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:
101 List of floating numbers: A list of floating numbers, separated by
106 May be used to override the job name. On the command line, this parameter
107 has the special purpose of signalling the start of a new job.
109 .BI description \fR=\fPstr
110 Human-readable description of the job. It is printed when the job is run, but
111 otherwise has no special purpose.
113 .BI directory \fR=\fPstr
114 Prefix filenames with this directory. Used to place files in a location other
117 .BI filename \fR=\fPstr
119 normally makes up a file name based on the job name, thread number, and file
120 number. If you want to share files between threads in a job or several jobs,
121 specify a \fIfilename\fR for each of them to override the default. If the I/O
122 engine used is `net', \fIfilename\fR is the host and port to connect to in the
123 format \fIhost\fR/\fIport\fR. If the I/O engine is file-based, you can specify
124 a number of files by separating the names with a `:' character. `\-' is a
125 reserved name, meaning stdin or stdout, depending on the read/write direction
128 .BI lockfile \fR=\fPstr
129 Fio defaults to not locking any files before it does IO to them. If a file or
130 file descriptor is shared, fio can serialize IO to that file to make the end
131 result consistent. This is usual for emulating real workloads that share files.
137 No locking. This is the default.
140 Only one thread or process may do IO at the time, excluding all others.
143 Read-write locking on the file. Many readers may access the file at the same
144 time, but writes get exclusive access.
147 The option may be post-fixed with a lock batch number. If set, then each
148 thread/process may do that amount of IOs to the file before giving up the lock.
149 Since lock acquisition is expensive, batching the lock/unlocks will speed up IO.
152 .BI opendir \fR=\fPstr
153 Recursively open any files below directory \fIstr\fR.
155 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
156 Type of I/O pattern. Accepted values are:
173 Mixed sequential reads and writes.
176 Mixed random reads and writes.
179 For mixed I/O, the default split is 50/50. For certain types of io the result
180 may still be skewed a bit, since the speed may be different. It is possible to
181 specify a number of IO's to do before getting a new offset, this is one by
182 appending a `:\fI<nr>\fR to the end of the string given. For a random read, it
183 would look like \fBrw=randread:8\fR for passing in an offset modifier with a
184 value of 8. See the \fBrw_sequencer\fR option.
187 .BI rw_sequencer \fR=\fPstr
188 If an offset modifier is given by appending a number to the \fBrw=<str>\fR line,
189 then this option controls how that number modifies the IO offset being
190 generated. Accepted values are:
195 Generate sequential offset
198 Generate the same offset
201 \fBsequential\fR is only useful for random IO, where fio would normally
202 generate a new random offset for every IO. If you append eg 8 to randread, you
203 would get a new random offset for every 8 IO's. The result would be a seek for
204 only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify
205 that. As sequential IO is already sequential, setting \fBsequential\fR for that
206 would not result in any differences. \fBidentical\fR behaves in a similar
207 fashion, except it sends the same offset 8 number of times before generating a
212 .BI kb_base \fR=\fPint
213 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
214 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
215 reasons. Allow values are 1024 or 1000, with 1024 being the default.
217 .BI randrepeat \fR=\fPbool
218 Seed the random number generator in a predictable way so results are repeatable
219 across runs. Default: true.
221 .BI use_os_rand \fR=\fPbool
222 Fio can either use the random generator supplied by the OS to generator random
223 offsets, or it can use it's own internal generator (based on Tausworthe).
224 Default is to use the internal generator, which is often of better quality and
225 faster. Default: false.
227 .BI fallocate \fR=\fPstr
228 Whether pre-allocation is performed when laying down files. Accepted values
234 Do not pre-allocate space.
237 Pre-allocate via posix_fallocate().
240 Pre-allocate via fallocate() with FALLOC_FL_KEEP_SIZE set.
243 Backward-compatible alias for 'none'.
246 Backward-compatible alias for 'posix'.
249 May not be available on all supported platforms. 'keep' is only
250 available on Linux. If using ZFS on Solaris this must be set to 'none'
251 because ZFS doesn't support it. Default: 'posix'.
254 .BI fadvise_hint \fR=\fPbool
255 Disable use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
256 are likely to be issued. Default: true.
259 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
260 been transfered, unless limited by other options (\fBruntime\fR, for instance).
261 Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be
262 divided between the available files for the job. If not set, fio will use the
263 full size of the given files or devices. If the the files do not exist, size
264 must be given. It is also possible to give size as a percentage between 1 and
265 100. If size=20% is given, fio will use 20% of the full size of the given files
268 .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool
269 Sets size to something really large and waits for ENOSPC (no space left on
270 device) as the terminating condition. Only makes sense with sequential write.
271 For a read workload, the mount point will be filled first then IO started on
272 the result. This option doesn't make sense if operating on a raw device node,
273 since the size of that is already known by the file system. Additionally,
274 writing beyond end-of-device will not return ENOSPC there.
276 .BI filesize \fR=\fPirange
277 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
278 for files at random within the given range, limited to \fBsize\fR in total (if
279 that is given). If \fBfilesize\fR is not specified, each created file is the
282 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
283 Block size for I/O units. Default: 4k. Values for reads and writes can be
284 specified separately in the format \fIread\fR,\fIwrite\fR, either of
285 which may be empty to leave that value at its default.
287 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
288 Specify a range of I/O block sizes. The issued I/O unit will always be a
289 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
290 to both reads and writes if only one range is given, but can be specified
291 separately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k.
292 Also (see \fBblocksize\fR).
294 .BI bssplit \fR=\fPstr
295 This option allows even finer grained control of the block sizes issued,
296 not just even splits between them. With this option, you can weight various
297 block sizes for exact control of the issued IO for a job that has mixed
298 block sizes. The format of the option is bssplit=blocksize/percentage,
299 optionally adding as many definitions as needed separated by a colon.
300 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
301 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
302 splits to reads and writes. The format is identical to what the
303 \fBbs\fR option accepts, the read and write parts are separated with a
306 .B blocksize_unaligned\fR,\fP bs_unaligned
307 If set, any size in \fBblocksize_range\fR may be used. This typically won't
308 work with direct I/O, as that normally requires sector alignment.
310 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
311 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
312 the minimum blocksize given. Minimum alignment is typically 512b
313 for using direct IO, though it usually depends on the hardware block size.
314 This option is mutually exclusive with using a random map for files, so it
315 will turn off that option.
318 Initialise buffers with all zeros. Default: fill buffers with random data.
321 If this option is given, fio will refill the IO buffers on every submit. The
322 default is to only fill it at init time and reuse that data. Only makes sense
323 if zero_buffers isn't specified, naturally. If data verification is enabled,
324 refill_buffers is also automatically enabled.
326 .BI nrfiles \fR=\fPint
327 Number of files to use for this job. Default: 1.
329 .BI openfiles \fR=\fPint
330 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
332 .BI file_service_type \fR=\fPstr
333 Defines how files to service are selected. The following types are defined:
338 Choose a file at random
341 Round robin over open files (default).
343 Do each file in the set sequentially.
346 The number of I/Os to issue before switching a new file can be specified by
347 appending `:\fIint\fR' to the service type.
350 .BI ioengine \fR=\fPstr
351 Defines how the job issues I/O. The following types are defined:
356 Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
357 position the I/O location.
360 Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O.
363 Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by
364 coalescing adjacents IOs into a single submission.
367 Linux native asynchronous I/O.
370 POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
373 Solaris native asynchronous I/O.
376 Windows native asynchronous I/O.
379 File is memory mapped with \fImmap\fR\|(2) and data copied using
383 \fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
384 transfer data from user-space to the kernel.
387 Use the syslet system calls to make regular read/write asynchronous.
390 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
391 the target is an sg character device, we use \fIread\fR\|(2) and
392 \fIwrite\fR\|(2) for asynchronous I/O.
395 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
396 itself and for debugging and testing purposes.
399 Transfer over the network. \fBfilename\fR must be set appropriately to
400 `\fIhost\fR/\fIport\fR' regardless of data direction. If receiving, only the
401 \fIport\fR argument is used.
404 Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
408 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
409 \fBcpucycles\fR parameters.
412 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
413 approach to asycnronous I/O.
415 See <http://www.xmailserver.org/guasi\-lib.html>.
418 The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ)
419 and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols.
422 Loads an external I/O engine object file. Append the engine filename as
427 .BI iodepth \fR=\fPint
428 Number of I/O units to keep in flight against the file. Note that increasing
429 iodepth beyond 1 will not affect synchronous ioengines (except for small
430 degress when verify_async is in use). Even async engines my impose OS
431 restrictions causing the desired depth not to be achieved. This may happen on
432 Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is
433 not async on that OS. Keep an eye on the IO depth distribution in the
434 fio output to verify that the achieved depth is as expected. Default: 1.
436 .BI iodepth_batch \fR=\fPint
437 Number of I/Os to submit at once. Default: \fBiodepth\fR.
439 .BI iodepth_batch_complete \fR=\fPint
440 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
441 means that we'll ask for a minimum of 1 IO in the retrieval process from the
442 kernel. The IO retrieval will go on until we hit the limit set by
443 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
444 completed events before queuing more IO. This helps reduce IO latency, at the
445 cost of more retrieval system calls.
447 .BI iodepth_low \fR=\fPint
448 Low watermark indicating when to start filling the queue again. Default:
451 .BI direct \fR=\fPbool
452 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
454 .BI buffered \fR=\fPbool
455 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
458 .BI offset \fR=\fPint
459 Offset in the file to start I/O. Data before the offset will not be touched.
462 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
463 0, don't sync. Default: 0.
465 .BI fdatasync \fR=\fPint
466 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
467 data parts of the file. Default: 0.
469 .BI sync_file_range \fR=\fPstr:int
470 Use sync_file_range() for every \fRval\fP number of write operations. Fio will
471 track range of writes that have happened since the last sync_file_range() call.
472 \fRstr\fP can currently be one or more of:
476 SYNC_FILE_RANGE_WAIT_BEFORE
479 SYNC_FILE_RANGE_WRITE
482 SYNC_FILE_RANGE_WRITE
486 So if you do sync_file_range=wait_before,write:8, fio would use
487 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
488 Also see the sync_file_range(2) man page. This option is Linux specific.
490 .BI overwrite \fR=\fPbool
491 If writing, setup the file first and do overwrites. Default: false.
493 .BI end_fsync \fR=\fPbool
494 Sync file contents when job exits. Default: false.
496 .BI fsync_on_close \fR=\fPbool
497 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
498 it will happen on every close, not just at the end of the job. Default: false.
500 .BI rwmixcycle \fR=\fPint
501 How many milliseconds before switching between reads and writes for a mixed
502 workload. Default: 500ms.
504 .BI rwmixread \fR=\fPint
505 Percentage of a mixed workload that should be reads. Default: 50.
507 .BI rwmixwrite \fR=\fPint
508 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
509 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
510 overrides the first. This may interfere with a given rate setting, if fio is
511 asked to limit reads or writes to a certain rate. If that is the case, then
512 the distribution may be skewed. Default: 50.
515 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
516 this parameter is given, a new offset will be chosen without looking at past
517 I/O history. This parameter is mutually exclusive with \fBverify\fR.
519 .BI softrandommap \fR=\fPbool
520 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
521 fails to allocate the map, if this option is set it will continue without a
522 random block map. As coverage will not be as complete as with random maps, this
523 option is disabled by default.
526 Run job with given nice value. See \fInice\fR\|(2).
529 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
532 .BI prioclass \fR=\fPint
533 Set I/O priority class. See \fIionice\fR\|(1).
535 .BI thinktime \fR=\fPint
536 Stall job for given number of microseconds between issuing I/Os.
538 .BI thinktime_spin \fR=\fPint
539 Pretend to spend CPU time for given number of microseconds, sleeping the rest
540 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
542 .BI thinktime_blocks \fR=\fPint
543 Number of blocks to issue before waiting \fBthinktime\fR microseconds.
547 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
548 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
549 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
550 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
551 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
552 limit writes (to 500KB/sec), the latter will only limit reads.
554 .BI ratemin \fR=\fPint
555 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
556 Failing to meet this requirement will cause the job to exit. The same format
557 as \fBrate\fR is used for read vs write separation.
559 .BI rate_iops \fR=\fPint
560 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
561 specified independently of bandwidth. The same format as \fBrate\fR is used for
562 read vs write seperation. If \fBblocksize\fR is a range, the smallest block
563 size is used as the metric.
565 .BI rate_iops_min \fR=\fPint
566 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
567 is used for read vs write seperation.
569 .BI ratecycle \fR=\fPint
570 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
571 milliseconds. Default: 1000ms.
573 .BI cpumask \fR=\fPint
574 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
575 may run on. See \fBsched_setaffinity\fR\|(2).
577 .BI cpus_allowed \fR=\fPstr
578 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
580 .BI startdelay \fR=\fPint
581 Delay start of job for the specified number of seconds.
583 .BI runtime \fR=\fPint
584 Terminate processing after the specified number of seconds.
587 If given, run for the specified \fBruntime\fR duration even if the files are
588 completely read or written. The same workload will be repeated as many times
589 as \fBruntime\fR allows.
591 .BI ramp_time \fR=\fPint
592 If set, fio will run the specified workload for this amount of time before
593 logging any performance numbers. Useful for letting performance settle before
594 logging results, thus minimizing the runtime required for stable results. Note
595 that the \fBramp_time\fR is considered lead in time for a job, thus it will
596 increase the total runtime if a special timeout or runtime is specified.
598 .BI invalidate \fR=\fPbool
599 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
602 Use synchronous I/O for buffered writes. For the majority of I/O engines,
603 this means using O_SYNC. Default: false.
605 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
606 Allocation method for I/O unit buffer. Allowed values are:
611 Allocate memory with \fImalloc\fR\|(3).
614 Use shared memory buffers allocated through \fIshmget\fR\|(2).
617 Same as \fBshm\fR, but use huge pages as backing.
620 Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
621 is given after the option in the format `:\fIfile\fR'.
624 Same as \fBmmap\fR, but use huge files as backing.
627 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
628 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
629 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
630 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
631 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
632 and the documentation for that. Normally you just need to echo an appropriate
633 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
637 .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint
638 This indiciates the memory alignment of the IO memory buffers. Note that the
639 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
640 the alignment of the following buffers are given by the \fBbs\fR used. In
641 other words, if using a \fBbs\fR that is a multiple of the page sized in the
642 system, all buffers will be aligned to this value. If using a \fBbs\fR that
643 is not page aligned, the alignment of subsequent IO memory buffers is the
644 sum of the \fBiomem_align\fR and \fBbs\fR used.
646 .BI hugepage\-size \fR=\fPint
647 Defines the size of a huge page. Must be at least equal to the system setting.
648 Should be a multiple of 1MB. Default: 4MB.
651 Terminate all jobs when one finishes. Default: wait for each job to finish.
653 .BI bwavgtime \fR=\fPint
654 Average bandwidth calculations over the given time in milliseconds. Default:
657 .BI create_serialize \fR=\fPbool
658 If true, serialize file creation for the jobs. Default: true.
660 .BI create_fsync \fR=\fPbool
661 \fIfsync\fR\|(2) data file after creation. Default: true.
663 .BI create_on_open \fR=\fPbool
664 If true, the files are not created until they are opened for IO by the job.
666 .BI pre_read \fR=\fPbool
667 If this is given, files will be pre-read into memory before starting the given
668 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
669 pointless to pre-read and then drop the cache. This will only work for IO
670 engines that are seekable, since they allow you to read the same data
671 multiple times. Thus it will not work on eg network or splice IO.
673 .BI unlink \fR=\fPbool
674 Unlink job files when done. Default: false.
677 Specifies the number of iterations (runs of the same workload) of this job.
680 .BI do_verify \fR=\fPbool
681 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
684 .BI verify \fR=\fPstr
685 Method of verifying file contents after each iteration of the job. Allowed
690 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1
691 Store appropriate checksum in the header of each block. crc32c-intel is
692 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
693 not supported by the system.
696 Write extra information about each I/O (timestamp, block number, etc.). The
697 block number is verified. See \fBverify_pattern\fR as well.
700 Pretend to verify. Used for testing internals.
703 This option can be used for repeated burn-in tests of a system to make sure
704 that the written data is also correctly read back. If the data direction given
705 is a read or random read, fio will assume that it should verify a previously
706 written file. If the data direction includes any form of write, the verify will
707 be of the newly written data.
710 .BI verify_sort \fR=\fPbool
711 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
712 read them back in a sorted manner. Default: true.
714 .BI verify_offset \fR=\fPint
715 Swap the verification header with data somewhere else in the block before
716 writing. It is swapped back before verifying.
718 .BI verify_interval \fR=\fPint
719 Write the verification header for this number of bytes, which should divide
720 \fBblocksize\fR. Default: \fBblocksize\fR.
722 .BI verify_pattern \fR=\fPstr
723 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
724 with totally random bytes, but sometimes it's interesting to fill with a known
725 pattern for io verification purposes. Depending on the width of the pattern,
726 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
727 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
728 has to be a hex number that starts with either "0x" or "0X". Use with
731 .BI verify_fatal \fR=\fPbool
732 If true, exit the job on the first observed verification failure. Default:
735 .BI verify_dump \fR=\fPbool
736 If set, dump the contents of both the original data block and the data block we
737 read off disk to files. This allows later analysis to inspect just what kind of
738 data corruption occurred. On by default.
740 .BI verify_async \fR=\fPint
741 Fio will normally verify IO inline from the submitting thread. This option
742 takes an integer describing how many async offload threads to create for IO
743 verification instead, causing fio to offload the duty of verifying IO contents
744 to one or more separate threads. If using this offload option, even sync IO
745 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
746 allows them to have IO in flight while verifies are running.
748 .BI verify_async_cpus \fR=\fPstr
749 Tell fio to set the given CPU affinity on the async IO verification threads.
750 See \fBcpus_allowed\fP for the format used.
752 .BI verify_backlog \fR=\fPint
753 Fio will normally verify the written contents of a job that utilizes verify
754 once that job has completed. In other words, everything is written then
755 everything is read back and verified. You may want to verify continually
756 instead for a variety of reasons. Fio stores the meta data associated with an
757 IO block in memory, so for large verify workloads, quite a bit of memory would
758 be used up holding this meta data. If this option is enabled, fio will write
759 only N blocks before verifying these blocks.
761 .BI verify_backlog_batch \fR=\fPint
762 Control how many blocks fio will verify if verify_backlog is set. If not set,
763 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
764 read back and verified). If \fBverify_backlog_batch\fR is less than
765 \fBverify_backlog\fR then not all blocks will be verified, if
766 \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks
767 will be verified more than once.
769 .B stonewall "\fR,\fP wait_for_previous"
770 Wait for preceding jobs in the job file to exit before starting this one.
771 \fBstonewall\fR implies \fBnew_group\fR.
774 Start a new reporting group. If not given, all jobs in a file will be part
775 of the same reporting group, unless separated by a stonewall.
777 .BI numjobs \fR=\fPint
778 Number of clones (processes/threads performing the same workload) of this job.
782 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
786 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
787 with \fBfork\fR\|(2).
789 .BI zonesize \fR=\fPint
790 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
792 .BI zoneskip \fR=\fPint
793 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
796 .BI write_iolog \fR=\fPstr
797 Write the issued I/O patterns to the specified file. Specify a separate file
798 for each job, otherwise the iologs will be interspersed and the file may be
801 .BI read_iolog \fR=\fPstr
802 Replay the I/O patterns contained in the specified file generated by
803 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
805 .BI replay_no_stall \fR=\fPint
806 While replaying I/O patterns using \fBread_iolog\fR the default behavior
807 attempts to respect timing information between I/Os. Enabling
808 \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while
809 still respecting ordering.
811 .BI replay_redirect \fR=\fPstr
812 While replaying I/O patterns using \fBread_iolog\fR the default behavior
813 is to replay the IOPS onto the major/minor device that each IOP was recorded
814 from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the
815 single specified device regardless of the device it was recorded from.
817 .B write_bw_log \fR=\fPstr
818 If given, write a bandwidth log of the jobs in this job file. Can be used to
819 store data of the bandwidth of the jobs in their lifetime. The included
820 fio_generate_plots script uses gnuplot to turn these text files into nice
821 graphs. See \fBwrite_log_log\fR for behaviour of given filename. For this
822 option, the postfix is _bw.log.
824 .B write_lat_log \fR=\fPstr
825 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
826 filename is given with this option, the default filename of "jobname_type.log"
827 is used. Even if the filename is given, fio will still append the type of log.
829 .B disable_lat \fR=\fPbool
830 Disable measurements of total latency numbers. Useful only for cutting
831 back the number of calls to gettimeofday, as that does impact performance at
832 really high IOPS rates. Note that to really get rid of a large amount of these
833 calls, this option must be used with disable_slat and disable_bw as well.
835 .B disable_clat \fR=\fPbool
836 Disable measurements of completion latency numbers. See \fBdisable_lat\fR.
838 .B disable_slat \fR=\fPbool
839 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
841 .B disable_bw_measurement \fR=\fPbool
842 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
844 .BI lockmem \fR=\fPint
845 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
846 simulate a smaller amount of memory.
848 .BI exec_prerun \fR=\fPstr
849 Before running the job, execute the specified command with \fBsystem\fR\|(3).
851 .BI exec_postrun \fR=\fPstr
852 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
854 .BI ioscheduler \fR=\fPstr
855 Attempt to switch the device hosting the file to the specified I/O scheduler.
857 .BI cpuload \fR=\fPint
858 If the job is a CPU cycle-eater, attempt to use the specified percentage of
861 .BI cpuchunks \fR=\fPint
862 If the job is a CPU cycle-eater, split the load into cycles of the
863 given time in milliseconds.
865 .BI disk_util \fR=\fPbool
866 Generate disk utilization statistics if the platform supports it. Default: true.
868 .BI gtod_reduce \fR=\fPbool
869 Enable all of the gettimeofday() reducing options (disable_clat, disable_slat,
870 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
871 gettimeofday() call count. With this option enabled, we only do about 0.4% of
872 the gtod() calls we would have done if all time keeping was enabled.
874 .BI gtod_cpu \fR=\fPint
875 Sometimes it's cheaper to dedicate a single thread of execution to just getting
876 the current time. Fio (and databases, for instance) are very intensive on
877 gettimeofday() calls. With this option, you can set one CPU aside for doing
878 nothing but logging current time to a shared memory location. Then the other
879 threads/processes that run IO workloads need only copy that segment, instead of
880 entering the kernel with a gettimeofday() call. The CPU set aside for doing
881 these time calls will be excluded from other uses. Fio will manually clear it
882 from the CPU mask of other jobs.
884 .BI cgroup \fR=\fPstr
885 Add job to this control group. If it doesn't exist, it will be created.
886 The system must have a mounted cgroup blkio mount point for this to work. If
887 your system doesn't have it mounted, you can do so with:
889 # mount \-t cgroup \-o blkio none /cgroup
891 .BI cgroup_weight \fR=\fPint
892 Set the weight of the cgroup to this value. See the documentation that comes
893 with the kernel, allowed values are in the range of 100..1000.
895 .BI cgroup_nodelete \fR=\fPbool
896 Normally fio will delete the cgroups it has created after the job completion.
897 To override this behavior and to leave cgroups around after the job completion,
898 set cgroup_nodelete=1. This can be useful if one wants to inspect various
899 cgroup files after job completion. Default: false
902 Instead of running as the invoking user, set the user ID to this value before
903 the thread/process does any work.
906 Set group ID, see \fBuid\fR.
908 .BI clat_percentiles \fR=\fPbool
909 Enable the reporting of percentiles of completion latencies.
911 .BI percentile_list \fR=\fPfloat_list
912 Overwrite the default list of percentiles for completion
913 latencies. Each number is a floating number in the range (0,100], and
914 the maximum length of the list is 20. Use ':' to separate the
915 numbers. For example, --percentile_list=99.5:99.9 will cause fio to
916 report the values of completion latency below which 99.5% and 99.9% of
917 the observed latencies fell, respectively.
919 While running, \fBfio\fR will display the status of the created jobs. For
923 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
926 The characters in the first set of brackets denote the current status of each
927 threads. The possible values are:
933 Setup but not started.
939 Initialized, waiting.
942 Running, doing sequential reads.
945 Running, doing random reads.
948 Running, doing sequential writes.
951 Running, doing random writes.
954 Running, doing mixed sequential reads/writes.
957 Running, doing mixed random reads/writes.
960 Running, currently waiting for \fBfsync\fR\|(2).
963 Running, verifying written data.
966 Exited, not reaped by main thread.
969 Exited, thread reaped.
973 The second set of brackets shows the estimated completion percentage of
974 the current group. The third set shows the read and write I/O rate,
975 respectively. Finally, the estimated run time of the job is displayed.
977 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
978 for each thread, each group of threads, and each disk, in that order.
980 Per-thread statistics first show the threads client number, group-id, and
981 error code. The remaining figures are as follows:
985 Number of megabytes of I/O performed.
988 Average data rate (bandwidth).
994 Submission latency minimum, maximum, average and standard deviation. This is
995 the time it took to submit the I/O.
998 Completion latency minimum, maximum, average and standard deviation. This
999 is the time between submission and completion.
1002 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
1003 and standard deviation.
1006 CPU usage statistics. Includes user and system time, number of context switches
1007 this thread went through and number of major and minor page faults.
1010 Distribution of I/O depths. Each depth includes everything less than (or equal)
1011 to it, but greater than the previous depth.
1014 Number of read/write requests issued, and number of short read/write requests.
1017 Distribution of I/O completion latencies. The numbers follow the same pattern
1021 The group statistics show:
1026 Number of megabytes I/O performed.
1029 Aggregate bandwidth of threads in the group.
1032 Minimum average bandwidth a thread saw.
1035 Maximum average bandwidth a thread saw.
1038 Shortest runtime of threads in the group.
1041 Longest runtime of threads in the group.
1045 Finally, disk statistics are printed with reads first:
1050 Number of I/Os performed by all groups.
1053 Number of merges in the I/O scheduler.
1056 Number of ticks we kept the disk busy.
1059 Total time spent in the disk queue.
1066 If the \fB\-\-minimal\fR option is given, the results will be printed in a
1067 semicolon-delimited format suitable for scripted use - a job description
1068 (if provided) follows on a new line. Note that the first
1069 number in the line is the version number. If the output has to be changed
1070 for some reason, this number will be incremented by 1 to signify that
1071 change. The fields are:
1074 .B version, jobname, groupid, error
1078 .B KB I/O, bandwidth \fR(KB/s)\fP, runtime \fR(ms)\fP
1082 .B min, max, mean, standard deviation
1086 .B min, max, mean, standard deviation
1090 .B min, max, mean, standard deviation
1094 .B min, max, aggregate percentage of total, mean, standard deviation
1100 .B KB I/O, bandwidth \fR(KB/s)\fP, runtime \fR(ms)\fP
1104 .B min, max, mean, standard deviation
1108 .B min, max, mean, standard deviation
1112 .B min, max, mean, standard deviation
1116 .B min, max, aggregate percentage of total, mean, standard deviation
1122 .B user, system, context switches, major page faults, minor page faults
1125 IO depth distribution:
1127 .B <=1, 2, 4, 8, 16, 32, >=64
1130 IO latency distribution:
1134 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1138 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1142 Error Info (dependent on continue_on_error, default off):
1144 .B total # errors, first error code
1147 .B text description (if provided in config - appears on newline)
1151 was written by Jens Axboe <jens.axboe@oracle.com>,
1152 now Jens Axboe <jaxboe@fusionio.com>.
1154 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1155 on documentation by Jens Axboe.
1156 .SH "REPORTING BUGS"
1157 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1160 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1162 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob