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. If the
209 ioengine is file based, you can specify a number of files
210 by seperating the names with a ':' colon. So if you wanted
211 a job to open /dev/sda and /dev/sdb as the two working files,
212 you would use filename=/dev/sda:/dev/sdb
214 rw=str Type of io pattern. Accepted values are:
216 read Sequential reads
217 write Sequential writes
218 randwrite Random writes
219 randread Random reads
220 rw Sequential mixed reads and writes
221 randrw Random mixed reads and writes
223 For the mixed io types, the default is to split them 50/50.
224 For certain types of io the result may still be skewed a bit,
225 since the speed may be different.
227 randrepeat=bool For random IO workloads, seed the generator in a predictable
228 way so that results are repeatable across repetitions.
230 size=siint The total size of file io for this job. This may describe
231 the size of the single file the job uses, or it may be
232 divided between the number of files in the job. If the
233 file already exists, the file size will be adjusted to this
234 size if larger than the current file size. If this parameter
235 is not given and the file exists, the file size will be used.
237 bs=siint The block size used for the io units. Defaults to 4k. Values
238 can be given for both read and writes. If a single siint is
239 given, it will apply to both. If a second siint is specified
240 after a comma, it will apply to writes only. In other words,
241 the format is either bs=read_and_write or bs=read,write.
242 bs=4k,8k will thus use 4k blocks for reads, and 8k blocks
243 for writes. If you only wish to set the write size, you
244 can do so by passing an empty read size - bs=,8k will set
245 8k for writes and leave the read default value.
247 bsrange=irange Instead of giving a single block size, specify a range
248 and fio will mix the issued io block sizes. The issued
249 io unit will always be a multiple of the minimum value
250 given (also see bs_unaligned). Applies to both reads and
251 writes, however a second range can be given after a comma.
254 bs_unaligned If this option is given, any byte size value within bsrange
255 may be used as a block range. This typically wont work with
256 direct IO, as that normally requires sector alignment.
258 nrfiles=int Number of files to use for this job. Defaults to 1.
260 openfiles=int Number of files to keep open at the same time. Defaults to
261 the same as nrfiles, can be set smaller to limit the number
264 file_service_type=str Defines how fio decides which file from a job to
265 service next. The following types are defined:
267 random Just choose a file at random.
269 roundrobin Round robin over open files. This
272 The string can have a number appended, indicating how
273 often to switch to a new file. So if option random:4 is
274 given, fio will switch to a new random file after 4 ios
277 ioengine=str Defines how the job issues io to the file. The following
280 sync Basic read(2) or write(2) io. lseek(2) is
281 used to position the io location.
283 libaio Linux native asynchronous io.
285 posixaio glibc posix asynchronous io.
287 mmap File is memory mapped and data copied
288 to/from using memcpy(3).
290 splice splice(2) is used to transfer the data and
291 vmsplice(2) to transfer data from user
294 syslet-rw Use the syslet system calls to make
295 regular read/write async.
297 sg SCSI generic sg v3 io. May either be
298 synchronous using the SG_IO ioctl, or if
299 the target is an sg character device
300 we use read(2) and write(2) for asynchronous
303 null Doesn't transfer any data, just pretends
304 to. This is mainly used to exercise fio
305 itself and for debugging/testing purposes.
307 net Transfer over the network to given host:port.
308 'filename' must be set appropriately to
309 filename=host:port regardless of send
310 or receive, if the latter only the port
313 cpu Doesn't transfer any data, but burns CPU
314 cycles according to the cpuload= and
315 cpucycle= options. Setting cpuload=85
316 will cause that job to do nothing but burn
319 external Prefix to specify loading an external
320 IO engine object file. Append the engine
321 filename, eg ioengine=external:/tmp/foo.o
322 to load ioengine foo.o in /tmp.
324 iodepth=int This defines how many io units to keep in flight against
325 the file. The default is 1 for each file defined in this
326 job, can be overridden with a larger value for higher
329 iodepth_batch=int This defines how many pieces of IO to submit at once.
330 It defaults to the same as iodepth, but can be set lower
333 iodepth_low=int The low water mark indicating when to start filling
334 the queue again. Defaults to the same as iodepth, meaning
335 that fio will attempt to keep the queue full at all times.
336 If iodepth is set to eg 16 and iodepth_low is set to 4, then
337 after fio has filled the queue of 16 requests, it will let
338 the depth drain down to 4 before starting to fill it again.
340 direct=bool If value is true, use non-buffered io. This is usually
343 buffered=bool If value is true, use buffered io. This is the opposite
344 of the 'direct' option. Defaults to true.
346 offset=siint Start io at the given offset in the file. The data before
347 the given offset will not be touched. This effectively
348 caps the file size at real_size - offset.
350 fsync=int If writing to a file, issue a sync of the dirty data
351 for every number of blocks given. For example, if you give
352 32 as a parameter, fio will sync the file for every 32
353 writes issued. If fio is using non-buffered io, we may
354 not sync the file. The exception is the sg io engine, which
355 synchronizes the disk cache anyway.
357 overwrite=bool If writing to a file, setup the file first and do overwrites.
359 end_fsync=bool If true, fsync file contents when the job exits.
361 rwmixcycle=int Value in milliseconds describing how often to switch between
362 reads and writes for a mixed workload. The default is
365 rwmixread=int How large a percentage of the mix should be reads.
367 rwmixwrite=int How large a percentage of the mix should be writes. If both
368 rwmixread and rwmixwrite is given and the values do not add
369 up to 100%, the latter of the two will be used to override
372 norandommap Normally fio will cover every block of the file when doing
373 random IO. If this option is given, fio will just get a
374 new random offset without looking at past io history. This
375 means that some blocks may not be read or written, and that
376 some blocks may be read/written more than once. This option
377 is mutually exclusive with verify= for that reason.
379 nice=int Run the job with the given nice value. See man nice(2).
381 prio=int Set the io priority value of this job. Linux limits us to
382 a positive value between 0 and 7, with 0 being the highest.
385 prioclass=int Set the io priority class. See man ionice(1).
387 thinktime=int Stall the job x microseconds after an io has completed before
388 issuing the next. May be used to simulate processing being
389 done by an application. See thinktime_blocks and
393 Only valid if thinktime is set - pretend to spend CPU time
394 doing something with the data received, before falling back
395 to sleeping for the rest of the period specified by
399 Only valid if thinktime is set - control how many blocks
400 to issue, before waiting 'thinktime' usecs. If not set,
401 defaults to 1 which will make fio wait 'thinktime' usecs
404 rate=int Cap the bandwidth used by this job to this number of KiB/sec.
406 ratemin=int Tell fio to do whatever it can to maintain at least this
409 ratecycle=int Average bandwidth for 'rate' and 'ratemin' over this number
412 cpumask=int Set the CPU affinity of this job. The parameter given is a
413 bitmask of allowed CPU's the job may run on. See man
414 sched_setaffinity(2).
416 startdelay=int Start this job the specified number of seconds after fio
417 has started. Only useful if the job file contains several
418 jobs, and you want to delay starting some jobs to a certain
421 runtime=int Tell fio to terminate processing after the specified number
422 of seconds. It can be quite hard to determine for how long
423 a specified job will run, so this parameter is handy to
424 cap the total runtime to a given time.
426 invalidate=bool Invalidate the buffer/page cache parts for this file prior
427 to starting io. Defaults to true.
429 sync=bool Use sync io for buffered writes. For the majority of the
430 io engines, this means using O_SYNC.
432 mem=str Fio can use various types of memory as the io unit buffer.
433 The allowed values are:
435 malloc Use memory from malloc(3) as the buffers.
437 shm Use shared memory as the buffers. Allocated
440 shmhuge Same as shm, but use huge pages as backing.
442 mmap Use mmap to allocate buffers. May either be
443 anonymous memory, or can be file backed if
444 a filename is given after the option. The
445 format is mem=mmap:/path/to/file.
447 mmaphuge Use a memory mapped huge file as the buffer
448 backing. Append filename after mmaphuge, ala
449 mem=mmaphuge:/hugetlbfs/file
451 The area allocated is a function of the maximum allowed
452 bs size for the job, multiplied by the io depth given. Note
453 that for shmhuge and mmaphuge to work, the system must have
454 free huge pages allocated. This can normally be checked
455 and set by reading/writing /proc/sys/vm/nr_hugepages on a
456 Linux system. Fio assumes a huge page is 4MiB in size. So
457 to calculate the number of huge pages you need for a given
458 job file, add up the io depth of all jobs (normally one unless
459 iodepth= is used) and multiply by the maximum bs set. Then
460 divide that number by the huge page size. You can see the
461 size of the huge pages in /proc/meminfo. If no huge pages
462 are allocated by having a non-zero number in nr_hugepages,
463 using mmaphuge or shmhuge will fail. Also see hugepage-size.
465 mmaphuge also needs to have hugetlbfs mounted and the file
466 location should point there. So if it's mounted in /huge,
467 you would use mem=mmaphuge:/huge/somefile.
470 Defines the size of a huge page. Must at least be equal
471 to the system setting, see /proc/meminfo. Defaults to 4MiB.
472 Should probably always be a multiple of megabytes, so using
473 hugepage-size=Xm is the preferred way to set this to avoid
474 setting a non-pow-2 bad value.
476 exitall When one job finishes, terminate the rest. The default is
477 to wait for each job to finish, sometimes that is not the
480 bwavgtime=int Average the calculated bandwidth over the given time. Value
481 is specified in milliseconds.
483 create_serialize=bool If true, serialize the file creating for the jobs.
484 This may be handy to avoid interleaving of data
485 files, which may greatly depend on the filesystem
486 used and even the number of processors in the system.
488 create_fsync=bool fsync the data file after creation. This is the
491 unlink=bool Unlink the job files when done. Not the default, as repeated
492 runs of that job would then waste time recreating the fileset
495 loops=int Run the specified number of iterations of this job. Used
496 to repeat the same workload a given number of times. Defaults
499 verify=str If writing to a file, fio can verify the file contents
500 after each iteration of the job. The allowed values are:
502 md5 Use an md5 sum of the data area and store
503 it in the header of each block.
505 crc32 Use a crc32 sum of the data area and store
506 it in the header of each block.
508 This option can be used for repeated burn-in tests of a
509 system to make sure that the written data is also
512 stonewall Wait for preceeding jobs in the job file to exit, before
513 starting this one. Can be used to insert serialization
514 points in the job file.
516 numjobs=int Create the specified number of clones of this job. May be
517 used to setup a larger number of threads/processes doing
518 the same thing. We regard that grouping of jobs as a
521 group_reporting If 'numjobs' is set, it may be interesting to display
522 statistics for the group as a whole instead of for each
523 individual job. This is especially true of 'numjobs' is
524 large, looking at individual thread/process output quickly
525 becomes unwieldy. If 'group_reporting' is specified, fio
526 will show the final report per-group instead of per-job.
528 thread fio defaults to forking jobs, however if this option is
529 given, fio will use pthread_create(3) to create threads
532 zonesize=siint Divide a file into zones of the specified size. See zoneskip.
534 zoneskip=siint Skip the specified number of bytes when zonesize data has
535 been read. The two zone options can be used to only do
536 io on zones of a file.
538 write_iolog=str Write the issued io patterns to the specified file. See
541 read_iolog=str Open an iolog with the specified file name and replay the
542 io patterns it contains. This can be used to store a
543 workload and replay it sometime later.
545 write_bw_log If given, write a bandwidth log of the jobs in this job
546 file. Can be used to store data of the bandwidth of the
547 jobs in their lifetime. The included fio_generate_plots
548 script uses gnuplot to turn these text files into nice
551 write_lat_log Same as write_bw_log, except that this option stores io
552 completion latencies instead.
554 lockmem=siint Pin down the specified amount of memory with mlock(2). Can
555 potentially be used instead of removing memory or booting
556 with less memory to simulate a smaller amount of memory.
558 exec_prerun=str Before running this job, issue the command specified
561 exec_postrun=str After the job completes, issue the command specified
564 ioscheduler=str Attempt to switch the device hosting the file to the specified
565 io scheduler before running.
567 cpuload=int If the job is a CPU cycle eater, attempt to use the specified
568 percentage of CPU cycles.
570 cpuchunks=int If the job is a CPU cycle eater, split the load into
571 cycles of the given time. In milliseconds.
574 6.0 Interpreting the output
575 ---------------------------
577 fio spits out a lot of output. While running, fio will display the
578 status of the jobs created. An example of that would be:
580 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
582 The characters inside the square brackets denote the current status of
583 each thread. The possible values (in typical life cycle order) are:
587 P Thread setup, but not started.
589 I Thread initialized, waiting.
590 R Running, doing sequential reads.
591 r Running, doing random reads.
592 W Running, doing sequential writes.
593 w Running, doing random writes.
594 M Running, doing mixed sequential reads/writes.
595 m Running, doing mixed random reads/writes.
596 F Running, currently waiting for fsync()
597 V Running, doing verification of written data.
598 E Thread exited, not reaped by main thread yet.
601 The other values are fairly self explanatory - number of threads
602 currently running and doing io, rate of io since last check, and the estimated
603 completion percentage and time for the running group. It's impossible to
604 estimate runtime of the following groups (if any).
606 When fio is done (or interrupted by ctrl-c), it will show the data for
607 each thread, group of threads, and disks in that order. For each data
608 direction, the output looks like:
610 Client1 (g=0): err= 0:
611 write: io= 32MiB, bw= 666KiB/s, runt= 50320msec
612 slat (msec): min= 0, max= 136, avg= 0.03, stdev= 1.92
613 clat (msec): min= 0, max= 631, avg=48.50, stdev=86.82
614 bw (KiB/s) : min= 0, max= 1196, per=51.00%, avg=664.02, stdev=681.68
615 cpu : usr=1.49%, sys=0.25%, ctx=7969
616 IO depths : 1=0.1%, 2=0.3%, 4=0.5%, 8=99.0%, 16=0.0%, 32=0.0%, >32=0.0%
617 lat (msec): 2=1.6%, 4=0.0%, 10=3.2%, 20=12.8%, 50=38.4%, 100=24.8%,
618 lat (msec): 250=15.2%, 500=0.0%, 750=0.0%, 1000=0.0%, >=2048=0.0%
620 The client number is printed, along with the group id and error of that
621 thread. Below is the io statistics, here for writes. In the order listed,
624 io= Number of megabytes io performed
625 bw= Average bandwidth rate
626 runt= The runtime of that thread
627 slat= Submission latency (avg being the average, dev being the
628 standard deviation). This is the time it took to submit
629 the io. For sync io, the slat is really the completion
630 latency, since queue/complete is one operation there.
631 clat= Completion latency. Same names as slat, this denotes the
632 time from submission to completion of the io pieces. For
633 sync io, clat will usually be equal (or very close) to 0,
634 as the time from submit to complete is basically just
635 CPU time (io has already been done, see slat explanation).
636 bw= Bandwidth. Same names as the xlat stats, but also includes
637 an approximate percentage of total aggregate bandwidth
638 this thread received in this group. This last value is
639 only really useful if the threads in this group are on the
640 same disk, since they are then competing for disk access.
641 cpu= CPU usage. User and system time, along with the number
642 of context switches this thread went through.
643 IO depths= The distribution of io depths over the job life time. The
644 numbers are divided into powers of 2, so for example the
645 16= entries includes depths up to that value but higher
646 than the previous entry. In other words, it covers the
648 IO latencies= The distribution of IO completion latencies. This is the
649 time from when IO leaves fio and when it gets completed.
650 The numbers follow the same pattern as the IO depths,
651 meaning that 2=1.6% means that 1.6% of the IO completed
652 within 2 msecs, 20=12.8% means that 12.8% of the IO
653 took more than 10 msecs, but less than (or equal to) 20 msecs.
655 After each client has been listed, the group statistics are printed. They
658 Run status group 0 (all jobs):
659 READ: io=64MiB, aggrb=22178, minb=11355, maxb=11814, mint=2840msec, maxt=2955msec
660 WRITE: io=64MiB, aggrb=1302, minb=666, maxb=669, mint=50093msec, maxt=50320msec
662 For each data direction, it prints:
664 io= Number of megabytes io performed.
665 aggrb= Aggregate bandwidth of threads in this group.
666 minb= The minimum average bandwidth a thread saw.
667 maxb= The maximum average bandwidth a thread saw.
668 mint= The smallest runtime of the threads in that group.
669 maxt= The longest runtime of the threads in that group.
671 And finally, the disk statistics are printed. They will look like this:
673 Disk stats (read/write):
674 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
676 Each value is printed for both reads and writes, with reads first. The
679 ios= Number of ios performed by all groups.
680 merge= Number of merges io the io scheduler.
681 ticks= Number of ticks we kept the disk busy.
682 io_queue= Total time spent in the disk queue.
683 util= The disk utilization. A value of 100% means we kept the disk
684 busy constantly, 50% would be a disk idling half of the time.
690 For scripted usage where you typically want to generate tables or graphs
691 of the results, fio can output the results in a semicolon separated format.
692 The format is one long line of values, such as:
694 client1;0;0;1906777;1090804;1790;0;0;0.000000;0.000000;0;0;0.000000;0.000000;929380;1152890;25.510151%;1078276.333333;128948.113404;0;0;0;0;0;0.000000;0.000000;0;0;0.000000;0.000000;0;0;0.000000%;0.000000;0.000000;100.000000%;0.000000%;324;100.0%;0.0%;0.0%;0.0%;0.0%;0.0%;0.0%;100.0%;0.0%;0.0%;0.0%;0.0%;0.0%
695 ;0.0%;0.0%;0.0%;0.0%;0.0%
697 Split up, the format is as follows:
699 jobname, groupid, error
701 KiB IO, bandwidth (KiB/sec), runtime (msec)
702 Submission latency: min, max, mean, deviation
703 Completion latency: min, max, mean, deviation
704 Bw: min, max, aggregate percentage of total, mean, deviation
706 KiB IO, bandwidth (KiB/sec), runtime (msec)
707 Submission latency: min, max, mean, deviation
708 Completion latency: min, max, mean, deviation
709 Bw: min, max, aggregate percentage of total, mean, deviation
710 CPU usage: user, system, context switches
711 IO depths: <=1, 2, 4, 8, 16, 32, >=64
712 IO latencies: <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, >=2000