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 \-\-output \fR=\fPfilename
16 Write output to \fIfilename\fR.
18 .BI \-\-timeout \fR=\fPtimeout
19 Limit run time to \fItimeout\fR seconds.
22 Generate per-job latency logs.
25 Generate per-job bandwidth logs.
28 Print statistics in a terse, semicolon-delimited format.
30 .BI \-\-showcmd \fR=\fPjobfile
31 Convert \fIjobfile\fR to a set of command-line options.
34 Enable read-only safety checks.
36 .BI \-\-eta \fR=\fPwhen
37 Specifies when real-time ETA estimate should be printed. \fIwhen\fR may
38 be one of `always', `never' or `auto'.
40 .BI \-\-section \fR=\fPsec
41 Only run section \fIsec\fR from job file.
43 .BI \-\-cmdhelp \fR=\fPcommand
44 Print help information for \fIcommand\fR. May be `all' for all commands.
46 .BI \-\-debug \fR=\fPtype
47 Enable verbose tracing of various fio actions. May be `all' for all types
48 or individual types separated by a comma (eg \-\-debug=io,file). `help' will
49 list all available tracing options.
52 Display usage information and exit.
55 Display version information and exit.
57 Job files are in `ini' format. They consist of one or more
58 job definitions, which begin with a job name in square brackets and
59 extend to the next job name. The job name can be any ASCII string
60 except `global', which has a special meaning. Following the job name is
61 a sequence of zero or more parameters, one per line, that define the
62 behavior of the job. Any line starting with a `;' or `#' character is
63 considered a comment and ignored.
65 If \fIjobfile\fR is specified as `-', the job file will be read from
68 The global section contains default parameters for jobs specified in the
69 job file. A job is only affected by global sections residing above it,
70 and there may be any number of global sections. Specific job definitions
71 may override any parameter set in global sections.
74 Some parameters may take arguments of a specific type. The types used are:
77 String: a sequence of alphanumeric characters.
80 SI integer: a whole number, possibly containing a suffix denoting the base unit
81 of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting
82 kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5)
83 respectively. The suffix is not case sensitive. If prefixed with '0x', the
84 value is assumed to be base 16 (hexadecimal). A suffix may include a trailing 'b',
85 for instance 'kb' is identical to 'k'. You can specify a base 10 value
86 by using 'KiB', 'MiB', 'GiB', etc. This is useful for disk drives where
87 values are often given in base 10 values. Specifying '30GiB' will get you
91 Boolean: a true or false value. `0' denotes false, `1' denotes true.
94 Integer range: a range of integers specified in the format
95 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
96 \fIupper\fR may contain a suffix as described above. If an option allows two
97 sets of ranges, they are separated with a `,' or `/' character. For example:
102 May be used to override the job name. On the command line, this parameter
103 has the special purpose of signalling the start of a new job.
105 .BI description \fR=\fPstr
106 Human-readable description of the job. It is printed when the job is run, but
107 otherwise has no special purpose.
109 .BI directory \fR=\fPstr
110 Prefix filenames with this directory. Used to place files in a location other
113 .BI filename \fR=\fPstr
115 normally makes up a file name based on the job name, thread number, and file
116 number. If you want to share files between threads in a job or several jobs,
117 specify a \fIfilename\fR for each of them to override the default. If the I/O
118 engine used is `net', \fIfilename\fR is the host and port to connect to in the
119 format \fIhost\fR/\fIport\fR. If the I/O engine is file-based, you can specify
120 a number of files by separating the names with a `:' character. `\-' is a
121 reserved name, meaning stdin or stdout, depending on the read/write direction
124 .BI lockfile \fR=\fPstr
125 Fio defaults to not locking any files before it does IO to them. If a file or
126 file descriptor is shared, fio can serialize IO to that file to make the end
127 result consistent. This is usual for emulating real workloads that share files.
133 No locking. This is the default.
136 Only one thread or process may do IO at the time, excluding all others.
139 Read-write locking on the file. Many readers may access the file at the same
140 time, but writes get exclusive access.
143 The option may be post-fixed with a lock batch number. If set, then each
144 thread/process may do that amount of IOs to the file before giving up the lock.
145 Since lock acquisition is expensive, batching the lock/unlocks will speed up IO.
148 .BI opendir \fR=\fPstr
149 Recursively open any files below directory \fIstr\fR.
151 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
152 Type of I/O pattern. Accepted values are:
169 Mixed sequential reads and writes.
172 Mixed random reads and writes.
175 For mixed I/O, the default split is 50/50. For certain types of io the result
176 may still be skewed a bit, since the speed may be different. It is possible to
177 specify a number of IO's to do before getting a new offset, this is one by
178 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
179 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
180 value of 8. See the \fBrw_sequencer\fR option.
183 .BI rw_sequencer \fR=\fPstr
184 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
185 then this option controls how that number modifies the IO offset being
186 generated. Accepted values are:
191 Generate sequential offset
194 Generate the same offset
197 \fBsequential\fR is only useful for random IO, where fio would normally
198 generate a new random offset for every IO. If you append eg 8 to randread, you
199 would get a new random offset for every 8 IO's. The result would be a seek for
200 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
201 that. As sequential IO is already sequential, setting \fBsequential\fR for that
202 would not result in any differences. \fBidentical\fR behaves in a similar
203 fashion, except it sends the same offset 8 number of times before generating a
208 .BI kb_base \fR=\fPint
209 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
210 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
211 reasons. Allow values are 1024 or 1000, with 1024 being the default.
213 .BI randrepeat \fR=\fPbool
214 Seed the random number generator in a predictable way so results are repeatable
215 across runs. Default: true.
217 .BI use_os_rand \fR=\fPbool
218 Fio can either use the random generator supplied by the OS to generator random
219 offsets, or it can use it's own internal generator (based on Tausworthe).
220 Default is to use the internal generator, which is often of better quality and
221 faster. Default: false.
223 .BI fallocate \fR=\fPstr
224 Whether pre-allocation is performed when laying down files. Accepted values
230 Do not pre-allocate space.
233 Pre-allocate via posix_fallocate().
236 Pre-allocate via fallocate() with FALLOC_FL_KEEP_SIZE set.
239 Backward-compatible alias for 'none'.
242 Backward-compatible alias for 'posix'.
245 May not be available on all supported platforms. 'keep' is only
246 available on Linux. If using ZFS on Solaris this must be set to 'none'
247 because ZFS doesn't support it. Default: 'posix'.
250 .BI fadvise_hint \fR=\fPbool
251 Disable use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
252 are likely to be issued. Default: true.
255 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
256 been transfered, unless limited by other options (\fBruntime\fR, for instance).
257 Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be
258 divided between the available files for the job. If not set, fio will use the
259 full size of the given files or devices. If the the files do not exist, size
260 must be given. It is also possible to give size as a percentage between 1 and
261 100. If size=20% is given, fio will use 20% of the full size of the given files
264 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
265 Sets size to something really large and waits for ENOSPC (no space left on
266 device) as the terminating condition. Only makes sense with sequential write.
267 For a read workload, the mount point will be filled first then IO started on
268 the result. This option doesn't make sense if operating on a raw device node,
269 since the size of that is already known by the file system. Additionally,
270 writing beyond end-of-device will not return ENOSPC there.
272 .BI filesize \fR=\fPirange
273 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
274 for files at random within the given range, limited to \fBsize\fR in total (if
275 that is given). If \fBfilesize\fR is not specified, each created file is the
278 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
279 Block size for I/O units. Default: 4k. Values for reads and writes can be
280 specified separately in the format \fIread\fR,\fIwrite\fR, either of
281 which may be empty to leave that value at its default.
283 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
284 Specify a range of I/O block sizes. The issued I/O unit will always be a
285 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
286 to both reads and writes if only one range is given, but can be specified
287 separately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k.
288 Also (see \fBblocksize\fR).
290 .BI bssplit \fR=\fPstr
291 This option allows even finer grained control of the block sizes issued,
292 not just even splits between them. With this option, you can weight various
293 block sizes for exact control of the issued IO for a job that has mixed
294 block sizes. The format of the option is bssplit=blocksize/percentage,
295 optionally adding as many definitions as needed separated by a colon.
296 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
297 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
298 splits to reads and writes. The format is identical to what the
299 \fBbs\fR option accepts, the read and write parts are separated with a
302 .B blocksize_unaligned\fR,\fP bs_unaligned
303 If set, any size in \fBblocksize_range\fR may be used. This typically won't
304 work with direct I/O, as that normally requires sector alignment.
306 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
307 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
308 the minimum blocksize given. Minimum alignment is typically 512b
309 for using direct IO, though it usually depends on the hardware block size.
310 This option is mutually exclusive with using a random map for files, so it
311 will turn off that option.
314 Initialise buffers with all zeros. Default: fill buffers with random data.
317 If this option is given, fio will refill the IO buffers on every submit. The
318 default is to only fill it at init time and reuse that data. Only makes sense
319 if zero_buffers isn't specified, naturally. If data verification is enabled,
320 refill_buffers is also automatically enabled.
322 .BI nrfiles \fR=\fPint
323 Number of files to use for this job. Default: 1.
325 .BI openfiles \fR=\fPint
326 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
328 .BI file_service_type \fR=\fPstr
329 Defines how files to service are selected. The following types are defined:
334 Choose a file at random
337 Round robin over open files (default).
339 Do each file in the set sequentially.
342 The number of I/Os to issue before switching a new file can be specified by
343 appending `:\fIint\fR' to the service type.
346 .BI ioengine \fR=\fPstr
347 Defines how the job issues I/O. The following types are defined:
352 Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
353 position the I/O location.
356 Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O.
359 Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by
360 coalescing adjacents IOs into a single submission.
363 Linux native asynchronous I/O.
366 POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
369 Solaris native asynchronous I/O.
372 Windows native asynchronous I/O.
375 File is memory mapped with \fImmap\fR\|(2) and data copied using
379 \fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
380 transfer data from user-space to the kernel.
383 Use the syslet system calls to make regular read/write asynchronous.
386 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
387 the target is an sg character device, we use \fIread\fR\|(2) and
388 \fIwrite\fR\|(2) for asynchronous I/O.
391 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
392 itself and for debugging and testing purposes.
395 Transfer over the network. \fBfilename\fR must be set appropriately to
396 `\fIhost\fR/\fIport\fR' regardless of data direction. If receiving, only the
397 \fIport\fR argument is used.
400 Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
404 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
405 \fBcpucycles\fR parameters.
408 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
409 approach to asycnronous I/O.
411 See <http://www.xmailserver.org/guasi\-lib.html>.
414 The RDMA I/O engine supports both RDMA memory semantic(RDMA_WRITE/RDMA_READ)
415 and channel semantic(Send/Recv) in InfiniBand, RoCE and iWarp environment.
418 Loads an external I/O engine object file. Append the engine filename as
423 .BI iodepth \fR=\fPint
424 Number of I/O units to keep in flight against the file. Note that increasing
425 iodepth beyond 1 will not affect synchronous ioengines (except for small
426 degress when verify_async is in use). Even async engines my impose OS
427 restrictions causing the desired depth not to be achieved. This may happen on
428 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
429 not async on that OS. Keep an eye on the IO depth distribution in the
430 fio output to verify that the achieved depth is as expected. Default: 1.
432 .BI iodepth_batch \fR=\fPint
433 Number of I/Os to submit at once. Default: \fBiodepth\fR.
435 .BI iodepth_batch_complete \fR=\fPint
436 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
437 means that we'll ask for a minimum of 1 IO in the retrieval process from the
438 kernel. The IO retrieval will go on until we hit the limit set by
439 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
440 completed events before queuing more IO. This helps reduce IO latency, at the
441 cost of more retrieval system calls.
443 .BI iodepth_low \fR=\fPint
444 Low watermark indicating when to start filling the queue again. Default:
447 .BI direct \fR=\fPbool
448 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
450 .BI buffered \fR=\fPbool
451 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
454 .BI offset \fR=\fPint
455 Offset in the file to start I/O. Data before the offset will not be touched.
458 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
459 0, don't sync. Default: 0.
461 .BI fdatasync \fR=\fPint
462 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
463 data parts of the file. Default: 0.
465 .BI sync_file_range \fR=\fPstr:int
466 Use sync_file_range() for every \fRval\fP number of write operations. Fio will
467 track range of writes that have happened since the last sync_file_range() call.
468 \fRstr\fP can currently be one or more of:
472 SYNC_FILE_RANGE_WAIT_BEFORE
475 SYNC_FILE_RANGE_WRITE
478 SYNC_FILE_RANGE_WRITE
482 So if you do sync_file_range=wait_before,write:8, fio would use
483 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
484 Also see the sync_file_range(2) man page. This option is Linux specific.
486 .BI overwrite \fR=\fPbool
487 If writing, setup the file first and do overwrites. Default: false.
489 .BI end_fsync \fR=\fPbool
490 Sync file contents when job exits. Default: false.
492 .BI fsync_on_close \fR=\fPbool
493 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
494 it will happen on every close, not just at the end of the job. Default: false.
496 .BI rwmixcycle \fR=\fPint
497 How many milliseconds before switching between reads and writes for a mixed
498 workload. Default: 500ms.
500 .BI rwmixread \fR=\fPint
501 Percentage of a mixed workload that should be reads. Default: 50.
503 .BI rwmixwrite \fR=\fPint
504 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
505 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
506 overrides the first. This may interfere with a given rate setting, if fio is
507 asked to limit reads or writes to a certain rate. If that is the case, then
508 the distribution may be skewed. Default: 50.
511 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
512 this parameter is given, a new offset will be chosen without looking at past
513 I/O history. This parameter is mutually exclusive with \fBverify\fR.
516 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
517 fails to allocate the map, if this option is set it will continue without a
518 random block map. As coverage will not be as complete as with random maps, this
519 option is disabled by default.
522 Run job with given nice value. See \fInice\fR\|(2).
525 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
528 .BI prioclass \fR=\fPint
529 Set I/O priority class. See \fIionice\fR\|(1).
531 .BI thinktime \fR=\fPint
532 Stall job for given number of microseconds between issuing I/Os.
534 .BI thinktime_spin \fR=\fPint
535 Pretend to spend CPU time for given number of microseconds, sleeping the rest
536 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
538 .BI thinktime_blocks \fR=\fPint
539 Number of blocks to issue before waiting \fBthinktime\fR microseconds.
543 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
544 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
545 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
546 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
547 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
548 limit writes (to 500KB/sec), the latter will only limit reads.
550 .BI ratemin \fR=\fPint
551 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
552 Failing to meet this requirement will cause the job to exit. The same format
553 as \fBrate\fR is used for read vs write separation.
555 .BI rate_iops \fR=\fPint
556 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
557 specified independently of bandwidth. The same format as \fBrate\fR is used for
558 read vs write seperation. If \fBblocksize\fR is a range, the smallest block
559 size is used as the metric.
561 .BI rate_iops_min \fR=\fPint
562 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
563 is used for read vs write seperation.
565 .BI ratecycle \fR=\fPint
566 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
567 milliseconds. Default: 1000ms.
569 .BI cpumask \fR=\fPint
570 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
571 may run on. See \fBsched_setaffinity\fR\|(2).
573 .BI cpus_allowed \fR=\fPstr
574 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
576 .BI startdelay \fR=\fPint
577 Delay start of job for the specified number of seconds.
579 .BI runtime \fR=\fPint
580 Terminate processing after the specified number of seconds.
583 If given, run for the specified \fBruntime\fR duration even if the files are
584 completely read or written. The same workload will be repeated as many times
585 as \fBruntime\fR allows.
587 .BI ramp_time \fR=\fPint
588 If set, fio will run the specified workload for this amount of time before
589 logging any performance numbers. Useful for letting performance settle before
590 logging results, thus minimizing the runtime required for stable results. Note
591 that the \fBramp_time\fR is considered lead in time for a job, thus it will
592 increase the total runtime if a special timeout or runtime is specified.
594 .BI invalidate \fR=\fPbool
595 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
598 Use synchronous I/O for buffered writes. For the majority of I/O engines,
599 this means using O_SYNC. Default: false.
601 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
602 Allocation method for I/O unit buffer. Allowed values are:
607 Allocate memory with \fImalloc\fR\|(3).
610 Use shared memory buffers allocated through \fIshmget\fR\|(2).
613 Same as \fBshm\fR, but use huge pages as backing.
616 Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
617 is given after the option in the format `:\fIfile\fR'.
620 Same as \fBmmap\fR, but use huge files as backing.
623 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
624 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
625 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
626 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
627 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
628 and the documentation for that. Normally you just need to echo an appropriate
629 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
633 .BI iomem_align \fR=\fPint
634 This indiciates the memory alignment of the IO memory buffers. Note that the
635 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
636 the alignment of the following buffers are given by the \fBbs\fR used. In
637 other words, if using a \fBbs\fR that is a multiple of the page sized in the
638 system, all buffers will be aligned to this value. If using a \fBbs\fR that
639 is not page aligned, the alignment of subsequent IO memory buffers is the
640 sum of the \fBiomem_align\fR and \fBbs\fR used.
642 .BI hugepage\-size \fR=\fPint
643 Defines the size of a huge page. Must be at least equal to the system setting.
644 Should be a multiple of 1MB. Default: 4MB.
647 Terminate all jobs when one finishes. Default: wait for each job to finish.
649 .BI bwavgtime \fR=\fPint
650 Average bandwidth calculations over the given time in milliseconds. Default:
653 .BI create_serialize \fR=\fPbool
654 If true, serialize file creation for the jobs. Default: true.
656 .BI create_fsync \fR=\fPbool
657 \fIfsync\fR\|(2) data file after creation. Default: true.
659 .BI create_on_open \fR=\fPbool
660 If true, the files are not created until they are opened for IO by the job.
662 .BI pre_read \fR=\fPbool
663 If this is given, files will be pre-read into memory before starting the given
664 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
665 pointless to pre-read and then drop the cache. This will only work for IO
666 engines that are seekable, since they allow you to read the same data
667 multiple times. Thus it will not work on eg network or splice IO.
669 .BI unlink \fR=\fPbool
670 Unlink job files when done. Default: false.
673 Specifies the number of iterations (runs of the same workload) of this job.
676 .BI do_verify \fR=\fPbool
677 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
680 .BI verify \fR=\fPstr
681 Method of verifying file contents after each iteration of the job. Allowed
686 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1
687 Store appropriate checksum in the header of each block. crc32c-intel is
688 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
689 not supported by the system.
692 Write extra information about each I/O (timestamp, block number, etc.). The
693 block number is verified. See \fBverify_pattern\fR as well.
696 Pretend to verify. Used for testing internals.
699 This option can be used for repeated burn-in tests of a system to make sure
700 that the written data is also correctly read back. If the data direction given
701 is a read or random read, fio will assume that it should verify a previously
702 written file. If the data direction includes any form of write, the verify will
703 be of the newly written data.
706 .BI verify_sort \fR=\fPbool
707 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
708 read them back in a sorted manner. Default: true.
710 .BI verify_offset \fR=\fPint
711 Swap the verification header with data somewhere else in the block before
712 writing. It is swapped back before verifying.
714 .BI verify_interval \fR=\fPint
715 Write the verification header for this number of bytes, which should divide
716 \fBblocksize\fR. Default: \fBblocksize\fR.
718 .BI verify_pattern \fR=\fPstr
719 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
720 with totally random bytes, but sometimes it's interesting to fill with a known
721 pattern for io verification purposes. Depending on the width of the pattern,
722 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
723 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
724 has to be a hex number that starts with either "0x" or "0X". Use with
727 .BI verify_fatal \fR=\fPbool
728 If true, exit the job on the first observed verification failure. Default:
731 .BI verify_dump \fR=\fPbool
732 If set, dump the contents of both the original data block and the data block we
733 read off disk to files. This allows later analysis to inspect just what kind of
734 data corruption occurred. On by default.
736 .BI verify_async \fR=\fPint
737 Fio will normally verify IO inline from the submitting thread. This option
738 takes an integer describing how many async offload threads to create for IO
739 verification instead, causing fio to offload the duty of verifying IO contents
740 to one or more separate threads. If using this offload option, even sync IO
741 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
742 allows them to have IO in flight while verifies are running.
744 .BI verify_async_cpus \fR=\fPstr
745 Tell fio to set the given CPU affinity on the async IO verification threads.
746 See \fBcpus_allowed\fP for the format used.
748 .BI verify_backlog \fR=\fPint
749 Fio will normally verify the written contents of a job that utilizes verify
750 once that job has completed. In other words, everything is written then
751 everything is read back and verified. You may want to verify continually
752 instead for a variety of reasons. Fio stores the meta data associated with an
753 IO block in memory, so for large verify workloads, quite a bit of memory would
754 be used up holding this meta data. If this option is enabled, fio will write
755 only N blocks before verifying these blocks.
757 .BI verify_backlog_batch \fR=\fPint
758 Control how many blocks fio will verify if verify_backlog is set. If not set,
759 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
760 read back and verified). If \fBverify_backlog_batch\fR is less than
761 \fBverify_backlog\fR then not all blocks will be verified, if
762 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
763 will be verified more than once.
766 Wait for preceding jobs in the job file to exit before starting this one.
767 \fBstonewall\fR implies \fBnew_group\fR.
770 Start a new reporting group. If not given, all jobs in a file will be part
771 of the same reporting group, unless separated by a stonewall.
773 .BI numjobs \fR=\fPint
774 Number of clones (processes/threads performing the same workload) of this job.
778 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
782 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
783 with \fBfork\fR\|(2).
785 .BI zonesize \fR=\fPint
786 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
788 .BI zoneskip \fR=\fPint
789 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
792 .BI write_iolog \fR=\fPstr
793 Write the issued I/O patterns to the specified file. Specify a separate file
794 for each job, otherwise the iologs will be interspersed and the file may be
797 .BI read_iolog \fR=\fPstr
798 Replay the I/O patterns contained in the specified file generated by
799 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
801 .BI replay_no_stall \fR=\fPint
802 While replaying I/O patterns using \fBread_iolog\fR the default behavior
803 attempts to respect timing information between I/Os. Enabling
804 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
805 still respecting ordering.
807 .BI replay_redirect \fR=\fPstr
808 While replaying I/O patterns using \fBread_iolog\fR the default behavior
809 is to replay the IOPS onto the major/minor device that each IOP was recorded
810 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
811 single specified device regardless of the device it was recorded from.
813 .B write_bw_log \fR=\fPstr
814 If given, write a bandwidth log of the jobs in this job file. Can be used to
815 store data of the bandwidth of the jobs in their lifetime. The included
816 fio_generate_plots script uses gnuplot to turn these text files into nice
817 graphs. See \fBwrite_log_log\fR for behaviour of given filename. For this
818 option, the postfix is _bw.log.
820 .B write_lat_log \fR=\fPstr
821 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
822 filename is given with this option, the default filename of "jobname_type.log"
823 is used. Even if the filename is given, fio will still append the type of log.
825 .B disable_lat \fR=\fPbool
826 Disable measurements of total latency numbers. Useful only for cutting
827 back the number of calls to gettimeofday, as that does impact performance at
828 really high IOPS rates. Note that to really get rid of a large amount of these
829 calls, this option must be used with disable_slat and disable_bw as well.
831 .B disable_clat \fR=\fPbool
832 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
834 .B disable_slat \fR=\fPbool
835 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
837 .B disable_bw_measurement \fR=\fPbool
838 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
840 .BI lockmem \fR=\fPint
841 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
842 simulate a smaller amount of memory.
844 .BI exec_prerun \fR=\fPstr
845 Before running the job, execute the specified command with \fBsystem\fR\|(3).
847 .BI exec_postrun \fR=\fPstr
848 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
850 .BI ioscheduler \fR=\fPstr
851 Attempt to switch the device hosting the file to the specified I/O scheduler.
853 .BI cpuload \fR=\fPint
854 If the job is a CPU cycle-eater, attempt to use the specified percentage of
857 .BI cpuchunks \fR=\fPint
858 If the job is a CPU cycle-eater, split the load into cycles of the
859 given time in milliseconds.
861 .BI disk_util \fR=\fPbool
862 Generate disk utilization statistics if the platform supports it. Default: true.
864 .BI gtod_reduce \fR=\fPbool
865 Enable all of the gettimeofday() reducing options (disable_clat, disable_slat,
866 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
867 gettimeofday() call count. With this option enabled, we only do about 0.4% of
868 the gtod() calls we would have done if all time keeping was enabled.
870 .BI gtod_cpu \fR=\fPint
871 Sometimes it's cheaper to dedicate a single thread of execution to just getting
872 the current time. Fio (and databases, for instance) are very intensive on
873 gettimeofday() calls. With this option, you can set one CPU aside for doing
874 nothing but logging current time to a shared memory location. Then the other
875 threads/processes that run IO workloads need only copy that segment, instead of
876 entering the kernel with a gettimeofday() call. The CPU set aside for doing
877 these time calls will be excluded from other uses. Fio will manually clear it
878 from the CPU mask of other jobs.
880 .BI cgroup \fR=\fPstr
881 Add job to this control group. If it doesn't exist, it will be created.
882 The system must have a mounted cgroup blkio mount point for this to work. If
883 your system doesn't have it mounted, you can do so with:
885 # mount \-t cgroup \-o blkio none /cgroup
887 .BI cgroup_weight \fR=\fPint
888 Set the weight of the cgroup to this value. See the documentation that comes
889 with the kernel, allowed values are in the range of 100..1000.
891 .BI cgroup_nodelete \fR=\fPbool
892 Normally fio will delete the cgroups it has created after the job completion.
893 To override this behavior and to leave cgroups around after the job completion,
894 set cgroup_nodelete=1. This can be useful if one wants to inspect various
895 cgroup files after job completion. Default: false
898 Instead of running as the invoking user, set the user ID to this value before
899 the thread/process does any work.
902 Set group ID, see \fBuid\fR.
904 While running, \fBfio\fR will display the status of the created jobs. For
908 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
911 The characters in the first set of brackets denote the current status of each
912 threads. The possible values are:
918 Setup but not started.
924 Initialized, waiting.
927 Running, doing sequential reads.
930 Running, doing random reads.
933 Running, doing sequential writes.
936 Running, doing random writes.
939 Running, doing mixed sequential reads/writes.
942 Running, doing mixed random reads/writes.
945 Running, currently waiting for \fBfsync\fR\|(2).
948 Running, verifying written data.
951 Exited, not reaped by main thread.
954 Exited, thread reaped.
958 The second set of brackets shows the estimated completion percentage of
959 the current group. The third set shows the read and write I/O rate,
960 respectively. Finally, the estimated run time of the job is displayed.
962 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
963 for each thread, each group of threads, and each disk, in that order.
965 Per-thread statistics first show the threads client number, group-id, and
966 error code. The remaining figures are as follows:
970 Number of megabytes of I/O performed.
973 Average data rate (bandwidth).
979 Submission latency minimum, maximum, average and standard deviation. This is
980 the time it took to submit the I/O.
983 Completion latency minimum, maximum, average and standard deviation. This
984 is the time between submission and completion.
987 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
988 and standard deviation.
991 CPU usage statistics. Includes user and system time, number of context switches
992 this thread went through and number of major and minor page faults.
995 Distribution of I/O depths. Each depth includes everything less than (or equal)
996 to it, but greater than the previous depth.
999 Number of read/write requests issued, and number of short read/write requests.
1002 Distribution of I/O completion latencies. The numbers follow the same pattern
1006 The group statistics show:
1011 Number of megabytes I/O performed.
1014 Aggregate bandwidth of threads in the group.
1017 Minimum average bandwidth a thread saw.
1020 Maximum average bandwidth a thread saw.
1023 Shortest runtime of threads in the group.
1026 Longest runtime of threads in the group.
1030 Finally, disk statistics are printed with reads first:
1035 Number of I/Os performed by all groups.
1038 Number of merges in the I/O scheduler.
1041 Number of ticks we kept the disk busy.
1044 Total time spent in the disk queue.
1051 If the \fB\-\-minimal\fR option is given, the results will be printed in a
1052 semicolon-delimited format suitable for scripted use - a job description
1053 (if provided) follows on a new line. Note that the first
1054 number in the line is the version number. If the output has to be changed
1055 for some reason, this number will be incremented by 1 to signify that
1056 change. The fields are:
1059 .B version, jobname, groupid, error
1063 .B KB I/O, bandwidth \fR(KB/s)\fP, runtime \fR(ms)\fP
1067 .B min, max, mean, standard deviation
1071 .B min, max, mean, standard deviation
1075 .B min, max, mean, standard deviation
1079 .B min, max, aggregate percentage of total, mean, standard deviation
1085 .B KB I/O, bandwidth \fR(KB/s)\fP, runtime \fR(ms)\fP
1089 .B min, max, mean, standard deviation
1093 .B min, max, mean, standard deviation
1097 .B min, max, mean, standard deviation
1101 .B min, max, aggregate percentage of total, mean, standard deviation
1107 .B user, system, context switches, major page faults, minor page faults
1110 IO depth distribution:
1112 .B <=1, 2, 4, 8, 16, 32, >=64
1115 IO latency distribution:
1119 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1123 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1127 Error Info (dependent on continue_on_error, default off):
1129 .B total # errors, first error code
1132 .B text description (if provided in config - appears on newline)
1136 was written by Jens Axboe <jens.axboe@oracle.com>,
1137 now Jens Axboe <jaxboe@fusionio.com>.
1139 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1140 on documentation by Jens Axboe.
1141 .SH "REPORTING BUGS"
1142 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1145 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1147 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob