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 seperated 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
85 'b', 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 random I/O, the number of I/Os
176 to perform before getting a new offset can be specified by appending
177 `:\fIint\fR' to the pattern type. The default is 1.
180 .BI kb_base \fR=\fPint
181 The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage
182 manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious
183 reasons. Allow values are 1024 or 1000, with 1024 being the default.
185 .BI randrepeat \fR=\fPbool
186 Seed the random number generator in a predictable way so results are repeatable
187 across runs. Default: true.
189 .BI fallocate \fR=\fPbool
190 By default, fio will use fallocate() to advise the system of the size of the
191 file we are going to write. This can be turned off with fallocate=0. May not
192 be available on all supported platforms.
194 .BI fadvise_hint \fR=\fPbool
195 Disable use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
196 are likely to be issued. Default: true.
199 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
200 been transfered, unless limited by other options (\fBruntime\fR, for instance).
201 Unless \fBnr_files\fR and \fBfilesize\fR options are given, this amount will be
202 divided between the available files for the job. If not set, fio will use the
203 full size of the given files or devices. If the the files do not exist, size
206 .BI fill_device \fR=\fPbool
207 Sets size to something really large and waits for ENOSPC (no space left on
208 device) as the terminating condition. Only makes sense with sequential write.
209 For a read workload, the mount point will be filled first then IO started on
212 .BI filesize \fR=\fPirange
213 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
214 for files at random within the given range, limited to \fBsize\fR in total (if
215 that is given). If \fBfilesize\fR is not specified, each created file is the
218 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
219 Block size for I/O units. Default: 4k. Values for reads and writes can be
220 specified separately in the format \fIread\fR,\fIwrite\fR, either of
221 which may be empty to leave that value at its default.
223 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
224 Specify a range of I/O block sizes. The issued I/O unit will always be a
225 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
226 to both reads and writes if only one range is given, but can be specified
227 separately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k.
228 Also (see \fBblocksize\fR).
230 .BI bssplit \fR=\fPstr
231 This option allows even finer grained control of the block sizes issued,
232 not just even splits between them. With this option, you can weight various
233 block sizes for exact control of the issued IO for a job that has mixed
234 block sizes. The format of the option is bssplit=blocksize/percentage,
235 optionally adding as many definitions as needed seperated by a colon.
236 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
237 blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate
238 splits to reads and writes. The format is identical to what the
239 \fBbs\fR option accepts, the read and write parts are separated with a
242 .B blocksize_unaligned\fR,\fP bs_unaligned
243 If set, any size in \fBblocksize_range\fR may be used. This typically won't
244 work with direct I/O, as that normally requires sector alignment.
246 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
247 At what boundary to align random IO offsets. Defaults to the same as 'blocksize'
248 the minimum blocksize given. Minimum alignment is typically 512b
249 for using direct IO, though it usually depends on the hardware block size.
250 This option is mutually exclusive with using a random map for files, so it
251 will turn off that option.
254 Initialise buffers with all zeros. Default: fill buffers with random data.
257 If this option is given, fio will refill the IO buffers on every submit. The
258 default is to only fill it at init time and reuse that data. Only makes sense
259 if zero_buffers isn't specified, naturally. If data verification is enabled,
260 refill_buffers is also automatically enabled.
262 .BI nrfiles \fR=\fPint
263 Number of files to use for this job. Default: 1.
265 .BI openfiles \fR=\fPint
266 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
268 .BI file_service_type \fR=\fPstr
269 Defines how files to service are selected. The following types are defined:
274 Choose a file at random
277 Round robin over open files (default).
279 Do each file in the set sequentially.
282 The number of I/Os to issue before switching a new file can be specified by
283 appending `:\fIint\fR' to the service type.
286 .BI ioengine \fR=\fPstr
287 Defines how the job issues I/O. The following types are defined:
292 Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
293 position the I/O location.
296 Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O.
299 Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by
300 coalescing adjacents IOs into a single submission.
