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://git.kernel.dk/fio.git
19 The http protocol also works, path is the same.
21 Snapshots are frequently generated and they include the git meta data as
22 well. You can download them here:
24 http://brick.kernel.dk/snaps/
31 Starting with Debian "Squeeze", fio packages are part of the official
32 Debian repository. http://packages.debian.org/search?keywords=fio
35 Starting with Ubuntu 10.04 LTS (aka "Lucid Lynx"), fio packages are part
36 of the Ubuntu "universe" repository.
37 http://packages.ubuntu.com/search?keywords=fio
40 Pascal Bleser <guru@unixtech.be> has fio RPMs in his repository for SUSE
41 variants, you can find them here:
42 http://linux01.gwdg.de/~pbleser/rpm-navigation.php?cat=System/fio
45 Dag Wieƫrs has RPMs for Red Hat related distros, find them here:
46 http://dag.wieers.com/rpm/packages/fio/
49 Mandriva has integrated fio into their package repository, so installing
50 on that distro should be as easy as typing 'urpmi fio'.
53 Packages for Solaris are available from OpenCSW. Install their pkgutil
54 tool (http://www.opencsw.org/get-it/pkgutil/) and then install fio via
58 Bruce Cran <bruce@cran.org.uk> has fio packages for Windows at
59 http://www.bluestop.org/fio .
65 There's a mailing list associated with fio. It's meant for general
66 discussion, bug reporting, questions, and development - basically anything
67 that has to do with fio. An automated mail detailing recent commits is
68 automatically sent to the list at most daily. The list address is
69 fio@vger.kernel.org, subscribe by sending an email to
70 majordomo@vger.kernel.org with
74 in the body of the email. Archives can be found here:
76 http://www.spinics.net/lists/fio/
78 and archives for the old list can be found here:
80 http://maillist.kernel.dk/fio-devel/
86 Just type 'make' and 'make install'. If on BSD, for now you have to
87 specify the BSD Makefile with -f and use gmake (not make), eg:
89 $ gmake -f Makefile.FreeBSD && gmake -f Makefile.FreeBSD install
93 $ gmake -f Makefile.aix && gmake -f Makefile.aix install
95 Likewise with OpenSolaris, use the Makefile.solaris to compile there.
96 The OpenSolaris make should work fine. This might change in the
97 future if I opt for an autoconf type setup.
99 If your compile fails with an error like this:
102 In file included from fio.h:23,
104 os/os.h:15:20: error: libaio.h: No such file or directory
105 In file included from gettime.c:8:
106 fio.h:119: error: field 'iocb' has incomplete type
107 make: *** [gettime.o] Error 1
109 Check that you have the libaio development package installed. On RPM
110 based distros, it's typically called libaio-devel.
116 On Windows Cygwin (http://www.cygwin.com) is required with at least
117 devel/gcc4 and devel/make installed in order to build fio, and
118 admin/cygrunsrv to run it. You can also install devel/git to fetch/update
119 the source files. To create an MSI installer package, install WiX 3.6 from
120 http://wix.sourceforge.net/releases/ and run dobuild.cmd from the
121 os/windows directory.
123 Before running fio you'll need to have a copy of cygserver running. Run
124 "/usr/bin/cygserver-config" from an elevated Cygwin shell (i.e. launch the
125 Cygwin shell under the Administrator account) to configure it. Once
126 configured, run "net start cygserver" to start it, or type
127 "/usr/sbin/cygserver &" in the Cygwin shell to start a local copy.
129 If fio exits with the message "Bad system call" it normally means that
130 Cygserver isn't running.
137 --debug Enable some debugging options (see below)
138 --output Write output to file
139 --timeout Runtime in seconds
140 --latency-log Generate per-job latency logs
141 --bandwidth-log Generate per-job bandwidth logs
142 --minimal Minimal (terse) output
143 --version Print version info and exit
144 --help Print this page
145 --cmdhelp=cmd Print command help, "all" for all of them
146 --showcmd Turn a job file into command line options
147 --readonly Turn on safety read-only checks, preventing writes
148 --eta=when When ETA estimate should be printed
149 May be "always", "never" or "auto"
150 --section=name Only run specified section in job file
151 --alloc-size=kb Set smalloc pool to this size in kb (def 1024)
152 --warnings-fatal Fio parser warnings are fatal
155 Any parameters following the options will be assumed to be job files,
156 unless they match a job file parameter. You can add as many as you want,
157 each job file will be regarded as a separate group and fio will stonewall
160 The --readonly switch is an extra safety guard to prevent accidentally
161 turning on a write setting when that is not desired. Fio will only write
162 if rw=write/randwrite/rw/randrw is given, but this extra safety net can
163 be used as an extra precaution. It will also enable a write check in the
164 io engine core to prevent an accidental write due to a fio bug.
166 The debug switch allows adding options that trigger certain logging
167 options in fio. Currently the options are:
169 process Dump info related to processes
170 file Dump info related to file actions
171 io Dump info related to IO queuing
172 mem Dump info related to memory allocations
173 blktrace Dump info related to blktrace setup
174 verify Dump info related to IO verification
175 all Enable all debug options
176 random Dump info related to random offset generation
177 parse Dump info related to option matching and parsing
178 diskutil Dump info related to disk utilization updates
179 job:x Dump info only related to job number x
180 mutex Dump info only related to mutex up/down ops
181 profile Dump info related to profile extensions
182 time Dump info related to internal time keeping
183 ? or help Show available debug options.
185 You can specify as many as you want, eg --debug=file,mem will enable
186 file and memory debugging.
188 The section switch is meant to make it easier to ship a bigger job file
189 instead of several smaller ones. Say you define a job file with light,
190 moderate, and heavy parts. Then you can ask fio to run the given part
191 only by giving it a --section=heavy command line option. The section
192 option only applies to job sections, the reserved 'global' section is
193 always parsed and taken into account.
195 Fio has an internal allocator for shared memory called smalloc. It
196 allocates shared structures from this pool. The pool defaults to 1024k
197 in size, and can grow to 128 pools. If running large jobs with randommap
198 enabled it can run out of memory, in which case the --alloc-size switch
199 is handy for starting with a larger pool size. The backing store is
200 files in /tmp. Fio cleans up after itself, while it is running you
201 may see .fio_smalloc.* files in /tmp.
207 See the HOWTO file for a more detailed description of parameters and what
208 they mean. This file contains the terse version. You can describe big and
209 complex setups with the command line, but generally it's a lot easier to
210 just write a simple job file to describe the workload. The job file format
211 is in the ini style format, as that is easy to read and write for the user.
213 The job file parameters are:
215 name=x Use 'x' as the identifier for this job.
216 description=x 'x' is a text description of the job.
217 directory=x Use 'x' as the top level directory for storing files
218 filename=x Force the use of 'x' as the filename for all files
219 in this thread. If not given, fio will make up
220 a suitable filename based on the thread and file
222 rw=x 'x' may be: read, randread, write, randwrite,
223 rw (read-write mix), randrw (read-write random mix)
224 rwmixcycle=x Base cycle for switching between read and write
226 rwmixread=x 'x' percentage of rw mix ios will be reads. If
227 rwmixwrite is also given, the last of the two will
228 be used if they don't add up to 100%.
229 rwmixwrite=x 'x' percentage of rw mix ios will be writes. See
231 rand_repeatable=x The sequence of random io blocks can be repeatable
232 across runs, if 'x' is 1.
233 size=x Set file size to x bytes (x string can include k/m/g)
234 ioengine=x 'x' may be: aio/libaio/linuxaio for Linux aio,
235 posixaio for POSIX aio, solarisaio for Solaris
236 native async IO, windowsaio for Windows native async IO,
237 sync for regular read/write io,
238 psync for regular pread/pwrite io, vsync for regular
239 readv/writev (with queuing emulation) mmap for mmap'ed
240 io, syslet-rw for syslet driven read/write, splice for
241 using splice/vmsplice, sg for direct SG_IO io, net
242 for network io, or cpuio for a cycler burner load. sg
243 only works on Linux on SCSI (or SCSI-like devices, such
244 as usb-storage or sata/libata driven) devices. Fio also
245 has a null io engine, which is mainly used for testing
248 iodepth=x For async io, allow 'x' ios in flight
249 overwrite=x If 'x', layout a write file first.
250 nrfiles=x Spread io load over 'x' number of files per job,
252 prio=x Run io at prio X, 0-7 is the kernel allowed range
253 prioclass=x Run io at prio class X
254 bs=x Use 'x' for thread blocksize. May include k/m postfix.
255 bsrange=x-y Mix thread block sizes randomly between x and y. May
256 also include k/m postfix.
257 direct=x 1 for direct IO, 0 for buffered IO
258 thinktime=x "Think" x usec after each io
259 rate=x Throttle rate to x KB/sec
260 ratemin=x Quit if rate of x KB/sec can't be met
261 ratecycle=x ratemin averaged over x msecs
262 cpumask=x Only allow job to run on CPUs defined by mask.
263 cpus_allowed=x Like 'cpumask', but allow text setting of CPU affinity.
264 fsync=x If writing with buffered IO, fsync after every
265 'x' blocks have been written.
266 end_fsync=x If 'x', run fsync() after end-of-job.
267 startdelay=x Start this thread x seconds after startup
268 runtime=x Terminate x seconds after startup. Can include a
269 normal time suffix if not given in seconds, such as
270 'm' for minutes, 'h' for hours, and 'd' for days.
271 offset=x Start io at offset x (x string can include k/m/g)
272 invalidate=x Invalidate page cache for file prior to doing io
273 sync=x Use sync writes if x and writing buffered IO.
274 mem=x If x == malloc, use malloc for buffers. If x == shm,
275 use shared memory for buffers. If x == mmap, use
277 exitall When one thread quits, terminate the others
278 bwavgtime=x Average bandwidth stats over an x msec window.
279 create_serialize=x If 'x', serialize file creation.
280 create_fsync=x If 'x', run fsync() after file creation.
281 unlink If set, unlink files when done.
282 loops=x Run the job 'x' number of times.
283 verify=x If 'x' == md5, use md5 for verifies. If 'x' == crc32,
284 use crc32 for verifies. md5 is 'safer', but crc32 is
285 a lot faster. Only makes sense for writing to a file.
286 For other types of checksumming, see HOWTO.
287 stonewall Wait for preceeding jobs to end before running.
288 numjobs=x Create 'x' similar entries for this job
289 thread Use pthreads instead of forked jobs
291 zoneskip=y Zone options must be paired. If given, the job
292 will skip y bytes for every x read/written. This
293 can be used to gauge hard drive speed over the entire
294 platter, without reading everything. Both x/y can
295 include k/m/g suffix.
296 iolog=x Open and read io pattern from file 'x'. The file must
297 contain one io action per line in the following format:
299 where with rw=0/1 for read/write, and the offset
300 and length entries being in bytes.
301 write_iolog=x Write an iolog to file 'x' in the same format as iolog.
302 The iolog options are exclusive, if both given the
303 read iolog will be performed. Specify a separate file
304 for each job, otherwise the iologs will be interspersed
305 and the file may be corrupt.
306 write_bw_log Write a bandwidth log.
307 write_lat_log Write a latency log.
308 lockmem=x Lock down x amount of memory on the machine, to
309 simulate a machine with less memory available. x can
310 include k/m/g suffix.
311 nice=x Run job at given nice value.
312 exec_prerun=x Run 'x' before job io is begun.
313 exec_postrun=x Run 'x' after job io has finished.
314 ioscheduler=x Use ioscheduler 'x' for this job.
315 cpuload=x For a CPU io thread, percentage of CPU time to attempt
317 cpuchunks=x Split burn cycles into pieces of x usecs.
324 Fio works on (at least) Linux, Solaris, AIX, OSX, NetBSD, Windows and FreeBSD.
325 Some features and/or options may only be available on some of the platforms,
326 typically because those features only apply to that platform (like the
327 solarisaio engine, or the splice engine on Linux).
329 Some features are not available on FreeBSD/Solaris even if they could be
330 implemented, I'd be happy to take patches for that. An example of that is
331 disk utility statistics and (I think) huge page support, support for that
332 does exist in FreeBSD/Solaris.
334 Fio uses pthread mutexes for signalling and locking and FreeBSD does not
335 support process shared pthread mutexes. As a result, only threads are
336 supported on FreeBSD. This could be fixed with sysv ipc locking or
337 other locking alternatives.
339 Other *BSD platforms are untested, but fio should work there almost out
340 of the box. Since I don't do test runs or even compiles on those platforms,
341 your mileage may vary. Sending me patches for other platforms is greatly
342 appreciated. There's a lot of value in having the same test/benchmark tool
343 available on all platforms.
345 Note that POSIX aio is not enabled by default on AIX. If you get messages like:
347 Symbol resolution failed for /usr/lib/libc.a(posix_aio.o) because:
348 Symbol _posix_kaio_rdwr (number 2) is not exported from dependent module /unix.
350 you need to enable POSIX aio. Run the following commands as root:
352 # lsdev -C -l posix_aio0
353 posix_aio0 Defined Posix Asynchronous I/O
354 # cfgmgr -l posix_aio0
355 # lsdev -C -l posix_aio0
356 posix_aio0 Available Posix Asynchronous I/O
358 POSIX aio should work now. To make the change permanent:
360 # chdev -l posix_aio0 -P -a autoconfig='available'
367 Fio was written by Jens Axboe <axboe@kernel.dk> to enable flexible testing
368 of the Linux IO subsystem and schedulers. He got tired of writing
369 specific test applications to simulate a given workload, and found that
370 the existing io benchmark/test tools out there weren't flexible enough
371 to do what he wanted.
373 Jens Axboe <axboe@kernel.dk> 20060905