1 .TH fio 1 "September 2007" "User Manual"
3 fio \- flexible I/O tester
6 [\fIoptions\fR] [\fIjobfile\fR]...
9 is a tool that will spawn a number of threads or processes doing a
10 particular type of I/O action as specified by the user.
11 The typical use of fio is to write a job file matching the I/O load
12 one wants to simulate.
15 .BI \-\-debug \fR=\fPtype
16 Enable verbose tracing of various fio actions. May be `all' for all types
17 or individual types separated by a comma (eg \-\-debug=io,file). `help' will
18 list all available tracing options.
20 .BI \-\-output \fR=\fPfilename
21 Write output to \fIfilename\fR.
23 .BI \-\-runtime \fR=\fPruntime
24 Limit run time to \fIruntime\fR seconds.
27 Generate per-job latency logs.
30 Generate per-job bandwidth logs.
33 Print statistics in a terse, semicolon-delimited format.
36 Display version information and exit.
38 .BI \-\-terse\-version \fR=\fPversion
39 Set terse version output format (Current version 3, or older version 2).
42 Display usage information and exit.
44 .BI \-\-cmdhelp \fR=\fPcommand
45 Print help information for \fIcommand\fR. May be `all' for all commands.
47 .BI \-\-enghelp \fR=\fPioengine[,command]
48 List all commands defined by \fIioengine\fR, or print help for \fIcommand\fR defined by \fIioengine\fR.
50 .BI \-\-showcmd \fR=\fPjobfile
51 Convert \fIjobfile\fR to a set of command-line options.
53 .BI \-\-eta \fR=\fPwhen
54 Specifies when real-time ETA estimate should be printed. \fIwhen\fR may
55 be one of `always', `never' or `auto'.
58 Turn on safety read-only checks, preventing any attempted write.
60 .BI \-\-section \fR=\fPsec
61 Only run section \fIsec\fR from job file. Multiple of these options can be given, adding more sections to run.
63 .BI \-\-alloc\-size \fR=\fPkb
64 Set the internal smalloc pool size to \fIkb\fP kilobytes.
66 .BI \-\-warnings\-fatal
67 All fio parser warnings are fatal, causing fio to exit with an error.
69 .BI \-\-max\-jobs \fR=\fPnr
70 Set the maximum allowed number of jobs (threads/processes) to support.
72 .BI \-\-server \fR=\fPargs
73 Start a backend server, with \fIargs\fP specifying what to listen to. See client/server section.
75 .BI \-\-daemonize \fR=\fPpidfile
76 Background a fio server, writing the pid to the given pid file.
78 .BI \-\-client \fR=\fPhost
79 Instead of running the jobs locally, send and run them on the given host.
81 .BI \-\-idle\-prof \fR=\fPoption
82 Report cpu idleness on a system or percpu basis (\fIoption\fP=system,percpu) or run unit work calibration only (\fIoption\fP=calibrate).
84 Job files are in `ini' format. They consist of one or more
85 job definitions, which begin with a job name in square brackets and
86 extend to the next job name. The job name can be any ASCII string
87 except `global', which has a special meaning. Following the job name is
88 a sequence of zero or more parameters, one per line, that define the
89 behavior of the job. Any line starting with a `;' or `#' character is
90 considered a comment and ignored.
92 If \fIjobfile\fR is specified as `-', the job file will be read from
95 The global section contains default parameters for jobs specified in the
96 job file. A job is only affected by global sections residing above it,
97 and there may be any number of global sections. Specific job definitions
98 may override any parameter set in global sections.
101 Some parameters may take arguments of a specific type. The types used are:
104 String: a sequence of alphanumeric characters.
107 SI integer: a whole number, possibly containing a suffix denoting the base unit
108 of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting
109 kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5)
110 respectively. The suffix is not case sensitive. If prefixed with '0x', the
111 value is assumed to be base 16 (hexadecimal). A suffix may include a trailing 'b',
112 for instance 'kb' is identical to 'k'. You can specify a base 10 value
113 by using 'KiB', 'MiB', 'GiB', etc. This is useful for disk drives where
114 values are often given in base 10 values. Specifying '30GiB' will get you
118 Boolean: a true or false value. `0' denotes false, `1' denotes true.
121 Integer range: a range of integers specified in the format
122 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
123 \fIupper\fR may contain a suffix as described above. If an option allows two
124 sets of ranges, they are separated with a `,' or `/' character. For example:
128 List of floating numbers: A list of floating numbers, separated by
133 May be used to override the job name. On the command line, this parameter
134 has the special purpose of signalling the start of a new job.
136 .BI description \fR=\fPstr
137 Human-readable description of the job. It is printed when the job is run, but
138 otherwise has no special purpose.
140 .BI directory \fR=\fPstr
141 Prefix filenames with this directory. Used to place files in a location other
144 .BI filename \fR=\fPstr
146 normally makes up a file name based on the job name, thread number, and file
147 number. If you want to share files between threads in a job or several jobs,
148 specify a \fIfilename\fR for each of them to override the default.
149 If the I/O engine is file-based, you can specify
150 a number of files by separating the names with a `:' character. `\-' is a
151 reserved name, meaning stdin or stdout, depending on the read/write direction
154 .BI filename_format \fR=\fPstr
155 .B If sharing multiple files between jobs, it is usually necessary to have
156 fio generate the exact names that you want. By default, fio will name a file
157 based on the default file format specification of
158 \fBjobname.jobnumber.filenumber\fP. With this option, that can be
159 customized. Fio will recognize and replace the following keywords in this
165 The name of the worker thread or process.
168 The incremental number of the worker thread or process.
171 The incremental number of the file for that worker thread or process.
174 To have dependent jobs share a set of files, this option can be set to
175 have fio generate filenames that are shared between the two. For instance,
176 if \fBtestfiles.$filenum\fR is specified, file number 4 for any job will
177 be named \fBtestfiles.4\fR. The default of \fB$jobname.$jobnum.$filenum\fR
178 will be used if no other format specifier is given.
182 .BI lockfile \fR=\fPstr
183 Fio defaults to not locking any files before it does IO to them. If a file or
184 file descriptor is shared, fio can serialize IO to that file to make the end
185 result consistent. This is usual for emulating real workloads that share files.
191 No locking. This is the default.
194 Only one thread or process may do IO at the time, excluding all others.
197 Read-write locking on the file. Many readers may access the file at the same
198 time, but writes get exclusive access.
201 .BI opendir \fR=\fPstr
202 Recursively open any files below directory \fIstr\fR.
204 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
205 Type of I/O pattern. Accepted values are:
222 Mixed sequential reads and writes.
225 Mixed random reads and writes.
228 For mixed I/O, the default split is 50/50. For certain types of io the result
229 may still be skewed a bit, since the speed may be different. It is possible to
230 specify a number of IO's to do before getting a new offset, this is done by
231 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
232 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
233 value of 8. If the postfix is used with a sequential IO pattern, then the value
234 specified will be added to the generated offset for each IO. For instance,
235 using \fBrw=write:4k\fR will skip 4k for every write. It turns sequential IO
236 into sequential IO with holes. See the \fBrw_sequencer\fR option.
239 .BI rw_sequencer \fR=\fPstr
240 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
241 then this option controls how that number modifies the IO offset being
242 generated. Accepted values are:
247 Generate sequential offset
250 Generate the same offset
253 \fBsequential\fR is only useful for random IO, where fio would normally
254 generate a new random offset for every IO. If you append eg 8 to randread, you
255 would get a new random offset for every 8 IO's. The result would be a seek for
256 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
257 that. As sequential IO is already sequential, setting \fBsequential\fR for that
258 would not result in any differences. \fBidentical\fR behaves in a similar
259 fashion, except it sends the same offset 8 number of times before generating a
264 .BI kb_base \fR=\fPint
265 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
266 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
267 reasons. Allow values are 1024 or 1000, with 1024 being the default.
269 .BI unified_rw_reporting \fR=\fPbool
270 Fio normally reports statistics on a per data direction basis, meaning that
271 read, write, and trim are accounted and reported separately. If this option is
272 set, the fio will sum the results and report them as "mixed" instead.
274 .BI randrepeat \fR=\fPbool
275 Seed the random number generator in a predictable way so results are repeatable
276 across runs. Default: true.
278 .BI use_os_rand \fR=\fPbool
279 Fio can either use the random generator supplied by the OS to generator random
280 offsets, or it can use it's own internal generator (based on Tausworthe).
281 Default is to use the internal generator, which is often of better quality and
282 faster. Default: false.
284 .BI fallocate \fR=\fPstr
285 Whether pre-allocation is performed when laying down files. Accepted values
291 Do not pre-allocate space.
294 Pre-allocate via posix_fallocate().
297 Pre-allocate via fallocate() with FALLOC_FL_KEEP_SIZE set.
300 Backward-compatible alias for 'none'.
303 Backward-compatible alias for 'posix'.
306 May not be available on all supported platforms. 'keep' is only
307 available on Linux. If using ZFS on Solaris this must be set to 'none'
308 because ZFS doesn't support it. Default: 'posix'.
311 .BI fadvise_hint \fR=\fPbool
312 Use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
313 are likely to be issued. Default: true.
316 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
317 been transfered, unless limited by other options (\fBruntime\fR, for instance).
318 Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be
319 divided between the available files for the job. If not set, fio will use the
320 full size of the given files or devices. If the the files do not exist, size
321 must be given. It is also possible to give size as a percentage between 1 and
322 100. If size=20% is given, fio will use 20% of the full size of the given files
325 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
326 Sets size to something really large and waits for ENOSPC (no space left on
327 device) as the terminating condition. Only makes sense with sequential write.
328 For a read workload, the mount point will be filled first then IO started on
329 the result. This option doesn't make sense if operating on a raw device node,
330 since the size of that is already known by the file system. Additionally,
331 writing beyond end-of-device will not return ENOSPC there.
333 .BI filesize \fR=\fPirange
334 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
335 for files at random within the given range, limited to \fBsize\fR in total (if
336 that is given). If \fBfilesize\fR is not specified, each created file is the
339 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
340 Block size for I/O units. Default: 4k. Values for reads and writes can be
341 specified separately in the format \fIread\fR,\fIwrite\fR, either of
342 which may be empty to leave that value at its default.
344 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
345 Specify a range of I/O block sizes. The issued I/O unit will always be a
346 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
347 to both reads and writes if only one range is given, but can be specified
348 separately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k.
349 Also (see \fBblocksize\fR).
351 .BI bssplit \fR=\fPstr
352 This option allows even finer grained control of the block sizes issued,
353 not just even splits between them. With this option, you can weight various
354 block sizes for exact control of the issued IO for a job that has mixed
355 block sizes. The format of the option is bssplit=blocksize/percentage,
356 optionally adding as many definitions as needed separated by a colon.
357 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
358 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
359 splits to reads and writes. The format is identical to what the
360 \fBbs\fR option accepts, the read and write parts are separated with a
363 .B blocksize_unaligned\fR,\fP bs_unaligned
364 If set, any size in \fBblocksize_range\fR may be used. This typically won't
365 work with direct I/O, as that normally requires sector alignment.
367 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
368 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
369 the minimum blocksize given. Minimum alignment is typically 512b
370 for using direct IO, though it usually depends on the hardware block size.
371 This option is mutually exclusive with using a random map for files, so it
372 will turn off that option.
375 Initialise buffers with all zeros. Default: fill buffers with random data.
378 If this option is given, fio will refill the IO buffers on every submit. The
379 default is to only fill it at init time and reuse that data. Only makes sense
380 if zero_buffers isn't specified, naturally. If data verification is enabled,
381 refill_buffers is also automatically enabled.
383 .BI scramble_buffers \fR=\fPbool
384 If \fBrefill_buffers\fR is too costly and the target is using data
385 deduplication, then setting this option will slightly modify the IO buffer
386 contents to defeat normal de-dupe attempts. This is not enough to defeat
387 more clever block compression attempts, but it will stop naive dedupe
388 of blocks. Default: true.
390 .BI buffer_compress_percentage \fR=\fPint
391 If this is set, then fio will attempt to provide IO buffer content (on WRITEs)
392 that compress to the specified level. Fio does this by providing a mix of
393 random data and zeroes. Note that this is per block size unit, for file/disk
394 wide compression level that matches this setting, you'll also want to set
395 \fBrefill_buffers\fR.
397 .BI buffer_compress_chunk \fR=\fPint
398 See \fBbuffer_compress_percentage\fR. This setting allows fio to manage how
399 big the ranges of random data and zeroed data is. Without this set, fio will
400 provide \fBbuffer_compress_percentage\fR of blocksize random data, followed by
401 the remaining zeroed. With this set to some chunk size smaller than the block
402 size, fio can alternate random and zeroed data throughout the IO buffer.
404 .BI nrfiles \fR=\fPint
405 Number of files to use for this job. Default: 1.
407 .BI openfiles \fR=\fPint
408 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
410 .BI file_service_type \fR=\fPstr
411 Defines how files to service are selected. The following types are defined:
416 Choose a file at random
419 Round robin over open files (default).
421 Do each file in the set sequentially.
424 The number of I/Os to issue before switching a new file can be specified by
425 appending `:\fIint\fR' to the service type.
428 .BI ioengine \fR=\fPstr
429 Defines how the job issues I/O. The following types are defined:
434 Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
435 position the I/O location.
438 Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O.
441 Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by
442 coalescing adjacents IOs into a single submission.
445 Linux native asynchronous I/O. This ioengine defines engine specific options.
448 POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
451 Solaris native asynchronous I/O.
454 Windows native asynchronous I/O.
457 File is memory mapped with \fImmap\fR\|(2) and data copied using
461 \fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
462 transfer data from user-space to the kernel.
465 Use the syslet system calls to make regular read/write asynchronous.
468 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
469 the target is an sg character device, we use \fIread\fR\|(2) and
470 \fIwrite\fR\|(2) for asynchronous I/O.
473 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
474 itself and for debugging and testing purposes.
477 Transfer over the network. The protocol to be used can be defined with the
478 \fBprotocol\fR parameter. Depending on the protocol, \fBfilename\fR,
479 \fBhostname\fR, \fBport\fR, or \fBlisten\fR must be specified.
480 This ioengine defines engine specific options.
483 Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
484 and send/receive. This ioengine defines engine specific options.
487 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
488 \fBcpucycles\fR parameters.
491 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
492 approach to asycnronous I/O.
494 See <http://www.xmailserver.org/guasi\-lib.html>.
497 The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ)
498 and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols.
501 Loads an external I/O engine object file. Append the engine filename as
505 IO engine that does regular linux native fallocate callt to simulate data
506 transfer as fio ioengine
508 DDIR_READ does fallocate(,mode = FALLOC_FL_KEEP_SIZE,)
510 DIR_WRITE does fallocate(,mode = 0)
512 DDIR_TRIM does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE)
515 IO engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate defragment activity
516 request to DDIR_WRITE event
521 .BI iodepth \fR=\fPint
522 Number of I/O units to keep in flight against the file. Note that increasing
523 iodepth beyond 1 will not affect synchronous ioengines (except for small
524 degress when verify_async is in use). Even async engines my impose OS
525 restrictions causing the desired depth not to be achieved. This may happen on
526 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
527 not async on that OS. Keep an eye on the IO depth distribution in the
528 fio output to verify that the achieved depth is as expected. Default: 1.
530 .BI iodepth_batch \fR=\fPint
531 Number of I/Os to submit at once. Default: \fBiodepth\fR.
533 .BI iodepth_batch_complete \fR=\fPint
534 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
535 means that we'll ask for a minimum of 1 IO in the retrieval process from the
536 kernel. The IO retrieval will go on until we hit the limit set by
537 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
538 completed events before queuing more IO. This helps reduce IO latency, at the
539 cost of more retrieval system calls.
541 .BI iodepth_low \fR=\fPint
542 Low watermark indicating when to start filling the queue again. Default:
545 .BI direct \fR=\fPbool
546 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
548 .BI buffered \fR=\fPbool
549 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
552 .BI offset \fR=\fPint
553 Offset in the file to start I/O. Data before the offset will not be touched.
555 .BI offset_increment \fR=\fPint
556 If this is provided, then the real offset becomes the
557 offset + offset_increment * thread_number, where the thread number is a counter
558 that starts at 0 and is incremented for each job. This option is useful if
559 there are several jobs which are intended to operate on a file in parallel in
560 disjoint segments, with even spacing between the starting points.
563 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
564 0, don't sync. Default: 0.
566 .BI fdatasync \fR=\fPint
567 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
568 data parts of the file. Default: 0.
570 .BI sync_file_range \fR=\fPstr:int
571 Use sync_file_range() for every \fRval\fP number of write operations. Fio will
572 track range of writes that have happened since the last sync_file_range() call.
573 \fRstr\fP can currently be one or more of:
577 SYNC_FILE_RANGE_WAIT_BEFORE
580 SYNC_FILE_RANGE_WRITE
583 SYNC_FILE_RANGE_WRITE
587 So if you do sync_file_range=wait_before,write:8, fio would use
588 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
589 Also see the sync_file_range(2) man page. This option is Linux specific.
591 .BI overwrite \fR=\fPbool
592 If writing, setup the file first and do overwrites. Default: false.
594 .BI end_fsync \fR=\fPbool
595 Sync file contents when a write stage has completed. Default: false.
597 .BI fsync_on_close \fR=\fPbool
598 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
599 it will happen on every close, not just at the end of the job. Default: false.
601 .BI rwmixread \fR=\fPint
602 Percentage of a mixed workload that should be reads. Default: 50.
604 .BI rwmixwrite \fR=\fPint
605 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
606 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
607 overrides the first. This may interfere with a given rate setting, if fio is
608 asked to limit reads or writes to a certain rate. If that is the case, then
609 the distribution may be skewed. Default: 50.
611 .BI random_distribution \fR=\fPstr:float
612 By default, fio will use a completely uniform random distribution when asked
613 to perform random IO. Sometimes it is useful to skew the distribution in
614 specific ways, ensuring that some parts of the data is more hot than others.
615 Fio includes the following distribution models:
619 Uniform random distribution
629 When using a zipf or pareto distribution, an input value is also needed to
630 define the access pattern. For zipf, this is the zipf theta. For pareto,
631 it's the pareto power. Fio includes a test program, genzipf, that can be
632 used visualize what the given input values will yield in terms of hit rates.
633 If you wanted to use zipf with a theta of 1.2, you would use
634 random_distribution=zipf:1.2 as the option. If a non-uniform model is used,
635 fio will disable use of the random map.
638 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
639 this parameter is given, a new offset will be chosen without looking at past
640 I/O history. This parameter is mutually exclusive with \fBverify\fR.
642 .BI softrandommap \fR=\fPbool
643 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
644 fails to allocate the map, if this option is set it will continue without a
645 random block map. As coverage will not be as complete as with random maps, this
646 option is disabled by default.
648 .BI random_generator \fR=\fPstr
649 Fio supports the following engines for generating IO offsets for random IO:
653 Strong 2^88 cycle random number generator
656 Linear feedback shift register generator
660 Tausworthe is a strong random number generator, but it requires tracking on the
661 side if we want to ensure that blocks are only read or written once. LFSR
662 guarantees that we never generate the same offset twice, and it's also less
663 computationally expensive. It's not a true random generator, however, though
664 for IO purposes it's typically good enough. LFSR only works with single block
665 sizes, not with workloads that use multiple block sizes. If used with such a
666 workload, fio may read or write some blocks multiple times.
669 Run job with given nice value. See \fInice\fR\|(2).
672 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
675 .BI prioclass \fR=\fPint
676 Set I/O priority class. See \fIionice\fR\|(1).
678 .BI thinktime \fR=\fPint
679 Stall job for given number of microseconds between issuing I/Os.
681 .BI thinktime_spin \fR=\fPint
682 Pretend to spend CPU time for given number of microseconds, sleeping the rest
683 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
685 .BI thinktime_blocks \fR=\fPint
686 Number of blocks to issue before waiting \fBthinktime\fR microseconds.
690 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
691 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
692 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
693 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
694 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
695 limit writes (to 500KB/sec), the latter will only limit reads.
697 .BI ratemin \fR=\fPint
698 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
699 Failing to meet this requirement will cause the job to exit. The same format
700 as \fBrate\fR is used for read vs write separation.
702 .BI rate_iops \fR=\fPint
703 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
704 specified independently of bandwidth. The same format as \fBrate\fR is used for
705 read vs write seperation. If \fBblocksize\fR is a range, the smallest block
706 size is used as the metric.
708 .BI rate_iops_min \fR=\fPint
709 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
710 is used for read vs write seperation.
712 .BI ratecycle \fR=\fPint
713 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
714 milliseconds. Default: 1000ms.
716 .BI max_latency \fR=\fPint
717 If set, fio will exit the job if it exceeds this maximum latency. It will exit
720 .BI cpumask \fR=\fPint
721 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
722 may run on. See \fBsched_setaffinity\fR\|(2).
724 .BI cpus_allowed \fR=\fPstr
725 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
727 .BI numa_cpu_nodes \fR=\fPstr
728 Set this job running on spcified NUMA nodes' CPUs. The arguments allow
729 comma delimited list of cpu numbers, A-B ranges, or 'all'.
731 .BI numa_mem_policy \fR=\fPstr
732 Set this job's memory policy and corresponding NUMA nodes. Format of
736 .B <mode>[:<nodelist>]
739 is one of the following memory policy:
741 .B default, prefer, bind, interleave, local
744 For \fBdefault\fR and \fBlocal\fR memory policy, no \fBnodelist\fR is
745 needed to be specified. For \fBprefer\fR, only one node is
746 allowed. For \fBbind\fR and \fBinterleave\fR, \fBnodelist\fR allows
747 comma delimited list of numbers, A-B ranges, or 'all'.
749 .BI startdelay \fR=\fPint
750 Delay start of job for the specified number of seconds.
752 .BI runtime \fR=\fPint
753 Terminate processing after the specified number of seconds.
756 If given, run for the specified \fBruntime\fR duration even if the files are
757 completely read or written. The same workload will be repeated as many times
758 as \fBruntime\fR allows.
760 .BI ramp_time \fR=\fPint
761 If set, fio will run the specified workload for this amount of time before
762 logging any performance numbers. Useful for letting performance settle before
763 logging results, thus minimizing the runtime required for stable results. Note
764 that the \fBramp_time\fR is considered lead in time for a job, thus it will
765 increase the total runtime if a special timeout or runtime is specified.
767 .BI invalidate \fR=\fPbool
768 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
771 Use synchronous I/O for buffered writes. For the majority of I/O engines,
772 this means using O_SYNC. Default: false.
774 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
775 Allocation method for I/O unit buffer. Allowed values are:
780 Allocate memory with \fImalloc\fR\|(3).
783 Use shared memory buffers allocated through \fIshmget\fR\|(2).
786 Same as \fBshm\fR, but use huge pages as backing.
789 Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
790 is given after the option in the format `:\fIfile\fR'.
793 Same as \fBmmap\fR, but use huge files as backing.
796 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
797 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
798 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
799 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
800 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
801 and the documentation for that. Normally you just need to echo an appropriate
802 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
806 .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint
807 This indiciates the memory alignment of the IO memory buffers. Note that the
808 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
809 the alignment of the following buffers are given by the \fBbs\fR used. In
810 other words, if using a \fBbs\fR that is a multiple of the page sized in the
811 system, all buffers will be aligned to this value. If using a \fBbs\fR that
812 is not page aligned, the alignment of subsequent IO memory buffers is the
813 sum of the \fBiomem_align\fR and \fBbs\fR used.
815 .BI hugepage\-size \fR=\fPint
816 Defines the size of a huge page. Must be at least equal to the system setting.
817 Should be a multiple of 1MB. Default: 4MB.
820 Terminate all jobs when one finishes. Default: wait for each job to finish.
822 .BI bwavgtime \fR=\fPint
823 Average bandwidth calculations over the given time in milliseconds. Default:
826 .BI iopsavgtime \fR=\fPint
827 Average IOPS calculations over the given time in milliseconds. Default:
830 .BI create_serialize \fR=\fPbool
831 If true, serialize file creation for the jobs. Default: true.
833 .BI create_fsync \fR=\fPbool
834 \fIfsync\fR\|(2) data file after creation. Default: true.
836 .BI create_on_open \fR=\fPbool
837 If true, the files are not created until they are opened for IO by the job.
839 .BI create_only \fR=\fPbool
840 If true, fio will only run the setup phase of the job. If files need to be
841 laid out or updated on disk, only that will be done. The actual job contents
844 .BI pre_read \fR=\fPbool
845 If this is given, files will be pre-read into memory before starting the given
846 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
847 pointless to pre-read and then drop the cache. This will only work for IO
848 engines that are seekable, since they allow you to read the same data
849 multiple times. Thus it will not work on eg network or splice IO.
851 .BI unlink \fR=\fPbool
852 Unlink job files when done. Default: false.
855 Specifies the number of iterations (runs of the same workload) of this job.
858 .BI do_verify \fR=\fPbool
859 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
862 .BI verify \fR=\fPstr
863 Method of verifying file contents after each iteration of the job. Allowed
868 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1
869 Store appropriate checksum in the header of each block. crc32c-intel is
870 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
871 not supported by the system.
874 Write extra information about each I/O (timestamp, block number, etc.). The
875 block number is verified. See \fBverify_pattern\fR as well.
878 Pretend to verify. Used for testing internals.
881 This option can be used for repeated burn-in tests of a system to make sure
882 that the written data is also correctly read back. If the data direction given
883 is a read or random read, fio will assume that it should verify a previously
884 written file. If the data direction includes any form of write, the verify will
885 be of the newly written data.
888 .BI verify_sort \fR=\fPbool
889 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
890 read them back in a sorted manner. Default: true.
892 .BI verify_offset \fR=\fPint
893 Swap the verification header with data somewhere else in the block before
894 writing. It is swapped back before verifying.
896 .BI verify_interval \fR=\fPint
897 Write the verification header for this number of bytes, which should divide
898 \fBblocksize\fR. Default: \fBblocksize\fR.
900 .BI verify_pattern \fR=\fPstr
901 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
902 with totally random bytes, but sometimes it's interesting to fill with a known
903 pattern for io verification purposes. Depending on the width of the pattern,
904 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
905 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
906 has to be a hex number that starts with either "0x" or "0X". Use with
909 .BI verify_fatal \fR=\fPbool
910 If true, exit the job on the first observed verification failure. Default:
913 .BI verify_dump \fR=\fPbool
914 If set, dump the contents of both the original data block and the data block we
915 read off disk to files. This allows later analysis to inspect just what kind of
916 data corruption occurred. Off by default.
918 .BI verify_async \fR=\fPint
919 Fio will normally verify IO inline from the submitting thread. This option
920 takes an integer describing how many async offload threads to create for IO
921 verification instead, causing fio to offload the duty of verifying IO contents
922 to one or more separate threads. If using this offload option, even sync IO
923 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
924 allows them to have IO in flight while verifies are running.
926 .BI verify_async_cpus \fR=\fPstr
927 Tell fio to set the given CPU affinity on the async IO verification threads.
928 See \fBcpus_allowed\fP for the format used.
930 .BI verify_backlog \fR=\fPint
931 Fio will normally verify the written contents of a job that utilizes verify
932 once that job has completed. In other words, everything is written then
933 everything is read back and verified. You may want to verify continually
934 instead for a variety of reasons. Fio stores the meta data associated with an
935 IO block in memory, so for large verify workloads, quite a bit of memory would
936 be used up holding this meta data. If this option is enabled, fio will write
937 only N blocks before verifying these blocks.
939 .BI verify_backlog_batch \fR=\fPint
940 Control how many blocks fio will verify if verify_backlog is set. If not set,
941 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
942 read back and verified). If \fBverify_backlog_batch\fR is less than
943 \fBverify_backlog\fR then not all blocks will be verified, if
944 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
945 will be verified more than once.
947 .B stonewall "\fR,\fP wait_for_previous"
948 Wait for preceding jobs in the job file to exit before starting this one.
949 \fBstonewall\fR implies \fBnew_group\fR.
952 Start a new reporting group. If not given, all jobs in a file will be part
953 of the same reporting group, unless separated by a stonewall.
955 .BI numjobs \fR=\fPint
956 Number of clones (processes/threads performing the same workload) of this job.
960 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
964 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
965 with \fBfork\fR\|(2).
967 .BI zonesize \fR=\fPint
968 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
970 .BI zoneskip \fR=\fPint
971 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
974 .BI write_iolog \fR=\fPstr
975 Write the issued I/O patterns to the specified file. Specify a separate file
976 for each job, otherwise the iologs will be interspersed and the file may be
979 .BI read_iolog \fR=\fPstr
980 Replay the I/O patterns contained in the specified file generated by
981 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
983 .BI replay_no_stall \fR=\fPint
984 While replaying I/O patterns using \fBread_iolog\fR the default behavior
985 attempts to respect timing information between I/Os. Enabling
986 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
987 still respecting ordering.
989 .BI replay_redirect \fR=\fPstr
990 While replaying I/O patterns using \fBread_iolog\fR the default behavior
991 is to replay the IOPS onto the major/minor device that each IOP was recorded
992 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
993 single specified device regardless of the device it was recorded from.
995 .BI write_bw_log \fR=\fPstr
996 If given, write a bandwidth log of the jobs in this job file. Can be used to
997 store data of the bandwidth of the jobs in their lifetime. The included
998 fio_generate_plots script uses gnuplot to turn these text files into nice
999 graphs. See \fBwrite_log_log\fR for behaviour of given filename. For this
1000 option, the postfix is _bw.log.
1002 .BI write_lat_log \fR=\fPstr
1003 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
1004 filename is given with this option, the default filename of "jobname_type.log"
1005 is used. Even if the filename is given, fio will still append the type of log.
1007 .BI write_iops_log \fR=\fPstr
1008 Same as \fBwrite_bw_log\fR, but writes IOPS. If no filename is given with this
1009 option, the default filename of "jobname_type.log" is used. Even if the
1010 filename is given, fio will still append the type of log.
1012 .BI log_avg_msec \fR=\fPint
1013 By default, fio will log an entry in the iops, latency, or bw log for every
1014 IO that completes. When writing to the disk log, that can quickly grow to a
1015 very large size. Setting this option makes fio average the each log entry
1016 over the specified period of time, reducing the resolution of the log.
1019 .BI disable_lat \fR=\fPbool
1020 Disable measurements of total latency numbers. Useful only for cutting
1021 back the number of calls to gettimeofday, as that does impact performance at
1022 really high IOPS rates. Note that to really get rid of a large amount of these
1023 calls, this option must be used with disable_slat and disable_bw as well.
1025 .BI disable_clat \fR=\fPbool
1026 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
1028 .BI disable_slat \fR=\fPbool
1029 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
1031 .BI disable_bw_measurement \fR=\fPbool
1032 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
1034 .BI lockmem \fR=\fPint
1035 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
1036 simulate a smaller amount of memory.
1038 .BI exec_prerun \fR=\fPstr
1039 Before running the job, execute the specified command with \fBsystem\fR\|(3).
1041 .BI exec_postrun \fR=\fPstr
1042 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
1044 .BI ioscheduler \fR=\fPstr
1045 Attempt to switch the device hosting the file to the specified I/O scheduler.
1047 .BI cpuload \fR=\fPint
1048 If the job is a CPU cycle-eater, attempt to use the specified percentage of
1051 .BI cpuchunks \fR=\fPint
1052 If the job is a CPU cycle-eater, split the load into cycles of the
1053 given time in milliseconds.
1055 .BI disk_util \fR=\fPbool
1056 Generate disk utilization statistics if the platform supports it. Default: true.
1058 .BI clocksource \fR=\fPstr
1059 Use the given clocksource as the base of timing. The supported options are:
1069 Internal CPU clock source
1073 \fBcpu\fR is the preferred clocksource if it is reliable, as it is very fast
1074 (and fio is heavy on time calls). Fio will automatically use this clocksource
1075 if it's supported and considered reliable on the system it is running on,
1076 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
1077 means supporting TSC Invariant.
1079 .BI gtod_reduce \fR=\fPbool
1080 Enable all of the gettimeofday() reducing options (disable_clat, disable_slat,
1081 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
1082 gettimeofday() call count. With this option enabled, we only do about 0.4% of
1083 the gtod() calls we would have done if all time keeping was enabled.
1085 .BI gtod_cpu \fR=\fPint
1086 Sometimes it's cheaper to dedicate a single thread of execution to just getting
1087 the current time. Fio (and databases, for instance) are very intensive on
1088 gettimeofday() calls. With this option, you can set one CPU aside for doing
1089 nothing but logging current time to a shared memory location. Then the other
1090 threads/processes that run IO workloads need only copy that segment, instead of
1091 entering the kernel with a gettimeofday() call. The CPU set aside for doing
1092 these time calls will be excluded from other uses. Fio will manually clear it
1093 from the CPU mask of other jobs.
1095 .BI ignore_error \fR=\fPstr
1096 Sometimes you want to ignore some errors during test in that case you can specify
1097 error list for each error type.
1099 ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST
1101 errors for given error type is separated with ':'.
1102 Error may be symbol ('ENOSPC', 'ENOMEM') or an integer.
1104 Example: ignore_error=EAGAIN,ENOSPC:122 .
1106 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from WRITE.
1108 .BI error_dump \fR=\fPbool
1109 If set dump every error even if it is non fatal, true by default. If disabled
1110 only fatal error will be dumped
1112 .BI cgroup \fR=\fPstr
1113 Add job to this control group. If it doesn't exist, it will be created.
1114 The system must have a mounted cgroup blkio mount point for this to work. If
1115 your system doesn't have it mounted, you can do so with:
1117 # mount \-t cgroup \-o blkio none /cgroup
1119 .BI cgroup_weight \fR=\fPint
1120 Set the weight of the cgroup to this value. See the documentation that comes
1121 with the kernel, allowed values are in the range of 100..1000.
1123 .BI cgroup_nodelete \fR=\fPbool
1124 Normally fio will delete the cgroups it has created after the job completion.
1125 To override this behavior and to leave cgroups around after the job completion,
1126 set cgroup_nodelete=1. This can be useful if one wants to inspect various
1127 cgroup files after job completion. Default: false
1130 Instead of running as the invoking user, set the user ID to this value before
1131 the thread/process does any work.
1134 Set group ID, see \fBuid\fR.
1136 .BI flow_id \fR=\fPint
1137 The ID of the flow. If not specified, it defaults to being a global flow. See
1141 Weight in token-based flow control. If this value is used, then there is a
1142 \fBflow counter\fR which is used to regulate the proportion of activity between
1143 two or more jobs. fio attempts to keep this flow counter near zero. The
1144 \fBflow\fR parameter stands for how much should be added or subtracted to the
1145 flow counter on each iteration of the main I/O loop. That is, if one job has
1146 \fBflow=8\fR and another job has \fBflow=-1\fR, then there will be a roughly
1147 1:8 ratio in how much one runs vs the other.
1149 .BI flow_watermark \fR=\fPint
1150 The maximum value that the absolute value of the flow counter is allowed to
1151 reach before the job must wait for a lower value of the counter.
1153 .BI flow_sleep \fR=\fPint
1154 The period of time, in microseconds, to wait after the flow watermark has been
1155 exceeded before retrying operations
1157 .BI clat_percentiles \fR=\fPbool
1158 Enable the reporting of percentiles of completion latencies.
1160 .BI percentile_list \fR=\fPfloat_list
1161 Overwrite the default list of percentiles for completion
1162 latencies. Each number is a floating number in the range (0,100], and
1163 the maximum length of the list is 20. Use ':' to separate the
1164 numbers. For example, \-\-percentile_list=99.5:99.9 will cause fio to
1165 report the values of completion latency below which 99.5% and 99.9% of
1166 the observed latencies fell, respectively.
1167 .SS "Ioengine Parameters List"
1168 Some parameters are only valid when a specific ioengine is in use. These are
1169 used identically to normal parameters, with the caveat that when used on the
1170 command line, the must come after the ioengine that defines them is selected.
1172 .BI (libaio)userspace_reap
1173 Normally, with the libaio engine in use, fio will use
1174 the io_getevents system call to reap newly returned events.
1175 With this flag turned on, the AIO ring will be read directly
1176 from user-space to reap events. The reaping mode is only
1177 enabled when polling for a minimum of 0 events (eg when
1178 iodepth_batch_complete=0).
1180 .BI (net,netsplice)hostname \fR=\fPstr
1181 The host name or IP address to use for TCP or UDP based IO.
1182 If the job is a TCP listener or UDP reader, the hostname is not
1183 used and must be omitted.
1185 .BI (net,netsplice)port \fR=\fPint
1186 The TCP or UDP port to bind to or connect to.
1188 .BI (net,netsplice)nodelay \fR=\fPbool
1189 Set TCP_NODELAY on TCP connections.
1191 .BI (net,netsplice)protocol \fR=\fPstr "\fR,\fP proto" \fR=\fPstr
1192 The network protocol to use. Accepted values are:
1197 Transmission control protocol
1200 User datagram protocol
1206 When the protocol is TCP or UDP, the port must also be given,
1207 as well as the hostname if the job is a TCP listener or UDP
1208 reader. For unix sockets, the normal filename option should be
1209 used and the port is invalid.
1212 .BI (net,netsplice)listen
1213 For TCP network connections, tell fio to listen for incoming
1214 connections rather than initiating an outgoing connection. The
1215 hostname must be omitted if this option is used.
1217 .BI (net, pingpong) \fR=\fPbool
1218 Normal a network writer will just continue writing data, and a network reader
1219 will just consume packages. If pingpong=1 is set, a writer will send its normal
1220 payload to the reader, then wait for the reader to send the same payload back.
1221 This allows fio to measure network latencies. The submission and completion
1222 latencies then measure local time spent sending or receiving, and the
1223 completion latency measures how long it took for the other end to receive and
1226 .BI (e4defrag,donorname) \fR=\fPstr
1227 File will be used as a block donor (swap extents between files)
1229 .BI (e4defrag,inplace) \fR=\fPint
1230 Configure donor file block allocation strategy
1233 Preallocate donor's file on init
1236 allocate space immidietly inside defragment event, and free right after event
1240 While running, \fBfio\fR will display the status of the created jobs. For
1244 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
1247 The characters in the first set of brackets denote the current status of each
1248 threads. The possible values are:
1254 Setup but not started.
1260 Initialized, waiting.
1263 Running, doing sequential reads.
1266 Running, doing random reads.
1269 Running, doing sequential writes.
1272 Running, doing random writes.
1275 Running, doing mixed sequential reads/writes.
1278 Running, doing mixed random reads/writes.
1281 Running, currently waiting for \fBfsync\fR\|(2).
1284 Running, verifying written data.
1287 Exited, not reaped by main thread.
1290 Exited, thread reaped.
1294 The second set of brackets shows the estimated completion percentage of
1295 the current group. The third set shows the read and write I/O rate,
1296 respectively. Finally, the estimated run time of the job is displayed.
1298 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
1299 for each thread, each group of threads, and each disk, in that order.
1301 Per-thread statistics first show the threads client number, group-id, and
1302 error code. The remaining figures are as follows:
1306 Number of megabytes of I/O performed.
1309 Average data rate (bandwidth).
1315 Submission latency minimum, maximum, average and standard deviation. This is
1316 the time it took to submit the I/O.
1319 Completion latency minimum, maximum, average and standard deviation. This
1320 is the time between submission and completion.
1323 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
1324 and standard deviation.
1327 CPU usage statistics. Includes user and system time, number of context switches
1328 this thread went through and number of major and minor page faults.
1331 Distribution of I/O depths. Each depth includes everything less than (or equal)
1332 to it, but greater than the previous depth.
1335 Number of read/write requests issued, and number of short read/write requests.
1338 Distribution of I/O completion latencies. The numbers follow the same pattern
1342 The group statistics show:
1347 Number of megabytes I/O performed.
1350 Aggregate bandwidth of threads in the group.
1353 Minimum average bandwidth a thread saw.
1356 Maximum average bandwidth a thread saw.
1359 Shortest runtime of threads in the group.
1362 Longest runtime of threads in the group.
1366 Finally, disk statistics are printed with reads first:
1371 Number of I/Os performed by all groups.
1374 Number of merges in the I/O scheduler.
1377 Number of ticks we kept the disk busy.
1380 Total time spent in the disk queue.
1387 It is also possible to get fio to dump the current output while it is
1388 running, without terminating the job. To do that, send fio the \fBUSR1\fR
1391 If the \fB\-\-minimal\fR option is given, the results will be printed in a
1392 semicolon-delimited format suitable for scripted use - a job description
1393 (if provided) follows on a new line. Note that the first
1394 number in the line is the version number. If the output has to be changed
1395 for some reason, this number will be incremented by 1 to signify that
1396 change. The fields are:
1399 .B terse version, fio version, jobname, groupid, error
1403 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1407 .B min, max, mean, standard deviation
1411 .B min, max, mean, standard deviation
1413 Completion latency percentiles (20 fields):
1415 .B Xth percentile=usec
1419 .B min, max, mean, standard deviation
1423 .B min, max, aggregate percentage of total, mean, standard deviation
1429 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1433 .B min, max, mean, standard deviation
1437 .B min, max, mean, standard deviation
1439 Completion latency percentiles (20 fields):
1441 .B Xth percentile=usec
1445 .B min, max, mean, standard deviation
1449 .B min, max, aggregate percentage of total, mean, standard deviation
1455 .B user, system, context switches, major page faults, minor page faults
1458 IO depth distribution:
1460 .B <=1, 2, 4, 8, 16, 32, >=64
1463 IO latency distribution:
1467 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1471 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1475 Disk utilization (1 for each disk used):
1477 .B name, read ios, write ios, read merges, write merges, read ticks, write ticks, read in-queue time, write in-queue time, disk utilization percentage
1480 Error Info (dependent on continue_on_error, default off):
1482 .B total # errors, first error code
1485 .B text description (if provided in config - appears on newline)
1488 Normally you would run fio as a stand-alone application on the machine
1489 where the IO workload should be generated. However, it is also possible to
1490 run the frontend and backend of fio separately. This makes it possible to
1491 have a fio server running on the machine(s) where the IO workload should
1492 be running, while controlling it from another machine.
1494 To start the server, you would do:
1496 \fBfio \-\-server=args\fR
1498 on that machine, where args defines what fio listens to. The arguments
1499 are of the form 'type:hostname or IP:port'. 'type' is either 'ip' (or ip4)
1500 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain
1501 socket. 'hostname' is either a hostname or IP address, and 'port' is the port to
1502 listen to (only valid for TCP/IP, not a local socket). Some examples:
1506 Start a fio server, listening on all interfaces on the default port (8765).
1508 2) fio \-\-server=ip:hostname,4444
1510 Start a fio server, listening on IP belonging to hostname and on port 4444.
1512 3) fio \-\-server=ip6:::1,4444
1514 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
1516 4) fio \-\-server=,4444
1518 Start a fio server, listening on all interfaces on port 4444.
1520 5) fio \-\-server=1.2.3.4
1522 Start a fio server, listening on IP 1.2.3.4 on the default port.
1524 6) fio \-\-server=sock:/tmp/fio.sock
1526 Start a fio server, listening on the local socket /tmp/fio.sock.
1528 When a server is running, you can connect to it from a client. The client
1531 fio \-\-local-args \-\-client=server \-\-remote-args <job file(s)>
1533 where \-\-local-args are arguments that are local to the client where it is
1534 running, 'server' is the connect string, and \-\-remote-args and <job file(s)>
1535 are sent to the server. The 'server' string follows the same format as it
1536 does on the server side, to allow IP/hostname/socket and port strings.
1537 You can connect to multiple clients as well, to do that you could run:
1539 fio \-\-client=server2 \-\-client=server2 <job file(s)>
1543 was written by Jens Axboe <jens.axboe@oracle.com>,
1544 now Jens Axboe <jaxboe@fusionio.com>.
1546 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1547 on documentation by Jens Axboe.
1548 .SH "REPORTING BUGS"
1549 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1552 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1554 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob