobs in that reporting group, while the
3055 context and fault counters are summed.
3058 The distribution of I/O depths over the job life time. The numbers are
3059 divided into powers of 2, so for example the 16= entries includes depths
3060 up to that value but higher than the previous entry. In other words, it
3061 covers the range from 16 to 31.
3064 How many pieces of I/O were submitting in a single submit call. Each
3065 entry denotes that amount and below, until the previous entry -- e.g.,
3066 8=100% mean that we submitted anywhere in between 5-8 I/Os per submit
3070 Like the above submit number, but for completions instead.
3073 The number of read/write requests issued, and how many of them were
3077 The distribution of I/O completion latencies. This is the time from when
3078 I/O leaves fio and when it gets completed. The numbers follow the same
3079 pattern as the I/O depths, meaning that 2=1.6% means that 1.6% of the
3080 I/O completed within 2 msecs, 20=12.8% means that 12.8% of the I/O took
3081 more than 10 msecs, but less than (or equal to) 20 msecs.
3083 After each client has been listed, the group statistics are printed. They
3084 will look like this::
3086 Run status group 0 (all jobs):
3087 READ: io=64MB, aggrb=22178, minb=11355, maxb=11814, mint=2840msec, maxt=2955msec
3088 WRITE: io=64MB, aggrb=1302, minb=666, maxb=669, mint=50093msec, maxt=50320msec
3090 For each data direction, it prints:
3093 Number of megabytes I/O performed.
3095 Aggregate bandwidth of threads in this group.
3097 The minimum average bandwidth a thread saw.
3099 The maximum average bandwidth a thread saw.
3101 The smallest runtime of the threads in that group.
3103 The longest runtime of the threads in that group.
3105 And finally, the disk statistics are printed. They will look like this::
3107 Disk stats (read/write):
3108 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
3110 Each value is printed for both reads and writes, with reads first. The
3114 Number of I/Os performed by all groups.
3116 Number of merges I/O the I/O scheduler.
3118 Number of ticks we kept the disk busy.
3120 Total time spent in the disk queue.
3122 The disk utilization. A value of 100% means we kept the disk
3123 busy constantly, 50% would be a disk idling half of the time.
3125 It is also possible to get fio to dump the current output while it is running,
3126 without terminating the job. To do that, send fio the **USR1** signal. You can
3127 also get regularly timed dumps by using the :option:`--status-interval`
3128 parameter, or by creating a file in :file:`/tmp` named
3129 :file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
3130 current output status.
3136 For scripted usage where you typically want to generate tables or graphs of the
3137 results, fio can output the results in a semicolon separated format. The format
3138 is one long line of values, such as::
3140 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%
3141 A description of this job goes here.
3143 The job description (if provided) follows on a second line.
3145 To enable terse output, use the :option:`--minimal` command line option. The
3146 first value is the version of the terse output format. If the output has to be
3147 changed for some reason, this number will be incremented by 1 to signify that
3150 Split up, the format is as follows:
3154 terse version, fio version, jobname, groupid, error
3158 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3159 Submission latency: min, max, mean, stdev (usec)
3160 Completion latency: min, max, mean, stdev (usec)
3161 Completion latency percentiles: 20 fields (see below)
3162 Total latency: min, max, mean, stdev (usec)
3163 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev
3169 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3170 Submission latency: min, max, mean, stdev (usec)
3171 Completion latency: min, max, mean, stdev(usec)
3172 Completion latency percentiles: 20 fields (see below)
3173 Total latency: min, max, mean, stdev (usec)
3174 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev
3178 user, system, context switches, major faults, minor faults
3182 <=1, 2, 4, 8, 16, 32, >=64
3184 I/O latencies microseconds::
3186 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
3188 I/O latencies milliseconds::
3190 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
3194 Disk name, Read ios, write ios,
3195 Read merges, write merges,
3196 Read ticks, write ticks,
3197 Time spent in queue, disk utilization percentage
3199 Additional Info (dependent on continue_on_error, default off)::
3201 total # errors, first error code
3203 Additional Info (dependent on description being set)::
3207 Completion latency percentiles can be a grouping of up to 20 sets, so for the
3208 terse output fio writes all of them. Each field will look like this::
3212 which is the Xth percentile, and the `usec` latency associated with it.
3214 For disk utilization, all disks used by fio are shown. So for each disk there
3215 will be a disk utilization section.
3221 There are two trace file format that you can encounter. The older (v1) format is
3222 unsupported since version 1.20-rc3 (March 2008). It will still be described
3223 below in case that you get an old trace and want to understand it.
3225 In any case the trace is a simple text file with a single action per line.
3228 Trace file format v1
3229 ~~~~~~~~~~~~~~~~~~~~
3231 Each line represents a single I/O action in the following format::
3235 where `rw=0/1` for read/write, and the offset and length entries being in bytes.
3237 This format is not supported in fio versions => 1.20-rc3.
3240 Trace file format v2
3241 ~~~~~~~~~~~~~~~~~~~~
3243 The second version of the trace file format was added in fio version 1.17. It
3244 allows to access more then one file per trace and has a bigger set of possible
3247 The first line of the trace file has to be::
3251 Following this can be lines in two different formats, which are described below.
3253 The file management format::
3257 The filename is given as an absolute path. The action can be one of these:
3260 Add the given filename to the trace.
3262 Open the file with the given filename. The filename has to have
3263 been added with the **add** action before.
3265 Close the file with the given filename. The file has to have been
3269 The file I/O action format::
3271 filename action offset length
3273 The `filename` is given as an absolute path, and has to have been added and
3274 opened before it can be used with this format. The `offset` and `length` are
3275 given in bytes. The `action` can be one of these:
3278 Wait for `offset` microseconds. Everything below 100 is discarded.
3279 The time is relative to the previous `wait` statement.
3281 Read `length` bytes beginning from `offset`.
3283 Write `length` bytes beginning from `offset`.
3285 :manpage:`fsync(2)` the file.
3287 :manpage:`fdatasync(2)` the file.
3289 Trim the given file from the given `offset` for `length` bytes.
3291 CPU idleness profiling
3292 ----------------------
3294 In some cases, we want to understand CPU overhead in a test. For example, we
3295 test patches for the specific goodness of whether they reduce CPU usage.
3296 Fio implements a balloon approach to create a thread per CPU that runs at idle
3297 priority, meaning that it only runs when nobody else needs the cpu.
3298 By measuring the amount of work completed by the thread, idleness of each CPU
3299 can be derived accordingly.
3301 An unit work is defined as touching a full page of unsigned characters. Mean and
3302 standard deviation of time to complete an unit work is reported in "unit work"
3303 section. Options can be chosen to report detailed percpu idleness or overall
3304 system idleness by aggregating percpu stats.
3307 Verification and triggers
3308 -------------------------
3310 Fio is usually run in one of two ways, when data verification is done. The first
3311 is a normal write job of some sort with verify enabled. When the write phase has
3312 completed, fio switches to reads and verifies everything it wrote. The second
3313 model is running just the write phase, and then later on running the same job
3314 (but with reads instead of writes) to repeat the same I/O patterns and verify
3315 the contents. Both of these methods depend on the write phase being completed,
3316 as fio otherwise has no idea how much data was written.
3318 With verification triggers, fio supports dumping the current write state to
3319 local files. Then a subsequent read verify workload can load this state and know
3320 exactly where to stop. This is useful for testing cases where power is cut to a
3321 server in a managed fashion, for instance.
3323 A verification trigger consists of two things:
3325 1) Storing the write state of each job.
3326 2) Executing a trigger command.
3328 The write state is relatively small, on the order of hundreds of bytes to single
3329 kilobytes. It contains information on the number of completions done, the last X
3332 A trigger is invoked either through creation ('touch') of a specified file in
3333 the system, or through a timeout setting. If fio is run with
3334 :option:`--trigger-file` = :file:`/tmp/trigger-file`, then it will continually
3335 check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
3336 will fire off the trigger (thus saving state, and executing the trigger
3339 For client/server runs, there's both a local and remote trigger. If fio is
3340 running as a server backend, it will send the job states back to the client for
3341 safe storage, then execute the remote trigger, if specified. If a local trigger
3342 is specified, the server will still send back the write state, but the client
3343 will then execute the trigger.
3345 Verification trigger example
3346 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3348 Lets say we want to run a powercut test on the remote machine 'server'. Our
3349 write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
3350 some point during the run, and we'll run this test from the safety or our local
3351 machine, 'localbox'. On the server, we'll start the fio backend normally::
3353 server# fio --server
3355 and on the client, we'll fire off the workload::
3357 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
3359 We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
3361 echo b > /proc/sysrq-trigger
3363 on the server once it has received the trigger and sent us the write state. This
3364 will work, but it's not **really** cutting power to the server, it's merely
3365 abruptly rebooting it. If we have a remote way of cutting power to the server
3366 through IPMI or similar, we could do that through a local trigger command
3367 instead. Lets assume we have a script that does IPMI reboot of a given hostname,
3368 ipmi-reboot. On localbox, we could then have run fio with a local trigger
3371 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
3373 For this case, fio would wait for the server to send us the write state, then
3374 execute ``ipmi-reboot server`` when that happened.
3376 Loading verify state
3377 ~~~~~~~~~~~~~~~~~~~~
3379 To load store write state, read verification job file must contain the
3380 :option:`verify_state_load` option. If that is set, fio will load the previously
3381 stored state. For a local fio run this is done by loading the files directly,
3382 and on a client/server run, the server backend will ask the client to send the
3383 files over and load them from there.
3389 Fio supports a variety of log file formats, for logging latencies, bandwidth,
3390 and IOPS. The logs share a common format, which looks like this:
3392 *time* (`msec`), *value*, *data direction*, *offset*
3394 Time for the log entry is always in milliseconds. The *value* logged depends
3395 on the type of log, it will be one of the following:
3398 Value is latency in usecs
3404 *Data direction* is one of the following:
3413 The *offset* is the offset, in bytes, from the start of the file, for that
3414 particular I/O. The logging of the offset can be toggled with
3415 :option:`log_offset`.
3417 If windowed logging is enabled through :option:`log_avg_msec` then fio doesn't
3418 log individual I/Os. Instead of logs the average values over the specified period
3419 of time. Since 'data direction' and 'offset' are per-I/O values, they aren't
3420 applicable if windowed logging is enabled. If windowed logging is enabled and
3421 :option:`log_max_value` is set, then fio logs maximum values in that window
3422 instead of averages.
3428 Normally fio is invoked as a stand-alone application on the machine where the
3429 I/O workload should be generated. However, the frontend and backend of fio can
3430 be run separately. Ie the fio server can generate an I/O workload on the "Device
3431 Under Test" while being controlled from another machine.
3433 Start the server on the machine which has access to the storage DUT::
3437 where args defines what fio listens to. The arguments are of the form
3438 ``type,hostname`` or ``IP,port``. *type* is either ``ip`` (or ip4) for TCP/IP
3439 v4, ``ip6`` for TCP/IP v6, or ``sock`` for a local unix domain socket.
3440 *hostname* is either a hostname or IP address, and *port* is the port to listen
3441 to (only valid for TCP/IP, not a local socket). Some examples:
3445 Start a fio server, listening on all interfaces on the default port (8765).
3447 2) ``fio --server=ip:hostname,4444``
3449 Start a fio server, listening on IP belonging to hostname and on port 4444.
3451 3) ``fio --server=ip6:::1,4444``
3453 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
3455 4) ``fio --server=,4444``
3457 Start a fio server, listening on all interfaces on port 4444.
3459 5) ``fio --server=1.2.3.4``
3461 Start a fio server, listening on IP 1.2.3.4 on the default port.
3463 6) ``fio --server=sock:/tmp/fio.sock``
3465 Start a fio server, listening on the local socket /tmp/fio.sock.
3467 Once a server is running, a "client" can connect to the fio server with::
3469 fio <local-args> --client=<server> <remote-args> <job file(s)>
3471 where `local-args` are arguments for the client where it is running, `server`
3472 is the connect string, and `remote-args` and `job file(s)` are sent to the
3473 server. The `server` string follows the same format as it does on the server
3474 side, to allow IP/hostname/socket and port strings.
3476 Fio can connect to multiple servers this way::
3478 fio --client=<server1> <job file(s)> --client=<server2> <job file(s)>
3480 If the job file is located on the fio server, then you can tell the server to
3481 load a local file as well. This is done by using :option:`--remote-config` ::
3483 fio --client=server --remote-config /path/to/file.fio
3485 Then fio will open this local (to the server) job file instead of being passed
3486 one from the client.
3488 If you have many servers (example: 100 VMs/containers), you can input a pathname
3489 of a file containing host IPs/names as the parameter value for the
3490 :option:`--client` option. For example, here is an example :file:`host.list`
3491 file containing 2 hostnames::
3493 host1.your.dns.domain
3494 host2.your.dns.domain
3496 The fio command would then be::
3498 fio --client=host.list <job file(s)>
3500 In this mode, you cannot input server-specific parameters or job files -- all
3501 servers receive the same job file.
3503 In order to let ``fio --client`` runs use a shared filesystem from multiple
3504 hosts, ``fio --client`` now prepends the IP address of the server to the
3505 filename. For example, if fio is using directory :file:`/mnt/nfs/fio` and is
3506 writing filename :file:`fileio.tmp`, with a :option:`--client` `hostfile`
3507 containing two hostnames ``h1`` and ``h2`` with IP addresses 192.168.10.120 and
3508 192.168.10.121, then fio will create two files::
3510 /mnt/nfs/fio/192.168.10.120.fileio.tmp
3511 /mnt/nfs/fio/192.168.10.121.fileio.tmp