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.
45 Perform test and validation of internal CPU clock
47 .BI \-\-crctest[\fR=\fPtest]
48 Test the speed of the builtin checksumming functions. If no argument is given,
49 all of them are tested. Or a comma separated list can be passed, in which
50 case the given ones are tested.
52 .BI \-\-cmdhelp \fR=\fPcommand
53 Print help information for \fIcommand\fR. May be `all' for all commands.
55 .BI \-\-enghelp \fR=\fPioengine[,command]
56 List all commands defined by \fIioengine\fR, or print help for \fIcommand\fR defined by \fIioengine\fR.
58 .BI \-\-showcmd \fR=\fPjobfile
59 Convert \fIjobfile\fR to a set of command-line options.
61 .BI \-\-eta \fR=\fPwhen
62 Specifies when real-time ETA estimate should be printed. \fIwhen\fR may
63 be one of `always', `never' or `auto'.
65 .BI \-\-eta\-newline \fR=\fPtime
66 Force an ETA newline for every `time` period passed.
68 .BI \-\-status\-interval \fR=\fPtime
69 Report full output status every `time` period passed.
72 Turn on safety read-only checks, preventing any attempted write.
74 .BI \-\-section \fR=\fPsec
75 Only run section \fIsec\fR from job file. Multiple of these options can be given, adding more sections to run.
77 .BI \-\-alloc\-size \fR=\fPkb
78 Set the internal smalloc pool size to \fIkb\fP kilobytes.
80 .BI \-\-warnings\-fatal
81 All fio parser warnings are fatal, causing fio to exit with an error.
83 .BI \-\-max\-jobs \fR=\fPnr
84 Set the maximum allowed number of jobs (threads/processes) to support.
86 .BI \-\-server \fR=\fPargs
87 Start a backend server, with \fIargs\fP specifying what to listen to. See client/server section.
89 .BI \-\-daemonize \fR=\fPpidfile
90 Background a fio server, writing the pid to the given pid file.
92 .BI \-\-client \fR=\fPhost
93 Instead of running the jobs locally, send and run them on the given host.
95 .BI \-\-idle\-prof \fR=\fPoption
96 Report cpu idleness on a system or percpu basis (\fIoption\fP=system,percpu) or run unit work calibration only (\fIoption\fP=calibrate).
98 Job files are in `ini' format. They consist of one or more
99 job definitions, which begin with a job name in square brackets and
100 extend to the next job name. The job name can be any ASCII string
101 except `global', which has a special meaning. Following the job name is
102 a sequence of zero or more parameters, one per line, that define the
103 behavior of the job. Any line starting with a `;' or `#' character is
104 considered a comment and ignored.
106 If \fIjobfile\fR is specified as `-', the job file will be read from
109 The global section contains default parameters for jobs specified in the
110 job file. A job is only affected by global sections residing above it,
111 and there may be any number of global sections. Specific job definitions
112 may override any parameter set in global sections.
115 Some parameters may take arguments of a specific type. The types used are:
118 String: a sequence of alphanumeric characters.
121 SI integer: a whole number, possibly containing a suffix denoting the base unit
122 of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting
123 kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5)
124 respectively. The suffix is not case sensitive. If prefixed with '0x', the
125 value is assumed to be base 16 (hexadecimal). A suffix may include a trailing 'b',
126 for instance 'kb' is identical to 'k'. You can specify a base 10 value
127 by using 'KiB', 'MiB', 'GiB', etc. This is useful for disk drives where
128 values are often given in base 10 values. Specifying '30GiB' will get you
132 Boolean: a true or false value. `0' denotes false, `1' denotes true.
135 Integer range: a range of integers specified in the format
136 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
137 \fIupper\fR may contain a suffix as described above. If an option allows two
138 sets of ranges, they are separated with a `,' or `/' character. For example:
142 List of floating numbers: A list of floating numbers, separated by
147 May be used to override the job name. On the command line, this parameter
148 has the special purpose of signalling the start of a new job.
150 .BI description \fR=\fPstr
151 Human-readable description of the job. It is printed when the job is run, but
152 otherwise has no special purpose.
154 .BI directory \fR=\fPstr
155 Prefix filenames with this directory. Used to place files in a location other
158 .BI filename \fR=\fPstr
160 normally makes up a file name based on the job name, thread number, and file
161 number. If you want to share files between threads in a job or several jobs,
162 specify a \fIfilename\fR for each of them to override the default.
163 If the I/O engine is file-based, you can specify
164 a number of files by separating the names with a `:' character. `\-' is a
165 reserved name, meaning stdin or stdout, depending on the read/write direction
168 .BI filename_format \fR=\fPstr
169 If sharing multiple files between jobs, it is usually necessary to have
170 fio generate the exact names that you want. By default, fio will name a file
171 based on the default file format specification of
172 \fBjobname.jobnumber.filenumber\fP. With this option, that can be
173 customized. Fio will recognize and replace the following keywords in this
179 The name of the worker thread or process.
182 The incremental number of the worker thread or process.
185 The incremental number of the file for that worker thread or process.
188 To have dependent jobs share a set of files, this option can be set to
189 have fio generate filenames that are shared between the two. For instance,
190 if \fBtestfiles.$filenum\fR is specified, file number 4 for any job will
191 be named \fBtestfiles.4\fR. The default of \fB$jobname.$jobnum.$filenum\fR
192 will be used if no other format specifier is given.
196 .BI lockfile \fR=\fPstr
197 Fio defaults to not locking any files before it does IO to them. If a file or
198 file descriptor is shared, fio can serialize IO to that file to make the end
199 result consistent. This is usual for emulating real workloads that share files.
205 No locking. This is the default.
208 Only one thread or process may do IO at the time, excluding all others.
211 Read-write locking on the file. Many readers may access the file at the same
212 time, but writes get exclusive access.
216 .BI opendir \fR=\fPstr
217 Recursively open any files below directory \fIstr\fR.
219 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
220 Type of I/O pattern. Accepted values are:
231 Sequential trim (Linux block devices only).
240 Random trim (Linux block devices only).
243 Mixed sequential reads and writes.
246 Mixed random reads and writes.
249 For mixed I/O, the default split is 50/50. For certain types of io the result
250 may still be skewed a bit, since the speed may be different. It is possible to
251 specify a number of IO's to do before getting a new offset, this is done by
252 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
253 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
254 value of 8. If the postfix is used with a sequential IO pattern, then the value
255 specified will be added to the generated offset for each IO. For instance,
256 using \fBrw=write:4k\fR will skip 4k for every write. It turns sequential IO
257 into sequential IO with holes. See the \fBrw_sequencer\fR option.
260 .BI rw_sequencer \fR=\fPstr
261 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
262 then this option controls how that number modifies the IO offset being
263 generated. Accepted values are:
268 Generate sequential offset
271 Generate the same offset
274 \fBsequential\fR is only useful for random IO, where fio would normally
275 generate a new random offset for every IO. If you append eg 8 to randread, you
276 would get a new random offset for every 8 IO's. The result would be a seek for
277 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
278 that. As sequential IO is already sequential, setting \fBsequential\fR for that
279 would not result in any differences. \fBidentical\fR behaves in a similar
280 fashion, except it sends the same offset 8 number of times before generating a
285 .BI kb_base \fR=\fPint
286 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
287 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
288 reasons. Allowed values are 1024 or 1000, with 1024 being the default.
290 .BI unified_rw_reporting \fR=\fPbool
291 Fio normally reports statistics on a per data direction basis, meaning that
292 read, write, and trim are accounted and reported separately. If this option is
293 set, the fio will sum the results and report them as "mixed" instead.
295 .BI randrepeat \fR=\fPbool
296 Seed the random number generator used for random I/O patterns in a predictable
297 way so the pattern is repeatable across runs. Default: true.
299 .BI allrandrepeat \fR=\fPbool
300 Seed all random number generators in a predictable way so results are
301 repeatable across runs. Default: false.
303 .BI randseed \fR=\fPint
304 Seed the random number generators based on this seed value, to be able to
305 control what sequence of output is being generated. If not set, the random
306 sequence depends on the \fBrandrepeat\fR setting.
308 .BI use_os_rand \fR=\fPbool
309 Fio can either use the random generator supplied by the OS to generator random
310 offsets, or it can use it's own internal generator (based on Tausworthe).
311 Default is to use the internal generator, which is often of better quality and
312 faster. Default: false.
314 .BI fallocate \fR=\fPstr
315 Whether pre-allocation is performed when laying down files. Accepted values
321 Do not pre-allocate space.
324 Pre-allocate via \fBposix_fallocate\fR\|(3).
327 Pre-allocate via \fBfallocate\fR\|(2) with FALLOC_FL_KEEP_SIZE set.
330 Backward-compatible alias for 'none'.
333 Backward-compatible alias for 'posix'.
336 May not be available on all supported platforms. 'keep' is only
337 available on Linux. If using ZFS on Solaris this must be set to 'none'
338 because ZFS doesn't support it. Default: 'posix'.
341 .BI fadvise_hint \fR=\fPbool
342 Use of \fBposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
343 are likely to be issued. Default: true.
346 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
347 been transferred, unless limited by other options (\fBruntime\fR, for instance).
348 Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be
349 divided between the available files for the job. If not set, fio will use the
350 full size of the given files or devices. If the files do not exist, size
351 must be given. It is also possible to give size as a percentage between 1 and
352 100. If size=20% is given, fio will use 20% of the full size of the given files
355 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
356 Sets size to something really large and waits for ENOSPC (no space left on
357 device) as the terminating condition. Only makes sense with sequential write.
358 For a read workload, the mount point will be filled first then IO started on
359 the result. This option doesn't make sense if operating on a raw device node,
360 since the size of that is already known by the file system. Additionally,
361 writing beyond end-of-device will not return ENOSPC there.
363 .BI filesize \fR=\fPirange
364 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
365 for files at random within the given range, limited to \fBsize\fR in total (if
366 that is given). If \fBfilesize\fR is not specified, each created file is the
369 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
370 Block size for I/O units. Default: 4k. Values for reads, writes, and trims
371 can be specified separately in the format \fIread\fR,\fIwrite\fR,\fItrim\fR
372 either of which may be empty to leave that value at its default. If a trailing
373 comma isn't given, the remainder will inherit the last value set.
375 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
376 Specify a range of I/O block sizes. The issued I/O unit will always be a
377 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
378 to both reads and writes if only one range is given, but can be specified
379 separately with a comma separating the values. Example: bsrange=1k-4k,2k-8k.
380 Also (see \fBblocksize\fR).
382 .BI bssplit \fR=\fPstr
383 This option allows even finer grained control of the block sizes issued,
384 not just even splits between them. With this option, you can weight various
385 block sizes for exact control of the issued IO for a job that has mixed
386 block sizes. The format of the option is bssplit=blocksize/percentage,
387 optionally adding as many definitions as needed separated by a colon.
388 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
389 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
390 splits to reads and writes. The format is identical to what the
391 \fBbs\fR option accepts, the read and write parts are separated with a
394 .B blocksize_unaligned\fR,\fP bs_unaligned
395 If set, any size in \fBblocksize_range\fR may be used. This typically won't
396 work with direct I/O, as that normally requires sector alignment.
398 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
399 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
400 the minimum blocksize given. Minimum alignment is typically 512b
401 for using direct IO, though it usually depends on the hardware block size.
402 This option is mutually exclusive with using a random map for files, so it
403 will turn off that option.
405 .BI bs_is_seq_rand \fR=\fPbool
406 If this option is set, fio will use the normal read,write blocksize settings as
407 sequential,random instead. Any random read or write will use the WRITE
408 blocksize settings, and any sequential read or write will use the READ
412 Initialise buffers with all zeros. Default: fill buffers with random data.
415 If this option is given, fio will refill the IO buffers on every submit. The
416 default is to only fill it at init time and reuse that data. Only makes sense
417 if zero_buffers isn't specified, naturally. If data verification is enabled,
418 refill_buffers is also automatically enabled.
420 .BI scramble_buffers \fR=\fPbool
421 If \fBrefill_buffers\fR is too costly and the target is using data
422 deduplication, then setting this option will slightly modify the IO buffer
423 contents to defeat normal de-dupe attempts. This is not enough to defeat
424 more clever block compression attempts, but it will stop naive dedupe
425 of blocks. Default: true.
427 .BI buffer_compress_percentage \fR=\fPint
428 If this is set, then fio will attempt to provide IO buffer content (on WRITEs)
429 that compress to the specified level. Fio does this by providing a mix of
430 random data and zeroes. Note that this is per block size unit, for file/disk
431 wide compression level that matches this setting, you'll also want to set
432 \fBrefill_buffers\fR.
434 .BI buffer_compress_chunk \fR=\fPint
435 See \fBbuffer_compress_percentage\fR. This setting allows fio to manage how
436 big the ranges of random data and zeroed data is. Without this set, fio will
437 provide \fBbuffer_compress_percentage\fR of blocksize random data, followed by
438 the remaining zeroed. With this set to some chunk size smaller than the block
439 size, fio can alternate random and zeroed data throughout the IO buffer.
441 .BI buffer_pattern \fR=\fPstr
442 If set, fio will fill the io buffers with this pattern. If not set, the contents
443 of io buffers is defined by the other options related to buffer contents. The
444 setting can be any pattern of bytes, and can be prefixed with 0x for hex
447 .BI nrfiles \fR=\fPint
448 Number of files to use for this job. Default: 1.
450 .BI openfiles \fR=\fPint
451 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
453 .BI file_service_type \fR=\fPstr
454 Defines how files to service are selected. The following types are defined:
459 Choose a file at random.
462 Round robin over open files (default).
465 Do each file in the set sequentially.
468 The number of I/Os to issue before switching a new file can be specified by
469 appending `:\fIint\fR' to the service type.
472 .BI ioengine \fR=\fPstr
473 Defines how the job issues I/O. The following types are defined:
478 Basic \fBread\fR\|(2) or \fBwrite\fR\|(2) I/O. \fBfseek\fR\|(2) is used to
479 position the I/O location.
482 Basic \fBpread\fR\|(2) or \fBpwrite\fR\|(2) I/O.
485 Basic \fBreadv\fR\|(2) or \fBwritev\fR\|(2) I/O. Will emulate queuing by
486 coalescing adjacent IOs into a single submission.
489 Basic \fBpreadv\fR\|(2) or \fBpwritev\fR\|(2) I/O.
492 Linux native asynchronous I/O. This ioengine defines engine specific options.
495 POSIX asynchronous I/O using \fBaio_read\fR\|(3) and \fBaio_write\fR\|(3).
498 Solaris native asynchronous I/O.
501 Windows native asynchronous I/O.
504 File is memory mapped with \fBmmap\fR\|(2) and data copied using
508 \fBsplice\fR\|(2) is used to transfer the data and \fBvmsplice\fR\|(2) to
509 transfer data from user-space to the kernel.
512 Use the syslet system calls to make regular read/write asynchronous.
515 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
516 the target is an sg character device, we use \fBread\fR\|(2) and
517 \fBwrite\fR\|(2) for asynchronous I/O.
520 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
521 itself and for debugging and testing purposes.
524 Transfer over the network. The protocol to be used can be defined with the
525 \fBprotocol\fR parameter. Depending on the protocol, \fBfilename\fR,
526 \fBhostname\fR, \fBport\fR, or \fBlisten\fR must be specified.
527 This ioengine defines engine specific options.
530 Like \fBnet\fR, but uses \fBsplice\fR\|(2) and \fBvmsplice\fR\|(2) to map data
531 and send/receive. This ioengine defines engine specific options.
534 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
535 \fBcpucycles\fR parameters.
538 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
539 approach to asynchronous I/O.
541 See <http://www.xmailserver.org/guasi\-lib.html>.
544 The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ)
545 and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols.
548 Loads an external I/O engine object file. Append the engine filename as
552 IO engine that does regular linux native fallocate call to simulate data
553 transfer as fio ioengine
555 DDIR_READ does fallocate(,mode = FALLOC_FL_KEEP_SIZE,)
557 DIR_WRITE does fallocate(,mode = 0)
559 DDIR_TRIM does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE)
562 IO engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate defragment activity
563 request to DDIR_WRITE event
566 IO engine supporting direct access to Ceph Rados Block Devices (RBD) via librbd
567 without the need to use the kernel rbd driver. This ioengine defines engine specific
573 .BI iodepth \fR=\fPint
574 Number of I/O units to keep in flight against the file. Note that increasing
575 iodepth beyond 1 will not affect synchronous ioengines (except for small
576 degress when verify_async is in use). Even async engines my impose OS
577 restrictions causing the desired depth not to be achieved. This may happen on
578 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
579 not async on that OS. Keep an eye on the IO depth distribution in the
580 fio output to verify that the achieved depth is as expected. Default: 1.
582 .BI iodepth_batch \fR=\fPint
583 Number of I/Os to submit at once. Default: \fBiodepth\fR.
585 .BI iodepth_batch_complete \fR=\fPint
586 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
587 means that we'll ask for a minimum of 1 IO in the retrieval process from the
588 kernel. The IO retrieval will go on until we hit the limit set by
589 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
590 completed events before queuing more IO. This helps reduce IO latency, at the
591 cost of more retrieval system calls.
593 .BI iodepth_low \fR=\fPint
594 Low watermark indicating when to start filling the queue again. Default:
597 .BI direct \fR=\fPbool
598 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
600 .BI atomic \fR=\fPbool
601 If value is true, attempt to use atomic direct IO. Atomic writes are guaranteed
602 to be stable once acknowledged by the operating system. Only Linux supports
605 .BI buffered \fR=\fPbool
606 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
609 .BI offset \fR=\fPint
610 Offset in the file to start I/O. Data before the offset will not be touched.
612 .BI offset_increment \fR=\fPint
613 If this is provided, then the real offset becomes the
614 offset + offset_increment * thread_number, where the thread number is a counter
615 that starts at 0 and is incremented for each job. This option is useful if
616 there are several jobs which are intended to operate on a file in parallel in
617 disjoint segments, with even spacing between the starting points.
619 .BI number_ios \fR=\fPint
620 Fio will normally perform IOs until it has exhausted the size of the region
621 set by \fBsize\fR, or if it exhaust the allocated time (or hits an error
622 condition). With this setting, the range/size can be set independently of
623 the number of IOs to perform. When fio reaches this number, it will exit
624 normally and report status.
627 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
628 0, don't sync. Default: 0.
630 .BI fdatasync \fR=\fPint
631 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
632 data parts of the file. Default: 0.
634 .BI write_barrier \fR=\fPint
635 Make every Nth write a barrier write.
637 .BI sync_file_range \fR=\fPstr:int
638 Use \fBsync_file_range\fR\|(2) for every \fRval\fP number of write operations. Fio will
639 track range of writes that have happened since the last \fBsync_file_range\fR\|(2) call.
640 \fRstr\fP can currently be one or more of:
644 SYNC_FILE_RANGE_WAIT_BEFORE
647 SYNC_FILE_RANGE_WRITE
650 SYNC_FILE_RANGE_WRITE
654 So if you do sync_file_range=wait_before,write:8, fio would use
655 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
656 Also see the \fBsync_file_range\fR\|(2) man page. This option is Linux specific.
658 .BI overwrite \fR=\fPbool
659 If writing, setup the file first and do overwrites. Default: false.
661 .BI end_fsync \fR=\fPbool
662 Sync file contents when a write stage has completed. Default: false.
664 .BI fsync_on_close \fR=\fPbool
665 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
666 it will happen on every close, not just at the end of the job. Default: false.
668 .BI rwmixread \fR=\fPint
669 Percentage of a mixed workload that should be reads. Default: 50.
671 .BI rwmixwrite \fR=\fPint
672 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
673 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
674 overrides the first. This may interfere with a given rate setting, if fio is
675 asked to limit reads or writes to a certain rate. If that is the case, then
676 the distribution may be skewed. Default: 50.
678 .BI random_distribution \fR=\fPstr:float
679 By default, fio will use a completely uniform random distribution when asked
680 to perform random IO. Sometimes it is useful to skew the distribution in
681 specific ways, ensuring that some parts of the data is more hot than others.
682 Fio includes the following distribution models:
686 Uniform random distribution
696 When using a zipf or pareto distribution, an input value is also needed to
697 define the access pattern. For zipf, this is the zipf theta. For pareto,
698 it's the pareto power. Fio includes a test program, genzipf, that can be
699 used visualize what the given input values will yield in terms of hit rates.
700 If you wanted to use zipf with a theta of 1.2, you would use
701 random_distribution=zipf:1.2 as the option. If a non-uniform model is used,
702 fio will disable use of the random map.
704 .BI percentage_random \fR=\fPint
705 For a random workload, set how big a percentage should be random. This defaults
706 to 100%, in which case the workload is fully random. It can be set from
707 anywhere from 0 to 100. Setting it to 0 would make the workload fully
708 sequential. It is possible to set different values for reads, writes, and
709 trim. To do so, simply use a comma separated list. See \fBblocksize\fR.
712 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
713 this parameter is given, a new offset will be chosen without looking at past
714 I/O history. This parameter is mutually exclusive with \fBverify\fR.
716 .BI softrandommap \fR=\fPbool
717 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
718 fails to allocate the map, if this option is set it will continue without a
719 random block map. As coverage will not be as complete as with random maps, this
720 option is disabled by default.
722 .BI random_generator \fR=\fPstr
723 Fio supports the following engines for generating IO offsets for random IO:
727 Strong 2^88 cycle random number generator
730 Linear feedback shift register generator
734 Tausworthe is a strong random number generator, but it requires tracking on the
735 side if we want to ensure that blocks are only read or written once. LFSR
736 guarantees that we never generate the same offset twice, and it's also less
737 computationally expensive. It's not a true random generator, however, though
738 for IO purposes it's typically good enough. LFSR only works with single block
739 sizes, not with workloads that use multiple block sizes. If used with such a
740 workload, fio may read or write some blocks multiple times.
743 Run job with given nice value. See \fBnice\fR\|(2).
746 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
749 .BI prioclass \fR=\fPint
750 Set I/O priority class. See \fBionice\fR\|(1).
752 .BI thinktime \fR=\fPint
753 Stall job for given number of microseconds between issuing I/Os.
755 .BI thinktime_spin \fR=\fPint
756 Pretend to spend CPU time for given number of microseconds, sleeping the rest
757 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
759 .BI thinktime_blocks \fR=\fPint
760 Only valid if thinktime is set - control how many blocks to issue, before
761 waiting \fBthinktime\fR microseconds. If not set, defaults to 1 which will
762 make fio wait \fBthinktime\fR microseconds after every block. This
763 effectively makes any queue depth setting redundant, since no more than 1 IO
764 will be queued before we have to complete it and do our thinktime. In other
765 words, this setting effectively caps the queue depth if the latter is larger.
769 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
770 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
771 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
772 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
773 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
774 limit writes (to 500KB/sec), the latter will only limit reads.
776 .BI ratemin \fR=\fPint
777 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
778 Failing to meet this requirement will cause the job to exit. The same format
779 as \fBrate\fR is used for read vs write separation.
781 .BI rate_iops \fR=\fPint
782 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
783 specified independently of bandwidth. The same format as \fBrate\fR is used for
784 read vs write separation. If \fBblocksize\fR is a range, the smallest block
785 size is used as the metric.
787 .BI rate_iops_min \fR=\fPint
788 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
789 is used for read vs write separation.
791 .BI ratecycle \fR=\fPint
792 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
793 milliseconds. Default: 1000ms.
795 .BI latency_target \fR=\fPint
796 If set, fio will attempt to find the max performance point that the given
797 workload will run at while maintaining a latency below this target. The
798 values is given in microseconds. See \fBlatency_window\fR and
799 \fBlatency_percentile\fR.
801 .BI latency_window \fR=\fPint
802 Used with \fBlatency_target\fR to specify the sample window that the job
803 is run at varying queue depths to test the performance. The value is given
806 .BI latency_percentile \fR=\fPfloat
807 The percentage of IOs that must fall within the criteria specified by
808 \fBlatency_target\fR and \fBlatency_window\fR. If not set, this defaults
809 to 100.0, meaning that all IOs must be equal or below to the value set
810 by \fBlatency_target\fR.
812 .BI max_latency \fR=\fPint
813 If set, fio will exit the job if it exceeds this maximum latency. It will exit
816 .BI cpumask \fR=\fPint
817 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
818 may run on. See \fBsched_setaffinity\fR\|(2).
820 .BI cpus_allowed \fR=\fPstr
821 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
823 .BI numa_cpu_nodes \fR=\fPstr
824 Set this job running on specified NUMA nodes' CPUs. The arguments allow
825 comma delimited list of cpu numbers, A-B ranges, or 'all'.
827 .BI numa_mem_policy \fR=\fPstr
828 Set this job's memory policy and corresponding NUMA nodes. Format of
832 .B <mode>[:<nodelist>]
835 is one of the following memory policy:
837 .B default, prefer, bind, interleave, local
840 For \fBdefault\fR and \fBlocal\fR memory policy, no \fBnodelist\fR is
841 needed to be specified. For \fBprefer\fR, only one node is
842 allowed. For \fBbind\fR and \fBinterleave\fR, \fBnodelist\fR allows
843 comma delimited list of numbers, A-B ranges, or 'all'.
845 .BI startdelay \fR=\fPint
846 Delay start of job for the specified number of seconds.
848 .BI runtime \fR=\fPint
849 Terminate processing after the specified number of seconds.
852 If given, run for the specified \fBruntime\fR duration even if the files are
853 completely read or written. The same workload will be repeated as many times
854 as \fBruntime\fR allows.
856 .BI ramp_time \fR=\fPint
857 If set, fio will run the specified workload for this amount of time before
858 logging any performance numbers. Useful for letting performance settle before
859 logging results, thus minimizing the runtime required for stable results. Note
860 that the \fBramp_time\fR is considered lead in time for a job, thus it will
861 increase the total runtime if a special timeout or runtime is specified.
863 .BI invalidate \fR=\fPbool
864 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
867 Use synchronous I/O for buffered writes. For the majority of I/O engines,
868 this means using O_SYNC. Default: false.
870 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
871 Allocation method for I/O unit buffer. Allowed values are:
876 Allocate memory with \fBmalloc\fR\|(3).
879 Use shared memory buffers allocated through \fBshmget\fR\|(2).
882 Same as \fBshm\fR, but use huge pages as backing.
885 Use \fBmmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
886 is given after the option in the format `:\fIfile\fR'.
889 Same as \fBmmap\fR, but use huge files as backing.
892 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
893 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
894 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
895 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
896 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
897 and the documentation for that. Normally you just need to echo an appropriate
898 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
902 .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint
903 This indicates the memory alignment of the IO memory buffers. Note that the
904 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
905 the alignment of the following buffers are given by the \fBbs\fR used. In
906 other words, if using a \fBbs\fR that is a multiple of the page sized in the
907 system, all buffers will be aligned to this value. If using a \fBbs\fR that
908 is not page aligned, the alignment of subsequent IO memory buffers is the
909 sum of the \fBiomem_align\fR and \fBbs\fR used.
911 .BI hugepage\-size \fR=\fPint
912 Defines the size of a huge page. Must be at least equal to the system setting.
913 Should be a multiple of 1MB. Default: 4MB.
916 Terminate all jobs when one finishes. Default: wait for each job to finish.
918 .BI bwavgtime \fR=\fPint
919 Average bandwidth calculations over the given time in milliseconds. Default:
922 .BI iopsavgtime \fR=\fPint
923 Average IOPS calculations over the given time in milliseconds. Default:
926 .BI create_serialize \fR=\fPbool
927 If true, serialize file creation for the jobs. Default: true.
929 .BI create_fsync \fR=\fPbool
930 \fBfsync\fR\|(2) data file after creation. Default: true.
932 .BI create_on_open \fR=\fPbool
933 If true, the files are not created until they are opened for IO by the job.
935 .BI create_only \fR=\fPbool
936 If true, fio will only run the setup phase of the job. If files need to be
937 laid out or updated on disk, only that will be done. The actual job contents
940 .BI pre_read \fR=\fPbool
941 If this is given, files will be pre-read into memory before starting the given
942 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
943 pointless to pre-read and then drop the cache. This will only work for IO
944 engines that are seekable, since they allow you to read the same data
945 multiple times. Thus it will not work on eg network or splice IO.
947 .BI unlink \fR=\fPbool
948 Unlink job files when done. Default: false.
951 Specifies the number of iterations (runs of the same workload) of this job.
954 .BI verify_only \fR=\fPbool
955 Do not perform the specified workload, only verify data still matches previous
956 invocation of this workload. This option allows one to check data multiple
957 times at a later date without overwriting it. This option makes sense only for
958 workloads that write data, and does not support workloads with the
959 \fBtime_based\fR option set.
961 .BI do_verify \fR=\fPbool
962 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
965 .BI verify \fR=\fPstr
966 Method of verifying file contents after each iteration of the job. Allowed
971 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1
972 Store appropriate checksum in the header of each block. crc32c-intel is
973 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
974 not supported by the system.
977 Write extra information about each I/O (timestamp, block number, etc.). The
978 block number is verified. See \fBverify_pattern\fR as well.
981 Pretend to verify. Used for testing internals.
984 This option can be used for repeated burn-in tests of a system to make sure
985 that the written data is also correctly read back. If the data direction given
986 is a read or random read, fio will assume that it should verify a previously
987 written file. If the data direction includes any form of write, the verify will
988 be of the newly written data.
991 .BI verifysort \fR=\fPbool
992 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
993 read them back in a sorted manner. Default: true.
995 .BI verifysort_nr \fR=\fPint
996 Pre-load and sort verify blocks for a read workload.
998 .BI verify_offset \fR=\fPint
999 Swap the verification header with data somewhere else in the block before
1000 writing. It is swapped back before verifying.
1002 .BI verify_interval \fR=\fPint
1003 Write the verification header for this number of bytes, which should divide
1004 \fBblocksize\fR. Default: \fBblocksize\fR.
1006 .BI verify_pattern \fR=\fPstr
1007 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
1008 with totally random bytes, but sometimes it's interesting to fill with a known
1009 pattern for io verification purposes. Depending on the width of the pattern,
1010 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
1011 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
1012 has to be a hex number that starts with either "0x" or "0X". Use with
1015 .BI verify_fatal \fR=\fPbool
1016 If true, exit the job on the first observed verification failure. Default:
1019 .BI verify_dump \fR=\fPbool
1020 If set, dump the contents of both the original data block and the data block we
1021 read off disk to files. This allows later analysis to inspect just what kind of
1022 data corruption occurred. Off by default.
1024 .BI verify_async \fR=\fPint
1025 Fio will normally verify IO inline from the submitting thread. This option
1026 takes an integer describing how many async offload threads to create for IO
1027 verification instead, causing fio to offload the duty of verifying IO contents
1028 to one or more separate threads. If using this offload option, even sync IO
1029 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
1030 allows them to have IO in flight while verifies are running.
1032 .BI verify_async_cpus \fR=\fPstr
1033 Tell fio to set the given CPU affinity on the async IO verification threads.
1034 See \fBcpus_allowed\fP for the format used.
1036 .BI verify_backlog \fR=\fPint
1037 Fio will normally verify the written contents of a job that utilizes verify
1038 once that job has completed. In other words, everything is written then
1039 everything is read back and verified. You may want to verify continually
1040 instead for a variety of reasons. Fio stores the meta data associated with an
1041 IO block in memory, so for large verify workloads, quite a bit of memory would
1042 be used up holding this meta data. If this option is enabled, fio will write
1043 only N blocks before verifying these blocks.
1045 .BI verify_backlog_batch \fR=\fPint
1046 Control how many blocks fio will verify if verify_backlog is set. If not set,
1047 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
1048 read back and verified). If \fBverify_backlog_batch\fR is less than
1049 \fBverify_backlog\fR then not all blocks will be verified, if
1050 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
1051 will be verified more than once.
1053 .BI trim_percentage \fR=\fPint
1054 Number of verify blocks to discard/trim.
1056 .BI trim_verify_zero \fR=\fPbool
1057 Verify that trim/discarded blocks are returned as zeroes.
1059 .BI trim_backlog \fR=\fPint
1060 Trim after this number of blocks are written.
1062 .BI trim_backlog_batch \fR=\fPint
1063 Trim this number of IO blocks.
1065 .BI experimental_verify \fR=\fPbool
1066 Enable experimental verification.
1068 .B stonewall "\fR,\fP wait_for_previous"
1069 Wait for preceding jobs in the job file to exit before starting this one.
1070 \fBstonewall\fR implies \fBnew_group\fR.
1073 Start a new reporting group. If not given, all jobs in a file will be part
1074 of the same reporting group, unless separated by a stonewall.
1076 .BI numjobs \fR=\fPint
1077 Number of clones (processes/threads performing the same workload) of this job.
1081 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
1085 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
1086 with \fBfork\fR\|(2).
1088 .BI zonesize \fR=\fPint
1089 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
1091 .BI zonerange \fR=\fPint
1092 Give size of an IO zone. See \fBzoneskip\fR.
1094 .BI zoneskip \fR=\fPint
1095 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
1098 .BI write_iolog \fR=\fPstr
1099 Write the issued I/O patterns to the specified file. Specify a separate file
1100 for each job, otherwise the iologs will be interspersed and the file may be
1103 .BI read_iolog \fR=\fPstr
1104 Replay the I/O patterns contained in the specified file generated by
1105 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
1107 .BI replay_no_stall \fR=\fPint
1108 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1109 attempts to respect timing information between I/Os. Enabling
1110 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
1111 still respecting ordering.
1113 .BI replay_redirect \fR=\fPstr
1114 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1115 is to replay the IOPS onto the major/minor device that each IOP was recorded
1116 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
1117 single specified device regardless of the device it was recorded from.
1119 .BI write_bw_log \fR=\fPstr
1120 If given, write a bandwidth log of the jobs in this job file. Can be used to
1121 store data of the bandwidth of the jobs in their lifetime. The included
1122 fio_generate_plots script uses gnuplot to turn these text files into nice
1123 graphs. See \fBwrite_lat_log\fR for behaviour of given filename. For this
1124 option, the postfix is _bw.log.
1126 .BI write_lat_log \fR=\fPstr
1127 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
1128 filename is given with this option, the default filename of "jobname_type.log"
1129 is used. Even if the filename is given, fio will still append the type of log.
1131 .BI write_iops_log \fR=\fPstr
1132 Same as \fBwrite_bw_log\fR, but writes IOPS. If no filename is given with this
1133 option, the default filename of "jobname_type.log" is used. Even if the
1134 filename is given, fio will still append the type of log.
1136 .BI log_avg_msec \fR=\fPint
1137 By default, fio will log an entry in the iops, latency, or bw log for every
1138 IO that completes. When writing to the disk log, that can quickly grow to a
1139 very large size. Setting this option makes fio average the each log entry
1140 over the specified period of time, reducing the resolution of the log.
1143 .BI disable_lat \fR=\fPbool
1144 Disable measurements of total latency numbers. Useful only for cutting
1145 back the number of calls to \fBgettimeofday\fR\|(2), as that does impact performance at
1146 really high IOPS rates. Note that to really get rid of a large amount of these
1147 calls, this option must be used with disable_slat and disable_bw as well.
1149 .BI disable_clat \fR=\fPbool
1150 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
1152 .BI disable_slat \fR=\fPbool
1153 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
1155 .BI disable_bw_measurement \fR=\fPbool
1156 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
1158 .BI lockmem \fR=\fPint
1159 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
1160 simulate a smaller amount of memory. The amount specified is per worker.
1162 .BI exec_prerun \fR=\fPstr
1163 Before running the job, execute the specified command with \fBsystem\fR\|(3).
1165 Output is redirected in a file called \fBjobname.prerun.txt\fR
1168 .BI exec_postrun \fR=\fPstr
1169 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
1171 Output is redirected in a file called \fBjobname.postrun.txt\fR
1174 .BI ioscheduler \fR=\fPstr
1175 Attempt to switch the device hosting the file to the specified I/O scheduler.
1177 .BI cpuload \fR=\fPint
1178 If the job is a CPU cycle-eater, attempt to use the specified percentage of
1181 .BI cpuchunks \fR=\fPint
1182 If the job is a CPU cycle-eater, split the load into cycles of the
1183 given time in milliseconds.
1185 .BI disk_util \fR=\fPbool
1186 Generate disk utilization statistics if the platform supports it. Default: true.
1188 .BI clocksource \fR=\fPstr
1189 Use the given clocksource as the base of timing. The supported options are:
1193 \fBgettimeofday\fR\|(2)
1196 \fBclock_gettime\fR\|(2)
1199 Internal CPU clock source
1203 \fBcpu\fR is the preferred clocksource if it is reliable, as it is very fast
1204 (and fio is heavy on time calls). Fio will automatically use this clocksource
1205 if it's supported and considered reliable on the system it is running on,
1206 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
1207 means supporting TSC Invariant.
1209 .BI gtod_reduce \fR=\fPbool
1210 Enable all of the \fBgettimeofday\fR\|(2) reducing options (disable_clat, disable_slat,
1211 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
1212 \fBgettimeofday\fR\|(2) call count. With this option enabled, we only do about 0.4% of
1213 the gtod() calls we would have done if all time keeping was enabled.
1215 .BI gtod_cpu \fR=\fPint
1216 Sometimes it's cheaper to dedicate a single thread of execution to just getting
1217 the current time. Fio (and databases, for instance) are very intensive on
1218 \fBgettimeofday\fR\|(2) calls. With this option, you can set one CPU aside for doing
1219 nothing but logging current time to a shared memory location. Then the other
1220 threads/processes that run IO workloads need only copy that segment, instead of
1221 entering the kernel with a \fBgettimeofday\fR\|(2) call. The CPU set aside for doing
1222 these time calls will be excluded from other uses. Fio will manually clear it
1223 from the CPU mask of other jobs.
1225 .BI ignore_error \fR=\fPstr
1226 Sometimes you want to ignore some errors during test in that case you can specify
1227 error list for each error type.
1229 ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST
1231 errors for given error type is separated with ':'.
1232 Error may be symbol ('ENOSPC', 'ENOMEM') or an integer.
1234 Example: ignore_error=EAGAIN,ENOSPC:122 .
1236 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from WRITE.
1238 .BI error_dump \fR=\fPbool
1239 If set dump every error even if it is non fatal, true by default. If disabled
1240 only fatal error will be dumped
1242 .BI profile \fR=\fPstr
1243 Select a specific builtin performance test.
1245 .BI cgroup \fR=\fPstr
1246 Add job to this control group. If it doesn't exist, it will be created.
1247 The system must have a mounted cgroup blkio mount point for this to work. If
1248 your system doesn't have it mounted, you can do so with:
1250 # mount \-t cgroup \-o blkio none /cgroup
1252 .BI cgroup_weight \fR=\fPint
1253 Set the weight of the cgroup to this value. See the documentation that comes
1254 with the kernel, allowed values are in the range of 100..1000.
1256 .BI cgroup_nodelete \fR=\fPbool
1257 Normally fio will delete the cgroups it has created after the job completion.
1258 To override this behavior and to leave cgroups around after the job completion,
1259 set cgroup_nodelete=1. This can be useful if one wants to inspect various
1260 cgroup files after job completion. Default: false
1263 Instead of running as the invoking user, set the user ID to this value before
1264 the thread/process does any work.
1267 Set group ID, see \fBuid\fR.
1269 .BI unit_base \fR=\fPint
1270 Base unit for reporting. Allowed values are:
1274 Use auto-detection (default).
1284 .BI flow_id \fR=\fPint
1285 The ID of the flow. If not specified, it defaults to being a global flow. See
1289 Weight in token-based flow control. If this value is used, then there is a
1290 \fBflow counter\fR which is used to regulate the proportion of activity between
1291 two or more jobs. fio attempts to keep this flow counter near zero. The
1292 \fBflow\fR parameter stands for how much should be added or subtracted to the
1293 flow counter on each iteration of the main I/O loop. That is, if one job has
1294 \fBflow=8\fR and another job has \fBflow=-1\fR, then there will be a roughly
1295 1:8 ratio in how much one runs vs the other.
1297 .BI flow_watermark \fR=\fPint
1298 The maximum value that the absolute value of the flow counter is allowed to
1299 reach before the job must wait for a lower value of the counter.
1301 .BI flow_sleep \fR=\fPint
1302 The period of time, in microseconds, to wait after the flow watermark has been
1303 exceeded before retrying operations
1305 .BI clat_percentiles \fR=\fPbool
1306 Enable the reporting of percentiles of completion latencies.
1308 .BI percentile_list \fR=\fPfloat_list
1309 Overwrite the default list of percentiles for completion
1310 latencies. Each number is a floating number in the range (0,100], and
1311 the maximum length of the list is 20. Use ':' to separate the
1312 numbers. For example, \-\-percentile_list=99.5:99.9 will cause fio to
1313 report the values of completion latency below which 99.5% and 99.9% of
1314 the observed latencies fell, respectively.
1315 .SS "Ioengine Parameters List"
1316 Some parameters are only valid when a specific ioengine is in use. These are
1317 used identically to normal parameters, with the caveat that when used on the
1318 command line, the must come after the ioengine that defines them is selected.
1320 .BI (cpu)cpuload \fR=\fPint
1321 Attempt to use the specified percentage of CPU cycles.
1323 .BI (cpu)cpuchunks \fR=\fPint
1324 Split the load into cycles of the given time. In microseconds.
1326 .BI (libaio)userspace_reap
1327 Normally, with the libaio engine in use, fio will use
1328 the io_getevents system call to reap newly returned events.
1329 With this flag turned on, the AIO ring will be read directly
1330 from user-space to reap events. The reaping mode is only
1331 enabled when polling for a minimum of 0 events (eg when
1332 iodepth_batch_complete=0).
1334 .BI (net,netsplice)hostname \fR=\fPstr
1335 The host name or IP address to use for TCP or UDP based IO.
1336 If the job is a TCP listener or UDP reader, the hostname is not
1337 used and must be omitted unless it is a valid UDP multicast address.
1339 .BI (net,netsplice)port \fR=\fPint
1340 The TCP or UDP port to bind to or connect to.
1342 .BI (net,netsplice)interface \fR=\fPstr
1343 The IP address of the network interface used to send or receive UDP multicast
1346 .BI (net,netsplice)ttl \fR=\fPint
1347 Time-to-live value for outgoing UDP multicast packets. Default: 1
1349 .BI (net,netsplice)nodelay \fR=\fPbool
1350 Set TCP_NODELAY on TCP connections.
1352 .BI (net,netsplice)protocol \fR=\fPstr "\fR,\fP proto" \fR=\fPstr
1353 The network protocol to use. Accepted values are:
1358 Transmission control protocol
1361 Transmission control protocol V6
1364 User datagram protocol
1367 User datagram protocol V6
1373 When the protocol is TCP or UDP, the port must also be given,
1374 as well as the hostname if the job is a TCP listener or UDP
1375 reader. For unix sockets, the normal filename option should be
1376 used and the port is invalid.
1379 .BI (net,netsplice)listen
1380 For TCP network connections, tell fio to listen for incoming
1381 connections rather than initiating an outgoing connection. The
1382 hostname must be omitted if this option is used.
1384 .BI (net, pingpong) \fR=\fPbool
1385 Normally a network writer will just continue writing data, and a network reader
1386 will just consume packages. If pingpong=1 is set, a writer will send its normal
1387 payload to the reader, then wait for the reader to send the same payload back.
1388 This allows fio to measure network latencies. The submission and completion
1389 latencies then measure local time spent sending or receiving, and the
1390 completion latency measures how long it took for the other end to receive and
1391 send back. For UDP multicast traffic pingpong=1 should only be set for a single
1392 reader when multiple readers are listening to the same address.
1394 .BI (e4defrag,donorname) \fR=\fPstr
1395 File will be used as a block donor (swap extents between files)
1397 .BI (e4defrag,inplace) \fR=\fPint
1398 Configure donor file block allocation strategy
1401 Preallocate donor's file on init
1404 allocate space immediately inside defragment event, and free right after event
1407 .BI (rbd)rbdname \fR=\fPstr
1408 Specifies the name of the RBD.
1410 .BI (rbd)pool \fR=\fPstr
1411 Specifies the name of the Ceph pool containing the RBD.
1413 .BI (rbd)clientname \fR=\fPstr
1414 Specifies the username (without the 'client.' prefix) used to access the Ceph cluster.
1416 While running, \fBfio\fR will display the status of the created jobs. For
1420 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
1423 The characters in the first set of brackets denote the current status of each
1424 threads. The possible values are:
1430 Setup but not started.
1436 Initialized, waiting.
1439 Running, doing sequential reads.
1442 Running, doing random reads.
1445 Running, doing sequential writes.
1448 Running, doing random writes.
1451 Running, doing mixed sequential reads/writes.
1454 Running, doing mixed random reads/writes.
1457 Running, currently waiting for \fBfsync\fR\|(2).
1460 Running, verifying written data.
1463 Exited, not reaped by main thread.
1466 Exited, thread reaped.
1470 The second set of brackets shows the estimated completion percentage of
1471 the current group. The third set shows the read and write I/O rate,
1472 respectively. Finally, the estimated run time of the job is displayed.
1474 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
1475 for each thread, each group of threads, and each disk, in that order.
1477 Per-thread statistics first show the threads client number, group-id, and
1478 error code. The remaining figures are as follows:
1482 Number of megabytes of I/O performed.
1485 Average data rate (bandwidth).
1491 Submission latency minimum, maximum, average and standard deviation. This is
1492 the time it took to submit the I/O.
1495 Completion latency minimum, maximum, average and standard deviation. This
1496 is the time between submission and completion.
1499 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
1500 and standard deviation.
1503 CPU usage statistics. Includes user and system time, number of context switches
1504 this thread went through and number of major and minor page faults.
1507 Distribution of I/O depths. Each depth includes everything less than (or equal)
1508 to it, but greater than the previous depth.
1511 Number of read/write requests issued, and number of short read/write requests.
1514 Distribution of I/O completion latencies. The numbers follow the same pattern
1518 The group statistics show:
1523 Number of megabytes I/O performed.
1526 Aggregate bandwidth of threads in the group.
1529 Minimum average bandwidth a thread saw.
1532 Maximum average bandwidth a thread saw.
1535 Shortest runtime of threads in the group.
1538 Longest runtime of threads in the group.
1542 Finally, disk statistics are printed with reads first:
1547 Number of I/Os performed by all groups.
1550 Number of merges in the I/O scheduler.
1553 Number of ticks we kept the disk busy.
1556 Total time spent in the disk queue.
1563 It is also possible to get fio to dump the current output while it is
1564 running, without terminating the job. To do that, send fio the \fBUSR1\fR
1567 If the \fB\-\-minimal\fR option is given, the results will be printed in a
1568 semicolon-delimited format suitable for scripted use - a job description
1569 (if provided) follows on a new line. Note that the first
1570 number in the line is the version number. If the output has to be changed
1571 for some reason, this number will be incremented by 1 to signify that
1572 change. The fields are:
1575 .B terse version, fio version, jobname, groupid, error
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 Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1609 .B min, max, mean, standard deviation
1613 .B min, max, mean, standard deviation
1615 Completion latency percentiles (20 fields):
1617 .B Xth percentile=usec
1621 .B min, max, mean, standard deviation
1625 .B min, max, aggregate percentage of total, mean, standard deviation
1631 .B user, system, context switches, major page faults, minor page faults
1634 IO depth distribution:
1636 .B <=1, 2, 4, 8, 16, 32, >=64
1639 IO latency distribution:
1643 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1647 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1651 Disk utilization (1 for each disk used):
1653 .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
1656 Error Info (dependent on continue_on_error, default off):
1658 .B total # errors, first error code
1661 .B text description (if provided in config - appears on newline)
1664 Normally you would run fio as a stand-alone application on the machine
1665 where the IO workload should be generated. However, it is also possible to
1666 run the frontend and backend of fio separately. This makes it possible to
1667 have a fio server running on the machine(s) where the IO workload should
1668 be running, while controlling it from another machine.
1670 To start the server, you would do:
1672 \fBfio \-\-server=args\fR
1674 on that machine, where args defines what fio listens to. The arguments
1675 are of the form 'type:hostname or IP:port'. 'type' is either 'ip' (or ip4)
1676 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain
1677 socket. 'hostname' is either a hostname or IP address, and 'port' is the port to
1678 listen to (only valid for TCP/IP, not a local socket). Some examples:
1682 Start a fio server, listening on all interfaces on the default port (8765).
1684 2) fio \-\-server=ip:hostname,4444
1686 Start a fio server, listening on IP belonging to hostname and on port 4444.
1688 3) fio \-\-server=ip6:::1,4444
1690 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
1692 4) fio \-\-server=,4444
1694 Start a fio server, listening on all interfaces on port 4444.
1696 5) fio \-\-server=1.2.3.4
1698 Start a fio server, listening on IP 1.2.3.4 on the default port.
1700 6) fio \-\-server=sock:/tmp/fio.sock
1702 Start a fio server, listening on the local socket /tmp/fio.sock.
1704 When a server is running, you can connect to it from a client. The client
1707 fio \-\-local-args \-\-client=server \-\-remote-args <job file(s)>
1709 where \-\-local-args are arguments that are local to the client where it is
1710 running, 'server' is the connect string, and \-\-remote-args and <job file(s)>
1711 are sent to the server. The 'server' string follows the same format as it
1712 does on the server side, to allow IP/hostname/socket and port strings.
1713 You can connect to multiple clients as well, to do that you could run:
1715 fio \-\-client=server2 \-\-client=server2 <job file(s)>
1719 was written by Jens Axboe <jens.axboe@oracle.com>,
1720 now Jens Axboe <jaxboe@fusionio.com>.
1722 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1723 on documentation by Jens Axboe.
1724 .SH "REPORTING BUGS"
1725 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1728 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1730 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob