Fix crash with multiple files
[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
c0a5d35e
AC
40.BI \-\-section \fR=\fPsec
41Only run section \fIsec\fR from job file.
42.TP
d60e92d1
AC
43.BI \-\-cmdhelp \fR=\fPcommand
44Print help information for \fIcommand\fR. May be `all' for all commands.
45.TP
9183788d
JA
46.BI \-\-debug \fR=\fPtype
47Enable verbose tracing of various fio actions. May be `all' for all types
48or individual types seperated by a comma (eg --debug=io,file). `help' will
49list all available tracing options.
50.TP
d60e92d1
AC
51.B \-\-help
52Display usage information and exit.
53.TP
54.B \-\-version
55Display version information and exit.
56.SH "JOB FILE FORMAT"
57Job files are in `ini' format. They consist of one or more
58job definitions, which begin with a job name in square brackets and
59extend to the next job name. The job name can be any ASCII string
60except `global', which has a special meaning. Following the job name is
61a sequence of zero or more parameters, one per line, that define the
62behavior of the job. Any line starting with a `;' or `#' character is
d1429b5c 63considered a comment and ignored.
d9956b64
AC
64.P
65If \fIjobfile\fR is specified as `-', the job file will be read from
66standard input.
d60e92d1
AC
67.SS "Global Section"
68The global section contains default parameters for jobs specified in the
69job file. A job is only affected by global sections residing above it,
70and there may be any number of global sections. Specific job definitions
71may override any parameter set in global sections.
72.SH "JOB PARAMETERS"
73.SS Types
74Some parameters may take arguments of a specific type. The types used are:
75.TP
76.I str
77String: a sequence of alphanumeric characters.
78.TP
79.I int
d60e92d1
AC
80SI integer: a whole number, possibly containing a suffix denoting the base unit
81of the value. Accepted suffixes are `k', 'M' and 'G', denoting kilo (1024),
f7fa2653
JA
82mega (1024*1024) and giga (1024*1024*1024) respectively. If prefixed with '0x',
83the value is assumed to be base 16 (hexadecimal).
d60e92d1
AC
84.TP
85.I bool
86Boolean: a true or false value. `0' denotes false, `1' denotes true.
87.TP
88.I irange
89Integer range: a range of integers specified in the format
d1429b5c
AC
90\fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
91\fIupper\fR may contain a suffix as described above. If an option allows two
92sets of ranges, they are separated with a `,' or `/' character. For example:
93`8\-8k/8M\-4G'.
d60e92d1
AC
94.SS "Parameter List"
95.TP
96.BI name \fR=\fPstr
d9956b64 97May be used to override the job name. On the command line, this parameter
d60e92d1
AC
98has the special purpose of signalling the start of a new job.
99.TP
100.BI description \fR=\fPstr
101Human-readable description of the job. It is printed when the job is run, but
102otherwise has no special purpose.
103.TP
104.BI directory \fR=\fPstr
105Prefix filenames with this directory. Used to place files in a location other
106than `./'.
107.TP
108.BI filename \fR=\fPstr
109.B fio
110normally makes up a file name based on the job name, thread number, and file
d1429b5c
AC
111number. If you want to share files between threads in a job or several jobs,
112specify a \fIfilename\fR for each of them to override the default. If the I/O
113engine used is `net', \fIfilename\fR is the host and port to connect to in the
114format \fIhost\fR/\fIport\fR. If the I/O engine is file-based, you can specify
115a number of files by separating the names with a `:' character. `\-' is a
116reserved name, meaning stdin or stdout, depending on the read/write direction
117set.
d60e92d1
AC
118.TP
119.BI opendir \fR=\fPstr
120Recursively open any files below directory \fIstr\fR.
121.TP
122.BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
123Type of I/O pattern. Accepted values are:
124.RS
125.RS
126.TP
127.B read
d1429b5c 128Sequential reads.
d60e92d1
AC
129.TP
130.B write
d1429b5c 131Sequential writes.
d60e92d1
AC
132.TP
133.B randread
d1429b5c 134Random reads.
d60e92d1
AC
135.TP
136.B randwrite
d1429b5c 137Random writes.
d60e92d1
AC
138.TP
139.B rw
d1429b5c 140Mixed sequential reads and writes.
d60e92d1
AC
141.TP
142.B randrw
d1429b5c 143Mixed random reads and writes.
d60e92d1
AC
144.RE
145.P
d1429b5c
AC
146For mixed I/O, the default split is 50/50. For random I/O, the number of I/Os
147to perform before getting a new offset can be specified by appending
148`:\fIint\fR' to the pattern type. The default is 1.
d60e92d1
AC
149.RE
150.TP
151.BI randrepeat \fR=\fPbool
152Seed the random number generator in a predictable way so results are repeatable
d1429b5c 153across runs. Default: true.
d60e92d1
AC
154.TP
155.BI fadvise_hint \fR=\fPbool
d1429b5c
AC
156Disable use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
157are likely to be issued. Default: true.
d60e92d1 158.TP
f7fa2653 159.BI size \fR=\fPint
d60e92d1
AC
160Total size of I/O for this job. \fBfio\fR will run until this many bytes have
161been transfered, unless limited by other options (\fBruntime\fR, for instance).
162Unless \fBnr_files\fR and \fBfilesize\fR options are given, this amount will be
163divided between the available files for the job.
164.TP
165.BI filesize \fR=\fPirange
166Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
d1429b5c
AC
167for files at random within the given range, limited to \fBsize\fR in total (if
168that is given). If \fBfilesize\fR is not specified, each created file is the
169same size.
d60e92d1 170.TP
f7fa2653 171.BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
d60e92d1
AC
172Block size for I/O units. Default: 4k. Values for reads and writes can be
173specified seperately in the format \fIread\fR,\fIwrite\fR, either of
174which may be empty to leave that value at its default.
175.TP
9183788d 176.BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
d1429b5c
AC
177Specify a range of I/O block sizes. The issued I/O unit will always be a
178multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
9183788d
JA
179to both reads and writes if only one range is given, but can be specified
180seperately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k.
181Also (see \fBblocksize\fR).
182.TP
183.BI bssplit \fR=\fPstr
184This option allows even finer grained control of the block sizes issued,
185not just even splits between them. With this option, you can weight various
186block sizes for exact control of the issued IO for a job that has mixed
187block sizes. The format of the option is bssplit=blocksize/percentage,
188optionally adding as many definitions as needed seperated by a colon.
189Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
190blocks and 40% 32k blocks.
d60e92d1
AC
191.TP
192.B blocksize_unaligned\fR,\fP bs_unaligned
d1429b5c
AC
193If set, any size in \fBblocksize_range\fR may be used. This typically won't
194work with direct I/O, as that normally requires sector alignment.
d60e92d1 195.TP
2b7a01d0
JA
196.BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
197At what boundary to align random IO offsets. Defaults to the same as
198'blocksize' the minimum blocksize given. Minimum alignment is typically 512b
199for using direct IO, though it usually depends on the hardware block size.
200This option is mutually exclusive with using a random map for files, so it
201will turn off that option.
43602667 202.TP
d60e92d1
AC
203.B zero_buffers
204Initialise buffers with all zeros. Default: fill buffers with random data.
205.TP
901bb994
JA
206.B refill_buffers
207If this option is given, fio will refill the IO buffers on every submit. The
208default is to only fill it at init time and reuse that data. Only makes sense
209if zero_buffers isn't specified, naturally. If data verification is enabled,
210refill_buffers is also automatically enabled.
211.TP
d60e92d1
AC
212.BI nrfiles \fR=\fPint
213Number of files to use for this job. Default: 1.
214.TP
215.BI openfiles \fR=\fPint
216Number of files to keep open at the same time. Default: \fBnrfiles\fR.
217.TP
218.BI file_service_type \fR=\fPstr
219Defines how files to service are selected. The following types are defined:
220.RS
221.RS
222.TP
223.B random
224Choose a file at random
225.TP
226.B roundrobin
227Round robin over open files (default).
6b7f6851
JA
228.B sequential
229Do each file in the set sequentially.
d60e92d1
AC
230.RE
231.P
232The number of I/Os to issue before switching a new file can be specified by
233appending `:\fIint\fR' to the service type.
234.RE
235.TP
236.BI ioengine \fR=\fPstr
237Defines how the job issues I/O. The following types are defined:
238.RS
239.RS
240.TP
241.B sync
242Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
243position the I/O location.
244.TP
a31041ea 245.B psync
246Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O.
247.TP
9183788d
JA
248.B vsync
249Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by
250coalescing adjacents IOs into a single submission.
251.TP
d60e92d1
AC
252.B libaio
253Linux native asynchronous I/O.
254.TP
255.B posixaio
256glibc POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
257.TP
258.B mmap
d1429b5c
AC
259File is memory mapped with \fImmap\fR\|(2) and data copied using
260\fImemcpy\fR\|(3).
d60e92d1
AC
261.TP
262.B splice
d1429b5c
AC
263\fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
264transfer data from user-space to the kernel.
d60e92d1
AC
265.TP
266.B syslet-rw
267Use the syslet system calls to make regular read/write asynchronous.
268.TP
269.B sg
270SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
d1429b5c
AC
271the target is an sg character device, we use \fIread\fR\|(2) and
272\fIwrite\fR\|(2) for asynchronous I/O.
d60e92d1
AC
273.TP
274.B null
275Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
276itself and for debugging and testing purposes.
277.TP
278.B net
279Transfer over the network. \fBfilename\fR must be set appropriately to
280`\fIhost\fR/\fIport\fR' regardless of data direction. If receiving, only the
281\fIport\fR argument is used.
282.TP
283.B netsplice
284Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
285and send/receive.
286.TP
53aec0a4 287.B cpuio
d60e92d1
AC
288Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
289\fBcpucycles\fR parameters.
290.TP
291.B guasi
292The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
293approach to asycnronous I/O.
d1429b5c
AC
294.br
295See <http://www.xmailserver.org/guasi\-lib.html>.
d60e92d1
AC
296.TP
297.B external
298Loads an external I/O engine object file. Append the engine filename as
299`:\fIenginepath\fR'.
300.RE
301.RE
302.TP
303.BI iodepth \fR=\fPint
304Number of I/O units to keep in flight against the file. Default: 1.
305.TP
306.BI iodepth_batch \fR=\fPint
307Number of I/Os to submit at once. Default: \fBiodepth\fR.
308.TP
309.BI iodepth_low \fR=\fPint
310Low watermark indicating when to start filling the queue again. Default:
311\fBiodepth\fR.
312.TP
313.BI direct \fR=\fPbool
314If true, use non-buffered I/O (usually O_DIRECT). Default: false.
315.TP
316.BI buffered \fR=\fPbool
317If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
318Default: true.
319.TP
f7fa2653 320.BI offset \fR=\fPint
d60e92d1
AC
321Offset in the file to start I/O. Data before the offset will not be touched.
322.TP
323.BI fsync \fR=\fPint
d1429b5c
AC
324How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
3250, don't sync. Default: 0.
d60e92d1
AC
326.TP
327.BI overwrite \fR=\fPbool
d1429b5c 328If writing, setup the file first and do overwrites. Default: false.
d60e92d1
AC
329.TP
330.BI end_fsync \fR=\fPbool
d1429b5c 331Sync file contents when job exits. Default: false.
d60e92d1
AC
332.TP
333.BI fsync_on_close \fR=\fPbool
334If true, sync file contents on close. This differs from \fBend_fsync\fR in that
d1429b5c 335it will happen on every close, not just at the end of the job. Default: false.
d60e92d1
AC
336.TP
337.BI rwmixcycle \fR=\fPint
338How many milliseconds before switching between reads and writes for a mixed
339workload. Default: 500ms.
340.TP
341.BI rwmixread \fR=\fPint
342Percentage of a mixed workload that should be reads. Default: 50.
343.TP
344.BI rwmixwrite \fR=\fPint
d1429b5c 345Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
d60e92d1 346\fBwrmixwrite\fR are given and do not sum to 100%, the latter of the two
d1429b5c 347overrides the first. Default: 50.
d60e92d1
AC
348.TP
349.B norandommap
350Normally \fBfio\fR will cover every block of the file when doing random I/O. If
351this parameter is given, a new offset will be chosen without looking at past
352I/O history. This parameter is mutually exclusive with \fBverify\fR.
353.TP
354.BI nice \fR=\fPint
355Run job with given nice value. See \fInice\fR\|(2).
356.TP
357.BI prio \fR=\fPint
358Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
359\fIionice\fR\|(1).
360.TP
361.BI prioclass \fR=\fPint
362Set I/O priority class. See \fIionice\fR\|(1).
363.TP
364.BI thinktime \fR=\fPint
365Stall job for given number of microseconds between issuing I/Os.
366.TP
367.BI thinktime_spin \fR=\fPint
368Pretend to spend CPU time for given number of microseconds, sleeping the rest
369of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
370.TP
371.BI thinktime_blocks \fR=\fPint
372Number of blocks to issue before waiting \fBthinktime\fR microseconds.
373Default: 1.
374.TP
375.BI rate \fR=\fPint
376Cap bandwidth used by this job to this number of KiB/s.
377.TP
378.BI ratemin \fR=\fPint
379Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
380Failing to meet this requirement will cause the job to exit.
381.TP
382.BI rate_iops \fR=\fPint
383Cap the bandwidth to this number of IOPS. If \fBblocksize\fR is a range, the
384smallest block size is used as the metric.
385.TP
386.BI rate_iops_min \fR=\fPint
387If this rate of I/O is not met, the job will exit.
388.TP
389.BI ratecycle \fR=\fPint
390Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
391milliseconds. Default: 1000ms.
392.TP
393.BI cpumask \fR=\fPint
394Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
395may run on. See \fBsched_setaffinity\fR\|(2).
396.TP
397.BI cpus_allowed \fR=\fPstr
398Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
399.TP
400.BI startdelay \fR=\fPint
401Delay start of job for the specified number of seconds.
402.TP
403.BI runtime \fR=\fPint
404Terminate processing after the specified number of seconds.
405.TP
406.B time_based
407If given, run for the specified \fBruntime\fR duration even if the files are
408completely read or written. The same workload will be repeated as many times
409as \fBruntime\fR allows.
410.TP
901bb994
JA
411.BI ramp_time \fR=\fPint
412If set, fio will run the specified workload for this amount of time before
413logging any performance numbers. Useful for letting performance settle before
414logging results, thus minimizing the runtime required for stable results. Note
415that the ramp_time is considered lead in time for a job, thus it will increase
416the total runtime if a special timeout or runtime is specified.
417.TP
d60e92d1
AC
418.BI invalidate \fR=\fPbool
419Invalidate buffer-cache for the file prior to starting I/O. Default: true.
420.TP
421.BI sync \fR=\fPbool
422Use synchronous I/O for buffered writes. For the majority of I/O engines,
d1429b5c 423this means using O_SYNC. Default: false.
d60e92d1
AC
424.TP
425.BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
426Allocation method for I/O unit buffer. Allowed values are:
427.RS
428.RS
429.TP
430.B malloc
431Allocate memory with \fImalloc\fR\|(3).
432.TP
433.B shm
434Use shared memory buffers allocated through \fIshmget\fR\|(2).
435.TP
436.B shmhuge
437Same as \fBshm\fR, but use huge pages as backing.
438.TP
439.B mmap
440Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
441is given after the option in the format `:\fIfile\fR'.
442.TP
443.B mmaphuge
444Same as \fBmmap\fR, but use huge files as backing.
445.RE
446.P
447The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
448job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
449the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
450have hugetlbfs mounted, and \fIfile\fR must point there.
451.RE
452.TP
f7fa2653 453.BI hugepage\-size \fR=\fPint
d60e92d1
AC
454Defines the size of a huge page. Must be at least equal to the system setting.
455Should be a multiple of 1MiB. Default: 4MiB.
456.TP
457.B exitall
458Terminate all jobs when one finishes. Default: wait for each job to finish.
459.TP
460.BI bwavgtime \fR=\fPint
461Average bandwidth calculations over the given time in milliseconds. Default:
462500ms.
463.TP
464.BI create_serialize \fR=\fPbool
d1429b5c 465If true, serialize file creation for the jobs. Default: true.
d60e92d1
AC
466.TP
467.BI create_fsync \fR=\fPbool
468\fIfsync\fR\|(2) data file after creation. Default: true.
469.TP
6b7f6851
JA
470.BI create_on_open \fR=\fPbool
471If true, the files are not created until they are opened for IO by the job.
472.TP
d60e92d1
AC
473.BI unlink \fR=\fPbool
474Unlink job files when done. Default: false.
475.TP
476.BI loops \fR=\fPint
477Specifies the number of iterations (runs of the same workload) of this job.
478Default: 1.
479.TP
480.BI do_verify \fR=\fPbool
481Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
482Default: true.
483.TP
484.BI verify \fR=\fPstr
485Method of verifying file contents after each iteration of the job. Allowed
486values are:
487.RS
488.RS
489.TP
490.B md5 crc16 crc32 crc64 crc7 sha256 sha512
491Store appropriate checksum in the header of each block.
492.TP
493.B meta
494Write extra information about each I/O (timestamp, block number, etc.). The
495block number is verified.
496.TP
497.B pattern
498Fill I/O buffers with a specific pattern that is used to verify. The pattern is
499specified by appending `:\fIint\fR' to the parameter. \fIint\fR cannot be larger
500than 32-bits.
501.TP
502.B null
503Pretend to verify. Used for testing internals.
504.RE
505.RE
506.TP
507.BI verify_sort \fR=\fPbool
508If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
509read them back in a sorted manner. Default: true.
510.TP
f7fa2653 511.BI verify_offset \fR=\fPint
d60e92d1 512Swap the verification header with data somewhere else in the block before
d1429b5c 513writing. It is swapped back before verifying.
d60e92d1 514.TP
f7fa2653 515.BI verify_interval \fR=\fPint
d60e92d1
AC
516Write the verification header for this number of bytes, which should divide
517\fBblocksize\fR. Default: \fBblocksize\fR.
518.TP
519.BI verify_fatal \fR=\fPbool
520If true, exit the job on the first observed verification failure. Default:
521false.
522.TP
523.B stonewall
d1429b5c 524Wait for preceeding jobs in the job file to exit before starting this one.
d60e92d1
AC
525\fBstonewall\fR implies \fBnew_group\fR.
526.TP
527.B new_group
528Start a new reporting group. If not given, all jobs in a file will be part
529of the same reporting group, unless separated by a stonewall.
530.TP
531.BI numjobs \fR=\fPint
532Number of clones (processes/threads performing the same workload) of this job.
533Default: 1.
534.TP
535.B group_reporting
536If set, display per-group reports instead of per-job when \fBnumjobs\fR is
537specified.
538.TP
539.B thread
540Use threads created with \fBpthread_create\fR\|(3) instead of processes created
541with \fBfork\fR\|(2).
542.TP
f7fa2653 543.BI zonesize \fR=\fPint
d60e92d1
AC
544Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
545.TP
f7fa2653 546.BI zoneskip \fR=\fPint
d1429b5c 547Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
d60e92d1
AC
548read.
549.TP
550.BI write_iolog \fR=\fPstr
551Write the issued I/O patterns to the specified file.
552.TP
553.BI read_iolog \fR=\fPstr
554Replay the I/O patterns contained in the specified file generated by
555\fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
556.TP
901bb994
JA
557.B write_bw_log \fR=\fPstr
558If given, write a bandwidth log of the jobs in this job file. Can be used to
559store data of the bandwidth of the jobs in their lifetime. The included
560fio_generate_plots script uses gnuplot to turn these text files into nice
561graphs. See \fBwrite_log_log\fR for behaviour of given filename. For this
562option, the postfix is _bw.log.
d60e92d1
AC
563.TP
564.B write_lat_log
901bb994
JA
565Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no
566filename is given with this option, the default filename of "jobname_type.log"
567is used. Even if the filename is given, fio will still append the type of log.
568.TP
569.B disable_clat \fR=\fPbool
570Disable measurements of completion latency numbers. Useful only for cutting
571back the number of calls to gettimeofday, as that does impact performance at
572really high IOPS rates. Note that to really get rid of a large amount of these
573calls, this option must be used with disable_slat and disable_bw as well.
574.TP
575.B disable_slat \fR=\fPbool
576Disable measurements of submission latency numbers. See \fBdisable_clat\fR.
577.TP
578.B disable_bw_measurement \fR=\fPbool
579Disable measurements of throughput/bandwidth numbers. See \fBdisable_clat\fR.
d60e92d1 580.TP
f7fa2653 581.BI lockmem \fR=\fPint
d60e92d1
AC
582Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
583simulate a smaller amount of memory.
584.TP
585.BI exec_prerun \fR=\fPstr
586Before running the job, execute the specified command with \fBsystem\fR\|(3).
587.TP
588.BI exec_postrun \fR=\fPstr
589Same as \fBexec_prerun\fR, but the command is executed after the job completes.
590.TP
591.BI ioscheduler \fR=\fPstr
592Attempt to switch the device hosting the file to the specified I/O scheduler.
593.TP
594.BI cpuload \fR=\fPint
595If the job is a CPU cycle-eater, attempt to use the specified percentage of
596CPU cycles.
597.TP
598.BI cpuchunks \fR=\fPint
599If the job is a CPU cycle-eater, split the load into cycles of the
600given time in milliseconds.
601.TP
602.BI disk_util \fR=\fPbool
d1429b5c 603Generate disk utilization statistics if the platform supports it. Default: true.
901bb994
JA
604.TP
605.BI gtod_reduce \fR=\fPbool
606Enable all of the gettimeofday() reducing options (disable_clat, disable_slat,
607disable_bw) plus reduce precision of the timeout somewhat to really shrink the
608gettimeofday() call count. With this option enabled, we only do about 0.4% of
609the gtod() calls we would have done if all time keeping was enabled.
610.TP
611.BI gtod_cpu \fR=\fPint
612Sometimes it's cheaper to dedicate a single thread of execution to just getting
613the current time. Fio (and databases, for instance) are very intensive on
614gettimeofday() calls. With this option, you can set one CPU aside for doing
615nothing but logging current time to a shared memory location. Then the other
616threads/processes that run IO workloads need only copy that segment, instead of
617entering the kernel with a gettimeofday() call. The CPU set aside for doing
618these time calls will be excluded from other uses. Fio will manually clear it
619from the CPU mask of other jobs.
d60e92d1 620.SH OUTPUT
d1429b5c
AC
621While running, \fBfio\fR will display the status of the created jobs. For
622example:
d60e92d1 623.RS
d1429b5c 624.P
d60e92d1
AC
625Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
626.RE
627.P
d1429b5c
AC
628The characters in the first set of brackets denote the current status of each
629threads. The possible values are:
630.P
631.PD 0
d60e92d1
AC
632.RS
633.TP
634.B P
635Setup but not started.
636.TP
637.B C
638Thread created.
639.TP
640.B I
641Initialized, waiting.
642.TP
643.B R
644Running, doing sequential reads.
645.TP
646.B r
647Running, doing random reads.
648.TP
649.B W
650Running, doing sequential writes.
651.TP
652.B w
653Running, doing random writes.
654.TP
655.B M
656Running, doing mixed sequential reads/writes.
657.TP
658.B m
659Running, doing mixed random reads/writes.
660.TP
661.B F
662Running, currently waiting for \fBfsync\fR\|(2).
663.TP
664.B V
665Running, verifying written data.
666.TP
667.B E
668Exited, not reaped by main thread.
669.TP
670.B \-
671Exited, thread reaped.
672.RE
d1429b5c 673.PD
d60e92d1
AC
674.P
675The second set of brackets shows the estimated completion percentage of
676the current group. The third set shows the read and write I/O rate,
677respectively. Finally, the estimated run time of the job is displayed.
678.P
679When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
680for each thread, each group of threads, and each disk, in that order.
681.P
682Per-thread statistics first show the threads client number, group-id, and
683error code. The remaining figures are as follows:
684.RS
d60e92d1
AC
685.TP
686.B io
687Number of megabytes of I/O performed.
688.TP
689.B bw
690Average data rate (bandwidth).
691.TP
692.B runt
693Threads run time.
694.TP
695.B slat
696Submission latency minimum, maximum, average and standard deviation. This is
697the time it took to submit the I/O.
698.TP
699.B clat
700Completion latency minimum, maximum, average and standard deviation. This
701is the time between submission and completion.
702.TP
703.B bw
704Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
705and standard deviation.
706.TP
707.B cpu
708CPU usage statistics. Includes user and system time, number of context switches
709this thread went through and number of major and minor page faults.
710.TP
711.B IO depths
712Distribution of I/O depths. Each depth includes everything less than (or equal)
713to it, but greater than the previous depth.
714.TP
715.B IO issued
716Number of read/write requests issued, and number of short read/write requests.
717.TP
718.B IO latencies
719Distribution of I/O completion latencies. The numbers follow the same pattern
720as \fBIO depths\fR.
721.RE
d60e92d1
AC
722.P
723The group statistics show:
d1429b5c 724.PD 0
d60e92d1
AC
725.RS
726.TP
727.B io
728Number of megabytes I/O performed.
729.TP
730.B aggrb
731Aggregate bandwidth of threads in the group.
732.TP
733.B minb
734Minimum average bandwidth a thread saw.
735.TP
736.B maxb
737Maximum average bandwidth a thread saw.
738.TP
739.B mint
d1429b5c 740Shortest runtime of threads in the group.
d60e92d1
AC
741.TP
742.B maxt
743Longest runtime of threads in the group.
744.RE
d1429b5c 745.PD
d60e92d1
AC
746.P
747Finally, disk statistics are printed with reads first:
d1429b5c 748.PD 0
d60e92d1
AC
749.RS
750.TP
751.B ios
752Number of I/Os performed by all groups.
753.TP
754.B merge
755Number of merges in the I/O scheduler.
756.TP
757.B ticks
758Number of ticks we kept the disk busy.
759.TP
760.B io_queue
761Total time spent in the disk queue.
762.TP
763.B util
764Disk utilization.
765.RE
d1429b5c 766.PD
d60e92d1
AC
767.SH TERSE OUTPUT
768If the \fB\-\-minimal\fR option is given, the results will be printed in a
769semicolon-delimited format suitable for scripted use. The fields are:
770.P
771.RS
772.B jobname, groupid, error
773.P
774Read status:
775.RS
776.B KiB I/O, bandwidth \fR(KiB/s)\fP, runtime \fR(ms)\fP
777.P
778Submission latency:
779.RS
780.B min, max, mean, standard deviation
781.RE
782Completion latency:
783.RS
784.B min, max, mean, standard deviation
785.RE
786Bandwidth:
787.RS
788.B min, max, aggregate percentage of total, mean, standard deviation
789.RE
790.RE
791.P
792Write status:
793.RS
794.B KiB I/O, bandwidth \fR(KiB/s)\fP, runtime \fR(ms)\fP
795.P
796Submission latency:
797.RS
798.B min, max, mean, standard deviation
799.RE
800Completion latency:
801.RS
802.B min, max, mean, standard deviation
803.RE
804Bandwidth:
805.RS
806.B min, max, aggregate percentage of total, mean, standard deviation
807.RE
808.RE
809.P
d1429b5c 810CPU usage:
d60e92d1 811.RS
bd2626f0 812.B user, system, context switches, major page faults, minor page faults
d60e92d1
AC
813.RE
814.P
815IO depth distribution:
816.RS
817.B <=1, 2, 4, 8, 16, 32, >=64
818.RE
819.P
820IO latency distribution (ms):
821.RS
822.B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, >=2000
823.RE
824.P
825.B text description
826.RE
827.SH AUTHORS
828.B fio
829was written by Jens Axboe <jens.axboe@oracle.com>.
d1429b5c
AC
830.br
831This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
d60e92d1
AC
832on documentation by Jens Axboe.
833.SH "REPORTING BUGS"
d1429b5c
AC
834Report bugs to the \fBfio\fR mailing list <fio-devel@kernel.dk>.
835See \fBREADME\fR.
d60e92d1 836.SH "SEE ALSO"
d1429b5c
AC
837For further documentation see \fBHOWTO\fR and \fBREADME\fR.
838.br
839Sample jobfiles are available in the \fBexamples\fR directory.
d60e92d1 840