303 Linux native asynchronous I/O.
306 glibc POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
309 File is memory mapped with \fImmap\fR\|(2) and data copied using
313 \fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
314 transfer data from user-space to the kernel.
317 Use the syslet system calls to make regular read/write asynchronous.
320 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
321 the target is an sg character device, we use \fIread\fR\|(2) and
322 \fIwrite\fR\|(2) for asynchronous I/O.
325 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
326 itself and for debugging and testing purposes.
329 Transfer over the network. \fBfilename\fR must be set appropriately to
330 `\fIhost\fR/\fIport\fR' regardless of data direction. If receiving, only the
331 \fIport\fR argument is used.
334 Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
338 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
339 \fBcpucycles\fR parameters.
342 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
343 approach to asycnronous I/O.
345 See <http://www.xmailserver.org/guasi\-lib.html>.
348 Loads an external I/O engine object file. Append the engine filename as
353 .BI iodepth \fR=\fPint
354 Number of I/O units to keep in flight against the file. Default: 1.
356 .BI iodepth_batch \fR=\fPint
357 Number of I/Os to submit at once. Default: \fBiodepth\fR.
359 .BI iodepth_batch_complete \fR=\fPint
360 This defines how many pieces of IO to retrieve at once. It defaults to 1 which
361 means that we'll ask for a minimum of 1 IO in the retrieval process from the
362 kernel. The IO retrieval will go on until we hit the limit set by
363 \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for
364 completed events before queuing more IO. This helps reduce IO latency, at the
365 cost of more retrieval system calls.
367 .BI iodepth_low \fR=\fPint
368 Low watermark indicating when to start filling the queue again. Default:
371 .BI direct \fR=\fPbool
372 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
374 .BI buffered \fR=\fPbool
375 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
378 .BI offset \fR=\fPint
379 Offset in the file to start I/O. Data before the offset will not be touched.
382 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
383 0, don't sync. Default: 0.
385 .BI fdatasync \fR=\fPint
386 Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the
387 data parts of the file. Default: 0.
389 .BI sync_file_range \fR=\fPstr:int
390 Use sync_file_range() for every \fRval\fP number of write operations. Fio will
391 track range of writes that have happened since the last sync_file_range() call.
392 \fRstr\fP can currently be one or more of:
396 SYNC_FILE_RANGE_WAIT_BEFORE
399 SYNC_FILE_RANGE_WRITE
402 SYNC_FILE_RANGE_WRITE
406 So if you do sync_file_range=wait_before,write:8, fio would use
407 \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes.
408 Also see the sync_file_range(2) man page. This option is Linux specific.
410 .BI overwrite \fR=\fPbool
411 If writing, setup the file first and do overwrites. Default: false.
413 .BI end_fsync \fR=\fPbool
414 Sync file contents when job exits. Default: false.
416 .BI fsync_on_close \fR=\fPbool
417 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
418 it will happen on every close, not just at the end of the job. Default: false.
420 .BI rwmixcycle \fR=\fPint
421 How many milliseconds before switching between reads and writes for a mixed
422 workload. Default: 500ms.
424 .BI rwmixread \fR=\fPint
425 Percentage of a mixed workload that should be reads. Default: 50.
427 .BI rwmixwrite \fR=\fPint
428 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
429 \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two
430 overrides the first. This may interfere with a given rate setting, if fio is
431 asked to limit reads or writes to a certain rate. If that is the case, then
432 the distribution may be skewed. Default: 50.
435 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
436 this parameter is given, a new offset will be chosen without looking at past
437 I/O history. This parameter is mutually exclusive with \fBverify\fR.
440 See \fBnorandommap\fR. If fio runs with the random block map enabled and it
441 fails to allocate the map, if this option is set it will continue without a
442 random block map. As coverage will not be as complete as with random maps, this
443 option is disabled by default.
446 Run job with given nice value. See \fInice\fR\|(2).
449 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
452 .BI prioclass \fR=\fPint
453 Set I/O priority class. See \fIionice\fR\|(1).
455 .BI thinktime \fR=\fPint
456 Stall job for given number of microseconds between issuing I/Os.
458 .BI thinktime_spin \fR=\fPint
459 Pretend to spend CPU time for given number of microseconds, sleeping the rest
460 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
462 .BI thinktime_blocks \fR=\fPint
463 Number of blocks to issue before waiting \fBthinktime\fR microseconds.
467 Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix
468 rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each,
469 or you can specify read and writes separately. Using \fBrate\fR=1m,500k would
470 limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes
471 can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only
472 limit writes (to 500KB/sec), the latter will only limit reads.
474 .BI ratemin \fR=\fPint
475 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
476 Failing to meet this requirement will cause the job to exit. The same format
477 as \fBrate\fR is used for read vs write separation.
479 .BI rate_iops \fR=\fPint
480 Cap the bandwidth to this number of IOPS. Basically the same as rate, just
481 specified independently of bandwidth. The same format as \fBrate\fR is used for
482 read vs write seperation. If \fBblocksize\fR is a range, the smallest block
483 size is used as the metric.
485 .BI rate_iops_min \fR=\fPint
486 If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR
487 is used for read vs write seperation.
489 .BI ratecycle \fR=\fPint
490 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
491 milliseconds. Default: 1000ms.
493 .BI cpumask \fR=\fPint
494 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
495 may run on. See \fBsched_setaffinity\fR\|(2).
497 .BI cpus_allowed \fR=\fPstr
498 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
500 .BI startdelay \fR=\fPint
501 Delay start of job for the specified number of seconds.
503 .BI runtime \fR=\fPint
504 Terminate processing after the specified number of seconds.
507 If given, run for the specified \fBruntime\fR duration even if the files are
508 completely read or written. The same workload will be repeated as many times
509 as \fBruntime\fR allows.
511 .BI ramp_time \fR=\fPint
512 If set, fio will run the specified workload for this amount of time before
513 logging any performance numbers. Useful for letting performance settle before
514 logging results, thus minimizing the runtime required for stable results. Note
515 that the \fBramp_time\fR is considered lead in time for a job, thus it will
516 increase the total runtime if a special timeout or runtime is specified.
518 .BI invalidate \fR=\fPbool
519 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
522 Use synchronous I/O for buffered writes. For the majority of I/O engines,
523 this means using O_SYNC. Default: false.
525 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
526 Allocation method for I/O unit buffer. Allowed values are:
531 Allocate memory with \fImalloc\fR\|(3).
534 Use shared memory buffers allocated through \fIshmget\fR\|(2).
537 Same as \fBshm\fR, but use huge pages as backing.
540 Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
541 is given after the option in the format `:\fIfile\fR'.
544 Same as \fBmmap\fR, but use huge files as backing.
547 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
548 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
549 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
550 have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux,
551 huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR
552 and the documentation for that. Normally you just need to echo an appropriate
553 number, eg echoing 8 will ensure that the OS has 8 huge pages ready for
557 .BI iomem_align \fR=\fPint
558 This indiciates the memory alignment of the IO memory buffers. Note that the
559 given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR
560 the alignment of the following buffers are given by the \fBbs\fR used. In
561 other words, if using a \fBbs\fR that is a multiple of the page sized in the
562 system, all buffers will be aligned to this value. If using a \fBbs\fR that
563 is not page aligned, the alignment of subsequent IO memory buffers is the
564 sum of the \fBiomem_align\fR and \fBbs\fR used.
566 .BI hugepage\-size \fR=\fPint
567 Defines the size of a huge page. Must be at least equal to the system setting.
568 Should be a multiple of 1MB. Default: 4MB.
571 Terminate all jobs when one finishes. Default: wait for each job to finish.
573 .BI bwavgtime \fR=\fPint
574 Average bandwidth calculations over the given time in milliseconds. Default:
577 .BI create_serialize \fR=\fPbool
578 If true, serialize file creation for the jobs. Default: true.
580 .BI create_fsync \fR=\fPbool
581 \fIfsync\fR\|(2) data file after creation. Default: true.
583 .BI create_on_open \fR=\fPbool
584 If true, the files are not created until they are opened for IO by the job.
586 .BI pre_read \fR=\fPbool
587 If this is given, files will be pre-read into memory before starting the given
588 IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is
589 pointless to pre-read and then drop the cache. This will only work for IO
590 engines that are seekable, since they allow you to read the same data
591 multiple times. Thus it will not work on eg network or splice IO.
593 .BI unlink \fR=\fPbool
594 Unlink job files when done. Default: false.
597 Specifies the number of iterations (runs of the same workload) of this job.
600 .BI do_verify \fR=\fPbool
601 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
604 .BI verify \fR=\fPstr
605 Method of verifying file contents after each iteration of the job. Allowed
610 .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1
611 Store appropriate checksum in the header of each block. crc32c-intel is
612 hardware accelerated SSE4.2 driven, falls back to regular crc32c if
613 not supported by the system.
616 Write extra information about each I/O (timestamp, block number, etc.). The
617 block number is verified. See \fBverify_pattern\fR as well.
620 Pretend to verify. Used for testing internals.
623 This option can be used for repeated burn-in tests of a system to make sure
624 that the written data is also correctly read back. If the data direction given
625 is a read or random read, fio will assume that it should verify a previously
626 written file. If the data direction includes any form of write, the verify will
627 be of the newly written data.
630 .BI verify_sort \fR=\fPbool
631 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
632 read them back in a sorted manner. Default: true.
634 .BI verify_offset \fR=\fPint
635 Swap the verification header with data somewhere else in the block before
636 writing. It is swapped back before verifying.
638 .BI verify_interval \fR=\fPint
639 Write the verification header for this number of bytes, which should divide
640 \fBblocksize\fR. Default: \fBblocksize\fR.
642 .BI verify_pattern \fR=\fPstr
643 If set, fio will fill the io buffers with this pattern. Fio defaults to filling
644 with totally random bytes, but sometimes it's interesting to fill with a known
645 pattern for io verification purposes. Depending on the width of the pattern,
646 fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a
647 decimal or a hex number). The verify_pattern if larger than a 32-bit quantity
648 has to be a hex number that starts with either "0x" or "0X". Use with
651 .BI verify_fatal \fR=\fPbool
652 If true, exit the job on the first observed verification failure. Default:
655 .BI verify_async \fR=\fPint
656 Fio will normally verify IO inline from the submitting thread. This option
657 takes an integer describing how many async offload threads to create for IO
658 verification instead, causing fio to offload the duty of verifying IO contents
659 to one or more separate threads. If using this offload option, even sync IO
660 engines can benefit from using an \fBiodepth\fR setting higher than 1, as it
661 allows them to have IO in flight while verifies are running.
663 .BI verify_async_cpus \fR=\fPstr
664 Tell fio to set the given CPU affinity on the async IO verification threads.
665 See \fBcpus_allowed\fP for the format used.
667 .BI verify_backlog \fR=\fPint
668 Fio will normally verify the written contents of a job that utilizes verify
669 once that job has completed. In other words, everything is written then
670 everything is read back and verified. You may want to verify continually
671 instead for a variety of reasons. Fio stores the meta data associated with an
672 IO block in memory, so for large verify workloads, quite a bit of memory would
673 be used up holding this meta data. If this option is enabled, fio will verify
674 the previously written blocks before continuing to write new ones.
676 .BI verify_backlog_batch \fR=\fPint
677 Control how many blocks fio will verify if verify_backlog is set. If not set,
678 will default to the value of \fBverify_backlog\fR (meaning the entire queue is
679 read back and verified).
682 Wait for preceeding jobs in the job file to exit before starting this one.
683 \fBstonewall\fR implies \fBnew_group\fR.
686 Start a new reporting group. If not given, all jobs in a file will be part
687 of the same reporting group, unless separated by a stonewall.
689 .BI numjobs \fR=\fPint
690 Number of clones (processes/threads performing the same workload) of this job.
694 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
698 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
699 with \fBfork\fR\|(2).
701 .BI zonesize \fR=\fPint
702 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
704 .BI zoneskip \fR=\fPint
705 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
708 .BI write_iolog \fR=\fPstr
709 Write the issued I/O patterns to the specified file.
711 .BI read_iolog \fR=\fPstr
712 Replay the I/O patterns contained in the specified file generated by
713 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
715 .B write_bw_log \fR=\fPstr
716 If given, write a bandwidth log of the jobs in this job file. Can be used to
717 store data of the bandwidth of the jobs in their lifetime. The included
718 fio_generate_plots script uses gnuplot to turn these text files into nice
719 graphs. See \fBwrite_log_log\fR for behaviour of given filename. For this
720 option, the postfix is _bw.log.
723 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
724 filename is given with this option, the default filename of "jobname_type.log"
725 is used. Even if the filename is given, fio will still append the type of log.
727 .B disable_lat \fR=\fPbool
728 Disable measurements of total latency numbers. Useful only for cutting
729 back the number of calls to gettimeofday, as that does impact performance at
730 really high IOPS rates. Note that to really get rid of a large amount of these
731 calls, this option must be used with disable_slat and disable_bw as well.
733 .B disable_clat \fR=\fPbool
734 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
736 .B disable_slat \fR=\fPbool
737 Disable measurements of submission latency numbers. See \fBdisable_lat\fR.
739 .B disable_bw_measurement \fR=\fPbool
740 Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR.
742 .BI lockmem \fR=\fPint
743 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
744 simulate a smaller amount of memory.
746 .BI exec_prerun \fR=\fPstr
747 Before running the job, execute the specified command with \fBsystem\fR\|(3).
749 .BI exec_postrun \fR=\fPstr
750 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
752 .BI ioscheduler \fR=\fPstr
753 Attempt to switch the device hosting the file to the specified I/O scheduler.
755 .BI cpuload \fR=\fPint
756 If the job is a CPU cycle-eater, attempt to use the specified percentage of
759 .BI cpuchunks \fR=\fPint
760 If the job is a CPU cycle-eater, split the load into cycles of the
761 given time in milliseconds.
763 .BI disk_util \fR=\fPbool
764 Generate disk utilization statistics if the platform supports it. Default: true.
766 .BI gtod_reduce \fR=\fPbool
767 Enable all of the gettimeofday() reducing options (disable_clat, disable_slat,
768 disable_bw) plus reduce precision of the timeout somewhat to really shrink the
769 gettimeofday() call count. With this option enabled, we only do about 0.4% of
770 the gtod() calls we would have done if all time keeping was enabled.
772 .BI gtod_cpu \fR=\fPint
773 Sometimes it's cheaper to dedicate a single thread of execution to just getting
774 the current time. Fio (and databases, for instance) are very intensive on
775 gettimeofday() calls. With this option, you can set one CPU aside for doing
776 nothing but logging current time to a shared memory location. Then the other
777 threads/processes that run IO workloads need only copy that segment, instead of
778 entering the kernel with a gettimeofday() call. The CPU set aside for doing
779 these time calls will be excluded from other uses. Fio will manually clear it
780 from the CPU mask of other jobs.
782 .BI cgroup \fR=\fPstr
783 Add job to this control group. If it doesn't exist, it will be created.
784 The system must have a mounted cgroup blkio mount point for this to work. If
785 your system doesn't have it mounted, you can do so with:
787 # mount -t cgroup -o blkio none /cgroup
789 .BI cgroup_weight \fR=\fPint
790 Set the weight of the cgroup to this value. See the documentation that comes
791 with the kernel, allowed values are in the range of 100..1000.
793 .BI cgroup_nodelete \fR=\fPbool
794 Normally fio will delete the cgroups it has created after the job completion.
795 To override this behavior and to leave cgroups around after the job completion,
796 set cgroup_nodelete=1. This can be useful if one wants to inspect various
797 cgroup files after job completion. Default: false
800 Instead of running as the invoking user, set the user ID to this value before
801 the thread/process does any work.
804 Set group ID, see \fBuid\fR.
806 While running, \fBfio\fR will display the status of the created jobs. For
810 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
813 The characters in the first set of brackets denote the current status of each
814 threads. The possible values are:
820 Setup but not started.
826 Initialized, waiting.
829 Running, doing sequential reads.
832 Running, doing random reads.
835 Running, doing sequential writes.
838 Running, doing random writes.
841 Running, doing mixed sequential reads/writes.
844 Running, doing mixed random reads/writes.
847 Running, currently waiting for \fBfsync\fR\|(2).
850 Running, verifying written data.
853 Exited, not reaped by main thread.
856 Exited, thread reaped.
860 The second set of brackets shows the estimated completion percentage of
861 the current group. The third set shows the read and write I/O rate,
862 respectively. Finally, the estimated run time of the job is displayed.
864 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
865 for each thread, each group of threads, and each disk, in that order.
867 Per-thread statistics first show the threads client number, group-id, and
868 error code. The remaining figures are as follows:
872 Number of megabytes of I/O performed.
875 Average data rate (bandwidth).
881 Submission latency minimum, maximum, average and standard deviation. This is
882 the time it took to submit the I/O.
885 Completion latency minimum, maximum, average and standard deviation. This
886 is the time between submission and completion.
889 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
890 and standard deviation.
893 CPU usage statistics. Includes user and system time, number of context switches
894 this thread went through and number of major and minor page faults.
897 Distribution of I/O depths. Each depth includes everything less than (or equal)
898 to it, but greater than the previous depth.
901 Number of read/write requests issued, and number of short read/write requests.
904 Distribution of I/O completion latencies. The numbers follow the same pattern
908 The group statistics show:
913 Number of megabytes I/O performed.
916 Aggregate bandwidth of threads in the group.
919 Minimum average bandwidth a thread saw.
922 Maximum average bandwidth a thread saw.
925 Shortest runtime of threads in the group.
928 Longest runtime of threads in the group.
932 Finally, disk statistics are printed with reads first:
937 Number of I/Os performed by all groups.
940 Number of merges in the I/O scheduler.
943 Number of ticks we kept the disk busy.
946 Total time spent in the disk queue.
953 If the \fB\-\-minimal\fR option is given, the results will be printed in a
954 semicolon-delimited format suitable for scripted use. The fields are:
957 .B jobname, groupid, error
961 .B KB I/O, bandwidth \fR(KB/s)\fP, runtime \fR(ms)\fP
965 .B min, max, mean, standard deviation
969 .B min, max, mean, standard deviation
973 .B min, max, aggregate percentage of total, mean, standard deviation
979 .B KB I/O, bandwidth \fR(KB/s)\fP, runtime \fR(ms)\fP
983 .B min, max, mean, standard deviation
987 .B min, max, mean, standard deviation
991 .B min, max, aggregate percentage of total, mean, standard deviation
997 .B user, system, context switches, major page faults, minor page faults
1000 IO depth distribution:
1002 .B <=1, 2, 4, 8, 16, 32, >=64
1005 IO latency distribution (ms):
1007 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, >=2000
1014 was written by Jens Axboe <jens.axboe@oracle.com>,
1015 now Jens Axboe <jaxboe@fusionio.com>.
1017 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
1018 on documentation by Jens Axboe.
1019 .SH "REPORTING BUGS"
1020 Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>.
1023 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
1025 Sample jobfiles are available in the \fBexamples\fR directory.