[PATCH] fsx example
[fio.git] / HOWTO
CommitLineData
71bfa161
JA
1Table of contents
2-----------------
3
41. Overview
52. How fio works
63. Running fio
74. Job file format
85. Detailed list of parameters
96. Normal output
107. Terse output
11
12
131.0 Overview and history
14------------------------
15fio was originally written to save me the hassle of writing special test
16case programs when I wanted to test a specific workload, either for
17performance reasons or to find/reproduce a bug. The process of writing
18such a test app can be tiresome, especially if you have to do it often.
19Hence I needed a tool that would be able to simulate a given io workload
20without resorting to writing a tailored test case again and again.
21
22A test work load is difficult to define, though. There can be any number
23of processes or threads involved, and they can each be using their own
24way of generating io. You could have someone dirtying large amounts of
25memory in an memory mapped file, or maybe several threads issuing
26reads using asynchronous io. fio needed to be flexible enough to
27simulate both of these cases, and many more.
28
292.0 How fio works
30-----------------
31The first step in getting fio to simulate a desired io workload, is
32writing a job file describing that specific setup. A job file may contain
33any number of threads and/or files - the typical contents of the job file
34is a global section defining shared parameters, and one or more job
35sections describing the jobs involved. When run, fio parses this file
36and sets everything up as described. If we break down a job from top to
37bottom, it contains the following basic parameters:
38
39 IO type Defines the io pattern issued to the file(s).
40 We may only be reading sequentially from this
41 file(s), or we may be writing randomly. Or even
42 mixing reads and writes, sequentially or randomly.
43
44 Block size In how large chunks are we issuing io? This may be
45 a single value, or it may describe a range of
46 block sizes.
47
48 IO size How much data are we going to be reading/writing.
49
50 IO engine How do we issue io? We could be memory mapping the
51 file, we could be using regular read/write, we
52 could be using splice, async io, or even
53 SG (SCSI generic sg).
54
55 IO depth If the io engine is async, how large a queueing
56 depth do we want to maintain?
57
58 IO type Should we be doing buffered io, or direct/raw io?
59
60 Num files How many files are we spreading the workload over.
61
62 Num threads How many threads or processes should we spread
63 this workload over.
64
65The above are the basic parameters defined for a workload, in addition
66there's a multitude of parameters that modify other aspects of how this
67job behaves.
68
69
703.0 Running fio
71---------------
72See the README file for command line parameters, there are only a few
73of them.
74
75Running fio is normally the easiest part - you just give it the job file
76(or job files) as parameters:
77
78$ fio job_file
79
80and it will start doing what the job_file tells it to do. You can give
81more than one job file on the command line, fio will serialize the running
82of those files. Internally that is the same as using the 'stonewall'
83parameter described the the parameter section.
84
b4692828
JA
85If the job file contains only one job, you may as well just give the
86parameters on the command line. The command line parameters are identical
87to the job parameters, with a few extra that control global parameters
88(see README). For example, for the job file parameter iodepth=2, the
c2b1e753
JA
89mirror command line option would be --iodepth 2 or --iodepth=2. You can
90also use the command line for giving more than one job entry. For each
91--name option that fio sees, it will start a new job with that name.
92Command line entries following a --name entry will apply to that job,
93until there are no more entries or a new --name entry is seen. This is
94similar to the job file options, where each option applies to the current
95job until a new [] job entry is seen.
b4692828 96
71bfa161
JA
97fio does not need to run as root, except if the files or devices specified
98in the job section requires that. Some other options may also be restricted,
99such as memory locking, io scheduler switching, and descreasing the nice value.
100
101
1024.0 Job file format
103-------------------
104As previously described, fio accepts one or more job files describing
105what it is supposed to do. The job file format is the classic ini file,
106where the names enclosed in [] brackets define the job name. You are free
107to use any ascii name you want, except 'global' which has special meaning.
108A global section sets defaults for the jobs described in that file. A job
109may override a global section parameter, and a job file may even have
110several global sections if so desired. A job is only affected by a global
111section residing above it. If the first character in a line is a ';', the
112entire line is discarded as a comment.
113
114So lets look at a really simple job file that define to threads, each
115randomly reading from a 128MiB file.
116
117; -- start job file --
118[global]
119rw=randread
120size=128m
121
122[job1]
123
124[job2]
125
126; -- end job file --
127
128As you can see, the job file sections themselves are empty as all the
129described parameters are shared. As no filename= option is given, fio
c2b1e753
JA
130makes up a filename for each of the jobs as it sees fit. On the command
131line, this job would look as follows:
132
133$ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
134
71bfa161
JA
135
136Lets look at an example that have a number of processes writing randomly
137to files.
138
139; -- start job file --
140[random-writers]
141ioengine=libaio
142iodepth=4
143rw=randwrite
144bs=32k
145direct=0
146size=64m
147numjobs=4
148
149; -- end job file --
150
151Here we have no global section, as we only have one job defined anyway.
152We want to use async io here, with a depth of 4 for each file. We also
153increased the buffer size used to 32KiB and define numjobs to 4 to
154fork 4 identical jobs. The result is 4 processes each randomly writing
b4692828
JA
155to their own 64MiB file. Instead of using the above job file, you could
156have given the parameters on the command line. For this case, you would
157specify:
158
159$ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
71bfa161
JA
160
161fio ships with a few example job files, you can also look there for
162inspiration.
163
164
1655.0 Detailed list of parameters
166-------------------------------
167
168This section describes in details each parameter associated with a job.
169Some parameters take an option of a given type, such as an integer or
170a string. The following types are used:
171
172str String. This is a sequence of alpha characters.
173int Integer. A whole number value, may be negative.
174siint SI integer. A whole number value, which may contain a postfix
175 describing the base of the number. Accepted postfixes are k/m/g,
176 meaning kilo, mega, and giga. So if you want to specifiy 4096,
177 you could either write out '4096' or just give 4k. The postfixes
178 signify base 2 values, so 1024 is 1k and 1024k is 1m and so on.
179bool Boolean. Usually parsed as an integer, however only defined for
180 true and false (1 and 0).
181irange Integer range with postfix. Allows value range to be given, such
182 as 1024-4096. Also see siint.
183
184With the above in mind, here follows the complete list of fio job
185parameters.
186
187name=str ASCII name of the job. This may be used to override the
188 name printed by fio for this job. Otherwise the job
c2b1e753
JA
189 name is used. On the command line this parameter has the
190 special purpose of also signalling the start of a new
191 job.
71bfa161
JA
192
193directory=str Prefix filenames with this directory. Used to places files
194 in a different location than "./".
195
196filename=str Fio normally makes up a filename based on the job name,
197 thread number, and file number. If you want to share
198 files between threads in a job or several jobs, specify
199 a filename for each of them to override the default.
200
201rw=str Type of io pattern. Accepted values are:
202
203 read Sequential reads
204 write Sequential writes
205 randwrite Random writes
206 randread Random reads
207 rw Sequential mixed reads and writes
208 randrw Random mixed reads and writes
209
210 For the mixed io types, the default is to split them 50/50.
211 For certain types of io the result may still be skewed a bit,
212 since the speed may be different.
213
214size=siint The total size of file io for this job. This may describe
215 the size of the single file the job uses, or it may be
216 divided between the number of files in the job. If the
217 file already exists, the file size will be adjusted to this
218 size if larger than the current file size. If this parameter
219 is not given and the file exists, the file size will be used.
220
221bs=siint The block size used for the io units. Defaults to 4k.
222
223bsrange=irange Instead of giving a single block size, specify a range
224 and fio will mix the issued io block sizes. The issued
225 io unit will always be a multiple of the minimum value
690adba3
JA
226 given (also see bs_unaligned).
227
228bs_unaligned If this option is given, any byte size value within bsrange
229 may be used as a block range. This typically wont work with
230 direct IO, as that normally requires sector alignment.
71bfa161
JA
231
232nrfiles=int Number of files to use for this job. Defaults to 1.
233
234ioengine=str Defines how the job issues io to the file. The following
235 types are defined:
236
237 sync Basic read(2) or write(2) io. lseek(2) is
238 used to position the io location.
239
240 libaio Linux native asynchronous io.
241
242 posixaio glibc posix asynchronous io.
243
244 mmap File is memory mapped and data copied
245 to/from using memcpy(3).
246
247 splice splice(2) is used to transfer the data and
248 vmsplice(2) to transfer data from user
249 space to the kernel.
250
251 sg SCSI generic sg v3 io. May either be
252 syncrhonous using the SG_IO ioctl, or if
253 the target is an sg character device
254 we use read(2) and write(2) for asynchronous
255 io.
256
257iodepth=int This defines how many io units to keep in flight against
258 the file. The default is 1 for each file defined in this
259 job, can be overridden with a larger value for higher
260 concurrency.
261
262direct=bool If value is true, use non-buffered io. This is usually
263 O_DIRECT. Defaults to true.
264
265offset=siint Start io at the given offset in the file. The data before
266 the given offset will not be touched. This effectively
267 caps the file size at real_size - offset.
268
269fsync=int If writing to a file, issue a sync of the dirty data
270 for every number of blocks given. For example, if you give
271 32 as a parameter, fio will sync the file for every 32
272 writes issued. If fio is using non-buffered io, we may
273 not sync the file. The exception is the sg io engine, which
274 syncronizes the disk cache anyway.
275
276overwrite=bool If writing to a file, setup the file first and do overwrites.
277
278end_fsync=bool If true, fsync file contents when the job exits.
279
280rwmixcycle=int Value in miliseconds describing how often to switch between
281 reads and writes for a mixed workload. The default is
282 500 msecs.
283
284rwmixread=int How large a percentage of the mix should be reads.
285
286rwmixwrite=int How large a percentage of the mix should be writes. If both
287 rwmixread and rwmixwrite is given and the values do not add
288 up to 100%, the latter of the two will be used to override
289 the first.
290
bb8895e0
JA
291norandommap Normally fio will cover every block of the file when doing
292 random IO. If this option is given, fio will just get a
293 new random offset without looking at past io history. This
294 means that some blocks may not be read or written, and that
295 some blocks may be read/written more than once. This option
296 is mutually exclusive with verify= for that reason.
297
71bfa161
JA
298nice=int Run the job with the given nice value. See man nice(2).
299
300prio=int Set the io priority value of this job. Linux limits us to
301 a positive value between 0 and 7, with 0 being the highest.
302 See man ionice(1).
303
304prioclass=int Set the io priority class. See man ionice(1).
305
306thinktime=int Stall the job x microseconds after an io has completed before
307 issuing the next. May be used to simulate processing being
308 done by an application.
309
310rate=int Cap the bandwidth used by this job to this number of KiB/sec.
311
312ratemin=int Tell fio to do whatever it can to maintain at least this
313 bandwidth.
314
315ratecycle=int Average bandwidth for 'rate' and 'ratemin' over this number
316 of miliseconds.
317
318cpumask=int Set the CPU affinity of this job. The parameter given is a
319 bitmask of allowed CPU's the job may run on. See man
320 sched_setaffinity(2).
321
322startdelay=int Start this job the specified number of seconds after fio
323 has started. Only useful if the job file contains several
324 jobs, and you want to delay starting some jobs to a certain
325 time.
326
327timeout=int Tell fio to terminate processing after the specified number
328 of seconds. It can be quite hard to determine for how long
329 a specified job will run, so this parameter is handy to
330 cap the total runtime to a given time.
331
332invalidate=bool Invalidate the buffer/page cache parts for this file prior
333 to starting io. Defaults to true.
334
335sync=bool Use sync io for buffered writes. For the majority of the
336 io engines, this means using O_SYNC.
337
338mem=str Fio can use various types of memory as the io unit buffer.
339 The allowed values are:
340
341 malloc Use memory from malloc(3) as the buffers.
342
343 shm Use shared memory as the buffers. Allocated
344 through shmget(2).
345
346 mmap Use anonymous memory maps as the buffers.
347 Allocated through mmap(2).
348
349 The area allocated is a function of the maximum allowed
350 bs size for the job, multiplied by the io depth given.
351
352exitall When one job finishes, terminate the rest. The default is
353 to wait for each job to finish, sometimes that is not the
354 desired action.
355
356bwavgtime=int Average the calculated bandwidth over the given time. Value
357 is specified in miliseconds.
358
359create_serialize=bool If true, serialize the file creating for the jobs.
360 This may be handy to avoid interleaving of data
361 files, which may greatly depend on the filesystem
362 used and even the number of processors in the system.
363
364create_fsync=bool fsync the data file after creation. This is the
365 default.
366
367unlink Unlink the job files when done. fio defaults to doing this,
368 if it created the file itself.
369
370loops=int Run the specified number of iterations of this job. Used
371 to repeat the same workload a given number of times. Defaults
372 to 1.
373
374verify=str If writing to a file, fio can verify the file contents
375 after each iteration of the job. The allowed values are:
376
377 md5 Use an md5 sum of the data area and store
378 it in the header of each block.
379
380 crc32 Use a crc32 sum of the data area and store
381 it in the header of each block.
382
383 This option can be used for repeated burnin tests of a
384 system to make sure that the written data is also
385 correctly read back.
386
387stonewall Wait for preceeding jobs in the job file to exit, before
388 starting this one. Can be used to insert serialization
389 points in the job file.
390
391numjobs=int Create the specified number of clones of this job. May be
392 used to setup a larger number of threads/processes doing
393 the same thing.
394
395thread fio defaults to forking jobs, however if this option is
396 given, fio will use pthread_create(3) to create threads
397 instead.
398
399zonesize=siint Divide a file into zones of the specified size. See zoneskip.
400
401zoneskip=siint Skip the specified number of bytes when zonesize data has
402 been read. The two zone options can be used to only do
403 io on zones of a file.
404
076efc7c
JA
405write_iolog=str Write the issued io patterns to the specified file. See
406 read_iolog.
71bfa161 407
076efc7c 408read_iolog=str Open an iolog with the specified file name and replay the
71bfa161
JA
409 io patterns it contains. This can be used to store a
410 workload and replay it sometime later.
411
412write_bw_log If given, write a bandwidth log of the jobs in this job
413 file. Can be used to store data of the bandwidth of the
e0da9bc2
JA
414 jobs in their lifetime. The included fio_generate_plots
415 script uses gnuplot to turn these text files into nice
416 graphs.
71bfa161
JA
417
418write_lat_log Same as write_bw_log, except that this option stores io
419 completion latencies instead.
420
421lockmem=siint Pin down the specified amount of memory with mlock(2). Can
422 potentially be used instead of removing memory or booting
423 with less memory to simulate a smaller amount of memory.
424
425exec_prerun=str Before running this job, issue the command specified
426 through system(3).
427
428exec_postrun=str After the job completes, issue the command specified
429 though system(3).
430
431ioscheduler=str Attempt to switch the device hosting the file to the specified
432 io scheduler before running.
433
434cpuload=int If the job is a CPU cycle eater, attempt to use the specified
435 percentage of CPU cycles.
436
437cpuchunks=int If the job is a CPU cycle eater, split the load into
438 cycles of the given time. In miliseconds.
439
440
4416.0 Interpreting the output
442---------------------------
443
444fio spits out a lot of output. While running, fio will display the
445status of the jobs created. An example of that would be:
446
447Threads running: 1: [_r] [24.79% done] [eta 00h:01m:31s]
448
449The characters inside the square brackets denote the current status of
450each thread. The possible values (in typical life cycle order) are:
451
452Idle Run
453---- ---
454P Thread setup, but not started.
455C Thread created.
456I Thread initialized, waiting.
457 R Running, doing sequential reads.
458 r Running, doing random reads.
459 W Running, doing sequential writes.
460 w Running, doing random writes.
461 M Running, doing mixed sequential reads/writes.
462 m Running, doing mixed random reads/writes.
463 F Running, currently waiting for fsync()
464V Running, doing verification of written data.
465E Thread exited, not reaped by main thread yet.
466_ Thread reaped.
467
468The other values are fairly self explanatory - number of threads
469currently running and doing io, and the estimated completion percentage
470and time for the running group. It's impossible to estimate runtime
471of the following groups (if any).
472
473When fio is done (or interrupted by ctrl-c), it will show the data for
474each thread, group of threads, and disks in that order. For each data
475direction, the output looks like:
476
477Client1 (g=0): err= 0:
478 write: io= 32MiB, bw= 666KiB/s, runt= 50320msec
479 slat (msec): min= 0, max= 136, avg= 0.03, dev= 1.92
480 clat (msec): min= 0, max= 631, avg=48.50, dev=86.82
481 bw (KiB/s) : min= 0, max= 1196, per=51.00%, avg=664.02, dev=681.68
482 cpu : usr=1.49%, sys=0.25%, ctx=7969
483
484The client number is printed, along with the group id and error of that
485thread. Below is the io statistics, here for writes. In the order listed,
486they denote:
487
488io= Number of megabytes io performed
489bw= Average bandwidth rate
490runt= The runtime of that thread
491 slat= Submission latency (avg being the average, dev being the
492 standard deviation). This is the time it took to submit
493 the io. For sync io, the slat is really the completion
494 latency, since queue/complete is one operation there.
495 clat= Completion latency. Same names as slat, this denotes the
496 time from submission to completion of the io pieces. For
497 sync io, clat will usually be equal (or very close) to 0,
498 as the time from submit to complete is basically just
499 CPU time (io has already been done, see slat explanation).
500 bw= Bandwidth. Same names as the xlat stats, but also includes
501 an approximate percentage of total aggregate bandwidth
502 this thread received in this group. This last value is
503 only really useful if the threads in this group are on the
504 same disk, since they are then competing for disk access.
505cpu= CPU usage. User and system time, along with the number
506 of context switches this thread went through.
507
508After each client has been listed, the group statistics are printed. They
509will look like this:
510
511Run status group 0 (all jobs):
512 READ: io=64MiB, aggrb=22178, minb=11355, maxb=11814, mint=2840msec, maxt=2955msec
513 WRITE: io=64MiB, aggrb=1302, minb=666, maxb=669, mint=50093msec, maxt=50320msec
514
515For each data direction, it prints:
516
517io= Number of megabytes io performed.
518aggrb= Aggregate bandwidth of threads in this group.
519minb= The minimum average bandwidth a thread saw.
520maxb= The maximum average bandwidth a thread saw.
521mint= The smallest runtime of the threads in that group.
522maxt= The longest runtime of the threads in that group.
523
524And finally, the disk statistics are printed. They will look like this:
525
526Disk stats (read/write):
527 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
528
529Each value is printed for both reads and writes, with reads first. The
530numbers denote:
531
532ios= Number of ios performed by all groups.
533merge= Number of merges io the io scheduler.
534ticks= Number of ticks we kept the disk busy.
535io_queue= Total time spent in the disk queue.
536util= The disk utilization. A value of 100% means we kept the disk
537 busy constantly, 50% would be a disk idling half of the time.
538
539
5407.0 Terse output
541----------------
542
543For scripted usage where you typically want to generate tables or graphs
544of the results, fio can output the results in a comma seperated format.
545The format is one long line of values, such as:
546
547client1,0,0,936,331,2894,0,0,0.000000,0.000000,1,170,22.115385,34.290410,16,714,84.252874%,366.500000,566.417819,3496,1237,2894,0,0,0.000000,0.000000,0,246,6.671625,21.436952,0,2534,55.465300%,1406.600000,2008.044216,0.000000%,0.431928%,1109
548
549Split up, the format is as follows:
550
551 jobname, groupid, error
552 READ status:
553 KiB IO, bandwidth (KiB/sec), runtime (msec)
554 Submission latency: min, max, mean, deviation
555 Completion latency: min, max, mean, deviation
556 Bw: min, max, aggreate percentage of total, mean, deviation
557 WRITE status:
558 KiB IO, bandwidth (KiB/sec), runtime (msec)
559 Submission latency: min, max, mean, deviation
560 Completion latency: min, max, mean, deviation
561 Bw: min, max, aggreate percentage of total, mean, deviation
562 CPU usage: user, system, context switches
563