4 fio is a tool that will spawn a number of threads or processes doing a
5 particular type of io action as specified by the user. fio takes a
6 number of global parameters, each inherited by the thread unless
7 otherwise parameters given to them overriding that setting is given.
8 The typical use of fio is to write a job file matching the io load
15 fio resides in a git repo, the canonical place is:
17 git://brick.kernel.dk/data/git/fio.git
19 Snapshots are frequently generated and they include the git meta data as
20 well. You can download them here:
22 http://brick.kernel.dk/snaps/
24 Pascal Bleser <guru@unixtech.be> has fio RPMs in his repository, you
27 http://linux01.gwdg.de/~pbleser/rpm-navigation.php?cat=System/fio
33 There's a mailing list associated with fio. It's meant for general
34 discussion, bug reporting, questions - basically anything that has to
35 do with fio. An automated mail detailing recent commits is automatically
36 sent to the list at most daily. The list address is fio-devel@kernel.dk,
37 subscribe by sending an empty email to fio-devel+subscribe@kernel.dk.
43 Just type 'make' and 'make install'. If on FreeBSD, for now you have to
44 specify the FreeBSD Makefile with -f, eg:
46 $ make -f Makefile.Freebsd && make -f Makefile.FreeBSD install
48 Likewise with OpenSolaris, use the Makefile.solaris to compile there.
49 This might change in the future if I opt for an autoconf type setup.
56 --output Write output to file
57 --runtime Runtime in seconds
58 --latency-log Generate per-job latency logs
59 --bandwidth-log Generate per-job bandwidth logs
60 --minimal Minimal (terse) output
61 --version Print version info and exit
62 --help Print this page
63 --cmdhelp=cmd Print command help, "all" for all of them
65 Any parameters following the options will be assumed to be job files,
66 unless they match a job file parameter. You can add as many as you want,
67 each job file will be regarded as a separate group and fio will stonewall
74 See the HOWTO file for a more detailed description of parameters and what
75 they mean. This file contains the terse version. Only a few options can
76 be controlled with command line parameters, generally it's a lot easier to
77 just write a simple job file to describe the workload. The job file format
78 is in the ini style format, as it's easy to read and write for the user.
80 The job file parameters are:
82 name=x Use 'x' as the identifier for this job.
83 description=x 'x' is a text description of the job.
84 directory=x Use 'x' as the top level directory for storing files
85 filename=x Force the use of 'x' as the filename for all files
86 in this thread. If not given, fio will make up
87 a suitable filename based on the thread and file
89 rw=x 'x' may be: read, randread, write, randwrite,
90 rw (read-write mix), randrw (read-write random mix)
91 rwmixcycle=x Base cycle for switching between read and write
93 rwmixread=x 'x' percentage of rw mix ios will be reads. If
94 rwmixwrite is also given, the last of the two will
95 be used if they don't add up to 100%.
96 rwmixwrite=x 'x' percentage of rw mix ios will be writes. See
98 rand_repeatable=x The sequence of random io blocks can be repeatable
99 across runs, if 'x' is 1.
100 size=x Set file size to x bytes (x string can include k/m/g)
101 ioengine=x 'x' may be: aio/libaio/linuxaio for Linux aio,
102 posixaio for POSIX aio, sync for regular read/write io,
103 mmap for mmap'ed io, splice for using splice/vmsplice,
104 or sgio for direct SG_IO io. The latter only works on
105 Linux on SCSI (or SCSI-like devices, such as
106 usb-storage or sata/libata driven) devices. Fio also
107 has a null io engine, which is mainly used for testing
109 iodepth=x For async io, allow 'x' ios in flight
110 overwrite=x If 'x', layout a write file first.
111 nrfiles=x Spread io load over 'x' number of files per job,
113 prio=x Run io at prio X, 0-7 is the kernel allowed range
114 prioclass=x Run io at prio class X
115 bs=x Use 'x' for thread blocksize. May include k/m postfix.
116 bsrange=x-y Mix thread block sizes randomly between x and y. May
117 also include k/m postfix.
118 direct=x 1 for direct IO, 0 for buffered IO
119 thinktime=x "Think" x usec after each io
120 rate=x Throttle rate to x KiB/sec
121 ratemin=x Quit if rate of x KiB/sec can't be met
122 ratecycle=x ratemin averaged over x msecs
123 cpumask=x Only allow job to run on CPUs defined by mask.
124 fsync=x If writing with buffered IO, fsync after every
125 'x' blocks have been written.
126 end_fsync=x If 'x', run fsync() after end-of-job.
127 startdelay=x Start this thread x seconds after startup
128 runtime=x Terminate x seconds after startup. Can include a
129 normal time suffix if not given in seconds, such as
130 'm' for minutes, 'h' for hours, and 'd' for days.
131 offset=x Start io at offset x (x string can include k/m/g)
132 invalidate=x Invalidate page cache for file prior to doing io
133 sync=x Use sync writes if x and writing buffered IO.
134 mem=x If x == malloc, use malloc for buffers. If x == shm,
135 use shared memory for buffers. If x == mmap, use
137 exitall When one thread quits, terminate the others
138 bwavgtime=x Average bandwidth stats over an x msec window.
139 create_serialize=x If 'x', serialize file creation.
140 create_fsync=x If 'x', run fsync() after file creation.
141 unlink If set, unlink files when done.
142 loops=x Run the job 'x' number of times.
143 verify=x If 'x' == md5, use md5 for verifies. If 'x' == crc32,
144 use crc32 for verifies. md5 is 'safer', but crc32 is
145 a lot faster. Only makes sense for writing to a file.
146 stonewall Wait for preceeding jobs to end before running.
147 numjobs=x Create 'x' similar entries for this job
148 thread Use pthreads instead of forked jobs
150 zoneskip=y Zone options must be paired. If given, the job
151 will skip y bytes for every x read/written. This
152 can be used to gauge hard drive speed over the entire
153 platter, without reading everything. Both x/y can
154 include k/m/g suffix.
155 iolog=x Open and read io pattern from file 'x'. The file must
156 contain one io action per line in the following format:
158 where with rw=0/1 for read/write, and the offset
159 and length entries being in bytes.
160 write_iolog=x Write an iolog to file 'x' in the same format as iolog.
161 The iolog options are exclusive, if both given the
162 read iolog will be performed.
163 write_bw_log Write a bandwidth log.
164 write_lat_log Write a latency log.
165 lockmem=x Lock down x amount of memory on the machine, to
166 simulate a machine with less memory available. x can
167 include k/m/g suffix.
168 nice=x Run job at given nice value.
169 exec_prerun=x Run 'x' before job io is begun.
170 exec_postrun=x Run 'x' after job io has finished.
171 ioscheduler=x Use ioscheduler 'x' for this job.
172 cpuload=x For a CPU io thread, percentage of CPU time to attempt
174 cpuchunks=x Split burn cycles into pieces of x.
180 Fio was written by Jens Axboe <axboe@kernel.dk> to enable flexible testing
181 of the Linux IO subsystem and schedulers. He got tired of writing
182 specific test applications to simulate a given workload, and found that
183 the existing io benchmark/test tools out there weren't flexible enough
184 to do what he wanted.
186 Jens Axboe <axboe@kernel.dk> 20060905