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 MinGW (http://www.mingw.org/) is required in order to
113 build fio. To create an MSI installer package install WiX 3.6 from
114 http://wix.sourceforge.net/releases/ and run dobuild.cmd from the
115 os/windows directory.
122 --debug Enable some debugging options (see below)
123 --output Write output to file
124 --runtime Runtime in seconds
125 --latency-log Generate per-job latency logs
126 --bandwidth-log Generate per-job bandwidth logs
127 --minimal Minimal (terse) output
128 --output-format=type Output format (terse,json,normal)
129 --terse-version=type Terse version output format (default 3, or 2 or 4).
130 --version Print version info and exit
131 --help Print this page
132 --cmdhelp=cmd Print command help, "all" for all of them
133 --enghelp=engine Print ioengine help, or list available ioengines
134 --enghelp=engine,cmd Print help for an ioengine cmd
135 --showcmd Turn a job file into command line options
136 --readonly Turn on safety read-only checks, preventing
138 --eta=when When ETA estimate should be printed
139 May be "always", "never" or "auto"
140 --section=name Only run specified section in job file.
141 Multiple sections can be specified.
142 --alloc-size=kb Set smalloc pool to this size in kb (def 1024)
143 --warnings-fatal Fio parser warnings are fatal
144 --max-jobs Maximum number of threads/processes to support
145 --server=args Start backend server. See Client/Server section.
146 --client=host Connect to specified backend.
149 Any parameters following the options will be assumed to be job files,
150 unless they match a job file parameter. You can add as many as you want,
151 each job file will be regarded as a separate group and fio will stonewall
154 The --readonly switch is an extra safety guard to prevent accidentally
155 turning on a write setting when that is not desired. Fio will only write
156 if rw=write/randwrite/rw/randrw is given, but this extra safety net can
157 be used as an extra precaution. It will also enable a write check in the
158 io engine core to prevent an accidental write due to a fio bug.
160 The debug switch allows adding options that trigger certain logging
161 options in fio. Currently the options are:
163 process Dump info related to processes
164 file Dump info related to file actions
165 io Dump info related to IO queuing
166 mem Dump info related to memory allocations
167 blktrace Dump info related to blktrace setup
168 verify Dump info related to IO verification
169 all Enable all debug options
170 random Dump info related to random offset generation
171 parse Dump info related to option matching and parsing
172 diskutil Dump info related to disk utilization updates
173 job:x Dump info only related to job number x
174 mutex Dump info only related to mutex up/down ops
175 profile Dump info related to profile extensions
176 time Dump info related to internal time keeping
177 ? or help Show available debug options.
179 You can specify as many as you want, eg --debug=file,mem will enable
180 file and memory debugging.
182 The section switch is meant to make it easier to ship a bigger job file
183 instead of several smaller ones. Say you define a job file with light,
184 moderate, and heavy parts. Then you can ask fio to run the given part
185 only by giving it a --section=heavy command line option. The section
186 option only applies to job sections, the reserved 'global' section is
187 always parsed and taken into account.
189 Fio has an internal allocator for shared memory called smalloc. It
190 allocates shared structures from this pool. The pool defaults to 1024k
191 in size, and can grow to 128 pools. If running large jobs with randommap
192 enabled it can run out of memory, in which case the --alloc-size switch
193 is handy for starting with a larger pool size. The backing store is
194 files in /tmp. Fio cleans up after itself, while it is running you
195 may see .fio_smalloc.* files in /tmp.
201 See the HOWTO file for a more detailed description of parameters and what
202 they mean. This file contains the terse version. You can describe big and
203 complex setups with the command line, but generally it's a lot easier to
204 just write a simple job file to describe the workload. The job file format
205 is in the ini style format, as that is easy to read and write for the user.
207 The job file parameters are:
209 name=x Use 'x' as the identifier for this job.
210 description=x 'x' is a text description of the job.
211 directory=x Use 'x' as the top level directory for storing files
212 filename=x Force the use of 'x' as the filename for all files
213 in this thread. If not given, fio will make up
214 a suitable filename based on the thread and file
216 rw=x 'x' may be: read, randread, write, randwrite,
217 rw (read-write mix), randrw (read-write random mix)
218 rwmixcycle=x Base cycle for switching between read and write
220 rwmixread=x 'x' percentage of rw mix ios will be reads. If
221 rwmixwrite is also given, the last of the two will
222 be used if they don't add up to 100%.
223 rwmixwrite=x 'x' percentage of rw mix ios will be writes. See
225 rand_repeatable=x The sequence of random io blocks can be repeatable
226 across runs, if 'x' is 1.
227 size=x Set file size to x bytes (x string can include k/m/g)
228 ioengine=x 'x' may be: aio/libaio/linuxaio for Linux aio,
229 posixaio for POSIX aio, solarisaio for Solaris
230 native async IO, windowsaio for Windows native async IO,
231 sync for regular read/write io,
232 psync for regular pread/pwrite io, vsync for regular
233 readv/writev (with queuing emulation) mmap for mmap'ed
234 io, syslet-rw for syslet driven read/write, splice for
235 using splice/vmsplice, sg for direct SG_IO io, net
236 for network io, or cpuio for a cycler burner load. sg
237 only works on Linux on SCSI (or SCSI-like devices, such
238 as usb-storage or sata/libata driven) devices. Fio also
239 has a null io engine, which is mainly used for testing
242 iodepth=x For async io, allow 'x' ios in flight
243 overwrite=x If 'x', layout a write file first.
244 nrfiles=x Spread io load over 'x' number of files per job,
246 prio=x Run io at prio X, 0-7 is the kernel allowed range
247 prioclass=x Run io at prio class X
248 bs=x Use 'x' for thread blocksize. May include k/m postfix.
249 bsrange=x-y Mix thread block sizes randomly between x and y. May
250 also include k/m postfix.
251 direct=x 1 for direct IO, 0 for buffered IO
252 thinktime=x "Think" x usec after each io
253 rate=x Throttle rate to x KB/sec
254 ratemin=x Quit if rate of x KB/sec can't be met
255 ratecycle=x ratemin averaged over x msecs
256 cpumask=x Only allow job to run on CPUs defined by mask.
257 cpus_allowed=x Like 'cpumask', but allow text setting of CPU affinity.
258 fsync=x If writing with buffered IO, fsync after every
259 'x' blocks have been written.
260 end_fsync=x If 'x', run fsync() after end-of-job.
261 startdelay=x Start this thread x seconds after startup
262 runtime=x Terminate x seconds after startup. Can include a
263 normal time suffix if not given in seconds, such as
264 'm' for minutes, 'h' for hours, and 'd' for days.
265 offset=x Start io at offset x (x string can include k/m/g)
266 invalidate=x Invalidate page cache for file prior to doing io
267 sync=x Use sync writes if x and writing buffered IO.
268 mem=x If x == malloc, use malloc for buffers. If x == shm,
269 use shared memory for buffers. If x == mmap, use
271 exitall When one thread quits, terminate the others
272 bwavgtime=x Average bandwidth stats over an x msec window.
273 create_serialize=x If 'x', serialize file creation.
274 create_fsync=x If 'x', run fsync() after file creation.
275 unlink If set, unlink files when done.
276 loops=x Run the job 'x' number of times.
277 verify=x If 'x' == md5, use md5 for verifies. If 'x' == crc32,
278 use crc32 for verifies. md5 is 'safer', but crc32 is
279 a lot faster. Only makes sense for writing to a file.
280 For other types of checksumming, see HOWTO.
281 stonewall Wait for preceeding jobs to end before running.
282 numjobs=x Create 'x' similar entries for this job
283 thread Use pthreads instead of forked jobs
285 zoneskip=y Zone options must be paired. If given, the job
286 will skip y bytes for every x read/written. This
287 can be used to gauge hard drive speed over the entire
288 platter, without reading everything. Both x/y can
289 include k/m/g suffix.
290 read_iolog=x Open and read io pattern from file 'x'. The file format
291 is described in the HOWTO.
292 write_iolog=x Write an iolog to file 'x' in the same format as iolog.
293 The iolog options are exclusive, if both given the
294 read iolog will be performed. Specify a separate file
295 for each job, otherwise the iologs will be interspersed
296 and the file may be corrupt.
297 write_bw_log Write a bandwidth log.
298 write_lat_log Write a latency log.
299 lockmem=x Lock down x amount of memory on the machine, to
300 simulate a machine with less memory available. x can
301 include k/m/g suffix.
302 nice=x Run job at given nice value.
303 exec_prerun=x Run 'x' before job io is begun.
304 exec_postrun=x Run 'x' after job io has finished.
305 ioscheduler=x Use ioscheduler 'x' for this job.
306 cpuload=x For a CPU io thread, percentage of CPU time to attempt
308 cpuchunks=x Split burn cycles into pieces of x usecs.
315 Normally you would run fio as a stand-alone application on the machine
316 where the IO workload should be generated. However, it is also possible to
317 run the frontend and backend of fio separately. This makes it possible to
318 have a fio server running on the machine(s) where the IO workload should
319 be running, while controlling it from another machine.
321 To start the server, you would do:
325 on that machine, where args defines what fio listens to. The arguments
326 are of the form 'type,hostname or IP,port'. 'type' is either 'ip' (or ip4)
327 for TCP/IP v4, 'ip6' for TCP/IP v6, or 'sock' for a local unix domain socket.
328 'hostname' is either a hostname or IP address, and 'port' is the port to
329 listen to (only valid for TCP/IP, not a local socket). Some examples:
333 Start a fio server, listening on all interfaces on the default port (8765).
335 2) fio --server=ip:hostname,4444
337 Start a fio server, listening on IP belonging to hostname and on port 4444.
339 3) fio --server=ip6:::1,4444
341 Start a fio server, listening on IPv6 localhost ::1 and on port 4444.
343 4) fio --server=,4444
345 Start a fio server, listening on all interfaces on port 4444.
347 5) fio --server=1.2.3.4
349 Start a fio server, listening on IP 1.2.3.4 on the default port.
351 6) fio --server=sock:/tmp/fio.sock
353 Start a fio server, listening on the local socket /tmp/fio.sock.
355 When a server is running, you can connect to it from a client. The client
358 fio --local-args --client=server --remote-args <job file(s)>
360 where --local-args are arguments that are local to the client where it is
361 running, 'server' is the connect string, and --remote-args and <job file(s)>
362 are sent to the server. The 'server' string follows the same format as it
363 does on the server side, to allow IP/hostname/socket and port strings.
364 You can connect to multiple clients as well, to do that you could run:
366 fio --client=server2 <job file(s)> --client=server2 <job file(s)>
372 Fio works on (at least) Linux, Solaris, AIX, HP-UX, OSX, NetBSD, Windows
373 and FreeBSD. Some features and/or options may only be available on some of
374 the platforms, typically because those features only apply to that platform
375 (like the solarisaio engine, or the splice engine on Linux).
377 Some features are not available on FreeBSD/Solaris even if they could be
378 implemented, I'd be happy to take patches for that. An example of that is
379 disk utility statistics and (I think) huge page support, support for that
380 does exist in FreeBSD/Solaris.
382 Fio uses pthread mutexes for signalling and locking and FreeBSD does not
383 support process shared pthread mutexes. As a result, only threads are
384 supported on FreeBSD. This could be fixed with sysv ipc locking or
385 other locking alternatives.
387 Other *BSD platforms are untested, but fio should work there almost out
388 of the box. Since I don't do test runs or even compiles on those platforms,
389 your mileage may vary. Sending me patches for other platforms is greatly
390 appreciated. There's a lot of value in having the same test/benchmark tool
391 available on all platforms.
393 Note that POSIX aio is not enabled by default on AIX. If you get messages like:
395 Symbol resolution failed for /usr/lib/libc.a(posix_aio.o) because:
396 Symbol _posix_kaio_rdwr (number 2) is not exported from dependent module /unix.
398 you need to enable POSIX aio. Run the following commands as root:
400 # lsdev -C -l posix_aio0
401 posix_aio0 Defined Posix Asynchronous I/O
402 # cfgmgr -l posix_aio0
403 # lsdev -C -l posix_aio0
404 posix_aio0 Available Posix Asynchronous I/O
406 POSIX aio should work now. To make the change permanent:
408 # chdev -l posix_aio0 -P -a autoconfig='available'
415 Fio was written by Jens Axboe <axboe@kernel.dk> to enable flexible testing
416 of the Linux IO subsystem and schedulers. He got tired of writing
417 specific test applications to simulate a given workload, and found that
418 the existing io benchmark/test tools out there weren't flexible enough
419 to do what he wanted.
421 Jens Axboe <axboe@kernel.dk> 20060905