1 .TH fio 1 "September 2007" "User Manual"
3 fio \- flexible I/O tester
6 [\fIoptions\fR] [\fIjobfile\fR]...
9 is a tool that will spawn a number of threads or processes doing a
10 particular type of I/O action as specified by the user.
11 The typical use of fio is to write a job file matching the I/O load
12 one wants to simulate.
15 .BI \-\-debug \fR=\fPtype
16 Enable verbose tracing of various fio actions. May be `all' for all types
17 or individual types separated by a comma (eg \-\-debug=io,file). `help' will
18 list all available tracing options.
20 .BI \-\-output \fR=\fPfilename
21 Write output to \fIfilename\fR.
23 .BI \-\-runtime \fR=\fPruntime
24 Limit run time to \fIruntime\fR seconds.
27 Generate per-job latency logs.
30 Generate per-job bandwidth logs.
33 Print statistics in a terse, semicolon-delimited format.
36 Display version information and exit.
38 .BI \-\-terse\-version \fR=\fPversion
39 Set terse version output format (Current version 3, or older version 2).
42 Display usage information and exit.
44 .BI \-\-cmdhelp \fR=\fPcommand
45 Print help information for \fIcommand\fR. May be `all' for all commands.
47 .BI \-\-enghelp \fR=\fPioengine[,command]
48 List all commands defined by \fIioengine\fR, or print help for \fIcommand\fR defined by \fIioengine\fR.
50 .BI \-\-showcmd \fR=\fPjobfile
51 Convert \fIjobfile\fR to a set of command-line options.
53 .BI \-\-eta \fR=\fPwhen
54 Specifies when real-time ETA estimate should be printed. \fIwhen\fR may
55 be one of `always', `never' or `auto'.
58 Turn on safety read-only checks, preventing any attempted write.
60 .BI \-\-section \fR=\fPsec
61 Only run section \fIsec\fR from job file. Multiple of these options can be given, adding more sections to run.
63 .BI \-\-alloc\-size \fR=\fPkb
64 Set the internal smalloc pool size to \fIkb\fP kilobytes.
66 .BI \-\-warnings\-fatal
67 All fio parser warnings are fatal, causing fio to exit with an error.
69 .BI \-\-max\-jobs \fR=\fPnr
70 Set the maximum allowed number of jobs (threads/processes) to support.
72 .BI \-\-server \fR=\fPargs
73 Start a backend server, with \fIargs\fP specifying what to listen to. See client/server section.
75 .BI \-\-daemonize \fR=\fPpidfile
76 Background a fio server, writing the pid to the given pid file.
78 .BI \-\-client \fR=\fPhost
79 Instead of running the jobs locally, send and run them on the given host.
81 Job files are in `ini' format. They consist of one or more
82 job definitions, which begin with a job name in square brackets and
83 extend to the next job name. The job name can be any ASCII string
84 except `global', which has a special meaning. Following the job name is
85 a sequence of zero or more parameters, one per line, that define the
86 behavior of the job. Any line starting with a `;' or `#' character is
87 considered a comment and ignored.
89 If \fIjobfile\fR is specified as `-', the job file will be read from
92 The global section contains default parameters for jobs specified in the
93 job file. A job is only affected by global sections residing above it,
94 and there may be any number of global sections. Specific job definitions
95 may override any parameter set in global sections.
98 Some parameters may take arguments of a specific type. The types used are:
101 String: a sequence of alphanumeric characters.
104 SI integer: a whole number, possibly containing a suffix denoting the base unit
105 of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting
106 kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5)
107 respectively. The suffix is not case sensitive. If prefixed with '0x', the
108 value is assumed to be base 16 (hexadecimal). A suffix may include a trailing 'b',
109 for instance 'kb' is identical to 'k'. You can specify a base 10 value
110 by using 'KiB', 'MiB', 'GiB', etc. This is useful for disk drives where
111 values are often given in base 10 values. Specifying '30GiB' will get you
115 Boolean: a true or false value. `0' denotes false, `1' denotes true.
118 Integer range: a range of integers specified in the format
119 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
120 \fIupper\fR may contain a suffix as described above. If an option allows two
121 sets of ranges, they are separated with a `,' or `/' character. For example:
125 List of floating numbers: A list of floating numbers, separated by
130 May be used to override the job name. On the command line, this parameter
131 has the special purpose of signalling the start of a new job.
133 .BI description \fR=\fPstr
134 Human-readable description of the job. It is printed when the job is run, but
135 otherwise has no special purpose.
137 .BI directory \fR=\fPstr
138 Prefix filenames with this directory. Used to place files in a location other
141 .BI filename \fR=\fPstr
143 normally makes up a file name based on the job name, thread number, and file
144 number. If you want to share files between threads in a job or several jobs,
145 specify a \fIfilename\fR for each of them to override the default.
146 If the I/O engine is file-based, you can specify
147 a number of files by separating the names with a `:' character. `\-' is a
148 reserved name, meaning stdin or stdout, depending on the read/write direction
151 .BI lockfile \fR=\fPstr
152 Fio defaults to not locking any files before it does IO to them. If a file or
153 file descriptor is shared, fio can serialize IO to that file to make the end
154 result consistent. This is usual for emulating real workloads that share files.
160 No locking. This is the default.
163 Only one thread or process may do IO at the time, excluding all others.
166 Read-write locking on the file. Many readers may access the file at the same
167 time, but writes get exclusive access.
170 The option may be post-fixed with a lock batch number. If set, then each
171 thread/process may do that amount of IOs to the file before giving up the lock.
172 Since lock acquisition is expensive, batching the lock/unlocks will speed up IO.
175 .BI opendir \fR=\fPstr
176 Recursively open any files below directory \fIstr\fR.
178 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
179 Type of I/O pattern. Accepted values are:
196 Mixed sequential reads and writes.
199 Mixed random reads and writes.
202 For mixed I/O, the default split is 50/50. For certain types of io the result
203 may still be skewed a bit, since the speed may be different. It is possible to
204 specify a number of IO's to do before getting a new offset, this is done by
205 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
206 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
207 value of 8. If the postfix is used with a sequential IO pattern, then the value
208 specified will be added to the generated offset for each IO. For instance,
209 using \fBrw=write:4k\fR will skip 4k for every write. It turns sequential IO
210 into sequential IO with holes. See the \fBrw_sequencer\fR option.
213 .BI rw_sequencer \fR=\fPstr
214 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
215 then this option controls how that number modifies the IO offset being
216 generated. Accepted values are:
221 Generate sequential offset
224 Generate the same offset
227 \fBsequential\fR is only useful for random IO, where fio would normally
228 generate a new random offset for every IO. If you append eg 8 to randread, you
229 would get a new random offset for every 8 IO's. The result would be a seek for
230 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
231 that. As sequential IO is already sequential, setting \fBsequential\fR for that
232 would not result in any differences. \fBidentical\fR behaves in a similar
233 fashion, except it sends the same offset 8 number of times before generating a
238 .BI kb_base \fR=\fPint
239 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
240 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
241 reasons. Allow values are 1024 or 1000, with 1024 being the default.
243 .BI unified_rw_reporting \fR=\fPbool
244 Fio normally reports statistics on a per data direction basis, meaning that
245 read, write, and trim are accounted and reported separately. If this option is
246 set, the fio will sum the results and report them as "mixed" instead.
248 .BI randrepeat \fR=\fPbool
249 Seed the random number generator in a predictable way so results are repeatable
250 across runs. Default: true.
252 .BI use_os_rand \fR=\fPbool
253 Fio can either use the random generator supplied by the OS to generator random
254 offsets, or it can use it's own internal generator (based on Tausworthe).
255 Default is to use the internal generator, which is often of better quality and
256 faster. Default: false.
258 .BI fallocate \fR=\fPstr
259 Whether pre-allocation is performed when laying down files. Accepted values
265 Do not pre-allocate space.
268 Pre-allocate via posix_fallocate().
271 Pre-allocate via fallocate() with FALLOC_FL_KEEP_SIZE set.
274 Backward-compatible alias for 'none'.
277 Backward-compatible alias for 'posix'.
280 May not be available on all supported platforms. 'keep' is only
281 available on Linux. If using ZFS on Solaris this must be set to 'none'
282 because ZFS doesn't support it. Default: 'posix'.
285 .BI fadvise_hint \fR=\fPbool
286 Use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
287 are likely to be issued. Default: true.
290 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
291 been transfered, unless limited by other options (\fBruntime\fR, for instance).
292 Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be
293 divided between the available files for the job. If not set, fio will use the
294 full size of the given files or devices. If the the files do not exist, size
295 must be given. It is also possible to give size as a percentage between 1 and
296 100. If size=20% is given, fio will use 20% of the full size of the given files
299 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
300 Sets size to something really large and waits for ENOSPC (no space left on
301 device) as the terminating condition. Only makes sense with sequential write.
302 For a read workload, the mount point will be filled first then IO started on
303 the result. This option doesn't make sense if operating on a raw device node,
304 since the size of that is already known by the file system. Additionally,
305 writing beyond end-of-device will not return ENOSPC there.
307 .BI filesize \fR=\fPirange
308 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
309 for files at random within the given range, limited to \fBsize\fR in total (if
310 that is given). If \fBfilesize\fR is not specified, each created file is the
313 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
314 Block size for I/O units. Default: 4k. Values for reads and writes can be
315 specified separately in the format \fIread\fR,\fIwrite\fR, either of
316 which may be empty to leave that value at its default.
318 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
319 Specify a range of I/O block sizes. The issued I/O unit will always be a
320 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
321 to both reads and writes if only one range is given, but can be specified
322 separately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k.
323 Also (see \fBblocksize\fR).
325 .BI bssplit \fR=\fPstr
326 This option allows even finer grained control of the block sizes issued,
327 not just even splits between them. With this option, you can weight various
328 block sizes for exact control of the issued IO for a job that has mixed
329 block sizes. The format of the option is bssplit=blocksize/percentage,
330 optionally adding as many definitions as needed separated by a colon.
331 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
332 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
333 splits to reads and writes. The format is identical to what the
334 \fBbs\fR option accepts, the read and write parts are separated with a
337 .B blocksize_unaligned\fR,\fP bs_unaligned
338 If set, any size in \fBblocksize_range\fR may be used. This typically won't
339 work with direct I/O, as that normally requires sector alignment.
341 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
342 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
343 the minimum blocksize given. Minimum alignment is typically 512b
344 for using direct IO, though it usually depends on the hardware block size.
345 This option is mutually exclusive with using a random map for files, so it
346 will turn off that option.
349 Initialise buffers with all zeros. Default: fill buffers with random data.
352 If this option is given, fio will refill the IO buffers on every submit. The
353 default is to only fill it at init time and reuse that data. Only makes sense
354 if zero_buffers isn't specified, naturally. If data verification is enabled,
355 refill_buffers is also automatically enabled.
