8 5. Detailed list of parameters
12 9. CPU idleness profiling
14 1.0 Overview and history
15 ------------------------
16 fio was originally written to save me the hassle of writing special test
17 case programs when I wanted to test a specific workload, either for
18 performance reasons or to find/reproduce a bug. The process of writing
19 such a test app can be tiresome, especially if you have to do it often.
20 Hence I needed a tool that would be able to simulate a given io workload
21 without resorting to writing a tailored test case again and again.
23 A test work load is difficult to define, though. There can be any number
24 of processes or threads involved, and they can each be using their own
25 way of generating io. You could have someone dirtying large amounts of
26 memory in an memory mapped file, or maybe several threads issuing
27 reads using asynchronous io. fio needed to be flexible enough to
28 simulate both of these cases, and many more.
32 The first step in getting fio to simulate a desired io workload, is
33 writing a job file describing that specific setup. A job file may contain
34 any number of threads and/or files - the typical contents of the job file
35 is a global section defining shared parameters, and one or more job
36 sections describing the jobs involved. When run, fio parses this file
37 and sets everything up as described. If we break down a job from top to
38 bottom, it contains the following basic parameters:
40 IO type Defines the io pattern issued to the file(s).
41 We may only be reading sequentially from this
42 file(s), or we may be writing randomly. Or even
43 mixing reads and writes, sequentially or randomly.
45 Block size In how large chunks are we issuing io? This may be
46 a single value, or it may describe a range of
49 IO size How much data are we going to be reading/writing.
51 IO engine How do we issue io? We could be memory mapping the
52 file, we could be using regular read/write, we
53 could be using splice, async io, syslet, or even
56 IO depth If the io engine is async, how large a queuing
57 depth do we want to maintain?
59 IO type Should we be doing buffered io, or direct/raw io?
61 Num files How many files are we spreading the workload over.
63 Num threads How many threads or processes should we spread
66 The above are the basic parameters defined for a workload, in addition
67 there's a multitude of parameters that modify other aspects of how this
73 See the README file for command line parameters, there are only a few
76 Running fio is normally the easiest part - you just give it the job file
77 (or job files) as parameters:
81 and it will start doing what the job_file tells it to do. You can give
82 more than one job file on the command line, fio will serialize the running
83 of those files. Internally that is the same as using the 'stonewall'
84 parameter described the the parameter section.
86 If the job file contains only one job, you may as well just give the
87 parameters on the command line. The command line parameters are identical
88 to the job parameters, with a few extra that control global parameters
89 (see README). For example, for the job file parameter iodepth=2, the
90 mirror command line option would be --iodepth 2 or --iodepth=2. You can
91 also use the command line for giving more than one job entry. For each
92 --name option that fio sees, it will start a new job with that name.
93 Command line entries following a --name entry will apply to that job,
94 until there are no more entries or a new --name entry is seen. This is
95 similar to the job file options, where each option applies to the current
96 job until a new [] job entry is seen.
98 fio does not need to run as root, except if the files or devices specified
99 in the job section requires that. Some other options may also be restricted,
100 such as memory locking, io scheduler switching, and decreasing the nice value.
105 As previously described, fio accepts one or more job files describing
106 what it is supposed to do. The job file format is the classic ini file,
107 where the names enclosed in [] brackets define the job name. You are free
108 to use any ascii name you want, except 'global' which has special meaning.
109 A global section sets defaults for the jobs described in that file. A job
110 may override a global section parameter, and a job file may even have
111 several global sections if so desired. A job is only affected by a global
112 section residing above it. If the first character in a line is a ';' or a
113 '#', the entire line is discarded as a comment.
115 So let's look at a really simple job file that defines two processes, each
116 randomly reading from a 128MB file.
118 ; -- start job file --
129 As you can see, the job file sections themselves are empty as all the
130 described parameters are shared. As no filename= option is given, fio
131 makes up a filename for each of the jobs as it sees fit. On the command
132 line, this job would look as follows:
134 $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
137 Let's look at an example that has a number of processes writing randomly
140 ; -- start job file --
152 Here we have no global section, as we only have one job defined anyway.
153 We want to use async io here, with a depth of 4 for each file. We also
154 increased the buffer size used to 32KB and define numjobs to 4 to
155 fork 4 identical jobs. The result is 4 processes each randomly writing
156 to their own 64MB file. Instead of using the above job file, you could
157 have given the parameters on the command line. For this case, you would
160 $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
162 4.1 Environment variables
163 -------------------------
165 fio also supports environment variable expansion in job files. Any
166 substring of the form "${VARNAME}" as part of an option value (in other
167 words, on the right of the `='), will be expanded to the value of the
168 environment variable called VARNAME. If no such environment variable
169 is defined, or VARNAME is the empty string, the empty string will be
172 As an example, let's look at a sample fio invocation and job file:
174 $ SIZE=64m NUMJOBS=4 fio jobfile.fio
176 ; -- start job file --
183 This will expand to the following equivalent job file at runtime:
185 ; -- start job file --
192 fio ships with a few example job files, you can also look there for
195 4.2 Reserved keywords
196 ---------------------
198 Additionally, fio has a set of reserved keywords that will be replaced
199 internally with the appropriate value. Those keywords are:
201 $pagesize The architecture page size of the running system
202 $mb_memory Megabytes of total memory in the system
203 $ncpus Number of online available CPUs
205 These can be used on the command line or in the job file, and will be
206 automatically substituted with the current system values when the job
207 is run. Simple math is also supported on these keywords, so you can
208 perform actions like:
212 and get that properly expanded to 8 times the size of memory in the
216 5.0 Detailed list of parameters
217 -------------------------------
219 This section describes in details each parameter associated with a job.
220 Some parameters take an option of a given type, such as an integer or
221 a string. The following types are used:
223 str String. This is a sequence of alpha characters.
224 time Integer with possible time suffix. In seconds unless otherwise
225 specified, use eg 10m for 10 minutes. Accepts s/m/h for seconds,
227 int SI integer. A whole number value, which may contain a suffix
228 describing the base of the number. Accepted suffixes are k/m/g/t/p,
229 meaning kilo, mega, giga, tera, and peta. The suffix is not case
230 sensitive, and you may also include trailing 'b' (eg 'kb' is the same
231 as 'k'). So if you want to specify 4096, you could either write
232 out '4096' or just give 4k. The suffixes signify base 2 values, so
233 1024 is 1k and 1024k is 1m and so on, unless the suffix is explicitly
234 set to a base 10 value using 'kib', 'mib', 'gib', etc. If that is the
235 case, then 1000 is used as the multiplier. This can be handy for
236 disks, since manufacturers generally use base 10 values when listing
237 the capacity of a drive. If the option accepts an upper and lower
238 range, use a colon ':' or minus '-' to separate such values. May also
239 include a prefix to indicate numbers base. If 0x is used, the number
240 is assumed to be hexadecimal. See irange.
241 bool Boolean. Usually parsed as an integer, however only defined for
242 true and false (1 and 0).
243 irange Integer range with suffix. Allows value range to be given, such
244 as 1024-4096. A colon may also be used as the separator, eg
245 1k:4k. If the option allows two sets of ranges, they can be
246 specified with a ',' or '/' delimiter: 1k-4k/8k-32k. Also see
248 float_list A list of floating numbers, separated by a ':' character.
250 With the above in mind, here follows the complete list of fio job
253 name=str ASCII name of the job. This may be used to override the
254 name printed by fio for this job. Otherwise the job
255 name is used. On the command line this parameter has the
256 special purpose of also signaling the start of a new
259 description=str Text description of the job. Doesn't do anything except
260 dump this text description when this job is run. It's
263 directory=str Prefix filenames with this directory. Used to place files
264 in a different location than "./".
266 filename=str Fio normally makes up a filename based on the job name,
267 thread number, and file number. If you want to share
268 files between threads in a job or several jobs, specify
269 a filename for each of them to override the default. If
270 the ioengine used is 'net', the filename is the host, port,
271 and protocol to use in the format of =host,port,protocol.
272 See ioengine=net for more. If the ioengine is file based, you
273 can specify a number of files by separating the names with a
274 ':' colon. So if you wanted a job to open /dev/sda and /dev/sdb
275 as the two working files, you would use
276 filename=/dev/sda:/dev/sdb. On Windows, disk devices are
277 accessed as \\.\PhysicalDrive0 for the first device,
278 \\.\PhysicalDrive1 for the second etc. Note: Windows and
279 FreeBSD prevent write access to areas of the disk containing
280 in-use data (e.g. filesystems).
281 If the wanted filename does need to include a colon, then
282 escape that with a '\' character. For instance, if the filename
283 is "/dev/dsk/foo@3,0:c", then you would use
284 filename="/dev/dsk/foo@3,0\:c". '-' is a reserved name, meaning
285 stdin or stdout. Which of the two depends on the read/write
289 If sharing multiple files between jobs, it is usually necessary
290 to have fio generate the exact names that you want. By default,
291 fio will name a file based on the default file format
292 specification of jobname.jobnumber.filenumber. With this
293 option, that can be customized. Fio will recognize and replace
294 the following keywords in this string:
297 The name of the worker thread or process.
300 The incremental number of the worker thread or
304 The incremental number of the file for that worker
307 To have dependent jobs share a set of files, this option can
308 be set to have fio generate filenames that are shared between
309 the two. For instance, if testfiles.$filenum is specified,
310 file number 4 for any job will be named testfiles.4. The
311 default of $jobname.$jobnum.$filenum will be used if
312 no other format specifier is given.
314 opendir=str Tell fio to recursively add any file it can find in this
315 directory and down the file system tree.
317 lockfile=str Fio defaults to not locking any files before it does
318 IO to them. If a file or file descriptor is shared, fio
319 can serialize IO to that file to make the end result
320 consistent. This is usual for emulating real workloads that
321 share files. The lock modes are:
323 none No locking. The default.
324 exclusive Only one thread/process may do IO,
325 excluding all others.
326 readwrite Read-write locking on the file. Many
327 readers may access the file at the
328 same time, but writes get exclusive
332 rw=str Type of io pattern. Accepted values are:
334 read Sequential reads
335 write Sequential writes
336 randwrite Random writes
337 randread Random reads
338 rw,readwrite Sequential mixed reads and writes
339 randrw Random mixed reads and writes
341 For the mixed io types, the default is to split them 50/50.
342 For certain types of io the result may still be skewed a bit,
343 since the speed may be different. It is possible to specify
344 a number of IO's to do before getting a new offset, this is
345 one by appending a ':<nr>' to the end of the string given.
346 For a random read, it would look like 'rw=randread:8' for
347 passing in an offset modifier with a value of 8. If the
348 suffix is used with a sequential IO pattern, then the value
349 specified will be added to the generated offset for each IO.
350 For instance, using rw=write:4k will skip 4k for every
351 write. It turns sequential IO into sequential IO with holes.
352 See the 'rw_sequencer' option.
354 rw_sequencer=str If an offset modifier is given by appending a number to
355 the rw=<str> line, then this option controls how that
356 number modifies the IO offset being generated. Accepted
359 sequential Generate sequential offset
360 identical Generate the same offset
362 'sequential' is only useful for random IO, where fio would
363 normally generate a new random offset for every IO. If you
364 append eg 8 to randread, you would get a new random offset for
365 every 8 IO's. The result would be a seek for only every 8
366 IO's, instead of for every IO. Use rw=randread:8 to specify
367 that. As sequential IO is already sequential, setting
368 'sequential' for that would not result in any differences.
369 'identical' behaves in a similar fashion, except it sends
370 the same offset 8 number of times before generating a new
373 kb_base=int The base unit for a kilobyte. The defacto base is 2^10, 1024.
374 Storage manufacturers like to use 10^3 or 1000 as a base
375 ten unit instead, for obvious reasons. Allow values are
376 1024 or 1000, with 1024 being the default.
378 unified_rw_reporting=bool Fio normally reports statistics on a per
379 data direction basis, meaning that read, write, and trim are
380 accounted and reported separately. If this option is set,
381 the fio will sum the results and report them as "mixed"
384 randrepeat=bool For random IO workloads, seed the generator in a predictable
385 way so that results are repeatable across repetitions.
387 use_os_rand=bool Fio can either use the random generator supplied by the OS
388 to generator random offsets, or it can use it's own internal
389 generator (based on Tausworthe). Default is to use the
390 internal generator, which is often of better quality and
393 fallocate=str Whether pre-allocation is performed when laying down files.
396 none Do not pre-allocate space
397 posix Pre-allocate via posix_fallocate()
398 keep Pre-allocate via fallocate() with
399 FALLOC_FL_KEEP_SIZE set
400 0 Backward-compatible alias for 'none'
401 1 Backward-compatible alias for 'posix'
403 May not be available on all supported platforms. 'keep' is only
404 available on Linux.If using ZFS on Solaris this must be set to
405 'none' because ZFS doesn't support it. Default: 'posix'.
407 fadvise_hint=bool By default, fio will use fadvise() to advise the kernel
408 on what IO patterns it is likely to issue. Sometimes you
409 want to test specific IO patterns without telling the
410 kernel about it, in which case you can disable this option.
411 If set, fio will use POSIX_FADV_SEQUENTIAL for sequential
412 IO and POSIX_FADV_RANDOM for random IO.
414 size=int The total size of file io for this job. Fio will run until
415 this many bytes has been transferred, unless runtime is
416 limited by other options (such as 'runtime', for instance).
417 Unless specific nrfiles and filesize options are given,
418 fio will divide this size between the available files
419 specified by the job. If not set, fio will use the full
420 size of the given files or devices. If the the files
421 do not exist, size must be given. It is also possible to
422 give size as a percentage between 1 and 100. If size=20%
423 is given, fio will use 20% of the full size of the given
426 filesize=int Individual file sizes. May be a range, in which case fio
427 will select sizes for files at random within the given range
428 and limited to 'size' in total (if that is given). If not
429 given, each created file is the same size.
432 fill_fs=bool Sets size to something really large and waits for ENOSPC (no
433 space left on device) as the terminating condition. Only makes
434 sense with sequential write. For a read workload, the mount
435 point will be filled first then IO started on the result. This
436 option doesn't make sense if operating on a raw device node,
437 since the size of that is already known by the file system.
438 Additionally, writing beyond end-of-device will not return
442 bs=int The block size used for the io units. Defaults to 4k. Values
443 can be given for both read and writes. If a single int is
444 given, it will apply to both. If a second int is specified
445 after a comma, it will apply to writes only. In other words,
446 the format is either bs=read_and_write or bs=read,write.
447 bs=4k,8k will thus use 4k blocks for reads, and 8k blocks
448 for writes. If you only wish to set the write size, you
449 can do so by passing an empty read size - bs=,8k will set
450 8k for writes and leave the read default value.
453 ba=int At what boundary to align random IO offsets. Defaults to
454 the same as 'blocksize' the minimum blocksize given.
455 Minimum alignment is typically 512b for using direct IO,
456 though it usually depends on the hardware block size. This
457 option is mutually exclusive with using a random map for
458 files, so it will turn off that option.
460 blocksize_range=irange
461 bsrange=irange Instead of giving a single block size, specify a range
462 and fio will mix the issued io block sizes. The issued
463 io unit will always be a multiple of the minimum value
464 given (also see bs_unaligned). Applies to both reads and
465 writes, however a second range can be given after a comma.
468 bssplit=str Sometimes you want even finer grained control of the
469 block sizes issued, not just an even split between them.
470 This option allows you to weight various block sizes,
471 so that you are able to define a specific amount of
472 block sizes issued. The format for this option is:
474 bssplit=blocksize/percentage:blocksize/percentage
476 for as many block sizes as needed. So if you want to define
477 a workload that has 50% 64k blocks, 10% 4k blocks, and
478 40% 32k blocks, you would write:
480 bssplit=4k/10:64k/50:32k/40
482 Ordering does not matter. If the percentage is left blank,
483 fio will fill in the remaining values evenly. So a bssplit
484 option like this one:
486 bssplit=4k/50:1k/:32k/
488 would have 50% 4k ios, and 25% 1k and 32k ios. The percentages
489 always add up to 100, if bssplit is given a range that adds
490 up to more, it will error out.
492 bssplit also supports giving separate splits to reads and
493 writes. The format is identical to what bs= accepts. You
494 have to separate the read and write parts with a comma. So
495 if you want a workload that has 50% 2k reads and 50% 4k reads,
496 while having 90% 4k writes and 10% 8k writes, you would
499 bssplit=2k/50:4k/50,4k/90,8k/10
502 bs_unaligned If this option is given, any byte size value within bsrange
503 may be used as a block range. This typically wont work with
504 direct IO, as that normally requires sector alignment.
506 zero_buffers If this option is given, fio will init the IO buffers to
507 all zeroes. The default is to fill them with random data.
509 refill_buffers If this option is given, fio will refill the IO buffers
510 on every submit. The default is to only fill it at init
511 time and reuse that data. Only makes sense if zero_buffers
512 isn't specified, naturally. If data verification is enabled,
513 refill_buffers is also automatically enabled.
515 scramble_buffers=bool If refill_buffers is too costly and the target is
516 using data deduplication, then setting this option will
517 slightly modify the IO buffer contents to defeat normal
518 de-dupe attempts. This is not enough to defeat more clever
519 block compression attempts, but it will stop naive dedupe of
520 blocks. Default: true.
522 buffer_compress_percentage=int If this is set, then fio will attempt to
523 provide IO buffer content (on WRITEs) that compress to
524 the specified level. Fio does this by providing a mix of
525 random data and zeroes. Note that this is per block size
526 unit, for file/disk wide compression level that matches
527 this setting, you'll also want to set refill_buffers.
529 buffer_compress_chunk=int See buffer_compress_percentage. This
530 setting allows fio to manage how big the ranges of random
531 data and zeroed data is. Without this set, fio will
532 provide buffer_compress_percentage of blocksize random
533 data, followed by the remaining zeroed. With this set
534 to some chunk size smaller than the block size, fio can
535 alternate random and zeroed data throughout the IO
538 nrfiles=int Number of files to use for this job. Defaults to 1.
540 openfiles=int Number of files to keep open at the same time. Defaults to
541 the same as nrfiles, can be set smaller to limit the number
544 file_service_type=str Defines how fio decides which file from a job to
545 service next. The following types are defined:
547 random Just choose a file at random.
549 roundrobin Round robin over open files. This
552 sequential Finish one file before moving on to
553 the next. Multiple files can still be
554 open depending on 'openfiles'.
556 The string can have a number appended, indicating how
557 often to switch to a new file. So if option random:4 is
558 given, fio will switch to a new random file after 4 ios
561 ioengine=str Defines how the job issues io to the file. The following
564 sync Basic read(2) or write(2) io. lseek(2) is
565 used to position the io location.
567 psync Basic pread(2) or pwrite(2) io.
569 vsync Basic readv(2) or writev(2) IO.
571 psyncv Basic preadv(2) or pwritev(2) IO.
573 libaio Linux native asynchronous io. Note that Linux
574 may only support queued behaviour with
575 non-buffered IO (set direct=1 or buffered=0).
576 This engine defines engine specific options.
578 posixaio glibc posix asynchronous io.
580 solarisaio Solaris native asynchronous io.
582 windowsaio Windows native asynchronous io.
584 mmap File is memory mapped and data copied
585 to/from using memcpy(3).
587 splice splice(2) is used to transfer the data and
588 vmsplice(2) to transfer data from user
591 syslet-rw Use the syslet system calls to make
592 regular read/write async.
594 sg SCSI generic sg v3 io. May either be
595 synchronous using the SG_IO ioctl, or if
596 the target is an sg character device
597 we use read(2) and write(2) for asynchronous
600 null Doesn't transfer any data, just pretends
601 to. This is mainly used to exercise fio
602 itself and for debugging/testing purposes.
604 net Transfer over the network to given host:port.
605 Depending on the protocol used, the hostname,
606 port, listen and filename options are used to
607 specify what sort of connection to make, while
608 the protocol option determines which protocol
610 This engine defines engine specific options.
612 netsplice Like net, but uses splice/vmsplice to
613 map data and send/receive.
614 This engine defines engine specific options.
616 cpuio Doesn't transfer any data, but burns CPU
617 cycles according to the cpuload= and
618 cpucycle= options. Setting cpuload=85
619 will cause that job to do nothing but burn
620 85% of the CPU. In case of SMP machines,
621 use numjobs=<no_of_cpu> to get desired CPU
622 usage, as the cpuload only loads a single
623 CPU at the desired rate.
625 guasi The GUASI IO engine is the Generic Userspace
626 Asyncronous Syscall Interface approach
629 http://www.xmailserver.org/guasi-lib.html
631 for more info on GUASI.
633 rdma The RDMA I/O engine supports both RDMA
634 memory semantics (RDMA_WRITE/RDMA_READ) and
635 channel semantics (Send/Recv) for the
636 InfiniBand, RoCE and iWARP protocols.
638 falloc IO engine that does regular fallocate to
639 simulate data transfer as fio ioengine.
640 DDIR_READ does fallocate(,mode = keep_size,)
641 DDIR_WRITE does fallocate(,mode = 0)
642 DDIR_TRIM does fallocate(,mode = punch_hole)
644 e4defrag IO engine that does regular EXT4_IOC_MOVE_EXT
645 ioctls to simulate defragment activity in
646 request to DDIR_WRITE event
648 external Prefix to specify loading an external
649 IO engine object file. Append the engine
650 filename, eg ioengine=external:/tmp/foo.o
651 to load ioengine foo.o in /tmp.
653 iodepth=int This defines how many io units to keep in flight against
654 the file. The default is 1 for each file defined in this
655 job, can be overridden with a larger value for higher
656 concurrency. Note that increasing iodepth beyond 1 will not
657 affect synchronous ioengines (except for small degress when
658 verify_async is in use). Even async engines may impose OS
659 restrictions causing the desired depth not to be achieved.
660 This may happen on Linux when using libaio and not setting
661 direct=1, since buffered IO is not async on that OS. Keep an
662 eye on the IO depth distribution in the fio output to verify
663 that the achieved depth is as expected. Default: 1.
665 iodepth_batch_submit=int
666 iodepth_batch=int This defines how many pieces of IO to submit at once.
667 It defaults to 1 which means that we submit each IO
668 as soon as it is available, but can be raised to submit
669 bigger batches of IO at the time.
671 iodepth_batch_complete=int This defines how many pieces of IO to retrieve
672 at once. It defaults to 1 which means that we'll ask
673 for a minimum of 1 IO in the retrieval process from
674 the kernel. The IO retrieval will go on until we
675 hit the limit set by iodepth_low. If this variable is
676 set to 0, then fio will always check for completed
677 events before queuing more IO. This helps reduce
678 IO latency, at the cost of more retrieval system calls.
680 iodepth_low=int The low water mark indicating when to start filling
681 the queue again. Defaults to the same as iodepth, meaning
682 that fio will attempt to keep the queue full at all times.
683 If iodepth is set to eg 16 and iodepth_low is set to 4, then
684 after fio has filled the queue of 16 requests, it will let
685 the depth drain down to 4 before starting to fill it again.
687 direct=bool If value is true, use non-buffered io. This is usually
688 O_DIRECT. Note that ZFS on Solaris doesn't support direct io.
689 On Windows the synchronous ioengines don't support direct io.
691 buffered=bool If value is true, use buffered io. This is the opposite
692 of the 'direct' option. Defaults to true.
694 offset=int Start io at the given offset in the file. The data before
695 the given offset will not be touched. This effectively
696 caps the file size at real_size - offset.
698 offset_increment=int If this is provided, then the real offset becomes
699 the offset + offset_increment * thread_number, where the
700 thread number is a counter that starts at 0 and is incremented
701 for each job. This option is useful if there are several jobs
702 which are intended to operate on a file in parallel in disjoint
703 segments, with even spacing between the starting points.
705 fsync=int If writing to a file, issue a sync of the dirty data
706 for every number of blocks given. For example, if you give
707 32 as a parameter, fio will sync the file for every 32
708 writes issued. If fio is using non-buffered io, we may
709 not sync the file. The exception is the sg io engine, which
710 synchronizes the disk cache anyway.
712 fdatasync=int Like fsync= but uses fdatasync() to only sync data and not
714 In FreeBSD and Windows there is no fdatasync(), this falls back to
717 sync_file_range=str:val Use sync_file_range() for every 'val' number of
718 write operations. Fio will track range of writes that
719 have happened since the last sync_file_range() call. 'str'
720 can currently be one or more of:
722 wait_before SYNC_FILE_RANGE_WAIT_BEFORE
723 write SYNC_FILE_RANGE_WRITE
724 wait_after SYNC_FILE_RANGE_WAIT_AFTER
726 So if you do sync_file_range=wait_before,write:8, fio would
727 use SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE for
728 every 8 writes. Also see the sync_file_range(2) man page.
729 This option is Linux specific.
731 overwrite=bool If true, writes to a file will always overwrite existing
732 data. If the file doesn't already exist, it will be
733 created before the write phase begins. If the file exists
734 and is large enough for the specified write phase, nothing
737 end_fsync=bool If true, fsync file contents when a write stage has completed.
739 fsync_on_close=bool If true, fio will fsync() a dirty file on close.
740 This differs from end_fsync in that it will happen on every
741 file close, not just at the end of the job.
743 rwmixread=int How large a percentage of the mix should be reads.
745 rwmixwrite=int How large a percentage of the mix should be writes. If both
746 rwmixread and rwmixwrite is given and the values do not add
747 up to 100%, the latter of the two will be used to override
748 the first. This may interfere with a given rate setting,
749 if fio is asked to limit reads or writes to a certain rate.
750 If that is the case, then the distribution may be skewed.
752 random_distribution=str:float By default, fio will use a completely uniform
753 random distribution when asked to perform random IO. Sometimes
754 it is useful to skew the distribution in specific ways,
755 ensuring that some parts of the data is more hot than others.
756 fio includes the following distribution models:
758 random Uniform random distribution
759 zipf Zipf distribution
760 pareto Pareto distribution
762 When using a zipf or pareto distribution, an input value
763 is also needed to define the access pattern. For zipf, this
764 is the zipf theta. For pareto, it's the pareto power. Fio
765 includes a test program, genzipf, that can be used visualize
766 what the given input values will yield in terms of hit rates.
767 If you wanted to use zipf with a theta of 1.2, you would use
768 random_distribution=zipf:1.2 as the option. If a non-uniform
769 model is used, fio will disable use of the random map.
771 percentage_random=int For a random workload, set how big a percentage should
772 be random. This defaults to 100%, in which case the workload
773 is fully random. It can be set from anywhere from 0 to 100.
774 Setting it to 0 would make the workload fully sequential. Any
775 setting in between will result in a random mix of sequential
776 and random IO, at the given percentages.
778 percentage_sequential=int See percentage_random. It is guaranteed that
779 they add up to 100. The later setting has priority, each
780 will adjust the other.
782 norandommap Normally fio will cover every block of the file when doing
783 random IO. If this option is given, fio will just get a
784 new random offset without looking at past io history. This
785 means that some blocks may not be read or written, and that
786 some blocks may be read/written more than once. This option
787 is mutually exclusive with verify= if and only if multiple
788 blocksizes (via bsrange=) are used, since fio only tracks
789 complete rewrites of blocks.
791 softrandommap=bool See norandommap. If fio runs with the random block map
792 enabled and it fails to allocate the map, if this option is
793 set it will continue without a random block map. As coverage
794 will not be as complete as with random maps, this option is
797 random_generator=str Fio supports the following engines for generating
798 IO offsets for random IO:
800 tausworthe Strong 2^88 cycle random number generator
801 lfsr Linear feedback shift register generator
803 Tausworthe is a strong random number generator, but it
804 requires tracking on the side if we want to ensure that
805 blocks are only read or written once. LFSR guarantees
806 that we never generate the same offset twice, and it's
807 also less computationally expensive. It's not a true
808 random generator, however, though for IO purposes it's
809 typically good enough. LFSR only works with single
810 block sizes, not with workloads that use multiple block
811 sizes. If used with such a workload, fio may read or write
812 some blocks multiple times.
814 nice=int Run the job with the given nice value. See man nice(2).
816 prio=int Set the io priority value of this job. Linux limits us to
817 a positive value between 0 and 7, with 0 being the highest.
820 prioclass=int Set the io priority class. See man ionice(1).
822 thinktime=int Stall the job x microseconds after an io has completed before
823 issuing the next. May be used to simulate processing being
824 done by an application. See thinktime_blocks and
828 Only valid if thinktime is set - pretend to spend CPU time
829 doing something with the data received, before falling back
830 to sleeping for the rest of the period specified by
834 Only valid if thinktime is set - control how many blocks
835 to issue, before waiting 'thinktime' usecs. If not set,
836 defaults to 1 which will make fio wait 'thinktime' usecs
837 after every block. This effectively makes any queue depth
838 setting redundant, since no more than 1 IO will be queued
839 before we have to complete it and do our thinktime. In
840 other words, this setting effectively caps the queue depth
841 if the latter is larger.
843 rate=int Cap the bandwidth used by this job. The number is in bytes/sec,
844 the normal suffix rules apply. You can use rate=500k to limit
845 reads and writes to 500k each, or you can specify read and
846 writes separately. Using rate=1m,500k would limit reads to
847 1MB/sec and writes to 500KB/sec. Capping only reads or
848 writes can be done with rate=,500k or rate=500k,. The former
849 will only limit writes (to 500KB/sec), the latter will only
852 ratemin=int Tell fio to do whatever it can to maintain at least this
853 bandwidth. Failing to meet this requirement, will cause
854 the job to exit. The same format as rate is used for
855 read vs write separation.
857 rate_iops=int Cap the bandwidth to this number of IOPS. Basically the same
858 as rate, just specified independently of bandwidth. If the
859 job is given a block size range instead of a fixed value,
860 the smallest block size is used as the metric. The same format
861 as rate is used for read vs write seperation.
863 rate_iops_min=int If fio doesn't meet this rate of IO, it will cause
864 the job to exit. The same format as rate is used for read vs
867 max_latency=int If set, fio will exit the job if it exceeds this maximum
868 latency. It will exit with an ETIME error.
870 ratecycle=int Average bandwidth for 'rate' and 'ratemin' over this number
873 cpumask=int Set the CPU affinity of this job. The parameter given is a
874 bitmask of allowed CPU's the job may run on. So if you want
875 the allowed CPUs to be 1 and 5, you would pass the decimal
876 value of (1 << 1 | 1 << 5), or 34. See man
877 sched_setaffinity(2). This may not work on all supported
878 operating systems or kernel versions. This option doesn't
879 work well for a higher CPU count than what you can store in
880 an integer mask, so it can only control cpus 1-32. For
881 boxes with larger CPU counts, use cpus_allowed.
883 cpus_allowed=str Controls the same options as cpumask, but it allows a text
884 setting of the permitted CPUs instead. So to use CPUs 1 and
885 5, you would specify cpus_allowed=1,5. This options also
886 allows a range of CPUs. Say you wanted a binding to CPUs
887 1, 5, and 8-15, you would set cpus_allowed=1,5,8-15.
889 numa_cpu_nodes=str Set this job running on spcified NUMA nodes' CPUs. The
890 arguments allow comma delimited list of cpu numbers,
891 A-B ranges, or 'all'. Note, to enable numa options support,
892 fio must be built on a system with libnuma-dev(el) installed.
894 numa_mem_policy=str Set this job's memory policy and corresponding NUMA
895 nodes. Format of the argements:
897 `mode' is one of the following memory policy:
898 default, prefer, bind, interleave, local
899 For `default' and `local' memory policy, no node is
900 needed to be specified.
901 For `prefer', only one node is allowed.
902 For `bind' and `interleave', it allow comma delimited
903 list of numbers, A-B ranges, or 'all'.
905 startdelay=time Start this job the specified number of seconds after fio
906 has started. Only useful if the job file contains several
907 jobs, and you want to delay starting some jobs to a certain
910 runtime=time Tell fio to terminate processing after the specified number
911 of seconds. It can be quite hard to determine for how long
912 a specified job will run, so this parameter is handy to
913 cap the total runtime to a given time.
915 time_based If set, fio will run for the duration of the runtime
916 specified even if the file(s) are completely read or
917 written. It will simply loop over the same workload
918 as many times as the runtime allows.
920 ramp_time=time If set, fio will run the specified workload for this amount
921 of time before logging any performance numbers. Useful for
922 letting performance settle before logging results, thus
923 minimizing the runtime required for stable results. Note
924 that the ramp_time is considered lead in time for a job,
925 thus it will increase the total runtime if a special timeout
926 or runtime is specified.
928 invalidate=bool Invalidate the buffer/page cache parts for this file prior
929 to starting io. Defaults to true.
931 sync=bool Use sync io for buffered writes. For the majority of the
932 io engines, this means using O_SYNC.
935 mem=str Fio can use various types of memory as the io unit buffer.
936 The allowed values are:
938 malloc Use memory from malloc(3) as the buffers.
940 shm Use shared memory as the buffers. Allocated
943 shmhuge Same as shm, but use huge pages as backing.
945 mmap Use mmap to allocate buffers. May either be
946 anonymous memory, or can be file backed if
947 a filename is given after the option. The
948 format is mem=mmap:/path/to/file.
950 mmaphuge Use a memory mapped huge file as the buffer
951 backing. Append filename after mmaphuge, ala
952 mem=mmaphuge:/hugetlbfs/file
954 The area allocated is a function of the maximum allowed
955 bs size for the job, multiplied by the io depth given. Note
956 that for shmhuge and mmaphuge to work, the system must have
957 free huge pages allocated. This can normally be checked
958 and set by reading/writing /proc/sys/vm/nr_hugepages on a
959 Linux system. Fio assumes a huge page is 4MB in size. So
960 to calculate the number of huge pages you need for a given
961 job file, add up the io depth of all jobs (normally one unless
962 iodepth= is used) and multiply by the maximum bs set. Then
963 divide that number by the huge page size. You can see the
964 size of the huge pages in /proc/meminfo. If no huge pages
965 are allocated by having a non-zero number in nr_hugepages,
966 using mmaphuge or shmhuge will fail. Also see hugepage-size.
968 mmaphuge also needs to have hugetlbfs mounted and the file
969 location should point there. So if it's mounted in /huge,
970 you would use mem=mmaphuge:/huge/somefile.
972 iomem_align=int This indiciates the memory alignment of the IO memory buffers.
973 Note that the given alignment is applied to the first IO unit
974 buffer, if using iodepth the alignment of the following buffers
975 are given by the bs used. In other words, if using a bs that is
976 a multiple of the page sized in the system, all buffers will
977 be aligned to this value. If using a bs that is not page
978 aligned, the alignment of subsequent IO memory buffers is the
979 sum of the iomem_align and bs used.
982 Defines the size of a huge page. Must at least be equal
983 to the system setting, see /proc/meminfo. Defaults to 4MB.
984 Should probably always be a multiple of megabytes, so using
985 hugepage-size=Xm is the preferred way to set this to avoid
986 setting a non-pow-2 bad value.
988 exitall When one job finishes, terminate the rest. The default is
989 to wait for each job to finish, sometimes that is not the
992 bwavgtime=int Average the calculated bandwidth over the given time. Value
993 is specified in milliseconds.
995 iopsavgtime=int Average the calculated IOPS over the given time. Value
996 is specified in milliseconds.
998 create_serialize=bool If true, serialize the file creating for the jobs.
999 This may be handy to avoid interleaving of data
1000 files, which may greatly depend on the filesystem
1001 used and even the number of processors in the system.
1003 create_fsync=bool fsync the data file after creation. This is the
1006 create_on_open=bool Don't pre-setup the files for IO, just create open()
1007 when it's time to do IO to that file.
1009 create_only=bool If true, fio will only run the setup phase of the job.
1010 If files need to be laid out or updated on disk, only
1011 that will be done. The actual job contents are not
1014 pre_read=bool If this is given, files will be pre-read into memory before
1015 starting the given IO operation. This will also clear
1016 the 'invalidate' flag, since it is pointless to pre-read
1017 and then drop the cache. This will only work for IO engines
1018 that are seekable, since they allow you to read the same data
1019 multiple times. Thus it will not work on eg network or splice
1022 unlink=bool Unlink the job files when done. Not the default, as repeated
1023 runs of that job would then waste time recreating the file
1024 set again and again.
1026 loops=int Run the specified number of iterations of this job. Used
1027 to repeat the same workload a given number of times. Defaults
1030 do_verify=bool Run the verify phase after a write phase. Only makes sense if
1031 verify is set. Defaults to 1.
1033 verify=str If writing to a file, fio can verify the file contents
1034 after each iteration of the job. The allowed values are:
1036 md5 Use an md5 sum of the data area and store
1037 it in the header of each block.
1039 crc64 Use an experimental crc64 sum of the data
1040 area and store it in the header of each
1043 crc32c Use a crc32c sum of the data area and store
1044 it in the header of each block.
1046 crc32c-intel Use hardware assisted crc32c calcuation
1047 provided on SSE4.2 enabled processors. Falls
1048 back to regular software crc32c, if not
1049 supported by the system.
1051 crc32 Use a crc32 sum of the data area and store
1052 it in the header of each block.
1054 crc16 Use a crc16 sum of the data area and store
1055 it in the header of each block.
1057 crc7 Use a crc7 sum of the data area and store
1058 it in the header of each block.
1060 sha512 Use sha512 as the checksum function.
1062 sha256 Use sha256 as the checksum function.
1064 sha1 Use optimized sha1 as the checksum function.
1066 meta Write extra information about each io
1067 (timestamp, block number etc.). The block
1068 number is verified. See also verify_pattern.
1070 null Only pretend to verify. Useful for testing
1071 internals with ioengine=null, not for much
1074 This option can be used for repeated burn-in tests of a
1075 system to make sure that the written data is also
1076 correctly read back. If the data direction given is
1077 a read or random read, fio will assume that it should
1078 verify a previously written file. If the data direction
1079 includes any form of write, the verify will be of the
1082 verifysort=bool If set, fio will sort written verify blocks when it deems
1083 it faster to read them back in a sorted manner. This is
1084 often the case when overwriting an existing file, since
1085 the blocks are already laid out in the file system. You
1086 can ignore this option unless doing huge amounts of really
1087 fast IO where the red-black tree sorting CPU time becomes
1090 verify_offset=int Swap the verification header with data somewhere else
1091 in the block before writing. Its swapped back before
1094 verify_interval=int Write the verification header at a finer granularity
1095 than the blocksize. It will be written for chunks the
1096 size of header_interval. blocksize should divide this
1099 verify_pattern=str If set, fio will fill the io buffers with this
1100 pattern. Fio defaults to filling with totally random
1101 bytes, but sometimes it's interesting to fill with a known
1102 pattern for io verification purposes. Depending on the
1103 width of the pattern, fio will fill 1/2/3/4 bytes of the
1104 buffer at the time(it can be either a decimal or a hex number).
1105 The verify_pattern if larger than a 32-bit quantity has to
1106 be a hex number that starts with either "0x" or "0X". Use
1109 verify_fatal=bool Normally fio will keep checking the entire contents
1110 before quitting on a block verification failure. If this
1111 option is set, fio will exit the job on the first observed
1114 verify_dump=bool If set, dump the contents of both the original data
1115 block and the data block we read off disk to files. This
1116 allows later analysis to inspect just what kind of data
1117 corruption occurred. Off by default.
1119 verify_async=int Fio will normally verify IO inline from the submitting
1120 thread. This option takes an integer describing how many
1121 async offload threads to create for IO verification instead,
1122 causing fio to offload the duty of verifying IO contents
1123 to one or more separate threads. If using this offload
1124 option, even sync IO engines can benefit from using an
1125 iodepth setting higher than 1, as it allows them to have
1126 IO in flight while verifies are running.
1128 verify_async_cpus=str Tell fio to set the given CPU affinity on the
1129 async IO verification threads. See cpus_allowed for the
1132 verify_backlog=int Fio will normally verify the written contents of a
1133 job that utilizes verify once that job has completed. In
1134 other words, everything is written then everything is read
1135 back and verified. You may want to verify continually
1136 instead for a variety of reasons. Fio stores the meta data
1137 associated with an IO block in memory, so for large
1138 verify workloads, quite a bit of memory would be used up
1139 holding this meta data. If this option is enabled, fio
1140 will write only N blocks before verifying these blocks.
1142 will verify the previously written blocks before continuing
1145 verify_backlog_batch=int Control how many blocks fio will verify
1146 if verify_backlog is set. If not set, will default to
1147 the value of verify_backlog (meaning the entire queue
1148 is read back and verified). If verify_backlog_batch is
1149 less than verify_backlog then not all blocks will be verified,
1150 if verify_backlog_batch is larger than verify_backlog, some
1151 blocks will be verified more than once.
1154 wait_for_previous Wait for preceeding jobs in the job file to exit, before
1155 starting this one. Can be used to insert serialization
1156 points in the job file. A stone wall also implies starting
1157 a new reporting group.
1159 new_group Start a new reporting group. See: group_reporting.
1161 numjobs=int Create the specified number of clones of this job. May be
1162 used to setup a larger number of threads/processes doing
1163 the same thing. Each thread is reported separately; to see
1164 statistics for all clones as a whole, use group_reporting in
1165 conjunction with new_group.
1167 group_reporting It may sometimes be interesting to display statistics for
1168 groups of jobs as a whole instead of for each individual job.
1169 This is especially true if 'numjobs' is used; looking at
1170 individual thread/process output quickly becomes unwieldy.
1171 To see the final report per-group instead of per-job, use
1172 'group_reporting'. Jobs in a file will be part of the same
1173 reporting group, unless if separated by a stonewall, or by
1176 thread fio defaults to forking jobs, however if this option is
1177 given, fio will use pthread_create(3) to create threads
1180 zonesize=int Divide a file into zones of the specified size. See zoneskip.
1182 zoneskip=int Skip the specified number of bytes when zonesize data has
1183 been read. The two zone options can be used to only do
1184 io on zones of a file.
1186 write_iolog=str Write the issued io patterns to the specified file. See
1187 read_iolog. Specify a separate file for each job, otherwise
1188 the iologs will be interspersed and the file may be corrupt.
1190 read_iolog=str Open an iolog with the specified file name and replay the
1191 io patterns it contains. This can be used to store a
1192 workload and replay it sometime later. The iolog given
1193 may also be a blktrace binary file, which allows fio
1194 to replay a workload captured by blktrace. See blktrace
1195 for how to capture such logging data. For blktrace replay,
1196 the file needs to be turned into a blkparse binary data
1197 file first (blkparse <device> -o /dev/null -d file_for_fio.bin).
1199 replay_no_stall=int When replaying I/O with read_iolog the default behavior
1200 is to attempt to respect the time stamps within the log and
1201 replay them with the appropriate delay between IOPS. By
1202 setting this variable fio will not respect the timestamps and
1203 attempt to replay them as fast as possible while still
1204 respecting ordering. The result is the same I/O pattern to a
1205 given device, but different timings.
1207 replay_redirect=str While replaying I/O patterns using read_iolog the
1208 default behavior is to replay the IOPS onto the major/minor
1209 device that each IOP was recorded from. This is sometimes
1210 undesireable because on a different machine those major/minor
1211 numbers can map to a different device. Changing hardware on
1212 the same system can also result in a different major/minor
1213 mapping. Replay_redirect causes all IOPS to be replayed onto
1214 the single specified device regardless of the device it was
1215 recorded from. i.e. replay_redirect=/dev/sdc would cause all
1216 IO in the blktrace to be replayed onto /dev/sdc. This means
1217 multiple devices will be replayed onto a single, if the trace
1218 contains multiple devices. If you want multiple devices to be
1219 replayed concurrently to multiple redirected devices you must
1220 blkparse your trace into separate traces and replay them with
1221 independent fio invocations. Unfortuantely this also breaks
1222 the strict time ordering between multiple device accesses.
1224 write_bw_log=str If given, write a bandwidth log of the jobs in this job
1225 file. Can be used to store data of the bandwidth of the
1226 jobs in their lifetime. The included fio_generate_plots
1227 script uses gnuplot to turn these text files into nice
1228 graphs. See write_lat_log for behaviour of given
1229 filename. For this option, the suffix is _bw.log.
1231 write_lat_log=str Same as write_bw_log, except that this option stores io
1232 submission, completion, and total latencies instead. If no
1233 filename is given with this option, the default filename of
1234 "jobname_type.log" is used. Even if the filename is given,
1235 fio will still append the type of log. So if one specifies
1239 The actual log names will be foo_slat.log, foo_clat.log,
1240 and foo_lat.log. This helps fio_generate_plot fine the logs
1243 write_bw_log=str If given, write an IOPS log of the jobs in this job
1244 file. See write_bw_log.
1246 write_iops_log=str Same as write_bw_log, but writes IOPS. If no filename is
1247 given with this option, the default filename of
1248 "jobname_type.log" is used. Even if the filename is given,
1249 fio will still append the type of log.
1251 log_avg_msec=int By default, fio will log an entry in the iops, latency,
1252 or bw log for every IO that completes. When writing to the
1253 disk log, that can quickly grow to a very large size. Setting
1254 this option makes fio average the each log entry over the
1255 specified period of time, reducing the resolution of the log.
1258 lockmem=int Pin down the specified amount of memory with mlock(2). Can
1259 potentially be used instead of removing memory or booting
1260 with less memory to simulate a smaller amount of memory.
1261 The amount specified is per worker.
1263 exec_prerun=str Before running this job, issue the command specified
1264 through system(3). Output is redirected in a file called
1267 exec_postrun=str After the job completes, issue the command specified
1268 though system(3). Output is redirected in a file called
1269 jobname.postrun.txt.
1271 ioscheduler=str Attempt to switch the device hosting the file to the specified
1272 io scheduler before running.
1274 disk_util=bool Generate disk utilization statistics, if the platform
1275 supports it. Defaults to on.
1277 disable_lat=bool Disable measurements of total latency numbers. Useful
1278 only for cutting back the number of calls to gettimeofday,
1279 as that does impact performance at really high IOPS rates.
1280 Note that to really get rid of a large amount of these
1281 calls, this option must be used with disable_slat and
1284 disable_clat=bool Disable measurements of completion latency numbers. See
1287 disable_slat=bool Disable measurements of submission latency numbers. See
1290 disable_bw=bool Disable measurements of throughput/bandwidth numbers. See
1293 clat_percentiles=bool Enable the reporting of percentiles of
1294 completion latencies.
1296 percentile_list=float_list Overwrite the default list of percentiles
1297 for completion latencies. Each number is a floating
1298 number in the range (0,100], and the maximum length of
1299 the list is 20. Use ':' to separate the numbers, and
1300 list the numbers in ascending order. For example,
1301 --percentile_list=99.5:99.9 will cause fio to report
1302 the values of completion latency below which 99.5% and
1303 99.9% of the observed latencies fell, respectively.
1305 clocksource=str Use the given clocksource as the base of timing. The
1306 supported options are:
1308 gettimeofday gettimeofday(2)
1310 clock_gettime clock_gettime(2)
1312 cpu Internal CPU clock source
1314 cpu is the preferred clocksource if it is reliable, as it
1315 is very fast (and fio is heavy on time calls). Fio will
1316 automatically use this clocksource if it's supported and
1317 considered reliable on the system it is running on, unless
1318 another clocksource is specifically set. For x86/x86-64 CPUs,
1319 this means supporting TSC Invariant.
1321 gtod_reduce=bool Enable all of the gettimeofday() reducing options
1322 (disable_clat, disable_slat, disable_bw) plus reduce
1323 precision of the timeout somewhat to really shrink
1324 the gettimeofday() call count. With this option enabled,
1325 we only do about 0.4% of the gtod() calls we would have
1326 done if all time keeping was enabled.
1328 gtod_cpu=int Sometimes it's cheaper to dedicate a single thread of
1329 execution to just getting the current time. Fio (and
1330 databases, for instance) are very intensive on gettimeofday()
1331 calls. With this option, you can set one CPU aside for
1332 doing nothing but logging current time to a shared memory
1333 location. Then the other threads/processes that run IO
1334 workloads need only copy that segment, instead of entering
1335 the kernel with a gettimeofday() call. The CPU set aside
1336 for doing these time calls will be excluded from other
1337 uses. Fio will manually clear it from the CPU mask of other
1340 continue_on_error=str Normally fio will exit the job on the first observed
1341 failure. If this option is set, fio will continue the job when
1342 there is a 'non-fatal error' (EIO or EILSEQ) until the runtime
1343 is exceeded or the I/O size specified is completed. If this
1344 option is used, there are two more stats that are appended,
1345 the total error count and the first error. The error field
1346 given in the stats is the first error that was hit during the
1349 The allowed values are:
1351 none Exit on any IO or verify errors.
1353 read Continue on read errors, exit on all others.
1355 write Continue on write errors, exit on all others.
1357 io Continue on any IO error, exit on all others.
1359 verify Continue on verify errors, exit on all others.
1361 all Continue on all errors.
1363 0 Backward-compatible alias for 'none'.
1365 1 Backward-compatible alias for 'all'.
1367 ignore_error=str Sometimes you want to ignore some errors during test
1368 in that case you can specify error list for each error type.
1369 ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST
1370 errors for given error type is separated with ':'. Error
1371 may be symbol ('ENOSPC', 'ENOMEM') or integer.
1373 ignore_error=EAGAIN,ENOSPC:122
1374 This option will ignore EAGAIN from READ, and ENOSPC and
1375 122(EDQUOT) from WRITE.
1377 error_dump=bool If set dump every error even if it is non fatal, true
1378 by default. If disabled only fatal error will be dumped
1380 cgroup=str Add job to this control group. If it doesn't exist, it will
1381 be created. The system must have a mounted cgroup blkio
1382 mount point for this to work. If your system doesn't have it
1383 mounted, you can do so with:
1385 # mount -t cgroup -o blkio none /cgroup
1387 cgroup_weight=int Set the weight of the cgroup to this value. See
1388 the documentation that comes with the kernel, allowed values
1389 are in the range of 100..1000.
1391 cgroup_nodelete=bool Normally fio will delete the cgroups it has created after
1392 the job completion. To override this behavior and to leave
1393 cgroups around after the job completion, set cgroup_nodelete=1.
1394 This can be useful if one wants to inspect various cgroup
1395 files after job completion. Default: false
1397 uid=int Instead of running as the invoking user, set the user ID to
1398 this value before the thread/process does any work.
1400 gid=int Set group ID, see uid.
1402 flow_id=int The ID of the flow. If not specified, it defaults to being a
1403 global flow. See flow.
1405 flow=int Weight in token-based flow control. If this value is used, then
1406 there is a 'flow counter' which is used to regulate the
1407 proportion of activity between two or more jobs. fio attempts
1408 to keep this flow counter near zero. The 'flow' parameter
1409 stands for how much should be added or subtracted to the flow
1410 counter on each iteration of the main I/O loop. That is, if
1411 one job has flow=8 and another job has flow=-1, then there
1412 will be a roughly 1:8 ratio in how much one runs vs the other.
1414 flow_watermark=int The maximum value that the absolute value of the flow
1415 counter is allowed to reach before the job must wait for a
1416 lower value of the counter.
1418 flow_sleep=int The period of time, in microseconds, to wait after the flow
1419 watermark has been exceeded before retrying operations
1421 In addition, there are some parameters which are only valid when a specific
1422 ioengine is in use. These are used identically to normal parameters, with the
1423 caveat that when used on the command line, they must come after the ioengine
1424 that defines them is selected.
1426 [libaio] userspace_reap Normally, with the libaio engine in use, fio will use
1427 the io_getevents system call to reap newly returned events.
1428 With this flag turned on, the AIO ring will be read directly
1429 from user-space to reap events. The reaping mode is only
1430 enabled when polling for a minimum of 0 events (eg when
1431 iodepth_batch_complete=0).
1433 [cpu] cpuload=int Attempt to use the specified percentage of CPU cycles.
1435 [cpu] cpuchunks=int Split the load into cycles of the given time. In
1438 [netsplice] hostname=str
1439 [net] hostname=str The host name or IP address to use for TCP or UDP based IO.
1440 If the job is a TCP listener or UDP reader, the hostname is not
1441 used and must be omitted.
1443 [netsplice] port=int
1444 [net] port=int The TCP or UDP port to bind to or connect to.
1446 [netsplice] nodelay=bool
1447 [net] nodelay=bool Set TCP_NODELAY on TCP connections.
1449 [netsplice] protocol=str
1450 [netsplice] proto=str
1452 [net] proto=str The network protocol to use. Accepted values are:
1454 tcp Transmission control protocol
1455 udp User datagram protocol
1456 unix UNIX domain socket
1458 When the protocol is TCP or UDP, the port must also be given,
1459 as well as the hostname if the job is a TCP listener or UDP
1460 reader. For unix sockets, the normal filename option should be
1461 used and the port is invalid.
1463 [net] listen For TCP network connections, tell fio to listen for incoming
1464 connections rather than initiating an outgoing connection. The
1465 hostname must be omitted if this option is used.
1466 [net] pingpong Normal a network writer will just continue writing data, and
1467 a network reader will just consume packages. If pingpong=1
1468 is set, a writer will send its normal payload to the reader,
1469 then wait for the reader to send the same payload back. This
1470 allows fio to measure network latencies. The submission
1471 and completion latencies then measure local time spent
1472 sending or receiving, and the completion latency measures
1473 how long it took for the other end to receive and send back.
1475 [e4defrag] donorname=str
1476 File will be used as a block donor(swap extents between files)
1477 [e4defrag] inplace=int
1478 Configure donor file blocks allocation strategy
1479 0(default): Preallocate donor's file on init
1480 1 : allocate space immidietly inside defragment event,
1481 and free right after event
1485 6.0 Interpreting the output
1486 ---------------------------
1488 fio spits out a lot of output. While running, fio will display the
1489 status of the jobs created. An example of that would be:
1491 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
1493 The characters inside the square brackets denote the current status of
1494 each thread. The possible values (in typical life cycle order) are:
1498 P Thread setup, but not started.
1500 I Thread initialized, waiting or generating necessary data.
1501 p Thread running pre-reading file(s).
1502 R Running, doing sequential reads.
1503 r Running, doing random reads.
1504 W Running, doing sequential writes.
1505 w Running, doing random writes.
1506 M Running, doing mixed sequential reads/writes.
1507 m Running, doing mixed random reads/writes.
1508 F Running, currently waiting for fsync()
1509 V Running, doing verification of written data.
1510 E Thread exited, not reaped by main thread yet.
1512 X Thread reaped, exited with an error.
1513 K Thread reaped, exited due to signal.
1515 The other values are fairly self explanatory - number of threads
1516 currently running and doing io, rate of io since last check (read speed
1517 listed first, then write speed), and the estimated completion percentage
1518 and time for the running group. It's impossible to estimate runtime of
1519 the following groups (if any). Note that the string is displayed in order,
1520 so it's possible to tell which of the jobs are currently doing what. The
1521 first character is the first job defined in the job file, and so forth.
1523 When fio is done (or interrupted by ctrl-c), it will show the data for
1524 each thread, group of threads, and disks in that order. For each data
1525 direction, the output looks like:
1527 Client1 (g=0): err= 0:
1528 write: io= 32MB, bw= 666KB/s, iops=89 , runt= 50320msec
1529 slat (msec): min= 0, max= 136, avg= 0.03, stdev= 1.92
1530 clat (msec): min= 0, max= 631, avg=48.50, stdev=86.82
1531 bw (KB/s) : min= 0, max= 1196, per=51.00%, avg=664.02, stdev=681.68
1532 cpu : usr=1.49%, sys=0.25%, ctx=7969, majf=0, minf=17
1533 IO depths : 1=0.1%, 2=0.3%, 4=0.5%, 8=99.0%, 16=0.0%, 32=0.0%, >32=0.0%
1534 submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
1535 complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
1536 issued r/w: total=0/32768, short=0/0
1537 lat (msec): 2=1.6%, 4=0.0%, 10=3.2%, 20=12.8%, 50=38.4%, 100=24.8%,
1538 lat (msec): 250=15.2%, 500=0.0%, 750=0.0%, 1000=0.0%, >=2048=0.0%
1540 The client number is printed, along with the group id and error of that
1541 thread. Below is the io statistics, here for writes. In the order listed,
1544 io= Number of megabytes io performed
1545 bw= Average bandwidth rate
1546 iops= Average IOs performed per second
1547 runt= The runtime of that thread
1548 slat= Submission latency (avg being the average, stdev being the
1549 standard deviation). This is the time it took to submit
1550 the io. For sync io, the slat is really the completion
1551 latency, since queue/complete is one operation there. This
1552 value can be in milliseconds or microseconds, fio will choose
1553 the most appropriate base and print that. In the example
1554 above, milliseconds is the best scale. Note: in --minimal mode
1555 latencies are always expressed in microseconds.
1556 clat= Completion latency. Same names as slat, this denotes the
1557 time from submission to completion of the io pieces. For
1558 sync io, clat will usually be equal (or very close) to 0,
1559 as the time from submit to complete is basically just
1560 CPU time (io has already been done, see slat explanation).
1561 bw= Bandwidth. Same names as the xlat stats, but also includes
1562 an approximate percentage of total aggregate bandwidth
1563 this thread received in this group. This last value is
1564 only really useful if the threads in this group are on the
1565 same disk, since they are then competing for disk access.
1566 cpu= CPU usage. User and system time, along with the number
1567 of context switches this thread went through, usage of
1568 system and user time, and finally the number of major
1569 and minor page faults.
1570 IO depths= The distribution of io depths over the job life time. The
1571 numbers are divided into powers of 2, so for example the
1572 16= entries includes depths up to that value but higher
1573 than the previous entry. In other words, it covers the
1574 range from 16 to 31.
1575 IO submit= How many pieces of IO were submitting in a single submit
1576 call. Each entry denotes that amount and below, until
1577 the previous entry - eg, 8=100% mean that we submitted
1578 anywhere in between 5-8 ios per submit call.
1579 IO complete= Like the above submit number, but for completions instead.
1580 IO issued= The number of read/write requests issued, and how many
1582 IO latencies= The distribution of IO completion latencies. This is the
1583 time from when IO leaves fio and when it gets completed.
1584 The numbers follow the same pattern as the IO depths,
1585 meaning that 2=1.6% means that 1.6% of the IO completed
1586 within 2 msecs, 20=12.8% means that 12.8% of the IO
1587 took more than 10 msecs, but less than (or equal to) 20 msecs.
1589 After each client has been listed, the group statistics are printed. They
1590 will look like this:
1592 Run status group 0 (all jobs):
1593 READ: io=64MB, aggrb=22178, minb=11355, maxb=11814, mint=2840msec, maxt=2955msec
1594 WRITE: io=64MB, aggrb=1302, minb=666, maxb=669, mint=50093msec, maxt=50320msec
1596 For each data direction, it prints:
1598 io= Number of megabytes io performed.
1599 aggrb= Aggregate bandwidth of threads in this group.
1600 minb= The minimum average bandwidth a thread saw.
1601 maxb= The maximum average bandwidth a thread saw.
1602 mint= The smallest runtime of the threads in that group.
1603 maxt= The longest runtime of the threads in that group.
1605 And finally, the disk statistics are printed. They will look like this:
1607 Disk stats (read/write):
1608 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
1610 Each value is printed for both reads and writes, with reads first. The
1613 ios= Number of ios performed by all groups.
1614 merge= Number of merges io the io scheduler.
1615 ticks= Number of ticks we kept the disk busy.
1616 io_queue= Total time spent in the disk queue.
1617 util= The disk utilization. A value of 100% means we kept the disk
1618 busy constantly, 50% would be a disk idling half of the time.
1620 It is also possible to get fio to dump the current output while it is
1621 running, without terminating the job. To do that, send fio the USR1 signal.
1622 You can also get regularly timed dumps by using the --status-interval
1623 parameter, or by creating a file in /tmp named fio-dump-status. If fio
1624 sees this file, it will unlink it and dump the current output status.
1630 For scripted usage where you typically want to generate tables or graphs
1631 of the results, fio can output the results in a semicolon separated format.
1632 The format is one long line of values, such as:
1634 2;card0;0;0;7139336;121836;60004;1;10109;27.932460;116.933948;220;126861;3495.446807;1085.368601;226;126864;3523.635629;1089.012448;24063;99944;50.275485%;59818.274627;5540.657370;7155060;122104;60004;1;8338;29.086342;117.839068;388;128077;5032.488518;1234.785715;391;128085;5061.839412;1236.909129;23436;100928;50.287926%;59964.832030;5644.844189;14.595833%;19.394167%;123706;0;7313;0.1%;0.1%;0.1%;0.1%;0.1%;0.1%;100.0%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.01%;0.02%;0.05%;0.16%;6.04%;40.40%;52.68%;0.64%;0.01%;0.00%;0.01%;0.00%;0.00%;0.00%;0.00%;0.00%
1635 A description of this job goes here.
1637 The job description (if provided) follows on a second line.
1639 To enable terse output, use the --minimal command line option. The first
1640 value is the version of the terse output format. If the output has to
1641 be changed for some reason, this number will be incremented by 1 to
1642 signify that change.
1644 Split up, the format is as follows:
1646 terse version, fio version, jobname, groupid, error
1648 Total IO (KB), bandwidth (KB/sec), IOPS, runtime (msec)
1649 Submission latency: min, max, mean, deviation (usec)
1650 Completion latency: min, max, mean, deviation (usec)
1651 Completion latency percentiles: 20 fields (see below)
1652 Total latency: min, max, mean, deviation (usec)
1653 Bw (KB/s): min, max, aggregate percentage of total, mean, deviation
1655 Total IO (KB), bandwidth (KB/sec), IOPS, runtime (msec)
1656 Submission latency: min, max, mean, deviation (usec)
1657 Completion latency: min, max, mean, deviation (usec)
1658 Completion latency percentiles: 20 fields (see below)
1659 Total latency: min, max, mean, deviation (usec)
1660 Bw (KB/s): min, max, aggregate percentage of total, mean, deviation
1661 CPU usage: user, system, context switches, major faults, minor faults
1662 IO depths: <=1, 2, 4, 8, 16, 32, >=64
1663 IO latencies microseconds: <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000
1664 IO latencies milliseconds: <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000
1665 Disk utilization: Disk name, Read ios, write ios,
1666 Read merges, write merges,
1667 Read ticks, write ticks,
1668 Time spent in queue, disk utilization percentage
1669 Additional Info (dependant on continue_on_error, default off): total # errors, first error code
1671 Additional Info (dependant on description being set): Text description
1673 Completion latency percentiles can be a grouping of up to 20 sets, so
1674 for the terse output fio writes all of them. Each field will look like this:
1678 which is the Xth percentile, and the usec latency associated with it.
1680 For disk utilization, all disks used by fio are shown. So for each disk
1681 there will be a disk utilization section.
1684 8.0 Trace file format
1685 ---------------------
1686 There are two trace file format that you can encounter. The older (v1) format
1687 is unsupported since version 1.20-rc3 (March 2008). It will still be described
1688 below in case that you get an old trace and want to understand it.
1690 In any case the trace is a simple text file with a single action per line.
1693 8.1 Trace file format v1
1694 ------------------------
1695 Each line represents a single io action in the following format:
1699 where rw=0/1 for read/write, and the offset and length entries being in bytes.
1701 This format is not supported in Fio versions => 1.20-rc3.
1704 8.2 Trace file format v2
1705 ------------------------
1706 The second version of the trace file format was added in Fio version 1.17.
1707 It allows to access more then one file per trace and has a bigger set of
1708 possible file actions.
1710 The first line of the trace file has to be:
1714 Following this can be lines in two different formats, which are described below.
1716 The file management format:
1720 The filename is given as an absolute path. The action can be one of these:
1722 add Add the given filename to the trace
1723 open Open the file with the given filename. The filename has to have
1724 been added with the add action before.
1725 close Close the file with the given filename. The file has to have been
1729 The file io action format:
1731 filename action offset length
1733 The filename is given as an absolute path, and has to have been added and opened
1734 before it can be used with this format. The offset and length are given in
1735 bytes. The action can be one of these:
1737 wait Wait for 'offset' microseconds. Everything below 100 is discarded.
1738 read Read 'length' bytes beginning from 'offset'
1739 write Write 'length' bytes beginning from 'offset'
1740 sync fsync() the file
1741 datasync fdatasync() the file
1742 trim trim the given file from the given 'offset' for 'length' bytes
1745 9.0 CPU idleness profiling
1746 --------------------------
1747 In some cases, we want to understand CPU overhead in a test. For example,
1748 we test patches for the specific goodness of whether they reduce CPU usage.
1749 fio implements a balloon approach to create a thread per CPU that runs at
1750 idle priority, meaning that it only runs when nobody else needs the cpu.
1751 By measuring the amount of work completed by the thread, idleness of each
1752 CPU can be derived accordingly.
1754 An unit work is defined as touching a full page of unsigned characters. Mean
1755 and standard deviation of time to complete an unit work is reported in "unit
1756 work" section. Options can be chosen to report detailed percpu idleness or
1757 overall system idleness by aggregating percpu stats.
projects / fio.git / blob