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 If you are inside a corporate firewall, git:// may not always work for
20 you. In that case you can use the http protocol, path is the same:
22 http://git.kernel.dk/fio.git
24 Snapshots are frequently generated and they include the git meta data as
25 well. You can download them here:
27 http://brick.kernel.dk/snaps/
34 Starting with Debian "Squeeze", fio packages are part of the official
35 Debian repository. http://packages.debian.org/search?keywords=fio
38 Starting with Ubuntu 10.04 LTS (aka "Lucid Lynx"), fio packages are part
39 of the Ubuntu "universe" repository.
40 http://packages.ubuntu.com/search?keywords=fio
43 Pascal Bleser <guru@unixtech.be> has fio RPMs in his repository for SUSE
44 variants, you can find them here:
45 http://linux01.gwdg.de/~pbleser/rpm-navigation.php?cat=System/fio
48 Dag Wieƫrs has RPMs for Red Hat related distros, find them here:
49 http://dag.wieers.com/rpm/packages/fio/
52 Mandriva has integrated fio into their package repository, so installing
53 on that distro should be as easy as typing 'urpmi fio'.
56 Packages for Solaris are available from OpenCSW. Install their pkgutil
57 tool (http://www.opencsw.org/get-it/pkgutil/) and then install fio via
61 Bruce Cran <bruce@cran.org.uk> has fio packages for Windows at
62 http://www.bluestop.org/fio .
68 There's a mailing list associated with fio. It's meant for general
69 discussion, bug reporting, questions, and development - basically anything
70 that has to do with fio. An automated mail detailing recent commits is
71 automatically sent to the list at most daily. The list address is
72 fio@vger.kernel.org, subscribe by sending an email to
73 majordomo@vger.kernel.org with
77 in the body of the email. Archives can be found here:
79 http://www.spinics.net/lists/fio/
81 and archives for the old list can be found here:
83 http://maillist.kernel.dk/fio-devel/
89 Just type 'make' and 'make install'.
91 Note that GNU make is required. On BSD it's available from devel/gmake;
92 on Solaris it's in the SUNWgmake package. On platforms where GNU make
93 isn't the default, type 'gmake' instead of 'make'.
95 If your compile fails with an error like this:
98 In file included from fio.h:23,
100 os/os.h:15:20: error: libaio.h: No such file or directory
101 In file included from gettime.c:8:
102 fio.h:119: error: field 'iocb' has incomplete type
103 make: *** [gettime.o] Error 1
105 Check that you have the libaio development package installed. On RPM
106 based distros, it's typically called libaio-devel.
112 On Windows Cygwin (http://www.cygwin.com) is required with at least
113 devel/gcc4 and devel/make installed in order to build fio, and
114 admin/cygrunsrv to run it. You can also install devel/git to fetch/update
115 the source files. To create an MSI installer package put a copy of Cygwin
116 in os\windows\fio, install WiX 3.6 from http://wix.sourceforge.net/releases/
117 and run dobuild.cmd from the os/windows directory.
119 Before running fio you'll need to have a copy of cygserver running. Run
120 "/usr/bin/cygserver-config" from an elevated Cygwin shell (i.e. launch the
121 Cygwin shell under the Administrator account) to configure it. Once
122 configured, run "net start cygserver" to start it, or type
123 "/usr/sbin/cygserver &" in the Cygwin shell to start a local copy.
125 If fio exits with the message "Bad system call" it normally means that
126 Cygserver isn't running.
133 --debug Enable some debugging options (see below)
134 --output Write output to file
135 --timeout Runtime in seconds
136 --latency-log Generate per-job latency logs
137 --bandwidth-log Generate per-job bandwidth logs
138 --minimal Minimal (terse) output
139 --version Print version info and exit
140 --terse-version=type Terse version output format (default 3, or 2).
141 --help Print this page
142 --cmdhelp=cmd Print command help, "all" for all of them
143 --showcmd Turn a job file into command line options
144 --readonly Turn on safety read-only checks, preventing
146 --eta=when When ETA estimate should be printed
147 May be "always", "never" or "auto"
148 --section=name Only run specified section in job file.
149 Multiple sections can be specified.
150 --alloc-size=kb Set smalloc pool to this size in kb (def 1024)
151 --warnings-fatal Fio parser warnings are fatal
152 --max-jobs Maximum number of threads/processes to support
153 --server=args Start backend server. See Client/Server section.
154 --client=host Connect to specified backend.
157 Any parameters following the options will be assumed to be job files,
158 unless they match a job file parameter. You can add as many as you want,
159 each job file will be regarded as a separate group and fio will stonewall
162 The --readonly switch is an extra safety guard to prevent accidentally
163 turning on a write setting when that is not desired. Fio will only write
164 if rw=write/randwrite/rw/randrw is given, but this extra safety net can
165 be used as an extra precaution. It will also enable a write check in the
166 io engine core to prevent an accidental write due to a fio bug.
168 The debug switch allows adding options that trigger certain logging
169 options in fio. Currently the options are:
171 process Dump info related to processes
172 file Dump info related to file actions
173 io Dump info related to IO queuing
174 mem Dump info related to memory allocations
175 blktrace Dump info related to blktrace setup
176 verify Dump info related to IO verification
177 all Enable all debug options
178 random Dump info related to random offset generation
179 parse Dump info related to option matching and parsing
180 diskutil Dump info related to disk utilization updates
181 job:x Dump info only related to job number x
182 mutex Dump info only related to mutex up/down ops
183 profile Dump info related to profile extensions
184 time Dump info related to internal time keeping
185 ? or help Show available debug options.
187 You can specify as many as you want, eg --debug=file,mem will enable
188 file and memory debugging.
190 The section switch is meant to make it easier to ship a bigger job file
191 instead of several smaller ones. Say you define a job file with light,
192 moderate, and heavy parts. Then you can ask fio to run the given part
193 only by giving it a --section=heavy command line option. The section
194 option only applies to job sections, the reserved 'global' section is
195 always parsed and taken into account.
197 Fio has an internal allocator for shared memory called smalloc. It
198 allocates shared structures from this pool. The pool defaults to 1024k
199 in size, and can grow to 128 pools. If running large jobs with randommap
200 enabled it can run out of memory, in which case the --alloc-size switch
201 is handy for starting with a larger pool size. The backing store is
202 files in /tmp. Fio cleans up after itself, while it is running you
203 may see .fio_smalloc.* files in /tmp.
209 See the HOWTO file for a more detailed description of parameters and what
210 they mean. This file contains the terse version. You can describe big and
211 complex setups with the command line, but generally it's a lot easier to
212 just write a simple job file to describe the workload. The job file format
213 is in the ini style format, as that is easy to read and write for the user.
215 The job file parameters are:
217 name=x Use 'x' as the identifier for this job.
218 description=x 'x' is a text description of the job.
219 directory=x Use 'x' as the top level directory for storing files
220 filename=x Force the use of 'x' as the filename for all files
221 in this thread. If not given, fio will make up
222 a suitable filename based on the thread and file
224 rw=x 'x' may be: read, randread, write, randwrite,
225 rw (read-write mix), randrw (read-write random mix)
226 rwmixcycle=x Base cycle for switching between read and write
228 rwmixread=x 'x' percentage of rw mix ios will be reads. If
229 rwmixwrite is also given, the last of the two will
230 be used if they don't add up to 100%.
231 rwmixwrite=x 'x' percentage of rw mix ios will be writes. See
233 rand_repeatable=x The sequence of random io blocks can be repeatable
234 across runs, if 'x' is 1.
235 size=x Set file size to x bytes (x string can include k/m/g)
236 ioengine=x 'x' may be: aio/libaio/linuxaio for Linux aio,
237 posixaio for POSIX aio, solarisaio for Solaris
238 native async IO, windowsaio for Windows native async IO,
239 sync for regular read/write io,
240 psync for regular pread/pwrite io, vsync for regular
241 readv/writev (with queuing emulation) mmap for mmap'ed
242 io, syslet-rw for syslet driven read/write, splice for
243 using splice/vmsplice, sg for direct SG_IO io, net
244 for network io, or cpuio for a cycler burner load. sg
245 only works on Linux on SCSI (or SCSI-like devices, such
246 as usb-storage or sata/libata driven) devices. Fio also
247 has a null io engine, which is mainly used for testing
250 iodepth=x For async io, allow 'x' ios in flight
251 overwrite=x If 'x', layout a write file first.
252 nrfiles=x Spread io load over 'x' number of files per job,
254 prio=x Run io at prio X, 0-7 is the kernel allowed range
255 prioclass=x Run io at prio class X
256 bs=x Use 'x' for thread blocksize. May include k/m postfix.
257 bsrange=x-y Mix thread block sizes randomly between x and y. May
258 also include k/m postfix.
259 direct=x 1 for direct IO, 0 for buffered IO
260 thinktime=x "Think" x usec after each io
261 rate=x Throttle rate to x KB/sec
262 ratemin=x Quit if rate of x KB/sec can't be met
263 ratecycle=x ratemin averaged over x msecs
264 cpumask=x Only allow job to run on CPUs defined by mask.
265 cpus_allowed=x Like 'cpumask', but allow text setting of CPU affinity.
266 fsync=x If writing with buffered IO, fsync after every
267 'x' blocks have been written.
268 end_fsync=x If 'x', run fsync() after end-of-job.
269 startdelay=x Start this thread x seconds after startup
270 runtime=x Terminate x seconds after startup. Can include a
271 normal time suffix if not given in seconds, such as
272 'm' for minutes, 'h' for hours, and 'd' for days.
273 offset=x Start io at offset x (x string can include k/m/g)
274 invalidate=x Invalidate page cache for file prior to doing io
275 sync=x Use sync writes if x and writing buffered IO.
276 mem=x If x == malloc, use malloc for buffers. If x == shm,
277 use shared memory for buffers. If x == mmap, use
279 exitall When one thread quits, terminate the others
280 bwavgtime=x Average bandwidth stats over an x msec window.
281 create_serialize=x If 'x', serialize file creation.
282 create_fsync=x If 'x', run fsync() after file creation.
283 unlink If set, unlink files when done.
284 loops=x Run the job 'x' number of times.
285 verify=x If 'x' == md5, use md5 for verifies. If 'x' == crc32,
286 use crc32 for verifies. md5 is 'safer', but crc32 is
287 a lot faster. Only makes sense for writing to a file.
288 For other types of checksumming, see HOWTO.
289 stonewall Wait for preceeding jobs to end before running.
290 numjobs=x Create 'x' similar entries for this job
291 thread Use pthreads instead of forked jobs
293 zoneskip=y Zone options must be paired. If given, the job
294 will skip y bytes for every x read/written. This
295 can be used to gauge hard drive speed over the entire
296 platter, without reading everything. Both x/y can
297 include k/m/g suffix.
298 read_iolog=x Open and read io pattern from file 'x'. The file format
299 is described in the HOWTO.
300 write_iolog=x Write an iolog to file 'x' in the same format as iolog.
301 The iolog options are exclusive, if both given the
302 read iolog will be performed. Specify a separate file
303 for each job, otherwise the iologs will be interspersed
304 and the file may be corrupt.
305 write_bw_log Write a bandwidth log.
306 write_lat_log Write a latency log.
307 lockmem=x Lock down x amount of memory on the machine, to
308 simulate a machine with less memory available. x can
309 include k/m/g suffix.
310 nice=x Run job at given nice value.
311 exec_prerun=x Run 'x' before job io is begun.
312 exec_postrun=x Run 'x' after job io has finished.
313 ioscheduler=x Use ioscheduler 'x' for this job.
314 cpuload=x For a CPU io thread, percentage of CPU time to attempt
316 cpuchunks=x Split burn cycles into pieces of x usecs.
323 Normally you would run fio as a stand-alone application on the machine
324 where the IO workload should be generated. However, it is also possible to
325 run the frontend and backend of fio separately. This makes it possible to
326 have a fio server running on the machine(s) where the IO workload should
327 be running, while controlling it from another machine.
329 To start the server, you would do:
333 on that machine, where args defines what fio listens to. The arguments
334 are of the form 'type,hostname or IP,port'. 'type' is either 'ip' (or ip4)
335 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain socket.
336 'hostname' is either a hostname or IP address, and 'port' is the port to
337 listen to (only valid for TCP/IP, not a local socket). Some examples:
341 Start a fio server, listening on all interfaces on the default port (8765).
343 2) fio --server=ip:hostname,4444
345 Start a fio server, listening on IP belonging to hostname and on port 4444.
347 3) fio --server=ip6:::1,4444
349 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
351 4) fio --server=,4444
353 Start a fio server, listening on all interfaces on port 4444.
355 5) fio --server=1.2.3.4
357 Start a fio server, listening on IP 1.2.3.4 on the default port.
359 6) fio --server=sock:/tmp/fio.sock
361 Start a fio server, listening on the local socket /tmp/fio.sock.
363 When a server is running, you can connect to it from a client. The client
366 fio --local-args --client=server --remote-args <job file(s)>
368 where --local-args are arguments that are local to the client where it is
369 running, 'server' is the connect string, and --remote-args and <job file(s)>
370 are sent to the server. The 'server' string follows the same format as it
371 does on the server side, to allow IP/hostname/socket and port strings.
372 You can connect to multiple clients as well, to do that you could run:
374 fio --client=server2 --client=server2 <job file(s)>
380 Fio works on (at least) Linux, Solaris, AIX, HP-UX, OSX, NetBSD, Windows
381 and FreeBSD. Some features and/or options may only be available on some of
382 the platforms, typically because those features only apply to that platform
383 (like the solarisaio engine, or the splice engine on Linux).
385 Some features are not available on FreeBSD/Solaris even if they could be
386 implemented, I'd be happy to take patches for that. An example of that is
387 disk utility statistics and (I think) huge page support, support for that
388 does exist in FreeBSD/Solaris.
390 Fio uses pthread mutexes for signalling and locking and FreeBSD does not
391 support process shared pthread mutexes. As a result, only threads are
392 supported on FreeBSD. This could be fixed with sysv ipc locking or
393 other locking alternatives.
395 Other *BSD platforms are untested, but fio should work there almost out
396 of the box. Since I don't do test runs or even compiles on those platforms,
397 your mileage may vary. Sending me patches for other platforms is greatly
398 appreciated. There's a lot of value in having the same test/benchmark tool
399 available on all platforms.
401 Note that POSIX aio is not enabled by default on AIX. If you get messages like:
403 Symbol resolution failed for /usr/lib/libc.a(posix_aio.o) because:
404 Symbol _posix_kaio_rdwr (number 2) is not exported from dependent module /unix.
406 you need to enable POSIX aio. Run the following commands as root:
408 # lsdev -C -l posix_aio0
409 posix_aio0 Defined Posix Asynchronous I/O
410 # cfgmgr -l posix_aio0
411 # lsdev -C -l posix_aio0
412 posix_aio0 Available Posix Asynchronous I/O
414 POSIX aio should work now. To make the change permanent:
416 # chdev -l posix_aio0 -P -a autoconfig='available'
423 Fio was written by Jens Axboe <axboe@kernel.dk> to enable flexible testing
424 of the Linux IO subsystem and schedulers. He got tired of writing
425 specific test applications to simulate a given workload, and found that
426 the existing io benchmark/test tools out there weren't flexible enough
427 to do what he wanted.
429 Jens Axboe <axboe@kernel.dk> 20060905