357 .BI scramble_buffers \fR=\fPbool
358 If \fBrefill_buffers\fR is too costly and the target is using data
359 deduplication, then setting this option will slightly modify the IO buffer
360 contents to defeat normal de-dupe attempts. This is not enough to defeat
361 more clever block compression attempts, but it will stop naive dedupe
362 of blocks. Default: true.
364 .BI buffer_compress_percentage \fR=\fPint
365 If this is set, then fio will attempt to provide IO buffer content (on WRITEs)
366 that compress to the specified level. Fio does this by providing a mix of
367 random data and zeroes. Note that this is per block size unit, for file/disk
368 wide compression level that matches this setting, you'll also want to set
369 \fBrefill_buffers\fR.
371 .BI buffer_compress_chunk \fR=\fPint
372 See \fBbuffer_compress_percentage\fR. This setting allows fio to manage how
373 big the ranges of random data and zeroed data is. Without this set, fio will
374 provide \fBbuffer_compress_percentage\fR of blocksize random data, followed by
375 the remaining zeroed. With this set to some chunk size smaller than the block
376 size, fio can alternate random and zeroed data throughout the IO buffer.
378 .BI nrfiles \fR=\fPint
379 Number of files to use for this job. Default: 1.
381 .BI openfiles \fR=\fPint
382 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
384 .BI file_service_type \fR=\fPstr
385 Defines how files to service are selected. The following types are defined:
390 Choose a file at random
393 Round robin over open files (default).
395 Do each file in the set sequentially.
398 The number of I/Os to issue before switching a new file can be specified by
399 appending `:\fIint\fR' to the service type.
402 .BI ioengine \fR=\fPstr
403 Defines how the job issues I/O. The following types are defined:
408 Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
409 position the I/O location.
412 Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O.
415 Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by
416 coalescing adjacents IOs into a single submission.
419 Linux native asynchronous I/O. This ioengine defines engine specific options.
422 POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
425 Solaris native asynchronous I/O.
428 Windows native asynchronous I/O.
431 File is memory mapped with \fImmap\fR\|(2) and data copied using
435 \fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
436 transfer data from user-space to the kernel.
439 Use the syslet system calls to make regular read/write asynchronous.
442 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
443 the target is an sg character device, we use \fIread\fR\|(2) and
444 \fIwrite\fR\|(2) for asynchronous I/O.
447 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
448 itself and for debugging and testing purposes.
451 Transfer over the network. The protocol to be used can be defined with the
452 \fBprotocol\fR parameter. Depending on the protocol, \fBfilename\fR,
453 \fBhostname\fR, \fBport\fR, or \fBlisten\fR must be specified.
454 This ioengine defines engine specific options.
457 Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
458 and send/receive. This ioengine defines engine specific options.
461 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
462 \fBcpucycles\fR parameters.
465 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
466 approach to asycnronous I/O.
468 See <http://www.xmailserver.org/guasi\-lib.html>.
471 The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ)
472 and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols.
475 Loads an external I/O engine object file. Append the engine filename as
479 IO engine that does regular linux native fallocate callt to simulate data
480 transfer as fio ioengine
482 DDIR_READ does fallocate(,mode = FALLOC_FL_KEEP_SIZE,)
484 DIR_WRITE does fallocate(,mode = 0)
486 DDIR_TRIM does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE)
489 IO engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate defragment activity
490 request to DDIR_WRITE event
495 .BI iodepth \fR=\fPint
496 Number of I/O units to keep in flight against the file. Note that increasing
497 iodepth beyond 1 will not affect synchronous ioengines (except for small
498 degress when verify_async is in use). Even async engines my impose OS
499 restrictions causing the desired depth not to be achieved. This may happen on
500 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
501 not async on that OS. Keep an eye on the IO depth distribution in the
502 fio output to verify that the achieved depth is as expected. Default: 1.
504 .BI iodepth_batch \fR=\fPint
505 Number of I/Os to submit at once. Default: \fBiodepth\fR.
507 .BI iodepth_batch_complete \fR=\fPint
508 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
509 means that we'll ask for a minimum of 1 IO in the retrieval process from the
510 kernel. The IO retrieval will go on until we hit the limit set by
511 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
512 completed events before queuing more IO. This helps reduce IO latency, at the
513 cost of more retrieval system calls.
515 .BI iodepth_low \fR=\fPint
516 Low watermark indicating when to start filling the queue again. Default:
519 .BI direct \fR=\fPbool
520 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
522 .BI buffered \fR=\fPbool
523 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
526 .BI offset \fR=\fPint
527 Offset in the file to start I/O. Data before the offset will not be touched.
529 .BI offset_increment \fR=\fPint
530 If this is provided, then the real offset becomes the
531 offset + offset_increment * thread_number, where the thread number is a counter
532 that starts at 0 and is incremented for each job. This option is useful if
533 there are several jobs which are intended to operate on a file in parallel in
534 disjoint segments, with even spacing between the starting points.
537 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
538 0, don't sync. Default: 0.
540 .BI fdatasync \fR=\fPint
541 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
542 data parts of the file. Default: 0.
544 .BI sync_file_range \fR=\fPstr:int
545 Use sync_file_range() for every \fRval\fP number of write operations. Fio will
546 track range of writes that have happened since the last sync_file_range() call.
547 \fRstr\fP can currently be one or more of:
551 SYNC_FILE_RANGE_WAIT_BEFORE
554 SYNC_FILE_RANGE_WRITE
557 SYNC_FILE_RANGE_WRITE
561 So if you do sync_file_range=wait_before,write:8, fio would use
562 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
563 Also see the sync_file_range(2) man page. This option is Linux specific.
565 .BI overwrite \fR=\fPbool
566 If writing, setup the file first and do overwrites. Default: false.
568 .BI end_fsync \fR=\fPbool
569 Sync file contents when a write stage has completed. Default: false.
571 .BI fsync_on_close \fR=\fPbool
572 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
573 it will happen on every close, not just at the end of the job. Default: false.
575 .BI rwmixread \fR=\fPint
576 Percentage of a mixed workload that should be reads. Default: 50.
578 .BI rwmixwrite \fR=\fPint
579 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
580 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
581 overrides the first. This may interfere with a given rate setting, if fio is
582 asked to limit reads or writes to a certain rate. If that is the case, then
583 the distribution may be skewed. Default: 50.
585 .BI random_distribution \fR=\fPstr:float
586 By default, fio will use a completely uniform random distribution when asked
587 to perform random IO. Sometimes it is useful to skew the distribution in
588 specific ways, ensuring that some parts of the data is more hot than others.
589 Fio includes the following distribution models:
593 Uniform random distribution
603 When using a zipf or pareto distribution, an input value is also needed to
604 define the access pattern. For zipf, this is the zipf theta. For pareto,
605 it's the pareto power. Fio includes a test program, genzipf, that can be
606 used visualize what the given input values will yield in terms of hit rates.
607 If you wanted to use zipf with a theta of 1.2, you would use
608 random_distribution=zipf:1.2 as the option. If a non-uniform model is used,
609 fio will disable use of the random map.
612 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
613 this parameter is given, a new offset will be chosen without looking at past
614 I/O history. This parameter is mutually exclusive with \fBverify\fR.
616 .BI softrandommap \fR=\fPbool
617 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
618 fails to allocate the map, if this option is set it will continue without a
619 random block map. As coverage will not be as complete as with random maps, this
620 option is disabled by default.
622 .BI random_generator \fR=\fPstr
623 Fio supports the following engines for generating IO offsets for random IO:
627 Strong 2^88 cycle random number generator
630 Linear feedback shift register generator
634 Tausworthe is a strong random number generator, but it requires tracking on the
635 side if we want to ensure that blocks are only read or written once. LFSR
636 guarantees that we never generate the same offset twice, and it's also less
637 computationally expensive. It's not a true random generator, however, though
638 for IO purposes it's typically good enough. LFSR only works with single block
639 sizes, not with workloads that use multiple block sizes. If used with such a
640 workload, fio may read or write some blocks multiple times.
643 Run job with given nice value. See \fInice\fR\|(2).
646 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
649 .BI prioclass \fR=\fPint
650 Set I/O priority class. See \fIionice\fR\|(1).
652 .BI thinktime \fR=\fPint
653 Stall job for given number of microseconds between issuing I/Os.
655 .BI thinktime_spin \fR=\fPint
656 Pretend to spend CPU time for given number of microseconds, sleeping the rest
657 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
659 .BI thinktime_blocks \fR=\fPint
660 Number of blocks to issue before waiting \fBthinktime\fR microseconds.
664 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
665 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
666 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
667 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
668 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
669 limit writes (to 500KB/sec), the latter will only limit reads.
671 .BI ratemin \fR=\fPint
672 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
673 Failing to meet this requirement will cause the job to exit. The same format
674 as \fBrate\fR is used for read vs write separation.
676 .BI rate_iops \fR=\fPint
677 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
678 specified independently of bandwidth. The same format as \fBrate\fR is used for
679 read vs write seperation. If \fBblocksize\fR is a range, the smallest block
680 size is used as the metric.
682 .BI rate_iops_min \fR=\fPint
683 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
684 is used for read vs write seperation.
686 .BI ratecycle \fR=\fPint
687 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
688 milliseconds. Default: 1000ms.
690 .BI max_latency \fR=\fPint
691 If set, fio will exit the job if it exceeds this maximum latency. It will exit
694 .BI cpumask \fR=\fPint
695 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
696 may run on. See \fBsched_setaffinity\fR\|(2).
698 .BI cpus_allowed \fR=\fPstr
699 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
701 .BI numa_cpu_nodes \fR=\fPstr
702 Set this job running on spcified NUMA nodes' CPUs. The arguments allow
703 comma delimited list of cpu numbers, A-B ranges, or 'all'.
705 .BI numa_mem_policy \fR=\fPstr
706 Set this job's memory policy and corresponding NUMA nodes. Format of
710 .B <mode>[:<nodelist>]
713 is one of the following memory policy:
715 .B default, prefer, bind, interleave, local
718 For \fBdefault\fR and \fBlocal\fR memory policy, no \fBnodelist\fR is
719 needed to be specified. For \fBprefer\fR, only one node is
720 allowed. For \fBbind\fR and \fBinterleave\fR, \fBnodelist\fR allows
721 comma delimited list of numbers, A-B ranges, or 'all'.
723 .BI startdelay \fR=\fPint
724 Delay start of job for the specified number of seconds.
726 .BI runtime \fR=\fPint
727 Terminate processing after the specified number of seconds.
730 If given, run for the specified \fBruntime\fR duration even if the files are
731 completely read or written. The same workload will be repeated as many times
732 as \fBruntime\fR allows.
734 .BI ramp_time \fR=\fPint
735 If set, fio will run the specified workload for this amount of time before
736 logging any performance numbers. Useful for letting performance settle before
737 logging results, thus minimizing the runtime required for stable results. Note
738 that the \fBramp_time\fR is considered lead in time for a job, thus it will
739 increase the total runtime if a special timeout or runtime is specified.
741 .BI invalidate \fR=\fPbool
742 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
745 Use synchronous I/O for buffered writes. For the majority of I/O engines,
746 this means using O_SYNC. Default: false.
748 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
749 Allocation method for I/O unit buffer. Allowed values are:
754 Allocate memory with \fImalloc\fR\|(3).
757 Use shared memory buffers allocated through \fIshmget\fR\|(2).
760 Same as \fBshm\fR, but use huge pages as backing.
763 Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
764 is given after the option in the format `:\fIfile\fR'.
767 Same as \fBmmap\fR, but use huge files as backing.
770 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
771 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
772 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
773 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
774 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
775 and the documentation for that. Normally you just need to echo an appropriate
776 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
780 .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint
781 This indiciates the memory alignment of the IO memory buffers. Note that the
782 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
783 the alignment of the following buffers are given by the \fBbs\fR used. In
784 other words, if using a \fBbs\fR that is a multiple of the page sized in the
785 system, all buffers will be aligned to this value. If using a \fBbs\fR that
786 is not page aligned, the alignment of subsequent IO memory buffers is the
787 sum of the \fBiomem_align\fR and \fBbs\fR used.
789 .BI hugepage\-size \fR=\fPint
790 Defines the size of a huge page. Must be at least equal to the system setting.
791 Should be a multiple of 1MB. Default: 4MB.
794 Terminate all jobs when one finishes. Default: wait for each job to finish.
796 .BI bwavgtime \fR=\fPint
797 Average bandwidth calculations over the given time in milliseconds. Default:
800 .BI iopsavgtime \fR=\fPint
801 Average IOPS calculations over the given time in milliseconds. Default:
804 .BI create_serialize \fR=\fPbool
805 If true, serialize file creation for the jobs. Default: true.
807 .BI create_fsync \fR=\fPbool
808 \fIfsync\fR\|(2) data file after creation. Default: true.
810 .BI create_on_open \fR=\fPbool
811 If true, the files are not created until they are opened for IO by the job.
813 .BI create_only \fR=\fPbool
814 If true, fio will only run the setup phase of the job. If files need to be
815 laid out or updated on disk, only that will be done. The actual job contents
818 .BI pre_read \fR=\fPbool
819 If this is given, files will be pre-read into memory before starting the given
820 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
821 pointless to pre-read and then drop the cache. This will only work for IO
822 engines that are seekable, since they allow you to read the same data
823 multiple times. Thus it will not work on eg network or splice IO.
825 .BI unlink \fR=\fPbool
826 Unlink job files when done. Default: false.
829 Specifies the number of iterations (runs of the same workload) of this job.
832 .BI do_verify \fR=\fPbool
833 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
836 .BI verify \fR=\fPstr
837 Method of verifying file contents after each iteration of the job. Allowed
842 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1
843 Store appropriate checksum in the header of each block. crc32c-intel is
844 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
845 not supported by the system.
848 Write extra information about each I/O (timestamp, block number, etc.). The
849 block number is verified. See \fBverify_pattern\fR as well.
852 Pretend to verify. Used for testing internals.
855 This option can be used for repeated burn-in tests of a system to make sure
856 that the written data is also correctly read back. If the data direction given
857 is a read or random read, fio will assume that it should verify a previously
858 written file. If the data direction includes any form of write, the verify will
859 be of the newly written data.
862 .BI verify_sort \fR=\fPbool
863 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
864 read them back in a sorted manner. Default: true.
866 .BI verify_offset \fR=\fPint
867 Swap the verification header with data somewhere else in the block before
868 writing. It is swapped back before verifying.
870 .BI verify_interval \fR=\fPint
871 Write the verification header for this number of bytes, which should divide
872 \fBblocksize\fR. Default: \fBblocksize\fR.
874 .BI verify_pattern \fR=\fPstr
875 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
876 with totally random bytes, but sometimes it's interesting to fill with a known
877 pattern for io verification purposes. Depending on the width of the pattern,
878 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
879 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
880 has to be a hex number that starts with either "0x" or "0X". Use with
883 .BI verify_fatal \fR=\fPbool
884 If true, exit the job on the first observed verification failure. Default:
887 .BI verify_dump \fR=\fPbool
888 If set, dump the contents of both the original data block and the data block we
889 read off disk to files. This allows later analysis to inspect just what kind of
890 data corruption occurred. Off by default.
892 .BI verify_async \fR=\fPint
893 Fio will normally verify IO inline from the submitting thread. This option
894 takes an integer describing how many async offload threads to create for IO
895 verification instead, causing fio to offload the duty of verifying IO contents
896 to one or more separate threads. If using this offload option, even sync IO
897 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
898 allows them to have IO in flight while verifies are running.
900 .BI verify_async_cpus \fR=\fPstr
901 Tell fio to set the given CPU affinity on the async IO verification threads.
902 See \fBcpus_allowed\fP for the format used.
904 .BI verify_backlog \fR=\fPint
905 Fio will normally verify the written contents of a job that utilizes verify
906 once that job has completed. In other words, everything is written then
907 everything is read back and verified. You may want to verify continually
908 instead for a variety of reasons. Fio stores the meta data associated with an
909 IO block in memory, so for large verify workloads, quite a bit of memory would
910 be used up holding this meta data. If this option is enabled, fio will write
911 only N blocks before verifying these blocks.
913 .BI verify_backlog_batch \fR=\fPint
914 Control how many blocks fio will verify if verify_backlog is set. If not set,
915 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
916 read back and verified). If \fBverify_backlog_batch\fR is less than
917 \fBverify_backlog\fR then not all blocks will be verified, if
918 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
919 will be verified more than once.
921 .B stonewall "\fR,\fP wait_for_previous"
922 Wait for preceding jobs in the job file to exit before starting this one.
923 \fBstonewall\fR implies \fBnew_group\fR.
926 Start a new reporting group. If not given, all jobs in a file will be part
927 of the same reporting group, unless separated by a stonewall.
929 .BI numjobs \fR=\fPint
930 Number of clones (processes/threads performing the same workload) of this job.
934 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
938 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
939 with \fBfork\fR\|(2).
941 .BI zonesize \fR=\fPint
942 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
944 .BI zoneskip \fR=\fPint
945 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
948 .BI write_iolog \fR=\fPstr
949 Write the issued I/O patterns to the specified file. Specify a separate file
950 for each job, otherwise the iologs will be interspersed and the file may be
953 .BI read_iolog \fR=\fPstr
954 Replay the I/O patterns contained in the specified file generated by
955 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
957 .BI replay_no_stall \fR=\fPint
958 While replaying I/O patterns using \fBread_iolog\fR the default behavior
959 attempts to respect timing information between I/Os. Enabling
960 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
961 still respecting ordering.
963 .BI replay_redirect \fR=\fPstr
964 While replaying I/O patterns using \fBread_iolog\fR the default behavior
965 is to replay the IOPS onto the major/minor device that each IOP was recorded
966 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
967 single specified device regardless of the device it was recorded from.
969 .BI write_bw_log \fR=\fPstr
970 If given, write a bandwidth log of the jobs in this job file. Can be used to
971 store data of the bandwidth of the jobs in their lifetime. The included
972 fio_generate_plots script uses gnuplot to turn these text files into nice
973 graphs. See \fBwrite_log_log\fR for behaviour of given filename. For this
974 option, the postfix is _bw.log.
976 .BI write_lat_log \fR=\fPstr
977 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
978 filename is given with this option, the default filename of "jobname_type.log"
979 is used. Even if the filename is given, fio will still append the type of log.
981 .BI write_iops_log \fR=\fPstr
982 Same as \fBwrite_bw_log\fR, but writes IOPS. If no filename is given with this
983 option, the default filename of "jobname_type.log" is used. Even if the
984 filename is given, fio will still append the type of log.
986 .BI log_avg_msec \fR=\fPint
987 By default, fio will log an entry in the iops, latency, or bw log for every
988 IO that completes. When writing to the disk log, that can quickly grow to a
989 very large size. Setting this option makes fio average the each log entry
990 over the specified period of time, reducing the resolution of the log.
993 .BI disable_lat \fR=\fPbool
994 Disable measurements of total latency numbers. Useful only for cutting
995 back the number of calls to gettimeofday, as that does impact performance at
996 really high IOPS rates. Note that to really get rid of a large amount of these
997 calls, this option must be used with disable_slat and disable_bw as well.
999 .BI disable_clat \fR=\fPbool
1000 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
1002 .BI disable_slat \fR=\fPbool
1003 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
1005 .BI disable_bw_measurement \fR=\fPbool
1006 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
1008 .BI lockmem \fR=\fPint
1009 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
1010 simulate a smaller amount of memory.
1012 .BI exec_prerun \fR=\fPstr
1013 Before running the job, execute the specified command with \fBsystem\fR\|(3).
1015 .BI exec_postrun \fR=\fPstr
1016 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
1018 .BI ioscheduler \fR=\fPstr
1019 Attempt to switch the device hosting the file to the specified I/O scheduler.
1021 .BI cpuload \fR=\fPint
1022 If the job is a CPU cycle-eater, attempt to use the specified percentage of
1025 .BI cpuchunks \fR=\fPint
1026 If the job is a CPU cycle-eater, split the load into cycles of the
1027 given time in milliseconds.
1029 .BI disk_util \fR=\fPbool
1030 Generate disk utilization statistics if the platform supports it. Default: true.
1032 .BI clocksource \fR=\fPstr
1033 Use the given clocksource as the base of timing. The supported options are:
1043 Internal CPU clock source
1047 \fBcpu\fR is the preferred clocksource if it is reliable, as it is very fast
1048 (and fio is heavy on time calls). Fio will automatically use this clocksource
1049 if it's supported and considered reliable on the system it is running on,
1050 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
1051 means supporting TSC Invariant.
1053 .BI gtod_reduce \fR=\fPbool
1054 Enable all of the gettimeofday() reducing options (disable_clat, disable_slat,
1055 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
1056 gettimeofday() call count. With this option enabled, we only do about 0.4% of
1057 the gtod() calls we would have done if all time keeping was enabled.
1059 .BI gtod_cpu \fR=\fPint
1060 Sometimes it's cheaper to dedicate a single thread of execution to just getting
1061 the current time. Fio (and databases, for instance) are very intensive on
1062 gettimeofday() calls. With this option, you can set one CPU aside for doing
1063 nothing but logging current time to a shared memory location. Then the other
1064 threads/processes that run IO workloads need only copy that segment, instead of
1065 entering the kernel with a gettimeofday() call. The CPU set aside for doing
1066 these time calls will be excluded from other uses. Fio will manually clear it
1067 from the CPU mask of other jobs.
1069 .BI ignore_error \fR=\fPstr
1070 Sometimes you want to ignore some errors during test in that case you can specify
1071 error list for each error type.
1073 ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST
1075 errors for given error type is separated with ':'.
1076 Error may be symbol ('ENOSPC', 'ENOMEM') or an integer.
1078 Example: ignore_error=EAGAIN,ENOSPC:122 .
1080 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from WRITE.
1082 .BI error_dump \fR=\fPbool
1083 If set dump every error even if it is non fatal, true by default. If disabled
1084 only fatal error will be dumped
1086 .BI cgroup \fR=\fPstr
1087 Add job to this control group. If it doesn't exist, it will be created.
1088 The system must have a mounted cgroup blkio mount point for this to work. If
1089 your system doesn't have it mounted, you can do so with:
1091 # mount \-t cgroup \-o blkio none /cgroup
1093 .BI cgroup_weight \fR=\fPint
1094 Set the weight of the cgroup to this value. See the documentation that comes
1095 with the kernel, allowed values are in the range of 100..1000.
1097 .BI cgroup_nodelete \fR=\fPbool
1098 Normally fio will delete the cgroups it has created after the job completion.
1099 To override this behavior and to leave cgroups around after the job completion,
1100 set cgroup_nodelete=1. This can be useful if one wants to inspect various
1101 cgroup files after job completion. Default: false
1104 Instead of running as the invoking user, set the user ID to this value before
1105 the thread/process does any work.
1108 Set group ID, see \fBuid\fR.
1110 .BI flow_id \fR=\fPint
1111 The ID of the flow. If not specified, it defaults to being a global flow. See
1115 Weight in token-based flow control. If this value is used, then there is a
1116 \fBflow counter\fR which is used to regulate the proportion of activity between
1117 two or more jobs. fio attempts to keep this flow counter near zero. The
1118 \fBflow\fR parameter stands for how much should be added or subtracted to the
1119 flow counter on each iteration of the main I/O loop. That is, if one job has
1120 \fBflow=8\fR and another job has \fBflow=-1\fR, then there will be a roughly
1121 1:8 ratio in how much one runs vs the other.
1123 .BI flow_watermark \fR=\fPint
1124 The maximum value that the absolute value of the flow counter is allowed to
1125 reach before the job must wait for a lower value of the counter.
1127 .BI flow_sleep \fR=\fPint
1128 The period of time, in microseconds, to wait after the flow watermark has been
1129 exceeded before retrying operations
1131 .BI clat_percentiles \fR=\fPbool
1132 Enable the reporting of percentiles of completion latencies.
1134 .BI percentile_list \fR=\fPfloat_list
1135 Overwrite the default list of percentiles for completion
1136 latencies. Each number is a floating number in the range (0,100], and
1137 the maximum length of the list is 20. Use ':' to separate the
1138 numbers. For example, \-\-percentile_list=99.5:99.9 will cause fio to
1139 report the values of completion latency below which 99.5% and 99.9% of
1140 the observed latencies fell, respectively.
1141 .SS "Ioengine Parameters List"
1142 Some parameters are only valid when a specific ioengine is in use. These are
1143 used identically to normal parameters, with the caveat that when used on the
1144 command line, the must come after the ioengine that defines them is selected.
1146 .BI (libaio)userspace_reap
1147 Normally, with the libaio engine in use, fio will use
1148 the io_getevents system call to reap newly returned events.
1149 With this flag turned on, the AIO ring will be read directly
1150 from user-space to reap events. The reaping mode is only
1151 enabled when polling for a minimum of 0 events (eg when
1152 iodepth_batch_complete=0).
1154 .BI (net,netsplice)hostname \fR=\fPstr
1155 The host name or IP address to use for TCP or UDP based IO.
1156 If the job is a TCP listener or UDP reader, the hostname is not
1157 used and must be omitted.
1159 .BI (net,netsplice)port \fR=\fPint
1160 The TCP or UDP port to bind to or connect to.
1162 .BI (net,netsplice)protocol \fR=\fPstr "\fR,\fP proto" \fR=\fPstr
1163 The network protocol to use. Accepted values are:
1168 Transmission control protocol
1171 User datagram protocol
1177 When the protocol is TCP or UDP, the port must also be given,
1178 as well as the hostname if the job is a TCP listener or UDP
1179 reader. For unix sockets, the normal filename option should be
1180 used and the port is invalid.
1183 .BI (net,netsplice)listen
1184 For TCP network connections, tell fio to listen for incoming
1185 connections rather than initiating an outgoing connection. The
1186 hostname must be omitted if this option is used.
1188 .BI (net, pingpong) \fR=\fPbool
1189 Normal a network writer will just continue writing data, and a network reader
1190 will just consume packages. If pingpong=1 is set, a writer will send its normal
1191 payload to the reader, then wait for the reader to send the same payload back.
1192 This allows fio to measure network latencies. The submission and completion
1193 latencies then measure local time spent sending or receiving, and the
1194 completion latency measures how long it took for the other end to receive and
1197 .BI (e4defrag,donorname) \fR=\fPstr
1198 File will be used as a block donor (swap extents between files)
1200 .BI (e4defrag,inplace) \fR=\fPint
1201 Configure donor file block allocation strategy
1204 Preallocate donor's file on init
1207 allocate space immidietly inside defragment event, and free right after event
1211 While running, \fBfio\fR will display the status of the created jobs. For
1215 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
1218 The characters in the first set of brackets denote the current status of each
1219 threads. The possible values are:
1225 Setup but not started.
1231 Initialized, waiting.
1234 Running, doing sequential reads.
1237 Running, doing random reads.
1240 Running, doing sequential writes.
1243 Running, doing random writes.
1246 Running, doing mixed sequential reads/writes.
1249 Running, doing mixed random reads/writes.
1252 Running, currently waiting for \fBfsync\fR\|(2).
1255 Running, verifying written data.
1258 Exited, not reaped by main thread.
1261 Exited, thread reaped.
1265 The second set of brackets shows the estimated completion percentage of
1266 the current group. The third set shows the read and write I/O rate,
1267 respectively. Finally, the estimated run time of the job is displayed.
1269 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
1270 for each thread, each group of threads, and each disk, in that order.
1272 Per-thread statistics first show the threads client number, group-id, and
1273 error code. The remaining figures are as follows:
1277 Number of megabytes of I/O performed.
1280 Average data rate (bandwidth).
1286 Submission latency minimum, maximum, average and standard deviation. This is
1287 the time it took to submit the I/O.
1290 Completion latency minimum, maximum, average and standard deviation. This
1291 is the time between submission and completion.
1294 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
1295 and standard deviation.
1298 CPU usage statistics. Includes user and system time, number of context switches
1299 this thread went through and number of major and minor page faults.
1302 Distribution of I/O depths. Each depth includes everything less than (or equal)
1303 to it, but greater than the previous depth.
1306 Number of read/write requests issued, and number of short read/write requests.
1309 Distribution of I/O completion latencies. The numbers follow the same pattern
1313 The group statistics show:
1318 Number of megabytes I/O performed.
1321 Aggregate bandwidth of threads in the group.
1324 Minimum average bandwidth a thread saw.
1327 Maximum average bandwidth a thread saw.
1330 Shortest runtime of threads in the group.
1333 Longest runtime of threads in the group.
1337 Finally, disk statistics are printed with reads first:
1342 Number of I/Os performed by all groups.
1345 Number of merges in the I/O scheduler.
1348 Number of ticks we kept the disk busy.
1351 Total time spent in the disk queue.
1358 It is also possible to get fio to dump the current output while it is
1359 running, without terminating the job. To do that, send fio the \fBUSR1\fR
1362 If the \fB\-\-minimal\fR option is given, the results will be printed in a
1363 semicolon-delimited format suitable for scripted use - a job description
1364 (if provided) follows on a new line. Note that the first
1365 number in the line is the version number. If the output has to be changed
1366 for some reason, this number will be incremented by 1 to signify that
1367 change. The fields are:
1370 .B terse version, fio version, jobname, groupid, error
1374 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1378 .B min, max, mean, standard deviation
1382 .B min, max, mean, standard deviation
1384 Completion latency percentiles (20 fields):
1386 .B Xth percentile=usec
1390 .B min, max, mean, standard deviation
1394 .B min, max, aggregate percentage of total, mean, standard deviation
1400 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1404 .B min, max, mean, standard deviation
1408 .B min, max, mean, standard deviation
1410 Completion latency percentiles (20 fields):
1412 .B Xth percentile=usec
1416 .B min, max, mean, standard deviation
1420 .B min, max, aggregate percentage of total, mean, standard deviation
1426 .B user, system, context switches, major page faults, minor page faults
1429 IO depth distribution:
1431 .B <=1, 2, 4, 8, 16, 32, >=64
1434 IO latency distribution:
1438 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1442 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1446 Disk utilization (1 for each disk used):
1448 .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
1451 Error Info (dependent on continue_on_error, default off):
1453 .B total # errors, first error code
1456 .B text description (if provided in config - appears on newline)
1459 Normally you would run fio as a stand-alone application on the machine
1460 where the IO workload should be generated. However, it is also possible to
1461 run the frontend and backend of fio separately. This makes it possible to
1462 have a fio server running on the machine(s) where the IO workload should
1463 be running, while controlling it from another machine.
1465 To start the server, you would do:
1467 \fBfio \-\-server=args\fR
1469 on that machine, where args defines what fio listens to. The arguments
1470 are of the form 'type:hostname or IP:port'. 'type' is either 'ip' (or ip4)
1471 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain
1472 socket. 'hostname' is either a hostname or IP address, and 'port' is the port to
1473 listen to (only valid for TCP/IP, not a local socket). Some examples:
1477 Start a fio server, listening on all interfaces on the default port (8765).
1479 2) fio \-\-server=ip:hostname,4444
1481 Start a fio server, listening on IP belonging to hostname and on port 4444.
1483 3) fio \-\-server=ip6:::1,4444
1485 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
1487 4) fio \-\-server=,4444
1489 Start a fio server, listening on all interfaces on port 4444.
1491 5) fio \-\-server=1.2.3.4
1493 Start a fio server, listening on IP 1.2.3.4 on the default port.
1495 6) fio \-\-server=sock:/tmp/fio.sock
1497 Start a fio server, listening on the local socket /tmp/fio.sock.
1499 When a server is running, you can connect to it from a client. The client
1502 fio \-\-local-args \-\-client=server \-\-remote-args <job file(s)>
1504 where \-\-local-args are arguments that are local to the client where it is
1505 running, 'server' is the connect string, and \-\-remote-args and <job file(s)>
1506 are sent to the server. The 'server' string follows the same format as it
1507 does on the server side, to allow IP/hostname/socket and port strings.
1508 You can connect to multiple clients as well, to do that you could run:
1510 fio \-\-client=server2 \-\-client=server2 <job file(s)>
1514 was written by Jens Axboe <jens.axboe@oracle.com>,
1515 now Jens Axboe <jaxboe@fusionio.com>.
1517 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1518 on documentation by Jens Axboe.
1519 .SH "REPORTING BUGS"
1520 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1523 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1525 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob