1 .TH fio 1 "October 2013" "User Manual"
3 fio \- flexible I/O tester
6 [\fIoptions\fR] [\fIjobfile\fR]...
9 is a tool that will spawn a number of threads or processes doing a
10 particular type of I/O action as specified by the user.
11 The typical use of fio is to write a job file matching the I/O load
12 one wants to simulate.
15 .BI \-\-debug \fR=\fPtype
16 Enable verbose tracing of various fio actions. May be `all' for all types
17 or individual types separated by a comma (eg \-\-debug=io,file). `help' will
18 list all available tracing options.
20 .BI \-\-output \fR=\fPfilename
21 Write output to \fIfilename\fR.
23 .BI \-\-runtime \fR=\fPruntime
24 Limit run time to \fIruntime\fR seconds.
27 Generate per-job latency logs.
30 Generate per-job bandwidth logs.
33 Print statistics in a terse, semicolon-delimited format.
36 Display version information and exit.
38 .BI \-\-terse\-version \fR=\fPversion
39 Set terse version output format (Current version 3, or older version 2).
42 Display usage information and exit.
44 .BI \-\-cmdhelp \fR=\fPcommand
45 Print help information for \fIcommand\fR. May be `all' for all commands.
47 .BI \-\-enghelp \fR=\fPioengine[,command]
48 List all commands defined by \fIioengine\fR, or print help for \fIcommand\fR defined by \fIioengine\fR.
50 .BI \-\-showcmd \fR=\fPjobfile
51 Convert \fIjobfile\fR to a set of command-line options.
53 .BI \-\-eta \fR=\fPwhen
54 Specifies when real-time ETA estimate should be printed. \fIwhen\fR may
55 be one of `always', `never' or `auto'.
57 .BI \-\-eta\-newline \fR=\fPtime
58 Force an ETA newline for every `time` period passed.
60 .BI \-\-status\-interval \fR=\fPtime
61 Report full output status every `time` period passed.
64 Turn on safety read-only checks, preventing any attempted write.
66 .BI \-\-section \fR=\fPsec
67 Only run section \fIsec\fR from job file. Multiple of these options can be given, adding more sections to run.
69 .BI \-\-alloc\-size \fR=\fPkb
70 Set the internal smalloc pool size to \fIkb\fP kilobytes.
72 .BI \-\-warnings\-fatal
73 All fio parser warnings are fatal, causing fio to exit with an error.
75 .BI \-\-max\-jobs \fR=\fPnr
76 Set the maximum allowed number of jobs (threads/processes) to support.
78 .BI \-\-server \fR=\fPargs
79 Start a backend server, with \fIargs\fP specifying what to listen to. See client/server section.
81 .BI \-\-daemonize \fR=\fPpidfile
82 Background a fio server, writing the pid to the given pid file.
84 .BI \-\-client \fR=\fPhost
85 Instead of running the jobs locally, send and run them on the given host.
87 .BI \-\-idle\-prof \fR=\fPoption
88 Report cpu idleness on a system or percpu basis (\fIoption\fP=system,percpu) or run unit work calibration only (\fIoption\fP=calibrate).
90 Job files are in `ini' format. They consist of one or more
91 job definitions, which begin with a job name in square brackets and
92 extend to the next job name. The job name can be any ASCII string
93 except `global', which has a special meaning. Following the job name is
94 a sequence of zero or more parameters, one per line, that define the
95 behavior of the job. Any line starting with a `;' or `#' character is
96 considered a comment and ignored.
98 If \fIjobfile\fR is specified as `-', the job file will be read from
101 The global section contains default parameters for jobs specified in the
102 job file. A job is only affected by global sections residing above it,
103 and there may be any number of global sections. Specific job definitions
104 may override any parameter set in global sections.
107 Some parameters may take arguments of a specific type. The types used are:
110 String: a sequence of alphanumeric characters.
113 SI integer: a whole number, possibly containing a suffix denoting the base unit
114 of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting
115 kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5)
116 respectively. The suffix is not case sensitive. If prefixed with '0x', the
117 value is assumed to be base 16 (hexadecimal). A suffix may include a trailing 'b',
118 for instance 'kb' is identical to 'k'. You can specify a base 10 value
119 by using 'KiB', 'MiB', 'GiB', etc. This is useful for disk drives where
120 values are often given in base 10 values. Specifying '30GiB' will get you
124 Boolean: a true or false value. `0' denotes false, `1' denotes true.
127 Integer range: a range of integers specified in the format
128 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
129 \fIupper\fR may contain a suffix as described above. If an option allows two
130 sets of ranges, they are separated with a `,' or `/' character. For example:
134 List of floating numbers: A list of floating numbers, separated by
139 May be used to override the job name. On the command line, this parameter
140 has the special purpose of signalling the start of a new job.
142 .BI description \fR=\fPstr
143 Human-readable description of the job. It is printed when the job is run, but
144 otherwise has no special purpose.
146 .BI directory \fR=\fPstr
147 Prefix filenames with this directory. Used to place files in a location other
150 .BI filename \fR=\fPstr
152 normally makes up a file name based on the job name, thread number, and file
153 number. If you want to share files between threads in a job or several jobs,
154 specify a \fIfilename\fR for each of them to override the default.
155 If the I/O engine is file-based, you can specify
156 a number of files by separating the names with a `:' character. `\-' is a
157 reserved name, meaning stdin or stdout, depending on the read/write direction
160 .BI filename_format \fR=\fPstr
161 If sharing multiple files between jobs, it is usually necessary to have
162 fio generate the exact names that you want. By default, fio will name a file
163 based on the default file format specification of
164 \fBjobname.jobnumber.filenumber\fP. With this option, that can be
165 customized. Fio will recognize and replace the following keywords in this
171 The name of the worker thread or process.
174 The incremental number of the worker thread or process.
177 The incremental number of the file for that worker thread or process.
180 To have dependent jobs share a set of files, this option can be set to
181 have fio generate filenames that are shared between the two. For instance,
182 if \fBtestfiles.$filenum\fR is specified, file number 4 for any job will
183 be named \fBtestfiles.4\fR. The default of \fB$jobname.$jobnum.$filenum\fR
184 will be used if no other format specifier is given.
188 .BI lockfile \fR=\fPstr
189 Fio defaults to not locking any files before it does IO to them. If a file or
190 file descriptor is shared, fio can serialize IO to that file to make the end
191 result consistent. This is usual for emulating real workloads that share files.
197 No locking. This is the default.
200 Only one thread or process may do IO at the time, excluding all others.
203 Read-write locking on the file. Many readers may access the file at the same
204 time, but writes get exclusive access.
208 .BI opendir \fR=\fPstr
209 Recursively open any files below directory \fIstr\fR.
211 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
212 Type of I/O pattern. Accepted values are:
223 Sequential trim (Linux block devices only).
232 Random trim (Linux block devices only).
235 Mixed sequential reads and writes.
238 Mixed random reads and writes.
241 For mixed I/O, the default split is 50/50. For certain types of io the result
242 may still be skewed a bit, since the speed may be different. It is possible to
243 specify a number of IO's to do before getting a new offset, this is done by
244 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
245 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
246 value of 8. If the postfix is used with a sequential IO pattern, then the value
247 specified will be added to the generated offset for each IO. For instance,
248 using \fBrw=write:4k\fR will skip 4k for every write. It turns sequential IO
249 into sequential IO with holes. See the \fBrw_sequencer\fR option.
252 .BI rw_sequencer \fR=\fPstr
253 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
254 then this option controls how that number modifies the IO offset being
255 generated. Accepted values are:
260 Generate sequential offset
263 Generate the same offset
266 \fBsequential\fR is only useful for random IO, where fio would normally
267 generate a new random offset for every IO. If you append eg 8 to randread, you
268 would get a new random offset for every 8 IO's. The result would be a seek for
269 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
270 that. As sequential IO is already sequential, setting \fBsequential\fR for that
271 would not result in any differences. \fBidentical\fR behaves in a similar
272 fashion, except it sends the same offset 8 number of times before generating a
277 .BI kb_base \fR=\fPint
278 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
279 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
280 reasons. Allowed values are 1024 or 1000, with 1024 being the default.
282 .BI unified_rw_reporting \fR=\fPbool
283 Fio normally reports statistics on a per data direction basis, meaning that
284 read, write, and trim are accounted and reported separately. If this option is
285 set, the fio will sum the results and report them as "mixed" instead.
287 .BI randrepeat \fR=\fPbool
288 Seed the random number generator in a predictable way so results are repeatable
289 across runs. Default: true.
291 .BI use_os_rand \fR=\fPbool
292 Fio can either use the random generator supplied by the OS to generator random
293 offsets, or it can use it's own internal generator (based on Tausworthe).
294 Default is to use the internal generator, which is often of better quality and
295 faster. Default: false.
297 .BI fallocate \fR=\fPstr
298 Whether pre-allocation is performed when laying down files. Accepted values
304 Do not pre-allocate space.
307 Pre-allocate via \fBposix_fallocate\fR\|(3).
310 Pre-allocate via \fBfallocate\fR\|(2) with FALLOC_FL_KEEP_SIZE set.
313 Backward-compatible alias for 'none'.
316 Backward-compatible alias for 'posix'.
319 May not be available on all supported platforms. 'keep' is only
320 available on Linux. If using ZFS on Solaris this must be set to 'none'
321 because ZFS doesn't support it. Default: 'posix'.
324 .BI fadvise_hint \fR=\fPbool
325 Use of \fBposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
326 are likely to be issued. Default: true.
329 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
330 been transferred, unless limited by other options (\fBruntime\fR, for instance).
331 Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be
332 divided between the available files for the job. If not set, fio will use the
333 full size of the given files or devices. If the files do not exist, size
334 must be given. It is also possible to give size as a percentage between 1 and
335 100. If size=20% is given, fio will use 20% of the full size of the given files
338 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
339 Sets size to something really large and waits for ENOSPC (no space left on
340 device) as the terminating condition. Only makes sense with sequential write.
341 For a read workload, the mount point will be filled first then IO started on
342 the result. This option doesn't make sense if operating on a raw device node,
343 since the size of that is already known by the file system. Additionally,
344 writing beyond end-of-device will not return ENOSPC there.
346 .BI filesize \fR=\fPirange
347 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
348 for files at random within the given range, limited to \fBsize\fR in total (if
349 that is given). If \fBfilesize\fR is not specified, each created file is the
352 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
353 Block size for I/O units. Default: 4k. Values for reads, writes, and trims
354 can be specified separately in the format \fIread\fR,\fIwrite\fR,\fItrim\fR
355 either of which may be empty to leave that value at its default. If a trailing
356 comma isn't given, the remainder will inherit the last value set.
358 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
359 Specify a range of I/O block sizes. The issued I/O unit will always be a
360 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
361 to both reads and writes if only one range is given, but can be specified
362 separately with a comma separating the values. Example: bsrange=1k-4k,2k-8k.
363 Also (see \fBblocksize\fR).
365 .BI bssplit \fR=\fPstr
366 This option allows even finer grained control of the block sizes issued,
367 not just even splits between them. With this option, you can weight various
368 block sizes for exact control of the issued IO for a job that has mixed
369 block sizes. The format of the option is bssplit=blocksize/percentage,
370 optionally adding as many definitions as needed separated by a colon.
371 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
372 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
373 splits to reads and writes. The format is identical to what the
374 \fBbs\fR option accepts, the read and write parts are separated with a
377 .B blocksize_unaligned\fR,\fP bs_unaligned
378 If set, any size in \fBblocksize_range\fR may be used. This typically won't
379 work with direct I/O, as that normally requires sector alignment.
381 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
382 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
383 the minimum blocksize given. Minimum alignment is typically 512b
384 for using direct IO, though it usually depends on the hardware block size.
385 This option is mutually exclusive with using a random map for files, so it
386 will turn off that option.
388 .BI bs_is_seq_rand \fR=\fPbool
389 If this option is set, fio will use the normal read,write blocksize settings as
390 sequential,random instead. Any random read or write will use the WRITE
391 blocksize settings, and any sequential read or write will use the READ
395 Initialise buffers with all zeros. Default: fill buffers with random data.
398 If this option is given, fio will refill the IO buffers on every submit. The
399 default is to only fill it at init time and reuse that data. Only makes sense
400 if zero_buffers isn't specified, naturally. If data verification is enabled,
401 refill_buffers is also automatically enabled.
403 .BI scramble_buffers \fR=\fPbool
404 If \fBrefill_buffers\fR is too costly and the target is using data
405 deduplication, then setting this option will slightly modify the IO buffer
406 contents to defeat normal de-dupe attempts. This is not enough to defeat
407 more clever block compression attempts, but it will stop naive dedupe
408 of blocks. Default: true.
410 .BI buffer_compress_percentage \fR=\fPint
411 If this is set, then fio will attempt to provide IO buffer content (on WRITEs)
412 that compress to the specified level. Fio does this by providing a mix of
413 random data and zeroes. Note that this is per block size unit, for file/disk
414 wide compression level that matches this setting, you'll also want to set
415 \fBrefill_buffers\fR.
417 .BI buffer_compress_chunk \fR=\fPint
418 See \fBbuffer_compress_percentage\fR. This setting allows fio to manage how
419 big the ranges of random data and zeroed data is. Without this set, fio will
420 provide \fBbuffer_compress_percentage\fR of blocksize random data, followed by
421 the remaining zeroed. With this set to some chunk size smaller than the block
422 size, fio can alternate random and zeroed data throughout the IO buffer.
424 .BI nrfiles \fR=\fPint
425 Number of files to use for this job. Default: 1.
427 .BI openfiles \fR=\fPint
428 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
430 .BI file_service_type \fR=\fPstr
431 Defines how files to service are selected. The following types are defined:
436 Choose a file at random.
439 Round robin over open files (default).
442 Do each file in the set sequentially.
445 The number of I/Os to issue before switching a new file can be specified by
446 appending `:\fIint\fR' to the service type.
449 .BI ioengine \fR=\fPstr
450 Defines how the job issues I/O. The following types are defined:
455 Basic \fBread\fR\|(2) or \fBwrite\fR\|(2) I/O. \fBfseek\fR\|(2) is used to
456 position the I/O location.
459 Basic \fBpread\fR\|(2) or \fBpwrite\fR\|(2) I/O.
462 Basic \fBreadv\fR\|(2) or \fBwritev\fR\|(2) I/O. Will emulate queuing by
463 coalescing adjacent IOs into a single submission.
466 Basic \fBpreadv\fR\|(2) or \fBpwritev\fR\|(2) I/O.
469 Linux native asynchronous I/O. This ioengine defines engine specific options.
472 POSIX asynchronous I/O using \fBaio_read\fR\|(3) and \fBaio_write\fR\|(3).
475 Solaris native asynchronous I/O.
478 Windows native asynchronous I/O.
481 File is memory mapped with \fBmmap\fR\|(2) and data copied using
485 \fBsplice\fR\|(2) is used to transfer the data and \fBvmsplice\fR\|(2) to
486 transfer data from user-space to the kernel.
489 Use the syslet system calls to make regular read/write asynchronous.
492 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
493 the target is an sg character device, we use \fBread\fR\|(2) and
494 \fBwrite\fR\|(2) for asynchronous I/O.
497 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
498 itself and for debugging and testing purposes.
501 Transfer over the network. The protocol to be used can be defined with the
502 \fBprotocol\fR parameter. Depending on the protocol, \fBfilename\fR,
503 \fBhostname\fR, \fBport\fR, or \fBlisten\fR must be specified.
504 This ioengine defines engine specific options.
507 Like \fBnet\fR, but uses \fBsplice\fR\|(2) and \fBvmsplice\fR\|(2) to map data
508 and send/receive. This ioengine defines engine specific options.
511 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
512 \fBcpucycles\fR parameters.
515 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
516 approach to asynchronous I/O.
518 See <http://www.xmailserver.org/guasi\-lib.html>.
521 The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ)
522 and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols.
525 Loads an external I/O engine object file. Append the engine filename as
529 IO engine that does regular linux native fallocate call to simulate data
530 transfer as fio ioengine
532 DDIR_READ does fallocate(,mode = FALLOC_FL_KEEP_SIZE,)
534 DIR_WRITE does fallocate(,mode = 0)
536 DDIR_TRIM does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE)
539 IO engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate defragment activity
540 request to DDIR_WRITE event
545 .BI iodepth \fR=\fPint
546 Number of I/O units to keep in flight against the file. Note that increasing
547 iodepth beyond 1 will not affect synchronous ioengines (except for small
548 degress when verify_async is in use). Even async engines my impose OS
549 restrictions causing the desired depth not to be achieved. This may happen on
550 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
551 not async on that OS. Keep an eye on the IO depth distribution in the
552 fio output to verify that the achieved depth is as expected. Default: 1.
554 .BI iodepth_batch \fR=\fPint
555 Number of I/Os to submit at once. Default: \fBiodepth\fR.
557 .BI iodepth_batch_complete \fR=\fPint
558 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
559 means that we'll ask for a minimum of 1 IO in the retrieval process from the
560 kernel. The IO retrieval will go on until we hit the limit set by
561 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
562 completed events before queuing more IO. This helps reduce IO latency, at the
563 cost of more retrieval system calls.
565 .BI iodepth_low \fR=\fPint
566 Low watermark indicating when to start filling the queue again. Default:
569 .BI direct \fR=\fPbool
570 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
572 .BI atomic \fR=\fPbool
573 If value is true, attempt to use atomic direct IO. Atomic writes are guaranteed
574 to be stable once acknowledged by the operating system. Only Linux supports
577 .BI buffered \fR=\fPbool
578 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
581 .BI offset \fR=\fPint
582 Offset in the file to start I/O. Data before the offset will not be touched.
584 .BI offset_increment \fR=\fPint
585 If this is provided, then the real offset becomes the
586 offset + offset_increment * thread_number, where the thread number is a counter
587 that starts at 0 and is incremented for each job. This option is useful if
588 there are several jobs which are intended to operate on a file in parallel in
589 disjoint segments, with even spacing between the starting points.
591 .BI number_ios \fR=\fPint
592 Fio will normally perform IOs until it has exhausted the size of the region
593 set by \fBsize\fR, or if it exhaust the allocated time (or hits an error
594 condition). With this setting, the range/size can be set independently of
595 the number of IOs to perform. When fio reaches this number, it will exit
596 normally and report status.
599 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
600 0, don't sync. Default: 0.
602 .BI fdatasync \fR=\fPint
603 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
604 data parts of the file. Default: 0.
606 .BI write_barrier \fR=\fPint
607 Make every Nth write a barrier write.
609 .BI sync_file_range \fR=\fPstr:int
610 Use \fBsync_file_range\fR\|(2) for every \fRval\fP number of write operations. Fio will
611 track range of writes that have happened since the last \fBsync_file_range\fR\|(2) call.
612 \fRstr\fP can currently be one or more of:
616 SYNC_FILE_RANGE_WAIT_BEFORE
619 SYNC_FILE_RANGE_WRITE
622 SYNC_FILE_RANGE_WRITE
626 So if you do sync_file_range=wait_before,write:8, fio would use
627 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
628 Also see the \fBsync_file_range\fR\|(2) man page. This option is Linux specific.
630 .BI overwrite \fR=\fPbool
631 If writing, setup the file first and do overwrites. Default: false.
633 .BI end_fsync \fR=\fPbool
634 Sync file contents when a write stage has completed. Default: false.
636 .BI fsync_on_close \fR=\fPbool
637 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
638 it will happen on every close, not just at the end of the job. Default: false.
640 .BI rwmixread \fR=\fPint
641 Percentage of a mixed workload that should be reads. Default: 50.
643 .BI rwmixwrite \fR=\fPint
644 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
645 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
646 overrides the first. This may interfere with a given rate setting, if fio is
647 asked to limit reads or writes to a certain rate. If that is the case, then
648 the distribution may be skewed. Default: 50.
650 .BI random_distribution \fR=\fPstr:float
651 By default, fio will use a completely uniform random distribution when asked
652 to perform random IO. Sometimes it is useful to skew the distribution in
653 specific ways, ensuring that some parts of the data is more hot than others.
654 Fio includes the following distribution models:
658 Uniform random distribution
668 When using a zipf or pareto distribution, an input value is also needed to
669 define the access pattern. For zipf, this is the zipf theta. For pareto,
670 it's the pareto power. Fio includes a test program, genzipf, that can be
671 used visualize what the given input values will yield in terms of hit rates.
672 If you wanted to use zipf with a theta of 1.2, you would use
673 random_distribution=zipf:1.2 as the option. If a non-uniform model is used,
674 fio will disable use of the random map.
676 .BI percentage_random \fR=\fPint
677 For a random workload, set how big a percentage should be random. This defaults
678 to 100%, in which case the workload is fully random. It can be set from
679 anywhere from 0 to 100. Setting it to 0 would make the workload fully
680 sequential. It is possible to set different values for reads, writes, and
681 trim. To do so, simply use a comma separated list. See \fBblocksize\fR.
684 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
685 this parameter is given, a new offset will be chosen without looking at past
686 I/O history. This parameter is mutually exclusive with \fBverify\fR.
688 .BI softrandommap \fR=\fPbool
689 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
690 fails to allocate the map, if this option is set it will continue without a
691 random block map. As coverage will not be as complete as with random maps, this
692 option is disabled by default.
694 .BI random_generator \fR=\fPstr
695 Fio supports the following engines for generating IO offsets for random IO:
699 Strong 2^88 cycle random number generator
702 Linear feedback shift register generator
706 Tausworthe is a strong random number generator, but it requires tracking on the
707 side if we want to ensure that blocks are only read or written once. LFSR
708 guarantees that we never generate the same offset twice, and it's also less
709 computationally expensive. It's not a true random generator, however, though
710 for IO purposes it's typically good enough. LFSR only works with single block
711 sizes, not with workloads that use multiple block sizes. If used with such a
712 workload, fio may read or write some blocks multiple times.
715 Run job with given nice value. See \fBnice\fR\|(2).
718 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
721 .BI prioclass \fR=\fPint
722 Set I/O priority class. See \fBionice\fR\|(1).
724 .BI thinktime \fR=\fPint
725 Stall job for given number of microseconds between issuing I/Os.
727 .BI thinktime_spin \fR=\fPint
728 Pretend to spend CPU time for given number of microseconds, sleeping the rest
729 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
731 .BI thinktime_blocks \fR=\fPint
732 Only valid if thinktime is set - control how many blocks to issue, before
733 waiting \fBthinktime\fR microseconds. If not set, defaults to 1 which will
734 make fio wait \fBthinktime\fR microseconds after every block. This
735 effectively makes any queue depth setting redundant, since no more than 1 IO
736 will be queued before we have to complete it and do our thinktime. In other
737 words, this setting effectively caps the queue depth if the latter is larger.
741 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
742 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
743 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
744 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
745 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
746 limit writes (to 500KB/sec), the latter will only limit reads.
748 .BI ratemin \fR=\fPint
749 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
750 Failing to meet this requirement will cause the job to exit. The same format
751 as \fBrate\fR is used for read vs write separation.
753 .BI rate_iops \fR=\fPint
754 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
755 specified independently of bandwidth. The same format as \fBrate\fR is used for
756 read vs write separation. If \fBblocksize\fR is a range, the smallest block
757 size is used as the metric.
759 .BI rate_iops_min \fR=\fPint
760 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
761 is used for read vs write separation.
763 .BI ratecycle \fR=\fPint
764 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
765 milliseconds. Default: 1000ms.
767 .BI max_latency \fR=\fPint
768 If set, fio will exit the job if it exceeds this maximum latency. It will exit
771 .BI cpumask \fR=\fPint
772 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
773 may run on. See \fBsched_setaffinity\fR\|(2).
775 .BI cpus_allowed \fR=\fPstr
776 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
778 .BI numa_cpu_nodes \fR=\fPstr
779 Set this job running on specified NUMA nodes' CPUs. The arguments allow
780 comma delimited list of cpu numbers, A-B ranges, or 'all'.
782 .BI numa_mem_policy \fR=\fPstr
783 Set this job's memory policy and corresponding NUMA nodes. Format of
787 .B <mode>[:<nodelist>]
790 is one of the following memory policy:
792 .B default, prefer, bind, interleave, local
795 For \fBdefault\fR and \fBlocal\fR memory policy, no \fBnodelist\fR is
796 needed to be specified. For \fBprefer\fR, only one node is
797 allowed. For \fBbind\fR and \fBinterleave\fR, \fBnodelist\fR allows
798 comma delimited list of numbers, A-B ranges, or 'all'.
800 .BI startdelay \fR=\fPint
801 Delay start of job for the specified number of seconds.
803 .BI runtime \fR=\fPint
804 Terminate processing after the specified number of seconds.
807 If given, run for the specified \fBruntime\fR duration even if the files are
808 completely read or written. The same workload will be repeated as many times
809 as \fBruntime\fR allows.
811 .BI ramp_time \fR=\fPint
812 If set, fio will run the specified workload for this amount of time before
813 logging any performance numbers. Useful for letting performance settle before
814 logging results, thus minimizing the runtime required for stable results. Note
815 that the \fBramp_time\fR is considered lead in time for a job, thus it will
816 increase the total runtime if a special timeout or runtime is specified.
818 .BI invalidate \fR=\fPbool
819 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
822 Use synchronous I/O for buffered writes. For the majority of I/O engines,
823 this means using O_SYNC. Default: false.
825 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
826 Allocation method for I/O unit buffer. Allowed values are:
831 Allocate memory with \fBmalloc\fR\|(3).
834 Use shared memory buffers allocated through \fBshmget\fR\|(2).
837 Same as \fBshm\fR, but use huge pages as backing.
840 Use \fBmmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
841 is given after the option in the format `:\fIfile\fR'.
844 Same as \fBmmap\fR, but use huge files as backing.
847 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
848 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
849 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
850 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
851 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
852 and the documentation for that. Normally you just need to echo an appropriate
853 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
857 .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint
858 This indicates the memory alignment of the IO memory buffers. Note that the
859 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
860 the alignment of the following buffers are given by the \fBbs\fR used. In
861 other words, if using a \fBbs\fR that is a multiple of the page sized in the
862 system, all buffers will be aligned to this value. If using a \fBbs\fR that
863 is not page aligned, the alignment of subsequent IO memory buffers is the
864 sum of the \fBiomem_align\fR and \fBbs\fR used.
866 .BI hugepage\-size \fR=\fPint
867 Defines the size of a huge page. Must be at least equal to the system setting.
868 Should be a multiple of 1MB. Default: 4MB.
871 Terminate all jobs when one finishes. Default: wait for each job to finish.
873 .BI bwavgtime \fR=\fPint
874 Average bandwidth calculations over the given time in milliseconds. Default:
877 .BI iopsavgtime \fR=\fPint
878 Average IOPS calculations over the given time in milliseconds. Default:
881 .BI create_serialize \fR=\fPbool
882 If true, serialize file creation for the jobs. Default: true.
884 .BI create_fsync \fR=\fPbool
885 \fBfsync\fR\|(2) data file after creation. Default: true.
887 .BI create_on_open \fR=\fPbool
888 If true, the files are not created until they are opened for IO by the job.
890 .BI create_only \fR=\fPbool
891 If true, fio will only run the setup phase of the job. If files need to be
892 laid out or updated on disk, only that will be done. The actual job contents
895 .BI pre_read \fR=\fPbool
896 If this is given, files will be pre-read into memory before starting the given
897 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
898 pointless to pre-read and then drop the cache. This will only work for IO
899 engines that are seekable, since they allow you to read the same data
900 multiple times. Thus it will not work on eg network or splice IO.
902 .BI unlink \fR=\fPbool
903 Unlink job files when done. Default: false.
906 Specifies the number of iterations (runs of the same workload) of this job.
909 .BI do_verify \fR=\fPbool
910 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
913 .BI verify \fR=\fPstr
914 Method of verifying file contents after each iteration of the job. Allowed
919 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1
920 Store appropriate checksum in the header of each block. crc32c-intel is
921 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
922 not supported by the system.
925 Write extra information about each I/O (timestamp, block number, etc.). The
926 block number is verified. See \fBverify_pattern\fR as well.
929 Pretend to verify. Used for testing internals.
932 This option can be used for repeated burn-in tests of a system to make sure
933 that the written data is also correctly read back. If the data direction given
934 is a read or random read, fio will assume that it should verify a previously
935 written file. If the data direction includes any form of write, the verify will
936 be of the newly written data.
939 .BI verifysort \fR=\fPbool
940 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
941 read them back in a sorted manner. Default: true.
943 .BI verifysort_nr \fR=\fPint
944 Pre-load and sort verify blocks for a read workload.
946 .BI verify_offset \fR=\fPint
947 Swap the verification header with data somewhere else in the block before
948 writing. It is swapped back before verifying.
950 .BI verify_interval \fR=\fPint
951 Write the verification header for this number of bytes, which should divide
952 \fBblocksize\fR. Default: \fBblocksize\fR.
954 .BI verify_pattern \fR=\fPstr
955 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
956 with totally random bytes, but sometimes it's interesting to fill with a known
957 pattern for io verification purposes. Depending on the width of the pattern,
958 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
959 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
960 has to be a hex number that starts with either "0x" or "0X". Use with
963 .BI verify_fatal \fR=\fPbool
964 If true, exit the job on the first observed verification failure. Default:
967 .BI verify_dump \fR=\fPbool
968 If set, dump the contents of both the original data block and the data block we
969 read off disk to files. This allows later analysis to inspect just what kind of
970 data corruption occurred. Off by default.
972 .BI verify_async \fR=\fPint
973 Fio will normally verify IO inline from the submitting thread. This option
974 takes an integer describing how many async offload threads to create for IO
975 verification instead, causing fio to offload the duty of verifying IO contents
976 to one or more separate threads. If using this offload option, even sync IO
977 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
978 allows them to have IO in flight while verifies are running.
980 .BI verify_async_cpus \fR=\fPstr
981 Tell fio to set the given CPU affinity on the async IO verification threads.
982 See \fBcpus_allowed\fP for the format used.
984 .BI verify_backlog \fR=\fPint
985 Fio will normally verify the written contents of a job that utilizes verify
986 once that job has completed. In other words, everything is written then
987 everything is read back and verified. You may want to verify continually
988 instead for a variety of reasons. Fio stores the meta data associated with an
989 IO block in memory, so for large verify workloads, quite a bit of memory would
990 be used up holding this meta data. If this option is enabled, fio will write
991 only N blocks before verifying these blocks.
993 .BI verify_backlog_batch \fR=\fPint
994 Control how many blocks fio will verify if verify_backlog is set. If not set,
995 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
996 read back and verified). If \fBverify_backlog_batch\fR is less than
997 \fBverify_backlog\fR then not all blocks will be verified, if
998 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
999 will be verified more than once.
1001 .BI trim_percentage \fR=\fPint
1002 Number of verify blocks to discard/trim.
1004 .BI trim_verify_zero \fR=\fPbool
1005 Verify that trim/discarded blocks are returned as zeroes.
1007 .BI trim_backlog \fR=\fPint
1008 Trim after this number of blocks are written.
1010 .BI trim_backlog_batch \fR=\fPint
1011 Trim this number of IO blocks.
1013 .BI experimental_verify \fR=\fPbool
1014 Enable experimental verification.
1016 .B stonewall "\fR,\fP wait_for_previous"
1017 Wait for preceding jobs in the job file to exit before starting this one.
1018 \fBstonewall\fR implies \fBnew_group\fR.
1021 Start a new reporting group. If not given, all jobs in a file will be part
1022 of the same reporting group, unless separated by a stonewall.
1024 .BI numjobs \fR=\fPint
1025 Number of clones (processes/threads performing the same workload) of this job.
1029 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
1033 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
1034 with \fBfork\fR\|(2).
1036 .BI zonesize \fR=\fPint
1037 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
1039 .BI zonerange \fR=\fPint
1040 Give size of an IO zone. See \fBzoneskip\fR.
1042 .BI zoneskip \fR=\fPint
1043 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
1046 .BI write_iolog \fR=\fPstr
1047 Write the issued I/O patterns to the specified file. Specify a separate file
1048 for each job, otherwise the iologs will be interspersed and the file may be
1051 .BI read_iolog \fR=\fPstr
1052 Replay the I/O patterns contained in the specified file generated by
1053 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
1055 .BI replay_no_stall \fR=\fPint
1056 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1057 attempts to respect timing information between I/Os. Enabling
1058 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
1059 still respecting ordering.
1061 .BI replay_redirect \fR=\fPstr
1062 While replaying I/O patterns using \fBread_iolog\fR the default behavior
1063 is to replay the IOPS onto the major/minor device that each IOP was recorded
1064 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
1065 single specified device regardless of the device it was recorded from.
1067 .BI write_bw_log \fR=\fPstr
1068 If given, write a bandwidth log of the jobs in this job file. Can be used to
1069 store data of the bandwidth of the jobs in their lifetime. The included
1070 fio_generate_plots script uses gnuplot to turn these text files into nice
1071 graphs. See \fBwrite_lat_log\fR for behaviour of given filename. For this
1072 option, the postfix is _bw.log.
1074 .BI write_lat_log \fR=\fPstr
1075 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
1076 filename is given with this option, the default filename of "jobname_type.log"
1077 is used. Even if the filename is given, fio will still append the type of log.
1079 .BI write_iops_log \fR=\fPstr
1080 Same as \fBwrite_bw_log\fR, but writes IOPS. If no filename is given with this
1081 option, the default filename of "jobname_type.log" is used. Even if the
1082 filename is given, fio will still append the type of log.
1084 .BI log_avg_msec \fR=\fPint
1085 By default, fio will log an entry in the iops, latency, or bw log for every
1086 IO that completes. When writing to the disk log, that can quickly grow to a
1087 very large size. Setting this option makes fio average the each log entry
1088 over the specified period of time, reducing the resolution of the log.
1091 .BI disable_lat \fR=\fPbool
1092 Disable measurements of total latency numbers. Useful only for cutting
1093 back the number of calls to \fBgettimeofday\fR\|(2), as that does impact performance at
1094 really high IOPS rates. Note that to really get rid of a large amount of these
1095 calls, this option must be used with disable_slat and disable_bw as well.
1097 .BI disable_clat \fR=\fPbool
1098 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
1100 .BI disable_slat \fR=\fPbool
1101 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
1103 .BI disable_bw_measurement \fR=\fPbool
1104 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
1106 .BI lockmem \fR=\fPint
1107 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
1108 simulate a smaller amount of memory. The amount specified is per worker.
1110 .BI exec_prerun \fR=\fPstr
1111 Before running the job, execute the specified command with \fBsystem\fR\|(3).
1113 Output is redirected in a file called \fBjobname.prerun.txt\fR
1116 .BI exec_postrun \fR=\fPstr
1117 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
1119 Output is redirected in a file called \fBjobname.postrun.txt\fR
1122 .BI ioscheduler \fR=\fPstr
1123 Attempt to switch the device hosting the file to the specified I/O scheduler.
1125 .BI cpuload \fR=\fPint
1126 If the job is a CPU cycle-eater, attempt to use the specified percentage of
1129 .BI cpuchunks \fR=\fPint
1130 If the job is a CPU cycle-eater, split the load into cycles of the
1131 given time in milliseconds.
1133 .BI disk_util \fR=\fPbool
1134 Generate disk utilization statistics if the platform supports it. Default: true.
1136 .BI clocksource \fR=\fPstr
1137 Use the given clocksource as the base of timing. The supported options are:
1141 \fBgettimeofday\fR\|(2)
1144 \fBclock_gettime\fR\|(2)
1147 Internal CPU clock source
1151 \fBcpu\fR is the preferred clocksource if it is reliable, as it is very fast
1152 (and fio is heavy on time calls). Fio will automatically use this clocksource
1153 if it's supported and considered reliable on the system it is running on,
1154 unless another clocksource is specifically set. For x86/x86-64 CPUs, this
1155 means supporting TSC Invariant.
1157 .BI gtod_reduce \fR=\fPbool
1158 Enable all of the \fBgettimeofday\fR\|(2) reducing options (disable_clat, disable_slat,
1159 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
1160 \fBgettimeofday\fR\|(2) call count. With this option enabled, we only do about 0.4% of
1161 the gtod() calls we would have done if all time keeping was enabled.
1163 .BI gtod_cpu \fR=\fPint
1164 Sometimes it's cheaper to dedicate a single thread of execution to just getting
1165 the current time. Fio (and databases, for instance) are very intensive on
1166 \fBgettimeofday\fR\|(2) calls. With this option, you can set one CPU aside for doing
1167 nothing but logging current time to a shared memory location. Then the other
1168 threads/processes that run IO workloads need only copy that segment, instead of
1169 entering the kernel with a \fBgettimeofday\fR\|(2) call. The CPU set aside for doing
1170 these time calls will be excluded from other uses. Fio will manually clear it
1171 from the CPU mask of other jobs.
1173 .BI ignore_error \fR=\fPstr
1174 Sometimes you want to ignore some errors during test in that case you can specify
1175 error list for each error type.
1177 ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST
1179 errors for given error type is separated with ':'.
1180 Error may be symbol ('ENOSPC', 'ENOMEM') or an integer.
1182 Example: ignore_error=EAGAIN,ENOSPC:122 .
1184 This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from WRITE.
1186 .BI error_dump \fR=\fPbool
1187 If set dump every error even if it is non fatal, true by default. If disabled
1188 only fatal error will be dumped
1190 .BI profile \fR=\fPstr
1191 Select a specific builtin performance test.
1193 .BI cgroup \fR=\fPstr
1194 Add job to this control group. If it doesn't exist, it will be created.
1195 The system must have a mounted cgroup blkio mount point for this to work. If
1196 your system doesn't have it mounted, you can do so with:
1198 # mount \-t cgroup \-o blkio none /cgroup
1200 .BI cgroup_weight \fR=\fPint
1201 Set the weight of the cgroup to this value. See the documentation that comes
1202 with the kernel, allowed values are in the range of 100..1000.
1204 .BI cgroup_nodelete \fR=\fPbool
1205 Normally fio will delete the cgroups it has created after the job completion.
1206 To override this behavior and to leave cgroups around after the job completion,
1207 set cgroup_nodelete=1. This can be useful if one wants to inspect various
1208 cgroup files after job completion. Default: false
1211 Instead of running as the invoking user, set the user ID to this value before
1212 the thread/process does any work.
1215 Set group ID, see \fBuid\fR.
1217 .BI unit_base \fR=\fPint
1218 Base unit for reporting. Allowed values are:
1222 Use auto-detection (default).
1232 .BI flow_id \fR=\fPint
1233 The ID of the flow. If not specified, it defaults to being a global flow. See
1237 Weight in token-based flow control. If this value is used, then there is a
1238 \fBflow counter\fR which is used to regulate the proportion of activity between
1239 two or more jobs. fio attempts to keep this flow counter near zero. The
1240 \fBflow\fR parameter stands for how much should be added or subtracted to the
1241 flow counter on each iteration of the main I/O loop. That is, if one job has
1242 \fBflow=8\fR and another job has \fBflow=-1\fR, then there will be a roughly
1243 1:8 ratio in how much one runs vs the other.
1245 .BI flow_watermark \fR=\fPint
1246 The maximum value that the absolute value of the flow counter is allowed to
1247 reach before the job must wait for a lower value of the counter.
1249 .BI flow_sleep \fR=\fPint
1250 The period of time, in microseconds, to wait after the flow watermark has been
1251 exceeded before retrying operations
1253 .BI clat_percentiles \fR=\fPbool
1254 Enable the reporting of percentiles of completion latencies.
1256 .BI percentile_list \fR=\fPfloat_list
1257 Overwrite the default list of percentiles for completion
1258 latencies. Each number is a floating number in the range (0,100], and
1259 the maximum length of the list is 20. Use ':' to separate the
1260 numbers. For example, \-\-percentile_list=99.5:99.9 will cause fio to
1261 report the values of completion latency below which 99.5% and 99.9% of
1262 the observed latencies fell, respectively.
1263 .SS "Ioengine Parameters List"
1264 Some parameters are only valid when a specific ioengine is in use. These are
1265 used identically to normal parameters, with the caveat that when used on the
1266 command line, the must come after the ioengine that defines them is selected.
1268 .BI (cpu)cpuload \fR=\fPint
1269 Attempt to use the specified percentage of CPU cycles.
1271 .BI (cpu)cpuchunks \fR=\fPint
1272 Split the load into cycles of the given time. In microseconds.
1274 .BI (libaio)userspace_reap
1275 Normally, with the libaio engine in use, fio will use
1276 the io_getevents system call to reap newly returned events.
1277 With this flag turned on, the AIO ring will be read directly
1278 from user-space to reap events. The reaping mode is only
1279 enabled when polling for a minimum of 0 events (eg when
1280 iodepth_batch_complete=0).
1282 .BI (net,netsplice)hostname \fR=\fPstr
1283 The host name or IP address to use for TCP or UDP based IO.
1284 If the job is a TCP listener or UDP reader, the hostname is not
1285 used and must be omitted unless it is a valid UDP multicast address.
1287 .BI (net,netsplice)port \fR=\fPint
1288 The TCP or UDP port to bind to or connect to.
1290 .BI (net,netsplice)interface \fR=\fPstr
1291 The IP address of the network interface used to send or receive UDP multicast
1294 .BI (net,netsplice)ttl \fR=\fPint
1295 Time-to-live value for outgoing UDP multicast packets. Default: 1
1297 .BI (net,netsplice)nodelay \fR=\fPbool
1298 Set TCP_NODELAY on TCP connections.
1300 .BI (net,netsplice)protocol \fR=\fPstr "\fR,\fP proto" \fR=\fPstr
1301 The network protocol to use. Accepted values are:
1306 Transmission control protocol
1309 User datagram protocol
1315 When the protocol is TCP or UDP, the port must also be given,
1316 as well as the hostname if the job is a TCP listener or UDP
1317 reader. For unix sockets, the normal filename option should be
1318 used and the port is invalid.
1321 .BI (net,netsplice)listen
1322 For TCP network connections, tell fio to listen for incoming
1323 connections rather than initiating an outgoing connection. The
1324 hostname must be omitted if this option is used.
1326 .BI (net, pingpong) \fR=\fPbool
1327 Normally a network writer will just continue writing data, and a network reader
1328 will just consume packages. If pingpong=1 is set, a writer will send its normal
1329 payload to the reader, then wait for the reader to send the same payload back.
1330 This allows fio to measure network latencies. The submission and completion
1331 latencies then measure local time spent sending or receiving, and the
1332 completion latency measures how long it took for the other end to receive and
1333 send back. For UDP multicast traffic pingpong=1 should only be set for a single
1334 reader when multiple readers are listening to the same address.
1336 .BI (e4defrag,donorname) \fR=\fPstr
1337 File will be used as a block donor (swap extents between files)
1339 .BI (e4defrag,inplace) \fR=\fPint
1340 Configure donor file block allocation strategy
1343 Preallocate donor's file on init
1346 allocate space immediately inside defragment event, and free right after event
1349 While running, \fBfio\fR will display the status of the created jobs. For
1353 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
1356 The characters in the first set of brackets denote the current status of each
1357 threads. The possible values are:
1363 Setup but not started.
1369 Initialized, waiting.
1372 Running, doing sequential reads.
1375 Running, doing random reads.
1378 Running, doing sequential writes.
1381 Running, doing random writes.
1384 Running, doing mixed sequential reads/writes.
1387 Running, doing mixed random reads/writes.
1390 Running, currently waiting for \fBfsync\fR\|(2).
1393 Running, verifying written data.
1396 Exited, not reaped by main thread.
1399 Exited, thread reaped.
1403 The second set of brackets shows the estimated completion percentage of
1404 the current group. The third set shows the read and write I/O rate,
1405 respectively. Finally, the estimated run time of the job is displayed.
1407 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
1408 for each thread, each group of threads, and each disk, in that order.
1410 Per-thread statistics first show the threads client number, group-id, and
1411 error code. The remaining figures are as follows:
1415 Number of megabytes of I/O performed.
1418 Average data rate (bandwidth).
1424 Submission latency minimum, maximum, average and standard deviation. This is
1425 the time it took to submit the I/O.
1428 Completion latency minimum, maximum, average and standard deviation. This
1429 is the time between submission and completion.
1432 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
1433 and standard deviation.
1436 CPU usage statistics. Includes user and system time, number of context switches
1437 this thread went through and number of major and minor page faults.
1440 Distribution of I/O depths. Each depth includes everything less than (or equal)
1441 to it, but greater than the previous depth.
1444 Number of read/write requests issued, and number of short read/write requests.
1447 Distribution of I/O completion latencies. The numbers follow the same pattern
1451 The group statistics show:
1456 Number of megabytes I/O performed.
1459 Aggregate bandwidth of threads in the group.
1462 Minimum average bandwidth a thread saw.
1465 Maximum average bandwidth a thread saw.
1468 Shortest runtime of threads in the group.
1471 Longest runtime of threads in the group.
1475 Finally, disk statistics are printed with reads first:
1480 Number of I/Os performed by all groups.
1483 Number of merges in the I/O scheduler.
1486 Number of ticks we kept the disk busy.
1489 Total time spent in the disk queue.
1496 It is also possible to get fio to dump the current output while it is
1497 running, without terminating the job. To do that, send fio the \fBUSR1\fR
1500 If the \fB\-\-minimal\fR option is given, the results will be printed in a
1501 semicolon-delimited format suitable for scripted use - a job description
1502 (if provided) follows on a new line. Note that the first
1503 number in the line is the version number. If the output has to be changed
1504 for some reason, this number will be incremented by 1 to signify that
1505 change. The fields are:
1508 .B terse version, fio version, jobname, groupid, error
1512 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1516 .B min, max, mean, standard deviation
1520 .B min, max, mean, standard deviation
1522 Completion latency percentiles (20 fields):
1524 .B Xth percentile=usec
1528 .B min, max, mean, standard deviation
1532 .B min, max, aggregate percentage of total, mean, standard deviation
1538 .B Total I/O \fR(KB)\fP, bandwidth \fR(KB/s)\fP, IOPS, runtime \fR(ms)\fP
1542 .B min, max, mean, standard deviation
1546 .B min, max, mean, standard deviation
1548 Completion latency percentiles (20 fields):
1550 .B Xth percentile=usec
1554 .B min, max, mean, standard deviation
1558 .B min, max, aggregate percentage of total, mean, standard deviation
1564 .B user, system, context switches, major page faults, minor page faults
1567 IO depth distribution:
1569 .B <=1, 2, 4, 8, 16, 32, >=64
1572 IO latency distribution:
1576 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1580 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1584 Disk utilization (1 for each disk used):
1586 .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
1589 Error Info (dependent on continue_on_error, default off):
1591 .B total # errors, first error code
1594 .B text description (if provided in config - appears on newline)
1597 Normally you would run fio as a stand-alone application on the machine
1598 where the IO workload should be generated. However, it is also possible to
1599 run the frontend and backend of fio separately. This makes it possible to
1600 have a fio server running on the machine(s) where the IO workload should
1601 be running, while controlling it from another machine.
1603 To start the server, you would do:
1605 \fBfio \-\-server=args\fR
1607 on that machine, where args defines what fio listens to. The arguments
1608 are of the form 'type:hostname or IP:port'. 'type' is either 'ip' (or ip4)
1609 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain
1610 socket. 'hostname' is either a hostname or IP address, and 'port' is the port to
1611 listen to (only valid for TCP/IP, not a local socket). Some examples:
1615 Start a fio server, listening on all interfaces on the default port (8765).
1617 2) fio \-\-server=ip:hostname,4444
1619 Start a fio server, listening on IP belonging to hostname and on port 4444.
1621 3) fio \-\-server=ip6:::1,4444
1623 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
1625 4) fio \-\-server=,4444
1627 Start a fio server, listening on all interfaces on port 4444.
1629 5) fio \-\-server=1.2.3.4
1631 Start a fio server, listening on IP 1.2.3.4 on the default port.
1633 6) fio \-\-server=sock:/tmp/fio.sock
1635 Start a fio server, listening on the local socket /tmp/fio.sock.
1637 When a server is running, you can connect to it from a client. The client
1640 fio \-\-local-args \-\-client=server \-\-remote-args <job file(s)>
1642 where \-\-local-args are arguments that are local to the client where it is
1643 running, 'server' is the connect string, and \-\-remote-args and <job file(s)>
1644 are sent to the server. The 'server' string follows the same format as it
1645 does on the server side, to allow IP/hostname/socket and port strings.
1646 You can connect to multiple clients as well, to do that you could run:
1648 fio \-\-client=server2 \-\-client=server2 <job file(s)>
1652 was written by Jens Axboe <jens.axboe@oracle.com>,
1653 now Jens Axboe <jaxboe@fusionio.com>.
1655 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1656 on documentation by Jens Axboe.
1657 .SH "REPORTING BUGS"
1658 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1661 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1663 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob