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' and 'G', denoting kilo (1024),
82 mega (1024*1024) and giga (1024*1024*1024) respectively. If prefixed with '0x',
83 the value is assumed to be base 16 (hexadecimal).
86 Boolean: a true or false value. `0' denotes false, `1' denotes true.
89 Integer range: a range of integers specified in the format
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
92 sets of ranges, they are separated with a `,' or `/' character. For example:
97 May be used to override the job name. On the command line, this parameter
98 has the special purpose of signalling the start of a new job.
100 .BI description \fR=\fPstr
101 Human-readable description of the job. It is printed when the job is run, but
102 otherwise has no special purpose.
104 .BI directory \fR=\fPstr
105 Prefix filenames with this directory. Used to place files in a location other
108 .BI filename \fR=\fPstr
110 normally makes up a file name based on the job name, thread number, and file
111 number. If you want to share files between threads in a job or several jobs,
112 specify a \fIfilename\fR for each of them to override the default. If the I/O
113 engine used is `net', \fIfilename\fR is the host and port to connect to in the
114 format \fIhost\fR/\fIport\fR. If the I/O engine is file-based, you can specify
115 a number of files by separating the names with a `:' character. `\-' is a
116 reserved name, meaning stdin or stdout, depending on the read/write direction
119 .BI opendir \fR=\fPstr
120 Recursively open any files below directory \fIstr\fR.
122 .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr
123 Type of I/O pattern. Accepted values are:
140 Mixed sequential reads and writes.
143 Mixed random reads and writes.
146 For mixed I/O, the default split is 50/50. For random I/O, the number of I/Os
147 to perform before getting a new offset can be specified by appending
148 `:\fIint\fR' to the pattern type. The default is 1.
151 .BI randrepeat \fR=\fPbool
152 Seed the random number generator in a predictable way so results are repeatable
153 across runs. Default: true.
155 .BI fadvise_hint \fR=\fPbool
156 Disable use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns
157 are likely to be issued. Default: true.
160 Total size of I/O for this job. \fBfio\fR will run until this many bytes have
161 been transfered, unless limited by other options (\fBruntime\fR, for instance).
162 Unless \fBnr_files\fR and \fBfilesize\fR options are given, this amount will be
163 divided between the available files for the job.
165 .BI filesize \fR=\fPirange
166 Individual file sizes. May be a range, in which case \fBfio\fR will select sizes
167 for files at random within the given range, limited to \fBsize\fR in total (if
168 that is given). If \fBfilesize\fR is not specified, each created file is the
171 .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int]
172 Block size for I/O units. Default: 4k. Values for reads and writes can be
173 specified seperately in the format \fIread\fR,\fIwrite\fR, either of
174 which may be empty to leave that value at its default.
176 .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange]
177 Specify a range of I/O block sizes. The issued I/O unit will always be a
178 multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies
179 to both reads and writes if only one range is given, but can be specified
180 seperately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k.
181 Also (see \fBblocksize\fR).
183 .BI bssplit \fR=\fPstr
184 This option allows even finer grained control of the block sizes issued,
185 not just even splits between them. With this option, you can weight various
186 block sizes for exact control of the issued IO for a job that has mixed
187 block sizes. The format of the option is bssplit=blocksize/percentage,
188 optionally adding as many definitions as needed seperated by a colon.
189 Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k
190 blocks and 40% 32k blocks.
192 .B blocksize_unaligned\fR,\fP bs_unaligned
193 If set, any size in \fBblocksize_range\fR may be used. This typically won't
194 work with direct I/O, as that normally requires sector alignment.
196 .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int]
197 At what boundary to align random IO offsets. Defaults to the same as
198 'blocksize' the minimum blocksize given. Minimum alignment is typically 512b
199 for using direct IO, though it usually depends on the hardware block size.
200 This option is mutually exclusive with using a random map for files, so it
201 will turn off that option.
203 Initialise buffers with all zeros. Default: fill buffers with random data.
205 .BI nrfiles \fR=\fPint
206 Number of files to use for this job. Default: 1.
208 .BI openfiles \fR=\fPint
209 Number of files to keep open at the same time. Default: \fBnrfiles\fR.
211 .BI file_service_type \fR=\fPstr
212 Defines how files to service are selected. The following types are defined:
217 Choose a file at random
220 Round robin over open files (default).
222 Do each file in the set sequentially.
225 The number of I/Os to issue before switching a new file can be specified by
226 appending `:\fIint\fR' to the service type.
229 .BI ioengine \fR=\fPstr
230 Defines how the job issues I/O. The following types are defined:
235 Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to
236 position the I/O location.
239 Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O.
242 Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by
243 coalescing adjacents IOs into a single submission.
246 Linux native asynchronous I/O.
249 glibc POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3).
252 File is memory mapped with \fImmap\fR\|(2) and data copied using
256 \fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to
257 transfer data from user-space to the kernel.
260 Use the syslet system calls to make regular read/write asynchronous.
263 SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if
264 the target is an sg character device, we use \fIread\fR\|(2) and
265 \fIwrite\fR\|(2) for asynchronous I/O.
268 Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR
269 itself and for debugging and testing purposes.
272 Transfer over the network. \fBfilename\fR must be set appropriately to
273 `\fIhost\fR/\fIport\fR' regardless of data direction. If receiving, only the
274 \fIport\fR argument is used.
277 Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data
281 Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and
282 \fBcpucycles\fR parameters.
285 The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface
286 approach to asycnronous I/O.
288 See <http://www.xmailserver.org/guasi\-lib.html>.
291 Loads an external I/O engine object file. Append the engine filename as
296 .BI iodepth \fR=\fPint
297 Number of I/O units to keep in flight against the file. Default: 1.
299 .BI iodepth_batch \fR=\fPint
300 Number of I/Os to submit at once. Default: \fBiodepth\fR.
302 .BI iodepth_low \fR=\fPint
303 Low watermark indicating when to start filling the queue again. Default:
306 .BI direct \fR=\fPbool
307 If true, use non-buffered I/O (usually O_DIRECT). Default: false.
309 .BI buffered \fR=\fPbool
310 If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter.
313 .BI offset \fR=\fPint
314 Offset in the file to start I/O. Data before the offset will not be touched.
317 How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If
318 0, don't sync. Default: 0.
320 .BI overwrite \fR=\fPbool
321 If writing, setup the file first and do overwrites. Default: false.
323 .BI end_fsync \fR=\fPbool
324 Sync file contents when job exits. Default: false.
326 .BI fsync_on_close \fR=\fPbool
327 If true, sync file contents on close. This differs from \fBend_fsync\fR in that
328 it will happen on every close, not just at the end of the job. Default: false.
330 .BI rwmixcycle \fR=\fPint
331 How many milliseconds before switching between reads and writes for a mixed
332 workload. Default: 500ms.
334 .BI rwmixread \fR=\fPint
335 Percentage of a mixed workload that should be reads. Default: 50.
337 .BI rwmixwrite \fR=\fPint
338 Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and
339 \fBwrmixwrite\fR are given and do not sum to 100%, the latter of the two
340 overrides the first. Default: 50.
343 Normally \fBfio\fR will cover every block of the file when doing random I/O. If
344 this parameter is given, a new offset will be chosen without looking at past
345 I/O history. This parameter is mutually exclusive with \fBverify\fR.
348 Run job with given nice value. See \fInice\fR\|(2).
351 Set I/O priority value of this job between 0 (highest) and 7 (lowest). See
354 .BI prioclass \fR=\fPint
355 Set I/O priority class. See \fIionice\fR\|(1).
357 .BI thinktime \fR=\fPint
358 Stall job for given number of microseconds between issuing I/Os.
360 .BI thinktime_spin \fR=\fPint
361 Pretend to spend CPU time for given number of microseconds, sleeping the rest
362 of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set.
364 .BI thinktime_blocks \fR=\fPint
365 Number of blocks to issue before waiting \fBthinktime\fR microseconds.
369 Cap bandwidth used by this job to this number of KiB/s.
371 .BI ratemin \fR=\fPint
372 Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth.
373 Failing to meet this requirement will cause the job to exit.
375 .BI rate_iops \fR=\fPint
376 Cap the bandwidth to this number of IOPS. If \fBblocksize\fR is a range, the
377 smallest block size is used as the metric.
379 .BI rate_iops_min \fR=\fPint
380 If this rate of I/O is not met, the job will exit.
382 .BI ratecycle \fR=\fPint
383 Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of
384 milliseconds. Default: 1000ms.
386 .BI cpumask \fR=\fPint
387 Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job
388 may run on. See \fBsched_setaffinity\fR\|(2).
390 .BI cpus_allowed \fR=\fPstr
391 Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers.
393 .BI startdelay \fR=\fPint
394 Delay start of job for the specified number of seconds.
396 .BI runtime \fR=\fPint
397 Terminate processing after the specified number of seconds.
400 If given, run for the specified \fBruntime\fR duration even if the files are
401 completely read or written. The same workload will be repeated as many times
402 as \fBruntime\fR allows.
404 .BI invalidate \fR=\fPbool
405 Invalidate buffer-cache for the file prior to starting I/O. Default: true.
408 Use synchronous I/O for buffered writes. For the majority of I/O engines,
409 this means using O_SYNC. Default: false.
411 .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr
412 Allocation method for I/O unit buffer. Allowed values are:
417 Allocate memory with \fImalloc\fR\|(3).
420 Use shared memory buffers allocated through \fIshmget\fR\|(2).
423 Same as \fBshm\fR, but use huge pages as backing.
426 Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename
427 is given after the option in the format `:\fIfile\fR'.
430 Same as \fBmmap\fR, but use huge files as backing.
433 The amount of memory allocated is the maximum allowed \fBblocksize\fR for the
434 job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work,
435 the system must have free huge pages allocated. \fBmmaphuge\fR also needs to
436 have hugetlbfs mounted, and \fIfile\fR must point there.
439 .BI hugepage\-size \fR=\fPint
440 Defines the size of a huge page. Must be at least equal to the system setting.
441 Should be a multiple of 1MiB. Default: 4MiB.
444 Terminate all jobs when one finishes. Default: wait for each job to finish.
446 .BI bwavgtime \fR=\fPint
447 Average bandwidth calculations over the given time in milliseconds. Default:
450 .BI create_serialize \fR=\fPbool
451 If true, serialize file creation for the jobs. Default: true.
453 .BI create_fsync \fR=\fPbool
454 \fIfsync\fR\|(2) data file after creation. Default: true.
456 .BI create_on_open \fR=\fPbool
457 If true, the files are not created until they are opened for IO by the job.
459 .BI unlink \fR=\fPbool
460 Unlink job files when done. Default: false.
463 Specifies the number of iterations (runs of the same workload) of this job.
466 .BI do_verify \fR=\fPbool
467 Run the verify phase after a write phase. Only valid if \fBverify\fR is set.
470 .BI verify \fR=\fPstr
471 Method of verifying file contents after each iteration of the job. Allowed
476 .B md5 crc16 crc32 crc64 crc7 sha256 sha512
477 Store appropriate checksum in the header of each block.
480 Write extra information about each I/O (timestamp, block number, etc.). The
481 block number is verified.
484 Fill I/O buffers with a specific pattern that is used to verify. The pattern is
485 specified by appending `:\fIint\fR' to the parameter. \fIint\fR cannot be larger
489 Pretend to verify. Used for testing internals.
493 .BI verify_sort \fR=\fPbool
494 If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to
495 read them back in a sorted manner. Default: true.
497 .BI verify_offset \fR=\fPint
498 Swap the verification header with data somewhere else in the block before
499 writing. It is swapped back before verifying.
501 .BI verify_interval \fR=\fPint
502 Write the verification header for this number of bytes, which should divide
503 \fBblocksize\fR. Default: \fBblocksize\fR.
505 .BI verify_fatal \fR=\fPbool
506 If true, exit the job on the first observed verification failure. Default:
510 Wait for preceeding jobs in the job file to exit before starting this one.
511 \fBstonewall\fR implies \fBnew_group\fR.
514 Start a new reporting group. If not given, all jobs in a file will be part
515 of the same reporting group, unless separated by a stonewall.
517 .BI numjobs \fR=\fPint
518 Number of clones (processes/threads performing the same workload) of this job.
522 If set, display per-group reports instead of per-job when \fBnumjobs\fR is
526 Use threads created with \fBpthread_create\fR\|(3) instead of processes created
527 with \fBfork\fR\|(2).
529 .BI zonesize \fR=\fPint
530 Divide file into zones of the specified size in bytes. See \fBzoneskip\fR.
532 .BI zoneskip \fR=\fPint
533 Skip the specified number of bytes when \fBzonesize\fR bytes of data have been
536 .BI write_iolog \fR=\fPstr
537 Write the issued I/O patterns to the specified file.
539 .BI read_iolog \fR=\fPstr
540 Replay the I/O patterns contained in the specified file generated by
541 \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file.
544 If given, write bandwidth logs of the jobs in this file.
547 Same as \fBwrite_bw_log\fR, but writes I/O completion latencies.
549 .BI lockmem \fR=\fPint
550 Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to
551 simulate a smaller amount of memory.
553 .BI exec_prerun \fR=\fPstr
554 Before running the job, execute the specified command with \fBsystem\fR\|(3).
556 .BI exec_postrun \fR=\fPstr
557 Same as \fBexec_prerun\fR, but the command is executed after the job completes.
559 .BI ioscheduler \fR=\fPstr
560 Attempt to switch the device hosting the file to the specified I/O scheduler.
562 .BI cpuload \fR=\fPint
563 If the job is a CPU cycle-eater, attempt to use the specified percentage of
566 .BI cpuchunks \fR=\fPint
567 If the job is a CPU cycle-eater, split the load into cycles of the
568 given time in milliseconds.
570 .BI disk_util \fR=\fPbool
571 Generate disk utilization statistics if the platform supports it. Default: true.
573 While running, \fBfio\fR will display the status of the created jobs. For
577 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
580 The characters in the first set of brackets denote the current status of each
581 threads. The possible values are:
587 Setup but not started.
593 Initialized, waiting.
596 Running, doing sequential reads.
599 Running, doing random reads.
602 Running, doing sequential writes.
605 Running, doing random writes.
608 Running, doing mixed sequential reads/writes.
611 Running, doing mixed random reads/writes.
614 Running, currently waiting for \fBfsync\fR\|(2).
617 Running, verifying written data.
620 Exited, not reaped by main thread.
623 Exited, thread reaped.
627 The second set of brackets shows the estimated completion percentage of
628 the current group. The third set shows the read and write I/O rate,
629 respectively. Finally, the estimated run time of the job is displayed.
631 When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data
632 for each thread, each group of threads, and each disk, in that order.
634 Per-thread statistics first show the threads client number, group-id, and
635 error code. The remaining figures are as follows:
639 Number of megabytes of I/O performed.
642 Average data rate (bandwidth).
648 Submission latency minimum, maximum, average and standard deviation. This is
649 the time it took to submit the I/O.
652 Completion latency minimum, maximum, average and standard deviation. This
653 is the time between submission and completion.
656 Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average
657 and standard deviation.
660 CPU usage statistics. Includes user and system time, number of context switches
661 this thread went through and number of major and minor page faults.
664 Distribution of I/O depths. Each depth includes everything less than (or equal)
665 to it, but greater than the previous depth.
668 Number of read/write requests issued, and number of short read/write requests.
671 Distribution of I/O completion latencies. The numbers follow the same pattern
675 The group statistics show:
680 Number of megabytes I/O performed.
683 Aggregate bandwidth of threads in the group.
686 Minimum average bandwidth a thread saw.
689 Maximum average bandwidth a thread saw.
692 Shortest runtime of threads in the group.
695 Longest runtime of threads in the group.
699 Finally, disk statistics are printed with reads first:
704 Number of I/Os performed by all groups.
707 Number of merges in the I/O scheduler.
710 Number of ticks we kept the disk busy.
713 Total time spent in the disk queue.
720 If the \fB\-\-minimal\fR option is given, the results will be printed in a
721 semicolon-delimited format suitable for scripted use. The fields are:
724 .B jobname, groupid, error
728 .B KiB I/O, bandwidth \fR(KiB/s)\fP, runtime \fR(ms)\fP
732 .B min, max, mean, standard deviation
736 .B min, max, mean, standard deviation
740 .B min, max, aggregate percentage of total, mean, standard deviation
746 .B KiB I/O, bandwidth \fR(KiB/s)\fP, runtime \fR(ms)\fP
750 .B min, max, mean, standard deviation
754 .B min, max, mean, standard deviation
758 .B min, max, aggregate percentage of total, mean, standard deviation
764 .B user, system, context switches, major page faults, minor page faults
767 IO depth distribution:
769 .B <=1, 2, 4, 8, 16, 32, >=64
772 IO latency distribution (ms):
774 .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, >=2000
781 was written by Jens Axboe <jens.axboe@oracle.com>.
783 This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based
784 on documentation by Jens Axboe.
786 Report bugs to the \fBfio\fR mailing list <fio-devel@kernel.dk>.
789 For further documentation see \fBHOWTO\fR and \fBREADME\fR.
791 Sample jobfiles are available in the \fBexamples\fR directory.