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 or 5).
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>`.
546 If the lower value specified happens to be larger than the upper value,
547 two values are swapped.
552 Boolean. Usually parsed as an integer, however only defined for
553 true and false (1 and 0).
558 Integer range with suffix. Allows value range to be given, such as
559 1024-4096. A colon may also be used as the separator, e.g. 1k:4k. If the
560 option allows two sets of ranges, they can be specified with a ',' or '/'
561 delimiter: 1k-4k/8k-32k. Also see :ref:`int <int>`.
564 A list of floating point numbers, separated by a ':' character.
570 .. option:: kb_base=int
572 Select the interpretation of unit prefixes in input parameters.
575 Inputs comply with IEC 80000-13 and the International
576 System of Units (SI). Use:
578 - power-of-2 values with IEC prefixes (e.g., KiB)
579 - power-of-10 values with SI prefixes (e.g., kB)
582 Compatibility mode (default). To avoid breaking old scripts:
584 - power-of-2 values with SI prefixes
585 - power-of-10 values with IEC prefixes
587 See :option:`bs` for more details on input parameters.
589 Outputs always use correct prefixes. Most outputs include both
592 bw=2383.3kB/s (2327.4KiB/s)
594 If only one value is reported, then kb_base selects the one to use:
596 **1000** -- SI prefixes
598 **1024** -- IEC prefixes
600 .. option:: unit_base=int
602 Base unit for reporting. Allowed values are:
605 Use auto-detection (default).
612 With the above in mind, here follows the complete list of fio job parameters.
620 ASCII name of the job. This may be used to override the name printed by fio
621 for this job. Otherwise the job name is used. On the command line this
622 parameter has the special purpose of also signaling the start of a new job.
624 .. option:: description=str
626 Text description of the job. Doesn't do anything except dump this text
627 description when this job is run. It's not parsed.
629 .. option:: loops=int
631 Run the specified number of iterations of this job. Used to repeat the same
632 workload a given number of times. Defaults to 1.
634 .. option:: numjobs=int
636 Create the specified number of clones of this job. Each clone of job
637 is spawned as an independent thread or process. May be used to setup a
638 larger number of threads/processes doing the same thing. Each thread is
639 reported separately; to see statistics for all clones as a whole, use
640 :option:`group_reporting` in conjunction with :option:`new_group`.
641 See :option:`--max-jobs`. Default: 1.
644 Time related parameters
645 ~~~~~~~~~~~~~~~~~~~~~~~
647 .. option:: runtime=time
649 Tell fio to terminate processing after the specified period of time. It
650 can be quite hard to determine for how long a specified job will run, so
651 this parameter is handy to cap the total runtime to a given time. When
652 the unit is omitted, the value is given in seconds.
654 .. option:: time_based
656 If set, fio will run for the duration of the :option:`runtime` specified
657 even if the file(s) are completely read or written. It will simply loop over
658 the same workload as many times as the :option:`runtime` allows.
660 .. option:: startdelay=irange(time)
662 Delay start of job for the specified number of seconds. Supports all time
663 suffixes to allow specification of hours, minutes, seconds and milliseconds
664 -- seconds are the default if a unit is omitted. Can be given as a range
665 which causes each thread to choose randomly out of the range.
667 .. option:: ramp_time=time
669 If set, fio will run the specified workload for this amount of time before
670 logging any performance numbers. Useful for letting performance settle
671 before logging results, thus minimizing the runtime required for stable
672 results. Note that the ``ramp_time`` is considered lead in time for a job,
673 thus it will increase the total runtime if a special timeout or
674 :option:`runtime` is specified. When the unit is omitted, the value is
677 .. option:: clocksource=str
679 Use the given clocksource as the base of timing. The supported options are:
682 :manpage:`gettimeofday(2)`
685 :manpage:`clock_gettime(2)`
688 Internal CPU clock source
690 cpu is the preferred clocksource if it is reliable, as it is very fast (and
691 fio is heavy on time calls). Fio will automatically use this clocksource if
692 it's supported and considered reliable on the system it is running on,
693 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
694 means supporting TSC Invariant.
696 .. option:: gtod_reduce=bool
698 Enable all of the :manpage:`gettimeofday(2)` reducing options
699 (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus
700 reduce precision of the timeout somewhat to really shrink the
701 :manpage:`gettimeofday(2)` call count. With this option enabled, we only do
702 about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all
703 time keeping was enabled.
705 .. option:: gtod_cpu=int
707 Sometimes it's cheaper to dedicate a single thread of execution to just
708 getting the current time. Fio (and databases, for instance) are very
709 intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set
710 one CPU aside for doing nothing but logging current time to a shared memory
711 location. Then the other threads/processes that run I/O workloads need only
712 copy that segment, instead of entering the kernel with a
713 :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time
714 calls will be excluded from other uses. Fio will manually clear it from the
715 CPU mask of other jobs.
721 .. option:: directory=str
723 Prefix filenames with this directory. Used to place files in a different
724 location than :file:`./`. You can specify a number of directories by
725 separating the names with a ':' character. These directories will be
726 assigned equally distributed to job clones creates with :option:`numjobs` as
727 long as they are using generated filenames. If specific `filename(s)` are
728 set fio will use the first listed directory, and thereby matching the
729 `filename` semantic which generates a file each clone if not specified, but
730 let all clones use the same if set.
732 See the :option:`filename` option for escaping certain characters.
734 .. option:: filename=str
736 Fio normally makes up a `filename` based on the job name, thread number, and
737 file number. If you want to share files between threads in a job or several
738 jobs with fixed file paths, specify a `filename` for each of them to override
739 the default. If the ioengine is file based, you can specify a number of files
740 by separating the names with a ':' colon. So if you wanted a job to open
741 :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use
742 ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is
743 specified, :option:`nrfiles` is ignored. The size of regular files specified
744 by this option will be :option:`size` divided by number of files unless
745 explicit size is specified by :option:`filesize`.
747 On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for
748 the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc.
749 Note: Windows and FreeBSD prevent write access to areas
750 of the disk containing in-use data (e.g. filesystems). If the wanted
751 `filename` does need to include a colon, then escape that with a ``\``
752 character. For instance, if the `filename` is :file:`/dev/dsk/foo@3,0:c`,
753 then you would use ``filename="/dev/dsk/foo@3,0\:c"``. The
754 :file:`-` is a reserved name, meaning stdin or stdout. Which of the two
755 depends on the read/write direction set.
757 .. option:: filename_format=str
759 If sharing multiple files between jobs, it is usually necessary to have fio
760 generate the exact names that you want. By default, fio will name a file
761 based on the default file format specification of
762 :file:`jobname.jobnumber.filenumber`. With this option, that can be
763 customized. Fio will recognize and replace the following keywords in this
767 The name of the worker thread or process.
769 The incremental number of the worker thread or process.
771 The incremental number of the file for that worker thread or
774 To have dependent jobs share a set of files, this option can be set to have
775 fio generate filenames that are shared between the two. For instance, if
776 :file:`testfiles.$filenum` is specified, file number 4 for any job will be
777 named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum`
778 will be used if no other format specifier is given.
780 .. option:: unique_filename=bool
782 To avoid collisions between networked clients, fio defaults to prefixing any
783 generated filenames (with a directory specified) with the source of the
784 client connecting. To disable this behavior, set this option to 0.
786 .. option:: opendir=str
788 Recursively open any files below directory `str`.
790 .. option:: lockfile=str
792 Fio defaults to not locking any files before it does I/O to them. If a file
793 or file descriptor is shared, fio can serialize I/O to that file to make the
794 end result consistent. This is usual for emulating real workloads that share
795 files. The lock modes are:
798 No locking. The default.
800 Only one thread or process may do I/O at a time, excluding all
803 Read-write locking on the file. Many readers may
804 access the file at the same time, but writes get exclusive access.
806 .. option:: nrfiles=int
808 Number of files to use for this job. Defaults to 1. The size of files
809 will be :option:`size` divided by this unless explicit size is specified by
810 :option:`filesize`. Files are created for each thread separately, and each
811 file will have a file number within its name by default, as explained in
812 :option:`filename` section.
815 .. option:: openfiles=int
817 Number of files to keep open at the same time. Defaults to the same as
818 :option:`nrfiles`, can be set smaller to limit the number simultaneous
821 .. option:: file_service_type=str
823 Defines how fio decides which file from a job to service next. The following
827 Choose a file at random.
830 Round robin over opened files. This is the default.
833 Finish one file before moving on to the next. Multiple files can
834 still be open depending on 'openfiles'.
837 Use a *Zipf* distribution to decide what file to access.
840 Use a *Pareto* distribution to decide what file to access.
843 Use a *Gaussian* (normal) distribution to decide what file to
846 For *random*, *roundrobin*, and *sequential*, a postfix can be appended to
847 tell fio how many I/Os to issue before switching to a new file. For example,
848 specifying ``file_service_type=random:8`` would cause fio to issue
849 8 I/Os before selecting a new file at random. For the non-uniform
850 distributions, a floating point postfix can be given to influence how the
851 distribution is skewed. See :option:`random_distribution` for a description
852 of how that would work.
854 .. option:: ioscheduler=str
856 Attempt to switch the device hosting the file to the specified I/O scheduler
859 .. option:: create_serialize=bool
861 If true, serialize the file creation for the jobs. This may be handy to
862 avoid interleaving of data files, which may greatly depend on the filesystem
863 used and even the number of processors in the system. Default: true.
865 .. option:: create_fsync=bool
867 fsync the data file after creation. This is the default.
869 .. option:: create_on_open=bool
871 Don't pre-setup the files for I/O, just create open() when it's time to do
872 I/O to that file. Default: false.
874 .. option:: create_only=bool
876 If true, fio will only run the setup phase of the job. If files need to be
877 laid out or updated on disk, only that will be done. The actual job contents
878 are not executed. Default: false.
880 .. option:: allow_file_create=bool
882 If true, fio is permitted to create files as part of its workload. This is
883 the default behavior. If this option is false, then fio will error out if
884 the files it needs to use don't already exist. Default: true.
886 .. option:: allow_mounted_write=bool
888 If this isn't set, fio will abort jobs that are destructive (e.g. that write)
889 to what appears to be a mounted device or partition. This should help catch
890 creating inadvertently destructive tests, not realizing that the test will
891 destroy data on the mounted file system. Note that some platforms don't allow
892 writing against a mounted device regardless of this option. Default: false.
894 .. option:: pre_read=bool
896 If this is given, files will be pre-read into memory before starting the
897 given I/O operation. This will also clear the :option:`invalidate` flag,
898 since it is pointless to pre-read and then drop the cache. This will only
899 work for I/O engines that are seek-able, since they allow you to read the
900 same data multiple times. Thus it will not work on non-seekable I/O engines
901 (e.g. network, splice). Default: false.
903 .. option:: unlink=bool
905 Unlink the job files when done. Not the default, as repeated runs of that
906 job would then waste time recreating the file set again and again. Default:
909 .. option:: unlink_each_loop=bool
911 Unlink job files after each iteration or loop. Default: false.
913 .. option:: zonesize=int
915 Divide a file into zones of the specified size. See :option:`zoneskip`.
917 .. option:: zonerange=int
919 Give size of an I/O zone. See :option:`zoneskip`.
921 .. option:: zoneskip=int
923 Skip the specified number of bytes when :option:`zonesize` data has been
924 read. The two zone options can be used to only do I/O on zones of a file.
930 .. option:: direct=bool
932 If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that
933 ZFS on Solaris doesn't support direct I/O. On Windows the synchronous
934 ioengines don't support direct I/O. Default: false.
936 .. option:: atomic=bool
938 If value is true, attempt to use atomic direct I/O. Atomic writes are
939 guaranteed to be stable once acknowledged by the operating system. Only
940 Linux supports O_ATOMIC right now.
942 .. option:: buffered=bool
944 If value is true, use buffered I/O. This is the opposite of the
945 :option:`direct` option. Defaults to true.
947 .. option:: readwrite=str, rw=str
949 Type of I/O pattern. Accepted values are:
956 Sequential trims (Linux block devices only).
962 Random trims (Linux block devices only).
964 Sequential mixed reads and writes.
966 Random mixed reads and writes.
968 Sequential trim+write sequences. Blocks will be trimmed first,
969 then the same blocks will be written to.
971 Fio defaults to read if the option is not specified. For the mixed I/O
972 types, the default is to split them 50/50. For certain types of I/O the
973 result may still be skewed a bit, since the speed may be different. It is
974 possible to specify a number of I/O's to do before getting a new offset,
975 this is done by appending a ``:<nr>`` to the end of the string given. For a
976 random read, it would look like ``rw=randread:8`` for passing in an offset
977 modifier with a value of 8. If the suffix is used with a sequential I/O
978 pattern, then the value specified will be added to the generated offset for
979 each I/O. For instance, using ``rw=write:4k`` will skip 4k for every
980 write. It turns sequential I/O into sequential I/O with holes. See the
981 :option:`rw_sequencer` option.
