1 .TH fio 1 "October 2013" "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'.
57 .BI \-\-eta\-newline \fR=\fPtime
58 Force an ETA newline for every `time` period passed.
60 .BI \-\-status\-interval \fR=\fPtime
61 Report full output status every `time` period passed.
64 Turn on safety read-only checks, preventing any attempted write.
66 .BI \-\-section \fR=\fPsec
67 Only run section \fIsec\fR from job file. Multiple of these options can be given, adding more sections to run.
69 .BI \-\-alloc\-size \fR=\fPkb
70 Set the internal smalloc pool size to \fIkb\fP kilobytes.
72 .BI \-\-warnings\-fatal
73 All fio parser warnings are fatal, causing fio to exit with an error.
75 .BI \-\-max\-jobs \fR=\fPnr
76 Set the maximum allowed number of jobs (threads/processes) to support.
78 .BI \-\-server \fR=\fPargs
79 Start a backend server, with \fIargs\fP specifying what to listen to. See client/server section.
81 .BI \-\-daemonize \fR=\fPpidfile
82 Background a fio server, writing the pid to the given pid file.
84 .BI \-\-client \fR=\fPhost
85 Instead of running the jobs locally, send and run them on the given host.
87 .BI \-\-idle\-prof \fR=\fPoption
88 Report cpu idleness on a system or percpu basis (\fIoption\fP=system,percpu) or run unit work calibration only (\fIoption\fP=calibrate).
90 Job files are in `ini' format. They consist of one or more
91 job definitions, which begin with a job name in square brackets and
92 extend to the next job name. The job name can be any ASCII string
93 except `global', which has a special meaning. Following the job name is
94 a sequence of zero or more parameters, one per line, that define the
95 behavior of the job. Any line starting with a `;' or `#' character is
96 considered a comment and ignored.
98 If \fIjobfile\fR is specified as `-', the job file will be read from
101 The global section contains default parameters for jobs specified in the
102 job file. A job is only affected by global sections residing above it,
103 and there may be any number of global sections. Specific job definitions
104 may override any parameter set in global sections.
107 Some parameters may take arguments of a specific type. The types used are:
110 String: a sequence of alphanumeric characters.
113 SI integer: a whole number, possibly containing a suffix denoting the base unit
114 of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting
115 kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5)
116 respectively. The suffix is not case sensitive. If prefixed with '0x', the
117 value is assumed to be base 16 (hexadecimal). A suffix may include a trailing 'b',
118 for instance 'kb' is identical to 'k'. You can specify a base 10 value
119 by using 'KiB', 'MiB', 'GiB', etc. This is useful for disk drives where
120 values are often given in base 10 values. Specifying '30GiB' will get you
124 Boolean: a true or false value. `0' denotes false, `1' denotes true.
127 Integer range: a range of integers specified in the format
128 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
129 \fIupper\fR may contain a suffix as described above. If an option allows two
130 sets of ranges, they are separated with a `,' or `/' character. For example:
134 List of floating numbers: A list of floating numbers, separated by
139 May be used to override the job name. On the command line, this parameter
140 has the special purpose of signalling the start of a new job.
142 .BI description \fR=\fPstr
143 Human-readable description of the job. It is printed when the job is run, but
144 otherwise has no special purpose.
146 .BI directory \fR=\fPstr
147 Prefix filenames with this directory. Used to place files in a location other
150 .BI filename \fR=\fPstr
152 normally makes up a file name based on the job name, thread number, and file
153 number. If you want to share files between threads in a job or several jobs,
154 specify a \fIfilename\fR for each of them to override the default.
155 If the I/O engine is file-based, you can specify
156 a number of files by separating the names with a `:' character. `\-' is a
157 reserved name, meaning stdin or stdout, depending on the read/write direction
160 .BI filename_format \fR=\fPstr
161 If sharing multiple files between jobs, it is usually necessary to have
162 fio generate the exact names that you want. By default, fio will name a file
163 based on the default file format specification of
164 \fBjobname.jobnumber.filenumber\fP. With this option, that can be
165 customized. Fio will recognize and replace the following keywords in this
171 The name of the worker thread or process.
174 The incremental number of the worker thread or process.
177 The incremental number of the file for that worker thread or process.
180 To have dependent jobs share a set of files, this option can be set to
181 have fio generate filenames that are shared between the two. For instance,
182 if \fBtestfiles.$filenum\fR is specified, file number 4 for any job will
183 be named \fBtestfiles.4\fR. The default of \fB$jobname.$jobnum.$filenum\fR
184 will be used if no other format specifier is given.
188 .BI lockfile \fR=\fPstr
189 Fio defaults to not locking any files before it does IO to them. If a file or
190 file descriptor is shared, fio can serialize IO to that file to make the end
191 result consistent. This is usual for emulating real workloads that share files.
197 No locking. This is the default.
200 Only one thread or process may do IO at the time, excluding all others.
203 Read-write locking on the file. Many readers may access the file at the same
204 time, but writes get exclusive access.
208 .BI opendir \fR=\fPstr
209 Recursively open any files below directory \fIstr\fR.
211 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
212 Type of I/O pattern. Accepted values are:
223 Sequential trim (Linux block devices only).
232 Random trim (Linux block devices only).
235 Mixed sequential reads and writes.
238 Mixed random reads and writes.
241 For mixed I/O, the default split is 50/50. For certain types of io the result
242 may still be skewed a bit, since the speed may be different. It is possible to
243 specify a number of IO's to do before getting a new offset, this is done by
244 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
245 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
246 value of 8. If the postfix is used with a sequential IO pattern, then the value
247 specified will be added to the generated offset for each IO. For instance,
248 using \fBrw=write:4k\fR will skip 4k for every write. It turns sequential IO
249 into sequential IO with holes. See the \fBrw_sequencer\fR option.
252 .BI rw_sequencer \fR=\fPstr
253 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
254 then this option controls how that number modifies the IO offset being
255 generated. Accepted values are:
260 Generate sequential offset
263 Generate the same offset
266 \fBsequential\fR is only useful for random IO, where fio would normally
267 generate a new random offset for every IO. If you append eg 8 to randread, you
268 would get a new random offset for every 8 IO's. The result would be a seek for
269 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
270 that. As sequential IO is already sequential, setting \fBsequential\fR for that
271 would not result in any differences. \fBidentical\fR behaves in a similar
272 fashion, except it sends the same offset 8 number of times before generating a
277 .BI kb_base \fR=\fPint
278 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
279 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
280 reasons. Allowed values are 1024 or 1000, with 1024 being the default.
282 .BI unified_rw_reporting \fR=\fPbool
283 Fio normally reports statistics on a per data direction basis, meaning that
284 read, write, and trim are accounted and reported separately. If this option is
285 set, the fio will sum the results and report them as "mixed" instead.
287 .BI randrepeat \fR=\fPbool
288 Seed the random number generator in a predictable way so results are repeatable
289 across runs. Default: true.
291 .BI randseed \fR=\fPint
292 Seed the random number generators based on this seed value, to be able to
293 control what sequence of output is being generated. If not set, the random
294 sequence depends on the \fBrandrepeat\fR setting.
296 .BI use_os_rand \fR=\fPbool
297 Fio can either use the random generator supplied by the OS to generator random
298 offsets, or it can use it's own internal generator (based on Tausworthe).
299 Default is to use the internal generator, which is often of better quality and
300 faster. Default: false.
302 .BI fallocate \fR=\fPstr
303 Whether pre-allocation is performed when laying down files. Accepted values
309 Do not pre-allocate space.
312 Pre-allocate via \fBposix_fallocate\fR\|(3).
315 Pre-allocate via \fBfallocate\fR\|(2) with FALLOC_FL_KEEP_SIZE set.
318 Backward-compatible alias for 'none'.
321 Backward-compatible alias for 'posix'.
324 May not be available on all supported platforms. 'keep' is only
325 available on Linux. If using ZFS on Solaris this must be set to 'none'
326 because ZFS doesn't support it. Default: 'posix'.
329 .BI fadvise_hint \fR=\fPbool
330 Use of \fBposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
331 are likely to be issued. Default: true.
334 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
335 been transferred, unless limited by other options (\fBruntime\fR, for instance).
336 Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be
337 divided between the available files for the job. If not set, fio will use the
338 full size of the given files or devices. If the files do not exist, size
339 must be given. It is also possible to give size as a percentage between 1 and
340 100. If size=20% is given, fio will use 20% of the full size of the given files
343 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
344 Sets size to something really large and waits for ENOSPC (no space left on
345 device) as the terminating condition. Only makes sense with sequential write.
346 For a read workload, the mount point will be filled first then IO started on
347 the result. This option doesn't make sense if operating on a raw device node,
348 since the size of that is already known by the file system. Additionally,
349 writing beyond end-of-device will not return ENOSPC there.
351 .BI filesize \fR=\fPirange
352 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
353 for files at random within the given range, limited to \fBsize\fR in total (if
354 that is given). If \fBfilesize\fR is not specified, each created file is the
357 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
358 Block size for I/O units. Default: 4k. Values for reads, writes, and trims
359 can be specified separately in the format \fIread\fR,\fIwrite\fR,\fItrim\fR
360 either of which may be empty to leave that value at its default. If a trailing
361 comma isn't given, the remainder will inherit the last value set.
363 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
364 Specify a range of I/O block sizes. The issued I/O unit will always be a
365 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
366 to both reads and writes if only one range is given, but can be specified
367 separately with a comma separating the values. Example: bsrange=1k-4k,2k-8k.
368 Also (see \fBblocksize\fR).
370 .BI bssplit \fR=\fPstr
371 This option allows even finer grained control of the block sizes issued,
372 not just even splits between them. With this option, you can weight various
373 block sizes for exact control of the issued IO for a job that has mixed
374 block sizes. The format of the option is bssplit=blocksize/percentage,
375 optionally adding as many definitions as needed separated by a colon.
376 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
377 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
378 splits to reads and writes. The format is identical to what the
379 \fBbs\fR option accepts, the read and write parts are separated with a
382 .B blocksize_unaligned\fR,\fP bs_unaligned
383 If set, any size in \fBblocksize_range\fR may be used. This typically won't
384 work with direct I/O, as that normally requires sector alignment.
386 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
387 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
388 the minimum blocksize given. Minimum alignment is typically 512b
389 for using direct IO, though it usually depends on the hardware block size.
390 This option is mutually exclusive with using a random map for files, so it
391 will turn off that option.
393 .BI bs_is_seq_rand \fR=\fPbool
394 If this option is set, fio will use the normal read,write blocksize settings as
395 sequential,random instead. Any random read or write will use the WRITE
396 blocksize settings, and any sequential read or write will use the READ
400 Initialise buffers with all zeros. Default: fill buffers with random data.
403 If this option is given, fio will refill the IO buffers on every submit. The
404 default is to only fill it at init time and reuse that data. Only makes sense
405 if zero_buffers isn't specified, naturally. If data verification is enabled,
406 refill_buffers is also automatically enabled.
408 .BI scramble_buffers \fR=\fPbool
409 If \fBrefill_buffers\fR is too costly and the target is using data
410 deduplication, then setting this option will slightly modify the IO buffer
411 contents to defeat normal de-dupe attempts. This is not enough to defeat
412 more clever block compression attempts, but it will stop naive dedupe
413 of blocks. Default: true.
415 .BI buffer_compress_percentage \fR=\fPint
416 If this is set, then fio will attempt to provide IO buffer content (on WRITEs)
417 that compress to the specified level. Fio does this by providing a mix of
418 random data and zeroes. Note that this is per block size unit, for file/disk
419 wide compression level that matches this setting, you'll also want to set
420 \fBrefill_buffers\fR.
422 .BI buffer_compress_chunk \fR=\fPint
423 See \fBbuffer_compress_percentage\fR. This setting allows fio to manage how
424 big the ranges of random data and zeroed data is. Without this set, fio will
425 provide \fBbuffer_compress_percentage\fR of blocksize random data, followed by
426 the remaining zeroed. With this set to some chunk size smaller than the block
427 size, fio can alternate random and zeroed data throughout the IO buffer.
429 .BI buffer_pattern \fR=\fPstr
430 If set, fio will fill the io buffers with this pattern. If not set, the contents
431 of io buffers is defined by the other options related to buffer contents. The
432 setting can be any pattern of bytes, and can be prefixed with 0x for hex
435 .BI nrfiles \fR=\fPint
436 Number of files to use for this job. Default: 1.
438 .BI openfiles \fR=\fPint
439 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
441 .BI file_service_type \fR=\fPstr
442 Defines how files to service are selected. The following types are defined:
447 Choose a file at random.
450 Round robin over open files (default).
453 Do each file in the set sequentially.
456 The number of I/Os to issue before switching a new file can be specified by
457 appending `:\fIint\fR' to the service type.
460 .BI ioengine \fR=\fPstr
461 Defines how the job issues I/O. The following types are defined:
466 Basic \fBread\fR\|(2) or \fBwrite\fR\|(2) I/O. \fBfseek\fR\|(2) is used to
467 position the I/O location.
470 Basic \fBpread\fR\|(2) or \fBpwrite\fR\|(2) I/O.
473 Basic \fBreadv\fR\|(2) or \fBwritev\fR\|(2) I/O. Will emulate queuing by
474 coalescing adjacent IOs into a single submission.
477 Basic \fBpreadv\fR\|(2) or \fBpwritev\fR\|(2) I/O.
480 Linux native asynchronous I/O. This ioengine defines engine specific options.
483 POSIX asynchronous I/O using \fBaio_read\fR\|(3) and \fBaio_write\fR\|(3).
486 Solaris native asynchronous I/O.
489 Windows native asynchronous I/O.
492 File is memory mapped with \fBmmap\fR\|(2) and data copied using
496 \fBsplice\fR\|(2) is used to transfer the data and \fBvmsplice\fR\|(2) to
497 transfer data from user-space to the kernel.
500 Use the syslet system calls to make regular read/write asynchronous.
503 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
504 the target is an sg character device, we use \fBread\fR\|(2) and
505 \fBwrite\fR\|(2) for asynchronous I/O.
508 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
509 itself and for debugging and testing purposes.
512 Transfer over the network. The protocol to be used can be defined with the
513 \fBprotocol\fR parameter. Depending on the protocol, \fBfilename\fR,
514 \fBhostname\fR, \fBport\fR, or \fBlisten\fR must be specified.
515 This ioengine defines engine specific options.
518 Like \fBnet\fR, but uses \fBsplice\fR\|(2) and \fBvmsplice\fR\|(2) to map data
519 and send/receive. This ioengine defines engine specific options.
522 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
523 \fBcpucycles\fR parameters.
526 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
527 approach to asynchronous I/O.
529 See <http://www.xmailserver.org/guasi\-lib.html>.
532 The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ)
533 and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols.
536 Loads an external I/O engine object file. Append the engine filename as
540 IO engine that does regular linux native fallocate call to simulate data
541 transfer as fio ioengine
543 DDIR_READ does fallocate(,mode = FALLOC_FL_KEEP_SIZE,)
545 DIR_WRITE does fallocate(,mode = 0)
547 DDIR_TRIM does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE)
550 IO engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate defragment activity
551 request to DDIR_WRITE event
556 .BI iodepth \fR=\fPint
557 Number of I/O units to keep in flight against the file. Note that increasing
558 iodepth beyond 1 will not affect synchronous ioengines (except for small
559 degress when verify_async is in use). Even async engines my impose OS
560 restrictions causing the desired depth not to be achieved. This may happen on
561 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
562 not async on that OS. Keep an eye on the IO depth distribution in the
563 fio output to verify that the achieved depth is as expected. Default: 1.
565 .BI iodepth_batch \fR=\fPint
566 Number of I/Os to submit at once. Default: \fBiodepth\fR.
568 .BI iodepth_batch_complete \fR=\fPint
569 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
570 means that we'll ask for a minimum of 1 IO in the retrieval process from the
571 kernel. The IO retrieval will go on until we hit the limit set by
572 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
573 completed events before queuing more IO. This helps reduce IO latency, at the
574 cost of more retrieval system calls.
576 .BI iodepth_low \fR=\fPint
577 Low watermark indicating when to start filling the queue again. Default:
580 .BI direct \fR=\fPbool
581 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
583 .BI atomic \fR=\fPbool
584 If value is true, attempt to use atomic direct IO. Atomic writes are guaranteed
585 to be stable once acknowledged by the operating system. Only Linux supports
588 .BI buffered \fR=\fPbool
589 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
592 .BI offset \fR=\fPint
593 Offset in the file to start I/O. Data before the offset will not be touched.
595 .BI offset_increment \fR=\fPint
596 If this is provided, then the real offset becomes the
597 offset + offset_increment * thread_number, where the thread number is a counter
598 that starts at 0 and is incremented for each job. This option is useful if
599 there are several jobs which are intended to operate on a file in parallel in
600 disjoint segments, with even spacing between the starting points.
602 .BI number_ios \fR=\fPint
603 Fio will normally perform IOs until it has exhausted the size of the region
604 set by \fBsize\fR, or if it exhaust the allocated time (or hits an error
605 condition). With this setting, the range/size can be set independently of
606 the number of IOs to perform. When fio reaches this number, it will exit
607 normally and report status.
610 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
611 0, don't sync. Default: 0.
613 .BI fdatasync \fR=\fPint
614 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
615 data parts of the file. Default: 0.
617 .BI write_barrier \fR=\fPint
618 Make every Nth write a barrier write.
620 .BI sync_file_range \fR=\fPstr:int
621 Use \fBsync_file_range\fR\|(2) for every \fRval\fP number of write operations. Fio will
622 track range of writes that have happened since the last \fBsync_file_range\fR\|(2) call.
623 \fRstr\fP can currently be one or more of:
627 SYNC_FILE_RANGE_WAIT_BEFORE
630 SYNC_FILE_RANGE_WRITE
633 SYNC_FILE_RANGE_WRITE
637 So if you do sync_file_range=wait_before,write:8, fio would use
638 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
639 Also see the \fBsync_file_range\fR\|(2) man page. This option is Linux specific.
641 .BI overwrite \fR=\fPbool
642 If writing, setup the file first and do overwrites. Default: false.
644 .BI end_fsync \fR=\fPbool
645 Sync file contents when a write stage has completed. Default: false.
647 .BI fsync_on_close \fR=\fPbool
648 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
649 it will happen on every close, not just at the end of the job. Default: false.
651 .BI rwmixread \fR=\fPint
652 Percentage of a mixed workload that should be reads. Default: 50.
654 .BI rwmixwrite \fR=\fPint
655 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
656 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
657 overrides the first. This may interfere with a given rate setting, if fio is
658 asked to limit reads or writes to a certain rate. If that is the case, then
659 the distribution may be skewed. Default: 50.
661 .BI random_distribution \fR=\fPstr:float
662 By default, fio will use a completely uniform random distribution when asked
663 to perform random IO. Sometimes it is useful to skew the distribution in
664 specific ways, ensuring that some parts of the data is more hot than others.
665 Fio includes the following distribution models:
669 Uniform random distribution
679 When using a zipf or pareto distribution, an input value is also needed to
680 define the access pattern. For zipf, this is the zipf theta. For pareto,
681 it's the pareto power. Fio includes a test program, genzipf, that can be
682 used visualize what the given input values will yield in terms of hit rates.
683 If you wanted to use zipf with a theta of 1.2, you would use
684 random_distribution=zipf:1.2 as the option. If a non-uniform model is used,
685 fio will disable use of the random map.
687 .BI percentage_random \fR=\fPint
688 For a random workload, set how big a percentage should be random. This defaults
689 to 100%, in which case the workload is fully random. It can be set from
690 anywhere from 0 to 100. Setting it to 0 would make the workload fully
691 sequential. It is possible to set different values for reads, writes, and
692 trim. To do so, simply use a comma separated list. See \fBblocksize\fR.
695 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
696 this parameter is given, a new offset will be chosen without looking at past
697 I/O history. This parameter is mutually exclusive with \fBverify\fR.
699 .BI softrandommap \fR=\fPbool
700 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
701 fails to allocate the map, if this option is set it will continue without a
702 random block map. As coverage will not be as complete as with random maps, this
703 option is disabled by default.
705 .BI random_generator \fR=\fPstr
706 Fio supports the following engines for generating IO offsets for random IO:
710 Strong 2^88 cycle random number generator
713 Linear feedback shift register generator
717 Tausworthe is a strong random number generator, but it requires tracking on the
718 side if we want to ensure that blocks are only read or written once. LFSR
719 guarantees that we never generate the same offset twice, and it's also less
720 computationally expensive. It's not a true random generator, however, though
721 for IO purposes it's typically good enough. LFSR only works with single block
722 sizes, not with workloads that use multiple block sizes. If used with such a
723 workload, fio may read or write some blocks multiple times.
726 Run job with given nice value. See \fBnice\fR\|(2).
729 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
732 .BI prioclass \fR=\fPint
733 Set I/O priority class. See \fBionice\fR\|(1).
735 .BI thinktime \fR=\fPint
736 Stall job for given number of microseconds between issuing I/Os.
738 .BI thinktime_spin \fR=\fPint
739 Pretend to spend CPU time for given number of microseconds, sleeping the rest
740 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
742 .BI thinktime_blocks \fR=\fPint
743 Only valid if thinktime is set - control how many blocks to issue, before
744 waiting \fBthinktime\fR microseconds. If not set, defaults to 1 which will
745 make fio wait \fBthinktime\fR microseconds after every block. This
746 effectively makes any queue depth setting redundant, since no more than 1 IO
747 will be queued before we have to complete it and do our thinktime. In other
748 words, this setting effectively caps the queue depth if the latter is larger.
752 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
753 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
754 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
755 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
756 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
757 limit writes (to 500KB/sec), the latter will only limit reads.
759 .BI ratemin \fR=\fPint
760 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
761 Failing to meet this requirement will cause the job to exit. The same format
762 as \fBrate\fR is used for read vs write separation.
764 .BI rate_iops \fR=\fPint
765 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
766 specified independently of bandwidth. The same format as \fBrate\fR is used for
767 read vs write separation. If \fBblocksize\fR is a range, the smallest block
768 size is used as the metric.
770 .BI rate_iops_min \fR=\fPint
771 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
772 is used for read vs write separation.
774 .BI ratecycle \fR=\fPint
775 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
776 milliseconds. Default: 1000ms.
778 .BI latency_target \fR=\fPint
779 If set, fio will attempt to find the max performance point that the given
780 workload will run at while maintaining a latency below this target. The
781 values is given in microseconds. See \fBlatency_window\fR and
782 \fBlatency_percentile\fR.
784 .BI latency_window \fR=\fPint
785 Used with \fBlatency_target\fR to specify the sample window that the job
786 is run at varying queue depths to test the performance. The value is given
789 .BI latency_percentile \fR=\fPfloat
790 The percentage of IOs that must fall within the criteria specified by
791 \fBlatency_target\fR and \fBlatency_window\fR. If not set, this defaults
792 to 100.0, meaning that all IOs must be equal or below to the value set
793 by \fBlatency_target\fR.
795 .BI max_latency \fR=\fPint
796 If set, fio will exit the job if it exceeds this maximum latency. It will exit
799 .BI cpumask \fR=\fPint
800 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
801 may run on. See \fBsched_setaffinity\fR\|(2).
803 .BI cpus_allowed \fR=\fPstr
804 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
806 .BI numa_cpu_nodes \fR=\fPstr
807 Set this job running on specified NUMA nodes' CPUs. The arguments allow
808 comma delimited list of cpu numbers, A-B ranges, or 'all'.
810 .BI numa_mem_policy \fR=\fPstr
811 Set this job's memory policy and corresponding NUMA nodes. Format of
815 .B <mode>[:<nodelist>]
818 is one of the following memory policy:
820 .B default, prefer, bind, interleave, local
823 For \fBdefault\fR and \fBlocal\fR memory policy, no \fBnodelist\fR is
824 needed to be specified. For \fBprefer\fR, only one node is
825 allowed. For \fBbind\fR and \fBinterleave\fR, \fBnodelist\fR allows
826 comma delimited list of numbers, A-B ranges, or 'all'.
828 .BI startdelay \fR=\fPint
829 Delay start of job for the specified number of seconds.
831 .BI runtime \fR=\fPint
832 Terminate processing after the specified number of seconds.
835 If given, run for the specified \fBruntime\fR duration even if the files are
836 completely read or written. The same workload will be repeated as many times
837 as \fBruntime\fR allows.
839 .BI ramp_time \fR=\fPint
840 If set, fio will run the specified workload for this amount of time before
841 logging any performance numbers. Useful for letting performance settle before
842 logging results, thus minimizing the runtime required for stable results. Note
843 that the \fBramp_time\fR is considered lead in time for a job, thus it will
844 increase the total runtime if a special timeout or runtime is specified.
846 .BI invalidate \fR=\fPbool
847 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
850 Use synchronous I/O for buffered writes. For the majority of I/O engines,
851 this means using O_SYNC. Default: false.
853 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
854 Allocation method for I/O unit buffer. Allowed values are:
859 Allocate memory with \fBmalloc\fR\|(3).
862 Use shared memory buffers allocated through \fBshmget\fR\|(2).
865 Same as \fBshm\fR, but use huge pages as backing.
868 Use \fBmmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
869 is given after the option in the format `:\fIfile\fR'.
872 Same as \fBmmap\fR, but use huge files as backing.
875 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
876 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
877 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
878 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
879 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
880 and the documentation for that. Normally you just need to echo an appropriate
881 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
885 .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint
886 This indicates the memory alignment of the IO memory buffers. Note that the
887 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
888 the alignment of the following buffers are given by the \fBbs\fR used. In
889 other words, if using a \fBbs\fR that is a multiple of the page sized in the
890 system, all buffers will be aligned to this value. If using a \fBbs\fR that
891 is not page aligned, the alignment of subsequent IO memory buffers is the
892 sum of the \fBiomem_align\fR and \fBbs\fR used.
894 .BI hugepage\-size \fR=\fPint
895 Defines the size of a huge page. Must be at least equal to the system setting.
896 Should be a multiple of 1MB. Default: 4MB.
899 Terminate all jobs when one finishes. Default: wait for each job to finish.
901 .BI bwavgtime \fR=\fPint
902 Average bandwidth calculations over the given time in milliseconds. Default:
905 .BI iopsavgtime \fR=\fPint
906 Average IOPS calculations over the given time in milliseconds. Default:
909 .BI create_serialize \fR=\fPbool
910 If true, serialize file creation for the jobs. Default: true.
912 .BI create_fsync \fR=\fPbool
913 \fBfsync\fR\|(2) data file after creation. Default: true.
915 .BI create_on_open \fR=\fPbool
916 If true, the files are not created until they are opened for IO by the job.
918 .BI create_only \fR=\fPbool
919 If true, fio will only run the setup phase of the job. If files need to be
920 laid out or updated on disk, only that will be done. The actual job contents
923 .BI pre_read \fR=\fPbool
924 If this is given, files will be pre-read into memory before starting the given
925 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
926 pointless to pre-read and then drop the cache. This will only work for IO
927 engines that are seekable, since they allow you to read the same data
928 multiple times. Thus it will not work on eg network or splice IO.
930 .BI unlink \fR=\fPbool
931 Unlink job files when done. Default: false.
934 Specifies the number of iterations (runs of the same workload) of this job.
937 .BI verify_only \fR=\fPbool
938 Do not perform the specified workload, only verify data still matches previous
939 invocation of this workload. This option allows one to check data multiple
940 times at a later date without overwriting it. This option makes sense only for
941 workloads that write data, and does not support workloads with the
942 \fBtime_based\fR option set.
944 .BI do_verify \fR=\fPbool
945 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
948 .BI verify \fR=\fPstr
949 Method of verifying file contents after each iteration of the job. Allowed
954 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1
955 Store appropriate checksum in the header of each block. crc32c-intel is
956 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
957 not supported by the system.
960 Write extra information about each I/O (timestamp, block number, etc.). The
961 block number is verified. See \fBverify_pattern\fR as well.
964 Pretend to verify. Used for testing internals.
967 This option can be used for repeated burn-in tests of a system to make sure
968 that the written data is also correctly read back. If the data direction given
969 is a read or random read, fio will assume that it should verify a previously
970 written file. If the data direction includes any form of write, the verify will
971 be of the newly written data.
974 .BI verifysort \fR=\fPbool
975 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
976 read them back in a sorted manner. Default: true.
978 .BI verifysort_nr \fR=\fPint
979 Pre-load and sort verify blocks for a read workload.
981 .BI verify_offset \fR=\fPint
982 Swap the verification header with data somewhere else in the block before
983 writing. It is swapped back before verifying.
985 .BI verify_interval \fR=\fPint
986 Write the verification header for this number of bytes, which should divide
987 \fBblocksize\fR. Default: \fBblocksize\fR.
989 .BI verify_pattern \fR=\fPstr
990 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
991 with totally random bytes, but sometimes it's interesting to fill with a known
992 pattern for io verification purposes. Depending on the width of the pattern,
993 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
994 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
995 has to be a hex number that starts with either "0x" or "0X". Use with
998 .BI verify_fatal \fR=\fPbool
999 If true, exit the job on the first observed verification failure. Default:
1002 .BI verify_dump \fR=\fPbool
1003 If set, dump the contents of both the original data block and the data block we
1004 read off disk to files. This allows later analysis to inspect just what kind of
1005 data corruption occurred. Off by default.
1007 .BI verify_async \fR=\fPint
1008 Fio will normally verify IO inline from the submitting thread. This option
1009 takes an integer describing how many async offload threads to create for IO
1010 verification instead, causing fio to offload the duty of verifying IO contents
1011 to one or more separate threads. If using this offload option, even sync IO
1012 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
1013 allows them to have IO in flight while verifies are running.
1015 .BI verify_async_cpus \fR=\fPstr
1016 Tell fio to set the given CPU affinity on the async IO verification threads.
1017 See \fBcpus_allowed\fP for the format used.
1019 .BI verify_backlog \fR=\fPint
1020 Fio will normally verify the written contents of a job that utilizes verify
1021 once that job has completed. In other words, everything is written then
1022 everything is read back and verified. You may want to verify continually
1023 instead for a variety of reasons. Fio stores the meta data associated with an
1024 IO block in memory, so for large verify workloads, quite a bit of memory would
1025 be used up holding this meta data. If this option is enabled, fio will write
1026 only N blocks before verifying these blocks.
1028 .BI verify_backlog_batch \fR=\fPint
1029 Control how many blocks fio will verify if verify_backlog is set. If not set,
1030 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
1031 read back and verified). If \fBverify_backlog_batch\fR is less than
1032 \fBverify_backlog\fR then not all blocks will be verified, if
1033 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
1034 will be verified more than once.
1036 .BI trim_percentage \fR=\fPint
1037 Number of verify blocks to discard/trim.
1039 .BI trim_verify_zero \fR=\fPbool
1040 Verify that trim/discarded blocks are returned as zeroes.
1042 .BI trim_backlog \fR=\fPint
1043 Trim after this number of blocks are written.
1045 .BI trim_backlog_batch \fR=\fPint
1046 Trim this number of IO blocks.
1048 .BI experimental_verify \fR=\fPbool
1049 Enable experimental verification.
1051 .B stonewall "\fR,\fP wait_for_previous"
1052 Wait for preceding jobs in the job file to exit before starting this one.
1053 \fBstonewall\fR implies \fBnew_group\fR.
1056 Start a new reporting group. If not given, all jobs in a file will be part
1057 of the same reporting group, unless separated by a stonewall.
1059 .BI numjobs \fR=\fPint
1060 Number of clones (processes/threads performing the same workload) of this job.
1064 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
1068 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
1069 with \fBfork\fR\|(2).
1071 .BI zonesize \fR=\fPint
1072 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
1074 .BI zonerange \fR=\fPint
1075 Give size of an IO zone. See \fBzoneskip\fR.
1077 .BI zoneskip \fR=\fPint
1078 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
1081 .BI write_iolog \fR=\fPstr
1082 Write the issued I/O patterns to the specified file. Specify a separate file
1083 for each job, otherwise the iologs will be interspersed and the file may be
1086 .BI read_iolog \fR=\fPstr
1087 Replay the I/O patterns contained in the specified file generated by
1088 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
1090 .BI replay_no_stall \fR=\fPint
1091 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1092 attempts to respect timing information between I/Os. Enabling
1093 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
1094 still respecting ordering.
1096 .BI replay_redirect \fR=\fPstr
1097 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1098 is to replay the IOPS onto the major/minor device that each IOP was recorded
1099 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
1100 single specified device regardless of the device it was recorded from.
1102 .BI write_bw_log \fR=\fPstr
1103 If given, write a bandwidth log of the jobs in this job file. Can be used to
1104 store data of the bandwidth of the jobs in their lifetime. The included
1105 fio_generate_plots script uses gnuplot to turn these text files into nice
1106 graphs. See \fBwrite_lat_log\fR for behaviour of given filename. For this
1107 option, the postfix is _bw.log.
1109 .BI write_lat_log \fR=\fPstr
1110 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
1111 filename is given with this option, the default filename of "jobname_type.log"
1112 is used. Even if the filename is given, fio will still append the type of log.
1114 .BI write_iops_log \fR=\fPstr
1115 Same as \fBwrite_bw_log\fR, but writes IOPS. If no filename is given with this
1116 option, the default filename of "jobname_type.log" is used. Even if the
1117 filename is given, fio will still append the type of log.
1119 .BI log_avg_msec \fR=\fPint
1120 By default, fio will log an entry in the iops, latency, or bw log for every
1121 IO that completes. When writing to the disk log, that can quickly grow to a
1122 very large size. Setting this option makes fio average the each log entry
1123 over the specified period of time, reducing the resolution of the log.
1126 .BI disable_lat \fR=\fPbool
1127 Disable measurements of total latency numbers. Useful only for cutting
1128 back the number of calls to \fBgettimeofday\fR\|(2), as that does impact performance at
1129 really high IOPS rates. Note that to really get rid of a large amount of these
1130 calls, this option must be used with disable_slat and disable_bw as well.
1132 .BI disable_clat \fR=\fPbool
1133 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
1135 .BI disable_slat \fR=\fPbool
1136 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
1138 .BI disable_bw_measurement \fR=\fPbool
1139 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
1141 .BI lockmem \fR=\fPint
1142 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
1143 simulate a smaller amount of memory. The amount specified is per worker.
1145 .BI exec_prerun \fR=\fPstr
1146 Before running the job, execute the specified command with \fBsystem\fR\|(3).
1148 Output is redirected in a file called \fBjobname.prerun.txt\fR
1151 .BI exec_postrun \fR=\fPstr
1152 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
1154 Output is redirected in a file called \fBjobname.postrun.txt\fR
1157 .BI ioscheduler \fR=\fPstr
1158 Attempt to switch the device hosting the file to the specified I/O scheduler.
1160 .BI cpuload \fR=\fPint
1161 If the job is a CPU cycle-eater, attempt to use the specified percentage of
1164 .BI cpuchunks \fR=\fPint
1165 If the job is a CPU cycle-eater, split the load into cycles of the
1166 given time in milliseconds.
1168 .BI disk_util \fR=\fPbool
1169 Generate disk utilization statistics if the platform supports it. Default: true.
1171 .BI clocksource \fR=\fPstr
1172 Use the given clocksource as the base of timing. The supported options are:
1176 \fBgettimeofday\fR\|(2)
1179 \fBclock_gettime\fR\|(2)
1182 Internal CPU clock source
1186 \fBcpu\fR is the preferred clocksource if it is reliable, as it is very fast
1187 (and fio is heavy on time calls). Fio will automatically use this clocksource
1188 if it's supported and considered reliable on the system it is running on,
1189 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
1190 means supporting TSC Invariant.
1192 .BI gtod_reduce \fR=\fPbool
1193 Enable all of the \fBgettimeofday\fR\|(2) reducing options (disable_clat, disable_slat,
1194 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
1195 \fBgettimeofday\fR\|(2) call count. With this option enabled, we only do about 0.4% of
1196 the gtod() calls we would have done if all time keeping was enabled.
1198 .BI gtod_cpu \fR=\fPint
1199 Sometimes it's cheaper to dedicate a single thread of execution to just getting
1200 the current time. Fio (and databases, for instance) are very intensive on
1201 \fBgettimeofday\fR\|(2) calls. With this option, you can set one CPU aside for doing
1202 nothing but logging current time to a shared memory location. Then the other
1203 threads/processes that run IO workloads need only copy that segment, instead of
1204 entering the kernel with a \fBgettimeofday\fR\|(2) call. The CPU set aside for doing
1205 these time calls will be excluded from other uses. Fio will manually clear it
1206 from the CPU mask of other jobs.
1208 .BI ignore_error \fR=\fPstr
1209 Sometimes you want to ignore some errors during test in that case you can specify
1210 error list for each error type.
1212 ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST
1214 errors for given error type is separated with ':'.
1215 Error may be symbol ('ENOSPC', 'ENOMEM') or an integer.
1217 Example: ignore_error=EAGAIN,ENOSPC:122 .
1219 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from WRITE.
1221 .BI error_dump \fR=\fPbool
1222 If set dump every error even if it is non fatal, true by default. If disabled
1223 only fatal error will be dumped
1225 .BI profile \fR=\fPstr
1226 Select a specific builtin performance test.
1228 .BI cgroup \fR=\fPstr
1229 Add job to this control group. If it doesn't exist, it will be created.
1230 The system must have a mounted cgroup blkio mount point for this to work. If
1231 your system doesn't have it mounted, you can do so with:
1233 # mount \-t cgroup \-o blkio none /cgroup
1235 .BI cgroup_weight \fR=\fPint
1236 Set the weight of the cgroup to this value. See the documentation that comes
1237 with the kernel, allowed values are in the range of 100..1000.
1239 .BI cgroup_nodelete \fR=\fPbool
1240 Normally fio will delete the cgroups it has created after the job completion.
1241 To override this behavior and to leave cgroups around after the job completion,
1242 set cgroup_nodelete=1. This can be useful if one wants to inspect various
1243 cgroup files after job completion. Default: false
1246 Instead of running as the invoking user, set the user ID to this value before
1247 the thread/process does any work.
1250 Set group ID, see \fBuid\fR.
1252 .BI unit_base \fR=\fPint
1253 Base unit for reporting. Allowed values are:
1257 Use auto-detection (default).
1267 .BI flow_id \fR=\fPint
1268 The ID of the flow. If not specified, it defaults to being a global flow. See
1272 Weight in token-based flow control. If this value is used, then there is a
1273 \fBflow counter\fR which is used to regulate the proportion of activity between
1274 two or more jobs. fio attempts to keep this flow counter near zero. The
1275 \fBflow\fR parameter stands for how much should be added or subtracted to the
1276 flow counter on each iteration of the main I/O loop. That is, if one job has
1277 \fBflow=8\fR and another job has \fBflow=-1\fR, then there will be a roughly
1278 1:8 ratio in how much one runs vs the other.
1280 .BI flow_watermark \fR=\fPint
1281 The maximum value that the absolute value of the flow counter is allowed to
1282 reach before the job must wait for a lower value of the counter.
1284 .BI flow_sleep \fR=\fPint
1285 The period of time, in microseconds, to wait after the flow watermark has been
1286 exceeded before retrying operations
1288 .BI clat_percentiles \fR=\fPbool
1289 Enable the reporting of percentiles of completion latencies.
1291 .BI percentile_list \fR=\fPfloat_list
1292 Overwrite the default list of percentiles for completion
1293 latencies. Each number is a floating number in the range (0,100], and
1294 the maximum length of the list is 20. Use ':' to separate the
1295 numbers. For example, \-\-percentile_list=99.5:99.9 will cause fio to
1296 report the values of completion latency below which 99.5% and 99.9% of
1297 the observed latencies fell, respectively.
1298 .SS "Ioengine Parameters List"
1299 Some parameters are only valid when a specific ioengine is in use. These are
1300 used identically to normal parameters, with the caveat that when used on the
1301 command line, the must come after the ioengine that defines them is selected.
1303 .BI (cpu)cpuload \fR=\fPint
1304 Attempt to use the specified percentage of CPU cycles.
1306 .BI (cpu)cpuchunks \fR=\fPint
1307 Split the load into cycles of the given time. In microseconds.
1309 .BI (libaio)userspace_reap
1310 Normally, with the libaio engine in use, fio will use
1311 the io_getevents system call to reap newly returned events.
1312 With this flag turned on, the AIO ring will be read directly
1313 from user-space to reap events. The reaping mode is only
1314 enabled when polling for a minimum of 0 events (eg when
1315 iodepth_batch_complete=0).
1317 .BI (net,netsplice)hostname \fR=\fPstr
1318 The host name or IP address to use for TCP or UDP based IO.
1319 If the job is a TCP listener or UDP reader, the hostname is not
1320 used and must be omitted unless it is a valid UDP multicast address.
1322 .BI (net,netsplice)port \fR=\fPint
1323 The TCP or UDP port to bind to or connect to.
1325 .BI (net,netsplice)interface \fR=\fPstr
1326 The IP address of the network interface used to send or receive UDP multicast
1329 .BI (net,netsplice)ttl \fR=\fPint
1330 Time-to-live value for outgoing UDP multicast packets. Default: 1
1332 .BI (net,netsplice)nodelay \fR=\fPbool
1333 Set TCP_NODELAY on TCP connections.
1335 .BI (net,netsplice)protocol \fR=\fPstr "\fR,\fP proto" \fR=\fPstr
1336 The network protocol to use. Accepted values are:
1341 Transmission control protocol
1344 Transmission control protocol V6
1347 User datagram protocol
1350 User datagram protocol V6
1356 When the protocol is TCP or UDP, the port must also be given,
1357 as well as the hostname if the job is a TCP listener or UDP
1358 reader. For unix sockets, the normal filename option should be
1359 used and the port is invalid.
1362 .BI (net,netsplice)listen
1363 For TCP network connections, tell fio to listen for incoming
1364 connections rather than initiating an outgoing connection. The
1365 hostname must be omitted if this option is used.
1367 .BI (net, pingpong) \fR=\fPbool
1368 Normally a network writer will just continue writing data, and a network reader
1369 will just consume packages. If pingpong=1 is set, a writer will send its normal
1370 payload to the reader, then wait for the reader to send the same payload back.
1371 This allows fio to measure network latencies. The submission and completion
1372 latencies then measure local time spent sending or receiving, and the
1373 completion latency measures how long it took for the other end to receive and
1374 send back. For UDP multicast traffic pingpong=1 should only be set for a single
1375 reader when multiple readers are listening to the same address.
1377 .BI (e4defrag,donorname) \fR=\fPstr
1378 File will be used as a block donor (swap extents between files)
1380 .BI (e4defrag,inplace) \fR=\fPint
1381 Configure donor file block allocation strategy
1384 Preallocate donor's file on init
1387 allocate space immediately inside defragment event, and free right after event
1390 While running, \fBfio\fR will display the status of the created jobs. For
1394 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
1397 The characters in the first set of brackets denote the current status of each
1398 threads. The possible values are:
1404 Setup but not started.
1410 Initialized, waiting.
1413 Running, doing sequential reads.
1416 Running, doing random reads.
1419 Running, doing sequential writes.
1422 Running, doing random writes.
1425 Running, doing mixed sequential reads/writes.
1428 Running, doing mixed random reads/writes.
1431 Running, currently waiting for \fBfsync\fR\|(2).
1434 Running, verifying written data.
1437 Exited, not reaped by main thread.
1440 Exited, thread reaped.
1444 The second set of brackets shows the estimated completion percentage of
1445 the current group. The third set shows the read and write I/O rate,
1446 respectively. Finally, the estimated run time of the job is displayed.
1448 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
1449 for each thread, each group of threads, and each disk, in that order.
1451 Per-thread statistics first show the threads client number, group-id, and
1452 error code. The remaining figures are as follows:
1456 Number of megabytes of I/O performed.
1459 Average data rate (bandwidth).
1465 Submission latency minimum, maximum, average and standard deviation. This is
1466 the time it took to submit the I/O.
1469 Completion latency minimum, maximum, average and standard deviation. This
1470 is the time between submission and completion.
1473 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
1474 and standard deviation.
1477 CPU usage statistics. Includes user and system time, number of context switches
1478 this thread went through and number of major and minor page faults.
1481 Distribution of I/O depths. Each depth includes everything less than (or equal)
1482 to it, but greater than the previous depth.
1485 Number of read/write requests issued, and number of short read/write requests.
1488 Distribution of I/O completion latencies. The numbers follow the same pattern
1492 The group statistics show:
1497 Number of megabytes I/O performed.
1500 Aggregate bandwidth of threads in the group.
1503 Minimum average bandwidth a thread saw.
1506 Maximum average bandwidth a thread saw.
1509 Shortest runtime of threads in the group.
1512 Longest runtime of threads in the group.
1516 Finally, disk statistics are printed with reads first:
1521 Number of I/Os performed by all groups.
1524 Number of merges in the I/O scheduler.
1527 Number of ticks we kept the disk busy.
1530 Total time spent in the disk queue.
1537 It is also possible to get fio to dump the current output while it is
1538 running, without terminating the job. To do that, send fio the \fBUSR1\fR
1541 If the \fB\-\-minimal\fR option is given, the results will be printed in a
1542 semicolon-delimited format suitable for scripted use - a job description
1543 (if provided) follows on a new line. Note that the first
1544 number in the line is the version number. If the output has to be changed
1545 for some reason, this number will be incremented by 1 to signify that
1546 change. The fields are:
1549 .B terse version, fio version, jobname, groupid, error
1553 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1557 .B min, max, mean, standard deviation
1561 .B min, max, mean, standard deviation
1563 Completion latency percentiles (20 fields):
1565 .B Xth percentile=usec
1569 .B min, max, mean, standard deviation
1573 .B min, max, aggregate percentage of total, mean, standard deviation
1579 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1583 .B min, max, mean, standard deviation
1587 .B min, max, mean, standard deviation
1589 Completion latency percentiles (20 fields):
1591 .B Xth percentile=usec
1595 .B min, max, mean, standard deviation
1599 .B min, max, aggregate percentage of total, mean, standard deviation
1605 .B user, system, context switches, major page faults, minor page faults
1608 IO depth distribution:
1610 .B <=1, 2, 4, 8, 16, 32, >=64
1613 IO latency distribution:
1617 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1621 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1625 Disk utilization (1 for each disk used):
1627 .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
1630 Error Info (dependent on continue_on_error, default off):
1632 .B total # errors, first error code
1635 .B text description (if provided in config - appears on newline)
1638 Normally you would run fio as a stand-alone application on the machine
1639 where the IO workload should be generated. However, it is also possible to
1640 run the frontend and backend of fio separately. This makes it possible to
1641 have a fio server running on the machine(s) where the IO workload should
1642 be running, while controlling it from another machine.
1644 To start the server, you would do:
1646 \fBfio \-\-server=args\fR
1648 on that machine, where args defines what fio listens to. The arguments
1649 are of the form 'type:hostname or IP:port'. 'type' is either 'ip' (or ip4)
1650 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain
1651 socket. 'hostname' is either a hostname or IP address, and 'port' is the port to
1652 listen to (only valid for TCP/IP, not a local socket). Some examples:
1656 Start a fio server, listening on all interfaces on the default port (8765).
1658 2) fio \-\-server=ip:hostname,4444
1660 Start a fio server, listening on IP belonging to hostname and on port 4444.
1662 3) fio \-\-server=ip6:::1,4444
1664 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
1666 4) fio \-\-server=,4444
1668 Start a fio server, listening on all interfaces on port 4444.
1670 5) fio \-\-server=1.2.3.4
1672 Start a fio server, listening on IP 1.2.3.4 on the default port.
1674 6) fio \-\-server=sock:/tmp/fio.sock
1676 Start a fio server, listening on the local socket /tmp/fio.sock.
1678 When a server is running, you can connect to it from a client. The client
1681 fio \-\-local-args \-\-client=server \-\-remote-args <job file(s)>
1683 where \-\-local-args are arguments that are local to the client where it is
1684 running, 'server' is the connect string, and \-\-remote-args and <job file(s)>
1685 are sent to the server. The 'server' string follows the same format as it
1686 does on the server side, to allow IP/hostname/socket and port strings.
1687 You can connect to multiple clients as well, to do that you could run:
1689 fio \-\-client=server2 \-\-client=server2 <job file(s)>
1693 was written by Jens Axboe <jens.axboe@oracle.com>,
1694 now Jens Axboe <jaxboe@fusionio.com>.
1696 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1697 on documentation by Jens Axboe.
1698 .SH "REPORTING BUGS"
1699 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1702 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1704 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob