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 Print statistics in selected mode AND terse, semicolon-delimited format.
39 Display version information and exit.
41 .BI \-\-terse\-version \fR=\fPversion
42 Set terse version output format (Current version 3, or older version 2).
45 Display usage information and exit.
48 Perform test and validation of internal CPU clock
50 .BI \-\-crctest[\fR=\fPtest]
51 Test the speed of the builtin checksumming functions. If no argument is given,
52 all of them are tested. Or a comma separated list can be passed, in which
53 case the given ones are tested.
55 .BI \-\-cmdhelp \fR=\fPcommand
56 Print help information for \fIcommand\fR. May be `all' for all commands.
58 .BI \-\-enghelp \fR=\fPioengine[,command]
59 List all commands defined by \fIioengine\fR, or print help for \fIcommand\fR defined by \fIioengine\fR.
61 .BI \-\-showcmd \fR=\fPjobfile
62 Convert \fIjobfile\fR to a set of command-line options.
64 .BI \-\-eta \fR=\fPwhen
65 Specifies when real-time ETA estimate should be printed. \fIwhen\fR may
66 be one of `always', `never' or `auto'.
68 .BI \-\-eta\-newline \fR=\fPtime
69 Force an ETA newline for every `time` period passed.
71 .BI \-\-status\-interval \fR=\fPtime
72 Report full output status every `time` period passed.
75 Turn on safety read-only checks, preventing any attempted write.
77 .BI \-\-section \fR=\fPsec
78 Only run section \fIsec\fR from job file. This option can be used multiple times to add more sections to run.
80 .BI \-\-alloc\-size \fR=\fPkb
81 Set the internal smalloc pool size to \fIkb\fP kilobytes.
83 .BI \-\-warnings\-fatal
84 All fio parser warnings are fatal, causing fio to exit with an error.
86 .BI \-\-max\-jobs \fR=\fPnr
87 Set the maximum allowed number of jobs (threads/processes) to support.
89 .BI \-\-server \fR=\fPargs
90 Start a backend server, with \fIargs\fP specifying what to listen to. See client/server section.
92 .BI \-\-daemonize \fR=\fPpidfile
93 Background a fio server, writing the pid to the given pid file.
95 .BI \-\-client \fR=\fPhost
96 Instead of running the jobs locally, send and run them on the given host.
98 .BI \-\-idle\-prof \fR=\fPoption
99 Report cpu idleness on a system or percpu basis (\fIoption\fP=system,percpu) or run unit work calibration only (\fIoption\fP=calibrate).
100 .SH "JOB FILE FORMAT"
101 Job files are in `ini' format. They consist of one or more
102 job definitions, which begin with a job name in square brackets and
103 extend to the next job name. The job name can be any ASCII string
104 except `global', which has a special meaning. Following the job name is
105 a sequence of zero or more parameters, one per line, that define the
106 behavior of the job. Any line starting with a `;' or `#' character is
107 considered a comment and ignored.
109 If \fIjobfile\fR is specified as `-', the job file will be read from
112 The global section contains default parameters for jobs specified in the
113 job file. A job is only affected by global sections residing above it,
114 and there may be any number of global sections. Specific job definitions
115 may override any parameter set in global sections.
118 Some parameters may take arguments of a specific type.
119 Anywhere a numeric value is required, an arithmetic expression may be used,
120 provided it is surrounded by parentheses. Supported operators are:
128 .B multiplication (*)
134 .B exponentiation (^)
138 For time values in expressions, units are microseconds by default. This is
139 different than for time values not in expressions (not enclosed in
140 parentheses). The types used are:
143 String: a sequence of alphanumeric characters.
146 SI integer: a whole number, possibly containing a suffix denoting the base unit
147 of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting
148 kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5)
149 respectively. If prefixed with '0x', the value is assumed to be base 16
150 (hexadecimal). A suffix may include a trailing 'b', for instance 'kb' is
151 identical to 'k'. You can specify a base 10 value by using 'KiB', 'MiB','GiB',
152 etc. This is useful for disk drives where values are often given in base 10
153 values. Specifying '30GiB' will get you 30*1000^3 bytes.
154 When specifying times the default suffix meaning changes, still denoting the
155 base unit of the value, but accepted suffixes are 'D' (days), 'H' (hours), 'M'
156 (minutes), 'S' Seconds, 'ms' (or msec) milli seconds, 'us' (or 'usec') micro
157 seconds. Time values without a unit specify seconds.
158 The suffixes are not case sensitive.
161 Boolean: a true or false value. `0' denotes false, `1' denotes true.
164 Integer range: a range of integers specified in the format
165 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
166 \fIupper\fR may contain a suffix as described above. If an option allows two
167 sets of ranges, they are separated with a `,' or `/' character. For example:
171 List of floating numbers: A list of floating numbers, separated by
176 May be used to override the job name. On the command line, this parameter
177 has the special purpose of signalling the start of a new job.
179 .BI description \fR=\fPstr
180 Human-readable description of the job. It is printed when the job is run, but
181 otherwise has no special purpose.
183 .BI directory \fR=\fPstr
184 Prefix filenames with this directory. Used to place files in a location other
186 You can specify a number of directories by separating the names with a ':'
187 character. These directories will be assigned equally distributed to job clones
188 creates with \fInumjobs\fR as long as they are using generated filenames.
189 If specific \fIfilename(s)\fR are set fio will use the first listed directory,
190 and thereby matching the \fIfilename\fR semantic which generates a file each
191 clone if not specified, but let all clones use the same if set. See
192 \fIfilename\fR for considerations regarding escaping certain characters on
195 .BI filename \fR=\fPstr
197 normally makes up a file name based on the job name, thread number, and file
198 number. If you want to share files between threads in a job or several jobs,
199 specify a \fIfilename\fR for each of them to override the default.
200 If the I/O engine is file-based, you can specify
201 a number of files by separating the names with a `:' character. `\-' is a
202 reserved name, meaning stdin or stdout, depending on the read/write direction
203 set. On Windows, disk devices are accessed as \\.\PhysicalDrive0 for the first
204 device, \\.\PhysicalDrive1 for the second etc. Note: Windows and FreeBSD
205 prevent write access to areas of the disk containing in-use data
206 (e.g. filesystems). If the wanted filename does need to include a colon, then
207 escape that with a '\\' character. For instance, if the filename is
208 "/dev/dsk/foo@3,0:c", then you would use filename="/dev/dsk/foo@3,0\\:c".
210 .BI filename_format \fR=\fPstr
211 If sharing multiple files between jobs, it is usually necessary to have
212 fio generate the exact names that you want. By default, fio will name a file
213 based on the default file format specification of
214 \fBjobname.jobnumber.filenumber\fP. With this option, that can be
215 customized. Fio will recognize and replace the following keywords in this
221 The name of the worker thread or process.
224 The incremental number of the worker thread or process.
227 The incremental number of the file for that worker thread or process.
230 To have dependent jobs share a set of files, this option can be set to
231 have fio generate filenames that are shared between the two. For instance,
232 if \fBtestfiles.$filenum\fR is specified, file number 4 for any job will
233 be named \fBtestfiles.4\fR. The default of \fB$jobname.$jobnum.$filenum\fR
234 will be used if no other format specifier is given.
238 .BI lockfile \fR=\fPstr
239 Fio defaults to not locking any files before it does IO to them. If a file or
240 file descriptor is shared, fio can serialize IO to that file to make the end
241 result consistent. This is usual for emulating real workloads that share files.
247 No locking. This is the default.
250 Only one thread or process may do IO at a time, excluding all others.
253 Read-write locking on the file. Many readers may access the file at the same
254 time, but writes get exclusive access.
258 .BI opendir \fR=\fPstr
259 Recursively open any files below directory \fIstr\fR.
261 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
262 Type of I/O pattern. Accepted values are:
273 Sequential trim (Linux block devices only).
282 Random trim (Linux block devices only).
285 Mixed sequential reads and writes.
288 Mixed random reads and writes.
291 For mixed I/O, the default split is 50/50. For certain types of io the result
292 may still be skewed a bit, since the speed may be different. It is possible to
293 specify a number of IO's to do before getting a new offset, this is done by
294 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
295 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
296 value of 8. If the postfix is used with a sequential IO pattern, then the value
297 specified will be added to the generated offset for each IO. For instance,
298 using \fBrw=write:4k\fR will skip 4k for every write. It turns sequential IO
299 into sequential IO with holes. See the \fBrw_sequencer\fR option.
302 .BI rw_sequencer \fR=\fPstr
303 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
304 then this option controls how that number modifies the IO offset being
305 generated. Accepted values are:
310 Generate sequential offset
313 Generate the same offset
316 \fBsequential\fR is only useful for random IO, where fio would normally
317 generate a new random offset for every IO. If you append eg 8 to randread, you
318 would get a new random offset for every 8 IO's. The result would be a seek for
319 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
320 that. As sequential IO is already sequential, setting \fBsequential\fR for that
321 would not result in any differences. \fBidentical\fR behaves in a similar
322 fashion, except it sends the same offset 8 number of times before generating a
327 .BI kb_base \fR=\fPint
328 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
329 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
330 reasons. Allowed values are 1024 or 1000, with 1024 being the default.
332 .BI unified_rw_reporting \fR=\fPbool
333 Fio normally reports statistics on a per data direction basis, meaning that
334 read, write, and trim are accounted and reported separately. If this option is
335 set fio sums the results and reports them as "mixed" instead.
337 .BI randrepeat \fR=\fPbool
338 Seed the random number generator used for random I/O patterns in a predictable
339 way so the pattern is repeatable across runs. Default: true.
341 .BI allrandrepeat \fR=\fPbool
342 Seed all random number generators in a predictable way so results are
343 repeatable across runs. Default: false.
345 .BI randseed \fR=\fPint
346 Seed the random number generators based on this seed value, to be able to
347 control what sequence of output is being generated. If not set, the random
348 sequence depends on the \fBrandrepeat\fR setting.
350 .BI use_os_rand \fR=\fPbool
351 Fio can either use the random generator supplied by the OS to generate random
352 offsets, or it can use its own internal generator (based on Tausworthe).
353 Default is to use the internal generator, which is often of better quality and
354 faster. Default: false.
356 .BI fallocate \fR=\fPstr
357 Whether pre-allocation is performed when laying down files. Accepted values
363 Do not pre-allocate space.
366 Pre-allocate via \fBposix_fallocate\fR\|(3).
369 Pre-allocate via \fBfallocate\fR\|(2) with FALLOC_FL_KEEP_SIZE set.
372 Backward-compatible alias for 'none'.
375 Backward-compatible alias for 'posix'.
378 May not be available on all supported platforms. 'keep' is only
379 available on Linux. If using ZFS on Solaris this must be set to 'none'
380 because ZFS doesn't support it. Default: 'posix'.
383 .BI fadvise_hint \fR=\fPbool
384 Use \fBposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
385 are likely to be issued. Default: true.
388 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
389 been transferred, unless limited by other options (\fBruntime\fR, for instance).
390 Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be
391 divided between the available files for the job. If not set, fio will use the
392 full size of the given files or devices. If the files do not exist, size
393 must be given. It is also possible to give size as a percentage between 1 and
394 100. If size=20% is given, fio will use 20% of the full size of the given
397 .BI io_limit \fR=\fPint
398 Normally fio operates within the region set by \fBsize\fR, which means that
399 the \fBsize\fR option sets both the region and size of IO to be performed.
400 Sometimes that is not what you want. With this option, it is possible to
401 define just the amount of IO that fio should do. For instance, if \fBsize\fR
402 is set to 20G and \fBio_limit\fR is set to 5G, fio will perform IO within
403 the first 20G but exit when 5G have been done.
405 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
406 Sets size to something really large and waits for ENOSPC (no space left on
407 device) as the terminating condition. Only makes sense with sequential write.
408 For a read workload, the mount point will be filled first then IO started on
409 the result. This option doesn't make sense if operating on a raw device node,
410 since the size of that is already known by the file system. Additionally,
411 writing beyond end-of-device will not return ENOSPC there.
413 .BI filesize \fR=\fPirange
414 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
415 for files at random within the given range, limited to \fBsize\fR in total (if
416 that is given). If \fBfilesize\fR is not specified, each created file is the
419 .BI file_append \fR=\fPbool
420 Perform IO after the end of the file. Normally fio will operate within the
421 size of a file. If this option is set, then fio will append to the file
422 instead. This has identical behavior to setting \fRoffset\fP to the size
423 of a file. This option is ignored on non-regular files.
425 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
426 Block size for I/O units. Default: 4k. Values for reads, writes, and trims
427 can be specified separately in the format \fIread\fR,\fIwrite\fR,\fItrim\fR
428 either of which may be empty to leave that value at its default. If a trailing
429 comma isn't given, the remainder will inherit the last value set.
431 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
432 Specify a range of I/O block sizes. The issued I/O unit will always be a
433 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
434 to both reads and writes if only one range is given, but can be specified
435 separately with a comma separating the values. Example: bsrange=1k-4k,2k-8k.
436 Also (see \fBblocksize\fR).
438 .BI bssplit \fR=\fPstr
439 This option allows even finer grained control of the block sizes issued,
440 not just even splits between them. With this option, you can weight various
441 block sizes for exact control of the issued IO for a job that has mixed
442 block sizes. The format of the option is bssplit=blocksize/percentage,
443 optionally adding as many definitions as needed separated by a colon.
444 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
445 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
446 splits to reads and writes. The format is identical to what the
447 \fBbs\fR option accepts, the read and write parts are separated with a
450 .B blocksize_unaligned\fR,\fP bs_unaligned
451 If set, any size in \fBblocksize_range\fR may be used. This typically won't
452 work with direct I/O, as that normally requires sector alignment.
454 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
455 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
456 the minimum blocksize given. Minimum alignment is typically 512b
457 for using direct IO, though it usually depends on the hardware block size.
458 This option is mutually exclusive with using a random map for files, so it
459 will turn off that option.
461 .BI bs_is_seq_rand \fR=\fPbool
462 If this option is set, fio will use the normal read,write blocksize settings as
463 sequential,random instead. Any random read or write will use the WRITE
464 blocksize settings, and any sequential read or write will use the READ
468 Initialize buffers with all zeros. Default: fill buffers with random data.
469 The resulting IO buffers will not be completely zeroed, unless
470 \fPscramble_buffers\fR is also turned off.
473 If this option is given, fio will refill the IO buffers on every submit. The
474 default is to only fill it at init time and reuse that data. Only makes sense
475 if zero_buffers isn't specified, naturally. If data verification is enabled,
476 refill_buffers is also automatically enabled.
478 .BI scramble_buffers \fR=\fPbool
479 If \fBrefill_buffers\fR is too costly and the target is using data
480 deduplication, then setting this option will slightly modify the IO buffer
481 contents to defeat normal de-dupe attempts. This is not enough to defeat
482 more clever block compression attempts, but it will stop naive dedupe
483 of blocks. Default: true.
485 .BI buffer_compress_percentage \fR=\fPint
486 If this is set, then fio will attempt to provide IO buffer content (on WRITEs)
487 that compress to the specified level. Fio does this by providing a mix of
488 random data and zeroes. Note that this is per block size unit, for file/disk
489 wide compression level that matches this setting, you'll also want to set
490 \fBrefill_buffers\fR.
492 .BI buffer_compress_chunk \fR=\fPint
493 See \fBbuffer_compress_percentage\fR. This setting allows fio to manage how
494 big the ranges of random data and zeroed data is. Without this set, fio will
495 provide \fBbuffer_compress_percentage\fR of blocksize random data, followed by
496 the remaining zeroed. With this set to some chunk size smaller than the block
497 size, fio can alternate random and zeroed data throughout the IO buffer.
499 .BI buffer_pattern \fR=\fPstr
500 If set, fio will fill the IO buffers with this pattern. If not set, the contents
501 of IO buffers is defined by the other options related to buffer contents. The
502 setting can be any pattern of bytes, and can be prefixed with 0x for hex
503 values. It may also be a string, where the string must then be wrapped with
506 .BI dedupe_percentage \fR=\fPint
507 If set, fio will generate this percentage of identical buffers when writing.
508 These buffers will be naturally dedupable. The contents of the buffers depend
509 on what other buffer compression settings have been set. It's possible to have
510 the individual buffers either fully compressible, or not at all. This option
511 only controls the distribution of unique buffers.
513 .BI nrfiles \fR=\fPint
514 Number of files to use for this job. Default: 1.
516 .BI openfiles \fR=\fPint
517 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
519 .BI file_service_type \fR=\fPstr
520 Defines how files to service are selected. The following types are defined:
525 Choose a file at random.
528 Round robin over opened files (default).
531 Do each file in the set sequentially.
534 The number of I/Os to issue before switching to a new file can be specified by
535 appending `:\fIint\fR' to the service type.
538 .BI ioengine \fR=\fPstr
539 Defines how the job issues I/O. The following types are defined:
544 Basic \fBread\fR\|(2) or \fBwrite\fR\|(2) I/O. \fBfseek\fR\|(2) is used to
545 position the I/O location.
548 Basic \fBpread\fR\|(2) or \fBpwrite\fR\|(2) I/O.
551 Basic \fBreadv\fR\|(2) or \fBwritev\fR\|(2) I/O. Will emulate queuing by
552 coalescing adjacent IOs into a single submission.
555 Basic \fBpreadv\fR\|(2) or \fBpwritev\fR\|(2) I/O.
558 Linux native asynchronous I/O. This ioengine defines engine specific options.
561 POSIX asynchronous I/O using \fBaio_read\fR\|(3) and \fBaio_write\fR\|(3).
564 Solaris native asynchronous I/O.
567 Windows native asynchronous I/O.
570 File is memory mapped with \fBmmap\fR\|(2) and data copied using
574 \fBsplice\fR\|(2) is used to transfer the data and \fBvmsplice\fR\|(2) to
575 transfer data from user-space to the kernel.
578 Use the syslet system calls to make regular read/write asynchronous.
581 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
582 the target is an sg character device, we use \fBread\fR\|(2) and
583 \fBwrite\fR\|(2) for asynchronous I/O.
586 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
587 itself and for debugging and testing purposes.
590 Transfer over the network. The protocol to be used can be defined with the
591 \fBprotocol\fR parameter. Depending on the protocol, \fBfilename\fR,
592 \fBhostname\fR, \fBport\fR, or \fBlisten\fR must be specified.
593 This ioengine defines engine specific options.
596 Like \fBnet\fR, but uses \fBsplice\fR\|(2) and \fBvmsplice\fR\|(2) to map data
597 and send/receive. This ioengine defines engine specific options.
600 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
601 \fBcpucycles\fR parameters.
604 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
605 approach to asynchronous I/O.
607 See <http://www.xmailserver.org/guasi\-lib.html>.
610 The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ)
611 and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols.
614 Loads an external I/O engine object file. Append the engine filename as
618 IO engine that does regular linux native fallocate call to simulate data
619 transfer as fio ioengine
621 DDIR_READ does fallocate(,mode = FALLOC_FL_KEEP_SIZE,)
623 DIR_WRITE does fallocate(,mode = 0)
625 DDIR_TRIM does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE)
628 IO engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate defragment activity
629 request to DDIR_WRITE event
632 IO engine supporting direct access to Ceph Rados Block Devices (RBD) via librbd
633 without the need to use the kernel rbd driver. This ioengine defines engine specific
637 Using Glusterfs libgfapi sync interface to direct access to Glusterfs volumes without
638 having to go through FUSE. This ioengine defines engine specific
642 Using Glusterfs libgfapi async interface to direct access to Glusterfs volumes without
643 having to go through FUSE. This ioengine defines engine specific
647 Read and write through Hadoop (HDFS). The \fBfilename\fR option is used to
648 specify host,port of the hdfs name-node to connect. This engine interprets
649 offsets a little differently. In HDFS, files once created cannot be modified.
650 So random writes are not possible. To imitate this, libhdfs engine expects
651 bunch of small files to be created over HDFS, and engine will randomly pick a
652 file out of those files based on the offset generated by fio backend. (see the
653 example job file to create such files, use rw=write option). Please note, you
654 might want to set necessary environment variables to work with hdfs/libhdfs
660 .BI iodepth \fR=\fPint
661 Number of I/O units to keep in flight against the file. Note that increasing
662 iodepth beyond 1 will not affect synchronous ioengines (except for small
663 degress when verify_async is in use). Even async engines may impose OS
664 restrictions causing the desired depth not to be achieved. This may happen on
665 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
666 not async on that OS. Keep an eye on the IO depth distribution in the
667 fio output to verify that the achieved depth is as expected. Default: 1.
669 .BI iodepth_batch \fR=\fPint
670 Number of I/Os to submit at once. Default: \fBiodepth\fR.
672 .BI iodepth_batch_complete \fR=\fPint
673 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
674 means that we'll ask for a minimum of 1 IO in the retrieval process from the
675 kernel. The IO retrieval will go on until we hit the limit set by
676 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
677 completed events before queuing more IO. This helps reduce IO latency, at the
678 cost of more retrieval system calls.
680 .BI iodepth_low \fR=\fPint
681 Low watermark indicating when to start filling the queue again. Default:
684 .BI direct \fR=\fPbool
685 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
687 .BI atomic \fR=\fPbool
688 If value is true, attempt to use atomic direct IO. Atomic writes are guaranteed
689 to be stable once acknowledged by the operating system. Only Linux supports
692 .BI buffered \fR=\fPbool
693 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
696 .BI offset \fR=\fPint
697 Offset in the file to start I/O. Data before the offset will not be touched.
699 .BI offset_increment \fR=\fPint
700 If this is provided, then the real offset becomes the
701 offset + offset_increment * thread_number, where the thread number is a
702 counter that starts at 0 and is incremented for each sub-job (i.e. when
703 numjobs option is specified). This option is useful if there are several jobs
704 which are intended to operate on a file in parallel disjoint segments, with
705 even spacing between the starting points.
707 .BI number_ios \fR=\fPint
708 Fio will normally perform IOs until it has exhausted the size of the region
709 set by \fBsize\fR, or if it exhaust the allocated time (or hits an error
710 condition). With this setting, the range/size can be set independently of
711 the number of IOs to perform. When fio reaches this number, it will exit
712 normally and report status. Note that this does not extend the amount
713 of IO that will be done, it will only stop fio if this condition is met
714 before other end-of-job criteria.
717 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
718 0, don't sync. Default: 0.
720 .BI fdatasync \fR=\fPint
721 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
722 data parts of the file. Default: 0.
724 .BI write_barrier \fR=\fPint
725 Make every Nth write a barrier write.
727 .BI sync_file_range \fR=\fPstr:int
728 Use \fBsync_file_range\fR\|(2) for every \fRval\fP number of write operations. Fio will
729 track range of writes that have happened since the last \fBsync_file_range\fR\|(2) call.
730 \fRstr\fP can currently be one or more of:
734 SYNC_FILE_RANGE_WAIT_BEFORE
737 SYNC_FILE_RANGE_WRITE
740 SYNC_FILE_RANGE_WRITE
744 So if you do sync_file_range=wait_before,write:8, fio would use
745 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
746 Also see the \fBsync_file_range\fR\|(2) man page. This option is Linux specific.
748 .BI overwrite \fR=\fPbool
749 If writing, setup the file first and do overwrites. Default: false.
751 .BI end_fsync \fR=\fPbool
752 Sync file contents when a write stage has completed. Default: false.
754 .BI fsync_on_close \fR=\fPbool
755 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
756 it will happen on every close, not just at the end of the job. Default: false.
758 .BI rwmixread \fR=\fPint
759 Percentage of a mixed workload that should be reads. Default: 50.
761 .BI rwmixwrite \fR=\fPint
762 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
763 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
764 overrides the first. This may interfere with a given rate setting, if fio is
765 asked to limit reads or writes to a certain rate. If that is the case, then
766 the distribution may be skewed. Default: 50.
768 .BI random_distribution \fR=\fPstr:float
769 By default, fio will use a completely uniform random distribution when asked
770 to perform random IO. Sometimes it is useful to skew the distribution in
771 specific ways, ensuring that some parts of the data is more hot than others.
772 Fio includes the following distribution models:
776 Uniform random distribution
786 When using a zipf or pareto distribution, an input value is also needed to
787 define the access pattern. For zipf, this is the zipf theta. For pareto,
788 it's the pareto power. Fio includes a test program, genzipf, that can be
789 used visualize what the given input values will yield in terms of hit rates.
790 If you wanted to use zipf with a theta of 1.2, you would use
791 random_distribution=zipf:1.2 as the option. If a non-uniform model is used,
792 fio will disable use of the random map.
794 .BI percentage_random \fR=\fPint
795 For a random workload, set how big a percentage should be random. This defaults
796 to 100%, in which case the workload is fully random. It can be set from
797 anywhere from 0 to 100. Setting it to 0 would make the workload fully
798 sequential. It is possible to set different values for reads, writes, and
799 trim. To do so, simply use a comma separated list. See \fBblocksize\fR.
802 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
803 this parameter is given, a new offset will be chosen without looking at past
804 I/O history. This parameter is mutually exclusive with \fBverify\fR.
806 .BI softrandommap \fR=\fPbool
807 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
808 fails to allocate the map, if this option is set it will continue without a
809 random block map. As coverage will not be as complete as with random maps, this
810 option is disabled by default.
812 .BI random_generator \fR=\fPstr
813 Fio supports the following engines for generating IO offsets for random IO:
817 Strong 2^88 cycle random number generator
820 Linear feedback shift register generator
824 Tausworthe is a strong random number generator, but it requires tracking on the
825 side if we want to ensure that blocks are only read or written once. LFSR
826 guarantees that we never generate the same offset twice, and it's also less
827 computationally expensive. It's not a true random generator, however, though
828 for IO purposes it's typically good enough. LFSR only works with single block
829 sizes, not with workloads that use multiple block sizes. If used with such a
830 workload, fio may read or write some blocks multiple times.
833 Run job with given nice value. See \fBnice\fR\|(2).
836 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
839 .BI prioclass \fR=\fPint
840 Set I/O priority class. See \fBionice\fR\|(1).
842 .BI thinktime \fR=\fPint
843 Stall job for given number of microseconds between issuing I/Os.
845 .BI thinktime_spin \fR=\fPint
846 Pretend to spend CPU time for given number of microseconds, sleeping the rest
847 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
849 .BI thinktime_blocks \fR=\fPint
850 Only valid if thinktime is set - control how many blocks to issue, before
851 waiting \fBthinktime\fR microseconds. If not set, defaults to 1 which will
852 make fio wait \fBthinktime\fR microseconds after every block. This
853 effectively makes any queue depth setting redundant, since no more than 1 IO
854 will be queued before we have to complete it and do our thinktime. In other
855 words, this setting effectively caps the queue depth if the latter is larger.
859 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
860 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
861 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
862 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
863 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
864 limit writes (to 500KB/sec), the latter will only limit reads.
866 .BI ratemin \fR=\fPint
867 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
868 Failing to meet this requirement will cause the job to exit. The same format
869 as \fBrate\fR is used for read vs write separation.
871 .BI rate_iops \fR=\fPint
872 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
873 specified independently of bandwidth. The same format as \fBrate\fR is used for
874 read vs write separation. If \fBblocksize\fR is a range, the smallest block
875 size is used as the metric.
877 .BI rate_iops_min \fR=\fPint
878 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
879 is used for read vs write separation.
881 .BI ratecycle \fR=\fPint
882 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
883 milliseconds. Default: 1000ms.
885 .BI latency_target \fR=\fPint
886 If set, fio will attempt to find the max performance point that the given
887 workload will run at while maintaining a latency below this target. The
888 values is given in microseconds. See \fBlatency_window\fR and
889 \fBlatency_percentile\fR.
891 .BI latency_window \fR=\fPint
892 Used with \fBlatency_target\fR to specify the sample window that the job
893 is run at varying queue depths to test the performance. The value is given
896 .BI latency_percentile \fR=\fPfloat
897 The percentage of IOs that must fall within the criteria specified by
898 \fBlatency_target\fR and \fBlatency_window\fR. If not set, this defaults
899 to 100.0, meaning that all IOs must be equal or below to the value set
900 by \fBlatency_target\fR.
902 .BI max_latency \fR=\fPint
903 If set, fio will exit the job if it exceeds this maximum latency. It will exit
906 .BI cpumask \fR=\fPint
907 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
908 may run on. See \fBsched_setaffinity\fR\|(2).
910 .BI cpus_allowed \fR=\fPstr
911 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
913 .BI cpus_allowed_policy \fR=\fPstr
914 Set the policy of how fio distributes the CPUs specified by \fBcpus_allowed\fR
915 or \fBcpumask\fR. Two policies are supported:
920 All jobs will share the CPU set specified.
923 Each job will get a unique CPU from the CPU set.
926 \fBshared\fR is the default behaviour, if the option isn't specified. If
927 \fBsplit\fR is specified, then fio will assign one cpu per job. If not enough
928 CPUs are given for the jobs listed, then fio will roundrobin the CPUs in
933 .BI numa_cpu_nodes \fR=\fPstr
934 Set this job running on specified NUMA nodes' CPUs. The arguments allow
935 comma delimited list of cpu numbers, A-B ranges, or 'all'.
937 .BI numa_mem_policy \fR=\fPstr
938 Set this job's memory policy and corresponding NUMA nodes. Format of
942 .B <mode>[:<nodelist>]
945 is one of the following memory policy:
947 .B default, prefer, bind, interleave, local
950 For \fBdefault\fR and \fBlocal\fR memory policy, no \fBnodelist\fR is
951 needed to be specified. For \fBprefer\fR, only one node is
952 allowed. For \fBbind\fR and \fBinterleave\fR, \fBnodelist\fR allows
953 comma delimited list of numbers, A-B ranges, or 'all'.
955 .BI startdelay \fR=\fPirange
956 Delay start of job for the specified number of seconds. Supports all time
957 suffixes to allow specification of hours, minutes, seconds and
958 milliseconds - seconds are the default if a unit is ommited.
959 Can be given as a range which causes each thread to choose randomly out of the
962 .BI runtime \fR=\fPint
963 Terminate processing after the specified number of seconds.
966 If given, run for the specified \fBruntime\fR duration even if the files are
967 completely read or written. The same workload will be repeated as many times
968 as \fBruntime\fR allows.
970 .BI ramp_time \fR=\fPint
971 If set, fio will run the specified workload for this amount of time before
972 logging any performance numbers. Useful for letting performance settle before
973 logging results, thus minimizing the runtime required for stable results. Note
974 that the \fBramp_time\fR is considered lead in time for a job, thus it will
975 increase the total runtime if a special timeout or runtime is specified.
977 .BI invalidate \fR=\fPbool
978 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
981 Use synchronous I/O for buffered writes. For the majority of I/O engines,
982 this means using O_SYNC. Default: false.
984 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
985 Allocation method for I/O unit buffer. Allowed values are:
990 Allocate memory with \fBmalloc\fR\|(3).
993 Use shared memory buffers allocated through \fBshmget\fR\|(2).
996 Same as \fBshm\fR, but use huge pages as backing.
999 Use \fBmmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
1000 is given after the option in the format `:\fIfile\fR'.
1003 Same as \fBmmap\fR, but use huge files as backing.
1006 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
1007 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
1008 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
1009 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
1010 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
1011 and the documentation for that. Normally you just need to echo an appropriate
1012 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
1016 .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint
1017 This indicates the memory alignment of the IO memory buffers. Note that the
1018 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
1019 the alignment of the following buffers are given by the \fBbs\fR used. In
1020 other words, if using a \fBbs\fR that is a multiple of the page sized in the
1021 system, all buffers will be aligned to this value. If using a \fBbs\fR that
1022 is not page aligned, the alignment of subsequent IO memory buffers is the
1023 sum of the \fBiomem_align\fR and \fBbs\fR used.
1025 .BI hugepage\-size \fR=\fPint
1026 Defines the size of a huge page. Must be at least equal to the system setting.
1027 Should be a multiple of 1MB. Default: 4MB.
1030 Terminate all jobs when one finishes. Default: wait for each job to finish.
1032 .BI bwavgtime \fR=\fPint
1033 Average bandwidth calculations over the given time in milliseconds. Default:
1036 .BI iopsavgtime \fR=\fPint
1037 Average IOPS calculations over the given time in milliseconds. Default:
1040 .BI create_serialize \fR=\fPbool
1041 If true, serialize file creation for the jobs. Default: true.
1043 .BI create_fsync \fR=\fPbool
1044 \fBfsync\fR\|(2) data file after creation. Default: true.
1046 .BI create_on_open \fR=\fPbool
1047 If true, the files are not created until they are opened for IO by the job.
1049 .BI create_only \fR=\fPbool
1050 If true, fio will only run the setup phase of the job. If files need to be
1051 laid out or updated on disk, only that will be done. The actual job contents
1054 .BI pre_read \fR=\fPbool
1055 If this is given, files will be pre-read into memory before starting the given
1056 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
1057 pointless to pre-read and then drop the cache. This will only work for IO
1058 engines that are seekable, since they allow you to read the same data
1059 multiple times. Thus it will not work on eg network or splice IO.
1061 .BI unlink \fR=\fPbool
1062 Unlink job files when done. Default: false.
1064 .BI loops \fR=\fPint
1065 Specifies the number of iterations (runs of the same workload) of this job.
1068 .BI verify_only \fR=\fPbool
1069 Do not perform the specified workload, only verify data still matches previous
1070 invocation of this workload. This option allows one to check data multiple
1071 times at a later date without overwriting it. This option makes sense only for
1072 workloads that write data, and does not support workloads with the
1073 \fBtime_based\fR option set.
1075 .BI do_verify \fR=\fPbool
1076 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
1079 .BI verify \fR=\fPstr
1080 Method of verifying file contents after each iteration of the job. Allowed
1085 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1 xxhash
1086 Store appropriate checksum in the header of each block. crc32c-intel is
1087 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
1088 not supported by the system.
1091 Write extra information about each I/O (timestamp, block number, etc.). The
1092 block number is verified. See \fBverify_pattern\fR as well.
1095 Pretend to verify. Used for testing internals.
1098 This option can be used for repeated burn-in tests of a system to make sure
1099 that the written data is also correctly read back. If the data direction given
1100 is a read or random read, fio will assume that it should verify a previously
1101 written file. If the data direction includes any form of write, the verify will
1102 be of the newly written data.
1105 .BI verifysort \fR=\fPbool
1106 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
1107 read them back in a sorted manner. Default: true.
1109 .BI verifysort_nr \fR=\fPint
1110 Pre-load and sort verify blocks for a read workload.
1112 .BI verify_offset \fR=\fPint
1113 Swap the verification header with data somewhere else in the block before
1114 writing. It is swapped back before verifying.
1116 .BI verify_interval \fR=\fPint
1117 Write the verification header for this number of bytes, which should divide
1118 \fBblocksize\fR. Default: \fBblocksize\fR.
1120 .BI verify_pattern \fR=\fPstr
1121 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
1122 with totally random bytes, but sometimes it's interesting to fill with a known
1123 pattern for io verification purposes. Depending on the width of the pattern,
1124 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
1125 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
1126 has to be a hex number that starts with either "0x" or "0X". Use with
1129 .BI verify_fatal \fR=\fPbool
1130 If true, exit the job on the first observed verification failure. Default:
1133 .BI verify_dump \fR=\fPbool
1134 If set, dump the contents of both the original data block and the data block we
1135 read off disk to files. This allows later analysis to inspect just what kind of
1136 data corruption occurred. Off by default.
1138 .BI verify_async \fR=\fPint
1139 Fio will normally verify IO inline from the submitting thread. This option
1140 takes an integer describing how many async offload threads to create for IO
1141 verification instead, causing fio to offload the duty of verifying IO contents
1142 to one or more separate threads. If using this offload option, even sync IO
1143 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
1144 allows them to have IO in flight while verifies are running.
1146 .BI verify_async_cpus \fR=\fPstr
1147 Tell fio to set the given CPU affinity on the async IO verification threads.
1148 See \fBcpus_allowed\fP for the format used.
1150 .BI verify_backlog \fR=\fPint
1151 Fio will normally verify the written contents of a job that utilizes verify
1152 once that job has completed. In other words, everything is written then
1153 everything is read back and verified. You may want to verify continually
1154 instead for a variety of reasons. Fio stores the meta data associated with an
1155 IO block in memory, so for large verify workloads, quite a bit of memory would
1156 be used up holding this meta data. If this option is enabled, fio will write
1157 only N blocks before verifying these blocks.
1159 .BI verify_backlog_batch \fR=\fPint
1160 Control how many blocks fio will verify if verify_backlog is set. If not set,
1161 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
1162 read back and verified). If \fBverify_backlog_batch\fR is less than
1163 \fBverify_backlog\fR then not all blocks will be verified, if
1164 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
1165 will be verified more than once.
1167 .BI trim_percentage \fR=\fPint
1168 Number of verify blocks to discard/trim.
1170 .BI trim_verify_zero \fR=\fPbool
1171 Verify that trim/discarded blocks are returned as zeroes.
1173 .BI trim_backlog \fR=\fPint
1174 Trim after this number of blocks are written.
1176 .BI trim_backlog_batch \fR=\fPint
1177 Trim this number of IO blocks.
1179 .BI experimental_verify \fR=\fPbool
1180 Enable experimental verification.
1182 .B stonewall "\fR,\fP wait_for_previous"
1183 Wait for preceding jobs in the job file to exit before starting this one.
1184 \fBstonewall\fR implies \fBnew_group\fR.
1187 Start a new reporting group. If not given, all jobs in a file will be part
1188 of the same reporting group, unless separated by a stonewall.
1190 .BI numjobs \fR=\fPint
1191 Number of clones (processes/threads performing the same workload) of this job.
1195 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
1199 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
1200 with \fBfork\fR\|(2).
1202 .BI zonesize \fR=\fPint
1203 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
1205 .BI zonerange \fR=\fPint
1206 Give size of an IO zone. See \fBzoneskip\fR.
1208 .BI zoneskip \fR=\fPint
1209 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
1212 .BI write_iolog \fR=\fPstr
1213 Write the issued I/O patterns to the specified file. Specify a separate file
1214 for each job, otherwise the iologs will be interspersed and the file may be
1217 .BI read_iolog \fR=\fPstr
1218 Replay the I/O patterns contained in the specified file generated by
1219 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
1221 .BI replay_no_stall \fR=\fPint
1222 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1223 attempts to respect timing information between I/Os. Enabling
1224 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
1225 still respecting ordering.
1227 .BI replay_redirect \fR=\fPstr
1228 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1229 is to replay the IOPS onto the major/minor device that each IOP was recorded
1230 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
1231 single specified device regardless of the device it was recorded from.
1233 .BI write_bw_log \fR=\fPstr
1234 If given, write a bandwidth log of the jobs in this job file. Can be used to
1235 store data of the bandwidth of the jobs in their lifetime. The included
1236 fio_generate_plots script uses gnuplot to turn these text files into nice
1237 graphs. See \fBwrite_lat_log\fR for behaviour of given filename. For this
1238 option, the postfix is _bw.x.log, where x is the index of the job (1..N,
1239 where N is the number of jobs)
1241 .BI write_lat_log \fR=\fPstr
1242 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
1243 filename is given with this option, the default filename of
1244 "jobname_type.x.log" is used, where x is the index of the job (1..N, where
1245 N is the number of jobs). Even if the filename is given, fio will still
1246 append the type of log.
1248 .BI write_iops_log \fR=\fPstr
1249 Same as \fBwrite_bw_log\fR, but writes IOPS. If no filename is given with this
1250 option, the default filename of "jobname_type.x.log" is used, where x is the
1251 index of the job (1..N, where N is the number of jobs). Even if the filename
1252 is given, fio will still append the type of log.
1254 .BI log_avg_msec \fR=\fPint
1255 By default, fio will log an entry in the iops, latency, or bw log for every
1256 IO that completes. When writing to the disk log, that can quickly grow to a
1257 very large size. Setting this option makes fio average the each log entry
1258 over the specified period of time, reducing the resolution of the log.
1261 .BI log_offset \fR=\fPbool
1262 If this is set, the iolog options will include the byte offset for the IO
1263 entry as well as the other data values.
1265 .BI log_compression \fR=\fPint
1266 If this is set, fio will compress the IO logs as it goes, to keep the memory
1267 footprint lower. When a log reaches the specified size, that chunk is removed
1268 and compressed in the background. Given that IO logs are fairly highly
1269 compressible, this yields a nice memory savings for longer runs. The downside
1270 is that the compression will consume some background CPU cycles, so it may
1271 impact the run. This, however, is also true if the logging ends up consuming
1272 most of the system memory. So pick your poison. The IO logs are saved
1273 normally at the end of a run, by decompressing the chunks and storing them
1274 in the specified log file. This feature depends on the availability of zlib.
1276 .BI log_store_compressed \fR=\fPbool
1277 If set, and \fBlog\fR_compression is also set, fio will store the log files in
1278 a compressed format. They can be decompressed with fio, using the
1279 \fB\-\-inflate-log\fR command line parameter. The files will be stored with a
1282 .BI disable_lat \fR=\fPbool
1283 Disable measurements of total latency numbers. Useful only for cutting
1284 back the number of calls to \fBgettimeofday\fR\|(2), as that does impact performance at
1285 really high IOPS rates. Note that to really get rid of a large amount of these
1286 calls, this option must be used with disable_slat and disable_bw as well.
1288 .BI disable_clat \fR=\fPbool
1289 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
1291 .BI disable_slat \fR=\fPbool
1292 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
1294 .BI disable_bw_measurement \fR=\fPbool
1295 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
1297 .BI lockmem \fR=\fPint
1298 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
1299 simulate a smaller amount of memory. The amount specified is per worker.
1301 .BI exec_prerun \fR=\fPstr
1302 Before running the job, execute the specified command with \fBsystem\fR\|(3).
1304 Output is redirected in a file called \fBjobname.prerun.txt\fR
1307 .BI exec_postrun \fR=\fPstr
1308 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
1310 Output is redirected in a file called \fBjobname.postrun.txt\fR
1313 .BI ioscheduler \fR=\fPstr
1314 Attempt to switch the device hosting the file to the specified I/O scheduler.
1316 .BI disk_util \fR=\fPbool
1317 Generate disk utilization statistics if the platform supports it. Default: true.
1319 .BI clocksource \fR=\fPstr
1320 Use the given clocksource as the base of timing. The supported options are:
1324 \fBgettimeofday\fR\|(2)
1327 \fBclock_gettime\fR\|(2)
1330 Internal CPU clock source
1334 \fBcpu\fR is the preferred clocksource if it is reliable, as it is very fast
1335 (and fio is heavy on time calls). Fio will automatically use this clocksource
1336 if it's supported and considered reliable on the system it is running on,
1337 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
1338 means supporting TSC Invariant.
1340 .BI gtod_reduce \fR=\fPbool
1341 Enable all of the \fBgettimeofday\fR\|(2) reducing options (disable_clat, disable_slat,
1342 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
1343 \fBgettimeofday\fR\|(2) call count. With this option enabled, we only do about 0.4% of
1344 the gtod() calls we would have done if all time keeping was enabled.
1346 .BI gtod_cpu \fR=\fPint
1347 Sometimes it's cheaper to dedicate a single thread of execution to just getting
1348 the current time. Fio (and databases, for instance) are very intensive on
1349 \fBgettimeofday\fR\|(2) calls. With this option, you can set one CPU aside for doing
1350 nothing but logging current time to a shared memory location. Then the other
1351 threads/processes that run IO workloads need only copy that segment, instead of
1352 entering the kernel with a \fBgettimeofday\fR\|(2) call. The CPU set aside for doing
1353 these time calls will be excluded from other uses. Fio will manually clear it
1354 from the CPU mask of other jobs.
1356 .BI ignore_error \fR=\fPstr
1357 Sometimes you want to ignore some errors during test in that case you can specify
1358 error list for each error type.
1360 ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST
1362 errors for given error type is separated with ':'.
1363 Error may be symbol ('ENOSPC', 'ENOMEM') or an integer.
1365 Example: ignore_error=EAGAIN,ENOSPC:122 .
1367 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from WRITE.
1369 .BI error_dump \fR=\fPbool
1370 If set dump every error even if it is non fatal, true by default. If disabled
1371 only fatal error will be dumped
1373 .BI profile \fR=\fPstr
1374 Select a specific builtin performance test.
1376 .BI cgroup \fR=\fPstr
1377 Add job to this control group. If it doesn't exist, it will be created.
1378 The system must have a mounted cgroup blkio mount point for this to work. If
1379 your system doesn't have it mounted, you can do so with:
1381 # mount \-t cgroup \-o blkio none /cgroup
1383 .BI cgroup_weight \fR=\fPint
1384 Set the weight of the cgroup to this value. See the documentation that comes
1385 with the kernel, allowed values are in the range of 100..1000.
1387 .BI cgroup_nodelete \fR=\fPbool
1388 Normally fio will delete the cgroups it has created after the job completion.
1389 To override this behavior and to leave cgroups around after the job completion,
1390 set cgroup_nodelete=1. This can be useful if one wants to inspect various
1391 cgroup files after job completion. Default: false
1394 Instead of running as the invoking user, set the user ID to this value before
1395 the thread/process does any work.
1398 Set group ID, see \fBuid\fR.
1400 .BI unit_base \fR=\fPint
1401 Base unit for reporting. Allowed values are:
1405 Use auto-detection (default).
1415 .BI flow_id \fR=\fPint
1416 The ID of the flow. If not specified, it defaults to being a global flow. See
1420 Weight in token-based flow control. If this value is used, then there is a
1421 \fBflow counter\fR which is used to regulate the proportion of activity between
1422 two or more jobs. fio attempts to keep this flow counter near zero. The
1423 \fBflow\fR parameter stands for how much should be added or subtracted to the
1424 flow counter on each iteration of the main I/O loop. That is, if one job has
1425 \fBflow=8\fR and another job has \fBflow=-1\fR, then there will be a roughly
1426 1:8 ratio in how much one runs vs the other.
1428 .BI flow_watermark \fR=\fPint
1429 The maximum value that the absolute value of the flow counter is allowed to
1430 reach before the job must wait for a lower value of the counter.
1432 .BI flow_sleep \fR=\fPint
1433 The period of time, in microseconds, to wait after the flow watermark has been
1434 exceeded before retrying operations
1436 .BI clat_percentiles \fR=\fPbool
1437 Enable the reporting of percentiles of completion latencies.
1439 .BI percentile_list \fR=\fPfloat_list
1440 Overwrite the default list of percentiles for completion
1441 latencies. Each number is a floating number in the range (0,100], and
1442 the maximum length of the list is 20. Use ':' to separate the
1443 numbers. For example, \-\-percentile_list=99.5:99.9 will cause fio to
1444 report the values of completion latency below which 99.5% and 99.9% of
1445 the observed latencies fell, respectively.
1446 .SS "Ioengine Parameters List"
1447 Some parameters are only valid when a specific ioengine is in use. These are
1448 used identically to normal parameters, with the caveat that when used on the
1449 command line, they must come after the ioengine.
1451 .BI (cpu)cpuload \fR=\fPint
1452 Attempt to use the specified percentage of CPU cycles.
1454 .BI (cpu)cpuchunks \fR=\fPint
1455 Split the load into cycles of the given time. In microseconds.
1457 .BI (cpu)exit_on_io_done \fR=\fPbool
1458 Detect when IO threads are done, then exit.
1460 .BI (libaio)userspace_reap
1461 Normally, with the libaio engine in use, fio will use
1462 the io_getevents system call to reap newly returned events.
1463 With this flag turned on, the AIO ring will be read directly
1464 from user-space to reap events. The reaping mode is only
1465 enabled when polling for a minimum of 0 events (eg when
1466 iodepth_batch_complete=0).
1468 .BI (net,netsplice)hostname \fR=\fPstr
1469 The host name or IP address to use for TCP or UDP based IO.
1470 If the job is a TCP listener or UDP reader, the hostname is not
1471 used and must be omitted unless it is a valid UDP multicast address.
1473 .BI (net,netsplice)port \fR=\fPint
1474 The TCP or UDP port to bind to or connect to. If this is used with
1475 \fBnumjobs\fR to spawn multiple instances of the same job type, then
1476 this will be the starting port number since fio will use a range of ports.
1478 .BI (net,netsplice)interface \fR=\fPstr
1479 The IP address of the network interface used to send or receive UDP multicast
1482 .BI (net,netsplice)ttl \fR=\fPint
1483 Time-to-live value for outgoing UDP multicast packets. Default: 1
1485 .BI (net,netsplice)nodelay \fR=\fPbool
1486 Set TCP_NODELAY on TCP connections.
1488 .BI (net,netsplice)protocol \fR=\fPstr "\fR,\fP proto" \fR=\fPstr
1489 The network protocol to use. Accepted values are:
1494 Transmission control protocol
1497 Transmission control protocol V6
1500 User datagram protocol
1503 User datagram protocol V6
1509 When the protocol is TCP or UDP, the port must also be given,
1510 as well as the hostname if the job is a TCP listener or UDP
1511 reader. For unix sockets, the normal filename option should be
1512 used and the port is invalid.
1515 .BI (net,netsplice)listen
1516 For TCP network connections, tell fio to listen for incoming
1517 connections rather than initiating an outgoing connection. The
1518 hostname must be omitted if this option is used.
1520 .BI (net, pingpong) \fR=\fPbool
1521 Normally a network writer will just continue writing data, and a network reader
1522 will just consume packets. If pingpong=1 is set, a writer will send its normal
1523 payload to the reader, then wait for the reader to send the same payload back.
1524 This allows fio to measure network latencies. The submission and completion
1525 latencies then measure local time spent sending or receiving, and the
1526 completion latency measures how long it took for the other end to receive and
1527 send back. For UDP multicast traffic pingpong=1 should only be set for a single
1528 reader when multiple readers are listening to the same address.
1530 .BI (net, window_size) \fR=\fPint
1531 Set the desired socket buffer size for the connection.
1533 .BI (net, mss) \fR=\fPint
1534 Set the TCP maximum segment size (TCP_MAXSEG).
1536 .BI (e4defrag,donorname) \fR=\fPstr
1537 File will be used as a block donor (swap extents between files)
1539 .BI (e4defrag,inplace) \fR=\fPint
1540 Configure donor file block allocation strategy
1543 Preallocate donor's file on init
1546 allocate space immediately inside defragment event, and free right after event
1549 .BI (rbd)rbdname \fR=\fPstr
1550 Specifies the name of the RBD.
1552 .BI (rbd)pool \fR=\fPstr
1553 Specifies the name of the Ceph pool containing the RBD.
1555 .BI (rbd)clientname \fR=\fPstr
1556 Specifies the username (without the 'client.' prefix) used to access the Ceph cluster.
1558 While running, \fBfio\fR will display the status of the created jobs. For
1562 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
1565 The characters in the first set of brackets denote the current status of each
1566 threads. The possible values are:
1572 Setup but not started.
1578 Initialized, waiting.
1581 Running, doing sequential reads.
1584 Running, doing random reads.
1587 Running, doing sequential writes.
1590 Running, doing random writes.
1593 Running, doing mixed sequential reads/writes.
1596 Running, doing mixed random reads/writes.
1599 Running, currently waiting for \fBfsync\fR\|(2).
1602 Running, verifying written data.
1605 Exited, not reaped by main thread.
1608 Exited, thread reaped.
1612 The second set of brackets shows the estimated completion percentage of
1613 the current group. The third set shows the read and write I/O rate,
1614 respectively. Finally, the estimated run time of the job is displayed.
1616 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
1617 for each thread, each group of threads, and each disk, in that order.
1619 Per-thread statistics first show the threads client number, group-id, and
1620 error code. The remaining figures are as follows:
1624 Number of megabytes of I/O performed.
1627 Average data rate (bandwidth).
1633 Submission latency minimum, maximum, average and standard deviation. This is
1634 the time it took to submit the I/O.
1637 Completion latency minimum, maximum, average and standard deviation. This
1638 is the time between submission and completion.
1641 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
1642 and standard deviation.
1645 CPU usage statistics. Includes user and system time, number of context switches
1646 this thread went through and number of major and minor page faults.
1649 Distribution of I/O depths. Each depth includes everything less than (or equal)
1650 to it, but greater than the previous depth.
1653 Number of read/write requests issued, and number of short read/write requests.
1656 Distribution of I/O completion latencies. The numbers follow the same pattern
1660 The group statistics show:
1665 Number of megabytes I/O performed.
1668 Aggregate bandwidth of threads in the group.
1671 Minimum average bandwidth a thread saw.
1674 Maximum average bandwidth a thread saw.
1677 Shortest runtime of threads in the group.
1680 Longest runtime of threads in the group.
1684 Finally, disk statistics are printed with reads first:
1689 Number of I/Os performed by all groups.
1692 Number of merges in the I/O scheduler.
1695 Number of ticks we kept the disk busy.
1698 Total time spent in the disk queue.
1705 It is also possible to get fio to dump the current output while it is
1706 running, without terminating the job. To do that, send fio the \fBUSR1\fR
1709 If the \fB\-\-minimal\fR / \fB\-\-append-terse\fR options are given, the
1710 results will be printed/appended in a semicolon-delimited format suitable for
1712 A job description (if provided) follows on a new line. Note that the first
1713 number in the line is the version number. If the output has to be changed
1714 for some reason, this number will be incremented by 1 to signify that
1715 change. The fields are:
1718 .B terse version, fio version, jobname, groupid, error
1722 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1726 .B min, max, mean, standard deviation
1730 .B min, max, mean, standard deviation
1732 Completion latency percentiles (20 fields):
1734 .B Xth percentile=usec
1738 .B min, max, mean, standard deviation
1742 .B min, max, aggregate percentage of total, mean, standard deviation
1748 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1752 .B min, max, mean, standard deviation
1756 .B min, max, mean, standard deviation
1758 Completion latency percentiles (20 fields):
1760 .B Xth percentile=usec
1764 .B min, max, mean, standard deviation
1768 .B min, max, aggregate percentage of total, mean, standard deviation
1774 .B user, system, context switches, major page faults, minor page faults
1777 IO depth distribution:
1779 .B <=1, 2, 4, 8, 16, 32, >=64
1782 IO latency distribution:
1786 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1790 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1794 Disk utilization (1 for each disk used):
1796 .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
1799 Error Info (dependent on continue_on_error, default off):
1801 .B total # errors, first error code
1804 .B text description (if provided in config - appears on newline)
1807 Normally you would run fio as a stand-alone application on the machine
1808 where the IO workload should be generated. However, it is also possible to
1809 run the frontend and backend of fio separately. This makes it possible to
1810 have a fio server running on the machine(s) where the IO workload should
1811 be running, while controlling it from another machine.
1813 To start the server, you would do:
1815 \fBfio \-\-server=args\fR
1817 on that machine, where args defines what fio listens to. The arguments
1818 are of the form 'type:hostname or IP:port'. 'type' is either 'ip' (or ip4)
1819 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain
1820 socket. 'hostname' is either a hostname or IP address, and 'port' is the port to
1821 listen to (only valid for TCP/IP, not a local socket). Some examples:
1825 Start a fio server, listening on all interfaces on the default port (8765).
1827 2) fio \-\-server=ip:hostname,4444
1829 Start a fio server, listening on IP belonging to hostname and on port 4444.
1831 3) fio \-\-server=ip6:::1,4444
1833 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
1835 4) fio \-\-server=,4444
1837 Start a fio server, listening on all interfaces on port 4444.
1839 5) fio \-\-server=1.2.3.4
1841 Start a fio server, listening on IP 1.2.3.4 on the default port.
1843 6) fio \-\-server=sock:/tmp/fio.sock
1845 Start a fio server, listening on the local socket /tmp/fio.sock.
1847 When a server is running, you can connect to it from a client. The client
1850 fio \-\-local-args \-\-client=server \-\-remote-args <job file(s)>
1852 where \-\-local-args are arguments that are local to the client where it is
1853 running, 'server' is the connect string, and \-\-remote-args and <job file(s)>
1854 are sent to the server. The 'server' string follows the same format as it
1855 does on the server side, to allow IP/hostname/socket and port strings.
1856 You can connect to multiple clients as well, to do that you could run:
1858 fio \-\-client=server2 \-\-client=server2 <job file(s)>
1860 If the job file is located on the fio server, then you can tell the server
1861 to load a local file as well. This is done by using \-\-remote-config:
1863 fio \-\-client=server \-\-remote-config /path/to/file.fio
1865 Then the fio serer will open this local (to the server) job file instead
1866 of being passed one from the client.
1870 was written by Jens Axboe <jens.axboe@oracle.com>,
1871 now Jens Axboe <jaxboe@fusionio.com>.
1873 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1874 on documentation by Jens Axboe.
1875 .SH "REPORTING BUGS"
1876 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1879 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1881 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob