8 5. Detailed list of parameters
13 1.0 Overview and history
14 ------------------------
15 fio was originally written to save me the hassle of writing special test
16 case programs when I wanted to test a specific workload, either for
17 performance reasons or to find/reproduce a bug. The process of writing
18 such a test app can be tiresome, especially if you have to do it often.
19 Hence I needed a tool that would be able to simulate a given io workload
20 without resorting to writing a tailored test case again and again.
22 A test work load is difficult to define, though. There can be any number
23 of processes or threads involved, and they can each be using their own
24 way of generating io. You could have someone dirtying large amounts of
25 memory in an memory mapped file, or maybe several threads issuing
26 reads using asynchronous io. fio needed to be flexible enough to
27 simulate both of these cases, and many more.
31 The first step in getting fio to simulate a desired io workload, is
32 writing a job file describing that specific setup. A job file may contain
33 any number of threads and/or files - the typical contents of the job file
34 is a global section defining shared parameters, and one or more job
35 sections describing the jobs involved. When run, fio parses this file
36 and sets everything up as described. If we break down a job from top to
37 bottom, it contains the following basic parameters:
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.
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
48 IO size How much data are we going to be reading/writing.
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, syslet, or even
55 IO depth If the io engine is async, how large a queuing
56 depth do we want to maintain?
58 IO type Should we be doing buffered io, or direct/raw io?
60 Num files How many files are we spreading the workload over.
62 Num threads How many threads or processes should we spread
65 The above are the basic parameters defined for a workload, in addition
66 there's a multitude of parameters that modify other aspects of how this
72 See the README file for command line parameters, there are only a few
75 Running fio is normally the easiest part - you just give it the job file
76 (or job files) as parameters:
80 and it will start doing what the job_file tells it to do. You can give
81 more than one job file on the command line, fio will serialize the running
82 of those files. Internally that is the same as using the 'stonewall'
83 parameter described the the parameter section.
85 If the job file contains only one job, you may as well just give the
86 parameters on the command line. The command line parameters are identical
87 to the job parameters, with a few extra that control global parameters
88 (see README). For example, for the job file parameter iodepth=2, the
89 mirror command line option would be --iodepth 2 or --iodepth=2. You can
90 also 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.
92 Command line entries following a --name entry will apply to that job,
93 until there are no more entries or a new --name entry is seen. This is
94 similar to the job file options, where each option applies to the current
95 job until a new [] job entry is seen.
97 fio does not need to run as root, except if the files or devices specified
98 in the job section requires that. Some other options may also be restricted,
99 such as memory locking, io scheduler switching, and decreasing the nice value.
104 As previously described, fio accepts one or more job files describing
105 what it is supposed to do. The job file format is the classic ini file,
106 where the names enclosed in [] brackets define the job name. You are free
107 to use any ascii name you want, except 'global' which has special meaning.
108 A global section sets defaults for the jobs described in that file. A job
109 may override a global section parameter, and a job file may even have
110 several global sections if so desired. A job is only affected by a global
111 section residing above it. If the first character in a line is a ';' or a
112 '#', the entire line is discarded as a comment.
114 So lets look at a really simple job file that define to threads, each
115 randomly reading from a 128MiB file.
117 ; -- start job file --
128 As you can see, the job file sections themselves are empty as all the
129 described parameters are shared. As no filename= option is given, fio
130 makes up a filename for each of the jobs as it sees fit. On the command
131 line, this job would look as follows:
133 $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
136 Lets look at an example that have a number of processes writing randomly
139 ; -- start job file --
151 Here we have no global section, as we only have one job defined anyway.
152 We want to use async io here, with a depth of 4 for each file. We also
153 increased the buffer size used to 32KiB and define numjobs to 4 to
154 fork 4 identical jobs. The result is 4 processes each randomly writing
155 to their own 64MiB file. Instead of using the above job file, you could
156 have given the parameters on the command line. For this case, you would
159 $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
161 fio ships with a few example job files, you can also look there for
165 5.0 Detailed list of parameters
166 -------------------------------
168 This section describes in details each parameter associated with a job.
169 Some parameters take an option of a given type, such as an integer or
170 a string. The following types are used:
172 str String. This is a sequence of alpha characters.
173 int Integer. A whole number value, may be negative.
174 siint 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 specify 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.
179 bool Boolean. Usually parsed as an integer, however only defined for
180 true and false (1 and 0).
181 irange Integer range with postfix. Allows value range to be given, such
182 as 1024-4096. A colon may also be used as the seperator, eg
183 1k:4k. If the option allows two sets of ranges, they can be
184 specified with a ',' or '/' delimiter: 1k-4k/8k-32k. Also see
187 With the above in mind, here follows the complete list of fio job
190 name=str ASCII name of the job. This may be used to override the
191 name printed by fio for this job. Otherwise the job
192 name is used. On the command line this parameter has the
193 special purpose of also signaling the start of a new
196 description=str Text description of the job. Doesn't do anything except
197 dump this text description when this job is run. It's
200 directory=str Prefix filenames with this directory. Used to places files
201 in a different location than "./".
203 filename=str Fio normally makes up a filename based on the job name,
204 thread number, and file number. If you want to share
205 files between threads in a job or several jobs, specify
206 a filename for each of them to override the default. If
207 the ioengine used is 'net', the filename is the host and
208 port to connect to in the format of =host:port.
210 rw=str Type of io pattern. Accepted values are:
212 read Sequential reads
213 write Sequential writes
214 randwrite Random writes
215 randread Random reads
216 rw Sequential mixed reads and writes
217 randrw Random mixed reads and writes
219 For the mixed io types, the default is to split them 50/50.
220 For certain types of io the result may still be skewed a bit,
221 since the speed may be different.
223 randrepeat=bool For random IO workloads, seed the generator in a predictable
224 way so that results are repeatable across repetitions.
226 size=siint The total size of file io for this job. This may describe
227 the size of the single file the job uses, or it may be
228 divided between the number of files in the job. If the
229 file already exists, the file size will be adjusted to this
230 size if larger than the current file size. If this parameter
231 is not given and the file exists, the file size will be used.
233 bs=siint The block size used for the io units. Defaults to 4k. Values
234 can be given for both read and writes. If a single siint is
235 given, it will apply to both. If a second siint is specified
236 after a comma, it will apply to writes only. In other words,
237 the format is either bs=read_and_write or bs=read,write.
238 bs=4k,8k will thus use 4k blocks for reads, and 8k blocks
239 for writes. If you only wish to set the write size, you
240 can do so by passing an empty read size - bs=,8k will set
241 8k for writes and leave the read default value.
243 bsrange=irange Instead of giving a single block size, specify a range
244 and fio will mix the issued io block sizes. The issued
245 io unit will always be a multiple of the minimum value
246 given (also see bs_unaligned). Applies to both reads and
247 writes, however a second range can be given after a comma.
250 bs_unaligned If this option is given, any byte size value within bsrange
251 may be used as a block range. This typically wont work with
252 direct IO, as that normally requires sector alignment.
254 nrfiles=int Number of files to use for this job. Defaults to 1.
256 ioengine=str Defines how the job issues io to the file. The following
259 sync Basic read(2) or write(2) io. lseek(2) is
260 used to position the io location.
262 libaio Linux native asynchronous io.
264 posixaio glibc posix asynchronous io.
266 mmap File is memory mapped and data copied
267 to/from using memcpy(3).
269 splice splice(2) is used to transfer the data and
270 vmsplice(2) to transfer data from user
273 syslet-rw Use the syslet system calls to make
274 regular read/write async.
276 sg SCSI generic sg v3 io. May either be
277 synchronous using the SG_IO ioctl, or if
278 the target is an sg character device
279 we use read(2) and write(2) for asynchronous
282 null Doesn't transfer any data, just pretends
283 to. This is mainly used to exercise fio
284 itself and for debugging/testing purposes.
286 net Transfer over the network to given host:port.
287 'filename' must be set appropriately to
288 filename=host:port regardless of send
289 or receive, if the latter only the port
292 iodepth=int This defines how many io units to keep in flight against
293 the file. The default is 1 for each file defined in this
294 job, can be overridden with a larger value for higher
297 iodepth_batch=int This defines how many pieces of IO to submit at once.
298 It defaults to the same as iodepth, but can be set lower
301 iodepth_low=int The low water mark indicating when to start filling
302 the queue again. Defaults to the same as iodepth, meaning
303 that fio will attempt to keep the queue full at all times.
304 If iodepth is set to eg 16 and iodepth_low is set to 4, then
305 after fio has filled the queue of 16 requests, it will let
306 the depth drain down to 4 before starting to fill it again.
308 direct=bool If value is true, use non-buffered io. This is usually
311 buffered=bool If value is true, use buffered io. This is the opposite
312 of the 'direct' option. Defaults to true.
314 offset=siint Start io at the given offset in the file. The data before
315 the given offset will not be touched. This effectively
316 caps the file size at real_size - offset.
318 fsync=int If writing to a file, issue a sync of the dirty data
319 for every number of blocks given. For example, if you give
320 32 as a parameter, fio will sync the file for every 32
321 writes issued. If fio is using non-buffered io, we may
322 not sync the file. The exception is the sg io engine, which
323 synchronizes the disk cache anyway.
325 overwrite=bool If writing to a file, setup the file first and do overwrites.
327 end_fsync=bool If true, fsync file contents when the job exits.
329 rwmixcycle=int Value in milliseconds describing how often to switch between
330 reads and writes for a mixed workload. The default is
333 rwmixread=int How large a percentage of the mix should be reads.
335 rwmixwrite=int How large a percentage of the mix should be writes. If both
336 rwmixread and rwmixwrite is given and the values do not add
337 up to 100%, the latter of the two will be used to override
340 norandommap Normally fio will cover every block of the file when doing
341 random IO. If this option is given, fio will just get a
342 new random offset without looking at past io history. This
343 means that some blocks may not be read or written, and that
344 some blocks may be read/written more than once. This option
345 is mutually exclusive with verify= for that reason.
347 nice=int Run the job with the given nice value. See man nice(2).
349 prio=int Set the io priority value of this job. Linux limits us to
350 a positive value between 0 and 7, with 0 being the highest.
353 prioclass=int Set the io priority class. See man ionice(1).
355 thinktime=int Stall the job x microseconds after an io has completed before
356 issuing the next. May be used to simulate processing being
357 done by an application. See thinktime_blocks and
361 Only valid if thinktime is set - pretend to spend CPU time
362 doing something with the data received, before falling back
363 to sleeping for the rest of the period specified by
367 Only valid if thinktime is set - control how many blocks
368 to issue, before waiting 'thinktime' usecs. If not set,
369 defaults to 1 which will make fio wait 'thinktime' usecs
372 rate=int Cap the bandwidth used by this job to this number of KiB/sec.
374 ratemin=int Tell fio to do whatever it can to maintain at least this
377 ratecycle=int Average bandwidth for 'rate' and 'ratemin' over this number
380 cpumask=int Set the CPU affinity of this job. The parameter given is a
381 bitmask of allowed CPU's the job may run on. See man
382 sched_setaffinity(2).
384 startdelay=int Start this job the specified number of seconds after fio
385 has started. Only useful if the job file contains several
386 jobs, and you want to delay starting some jobs to a certain
389 runtime=int Tell fio to terminate processing after the specified number
390 of seconds. It can be quite hard to determine for how long
391 a specified job will run, so this parameter is handy to
392 cap the total runtime to a given time.
394 invalidate=bool Invalidate the buffer/page cache parts for this file prior
395 to starting io. Defaults to true.
397 sync=bool Use sync io for buffered writes. For the majority of the
398 io engines, this means using O_SYNC.
400 mem=str Fio can use various types of memory as the io unit buffer.
401 The allowed values are:
403 malloc Use memory from malloc(3) as the buffers.
405 shm Use shared memory as the buffers. Allocated
408 shmhuge Same as shm, but use huge pages as backing.
410 mmap Use mmap to allocate buffers. May either be
411 anonymous memory, or can be file backed if
412 a filename is given after the option. The
413 format is mem=mmap:/path/to/file.
415 mmaphuge Use a memory mapped huge file as the buffer
416 backing. Append filename after mmaphuge, ala
417 mem=mmaphuge:/hugetlbfs/file
419 The area allocated is a function of the maximum allowed
420 bs size for the job, multiplied by the io depth given. Note
421 that for shmhuge and mmaphuge to work, the system must have
422 free huge pages allocated. This can normally be checked
423 and set by reading/writing /proc/sys/vm/nr_hugepages on a
424 Linux system. Fio assumes a huge page is 4MiB in size. So
425 to calculate the number of huge pages you need for a given
426 job file, add up the io depth of all jobs (normally one unless
427 iodepth= is used) and multiply by the maximum bs set. Then
428 divide that number by the huge page size. You can see the
429 size of the huge pages in /proc/meminfo. If no huge pages
430 are allocated by having a non-zero number in nr_hugepages,
431 using mmaphuge or shmhuge will fail. Also see hugepage-size.
433 mmaphuge also needs to have hugetlbfs mounted and the file
434 location should point there. So if it's mounted in /huge,
435 you would use mem=mmaphuge:/huge/somefile.
438 Defines the size of a huge page. Must at least be equal
439 to the system setting, see /proc/meminfo. Defaults to 4MiB.
440 Should probably always be a multiple of megabytes, so using
441 hugepage-size=Xm is the preferred way to set this to avoid
442 setting a non-pow-2 bad value.
444 exitall When one job finishes, terminate the rest. The default is
445 to wait for each job to finish, sometimes that is not the
448 bwavgtime=int Average the calculated bandwidth over the given time. Value
449 is specified in milliseconds.
451 create_serialize=bool If true, serialize the file creating for the jobs.
452 This may be handy to avoid interleaving of data
453 files, which may greatly depend on the filesystem
454 used and even the number of processors in the system.
456 create_fsync=bool fsync the data file after creation. This is the
459 unlink=bool Unlink the job files when done. Not the default, as repeated
460 runs of that job would then waste time recreating the fileset
463 loops=int Run the specified number of iterations of this job. Used
464 to repeat the same workload a given number of times. Defaults
467 verify=str If writing to a file, fio can verify the file contents
468 after each iteration of the job. The allowed values are:
470 md5 Use an md5 sum of the data area and store
471 it in the header of each block.
473 crc32 Use a crc32 sum of the data area and store
474 it in the header of each block.
476 This option can be used for repeated burn-in tests of a
477 system to make sure that the written data is also
480 stonewall Wait for preceeding jobs in the job file to exit, before
481 starting this one. Can be used to insert serialization
482 points in the job file.
484 numjobs=int Create the specified number of clones of this job. May be
485 used to setup a larger number of threads/processes doing
488 thread fio defaults to forking jobs, however if this option is
489 given, fio will use pthread_create(3) to create threads
492 zonesize=siint Divide a file into zones of the specified size. See zoneskip.
494 zoneskip=siint Skip the specified number of bytes when zonesize data has
495 been read. The two zone options can be used to only do
496 io on zones of a file.
498 write_iolog=str Write the issued io patterns to the specified file. See
501 read_iolog=str Open an iolog with the specified file name and replay the
502 io patterns it contains. This can be used to store a
503 workload and replay it sometime later.
505 write_bw_log If given, write a bandwidth log of the jobs in this job
506 file. Can be used to store data of the bandwidth of the
507 jobs in their lifetime. The included fio_generate_plots
508 script uses gnuplot to turn these text files into nice
511 write_lat_log Same as write_bw_log, except that this option stores io
512 completion latencies instead.
514 lockmem=siint Pin down the specified amount of memory with mlock(2). Can
515 potentially be used instead of removing memory or booting
516 with less memory to simulate a smaller amount of memory.
518 exec_prerun=str Before running this job, issue the command specified
521 exec_postrun=str After the job completes, issue the command specified
524 ioscheduler=str Attempt to switch the device hosting the file to the specified
525 io scheduler before running.
527 cpuload=int If the job is a CPU cycle eater, attempt to use the specified
528 percentage of CPU cycles.
530 cpuchunks=int If the job is a CPU cycle eater, split the load into
531 cycles of the given time. In milliseconds.
534 6.0 Interpreting the output
535 ---------------------------
537 fio spits out a lot of output. While running, fio will display the
538 status of the jobs created. An example of that would be:
540 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
542 The characters inside the square brackets denote the current status of
543 each thread. The possible values (in typical life cycle order) are:
547 P Thread setup, but not started.
549 I Thread initialized, waiting.
550 R Running, doing sequential reads.
551 r Running, doing random reads.
552 W Running, doing sequential writes.
553 w Running, doing random writes.
554 M Running, doing mixed sequential reads/writes.
555 m Running, doing mixed random reads/writes.
556 F Running, currently waiting for fsync()
557 V Running, doing verification of written data.
558 E Thread exited, not reaped by main thread yet.
561 The other values are fairly self explanatory - number of threads
562 currently running and doing io, rate of io since last check, and the estimated
563 completion percentage and time for the running group. It's impossible to
564 estimate runtime of the following groups (if any).
566 When fio is done (or interrupted by ctrl-c), it will show the data for
567 each thread, group of threads, and disks in that order. For each data
568 direction, the output looks like:
570 Client1 (g=0): err= 0:
571 write: io= 32MiB, bw= 666KiB/s, runt= 50320msec
572 slat (msec): min= 0, max= 136, avg= 0.03, stdev= 1.92
573 clat (msec): min= 0, max= 631, avg=48.50, stdev=86.82
574 bw (KiB/s) : min= 0, max= 1196, per=51.00%, avg=664.02, stdev=681.68
575 cpu : usr=1.49%, sys=0.25%, ctx=7969
576 IO depths : 1=0.1%, 2=0.3%, 4=0.5%, 8=99.0%, 16=0.0%, 32=0.0%, >32=0.0%
577 lat (msec): 2=1.6%, 4=0.0%, 10=3.2%, 20=12.8%, 50=38.4%, 100=24.8%,
578 lat (msec): 250=15.2%, 500=0.0%, 750=0.0%, 1000=0.0%, >=2048=0.0%
580 The client number is printed, along with the group id and error of that
581 thread. Below is the io statistics, here for writes. In the order listed,
584 io= Number of megabytes io performed
585 bw= Average bandwidth rate
586 runt= The runtime of that thread
587 slat= Submission latency (avg being the average, dev being the
588 standard deviation). This is the time it took to submit
589 the io. For sync io, the slat is really the completion
590 latency, since queue/complete is one operation there.
591 clat= Completion latency. Same names as slat, this denotes the
592 time from submission to completion of the io pieces. For
593 sync io, clat will usually be equal (or very close) to 0,
594 as the time from submit to complete is basically just
595 CPU time (io has already been done, see slat explanation).
596 bw= Bandwidth. Same names as the xlat stats, but also includes
597 an approximate percentage of total aggregate bandwidth
598 this thread received in this group. This last value is
599 only really useful if the threads in this group are on the
600 same disk, since they are then competing for disk access.
601 cpu= CPU usage. User and system time, along with the number
602 of context switches this thread went through.
603 IO depths= The distribution of io depths over the job life time. The
604 numbers are divided into powers of 2, so for example the
605 16= entries includes depths up to that value but higher
606 than the previous entry. In other words, it covers the
608 IO latencies= The distribution of IO completion latencies. This is the
609 time from when IO leaves fio and when it gets completed.
610 The numbers follow the same pattern as the IO depths,
611 meaning that 2=1.6% means that 1.6% of the IO completed
612 within 2 msecs, 20=12.8% means that 12.8% of the IO
613 took more than 10 msecs, but less than (or equal to) 20 msecs.
615 After each client has been listed, the group statistics are printed. They
618 Run status group 0 (all jobs):
619 READ: io=64MiB, aggrb=22178, minb=11355, maxb=11814, mint=2840msec, maxt=2955msec
620 WRITE: io=64MiB, aggrb=1302, minb=666, maxb=669, mint=50093msec, maxt=50320msec
622 For each data direction, it prints:
624 io= Number of megabytes io performed.
625 aggrb= Aggregate bandwidth of threads in this group.
626 minb= The minimum average bandwidth a thread saw.
627 maxb= The maximum average bandwidth a thread saw.
628 mint= The smallest runtime of the threads in that group.
629 maxt= The longest runtime of the threads in that group.
631 And finally, the disk statistics are printed. They will look like this:
633 Disk stats (read/write):
634 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
636 Each value is printed for both reads and writes, with reads first. The
639 ios= Number of ios performed by all groups.
640 merge= Number of merges io the io scheduler.
641 ticks= Number of ticks we kept the disk busy.
642 io_queue= Total time spent in the disk queue.
643 util= The disk utilization. A value of 100% means we kept the disk
644 busy constantly, 50% would be a disk idling half of the time.
650 For scripted usage where you typically want to generate tables or graphs
651 of the results, fio can output the results in a comma separated format.
652 The format is one long line of values, such as:
654 client1,0,0,936,331,2894,0,0,0.000000,0.000000,1,170,22.115385,34.290410,16,714,84.252874%,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
656 Split up, the format is as follows:
658 jobname, groupid, error
660 KiB IO, bandwidth (KiB/sec), runtime (msec)
661 Submission latency: min, max, mean, deviation
662 Completion latency: min, max, mean, deviation
663 Bw: min, max, aggregate percentage of total, mean, deviation
665 KiB IO, bandwidth (KiB/sec), runtime (msec)
666 Submission latency: min, max, mean, deviation
667 Completion latency: min, max, mean, deviation
668 Bw: min, max, aggregate percentage of total, mean, deviation
669 CPU usage: user, system, context switches