983 .. option:: rw_sequencer=str
985 If an offset modifier is given by appending a number to the ``rw=<str>``
986 line, then this option controls how that number modifies the I/O offset
987 being generated. Accepted values are:
990 Generate sequential offset.
992 Generate the same offset.
994 ``sequential`` is only useful for random I/O, where fio would normally
995 generate a new random offset for every I/O. If you append e.g. 8 to randread,
996 you would get a new random offset for every 8 I/O's. The result would be a
997 seek for only every 8 I/O's, instead of for every I/O. Use ``rw=randread:8``
998 to specify that. As sequential I/O is already sequential, setting
999 ``sequential`` for that would not result in any differences. ``identical``
1000 behaves in a similar fashion, except it sends the same offset 8 number of
1001 times before generating a new offset.
1003 .. option:: unified_rw_reporting=bool
1005 Fio normally reports statistics on a per data direction basis, meaning that
1006 reads, writes, and trims are accounted and reported separately. If this
1007 option is set fio sums the results and report them as "mixed" instead.
1009 .. option:: randrepeat=bool
1011 Seed the random number generator used for random I/O patterns in a
1012 predictable way so the pattern is repeatable across runs. Default: true.
1014 .. option:: allrandrepeat=bool
1016 Seed all random number generators in a predictable way so results are
1017 repeatable across runs. Default: false.
1019 .. option:: randseed=int
1021 Seed the random number generators based on this seed value, to be able to
1022 control what sequence of output is being generated. If not set, the random
1023 sequence depends on the :option:`randrepeat` setting.
1025 .. option:: fallocate=str
1027 Whether pre-allocation is performed when laying down files.
1028 Accepted values are:
1031 Do not pre-allocate space.
1034 Pre-allocate via :manpage:`posix_fallocate(3)`.
1037 Pre-allocate via :manpage:`fallocate(2)` with
1038 FALLOC_FL_KEEP_SIZE set.
1041 Backward-compatible alias for **none**.
1044 Backward-compatible alias for **posix**.
1046 May not be available on all supported platforms. **keep** is only available
1047 on Linux. If using ZFS on Solaris this must be set to **none** because ZFS
1048 doesn't support it. Default: **posix**.
1050 .. option:: fadvise_hint=str
1052 Use :manpage:`posix_fadvise(2)` to advise the kernel on what I/O patterns
1053 are likely to be issued. Accepted values are:
1056 Backwards-compatible hint for "no hint".
1059 Backwards compatible hint for "advise with fio workload type". This
1060 uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL**
1061 for a sequential workload.
1064 Advise using **FADV_SEQUENTIAL**.
1067 Advise using **FADV_RANDOM**.
1069 .. option:: fadvise_stream=int
1071 Use :manpage:`posix_fadvise(2)` to advise the kernel what stream ID the
1072 writes issued belong to. Only supported on Linux. Note, this option may
1073 change going forward.
1075 .. option:: offset=int
1077 Start I/O at the provided offset in the file, given as either a fixed size or
1078 a percentage. If a percentage is given, the next ``blockalign``-ed offset
1079 will be used. Data before the given offset will not be touched. This
1080 effectively caps the file size at `real_size - offset`. Can be combined with
1081 :option:`size` to constrain the start and end range of the I/O workload.
1083 .. option:: offset_increment=int
1085 If this is provided, then the real offset becomes `offset + offset_increment
1086 * thread_number`, where the thread number is a counter that starts at 0 and
1087 is incremented for each sub-job (i.e. when :option:`numjobs` option is
1088 specified). This option is useful if there are several jobs which are
1089 intended to operate on a file in parallel disjoint segments, with even
1090 spacing between the starting points.
1092 .. option:: number_ios=int
1094 Fio will normally perform I/Os until it has exhausted the size of the region
1095 set by :option:`size`, or if it exhaust the allocated time (or hits an error
1096 condition). With this setting, the range/size can be set independently of
1097 the number of I/Os to perform. When fio reaches this number, it will exit
1098 normally and report status. Note that this does not extend the amount of I/O
1099 that will be done, it will only stop fio if this condition is met before
1100 other end-of-job criteria.
1102 .. option:: fsync=int
1104 If writing to a file, issue a sync of the dirty data for every number of
1105 blocks given. For example, if you give 32 as a parameter, fio will sync the
1106 file for every 32 writes issued. If fio is using non-buffered I/O, we may
1107 not sync the file. The exception is the sg I/O engine, which synchronizes
1108 the disk cache anyway. Defaults to 0, which means no sync every certain
1111 .. option:: fdatasync=int
1113 Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and
1114 not metadata blocks. In Windows, FreeBSD, and DragonFlyBSD there is no
1115 :manpage:`fdatasync(2)`, this falls back to using :manpage:`fsync(2)`.
1116 Defaults to 0, which means no sync data every certain number of writes.
1118 .. option:: write_barrier=int
1120 Make every `N-th` write a barrier write.
1122 .. option:: sync_file_range=str:val
1124 Use :manpage:`sync_file_range(2)` for every `val` number of write
1125 operations. Fio will track range of writes that have happened since the last
1126 :manpage:`sync_file_range(2)` call. `str` can currently be one or more of:
1129 SYNC_FILE_RANGE_WAIT_BEFORE
1131 SYNC_FILE_RANGE_WRITE
1133 SYNC_FILE_RANGE_WAIT_AFTER
1135 So if you do ``sync_file_range=wait_before,write:8``, fio would use
1136 ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8
1137 writes. Also see the :manpage:`sync_file_range(2)` man page. This option is
1140 .. option:: overwrite=bool
1142 If true, writes to a file will always overwrite existing data. If the file
1143 doesn't already exist, it will be created before the write phase begins. If
1144 the file exists and is large enough for the specified write phase, nothing
1145 will be done. Default: false.
1147 .. option:: end_fsync=bool
1149 If true, :manpage:`fsync(2)` file contents when a write stage has completed.
1152 .. option:: fsync_on_close=bool
1154 If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs
1155 from :option:`end_fsync` in that it will happen on every file close, not
1156 just at the end of the job. Default: false.
1158 .. option:: rwmixread=int
1160 Percentage of a mixed workload that should be reads. Default: 50.
1162 .. option:: rwmixwrite=int
1164 Percentage of a mixed workload that should be writes. If both
1165 :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not
1166 add up to 100%, the latter of the two will be used to override the
1167 first. This may interfere with a given rate setting, if fio is asked to
1168 limit reads or writes to a certain rate. If that is the case, then the
1169 distribution may be skewed. Default: 50.
1171 .. option:: random_distribution=str:float[,str:float][,str:float]
1173 By default, fio will use a completely uniform random distribution when asked
1174 to perform random I/O. Sometimes it is useful to skew the distribution in
1175 specific ways, ensuring that some parts of the data is more hot than others.
1176 fio includes the following distribution models:
1179 Uniform random distribution
1188 Normal (Gaussian) distribution
1191 Zoned random distribution
1193 When using a **zipf** or **pareto** distribution, an input value is also
1194 needed to define the access pattern. For **zipf**, this is the `zipf
1195 theta`. For **pareto**, it's the `Pareto power`. Fio includes a test
1196 program, :command:`genzipf`, that can be used visualize what the given input
1197 values will yield in terms of hit rates. If you wanted to use **zipf** with
1198 a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the
1199 option. If a non-uniform model is used, fio will disable use of the random
1200 map. For the **gauss** distribution, a normal deviation is supplied as a
1201 value between 0 and 100.
1203 For a **zoned** distribution, fio supports specifying percentages of I/O
1204 access that should fall within what range of the file or device. For
1205 example, given a criteria of:
1207 * 60% of accesses should be to the first 10%
1208 * 30% of accesses should be to the next 20%
1209 * 8% of accesses should be to to the next 30%
1210 * 2% of accesses should be to the next 40%
1212 we can define that through zoning of the random accesses. For the above
1213 example, the user would do::
1215 random_distribution=zoned:60/10:30/20:8/30:2/40
1217 similarly to how :option:`bssplit` works for setting ranges and percentages
1218 of block sizes. Like :option:`bssplit`, it's possible to specify separate
1219 zones for reads, writes, and trims. If just one set is given, it'll apply to
1222 .. option:: percentage_random=int[,int][,int]
1224 For a random workload, set how big a percentage should be random. This
1225 defaults to 100%, in which case the workload is fully random. It can be set
1226 from anywhere from 0 to 100. Setting it to 0 would make the workload fully
1227 sequential. Any setting in between will result in a random mix of sequential
1228 and random I/O, at the given percentages. Comma-separated values may be
1229 specified for reads, writes, and trims as described in :option:`blocksize`.
1231 .. option:: norandommap
1233 Normally fio will cover every block of the file when doing random I/O. If
1234 this option is given, fio will just get a new random offset without looking
1235 at past I/O history. This means that some blocks may not be read or written,
1236 and that some blocks may be read/written more than once. If this option is
1237 used with :option:`verify` and multiple blocksizes (via :option:`bsrange`),
1238 only intact blocks are verified, i.e., partially-overwritten blocks are
1241 .. option:: softrandommap=bool
1243 See :option:`norandommap`. If fio runs with the random block map enabled and
1244 it fails to allocate the map, if this option is set it will continue without
1245 a random block map. As coverage will not be as complete as with random maps,
1246 this option is disabled by default.
1248 .. option:: random_generator=str
1250 Fio supports the following engines for generating
1251 I/O offsets for random I/O:
1254 Strong 2^88 cycle random number generator
1256 Linear feedback shift register generator
1258 Strong 64-bit 2^258 cycle random number generator
1260 **tausworthe** is a strong random number generator, but it requires tracking
1261 on the side if we want to ensure that blocks are only read or written
1262 once. **LFSR** guarantees that we never generate the same offset twice, and
1263 it's also less computationally expensive. It's not a true random generator,
1264 however, though for I/O purposes it's typically good enough. **LFSR** only
1265 works with single block sizes, not with workloads that use multiple block
1266 sizes. If used with such a workload, fio may read or write some blocks
1267 multiple times. The default value is **tausworthe**, unless the required
1268 space exceeds 2^32 blocks. If it does, then **tausworthe64** is
1269 selected automatically.
1275 .. option:: blocksize=int[,int][,int], bs=int[,int][,int]
1277 The block size in bytes used for I/O units. Default: 4096. A single value
1278 applies to reads, writes, and trims. Comma-separated values may be
1279 specified for reads, writes, and trims. A value not terminated in a comma
1280 applies to subsequent types.
1285 means 256k for reads, writes and trims.
1288 means 8k for reads, 32k for writes and trims.
1291 means 8k for reads, 32k for writes, and default for trims.
1294 means default for reads, 8k for writes and trims.
1297 means default for reads, 8k for writes, and default for trims.
1299 .. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange]
1301 A range of block sizes in bytes for I/O units. The issued I/O unit will
1302 always be a multiple of the minimum size, unless
1303 :option:`blocksize_unaligned` is set.
1305 Comma-separated ranges may be specified for reads, writes, and trims as
1306 described in :option:`blocksize`.
1308 Example: ``bsrange=1k-4k,2k-8k``.
1310 .. option:: bssplit=str[,str][,str]
1312 Sometimes you want even finer grained control of the block sizes issued, not
1313 just an even split between them. This option allows you to weight various
1314 block sizes, so that you are able to define a specific amount of block sizes
1315 issued. The format for this option is::
1317 bssplit=blocksize/percentage:blocksize/percentage
1319 for as many block sizes as needed. So if you want to define a workload that
1320 has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would write::
1322 bssplit=4k/10:64k/50:32k/40
1324 Ordering does not matter. If the percentage is left blank, fio will fill in
1325 the remaining values evenly. So a bssplit option like this one::
1327 bssplit=4k/50:1k/:32k/
1329 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always add up
1330 to 100, if bssplit is given a range that adds up to more, it will error out.
1332 Comma-separated values may be specified for reads, writes, and trims as
1333 described in :option:`blocksize`.
1335 If you want a workload that has 50% 2k reads and 50% 4k reads, while having
1336 90% 4k writes and 10% 8k writes, you would specify::
1338 bssplit=2k/50:4k/50,4k/90,8k/10
1340 .. option:: blocksize_unaligned, bs_unaligned
1342 If set, fio will issue I/O units with any size within
1343 :option:`blocksize_range`, not just multiples of the minimum size. This
1344 typically won't work with direct I/O, as that normally requires sector
1347 .. option:: bs_is_seq_rand
1349 If this option is set, fio will use the normal read,write blocksize settings
1350 as sequential,random blocksize settings instead. Any random read or write
1351 will use the WRITE blocksize settings, and any sequential read or write will
1352 use the READ blocksize settings.
1354 .. option:: blockalign=int[,int][,int], ba=int[,int][,int]
1356 Boundary to which fio will align random I/O units. Default:
1357 :option:`blocksize`. Minimum alignment is typically 512b for using direct
1358 I/O, though it usually depends on the hardware block size. This option is
1359 mutually exclusive with using a random map for files, so it will turn off
1360 that option. Comma-separated values may be specified for reads, writes, and
1361 trims as described in :option:`blocksize`.
1367 .. option:: zero_buffers
1369 Initialize buffers with all zeros. Default: fill buffers with random data.
1371 .. option:: refill_buffers
1373 If this option is given, fio will refill the I/O buffers on every
1374 submit. The default is to only fill it at init time and reuse that
1375 data. Only makes sense if zero_buffers isn't specified, naturally. If data
1376 verification is enabled, `refill_buffers` is also automatically enabled.
1378 .. option:: scramble_buffers=bool
1380 If :option:`refill_buffers` is too costly and the target is using data
1381 deduplication, then setting this option will slightly modify the I/O buffer
1382 contents to defeat normal de-dupe attempts. This is not enough to defeat
1383 more clever block compression attempts, but it will stop naive dedupe of
1384 blocks. Default: true.
1386 .. option:: buffer_compress_percentage=int
1388 If this is set, then fio will attempt to provide I/O buffer content (on
1389 WRITEs) that compress to the specified level. Fio does this by providing a
1390 mix of random data and a fixed pattern. The fixed pattern is either zeroes,
1391 or the pattern specified by :option:`buffer_pattern`. If the pattern option
1392 is used, it might skew the compression ratio slightly. Note that this is per
1393 block size unit, for file/disk wide compression level that matches this
1394 setting, you'll also want to set :option:`refill_buffers`.
1396 .. option:: buffer_compress_chunk=int
1398 See :option:`buffer_compress_percentage`. This setting allows fio to manage
1399 how big the ranges of random data and zeroed data is. Without this set, fio
1400 will provide :option:`buffer_compress_percentage` of blocksize random data,
1401 followed by the remaining zeroed. With this set to some chunk size smaller
1402 than the block size, fio can alternate random and zeroed data throughout the
1405 .. option:: buffer_pattern=str
1407 If set, fio will fill the I/O buffers with this pattern or with the contents
1408 of a file. If not set, the contents of I/O buffers are defined by the other
1409 options related to buffer contents. The setting can be any pattern of bytes,
1410 and can be prefixed with 0x for hex values. It may also be a string, where
1411 the string must then be wrapped with ``""``. Or it may also be a filename,
1412 where the filename must be wrapped with ``''`` in which case the file is
1413 opened and read. Note that not all the file contents will be read if that
1414 would cause the buffers to overflow. So, for example::
1416 buffer_pattern='filename'
1420 buffer_pattern="abcd"
1428 buffer_pattern=0xdeadface
1430 Also you can combine everything together in any order::
1432 buffer_pattern=0xdeadface"abcd"-12'filename'
1434 .. option:: dedupe_percentage=int
1436 If set, fio will generate this percentage of identical buffers when
1437 writing. These buffers will be naturally dedupable. The contents of the
1438 buffers depend on what other buffer compression settings have been set. It's
1439 possible to have the individual buffers either fully compressible, or not at
1440 all. This option only controls the distribution of unique buffers.
1442 .. option:: invalidate=bool
1444 Invalidate the buffer/page cache parts for this file prior to starting
1445 I/O if the platform and file type support it. Defaults to true.
1446 This will be ignored if :option:`pre_read` is also specified for the
1449 .. option:: sync=bool
1451 Use synchronous I/O for buffered writes. For the majority of I/O engines,
1452 this means using O_SYNC. Default: false.
1454 .. option:: iomem=str, mem=str
1456 Fio can use various types of memory as the I/O unit buffer. The allowed
1460 Use memory from :manpage:`malloc(3)` as the buffers. Default memory
1464 Use shared memory as the buffers. Allocated through
1465 :manpage:`shmget(2)`.
1468 Same as shm, but use huge pages as backing.
1471 Use mmap to allocate buffers. May either be anonymous memory, or can
1472 be file backed if a filename is given after the option. The format
1473 is `mem=mmap:/path/to/file`.
1476 Use a memory mapped huge file as the buffer backing. Append filename
1477 after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`.
1480 Same as mmap, but use a MMAP_SHARED mapping.
1483 Use GPU memory as the buffers for GPUDirect RDMA benchmark.
1485 The area allocated is a function of the maximum allowed bs size for the job,
1486 multiplied by the I/O depth given. Note that for **shmhuge** and
1487 **mmaphuge** to work, the system must have free huge pages allocated. This
1488 can normally be checked and set by reading/writing
1489 :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page
1490 is 4MiB in size. So to calculate the number of huge pages you need for a
1491 given job file, add up the I/O depth of all jobs (normally one unless
1492 :option:`iodepth` is used) and multiply by the maximum bs set. Then divide
1493 that number by the huge page size. You can see the size of the huge pages in
1494 :file:`/proc/meminfo`. If no huge pages are allocated by having a non-zero
1495 number in `nr_hugepages`, using **mmaphuge** or **shmhuge** will fail. Also
1496 see :option:`hugepage-size`.
1498 **mmaphuge** also needs to have hugetlbfs mounted and the file location
1499 should point there. So if it's mounted in :file:`/huge`, you would use
1500 `mem=mmaphuge:/huge/somefile`.
1502 .. option:: iomem_align=int
1504 This indicates the memory alignment of the I/O memory buffers. Note that
1505 the given alignment is applied to the first I/O unit buffer, if using
1506 :option:`iodepth` the alignment of the following buffers are given by the
1507 :option:`bs` used. In other words, if using a :option:`bs` that is a
1508 multiple of the page sized in the system, all buffers will be aligned to
1509 this value. If using a :option:`bs` that is not page aligned, the alignment
1510 of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and
1513 .. option:: hugepage-size=int
1515 Defines the size of a huge page. Must at least be equal to the system
1516 setting, see :file:`/proc/meminfo`. Defaults to 4MiB. Should probably
1517 always be a multiple of megabytes, so using ``hugepage-size=Xm`` is the
1518 preferred way to set this to avoid setting a non-pow-2 bad value.
1520 .. option:: lockmem=int
1522 Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to
1523 simulate a smaller amount of memory. The amount specified is per worker.
1529 .. option:: size=int
1531 The total size of file I/O for each thread of this job. Fio will run until
1532 this many bytes has been transferred, unless runtime is limited by other options
1533 (such as :option:`runtime`, for instance, or increased/decreased by :option:`io_size`).
1534 Fio will divide this size between the available files determined by options
1535 such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is
1536 specified by the job. If the result of division happens to be 0, the size is
1537 set to the physical size of the given files or devices if they exist.
1538 If this option is not specified, fio will use the full size of the given
1539 files or devices. If the files do not exist, size must be given. It is also
1540 possible to give size as a percentage between 1 and 100. If ``size=20%`` is
1541 given, fio will use 20% of the full size of the given files or devices.
1542 Can be combined with :option:`offset` to constrain the start and end range
1543 that I/O will be done within.
1545 .. option:: io_size=int, io_limit=int
1547 Normally fio operates within the region set by :option:`size`, which means
1548 that the :option:`size` option sets both the region and size of I/O to be
1549 performed. Sometimes that is not what you want. With this option, it is
1550 possible to define just the amount of I/O that fio should do. For instance,
1551 if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio
1552 will perform I/O within the first 20GiB but exit when 5GiB have been
1553 done. The opposite is also possible -- if :option:`size` is set to 20GiB,
1554 and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within
1555 the 0..20GiB region.
1557 .. option:: filesize=int
1559 Individual file sizes. May be a range, in which case fio will select sizes
1560 for files at random within the given range and limited to :option:`size` in
1561 total (if that is given). If not given, each created file is the same size.
1562 This option overrides :option:`size` in terms of file size, which means
1563 this value is used as a fixed size or possible range of each file.
1565 .. option:: file_append=bool
1567 Perform I/O after the end of the file. Normally fio will operate within the
1568 size of a file. If this option is set, then fio will append to the file
1569 instead. This has identical behavior to setting :option:`offset` to the size
1570 of a file. This option is ignored on non-regular files.
1572 .. option:: fill_device=bool, fill_fs=bool
1574 Sets size to something really large and waits for ENOSPC (no space left on
1575 device) as the terminating condition. Only makes sense with sequential
1576 write. For a read workload, the mount point will be filled first then I/O
1577 started on the result. This option doesn't make sense if operating on a raw
1578 device node, since the size of that is already known by the file system.
1579 Additionally, writing beyond end-of-device will not return ENOSPC there.
1585 .. option:: ioengine=str
1587 Defines how the job issues I/O to the file. The following types are defined:
1590 Basic :manpage:`read(2)` or :manpage:`write(2)`
1591 I/O. :manpage:`lseek(2)` is used to position the I/O location.
1592 See :option:`fsync` and :option:`fdatasync` for syncing write I/Os.
1595 Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on
1596 all supported operating systems except for Windows.
1599 Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate
1600 queuing by coalescing adjacent I/Os into a single submission.
1603 Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O.
1606 Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O.
1609 Linux native asynchronous I/O. Note that Linux may only support
1610 queued behaviour with non-buffered I/O (set ``direct=1`` or
1612 This engine defines engine specific options.
1615 POSIX asynchronous I/O using :manpage:`aio_read(3)` and
1616 :manpage:`aio_write(3)`.
1619 Solaris native asynchronous I/O.
1622 Windows native asynchronous I/O. Default on Windows.
1625 File is memory mapped with :manpage:`mmap(2)` and data copied
1626 to/from using :manpage:`memcpy(3)`.
1629 :manpage:`splice(2)` is used to transfer the data and
1630 :manpage:`vmsplice(2)` to transfer data from user space to the
1634 SCSI generic sg v3 I/O. May either be synchronous using the SG_IO
1635 ioctl, or if the target is an sg character device we use
1636 :manpage:`read(2)` and :manpage:`write(2)` for asynchronous
1637 I/O. Requires filename option to specify either block or character
1641 Doesn't transfer any data, just pretends to. This is mainly used to
1642 exercise fio itself and for debugging/testing purposes.
1645 Transfer over the network to given ``host:port``. Depending on the
1646 :option:`protocol` used, the :option:`hostname`, :option:`port`,
1647 :option:`listen` and :option:`filename` options are used to specify
1648 what sort of connection to make, while the :option:`protocol` option
1649 determines which protocol will be used. This engine defines engine
1653 Like **net**, but uses :manpage:`splice(2)` and
1654 :manpage:`vmsplice(2)` to map data and send/receive.
1655 This engine defines engine specific options.
1658 Doesn't transfer any data, but burns CPU cycles according to the
1659 :option:`cpuload` and :option:`cpuchunks` options. Setting
1660 :option:`cpuload` =85 will cause that job to do nothing but burn 85%
1661 of the CPU. In case of SMP machines, use :option:`numjobs`
1662 =<no_of_cpu> to get desired CPU usage, as the cpuload only loads a
1663 single CPU at the desired rate. A job never finishes unless there is
1664 at least one non-cpuio job.
1667 The GUASI I/O engine is the Generic Userspace Asyncronous Syscall
1668 Interface approach to async I/O. See
1670 http://www.xmailserver.org/guasi-lib.html
1672 for more info on GUASI.
1675 The RDMA I/O engine supports both RDMA memory semantics
1676 (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the
1677 InfiniBand, RoCE and iWARP protocols.
1680 I/O engine that does regular fallocate to simulate data transfer as
1684 does fallocate(,mode = FALLOC_FL_KEEP_SIZE,).
1687 does fallocate(,mode = 0).
1690 does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE).
1693 I/O engine that sends :manpage:`ftruncate(2)` operations in response
1694 to write (DDIR_WRITE) events. Each ftruncate issued sets the file's
1695 size to the current block offset. Block size is ignored.
1698 I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate
1699 defragment activity in request to DDIR_WRITE event.
1702 I/O engine supporting direct access to Ceph Rados Block Devices
1703 (RBD) via librbd without the need to use the kernel rbd driver. This
1704 ioengine defines engine specific options.
1707 Using Glusterfs libgfapi sync interface to direct access to
1708 Glusterfs volumes without having to go through FUSE. This ioengine
1709 defines engine specific options.
1712 Using Glusterfs libgfapi async interface to direct access to
1713 Glusterfs volumes without having to go through FUSE. This ioengine
1714 defines engine specific options.
1717 Read and write through Hadoop (HDFS). The :file:`filename` option
1718 is used to specify host,port of the hdfs name-node to connect. This
1719 engine interprets offsets a little differently. In HDFS, files once
1720 created cannot be modified. So random writes are not possible. To
1721 imitate this, libhdfs engine expects bunch of small files to be
1722 created over HDFS, and engine will randomly pick a file out of those
1723 files based on the offset generated by fio backend. (see the example
1724 job file to create such files, use ``rw=write`` option). Please
1725 note, you might want to set necessary environment variables to work
1726 with hdfs/libhdfs properly. Each job uses its own connection to
1730 Read, write and erase an MTD character device (e.g.,
1731 :file:`/dev/mtd0`). Discards are treated as erases. Depending on the
1732 underlying device type, the I/O may have to go in a certain pattern,
1733 e.g., on NAND, writing sequentially to erase blocks and discarding
1734 before overwriting. The writetrim mode works well for this
1738 Read and write using filesystem DAX to a file on a filesystem
1739 mounted with DAX on a persistent memory device through the NVML
1743 Read and write using device DAX to a persistent memory device (e.g.,
1744 /dev/dax0.0) through the NVML libpmem library.
1747 Prefix to specify loading an external I/O engine object file. Append
1748 the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load
1749 ioengine :file:`foo.o` in :file:`/tmp`.
1752 I/O engine specific parameters
1753 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1755 In addition, there are some parameters which are only valid when a specific
1756 ioengine is in use. These are used identically to normal parameters, with the
1757 caveat that when used on the command line, they must come after the
1758 :option:`ioengine` that defines them is selected.
1760 .. option:: userspace_reap : [libaio]
1762 Normally, with the libaio engine in use, fio will use the
1763 :manpage:`io_getevents(2)` system call to reap newly returned events. With
1764 this flag turned on, the AIO ring will be read directly from user-space to
1765 reap events. The reaping mode is only enabled when polling for a minimum of
1766 0 events (e.g. when :option:`iodepth_batch_complete` `=0`).
1768 .. option:: hipri : [pvsync2]
1770 Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority
1773 .. option:: cpuload=int : [cpuio]
1775 Attempt to use the specified percentage of CPU cycles. This is a mandatory
1776 option when using cpuio I/O engine.
1778 .. option:: cpuchunks=int : [cpuio]
1780 Split the load into cycles of the given time. In microseconds.
1782 .. option:: exit_on_io_done=bool : [cpuio]
1784 Detect when I/O threads are done, then exit.
1786 .. option:: hostname=str : [netsplice] [net]
1788 The host name or IP address to use for TCP or UDP based I/O. If the job is
1789 a TCP listener or UDP reader, the host name is not used and must be omitted
1790 unless it is a valid UDP multicast address.
1792 .. option:: namenode=str : [libhdfs]
1794 The host name or IP address of a HDFS cluster namenode to contact.
1796 .. option:: port=int
1800 The TCP or UDP port to bind to or connect to. If this is used with
1801 :option:`numjobs` to spawn multiple instances of the same job type, then
1802 this will be the starting port number since fio will use a range of
1807 the listening port of the HFDS cluster namenode.
1809 .. option:: interface=str : [netsplice] [net]
1811 The IP address of the network interface used to send or receive UDP
1814 .. option:: ttl=int : [netsplice] [net]
1816 Time-to-live value for outgoing UDP multicast packets. Default: 1.
1818 .. option:: nodelay=bool : [netsplice] [net]
1820 Set TCP_NODELAY on TCP connections.
1822 .. option:: protocol=str : [netsplice] [net]
1824 .. option:: proto=str : [netsplice] [net]
1826 The network protocol to use. Accepted values are:
1829 Transmission control protocol.
1831 Transmission control protocol V6.
1833 User datagram protocol.
1835 User datagram protocol V6.
1839 When the protocol is TCP or UDP, the port must also be given, as well as the
1840 hostname if the job is a TCP listener or UDP reader. For unix sockets, the
1841 normal filename option should be used and the port is invalid.
1843 .. option:: listen : [net]
1845 For TCP network connections, tell fio to listen for incoming connections
1846 rather than initiating an outgoing connection. The :option:`hostname` must
1847 be omitted if this option is used.
1849 .. option:: pingpong : [net]
1851 Normally a network writer will just continue writing data, and a network
1852 reader will just consume packages. If ``pingpong=1`` is set, a writer will
1853 send its normal payload to the reader, then wait for the reader to send the
1854 same payload back. This allows fio to measure network latencies. The
1855 submission and completion latencies then measure local time spent sending or
1856 receiving, and the completion latency measures how long it took for the
1857 other end to receive and send back. For UDP multicast traffic
1858 ``pingpong=1`` should only be set for a single reader when multiple readers
1859 are listening to the same address.
1861 .. option:: window_size : [net]
1863 Set the desired socket buffer size for the connection.
1865 .. option:: mss : [net]
1867 Set the TCP maximum segment size (TCP_MAXSEG).
1869 .. option:: donorname=str : [e4defrag]
1871 File will be used as a block donor(swap extents between files).
1873 .. option:: inplace=int : [e4defrag]
1875 Configure donor file blocks allocation strategy:
1878 Default. Preallocate donor's file on init.
1880 Allocate space immediately inside defragment event, and free right
1883 .. option:: clustername=str : [rbd]
1885 Specifies the name of the Ceph cluster.
1887 .. option:: rbdname=str : [rbd]
1889 Specifies the name of the RBD.
1891 .. option:: pool=str : [rbd]
1893 Specifies the name of the Ceph pool containing RBD.
1895 .. option:: clientname=str : [rbd]
1897 Specifies the username (without the 'client.' prefix) used to access the
1898 Ceph cluster. If the *clustername* is specified, the *clientname* shall be
1899 the full *type.id* string. If no type. prefix is given, fio will add
1900 'client.' by default.
1902 .. option:: skip_bad=bool : [mtd]
1904 Skip operations against known bad blocks.
1906 .. option:: hdfsdirectory : [libhdfs]
1908 libhdfs will create chunk in this HDFS directory.
1910 .. option:: chunk_size : [libhdfs]
1912 the size of the chunk to use for each file.
1918 .. option:: iodepth=int
1920 Number of I/O units to keep in flight against the file. Note that
1921 increasing *iodepth* beyond 1 will not affect synchronous ioengines (except
1922 for small degrees when :option:`verify_async` is in use). Even async
1923 engines may impose OS restrictions causing the desired depth not to be
1924 achieved. This may happen on Linux when using libaio and not setting
1925 :option:`direct` =1, since buffered I/O is not async on that OS. Keep an
1926 eye on the I/O depth distribution in the fio output to verify that the
1927 achieved depth is as expected. Default: 1.
1929 .. option:: iodepth_batch_submit=int, iodepth_batch=int
1931 This defines how many pieces of I/O to submit at once. It defaults to 1
1932 which means that we submit each I/O as soon as it is available, but can be
1933 raised to submit bigger batches of I/O at the time. If it is set to 0 the
1934 :option:`iodepth` value will be used.
1936 .. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int
1938 This defines how many pieces of I/O to retrieve at once. It defaults to 1
1939 which means that we'll ask for a minimum of 1 I/O in the retrieval process
1940 from the kernel. The I/O retrieval will go on until we hit the limit set by
1941 :option:`iodepth_low`. If this variable is set to 0, then fio will always
1942 check for completed events before queuing more I/O. This helps reduce I/O
1943 latency, at the cost of more retrieval system calls.
1945 .. option:: iodepth_batch_complete_max=int
1947 This defines maximum pieces of I/O to retrieve at once. This variable should
1948 be used along with :option:`iodepth_batch_complete_min` =int variable,
1949 specifying the range of min and max amount of I/O which should be
1950 retrieved. By default it is equal to :option:`iodepth_batch_complete_min`
1955 iodepth_batch_complete_min=1
1956 iodepth_batch_complete_max=<iodepth>
1958 which means that we will retrieve at least 1 I/O and up to the whole
1959 submitted queue depth. If none of I/O has been completed yet, we will wait.
1963 iodepth_batch_complete_min=0
1964 iodepth_batch_complete_max=<iodepth>
1966 which means that we can retrieve up to the whole submitted queue depth, but
1967 if none of I/O has been completed yet, we will NOT wait and immediately exit
1968 the system call. In this example we simply do polling.
1970 .. option:: iodepth_low=int
1972 The low water mark indicating when to start filling the queue
1973 again. Defaults to the same as :option:`iodepth`, meaning that fio will
1974 attempt to keep the queue full at all times. If :option:`iodepth` is set to
1975 e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of
1976 16 requests, it will let the depth drain down to 4 before starting to fill
1979 .. option:: io_submit_mode=str
1981 This option controls how fio submits the I/O to the I/O engine. The default
1982 is `inline`, which means that the fio job threads submit and reap I/O
1983 directly. If set to `offload`, the job threads will offload I/O submission
1984 to a dedicated pool of I/O threads. This requires some coordination and thus
1985 has a bit of extra overhead, especially for lower queue depth I/O where it
1986 can increase latencies. The benefit is that fio can manage submission rates
1987 independently of the device completion rates. This avoids skewed latency
1988 reporting if I/O gets back up on the device side (the coordinated omission
1995 .. option:: thinktime=time
1997 Stall the job for the specified period of time after an I/O has completed before issuing the
1998 next. May be used to simulate processing being done by an application.
1999 When the unit is omitted, the value is given in microseconds. See
2000 :option:`thinktime_blocks` and :option:`thinktime_spin`.
2002 .. option:: thinktime_spin=time
2004 Only valid if :option:`thinktime` is set - pretend to spend CPU time doing
2005 something with the data received, before falling back to sleeping for the
2006 rest of the period specified by :option:`thinktime`. When the unit is
2007 omitted, the value is given in microseconds.
2009 .. option:: thinktime_blocks=int
2011 Only valid if :option:`thinktime` is set - control how many blocks to issue,
2012 before waiting `thinktime` usecs. If not set, defaults to 1 which will make
2013 fio wait `thinktime` usecs after every block. This effectively makes any
2014 queue depth setting redundant, since no more than 1 I/O will be queued
2015 before we have to complete it and do our thinktime. In other words, this
2016 setting effectively caps the queue depth if the latter is larger.
2018 .. option:: rate=int[,int][,int]
2020 Cap the bandwidth used by this job. The number is in bytes/sec, the normal
2021 suffix rules apply. Comma-separated values may be specified for reads,
2022 writes, and trims as described in :option:`blocksize`.
2024 .. option:: rate_min=int[,int][,int]
2026 Tell fio to do whatever it can to maintain at least this bandwidth. Failing
2027 to meet this requirement will cause the job to exit. Comma-separated values
2028 may be specified for reads, writes, and trims as described in
2029 :option:`blocksize`.
2031 .. option:: rate_iops=int[,int][,int]
2033 Cap the bandwidth to this number of IOPS. Basically the same as
2034 :option:`rate`, just specified independently of bandwidth. If the job is
2035 given a block size range instead of a fixed value, the smallest block size
2036 is used as the metric. Comma-separated values may be specified for reads,
2037 writes, and trims as described in :option:`blocksize`.
2039 .. option:: rate_iops_min=int[,int][,int]
2041 If fio doesn't meet this rate of I/O, it will cause the job to exit.
2042 Comma-separated values may be specified for reads, writes, and trims as
2043 described in :option:`blocksize`.
2045 .. option:: rate_process=str
2047 This option controls how fio manages rated I/O submissions. The default is
2048 `linear`, which submits I/O in a linear fashion with fixed delays between
2049 I/Os that gets adjusted based on I/O completion rates. If this is set to
2050 `poisson`, fio will submit I/O based on a more real world random request
2051 flow, known as the Poisson process
2052 (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be
2053 10^6 / IOPS for the given workload.
2059 .. option:: latency_target=time
2061 If set, fio will attempt to find the max performance point that the given
2062 workload will run at while maintaining a latency below this target. When
2063 the unit is omitted, the value is given in microseconds. See
2064 :option:`latency_window` and :option:`latency_percentile`.
2066 .. option:: latency_window=time
2068 Used with :option:`latency_target` to specify the sample window that the job
2069 is run at varying queue depths to test the performance. When the unit is
2070 omitted, the value is given in microseconds.
2072 .. option:: latency_percentile=float
2074 The percentage of I/Os that must fall within the criteria specified by
2075 :option:`latency_target` and :option:`latency_window`. If not set, this
2076 defaults to 100.0, meaning that all I/Os must be equal or below to the value
2077 set by :option:`latency_target`.
2079 .. option:: max_latency=time
2081 If set, fio will exit the job with an ETIMEDOUT error if it exceeds this
2082 maximum latency. When the unit is omitted, the value is given in
2085 .. option:: rate_cycle=int
2087 Average bandwidth for :option:`rate` and :option:`rate_min` over this number
2088 of milliseconds. Defaults to 1000.
2094 .. option:: write_iolog=str
2096 Write the issued I/O patterns to the specified file. See
2097 :option:`read_iolog`. Specify a separate file for each job, otherwise the
2098 iologs will be interspersed and the file may be corrupt.
2100 .. option:: read_iolog=str
2102 Open an iolog with the specified file name and replay the I/O patterns it
2103 contains. This can be used to store a workload and replay it sometime
2104 later. The iolog given may also be a blktrace binary file, which allows fio
2105 to replay a workload captured by :command:`blktrace`. See
2106 :manpage:`blktrace(8)` for how to capture such logging data. For blktrace
2107 replay, the file needs to be turned into a blkparse binary data file first
2108 (``blkparse <device> -o /dev/null -d file_for_fio.bin``).
2110 .. option:: replay_no_stall=int
2112 When replaying I/O with :option:`read_iolog` the default behavior is to
2113 attempt to respect the time stamps within the log and replay them with the
2114 appropriate delay between IOPS. By setting this variable fio will not
2115 respect the timestamps and attempt to replay them as fast as possible while
2116 still respecting ordering. The result is the same I/O pattern to a given
2117 device, but different timings.
2119 .. option:: replay_redirect=str
2121 While replaying I/O patterns using :option:`read_iolog` the default behavior
2122 is to replay the IOPS onto the major/minor device that each IOP was recorded
2123 from. This is sometimes undesirable because on a different machine those
2124 major/minor numbers can map to a different device. Changing hardware on the
2125 same system can also result in a different major/minor mapping.
2126 ``replay_redirect`` causes all IOPS to be replayed onto the single specified
2127 device regardless of the device it was recorded
2128 from. i.e. :option:`replay_redirect` = :file:`/dev/sdc` would cause all I/O
2129 in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means
2130 multiple devices will be replayed onto a single device, if the trace
2131 contains multiple devices. If you want multiple devices to be replayed
2132 concurrently to multiple redirected devices you must blkparse your trace
2133 into separate traces and replay them with independent fio invocations.
2134 Unfortunately this also breaks the strict time ordering between multiple
2137 .. option:: replay_align=int
2139 Force alignment of I/O offsets and lengths in a trace to this power of 2
2142 .. option:: replay_scale=int
2144 Scale sector offsets down by this factor when replaying traces.
2147 Threads, processes and job synchronization
2148 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2152 Fio defaults to forking jobs, however if this option is given, fio will use
2153 POSIX Threads function :manpage:`pthread_create(3)` to create threads instead
2154 of forking processes.
2156 .. option:: wait_for=str
2158 Specifies the name of the already defined job to wait for. Single waitee
2159 name only may be specified. If set, the job won't be started until all
2160 workers of the waitee job are done.
2162 ``wait_for`` operates on the job name basis, so there are a few
2163 limitations. First, the waitee must be defined prior to the waiter job
2164 (meaning no forward references). Second, if a job is being referenced as a
2165 waitee, it must have a unique name (no duplicate waitees).
2167 .. option:: nice=int
2169 Run the job with the given nice value. See man :manpage:`nice(2)`.
2171 On Windows, values less than -15 set the process class to "High"; -1 through
2172 -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle"
2175 .. option:: prio=int
2177 Set the I/O priority value of this job. Linux limits us to a positive value
2178 between 0 and 7, with 0 being the highest. See man
2179 :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
2180 systems since meaning of priority may differ.
2182 .. option:: prioclass=int
2184 Set the I/O priority class. See man :manpage:`ionice(1)`.
2186 .. option:: cpumask=int
2188 Set the CPU affinity of this job. The parameter given is a bitmask of
2189 allowed CPU's the job may run on. So if you want the allowed CPUs to be 1
2190 and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man
2191 :manpage:`sched_setaffinity(2)`. This may not work on all supported
2192 operating systems or kernel versions. This option doesn't work well for a
2193 higher CPU count than what you can store in an integer mask, so it can only
2194 control cpus 1-32. For boxes with larger CPU counts, use
2195 :option:`cpus_allowed`.
2197 .. option:: cpus_allowed=str
2199 Controls the same options as :option:`cpumask`, but it allows a text setting
2200 of the permitted CPUs instead. So to use CPUs 1 and 5, you would specify
2201 ``cpus_allowed=1,5``. This options also allows a range of CPUs. Say you
2202 wanted a binding to CPUs 1, 5, and 8-15, you would set
2203 ``cpus_allowed=1,5,8-15``.
2205 .. option:: cpus_allowed_policy=str
2207 Set the policy of how fio distributes the CPUs specified by
2208 :option:`cpus_allowed` or cpumask. Two policies are supported:
2211 All jobs will share the CPU set specified.
2213 Each job will get a unique CPU from the CPU set.
2215 **shared** is the default behaviour, if the option isn't specified. If
2216 **split** is specified, then fio will will assign one cpu per job. If not
2217 enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs
2220 .. option:: numa_cpu_nodes=str
2222 Set this job running on specified NUMA nodes' CPUs. The arguments allow
2223 comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable
2224 numa options support, fio must be built on a system with libnuma-dev(el)
2227 .. option:: numa_mem_policy=str
2229 Set this job's memory policy and corresponding NUMA nodes. Format of the
2234 ``mode`` is one of the following memory policy: ``default``, ``prefer``,
2235 ``bind``, ``interleave``, ``local`` For ``default`` and ``local`` memory
2236 policy, no node is needed to be specified. For ``prefer``, only one node is
2237 allowed. For ``bind`` and ``interleave``, it allow comma delimited list of
2238 numbers, A-B ranges, or `all`.
2240 .. option:: cgroup=str
2242 Add job to this control group. If it doesn't exist, it will be created. The
2243 system must have a mounted cgroup blkio mount point for this to work. If
2244 your system doesn't have it mounted, you can do so with::
2246 # mount -t cgroup -o blkio none /cgroup
2248 .. option:: cgroup_weight=int
2250 Set the weight of the cgroup to this value. See the documentation that comes
2251 with the kernel, allowed values are in the range of 100..1000.
2253 .. option:: cgroup_nodelete=bool
2255 Normally fio will delete the cgroups it has created after the job
2256 completion. To override this behavior and to leave cgroups around after the
2257 job completion, set ``cgroup_nodelete=1``. This can be useful if one wants
2258 to inspect various cgroup files after job completion. Default: false.
2260 .. option:: flow_id=int
2262 The ID of the flow. If not specified, it defaults to being a global
2263 flow. See :option:`flow`.
2265 .. option:: flow=int
2267 Weight in token-based flow control. If this value is used, then there is a
2268 'flow counter' which is used to regulate the proportion of activity between
2269 two or more jobs. Fio attempts to keep this flow counter near zero. The
2270 ``flow`` parameter stands for how much should be added or subtracted to the
2271 flow counter on each iteration of the main I/O loop. That is, if one job has
2272 ``flow=8`` and another job has ``flow=-1``, then there will be a roughly 1:8
2273 ratio in how much one runs vs the other.
2275 .. option:: flow_watermark=int
2277 The maximum value that the absolute value of the flow counter is allowed to
2278 reach before the job must wait for a lower value of the counter.
2280 .. option:: flow_sleep=int
2282 The period of time, in microseconds, to wait after the flow watermark has
2283 been exceeded before retrying operations.
2285 .. option:: stonewall, wait_for_previous
2287 Wait for preceding jobs in the job file to exit, before starting this
2288 one. Can be used to insert serialization points in the job file. A stone
2289 wall also implies starting a new reporting group, see
2290 :option:`group_reporting`.
2294 When one job finishes, terminate the rest. The default is to wait for each
2295 job to finish, sometimes that is not the desired action.
2297 .. option:: exec_prerun=str
2299 Before running this job, issue the command specified through
2300 :manpage:`system(3)`. Output is redirected in a file called
2301 :file:`jobname.prerun.txt`.
2303 .. option:: exec_postrun=str
2305 After the job completes, issue the command specified though
2306 :manpage:`system(3)`. Output is redirected in a file called
2307 :file:`jobname.postrun.txt`.
2311 Instead of running as the invoking user, set the user ID to this value
2312 before the thread/process does any work.
2316 Set group ID, see :option:`uid`.
2322 .. option:: verify_only
2324 Do not perform specified workload, only verify data still matches previous
2325 invocation of this workload. This option allows one to check data multiple
2326 times at a later date without overwriting it. This option makes sense only
2327 for workloads that write data, and does not support workloads with the
2328 :option:`time_based` option set.
2330 .. option:: do_verify=bool
2332 Run the verify phase after a write phase. Only valid if :option:`verify` is
2335 .. option:: verify=str
2337 If writing to a file, fio can verify the file contents after each iteration
2338 of the job. Each verification method also implies verification of special
2339 header, which is written to the beginning of each block. This header also
2340 includes meta information, like offset of the block, block number, timestamp
2341 when block was written, etc. :option:`verify` can be combined with
2342 :option:`verify_pattern` option. The allowed values are:
2345 Use an md5 sum of the data area and store it in the header of
2349 Use an experimental crc64 sum of the data area and store it in the
2350 header of each block.
2353 Use a crc32c sum of the data area and store it in the header of each
2357 Use hardware assisted crc32c calculation provided on SSE4.2 enabled
2358 processors. Falls back to regular software crc32c, if not supported
2362 Use a crc32 sum of the data area and store it in the header of each
2366 Use a crc16 sum of the data area and store it in the header of each
2370 Use a crc7 sum of the data area and store it in the header of each
2374 Use xxhash as the checksum function. Generally the fastest software
2375 checksum that fio supports.
2378 Use sha512 as the checksum function.
2381 Use sha256 as the checksum function.
2384 Use optimized sha1 as the checksum function.
2387 Use optimized sha3-224 as the checksum function.
2390 Use optimized sha3-256 as the checksum function.
2393 Use optimized sha3-384 as the checksum function.
2396 Use optimized sha3-512 as the checksum function.
2399 This option is deprecated, since now meta information is included in
2400 generic verification header and meta verification happens by
2401 default. For detailed information see the description of the
2402 :option:`verify` setting. This option is kept because of
2403 compatibility's sake with old configurations. Do not use it.
2406 Verify a strict pattern. Normally fio includes a header with some
2407 basic information and checksumming, but if this option is set, only
2408 the specific pattern set with :option:`verify_pattern` is verified.
2411 Only pretend to verify. Useful for testing internals with
2412 :option:`ioengine` `=null`, not for much else.
2414 This option can be used for repeated burn-in tests of a system to make sure
2415 that the written data is also correctly read back. If the data direction
2416 given is a read or random read, fio will assume that it should verify a
2417 previously written file. If the data direction includes any form of write,
2418 the verify will be of the newly written data.
2420 .. option:: verifysort=bool
2422 If true, fio will sort written verify blocks when it deems it faster to read
2423 them back in a sorted manner. This is often the case when overwriting an
2424 existing file, since the blocks are already laid out in the file system. You
2425 can ignore this option unless doing huge amounts of really fast I/O where
2426 the red-black tree sorting CPU time becomes significant. Default: true.
2428 .. option:: verifysort_nr=int
2430 Pre-load and sort verify blocks for a read workload.
2432 .. option:: verify_offset=int
2434 Swap the verification header with data somewhere else in the block before
2435 writing. It is swapped back before verifying.
2437 .. option:: verify_interval=int
2439 Write the verification header at a finer granularity than the
2440 :option:`blocksize`. It will be written for chunks the size of
2441 ``verify_interval``. :option:`blocksize` should divide this evenly.
2443 .. option:: verify_pattern=str
2445 If set, fio will fill the I/O buffers with this pattern. Fio defaults to
2446 filling with totally random bytes, but sometimes it's interesting to fill
2447 with a known pattern for I/O verification purposes. Depending on the width
2448 of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time(it can
2449 be either a decimal or a hex number). The ``verify_pattern`` if larger than
2450 a 32-bit quantity has to be a hex number that starts with either "0x" or
2451 "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o
2452 format, which means that for each block offset will be written and then
2453 verified back, e.g.::
2457 Or use combination of everything::
2459 verify_pattern=0xff%o"abcd"-12
2461 .. option:: verify_fatal=bool
2463 Normally fio will keep checking the entire contents before quitting on a
2464 block verification failure. If this option is set, fio will exit the job on
2465 the first observed failure. Default: false.
2467 .. option:: verify_dump=bool
2469 If set, dump the contents of both the original data block and the data block
2470 we read off disk to files. This allows later analysis to inspect just what
2471 kind of data corruption occurred. Off by default.
2473 .. option:: verify_async=int
2475 Fio will normally verify I/O inline from the submitting thread. This option
2476 takes an integer describing how many async offload threads to create for I/O
2477 verification instead, causing fio to offload the duty of verifying I/O
2478 contents to one or more separate threads. If using this offload option, even
2479 sync I/O engines can benefit from using an :option:`iodepth` setting higher
2480 than 1, as it allows them to have I/O in flight while verifies are running.
2481 Defaults to 0 async threads, i.e. verification is not asynchronous.
2483 .. option:: verify_async_cpus=str
2485 Tell fio to set the given CPU affinity on the async I/O verification
2486 threads. See :option:`cpus_allowed` for the format used.
2488 .. option:: verify_backlog=int
2490 Fio will normally verify the written contents of a job that utilizes verify
2491 once that job has completed. In other words, everything is written then
2492 everything is read back and verified. You may want to verify continually
2493 instead for a variety of reasons. Fio stores the meta data associated with
2494 an I/O block in memory, so for large verify workloads, quite a bit of memory
2495 would be used up holding this meta data. If this option is enabled, fio will
2496 write only N blocks before verifying these blocks.
2498 .. option:: verify_backlog_batch=int
2500 Control how many blocks fio will verify if :option:`verify_backlog` is
2501 set. If not set, will default to the value of :option:`verify_backlog`
2502 (meaning the entire queue is read back and verified). If
2503 ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all
2504 blocks will be verified, if ``verify_backlog_batch`` is larger than
2505 :option:`verify_backlog`, some blocks will be verified more than once.
2507 .. option:: verify_state_save=bool
2509 When a job exits during the write phase of a verify workload, save its
2510 current state. This allows fio to replay up until that point, if the verify
2511 state is loaded for the verify read phase. The format of the filename is,
2514 <type>-<jobname>-<jobindex>-verify.state.
2516 <type> is "local" for a local run, "sock" for a client/server socket
2517 connection, and "ip" (192.168.0.1, for instance) for a networked
2518 client/server connection. Defaults to true.
2520 .. option:: verify_state_load=bool
2522 If a verify termination trigger was used, fio stores the current write state
2523 of each thread. This can be used at verification time so that fio knows how
2524 far it should verify. Without this information, fio will run a full
2525 verification pass, according to the settings in the job file used. Default
2528 .. option:: trim_percentage=int
2530 Number of verify blocks to discard/trim.
2532 .. option:: trim_verify_zero=bool
2534 Verify that trim/discarded blocks are returned as zeroes.
2536 .. option:: trim_backlog=int
2538 Verify that trim/discarded blocks are returned as zeroes.
2540 .. option:: trim_backlog_batch=int
2542 Trim this number of I/O blocks.
2544 .. option:: experimental_verify=bool
2546 Enable experimental verification.
2552 .. option:: steadystate=str:float, ss=str:float
2554 Define the criterion and limit for assessing steady state performance. The
2555 first parameter designates the criterion whereas the second parameter sets
2556 the threshold. When the criterion falls below the threshold for the
2557 specified duration, the job will stop. For example, `iops_slope:0.1%` will
2558 direct fio to terminate the job when the least squares regression slope
2559 falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled
2560 this will apply to all jobs in the group. Below is the list of available
2561 steady state assessment criteria. All assessments are carried out using only
2562 data from the rolling collection window. Threshold limits can be expressed
2563 as a fixed value or as a percentage of the mean in the collection window.
2566 Collect IOPS data. Stop the job if all individual IOPS measurements
2567 are within the specified limit of the mean IOPS (e.g., ``iops:2``
2568 means that all individual IOPS values must be within 2 of the mean,
2569 whereas ``iops:0.2%`` means that all individual IOPS values must be
2570 within 0.2% of the mean IOPS to terminate the job).
2573 Collect IOPS data and calculate the least squares regression
2574 slope. Stop the job if the slope falls below the specified limit.
2577 Collect bandwidth data. Stop the job if all individual bandwidth
2578 measurements are within the specified limit of the mean bandwidth.
2581 Collect bandwidth data and calculate the least squares regression
2582 slope. Stop the job if the slope falls below the specified limit.
2584 .. option:: steadystate_duration=time, ss_dur=time
2586 A rolling window of this duration will be used to judge whether steady state
2587 has been reached. Data will be collected once per second. The default is 0
2588 which disables steady state detection. When the unit is omitted, the
2589 value is given in seconds.
2591 .. option:: steadystate_ramp_time=time, ss_ramp=time
2593 Allow the job to run for the specified duration before beginning data
2594 collection for checking the steady state job termination criterion. The
2595 default is 0. When the unit is omitted, the value is given in seconds.
2598 Measurements and reporting
2599 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2601 .. option:: per_job_logs=bool
2603 If set, this generates bw/clat/iops log with per file private filenames. If
2604 not set, jobs with identical names will share the log filename. Default:
2607 .. option:: group_reporting
2609 It may sometimes be interesting to display statistics for groups of jobs as
2610 a whole instead of for each individual job. This is especially true if
2611 :option:`numjobs` is used; looking at individual thread/process output
2612 quickly becomes unwieldy. To see the final report per-group instead of
2613 per-job, use :option:`group_reporting`. Jobs in a file will be part of the
2614 same reporting group, unless if separated by a :option:`stonewall`, or by
2615 using :option:`new_group`.
2617 .. option:: new_group
2619 Start a new reporting group. See: :option:`group_reporting`. If not given,
2620 all jobs in a file will be part of the same reporting group, unless
2621 separated by a :option:`stonewall`.
2625 By default, fio collects and shows final output results for all jobs
2626 that run. If this option is set to 0, then fio will ignore it in
2627 the final stat output.
2629 .. option:: write_bw_log=str
2631 If given, write a bandwidth log for this job. Can be used to store data of
2632 the bandwidth of the jobs in their lifetime. The included
2633 :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these
2634 text files into nice graphs. See :option:`write_lat_log` for behaviour of
2635 given filename. For this option, the postfix is :file:`_bw.x.log`, where `x`
2636 is the index of the job (`1..N`, where `N` is the number of jobs). If
2637 :option:`per_job_logs` is false, then the filename will not include the job
2638 index. See `Log File Formats`_.
2640 .. option:: write_lat_log=str
2642 Same as :option:`write_bw_log`, except that this option stores I/O
2643 submission, completion, and total latencies instead. If no filename is given
2644 with this option, the default filename of :file:`jobname_type.log` is
2645 used. Even if the filename is given, fio will still append the type of
2646 log. So if one specifies::
2650 The actual log names will be :file:`foo_slat.x.log`, :file:`foo_clat.x.log`,
2651 and :file:`foo_lat.x.log`, where `x` is the index of the job (1..N, where N
2652 is the number of jobs). This helps :command:`fio_generate_plot` find the
2653 logs automatically. If :option:`per_job_logs` is false, then the filename
2654 will not include the job index. See `Log File Formats`_.
2656 .. option:: write_hist_log=str
2658 Same as :option:`write_lat_log`, but writes I/O completion latency
2659 histograms. If no filename is given with this option, the default filename
2660 of :file:`jobname_clat_hist.x.log` is used, where `x` is the index of the
2661 job (1..N, where `N` is the number of jobs). Even if the filename is given,
2662 fio will still append the type of log. If :option:`per_job_logs` is false,
2663 then the filename will not include the job index. See `Log File Formats`_.
2665 .. option:: write_iops_log=str
2667 Same as :option:`write_bw_log`, but writes IOPS. If no filename is given
2668 with this option, the default filename of :file:`jobname_type.x.log` is
2669 used,where `x` is the index of the job (1..N, where `N` is the number of
2670 jobs). Even if the filename is given, fio will still append the type of
2671 log. If :option:`per_job_logs` is false, then the filename will not include
2672 the job index. See `Log File Formats`_.
2674 .. option:: log_avg_msec=int
2676 By default, fio will log an entry in the iops, latency, or bw log for every
2677 I/O that completes. When writing to the disk log, that can quickly grow to a
2678 very large size. Setting this option makes fio average the each log entry
2679 over the specified period of time, reducing the resolution of the log. See
2680 :option:`log_max_value` as well. Defaults to 0, logging all entries.
2682 .. option:: log_hist_msec=int
2684 Same as :option:`log_avg_msec`, but logs entries for completion latency
2685 histograms. Computing latency percentiles from averages of intervals using
2686 :option:`log_avg_msec` is inaccurate. Setting this option makes fio log
2687 histogram entries over the specified period of time, reducing log sizes for
2688 high IOPS devices while retaining percentile accuracy. See
2689 :option:`log_hist_coarseness` as well. Defaults to 0, meaning histogram
2690 logging is disabled.
2692 .. option:: log_hist_coarseness=int
2694 Integer ranging from 0 to 6, defining the coarseness of the resolution of
2695 the histogram logs enabled with :option:`log_hist_msec`. For each increment
2696 in coarseness, fio outputs half as many bins. Defaults to 0, for which
2697 histogram logs contain 1216 latency bins. See `Log File Formats`_.
2699 .. option:: log_max_value=bool
2701 If :option:`log_avg_msec` is set, fio logs the average over that window. If
2702 you instead want to log the maximum value, set this option to 1. Defaults to
2703 0, meaning that averaged values are logged.
2705 .. option:: log_offset=int
2707 If this is set, the iolog options will include the byte offset for the I/O
2708 entry as well as the other data values.
2710 .. option:: log_compression=int
2712 If this is set, fio will compress the I/O logs as it goes, to keep the
2713 memory footprint lower. When a log reaches the specified size, that chunk is
2714 removed and compressed in the background. Given that I/O logs are fairly
2715 highly compressible, this yields a nice memory savings for longer runs. The
2716 downside is that the compression will consume some background CPU cycles, so
2717 it may impact the run. This, however, is also true if the logging ends up
2718 consuming most of the system memory. So pick your poison. The I/O logs are
2719 saved normally at the end of a run, by decompressing the chunks and storing
2720 them in the specified log file. This feature depends on the availability of
2723 .. option:: log_compression_cpus=str
2725 Define the set of CPUs that are allowed to handle online log compression for
2726 the I/O jobs. This can provide better isolation between performance
2727 sensitive jobs, and background compression work.
2729 .. option:: log_store_compressed=bool
2731 If set, fio will store the log files in a compressed format. They can be
2732 decompressed with fio, using the :option:`--inflate-log` command line
2733 parameter. The files will be stored with a :file:`.fz` suffix.
2735 .. option:: log_unix_epoch=bool
2737 If set, fio will log Unix timestamps to the log files produced by enabling
2738 write_type_log for each log type, instead of the default zero-based
2741 .. option:: block_error_percentiles=bool
2743 If set, record errors in trim block-sized units from writes and trims and
2744 output a histogram of how many trims it took to get to errors, and what kind
2745 of error was encountered.
2747 .. option:: bwavgtime=int
2749 Average the calculated bandwidth over the given time. Value is specified in
2750 milliseconds. If the job also does bandwidth logging through
2751 :option:`write_bw_log`, then the minimum of this option and
2752 :option:`log_avg_msec` will be used. Default: 500ms.
2754 .. option:: iopsavgtime=int
2756 Average the calculated IOPS over the given time. Value is specified in
2757 milliseconds. If the job also does IOPS logging through
2758 :option:`write_iops_log`, then the minimum of this option and
2759 :option:`log_avg_msec` will be used. Default: 500ms.
2761 .. option:: disk_util=bool
2763 Generate disk utilization statistics, if the platform supports it.
2766 .. option:: disable_lat=bool
2768 Disable measurements of total latency numbers. Useful only for cutting back
2769 the number of calls to :manpage:`gettimeofday(2)`, as that does impact
2770 performance at really high IOPS rates. Note that to really get rid of a
2771 large amount of these calls, this option must be used with
2772 :option:`disable_slat` and :option:`disable_bw_measurement` as well.
2774 .. option:: disable_clat=bool
2776 Disable measurements of completion latency numbers. See
2777 :option:`disable_lat`.
2779 .. option:: disable_slat=bool
2781 Disable measurements of submission latency numbers. See
2782 :option:`disable_slat`.
2784 .. option:: disable_bw_measurement=bool, disable_bw=bool
2786 Disable measurements of throughput/bandwidth numbers. See
2787 :option:`disable_lat`.
2789 .. option:: clat_percentiles=bool
2791 Enable the reporting of percentiles of completion latencies.
2793 .. option:: percentile_list=float_list
2795 Overwrite the default list of percentiles for completion latencies and the
2796 block error histogram. Each number is a floating number in the range
2797 (0,100], and the maximum length of the list is 20. Use ``:`` to separate the
2798 numbers, and list the numbers in ascending order. For example,
2799 ``--percentile_list=99.5:99.9`` will cause fio to report the values of
2800 completion latency below which 99.5% and 99.9% of the observed latencies
2807 .. option:: exitall_on_error
2809 When one job finishes in error, terminate the rest. The default is to wait
2810 for each job to finish.
2812 .. option:: continue_on_error=str
2814 Normally fio will exit the job on the first observed failure. If this option
2815 is set, fio will continue the job when there is a 'non-fatal error' (EIO or
2816 EILSEQ) until the runtime is exceeded or the I/O size specified is
2817 completed. If this option is used, there are two more stats that are
2818 appended, the total error count and the first error. The error field given
2819 in the stats is the first error that was hit during the run.
2821 The allowed values are:
2824 Exit on any I/O or verify errors.
2827 Continue on read errors, exit on all others.
2830 Continue on write errors, exit on all others.
2833 Continue on any I/O error, exit on all others.
2836 Continue on verify errors, exit on all others.
2839 Continue on all errors.
2842 Backward-compatible alias for 'none'.
2845 Backward-compatible alias for 'all'.
2847 .. option:: ignore_error=str
2849 Sometimes you want to ignore some errors during test in that case you can
2850 specify error list for each error type, instead of only being able to
2851 ignore the default 'non-fatal error' using :option:`continue_on_error`.
2852 ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for
2853 given error type is separated with ':'. Error may be symbol ('ENOSPC',
2854 'ENOMEM') or integer. Example::
2856 ignore_error=EAGAIN,ENOSPC:122
2858 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from
2859 WRITE. This option works by overriding :option:`continue_on_error` with
2860 the list of errors for each error type if any.
2862 .. option:: error_dump=bool
2864 If set dump every error even if it is non fatal, true by default. If
2865 disabled only fatal error will be dumped.
2867 Running predefined workloads
2868 ----------------------------
2870 Fio includes predefined profiles that mimic the I/O workloads generated by
2873 .. option:: profile=str
2875 The predefined workload to run. Current profiles are:
2878 Threaded I/O bench (tiotest/tiobench) like workload.
2881 Aerospike Certification Tool (ACT) like workload.
2883 To view a profile's additional options use :option:`--cmdhelp` after specifying
2884 the profile. For example::
2886 $ fio --profile=act --cmdhelp
2891 .. option:: device-names=str
2896 .. option:: load=int
2899 ACT load multiplier. Default: 1.
2901 .. option:: test-duration=time
2904 How long the entire test takes to run. Default: 24h.
2906 .. option:: threads-per-queue=int
2909 Number of read IO threads per device. Default: 8.
2911 .. option:: read-req-num-512-blocks=int
2914 Number of 512B blocks to read at the time. Default: 3.
2916 .. option:: large-block-op-kbytes=int
2919 Size of large block ops in KiB (writes). Default: 131072.
2924 Set to run ACT prep phase.
2926 Tiobench profile options
2927 ~~~~~~~~~~~~~~~~~~~~~~~~
2929 .. option:: size=str
2934 .. option:: block=int
2937 Block size in bytes. Default: 4096.
2939 .. option:: numruns=int
2949 .. option:: threads=int
2954 Interpreting the output
2955 -----------------------
2957 Fio spits out a lot of output. While running, fio will display the status of the
2958 jobs created. An example of that would be::
2960 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]
2962 The characters inside the square brackets denote the current status of each
2963 thread. The possible values (in typical life cycle order) are:
2965 +------+-----+-----------------------------------------------------------+
2967 +======+=====+===========================================================+
2968 | P | | Thread setup, but not started. |
2969 +------+-----+-----------------------------------------------------------+
2970 | C | | Thread created. |
2971 +------+-----+-----------------------------------------------------------+
2972 | I | | Thread initialized, waiting or generating necessary data. |
2973 +------+-----+-----------------------------------------------------------+
2974 | | p | Thread running pre-reading file(s). |
2975 +------+-----+-----------------------------------------------------------+
2976 | | R | Running, doing sequential reads. |
2977 +------+-----+-----------------------------------------------------------+
2978 | | r | Running, doing random reads. |
2979 +------+-----+-----------------------------------------------------------+
2980 | | W | Running, doing sequential writes. |
2981 +------+-----+-----------------------------------------------------------+
2982 | | w | Running, doing random writes. |
2983 +------+-----+-----------------------------------------------------------+
2984 | | M | Running, doing mixed sequential reads/writes. |
2985 +------+-----+-----------------------------------------------------------+
2986 | | m | Running, doing mixed random reads/writes. |
2987 +------+-----+-----------------------------------------------------------+
2988 | | F | Running, currently waiting for :manpage:`fsync(2)` |
2989 +------+-----+-----------------------------------------------------------+
2990 | | V | Running, doing verification of written data. |
2991 +------+-----+-----------------------------------------------------------+
2992 | E | | Thread exited, not reaped by main thread yet. |
2993 +------+-----+-----------------------------------------------------------+
2994 | _ | | Thread reaped, or |
2995 +------+-----+-----------------------------------------------------------+
2996 | X | | Thread reaped, exited with an error. |
2997 +------+-----+-----------------------------------------------------------+
2998 | K | | Thread reaped, exited due to signal. |
2999 +------+-----+-----------------------------------------------------------+
3001 Fio will condense the thread string as not to take up more space on the command
3002 line as is needed. For instance, if you have 10 readers and 10 writers running,
3003 the output would look like this::
3005 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]
3007 Fio will still maintain the ordering, though. So the above means that jobs 1..10
3008 are readers, and 11..20 are writers.
3010 The other values are fairly self explanatory -- number of threads currently
3011 running and doing I/O, the number of currently open files (f=), the rate of I/O
3012 since last check (read speed listed first, then write speed and optionally trim
3013 speed), and the estimated completion percentage and time for the current
3014 running group. It's impossible to estimate runtime of the following groups (if
3015 any). Note that the string is displayed in order, so it's possible to tell which
3016 of the jobs are currently doing what. The first character is the first job
3017 defined in the job file, and so forth.
3019 When fio is done (or interrupted by :kbd:`ctrl-c`), it will show the data for
3020 each thread, group of threads, and disks in that order. For each data direction,
3021 the output looks like::
3023 Client1 (g=0): err= 0:
3024 write: io= 32MiB, bw= 666KiB/s, iops=89 , runt= 50320msec
3025 slat (msec): min= 0, max= 136, avg= 0.03, stdev= 1.92
3026 clat (msec): min= 0, max= 631, avg=48.50, stdev=86.82
3027 bw (KiB/s) : min= 0, max= 1196, per=51.00%, avg=664.02, stdev=681.68
3028 cpu : usr=1.49%, sys=0.25%, ctx=7969, majf=0, minf=17
3029 IO depths : 1=0.1%, 2=0.3%, 4=0.5%, 8=99.0%, 16=0.0%, 32=0.0%, >32=0.0%
3030 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3031 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
3032 issued r/w: total=0/32768, short=0/0
3033 lat (msec): 2=1.6%, 4=0.0%, 10=3.2%, 20=12.8%, 50=38.4%, 100=24.8%,
3034 lat (msec): 250=15.2%, 500=0.0%, 750=0.0%, 1000=0.0%, >=2048=0.0%
3036 The client number is printed, along with the group id and error of that
3037 thread. Below is the I/O statistics, here for writes. In the order listed, they
3041 Number of megabytes I/O performed.
3044 Average bandwidth rate.
3047 Average I/Os performed per second.
3050 The runtime of that thread.
3053 Submission latency (avg being the average, stdev being the standard
3054 deviation). This is the time it took to submit the I/O. For sync I/O,
3055 the slat is really the completion latency, since queue/complete is one
3056 operation there. This value can be in milliseconds or microseconds, fio
3057 will choose the most appropriate base and print that. In the example
3058 above, milliseconds is the best scale. Note: in :option:`--minimal` mode
3059 latencies are always expressed in microseconds.
3062 Completion latency. Same names as slat, this denotes the time from
3063 submission to completion of the I/O pieces. For sync I/O, clat will
3064 usually be equal (or very close) to 0, as the time from submit to
3065 complete is basically just CPU time (I/O has already been done, see slat
3069 Bandwidth. Same names as the xlat stats, but also includes an
3070 approximate percentage of total aggregate bandwidth this thread received
3071 in this group. This last value is only really useful if the threads in
3072 this group are on the same disk, since they are then competing for disk
3076 CPU usage. User and system time, along with the number of context
3077 switches this thread went through, usage of system and user time, and
3078 finally the number of major and minor page faults. The CPU utilization
3079 numbers are averages for the jobs in that reporting group, while the
3080 context and fault counters are summed.
3083 The distribution of I/O depths over the job life time. The numbers are
3084 divided into powers of 2, so for example the 16= entries includes depths
3085 up to that value but higher than the previous entry. In other words, it
3086 covers the range from 16 to 31.
3089 How many pieces of I/O were submitting in a single submit call. Each
3090 entry denotes that amount and below, until the previous entry -- e.g.,
3091 8=100% mean that we submitted anywhere in between 5-8 I/Os per submit
3095 Like the above submit number, but for completions instead.
3098 The number of read/write requests issued, and how many of them were
3102 The distribution of I/O completion latencies. This is the time from when
3103 I/O leaves fio and when it gets completed. The numbers follow the same
3104 pattern as the I/O depths, meaning that 2=1.6% means that 1.6% of the
3105 I/O completed within 2 msecs, 20=12.8% means that 12.8% of the I/O took
3106 more than 10 msecs, but less than (or equal to) 20 msecs.
3108 After each client has been listed, the group statistics are printed. They
3109 will look like this::
3111 Run status group 0 (all jobs):
3112 READ: io=64MB, aggrb=22178, minb=11355, maxb=11814, mint=2840msec, maxt=2955msec
3113 WRITE: io=64MB, aggrb=1302, minb=666, maxb=669, mint=50093msec, maxt=50320msec
3115 For each data direction, it prints:
3118 Number of megabytes I/O performed.
3120 Aggregate bandwidth of threads in this group.
3122 The minimum average bandwidth a thread saw.
3124 The maximum average bandwidth a thread saw.
3126 The smallest runtime of the threads in that group.
3128 The longest runtime of the threads in that group.
3130 And finally, the disk statistics are printed. They will look like this::
3132 Disk stats (read/write):
3133 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
3135 Each value is printed for both reads and writes, with reads first. The
3139 Number of I/Os performed by all groups.
3141 Number of merges I/O the I/O scheduler.
3143 Number of ticks we kept the disk busy.
3145 Total time spent in the disk queue.
3147 The disk utilization. A value of 100% means we kept the disk
3148 busy constantly, 50% would be a disk idling half of the time.
3150 It is also possible to get fio to dump the current output while it is running,
3151 without terminating the job. To do that, send fio the **USR1** signal. You can
3152 also get regularly timed dumps by using the :option:`--status-interval`
3153 parameter, or by creating a file in :file:`/tmp` named
3154 :file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the
3155 current output status.
3161 For scripted usage where you typically want to generate tables or graphs of the
3162 results, fio can output the results in a semicolon separated format. The format
3163 is one long line of values, such as::
3165 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%
3166 A description of this job goes here.
3168 The job description (if provided) follows on a second line.
3170 To enable terse output, use the :option:`--minimal` command line option. The
3171 first value is the version of the terse output format. If the output has to be
3172 changed for some reason, this number will be incremented by 1 to signify that
3175 Split up, the format is as follows (comments in brackets denote when a
3176 field was introduced or whether its specific to some terse version):
3180 terse version, fio version [v3], jobname, groupid, error
3184 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3185 Submission latency: min, max, mean, stdev (usec)
3186 Completion latency: min, max, mean, stdev (usec)
3187 Completion latency percentiles: 20 fields (see below)
3188 Total latency: min, max, mean, stdev (usec)
3189 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3190 IOPS [v5]: min, max, mean, stdev, number of samples
3196 Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec)
3197 Submission latency: min, max, mean, stdev (usec)
3198 Completion latency: min, max, mean, stdev (usec)
3199 Completion latency percentiles: 20 fields (see below)
3200 Total latency: min, max, mean, stdev (usec)
3201 Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5]
3202 IOPS [v5]: min, max, mean, stdev, number of samples
3204 TRIM status [all but version 3]:
3206 Fields are similar to READ/WRITE status.
3210 user, system, context switches, major faults, minor faults
3214 <=1, 2, 4, 8, 16, 32, >=64
3216 I/O latencies microseconds::
3218 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
3220 I/O latencies milliseconds::
3222 <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
3224 Disk utilization [v3]::
3226 Disk name, Read ios, write ios,
3227 Read merges, write merges,
3228 Read ticks, write ticks,
3229 Time spent in queue, disk utilization percentage
3231 Additional Info (dependent on continue_on_error, default off)::
3233 total # errors, first error code
3235 Additional Info (dependent on description being set)::
3239 Completion latency percentiles can be a grouping of up to 20 sets, so for the
3240 terse output fio writes all of them. Each field will look like this::
3244 which is the Xth percentile, and the `usec` latency associated with it.
3246 For disk utilization, all disks used by fio are shown. So for each disk there
3247 will be a disk utilization section.
3249 Below is a single line containing short names for each of the fields in the
3250 minimal output v3, separated by semicolons:
3252 terse_version_3;fio_version;jobname;groupid;error;read_kb;read_bandwidth;read_iops;read_runtime_ms;read_slat_min;read_slat_max;read_slat_mean;read_slat_dev;read_clat_max;read_clat_min;read_clat_mean;read_clat_dev;read_clat_pct01;read_clat_pct02;read_clat_pct03;read_clat_pct04;read_clat_pct05;read_clat_pct06;read_clat_pct07;read_clat_pct08;read_clat_pct09;read_clat_pct10;read_clat_pct11;read_clat_pct12;read_clat_pct13;read_clat_pct14;read_clat_pct15;read_clat_pct16;read_clat_pct17;read_clat_pct18;read_clat_pct19;read_clat_pct20;read_tlat_min;read_lat_max;read_lat_mean;read_lat_dev;read_bw_min;read_bw_max;read_bw_agg_pct;read_bw_mean;read_bw_dev;write_kb;write_bandwidth;write_iops;write_runtime_ms;write_slat_min;write_slat_max;write_slat_mean;write_slat_dev;write_clat_max;write_clat_min;write_clat_mean;write_clat_dev;write_clat_pct01;write_clat_pct02;write_clat_pct03;write_clat_pct04;write_clat_pct05;write_clat_pct06;write_clat_pct07;write_clat_pct08;write_clat_pct09;write_clat_pct10;write_clat_pct11;write_clat_pct12;write_clat_pct13;write_clat_pct14;write_clat_pct15;write_clat_pct16;write_clat_pct17;write_clat_pct18;write_clat_pct19;write_clat_pct20;write_tlat_min;write_lat_max;write_lat_mean;write_lat_dev;write_bw_min;write_bw_max;write_bw_agg_pct;write_bw_mean;write_bw_dev;cpu_user;cpu_sys;cpu_csw;cpu_mjf;pu_minf;iodepth_1;iodepth_2;iodepth_4;iodepth_8;iodepth_16;iodepth_32;iodepth_64;lat_2us;lat_4us;lat_10us;lat_20us;lat_50us;lat_100us;lat_250us;lat_500us;lat_750us;lat_1000us;lat_2ms;lat_4ms;lat_10ms;lat_20ms;lat_50ms;lat_100ms;lat_250ms;lat_500ms;lat_750ms;lat_1000ms;lat_2000ms;lat_over_2000ms;disk_name;disk_read_iops;disk_write_iops;disk_read_merges;disk_write_merges;disk_read_ticks;write_ticks;disk_queue_time;disk_util
3258 There are two trace file format that you can encounter. The older (v1) format is
3259 unsupported since version 1.20-rc3 (March 2008). It will still be described
3260 below in case that you get an old trace and want to understand it.
3262 In any case the trace is a simple text file with a single action per line.
3265 Trace file format v1
3266 ~~~~~~~~~~~~~~~~~~~~
3268 Each line represents a single I/O action in the following format::
3272 where `rw=0/1` for read/write, and the offset and length entries being in bytes.
3274 This format is not supported in fio versions => 1.20-rc3.
3277 Trace file format v2
3278 ~~~~~~~~~~~~~~~~~~~~
3280 The second version of the trace file format was added in fio version 1.17. It
3281 allows to access more then one file per trace and has a bigger set of possible
3284 The first line of the trace file has to be::
3288 Following this can be lines in two different formats, which are described below.
3290 The file management format::
3294 The filename is given as an absolute path. The action can be one of these:
3297 Add the given filename to the trace.
3299 Open the file with the given filename. The filename has to have
3300 been added with the **add** action before.
3302 Close the file with the given filename. The file has to have been
3306 The file I/O action format::
3308 filename action offset length
3310 The `filename` is given as an absolute path, and has to have been added and
3311 opened before it can be used with this format. The `offset` and `length` are
3312 given in bytes. The `action` can be one of these:
3315 Wait for `offset` microseconds. Everything below 100 is discarded.
3316 The time is relative to the previous `wait` statement.
3318 Read `length` bytes beginning from `offset`.
3320 Write `length` bytes beginning from `offset`.
3322 :manpage:`fsync(2)` the file.
3324 :manpage:`fdatasync(2)` the file.
3326 Trim the given file from the given `offset` for `length` bytes.
3328 CPU idleness profiling
3329 ----------------------
3331 In some cases, we want to understand CPU overhead in a test. For example, we
3332 test patches for the specific goodness of whether they reduce CPU usage.
3333 Fio implements a balloon approach to create a thread per CPU that runs at idle
3334 priority, meaning that it only runs when nobody else needs the cpu.
3335 By measuring the amount of work completed by the thread, idleness of each CPU
3336 can be derived accordingly.
3338 An unit work is defined as touching a full page of unsigned characters. Mean and
3339 standard deviation of time to complete an unit work is reported in "unit work"
3340 section. Options can be chosen to report detailed percpu idleness or overall
3341 system idleness by aggregating percpu stats.
3344 Verification and triggers
3345 -------------------------
3347 Fio is usually run in one of two ways, when data verification is done. The first
3348 is a normal write job of some sort with verify enabled. When the write phase has
3349 completed, fio switches to reads and verifies everything it wrote. The second
3350 model is running just the write phase, and then later on running the same job
3351 (but with reads instead of writes) to repeat the same I/O patterns and verify
3352 the contents. Both of these methods depend on the write phase being completed,
3353 as fio otherwise has no idea how much data was written.
3355 With verification triggers, fio supports dumping the current write state to
3356 local files. Then a subsequent read verify workload can load this state and know
3357 exactly where to stop. This is useful for testing cases where power is cut to a
3358 server in a managed fashion, for instance.
3360 A verification trigger consists of two things:
3362 1) Storing the write state of each job.
3363 2) Executing a trigger command.
3365 The write state is relatively small, on the order of hundreds of bytes to single
3366 kilobytes. It contains information on the number of completions done, the last X
3369 A trigger is invoked either through creation ('touch') of a specified file in
3370 the system, or through a timeout setting. If fio is run with
3371 :option:`--trigger-file` = :file:`/tmp/trigger-file`, then it will continually
3372 check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it
3373 will fire off the trigger (thus saving state, and executing the trigger
3376 For client/server runs, there's both a local and remote trigger. If fio is
3377 running as a server backend, it will send the job states back to the client for
3378 safe storage, then execute the remote trigger, if specified. If a local trigger
3379 is specified, the server will still send back the write state, but the client
3380 will then execute the trigger.
3382 Verification trigger example
3383 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3385 Lets say we want to run a powercut test on the remote machine 'server'. Our
3386 write workload is in :file:`write-test.fio`. We want to cut power to 'server' at
3387 some point during the run, and we'll run this test from the safety or our local
3388 machine, 'localbox'. On the server, we'll start the fio backend normally::
3390 server# fio --server
3392 and on the client, we'll fire off the workload::
3394 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\""
3396 We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute::
3398 echo b > /proc/sysrq-trigger
3400 on the server once it has received the trigger and sent us the write state. This
3401 will work, but it's not **really** cutting power to the server, it's merely
3402 abruptly rebooting it. If we have a remote way of cutting power to the server
3403 through IPMI or similar, we could do that through a local trigger command
3404 instead. Lets assume we have a script that does IPMI reboot of a given hostname,
3405 ipmi-reboot. On localbox, we could then have run fio with a local trigger
3408 localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server"
3410 For this case, fio would wait for the server to send us the write state, then
3411 execute ``ipmi-reboot server`` when that happened.
3413 Loading verify state
3414 ~~~~~~~~~~~~~~~~~~~~
3416 To load store write state, read verification job file must contain the
3417 :option:`verify_state_load` option. If that is set, fio will load the previously
3418 stored state. For a local fio run this is done by loading the files directly,
3419 and on a client/server run, the server backend will ask the client to send the
3420 files over and load them from there.
3426 Fio supports a variety of log file formats, for logging latencies, bandwidth,
3427 and IOPS. The logs share a common format, which looks like this:
3429 *time* (`msec`), *value*, *data direction*, *offset*
3431 Time for the log entry is always in milliseconds. The *value* logged depends
3432 on the type of log, it will be one of the following:
3435 Value is latency in usecs
3441 *Data direction* is one of the following:
3450 The *offset* is the offset, in bytes, from the start of the file, for that
3451 particular I/O. The logging of the offset can be toggled with
3452 :option:`log_offset`.
3454 If windowed logging is enabled through :option:`log_avg_msec` then fio doesn't
3455 log individual I/Os. Instead of logs the average values over the specified period
3456 of time. Since 'data direction' and 'offset' are per-I/O values, they aren't
3457 applicable if windowed logging is enabled. If windowed logging is enabled and
3458 :option:`log_max_value` is set, then fio logs maximum values in that window
3459 instead of averages.
3465 Normally fio is invoked as a stand-alone application on the machine where the
3466 I/O workload should be generated. However, the frontend and backend of fio can
3467 be run separately. Ie the fio server can generate an I/O workload on the "Device
3468 Under Test" while being controlled from another machine.
3470 Start the server on the machine which has access to the storage DUT::
3474 where args defines what fio listens to. The arguments are of the form
3475 ``type,hostname`` or ``IP,port``. *type* is either ``ip`` (or ip4) for TCP/IP
3476 v4, ``ip6`` for TCP/IP v6, or ``sock`` for a local unix domain socket.
3477 *hostname* is either a hostname or IP address, and *port* is the port to listen
3478 to (only valid for TCP/IP, not a local socket). Some examples:
3482 Start a fio server, listening on all interfaces on the default port (8765).
3484 2) ``fio --server=ip:hostname,4444``
3486 Start a fio server, listening on IP belonging to hostname and on port 4444.
3488 3) ``fio --server=ip6:::1,4444``
3490 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
3492 4) ``fio --server=,4444``
3494 Start a fio server, listening on all interfaces on port 4444.
3496 5) ``fio --server=1.2.3.4``
3498 Start a fio server, listening on IP 1.2.3.4 on the default port.
3500 6) ``fio --server=sock:/tmp/fio.sock``
3502 Start a fio server, listening on the local socket /tmp/fio.sock.
3504 Once a server is running, a "client" can connect to the fio server with::
3506 fio <local-args> --client=<server> <remote-args> <job file(s)>
3508 where `local-args` are arguments for the client where it is running, `server`
3509 is the connect string, and `remote-args` and `job file(s)` are sent to the
3510 server. The `server` string follows the same format as it does on the server
3511 side, to allow IP/hostname/socket and port strings.
3513 Fio can connect to multiple servers this way::
3515 fio --client=<server1> <job file(s)> --client=<server2> <job file(s)>
3517 If the job file is located on the fio server, then you can tell the server to
3518 load a local file as well. This is done by using :option:`--remote-config` ::
3520 fio --client=server --remote-config /path/to/file.fio
3522 Then fio will open this local (to the server) job file instead of being passed
3523 one from the client.
3525 If you have many servers (example: 100 VMs/containers), you can input a pathname
3526 of a file containing host IPs/names as the parameter value for the
3527 :option:`--client` option. For example, here is an example :file:`host.list`
3528 file containing 2 hostnames::
3530 host1.your.dns.domain
3531 host2.your.dns.domain
3533 The fio command would then be::
3535 fio --client=host.list <job file(s)>
3537 In this mode, you cannot input server-specific parameters or job files -- all
3538 servers receive the same job file.
3540 In order to let ``fio --client`` runs use a shared filesystem from multiple
3541 hosts, ``fio --client`` now prepends the IP address of the server to the
3542 filename. For example, if fio is using directory :file:`/mnt/nfs/fio` and is
3543 writing filename :file:`fileio.tmp`, with a :option:`--client` `hostfile`
3544 containing two hostnames ``h1`` and ``h2`` with IP addresses 192.168.10.120 and
3545 192.168.10.121, then fio will create two files::
3547 /mnt/nfs/fio/192.168.10.120.fileio.tmp
3548 /mnt/nfs/fio/192.168.10.121.fileio.tmp