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 When inside a corporate firewall, git:// URL sometimes does not work.
20 If git:// does not work, use the http protocol instead:
22 http://git.kernel.dk/fio.git
24 Snapshots are frequently generated and include the git meta data as well.
25 Snapshots can download from:
27 http://brick.kernel.dk/snaps/
29 There are also two official mirrors. Both of these are synced within
30 an hour of commits landing at git.kernel.dk. So if the main repo is
31 down for some reason, either one of those is safe to use:
33 git://git.kernel.org/pub/scm/linux/kernel/git/axboe/fio.git
34 https://git.kernel.org/pub/scm/linux/kernel/git/axboe/fio.git
38 git://github.com/axboe/fio.git
39 https://github.com/axboe/fio.git
46 Starting with Debian "Squeeze", fio packages are part of the official
47 Debian repository. http://packages.debian.org/search?keywords=fio
50 Starting with Ubuntu 10.04 LTS (aka "Lucid Lynx"), fio packages are part
51 of the Ubuntu "universe" repository.
52 http://packages.ubuntu.com/search?keywords=fio
55 Dag Wieƫrs has RPMs for Red Hat related distros, find them here:
56 http://dag.wieers.com/rpm/packages/fio/
59 Mandriva has integrated fio into their package repository, so installing
60 on that distro should be as easy as typing 'urpmi fio'.
63 Packages for Solaris are available from OpenCSW. Install their pkgutil
64 tool (http://www.opencsw.org/get-it/pkgutil/) and then install fio via
68 Bruce Cran <bruce@cran.org.uk> has fio packages for Windows at
69 http://www.bluestop.org/fio/ .
75 The fio project mailing list is meant for anything related to fio including
76 general discussion, bug reporting, questions, and development.
78 An automated mail detailing recent commits is automatically sent to the
79 list at most daily. The list address is fio@vger.kernel.org, subscribe
80 by sending an email to majordomo@vger.kernel.org with
84 in the body of the email. Archives can be found here:
86 http://www.spinics.net/lists/fio/
88 and archives for the old list can be found here:
90 http://maillist.kernel.dk/fio-devel/
96 Just type 'configure', 'make' and 'make install'.
98 Note that GNU make is required. On BSD it's available from devel/gmake;
99 on Solaris it's in the SUNWgmake package. On platforms where GNU make
100 isn't the default, type 'gmake' instead of 'make'.
102 Configure will print the enabled options. Note that on Linux based
103 platforms, the libaio development packages must be installed to use
104 the libaio engine. Depending on distro, it is usually called
105 libaio-devel or libaio-dev.
107 For gfio, gtk 2.18 (or newer), associated glib threads, and cairo are required
108 to be installed. gfio isn't built automatically and can be enabled
109 with a --enable-gfio option to configure.
111 To build FIO with a cross-compiler:
113 $ make CROSS_COMPILE=/path/to/toolchain/prefix
114 Configure will attempt to determine the target platform automatically.
116 It's possible to build fio for ESX as well, use the --esx switch to
123 On Windows, Cygwin (http://www.cygwin.com/) is required in order to
124 build fio. To create an MSI installer package install WiX 3.8 from
125 http://wixtoolset.org and run dobuild.cmd from the
126 os/windows directory.
128 How to compile fio on 64-bit Windows:
130 1. Install Cygwin (http://www.cygwin.com/). Install 'make' and all
131 packages starting with 'mingw64-i686' and 'mingw64-x86_64'.
132 2. Open the Cygwin Terminal.
133 3. Go to the fio directory (source files).
134 4. Run 'make clean && make -j'.
136 To build fio on 32-bit Windows, run './configure --build-32bit-win' before 'make'.
138 It's recommended that once built or installed, fio be run in a Command Prompt
139 or other 'native' console such as console2, since there are known to be display
140 and signal issues when running it under a Cygwin shell
141 (see http://code.google.com/p/mintty/issues/detail?id=56 for details).
148 --debug Enable some debugging options (see below)
149 --parse-only Parse options only, don't start any IO
150 --output Write output to file
151 --runtime Runtime in seconds
152 --bandwidth-log Generate per-job bandwidth logs
153 --minimal Minimal (terse) output
154 --output-format=type Output format (terse,json,normal)
155 --terse-version=type Terse version output format (default 3, or 2 or 4).
156 --version Print version info and exit
157 --help Print this page
158 --cpuclock-test Perform test/validation of CPU clock
159 --crctest[=test] Test speed of checksum functions
160 --cmdhelp=cmd Print command help, "all" for all of them
161 --enghelp=engine Print ioengine help, or list available ioengines
162 --enghelp=engine,cmd Print help for an ioengine cmd
163 --showcmd Turn a job file into command line options
164 --readonly Turn on safety read-only checks, preventing
166 --eta=when When ETA estimate should be printed
167 May be "always", "never" or "auto"
168 --eta-newline=time Force a new line for every 'time' period passed
169 --status-interval=t Force full status dump every 't' period passed
170 --section=name Only run specified section in job file.
171 Multiple sections can be specified.
172 --alloc-size=kb Set smalloc pool to this size in kb (def 1024)
173 --warnings-fatal Fio parser warnings are fatal
174 --max-jobs Maximum number of threads/processes to support
175 --server=args Start backend server. See Client/Server section.
176 --client=host Connect to specified backend(s).
177 --remote-config=file Tell fio server to load this local file
178 --idle-prof=option Report cpu idleness on a system or percpu basis
179 (option=system,percpu) or run unit work
180 calibration only (option=calibrate).
181 --inflate-log=log Inflate and output compressed log
184 Any parameters following the options will be assumed to be job files,
185 unless they match a job file parameter. Multiple job files can be listed
186 and each job file will be regarded as a separate group. fio will stonewall
187 execution between each group.
189 The --readonly option is an extra safety guard to prevent users from
190 accidentally starting a write workload when that is not desired. Fio
191 will only write if rw=write/randwrite/rw/randrw is given. This extra
192 safety net can be used as an extra precaution as --readonly will also
193 enable a write check in the io engine core to prevent writes due to
194 unknown user space bug(s).
196 The --debug option triggers additional logging by fio.
197 Currently, additional logging is available for:
199 process Dump info related to processes
200 file Dump info related to file actions
201 io Dump info related to IO queuing
202 mem Dump info related to memory allocations
203 blktrace Dump info related to blktrace setup
204 verify Dump info related to IO verification
205 all Enable all debug options
206 random Dump info related to random offset generation
207 parse Dump info related to option matching and parsing
208 diskutil Dump info related to disk utilization updates
209 job:x Dump info only related to job number x
210 mutex Dump info only related to mutex up/down ops
211 profile Dump info related to profile extensions
212 time Dump info related to internal time keeping
213 net Dump info related to networking connections
214 rate Dump info related to IO rate switching
215 compress Dump info related to log compress/decompress
216 ? or help Show available debug options.
218 One can specify multiple debug options: e.g. --debug=file,mem will enable
219 file and memory debugging.
221 The --section option allows one to combine related jobs into one file.
222 E.g. one job file could define light, moderate, and heavy sections. Tell fio to
223 run only the "heavy" section by giving --section=heavy command line option.
224 One can also specify the "write" operations in one section and "verify"
225 operation in another section. The --section option only applies to job
226 sections. The reserved 'global' section is always parsed and used.
228 The --alloc-size switch allows one to use a larger pool size for smalloc.
229 If running large jobs with randommap enabled, fio can run out of memory.
230 Smalloc is an internal allocator for shared structures from a fixed size
231 memory pool. The pool size defaults to 1024k and can grow to 128 pools.
233 NOTE: While running .fio_smalloc.* backing store files are visible in /tmp.
239 See the HOWTO file for a complete description of job file syntax and
240 parameters. The --cmdhelp option also lists all options. If used with
241 an option argument, --cmdhelp will detail the given option. The job file
242 format is in the ini style format, as that is easy for the user to review
245 This README contains the terse version. Job files can describe big and
246 complex setups that are not possible with the command line. Job files
247 are a good practice even for simple jobs since the file provides an
248 easily accessed record of the workload and can include comments.
250 See the examples/ directory for inspiration on how to write job files. Note
251 the copyright and license requirements currently apply to examples/ files.
257 Normally fio is invoked as a stand-alone application on the machine
258 where the IO workload should be generated. However, the frontend and
259 backend of fio can be run separately. Ie the fio server can generate
260 an IO workload on the "Device Under Test" while being controlled from
263 Start the server on the machine which has access to the storage DUT:
267 where args defines what fio listens to. The arguments are of the form
268 'type,hostname or IP,port'. 'type' is either 'ip' (or ip4) for TCP/IP v4,
269 'ip6' for TCP/IP v6, or 'sock' for a local unix domain socket.
270 'hostname' is either a hostname or IP address, and 'port' is the port to
271 listen to (only valid for TCP/IP, not a local socket). Some examples:
275 Start a fio server, listening on all interfaces on the default port (8765).
277 2) fio --server=ip:hostname,4444
279 Start a fio server, listening on IP belonging to hostname and on port 4444.
281 3) fio --server=ip6:::1,4444
283 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
285 4) fio --server=,4444
287 Start a fio server, listening on all interfaces on port 4444.
289 5) fio --server=1.2.3.4
291 Start a fio server, listening on IP 1.2.3.4 on the default port.
293 6) fio --server=sock:/tmp/fio.sock
295 Start a fio server, listening on the local socket /tmp/fio.sock.
297 Once a server is running, a "client" can connect to the fio server with:
299 fio --local-args --client=<server> --remote-args <job file(s)>
301 where --local-args are arguments for the client where it is
302 running, 'server' is the connect string, and --remote-args and <job file(s)>
303 are sent to the server. The 'server' string follows the same format as it
304 does on the server side, to allow IP/hostname/socket and port strings.
306 Fio can connect to multiple servers this way:
308 fio --client=<server1> <job file(s)> --client=<server2> <job file(s)>
310 If the job file is located on the fio server, then you can tell the server
311 to load a local file as well. This is done by using --remote-config:
313 fio --client=server --remote-config /path/to/file.fio
315 Then fio will open this local (to the server) job file instead
316 of being passed one from the client.
318 If you have many servers (example: 100 VMs/containers),
319 you can input a pathname of a file containing host IPs/names as the parameter
320 value for the --client option. For example, here is an example "host.list"
321 file containing 2 hostnames:
323 host1.your.dns.domain
324 host2.your.dns.domain
326 The fio command would then be:
328 fio --client=host.list <job file(s)>
330 In this mode, you cannot input server-specific parameters or job files -- all
331 servers receive the same job file.
333 In order to let fio --client runs use a shared filesystem
334 from multiple hosts, fio --client now prepends the IP address of the
335 server to the filename. For example, if fio is using directory /mnt/nfs/fio
336 and is writing filename fileio.tmp, with a --client hostfile containing
337 two hostnames h1 and h2 with IP addresses 192.168.10.120 and 192.168.10.121,
338 then fio will create two files:
340 /mnt/nfs/fio/192.168.10.120.fileio.tmp
341 /mnt/nfs/fio/192.168.10.121.fileio.tmp
347 Fio works on (at least) Linux, Solaris, AIX, HP-UX, OSX, NetBSD, OpenBSD,
348 Windows and FreeBSD. Some features and/or options may only be available on
349 some of the platforms, typically because those features only apply to that
350 platform (like the solarisaio engine, or the splice engine on Linux).
352 Some features are not available on FreeBSD/Solaris even if they could be
353 implemented, I'd be happy to take patches for that. An example of that is
354 disk utility statistics and (I think) huge page support, support for that
355 does exist in FreeBSD/Solaris.
357 Fio uses pthread mutexes for signalling and locking and FreeBSD does not
358 support process shared pthread mutexes. As a result, only threads are
359 supported on FreeBSD. This could be fixed with sysv ipc locking or
360 other locking alternatives.
362 Other *BSD platforms are untested, but fio should work there almost out
363 of the box. Since I don't do test runs or even compiles on those platforms,
364 your mileage may vary. Sending me patches for other platforms is greatly
365 appreciated. There's a lot of value in having the same test/benchmark tool
366 available on all platforms.
368 Note that POSIX aio is not enabled by default on AIX. Messages like these:
370 Symbol resolution failed for /usr/lib/libc.a(posix_aio.o) because:
371 Symbol _posix_kaio_rdwr (number 2) is not exported from dependent module /unix.
373 indicate one needs to enable POSIX aio. Run the following commands as root:
375 # lsdev -C -l posix_aio0
376 posix_aio0 Defined Posix Asynchronous I/O
377 # cfgmgr -l posix_aio0
378 # lsdev -C -l posix_aio0
379 posix_aio0 Available Posix Asynchronous I/O
381 POSIX aio should work now. To make the change permanent:
383 # chdev -l posix_aio0 -P -a autoconfig='available'
390 Fio was written by Jens Axboe <axboe@kernel.dk> to enable flexible testing
391 of the Linux IO subsystem and schedulers. He got tired of writing
392 specific test applications to simulate a given workload, and found that
393 the existing io benchmark/test tools out there weren't flexible enough
394 to do what he wanted.
396 Jens Axboe <axboe@kernel.dk> 20060905