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 \-\-cmdhelp \fR=\fPcommand
41 Print help information for \fIcommand\fR. May be `all' for all commands.
44 Display usage information and exit.
47 Display version information and exit.
49 Job files are in `ini' format. They consist of one or more
50 job definitions, which begin with a job name in square brackets and
51 extend to the next job name. The job name can be any ASCII string
52 except `global', which has a special meaning. Following the job name is
53 a sequence of zero or more parameters, one per line, that define the
54 behavior of the job. Any line starting with a `;' or `#' character is
55 considered a comment and ignored.
58 The global section contains default parameters for jobs specified in the
59 job file. A job is only affected by global sections residing above it,
60 and there may be any number of global sections. Specific job definitions
61 may override any parameter set in global sections.
64 Some parameters may take arguments of a specific type. The types used are:
67 String: a sequence of alphanumeric characters.
70 Integer: a whole number, possibly negative. If prefixed with `0x', the value
71 is assumed to be base 16 (hexadecimal).
74 SI integer: a whole number, possibly containing a suffix denoting the base unit
75 of the value. Accepted suffixes are `k', 'M' and 'G', denoting kilo (1024),
76 mega (1024*1024) and giga (1024*1024*1024) respectively.
79 Boolean: a true or false value. `0' denotes false, `1' denotes true.
82 Integer range: a range of integers specified in the format
83 \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and
84 \fIupper\fR may contain a suffix as described above. If an option allows two
85 sets of ranges, they are separated with a `,' or `/' character. For example:
90 May be used to override the job name. On the command line, this paramter
91 has the special purpose of signalling the start of a new job.
93 .BI description \fR=\fPstr
94 Human-readable description of the job. It is printed when the job is run, but
95 otherwise has no special purpose.
97 .BI directory \fR=\fPstr
98 Prefix filenames with this directory. Used to place files in a location other
101 .BI filename \fR=\fPstr
103 normally makes up a file name based on the job name, thread number, and file
104 number. If you want to share files between threads in a job or several jobs,
105 specify a \fIfilename\fR for each of them to override the default. If the I/O
106 engine used is `net', \fIfilename\fR is the host and port to connect to in the
107 format \fIhost\fR/\fIport\fR. If the I/O engine is file-based, you can specify
108 a number of files by separating the names with a `:' character. `\-' is a
109 reserved name, meaning stdin or stdout, depending on the read/write direction
112 .BI opendir \fR=\fPstr
113 Recursively open any files below directory \fIstr\fR.
115 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
116 Type of I/O pattern. Accepted values are:
133 Mixed sequential reads and writes.
136 Mixed random reads and writes.
139 For mixed I/O, the default split is 50/50. For random I/O, the number of I/Os
140 to perform before getting a new offset can be specified by appending
141 `:\fIint\fR' to the pattern type. The default is 1.
144 .BI randrepeat \fR=\fPbool
145 Seed the random number generator in a predictable way so results are repeatable
146 across runs. Default: true.
148 .BI fadvise_hint \fR=\fPbool
149 Disable use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
150 are likely to be issued. Default: true.
152 .BI size \fR=\fPsiint
153 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
154 been transfered, unless limited by other options (\fBruntime\fR, for instance).
155 Unless \fBnr_files\fR and \fBfilesize\fR options are given, this amount will be
156 divided between the available files for the job.
158 .BI filesize \fR=\fPirange
159 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
160 for files at random within the given range, limited to \fBsize\fR in total (if
161 that is given). If \fBfilesize\fR is not specified, each created file is the
164 .BI blocksize \fR=\fPsiint "\fR,\fB bs" \fR=\fPsiint
165 Block size for I/O units. Default: 4k. Values for reads and writes can be
166 specified seperately in the format \fIread\fR,\fIwrite\fR, either of
167 which may be empty to leave that value at its default.
169 .BI blocksize_range \fR=\fPirange "\fR,\fB bsrange" \fR=\fPirange
170 Specify a range of I/O block sizes. The issued I/O unit will always be a
171 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
172 to both reads and writes, but can be specified seperately (see \fBblocksize\fR).
