Fio 1.19
[fio.git] / fio.1
CommitLineData
d60e92d1
AC
1.TH fio 1 "September 2007" "User Manual"
2.SH NAME
3fio \- flexible I/O tester
4.SH SYNOPSIS
5.B fio
6[\fIoptions\fR] [\fIjobfile\fR]...
7.SH DESCRIPTION
8.B fio
9is a tool that will spawn a number of threads or processes doing a
10particular type of I/O action as specified by the user.
11The typical use of fio is to write a job file matching the I/O load
12one wants to simulate.
13.SH OPTIONS
14.TP
15.BI \-\-output \fR=\fPfilename
16Write output to \fIfilename\fR.
17.TP
18.BI \-\-timeout \fR=\fPtimeout
19Limit run time to \fItimeout\fR seconds.
20.TP
21.B \-\-latency\-log
22Generate per-job latency logs.
23.TP
24.B \-\-bandwidth\-log
25Generate per-job bandwidth logs.
26.TP
27.B \-\-minimal
d1429b5c 28Print statistics in a terse, semicolon-delimited format.
d60e92d1
AC
29.TP
30.BI \-\-showcmd \fR=\fPjobfile
31Convert \fIjobfile\fR to a set of command-line options.
32.TP
33.B \-\-readonly
34Enable read-only safety checks.
35.TP
36.BI \-\-eta \fR=\fPwhen
37Specifies when real-time ETA estimate should be printed. \fIwhen\fR may
38be one of `always', `never' or `auto'.
39.TP
40.BI \-\-cmdhelp \fR=\fPcommand
41Print help information for \fIcommand\fR. May be `all' for all commands.
42.TP
9183788d
JA
43.BI \-\-debug \fR=\fPtype
44Enable verbose tracing of various fio actions. May be `all' for all types
45or individual types seperated by a comma (eg --debug=io,file). `help' will
46list all available tracing options.
47.TP
48Print help information for \fIcommand\fR. May be `all' for all commands.
49.TP
d60e92d1
AC
50.B \-\-help
51Display usage information and exit.
52.TP
53.B \-\-version
54Display version information and exit.
55.SH "JOB FILE FORMAT"
56Job files are in `ini' format. They consist of one or more
57job definitions, which begin with a job name in square brackets and
58extend to the next job name. The job name can be any ASCII string
59except `global', which has a special meaning. Following the job name is
60a sequence of zero or more parameters, one per line, that define the
61behavior of the job. Any line starting with a `;' or `#' character is
d1429b5c 62considered a comment and ignored.
d9956b64
AC
63.P
64If \fIjobfile\fR is specified as `-', the job file will be read from
65standard input.
d60e92d1
AC
66.SS "Global Section"
67The global section contains default parameters for jobs specified in the
68job file. A job is only affected by global sections residing above it,
69and there may be any number of global sections. Specific job definitions
70may override any parameter set in global sections.
71.SH "JOB PARAMETERS"
72.SS Types
73Some parameters may take arguments of a specific type. The types used are:
74.TP
75.I str
76String: a sequence of alphanumeric characters.
77.TP
78.I int
79Integer: a whole number, possibly negative. If prefixed with `0x', the value
80is assumed to be base 16 (hexadecimal).
81.TP
82.I siint
83SI integer: a whole number, possibly containing a suffix denoting the base unit
84of the value. Accepted suffixes are `k', 'M' and 'G', denoting kilo (1024),
85mega (1024*1024) and giga (1024*1024*1024) respectively.
86.TP
87.I bool
88Boolean: a true or false value. `0' denotes false, `1' denotes true.
89.TP
90.I irange
91Integer range: a range of integers specified in the format
d1429b5c
AC
92\fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
93\fIupper\fR may contain a suffix as described above. If an option allows two
94sets of ranges, they are separated with a `,' or `/' character. For example:
95`8\-8k/8M\-4G'.
d60e92d1
AC
96.SS "Parameter List"
97.TP
98.BI name \fR=\fPstr
d9956b64 99May be used to override the job name. On the command line, this parameter
d60e92d1
AC
100has the special purpose of signalling the start of a new job.
101.TP
102.BI description \fR=\fPstr
103Human-readable description of the job. It is printed when the job is run, but
104otherwise has no special purpose.
105.TP
106.BI directory \fR=\fPstr
107Prefix filenames with this directory. Used to place files in a location other
108than `./'.
109.TP
110.BI filename \fR=\fPstr
111.B fio
112normally makes up a file name based on the job name, thread number, and file
d1429b5c
AC
113number. If you want to share files between threads in a job or several jobs,
114specify a \fIfilename\fR for each of them to override the default. If the I/O
115engine used is `net', \fIfilename\fR is the host and port to connect to in the
116format \fIhost\fR/\fIport\fR. If the I/O engine is file-based, you can specify
117a number of files by separating the names with a `:' character. `\-' is a
118reserved name, meaning stdin or stdout, depending on the read/write direction
119set.
d60e92d1
AC
120.TP
121.BI opendir \fR=\fPstr
122Recursively open any files below directory \fIstr\fR.
123.TP
124.BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
125Type of I/O pattern. Accepted values are:
126.RS
127.RS
128.TP
129.B read
d1429b5c 130Sequential reads.
d60e92d1
AC
131.TP
132.B write
d1429b5c 133Sequential writes.
d60e92d1
AC
134.TP
135.B randread
d1429b5c 136Random reads.
d60e92d1
AC
137.TP
138.B randwrite
d1429b5c 139Random writes.
d60e92d1
AC
140.TP
141.B rw
d1429b5c 142Mixed sequential reads and writes.
d60e92d1
AC
143.TP
144.B randrw
d1429b5c 145Mixed random reads and writes.
d60e92d1
AC
146.RE
147.P
d1429b5c
AC
148For mixed I/O, the default split is 50/50. For random I/O, the number of I/Os
149to perform before getting a new offset can be specified by appending
150`:\fIint\fR' to the pattern type. The default is 1.
d60e92d1
AC
151.RE
152.TP
153.BI randrepeat \fR=\fPbool
154Seed the random number generator in a predictable way so results are repeatable
d1429b5c 155across runs. Default: true.
d60e92d1
AC
156.TP
157.BI fadvise_hint \fR=\fPbool
d1429b5c
AC
158Disable use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
159are likely to be issued. Default: true.
d60e92d1
AC
160.TP
161.BI size \fR=\fPsiint
162Total size of I/O for this job. \fBfio\fR will run until this many bytes have
163been transfered, unless limited by other options (\fBruntime\fR, for instance).
164Unless \fBnr_files\fR and \fBfilesize\fR options are given, this amount will be
165divided between the available files for the job.
166.TP
167.BI filesize \fR=\fPirange
168Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
d1429b5c
AC
169for files at random within the given range, limited to \fBsize\fR in total (if
170that is given). If \fBfilesize\fR is not specified, each created file is the
171same size.
d60e92d1 172.TP
9183788d 173.BI blocksize \fR=\fPsiint[,siint] "\fR,\fB bs" \fR=\fPsiint[,siint]
d60e92d1
AC
174Block size for I/O units. Default: 4k. Values for reads and writes can be
175specified seperately in the format \fIread\fR,\fIwrite\fR, either of
176which may be empty to leave that value at its default.
177.TP
9183788d 178.BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
d1429b5c
AC
179Specify a range of I/O block sizes. The issued I/O unit will always be a
180multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
9183788d
JA
181to both reads and writes if only one range is given, but can be specified
182seperately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k.
183Also (see \fBblocksize\fR).
184.TP
185.BI bssplit \fR=\fPstr
186This option allows even finer grained control of the block sizes issued,
187not just even splits between them. With this option, you can weight various
188block sizes for exact control of the issued IO for a job that has mixed
189block sizes. The format of the option is bssplit=blocksize/percentage,
190optionally adding as many definitions as needed seperated by a colon.
191Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
192blocks and 40% 32k blocks.
d60e92d1
AC
193.TP
194.B blocksize_unaligned\fR,\fP bs_unaligned
d1429b5c
AC
195If set, any size in \fBblocksize_range\fR may be used. This typically won't
196work with direct I/O, as that normally requires sector alignment.
d60e92d1
AC
197.TP
198.B zero_buffers
199Initialise buffers with all zeros. Default: fill buffers with random data.
200.TP
201.BI nrfiles \fR=\fPint
202Number of files to use for this job. Default: 1.
203.TP
204.BI openfiles \fR=\fPint
205Number of files to keep open at the same time. Default: \fBnrfiles\fR.
206.TP
207.BI file_service_type \fR=\fPstr
208Defines how files to service are selected. The following types are defined:
209.RS
210.RS
211.TP
212.B random
213Choose a file at random
214.TP
215.B roundrobin
216Round robin over open files (default).
217.RE
218.P
219The number of I/Os to issue before switching a new file can be specified by
220appending `:\fIint\fR' to the service type.
221.RE
222.TP
223.BI ioengine \fR=\fPstr
224Defines how the job issues I/O. The following types are defined:
225.RS
226.RS
227.TP
228.B sync
229Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
230position the I/O location.
231.TP
a31041ea 232.B psync
233Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O.
234.TP
9183788d
JA
235.B vsync
236Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by
237coalescing adjacents IOs into a single submission.
238.TP
d60e92d1
AC
239.B libaio
240Linux native asynchronous I/O.
241.TP
242.B posixaio
243glibc POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
244.TP
245.B mmap
d1429b5c
AC
246File is memory mapped with \fImmap\fR\|(2) and data copied using
247\fImemcpy\fR\|(3).
d60e92d1
AC
248.TP
249.B splice
d1429b5c
AC
250\fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
251transfer data from user-space to the kernel.
d60e92d1
AC
252.TP
253.B syslet-rw
254Use the syslet system calls to make regular read/write asynchronous.
255.TP
256.B sg
257SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
d1429b5c
AC
258the target is an sg character device, we use \fIread\fR\|(2) and
259\fIwrite\fR\|(2) for asynchronous I/O.
d60e92d1
AC
260.TP
261.B null
262Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
263itself and for debugging and testing purposes.
264.TP
265.B net
266Transfer over the network. \fBfilename\fR must be set appropriately to
267`\fIhost\fR/\fIport\fR' regardless of data direction. If receiving, only the
268\fIport\fR argument is used.
269.TP
270.B netsplice
271Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
272and send/receive.
273.TP
53aec0a4 274.B cpuio
d60e92d1
AC
275Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
276\fBcpucycles\fR parameters.
277.TP
278.B guasi
279The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
280approach to asycnronous I/O.
d1429b5c
AC
281.br
282See <http://www.xmailserver.org/guasi\-lib.html>.
d60e92d1
AC
283.TP
284.B external
285Loads an external I/O engine object file. Append the engine filename as
286`:\fIenginepath\fR'.
287.RE
288.RE
289.TP
290.BI iodepth \fR=\fPint
291Number of I/O units to keep in flight against the file. Default: 1.
292.TP
293.BI iodepth_batch \fR=\fPint
294Number of I/Os to submit at once. Default: \fBiodepth\fR.
295.TP
296.BI iodepth_low \fR=\fPint
297Low watermark indicating when to start filling the queue again. Default:
298\fBiodepth\fR.
299.TP
300.BI direct \fR=\fPbool
301If true, use non-buffered I/O (usually O_DIRECT). Default: false.
302.TP
303.BI buffered \fR=\fPbool
304If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
305Default: true.
306.TP
307.BI offset \fR=\fPsiint
308Offset in the file to start I/O. Data before the offset will not be touched.
309.TP
310.BI fsync \fR=\fPint
d1429b5c
AC
311How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
3120, don't sync. Default: 0.
d60e92d1
AC
313.TP
314.BI overwrite \fR=\fPbool
d1429b5c 315If writing, setup the file first and do overwrites. Default: false.
d60e92d1
AC
316.TP
317.BI end_fsync \fR=\fPbool
d1429b5c 318Sync file contents when job exits. Default: false.
d60e92d1
AC
319.TP
320.BI fsync_on_close \fR=\fPbool
321If true, sync file contents on close. This differs from \fBend_fsync\fR in that
d1429b5c 322it will happen on every close, not just at the end of the job. Default: false.
d60e92d1
AC
323.TP
324.BI rwmixcycle \fR=\fPint
325How many milliseconds before switching between reads and writes for a mixed
326workload. Default: 500ms.
327.TP
328.BI rwmixread \fR=\fPint
329Percentage of a mixed workload that should be reads. Default: 50.
330.TP
331.BI rwmixwrite \fR=\fPint
d1429b5c 332Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
d60e92d1 333\fBwrmixwrite\fR are given and do not sum to 100%, the latter of the two
d1429b5c 334overrides the first. Default: 50.
d60e92d1
AC
335.TP
336.B norandommap
337Normally \fBfio\fR will cover every block of the file when doing random I/O. If
338this parameter is given, a new offset will be chosen without looking at past
339I/O history. This parameter is mutually exclusive with \fBverify\fR.
340.TP
341.BI nice \fR=\fPint
342Run job with given nice value. See \fInice\fR\|(2).
343.TP
344.BI prio \fR=\fPint
345Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
346\fIionice\fR\|(1).
347.TP
348.BI prioclass \fR=\fPint
349Set I/O priority class. See \fIionice\fR\|(1).
350.TP
351.BI thinktime \fR=\fPint
352Stall job for given number of microseconds between issuing I/Os.
353.TP
354.BI thinktime_spin \fR=\fPint
355Pretend to spend CPU time for given number of microseconds, sleeping the rest
356of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
357.TP
358.BI thinktime_blocks \fR=\fPint
359Number of blocks to issue before waiting \fBthinktime\fR microseconds.
360Default: 1.
361.TP
362.BI rate \fR=\fPint
363Cap bandwidth used by this job to this number of KiB/s.
364.TP
365.BI ratemin \fR=\fPint
366Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
367Failing to meet this requirement will cause the job to exit.
368.TP
369.BI rate_iops \fR=\fPint
370Cap the bandwidth to this number of IOPS. If \fBblocksize\fR is a range, the
371smallest block size is used as the metric.
372.TP
373.BI rate_iops_min \fR=\fPint
374If this rate of I/O is not met, the job will exit.
375.TP
376.BI ratecycle \fR=\fPint
377Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
378milliseconds. Default: 1000ms.
379.TP
380.BI cpumask \fR=\fPint
381Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
382may run on. See \fBsched_setaffinity\fR\|(2).
383.TP
384.BI cpus_allowed \fR=\fPstr
385Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
386.TP
387.BI startdelay \fR=\fPint
388Delay start of job for the specified number of seconds.
389.TP
390.BI runtime \fR=\fPint
391Terminate processing after the specified number of seconds.
392.TP
393.B time_based
394If given, run for the specified \fBruntime\fR duration even if the files are
395completely read or written. The same workload will be repeated as many times
396as \fBruntime\fR allows.
397.TP
398.BI invalidate \fR=\fPbool
399Invalidate buffer-cache for the file prior to starting I/O. Default: true.
400.TP
401.BI sync \fR=\fPbool
402Use synchronous I/O for buffered writes. For the majority of I/O engines,
d1429b5c 403this means using O_SYNC. Default: false.
d60e92d1
AC
404.TP
405.BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
406Allocation method for I/O unit buffer. Allowed values are:
407.RS
408.RS
409.TP
410.B malloc
411Allocate memory with \fImalloc\fR\|(3).
412.TP
413.B shm
414Use shared memory buffers allocated through \fIshmget\fR\|(2).
415.TP
416.B shmhuge
417Same as \fBshm\fR, but use huge pages as backing.
418.TP
419.B mmap
420Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
421is given after the option in the format `:\fIfile\fR'.
422.TP
423.B mmaphuge
424Same as \fBmmap\fR, but use huge files as backing.
425.RE
426.P
427The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
428job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
429the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
430have hugetlbfs mounted, and \fIfile\fR must point there.
431.RE
432.TP
d1429b5c 433.BI hugepage\-size \fR=\fPsiint
d60e92d1
AC
434Defines the size of a huge page. Must be at least equal to the system setting.
435Should be a multiple of 1MiB. Default: 4MiB.
436.TP
437.B exitall
438Terminate all jobs when one finishes. Default: wait for each job to finish.
439.TP
440.BI bwavgtime \fR=\fPint
441Average bandwidth calculations over the given time in milliseconds. Default:
442500ms.
443.TP
444.BI create_serialize \fR=\fPbool
d1429b5c 445If true, serialize file creation for the jobs. Default: true.
d60e92d1
AC
446.TP
447.BI create_fsync \fR=\fPbool
448\fIfsync\fR\|(2) data file after creation. Default: true.
449.TP
450.BI unlink \fR=\fPbool
451Unlink job files when done. Default: false.
452.TP
453.BI loops \fR=\fPint
454Specifies the number of iterations (runs of the same workload) of this job.
455Default: 1.
456.TP
457.BI do_verify \fR=\fPbool
458Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
459Default: true.
460.TP
461.BI verify \fR=\fPstr
462Method of verifying file contents after each iteration of the job. Allowed
463values are:
464.RS
465.RS
466.TP
467.B md5 crc16 crc32 crc64 crc7 sha256 sha512
468Store appropriate checksum in the header of each block.
469.TP
470.B meta
471Write extra information about each I/O (timestamp, block number, etc.). The
472block number is verified.
473.TP
474.B pattern
475Fill I/O buffers with a specific pattern that is used to verify. The pattern is
476specified by appending `:\fIint\fR' to the parameter. \fIint\fR cannot be larger
477than 32-bits.
478.TP
479.B null
480Pretend to verify. Used for testing internals.
481.RE
482.RE
483.TP
484.BI verify_sort \fR=\fPbool
485If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
486read them back in a sorted manner. Default: true.
487.TP
488.BI verify_offset \fR=\fPsiint
489Swap the verification header with data somewhere else in the block before
d1429b5c 490writing. It is swapped back before verifying.
d60e92d1
AC
491.TP
492.BI verify_interval \fR=\fPsiint
493Write the verification header for this number of bytes, which should divide
494\fBblocksize\fR. Default: \fBblocksize\fR.
495.TP
496.BI verify_fatal \fR=\fPbool
497If true, exit the job on the first observed verification failure. Default:
498false.
499.TP
500.B stonewall
d1429b5c 501Wait for preceeding jobs in the job file to exit before starting this one.
d60e92d1
AC
502\fBstonewall\fR implies \fBnew_group\fR.
503.TP
504.B new_group
505Start a new reporting group. If not given, all jobs in a file will be part
506of the same reporting group, unless separated by a stonewall.
507.TP
508.BI numjobs \fR=\fPint
509Number of clones (processes/threads performing the same workload) of this job.
510Default: 1.
511.TP
512.B group_reporting
513If set, display per-group reports instead of per-job when \fBnumjobs\fR is
514specified.
515.TP
516.B thread
517Use threads created with \fBpthread_create\fR\|(3) instead of processes created
518with \fBfork\fR\|(2).
519.TP
520.BI zonesize \fR=\fPsiint
521Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
522.TP
523.BI zoneskip \fR=\fPsiint
d1429b5c 524Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
d60e92d1
AC
525read.
526.TP
527.BI write_iolog \fR=\fPstr
528Write the issued I/O patterns to the specified file.
529.TP
530.BI read_iolog \fR=\fPstr
531Replay the I/O patterns contained in the specified file generated by
532\fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
533.TP
534.B write_bw_log
535If given, write bandwidth logs of the jobs in this file.
536.TP
537.B write_lat_log
538Same as \fBwrite_bw_log\fR, but writes I/O completion latencies.
539.TP
540.BI lockmem \fR=\fPsiint
541Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
542simulate a smaller amount of memory.
543.TP
544.BI exec_prerun \fR=\fPstr
545Before running the job, execute the specified command with \fBsystem\fR\|(3).
546.TP
547.BI exec_postrun \fR=\fPstr
548Same as \fBexec_prerun\fR, but the command is executed after the job completes.
549.TP
550.BI ioscheduler \fR=\fPstr
551Attempt to switch the device hosting the file to the specified I/O scheduler.
552.TP
553.BI cpuload \fR=\fPint
554If the job is a CPU cycle-eater, attempt to use the specified percentage of
555CPU cycles.
556.TP
557.BI cpuchunks \fR=\fPint
558If the job is a CPU cycle-eater, split the load into cycles of the
559given time in milliseconds.
560.TP
561.BI disk_util \fR=\fPbool
d1429b5c 562Generate disk utilization statistics if the platform supports it. Default: true.
d60e92d1 563.SH OUTPUT
d1429b5c
AC
564While running, \fBfio\fR will display the status of the created jobs. For
565example:
d60e92d1 566.RS
d1429b5c 567.P
d60e92d1
AC
568Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
569.RE
570.P
d1429b5c
AC
571The characters in the first set of brackets denote the current status of each
572threads. The possible values are:
573.P
574.PD 0
d60e92d1
AC
575.RS
576.TP
577.B P
578Setup but not started.
579.TP
580.B C
581Thread created.
582.TP
583.B I
584Initialized, waiting.
585.TP
586.B R
587Running, doing sequential reads.
588.TP
589.B r
590Running, doing random reads.
591.TP
592.B W
593Running, doing sequential writes.
594.TP
595.B w
596Running, doing random writes.
597.TP
598.B M
599Running, doing mixed sequential reads/writes.
600.TP
601.B m
602Running, doing mixed random reads/writes.
603.TP
604.B F
605Running, currently waiting for \fBfsync\fR\|(2).
606.TP
607.B V
608Running, verifying written data.
609.TP
610.B E
611Exited, not reaped by main thread.
612.TP
613.B \-
614Exited, thread reaped.
615.RE
d1429b5c 616.PD
d60e92d1
AC
617.P
618The second set of brackets shows the estimated completion percentage of
619the current group. The third set shows the read and write I/O rate,
620respectively. Finally, the estimated run time of the job is displayed.
621.P
622When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
623for each thread, each group of threads, and each disk, in that order.
624.P
625Per-thread statistics first show the threads client number, group-id, and
626error code. The remaining figures are as follows:
627.RS
d60e92d1
AC
628.TP
629.B io
630Number of megabytes of I/O performed.
631.TP
632.B bw
633Average data rate (bandwidth).
634.TP
635.B runt
636Threads run time.
637.TP
638.B slat
639Submission latency minimum, maximum, average and standard deviation. This is
640the time it took to submit the I/O.
641.TP
642.B clat
643Completion latency minimum, maximum, average and standard deviation. This
644is the time between submission and completion.
645.TP
646.B bw
647Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
648and standard deviation.
649.TP
650.B cpu
651CPU usage statistics. Includes user and system time, number of context switches
652this thread went through and number of major and minor page faults.
653.TP
654.B IO depths
655Distribution of I/O depths. Each depth includes everything less than (or equal)
656to it, but greater than the previous depth.
657.TP
658.B IO issued
659Number of read/write requests issued, and number of short read/write requests.
660.TP
661.B IO latencies
662Distribution of I/O completion latencies. The numbers follow the same pattern
663as \fBIO depths\fR.
664.RE
d60e92d1
AC
665.P
666The group statistics show:
d1429b5c 667.PD 0
d60e92d1
AC
668.RS
669.TP
670.B io
671Number of megabytes I/O performed.
672.TP
673.B aggrb
674Aggregate bandwidth of threads in the group.
675.TP
676.B minb
677Minimum average bandwidth a thread saw.
678.TP
679.B maxb
680Maximum average bandwidth a thread saw.
681.TP
682.B mint
d1429b5c 683Shortest runtime of threads in the group.
d60e92d1
AC
684.TP
685.B maxt
686Longest runtime of threads in the group.
687.RE
d1429b5c 688.PD
d60e92d1
AC
689.P
690Finally, disk statistics are printed with reads first:
d1429b5c 691.PD 0
d60e92d1
AC
692.RS
693.TP
694.B ios
695Number of I/Os performed by all groups.
696.TP
697.B merge
698Number of merges in the I/O scheduler.
699.TP
700.B ticks
701Number of ticks we kept the disk busy.
702.TP
703.B io_queue
704Total time spent in the disk queue.
705.TP
706.B util
707Disk utilization.
708.RE
d1429b5c 709.PD
d60e92d1
AC
710.SH TERSE OUTPUT
711If the \fB\-\-minimal\fR option is given, the results will be printed in a
712semicolon-delimited format suitable for scripted use. The fields are:
713.P
714.RS
715.B jobname, groupid, error
716.P
717Read status:
718.RS
719.B KiB I/O, bandwidth \fR(KiB/s)\fP, runtime \fR(ms)\fP
720.P
721Submission latency:
722.RS
723.B min, max, mean, standard deviation
724.RE
725Completion latency:
726.RS
727.B min, max, mean, standard deviation
728.RE
729Bandwidth:
730.RS
731.B min, max, aggregate percentage of total, mean, standard deviation
732.RE
733.RE
734.P
735Write status:
736.RS
737.B KiB I/O, bandwidth \fR(KiB/s)\fP, runtime \fR(ms)\fP
738.P
739Submission latency:
740.RS
741.B min, max, mean, standard deviation
742.RE
743Completion latency:
744.RS
745.B min, max, mean, standard deviation
746.RE
747Bandwidth:
748.RS
749.B min, max, aggregate percentage of total, mean, standard deviation
750.RE
751.RE
752.P
d1429b5c 753CPU usage:
d60e92d1
AC
754.RS
755.B user, system, context switches
756.RE
757.P
758IO depth distribution:
759.RS
760.B <=1, 2, 4, 8, 16, 32, >=64
761.RE
762.P
763IO latency distribution (ms):
764.RS
765.B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, >=2000
766.RE
767.P
768.B text description
769.RE
770.SH AUTHORS
771.B fio
772was written by Jens Axboe <jens.axboe@oracle.com>.
d1429b5c
AC
773.br
774This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
d60e92d1
AC
775on documentation by Jens Axboe.
776.SH "REPORTING BUGS"
d1429b5c
AC
777Report bugs to the \fBfio\fR mailing list <fio-devel@kernel.dk>.
778See \fBREADME\fR.
d60e92d1 779.SH "SEE ALSO"
d1429b5c
AC
780For further documentation see \fBHOWTO\fR and \fBREADME\fR.
781.br
782Sample jobfiles are available in the \fBexamples\fR directory.
d60e92d1 783