174 .B blocksize_unaligned\fR,\fP bs_unaligned
175 If set, any size in \fBblocksize_range\fR may be used. This typically won't
176 work with direct I/O, as that normally requires sector alignment.
179 Initialise buffers with all zeros. Default: fill buffers with random data.
181 .BI nrfiles \fR=\fPint
182 Number of files to use for this job. Default: 1.
184 .BI openfiles \fR=\fPint
185 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
187 .BI file_service_type \fR=\fPstr
188 Defines how files to service are selected. The following types are defined:
193 Choose a file at random
196 Round robin over open files (default).
199 The number of I/Os to issue before switching a new file can be specified by
200 appending `:\fIint\fR' to the service type.
203 .BI ioengine \fR=\fPstr
204 Defines how the job issues I/O. The following types are defined:
209 Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
210 position the I/O location.
213 Linux native asynchronous I/O.
216 glibc POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
219 File is memory mapped with \fImmap\fR\|(2) and data copied using
223 \fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
224 transfer data from user-space to the kernel.
227 Use the syslet system calls to make regular read/write asynchronous.
230 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
231 the target is an sg character device, we use \fIread\fR\|(2) and
232 \fIwrite\fR\|(2) for asynchronous I/O.
235 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
236 itself and for debugging and testing purposes.
239 Transfer over the network. \fBfilename\fR must be set appropriately to
240 `\fIhost\fR/\fIport\fR' regardless of data direction. If receiving, only the
241 \fIport\fR argument is used.
244 Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
248 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
249 \fBcpucycles\fR parameters.
252 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
253 approach to asycnronous I/O.
255 See <http://www.xmailserver.org/guasi\-lib.html>.
258 Loads an external I/O engine object file. Append the engine filename as
263 .BI iodepth \fR=\fPint
264 Number of I/O units to keep in flight against the file. Default: 1.
266 .BI iodepth_batch \fR=\fPint
267 Number of I/Os to submit at once. Default: \fBiodepth\fR.
269 .BI iodepth_low \fR=\fPint
270 Low watermark indicating when to start filling the queue again. Default:
273 .BI direct \fR=\fPbool
274 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
276 .BI buffered \fR=\fPbool
277 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
280 .BI offset \fR=\fPsiint
281 Offset in the file to start I/O. Data before the offset will not be touched.
284 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
285 0, don't sync. Default: 0.
287 .BI overwrite \fR=\fPbool
288 If writing, setup the file first and do overwrites. Default: false.
290 .BI end_fsync \fR=\fPbool
291 Sync file contents when job exits. Default: false.
293 .BI fsync_on_close \fR=\fPbool
294 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
295 it will happen on every close, not just at the end of the job. Default: false.
297 .BI rwmixcycle \fR=\fPint
298 How many milliseconds before switching between reads and writes for a mixed
299 workload. Default: 500ms.
301 .BI rwmixread \fR=\fPint
302 Percentage of a mixed workload that should be reads. Default: 50.
304 .BI rwmixwrite \fR=\fPint
305 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
306 \fBwrmixwrite\fR are given and do not sum to 100%, the latter of the two
307 overrides the first. Default: 50.
310 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
311 this parameter is given, a new offset will be chosen without looking at past
312 I/O history. This parameter is mutually exclusive with \fBverify\fR.
315 Run job with given nice value. See \fInice\fR\|(2).
318 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
321 .BI prioclass \fR=\fPint
322 Set I/O priority class. See \fIionice\fR\|(1).
324 .BI thinktime \fR=\fPint
325 Stall job for given number of microseconds between issuing I/Os.
327 .BI thinktime_spin \fR=\fPint
328 Pretend to spend CPU time for given number of microseconds, sleeping the rest
329 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
331 .BI thinktime_blocks \fR=\fPint
332 Number of blocks to issue before waiting \fBthinktime\fR microseconds.
336 Cap bandwidth used by this job to this number of KiB/s.
338 .BI ratemin \fR=\fPint
339 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
340 Failing to meet this requirement will cause the job to exit.
342 .BI rate_iops \fR=\fPint
343 Cap the bandwidth to this number of IOPS. If \fBblocksize\fR is a range, the
344 smallest block size is used as the metric.
346 .BI rate_iops_min \fR=\fPint
347 If this rate of I/O is not met, the job will exit.
349 .BI ratecycle \fR=\fPint
350 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
351 milliseconds. Default: 1000ms.
353 .BI cpumask \fR=\fPint
354 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
355 may run on. See \fBsched_setaffinity\fR\|(2).
357 .BI cpus_allowed \fR=\fPstr
358 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
360 .BI startdelay \fR=\fPint
361 Delay start of job for the specified number of seconds.
363 .BI runtime \fR=\fPint
364 Terminate processing after the specified number of seconds.
367 If given, run for the specified \fBruntime\fR duration even if the files are
368 completely read or written. The same workload will be repeated as many times
369 as \fBruntime\fR allows.
371 .BI invalidate \fR=\fPbool
372 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
375 Use synchronous I/O for buffered writes. For the majority of I/O engines,
376 this means using O_SYNC. Default: false.
378 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
379 Allocation method for I/O unit buffer. Allowed values are:
384 Allocate memory with \fImalloc\fR\|(3).
387 Use shared memory buffers allocated through \fIshmget\fR\|(2).
390 Same as \fBshm\fR, but use huge pages as backing.
393 Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
394 is given after the option in the format `:\fIfile\fR'.
397 Same as \fBmmap\fR, but use huge files as backing.
400 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
401 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
402 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
403 have hugetlbfs mounted, and \fIfile\fR must point there.
406 .BI hugepage\-size \fR=\fPsiint
407 Defines the size of a huge page. Must be at least equal to the system setting.
408 Should be a multiple of 1MiB. Default: 4MiB.
411 Terminate all jobs when one finishes. Default: wait for each job to finish.
413 .BI bwavgtime \fR=\fPint
414 Average bandwidth calculations over the given time in milliseconds. Default:
417 .BI create_serialize \fR=\fPbool
418 If true, serialize file creation for the jobs. Default: true.
420 .BI create_fsync \fR=\fPbool
421 \fIfsync\fR\|(2) data file after creation. Default: true.
423 .BI unlink \fR=\fPbool
424 Unlink job files when done. Default: false.
427 Specifies the number of iterations (runs of the same workload) of this job.
430 .BI do_verify \fR=\fPbool
431 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
434 .BI verify \fR=\fPstr
435 Method of verifying file contents after each iteration of the job. Allowed
440 .B md5 crc16 crc32 crc64 crc7 sha256 sha512
441 Store appropriate checksum in the header of each block.
444 Write extra information about each I/O (timestamp, block number, etc.). The
445 block number is verified.
448 Fill I/O buffers with a specific pattern that is used to verify. The pattern is
449 specified by appending `:\fIint\fR' to the parameter. \fIint\fR cannot be larger
453 Pretend to verify. Used for testing internals.
457 .BI verify_sort \fR=\fPbool
458 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
459 read them back in a sorted manner. Default: true.
461 .BI verify_offset \fR=\fPsiint
462 Swap the verification header with data somewhere else in the block before
463 writing. It is swapped back before verifying.
465 .BI verify_interval \fR=\fPsiint
466 Write the verification header for this number of bytes, which should divide
467 \fBblocksize\fR. Default: \fBblocksize\fR.
469 .BI verify_fatal \fR=\fPbool
470 If true, exit the job on the first observed verification failure. Default:
474 Wait for preceeding jobs in the job file to exit before starting this one.
475 \fBstonewall\fR implies \fBnew_group\fR.
478 Start a new reporting group. If not given, all jobs in a file will be part
479 of the same reporting group, unless separated by a stonewall.
481 .BI numjobs \fR=\fPint
482 Number of clones (processes/threads performing the same workload) of this job.
486 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
490 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
491 with \fBfork\fR\|(2).
493 .BI zonesize \fR=\fPsiint
494 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
496 .BI zoneskip \fR=\fPsiint
497 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
500 .BI write_iolog \fR=\fPstr
501 Write the issued I/O patterns to the specified file.
503 .BI read_iolog \fR=\fPstr
504 Replay the I/O patterns contained in the specified file generated by
505 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
508 If given, write bandwidth logs of the jobs in this file.
511 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies.
513 .BI lockmem \fR=\fPsiint
514 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
515 simulate a smaller amount of memory.
517 .BI exec_prerun \fR=\fPstr
518 Before running the job, execute the specified command with \fBsystem\fR\|(3).
520 .BI exec_postrun \fR=\fPstr
521 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
523 .BI ioscheduler \fR=\fPstr
524 Attempt to switch the device hosting the file to the specified I/O scheduler.
526 .BI cpuload \fR=\fPint
527 If the job is a CPU cycle-eater, attempt to use the specified percentage of
530 .BI cpuchunks \fR=\fPint
531 If the job is a CPU cycle-eater, split the load into cycles of the
532 given time in milliseconds.
534 .BI disk_util \fR=\fPbool
535 Generate disk utilization statistics if the platform supports it. Default: true.
537 While running, \fBfio\fR will display the status of the created jobs. For
541 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
544 The characters in the first set of brackets denote the current status of each
545 threads. The possible values are:
551 Setup but not started.
557 Initialized, waiting.
560 Running, doing sequential reads.
563 Running, doing random reads.
566 Running, doing sequential writes.
569 Running, doing random writes.
572 Running, doing mixed sequential reads/writes.
575 Running, doing mixed random reads/writes.
578 Running, currently waiting for \fBfsync\fR\|(2).
581 Running, verifying written data.
584 Exited, not reaped by main thread.
587 Exited, thread reaped.
591 The second set of brackets shows the estimated completion percentage of
592 the current group. The third set shows the read and write I/O rate,
593 respectively. Finally, the estimated run time of the job is displayed.
595 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
596 for each thread, each group of threads, and each disk, in that order.
598 Per-thread statistics first show the threads client number, group-id, and
599 error code. The remaining figures are as follows:
603 Number of megabytes of I/O performed.
606 Average data rate (bandwidth).
612 Submission latency minimum, maximum, average and standard deviation. This is
613 the time it took to submit the I/O.
616 Completion latency minimum, maximum, average and standard deviation. This
617 is the time between submission and completion.
620 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
621 and standard deviation.
624 CPU usage statistics. Includes user and system time, number of context switches
625 this thread went through and number of major and minor page faults.
628 Distribution of I/O depths. Each depth includes everything less than (or equal)
629 to it, but greater than the previous depth.
632 Number of read/write requests issued, and number of short read/write requests.
635 Distribution of I/O completion latencies. The numbers follow the same pattern
639 The group statistics show:
644 Number of megabytes I/O performed.
647 Aggregate bandwidth of threads in the group.
650 Minimum average bandwidth a thread saw.
653 Maximum average bandwidth a thread saw.
656 Shortest runtime of threads in the group.
659 Longest runtime of threads in the group.
663 Finally, disk statistics are printed with reads first:
668 Number of I/Os performed by all groups.
671 Number of merges in the I/O scheduler.
674 Number of ticks we kept the disk busy.
677 Total time spent in the disk queue.
684 If the \fB\-\-minimal\fR option is given, the results will be printed in a
685 semicolon-delimited format suitable for scripted use. The fields are:
688 .B jobname, groupid, error
692 .B KiB I/O, bandwidth \fR(KiB/s)\fP, runtime \fR(ms)\fP
696 .B min, max, mean, standard deviation
700 .B min, max, mean, standard deviation
704 .B min, max, aggregate percentage of total, mean, standard deviation
710 .B KiB I/O, bandwidth \fR(KiB/s)\fP, runtime \fR(ms)\fP
714 .B min, max, mean, standard deviation
718 .B min, max, mean, standard deviation
722 .B min, max, aggregate percentage of total, mean, standard deviation
728 .B user, system, context switches
731 IO depth distribution:
733 .B <=1, 2, 4, 8, 16, 32, >=64
736 IO latency distribution (ms):
738 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, >=2000
745 was written by Jens Axboe <jens.axboe@oracle.com>.
747 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
748 on documentation by Jens Axboe.
750 Report bugs to the \fBfio\fR mailing list <fio-devel@kernel.dk>.
753 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
755 Sample jobfiles are available in the \fBexamples\fR directory.
projects / fio.git / blob