Commit | Line | Data |
---|---|---|
f80dba8d MT |
1 | How fio works |
2 | ------------- | |
3 | ||
4 | The first step in getting fio to simulate a desired I/O workload, is writing a | |
5 | job file describing that specific setup. A job file may contain any number of | |
6 | threads and/or files -- the typical contents of the job file is a *global* | |
7 | section defining shared parameters, and one or more job sections describing the | |
8 | jobs involved. When run, fio parses this file and sets everything up as | |
9 | described. If we break down a job from top to bottom, it contains the following | |
10 | basic parameters: | |
11 | ||
12 | `I/O type`_ | |
13 | ||
14 | Defines the I/O pattern issued to the file(s). We may only be reading | |
15 | sequentially from this file(s), or we may be writing randomly. Or even | |
16 | mixing reads and writes, sequentially or randomly. | |
17 | Should we be doing buffered I/O, or direct/raw I/O? | |
18 | ||
19 | `Block size`_ | |
20 | ||
21 | In how large chunks are we issuing I/O? This may be a single value, | |
22 | or it may describe a range of block sizes. | |
23 | ||
24 | `I/O size`_ | |
25 | ||
26 | How much data are we going to be reading/writing. | |
27 | ||
28 | `I/O engine`_ | |
29 | ||
30 | How do we issue I/O? We could be memory mapping the file, we could be | |
31 | using regular read/write, we could be using splice, async I/O, or even | |
32 | SG (SCSI generic sg). | |
33 | ||
34 | `I/O depth`_ | |
35 | ||
36 | If the I/O engine is async, how large a queuing depth do we want to | |
37 | maintain? | |
38 | ||
39 | ||
40 | `Target file/device`_ | |
41 | ||
42 | How many files are we spreading the workload over. | |
43 | ||
44 | `Threads, processes and job synchronization`_ | |
45 | ||
46 | How many threads or processes should we spread this workload over. | |
47 | ||
48 | The above are the basic parameters defined for a workload, in addition there's a | |
49 | multitude of parameters that modify other aspects of how this job behaves. | |
50 | ||
51 | ||
52 | Command line options | |
53 | -------------------- | |
54 | ||
55 | .. option:: --debug=type | |
56 | ||
f50fbdda | 57 | Enable verbose tracing `type` of various fio actions. May be ``all`` for all types |
b034c0dd TK |
58 | or individual types separated by a comma (e.g. ``--debug=file,mem`` will |
59 | enable file and memory debugging). Currently, additional logging is | |
60 | available for: | |
f80dba8d | 61 | |
b034c0dd | 62 | *process* |
f80dba8d | 63 | Dump info related to processes. |
b034c0dd | 64 | *file* |
f80dba8d | 65 | Dump info related to file actions. |
b034c0dd | 66 | *io* |
f80dba8d | 67 | Dump info related to I/O queuing. |
b034c0dd | 68 | *mem* |
f80dba8d | 69 | Dump info related to memory allocations. |
b034c0dd | 70 | *blktrace* |
f80dba8d | 71 | Dump info related to blktrace setup. |
b034c0dd | 72 | *verify* |
f80dba8d | 73 | Dump info related to I/O verification. |
b034c0dd | 74 | *all* |
f80dba8d | 75 | Enable all debug options. |
b034c0dd | 76 | *random* |
f80dba8d | 77 | Dump info related to random offset generation. |
b034c0dd | 78 | *parse* |
f80dba8d | 79 | Dump info related to option matching and parsing. |
b034c0dd | 80 | *diskutil* |
f80dba8d | 81 | Dump info related to disk utilization updates. |
b034c0dd | 82 | *job:x* |
f80dba8d | 83 | Dump info only related to job number x. |
b034c0dd | 84 | *mutex* |
f80dba8d | 85 | Dump info only related to mutex up/down ops. |
b034c0dd | 86 | *profile* |
f80dba8d | 87 | Dump info related to profile extensions. |
b034c0dd | 88 | *time* |
f80dba8d | 89 | Dump info related to internal time keeping. |
b034c0dd | 90 | *net* |
f80dba8d | 91 | Dump info related to networking connections. |
b034c0dd | 92 | *rate* |
f80dba8d | 93 | Dump info related to I/O rate switching. |
b034c0dd | 94 | *compress* |
f80dba8d | 95 | Dump info related to log compress/decompress. |
a02ec45a VF |
96 | *steadystate* |
97 | Dump info related to steadystate detection. | |
98 | *helperthread* | |
99 | Dump info related to the helper thread. | |
100 | *zbd* | |
101 | Dump info related to support for zoned block devices. | |
b034c0dd | 102 | *?* or *help* |
f80dba8d MT |
103 | Show available debug options. |
104 | ||
105 | .. option:: --parse-only | |
106 | ||
25cd4b95 | 107 | Parse options only, don't start any I/O. |
f80dba8d | 108 | |
b9921d1a DZ |
109 | .. option:: --merge-blktrace-only |
110 | ||
111 | Merge blktraces only, don't start any I/O. | |
112 | ||
f80dba8d MT |
113 | .. option:: --output=filename |
114 | ||
115 | Write output to file `filename`. | |
116 | ||
f50fbdda | 117 | .. option:: --output-format=format |
b8f7e412 | 118 | |
f50fbdda | 119 | Set the reporting `format` to `normal`, `terse`, `json`, or `json+`. Multiple |
b8f7e412 TK |
120 | formats can be selected, separated by a comma. `terse` is a CSV based |
121 | format. `json+` is like `json`, except it adds a full dump of the latency | |
122 | buckets. | |
123 | ||
f80dba8d MT |
124 | .. option:: --bandwidth-log |
125 | ||
126 | Generate aggregate bandwidth logs. | |
127 | ||
128 | .. option:: --minimal | |
129 | ||
130 | Print statistics in a terse, semicolon-delimited format. | |
131 | ||
132 | .. option:: --append-terse | |
133 | ||
b034c0dd TK |
134 | Print statistics in selected mode AND terse, semicolon-delimited format. |
135 | **Deprecated**, use :option:`--output-format` instead to select multiple | |
136 | formats. | |
f80dba8d | 137 | |
f50fbdda | 138 | .. option:: --terse-version=version |
f80dba8d | 139 | |
f50fbdda | 140 | Set terse `version` output format (default 3, or 2 or 4 or 5). |
f80dba8d MT |
141 | |
142 | .. option:: --version | |
143 | ||
b8f7e412 | 144 | Print version information and exit. |
f80dba8d MT |
145 | |
146 | .. option:: --help | |
147 | ||
113f0e7c | 148 | Print a summary of the command line options and exit. |
f80dba8d MT |
149 | |
150 | .. option:: --cpuclock-test | |
151 | ||
152 | Perform test and validation of internal CPU clock. | |
153 | ||
113f0e7c | 154 | .. option:: --crctest=[test] |
f80dba8d | 155 | |
b034c0dd TK |
156 | Test the speed of the built-in checksumming functions. If no argument is |
157 | given, all of them are tested. Alternatively, a comma separated list can | |
158 | be passed, in which case the given ones are tested. | |
f80dba8d MT |
159 | |
160 | .. option:: --cmdhelp=command | |
161 | ||
162 | Print help information for `command`. May be ``all`` for all commands. | |
163 | ||
164 | .. option:: --enghelp=[ioengine[,command]] | |
165 | ||
f50fbdda TK |
166 | List all commands defined by `ioengine`, or print help for `command` |
167 | defined by `ioengine`. If no `ioengine` is given, list all | |
b034c0dd | 168 | available ioengines. |
f80dba8d | 169 | |
57fd9225 | 170 | .. option:: --showcmd |
f80dba8d | 171 | |
57fd9225 | 172 | Convert given job files to a set of command-line options. |
f80dba8d MT |
173 | |
174 | .. option:: --readonly | |
175 | ||
4027b2a1 VF |
176 | Turn on safety read-only checks, preventing writes and trims. The |
177 | ``--readonly`` option is an extra safety guard to prevent users from | |
178 | accidentally starting a write or trim workload when that is not desired. | |
179 | Fio will only modify the device under test if | |
180 | `rw=write/randwrite/rw/randrw/trim/randtrim/trimwrite` is given. This | |
181 | safety net can be used as an extra precaution. | |
f80dba8d MT |
182 | |
183 | .. option:: --eta=when | |
184 | ||
b8f7e412 | 185 | Specifies when real-time ETA estimate should be printed. `when` may be |
db37d890 JA |
186 | `always`, `never` or `auto`. `auto` is the default, it prints ETA |
187 | when requested if the output is a TTY. `always` disregards the output | |
188 | type, and prints ETA when requested. `never` never prints ETA. | |
189 | ||
190 | .. option:: --eta-interval=time | |
191 | ||
192 | By default, fio requests client ETA status roughly every second. With | |
193 | this option, the interval is configurable. Fio imposes a minimum | |
194 | allowed time to avoid flooding the console, less than 250 msec is | |
195 | not supported. | |
f80dba8d MT |
196 | |
197 | .. option:: --eta-newline=time | |
198 | ||
947e0fe0 SW |
199 | Force a new line for every `time` period passed. When the unit is omitted, |
200 | the value is interpreted in seconds. | |
f80dba8d MT |
201 | |
202 | .. option:: --status-interval=time | |
203 | ||
aa6cb459 VF |
204 | Force a full status dump of cumulative (from job start) values at `time` |
205 | intervals. This option does *not* provide per-period measurements. So | |
206 | values such as bandwidth are running averages. When the time unit is omitted, | |
c1f4de8a JA |
207 | `time` is interpreted in seconds. Note that using this option with |
208 | ``--output-format=json`` will yield output that technically isn't valid | |
209 | json, since the output will be collated sets of valid json. It will need | |
210 | to be split into valid sets of json after the run. | |
f80dba8d MT |
211 | |
212 | .. option:: --section=name | |
213 | ||
b034c0dd TK |
214 | Only run specified section `name` in job file. Multiple sections can be specified. |
215 | The ``--section`` option allows one to combine related jobs into one file. | |
216 | E.g. one job file could define light, moderate, and heavy sections. Tell | |
217 | fio to run only the "heavy" section by giving ``--section=heavy`` | |
218 | command line option. One can also specify the "write" operations in one | |
219 | section and "verify" operation in another section. The ``--section`` option | |
220 | only applies to job sections. The reserved *global* section is always | |
221 | parsed and used. | |
f80dba8d MT |
222 | |
223 | .. option:: --alloc-size=kb | |
224 | ||
4a419903 VF |
225 | Allocate additional internal smalloc pools of size `kb` in KiB. The |
226 | ``--alloc-size`` option increases shared memory set aside for use by fio. | |
b034c0dd TK |
227 | If running large jobs with randommap enabled, fio can run out of memory. |
228 | Smalloc is an internal allocator for shared structures from a fixed size | |
229 | memory pool and can grow to 16 pools. The pool size defaults to 16MiB. | |
f80dba8d | 230 | |
b034c0dd TK |
231 | NOTE: While running :file:`.fio_smalloc.*` backing store files are visible |
232 | in :file:`/tmp`. | |
f80dba8d MT |
233 | |
234 | .. option:: --warnings-fatal | |
235 | ||
b034c0dd TK |
236 | All fio parser warnings are fatal, causing fio to exit with an |
237 | error. | |
f80dba8d MT |
238 | |
239 | .. option:: --max-jobs=nr | |
240 | ||
f50fbdda | 241 | Set the maximum number of threads/processes to support to `nr`. |
818322cc | 242 | NOTE: On Linux, it may be necessary to increase the shared-memory |
71aa48eb | 243 | limit (:file:`/proc/sys/kernel/shmmax`) if fio runs into errors while |
818322cc | 244 | creating jobs. |
f80dba8d MT |
245 | |
246 | .. option:: --server=args | |
247 | ||
b034c0dd TK |
248 | Start a backend server, with `args` specifying what to listen to. |
249 | See `Client/Server`_ section. | |
f80dba8d MT |
250 | |
251 | .. option:: --daemonize=pidfile | |
252 | ||
b034c0dd | 253 | Background a fio server, writing the pid to the given `pidfile` file. |
f80dba8d MT |
254 | |
255 | .. option:: --client=hostname | |
256 | ||
f50fbdda | 257 | Instead of running the jobs locally, send and run them on the given `hostname` |
71aa48eb | 258 | or set of `hostname`\s. See `Client/Server`_ section. |
f80dba8d MT |
259 | |
260 | .. option:: --remote-config=file | |
261 | ||
f50fbdda | 262 | Tell fio server to load this local `file`. |
f80dba8d MT |
263 | |
264 | .. option:: --idle-prof=option | |
265 | ||
b8f7e412 | 266 | Report CPU idleness. `option` is one of the following: |
113f0e7c SW |
267 | |
268 | **calibrate** | |
269 | Run unit work calibration only and exit. | |
270 | ||
271 | **system** | |
272 | Show aggregate system idleness and unit work. | |
273 | ||
274 | **percpu** | |
275 | As **system** but also show per CPU idleness. | |
f80dba8d MT |
276 | |
277 | .. option:: --inflate-log=log | |
278 | ||
f50fbdda | 279 | Inflate and output compressed `log`. |
f80dba8d MT |
280 | |
281 | .. option:: --trigger-file=file | |
282 | ||
f50fbdda | 283 | Execute trigger command when `file` exists. |
f80dba8d | 284 | |
f50fbdda | 285 | .. option:: --trigger-timeout=time |
f80dba8d | 286 | |
f50fbdda | 287 | Execute trigger at this `time`. |
f80dba8d | 288 | |
f50fbdda | 289 | .. option:: --trigger=command |
f80dba8d | 290 | |
f50fbdda | 291 | Set this `command` as local trigger. |
f80dba8d | 292 | |
f50fbdda | 293 | .. option:: --trigger-remote=command |
f80dba8d | 294 | |
f50fbdda | 295 | Set this `command` as remote trigger. |
f80dba8d MT |
296 | |
297 | .. option:: --aux-path=path | |
298 | ||
f4401bf8 SW |
299 | Use the directory specified by `path` for generated state files instead |
300 | of the current working directory. | |
f80dba8d MT |
301 | |
302 | Any parameters following the options will be assumed to be job files, unless | |
303 | they match a job file parameter. Multiple job files can be listed and each job | |
304 | file will be regarded as a separate group. Fio will :option:`stonewall` | |
305 | execution between each group. | |
306 | ||
307 | ||
308 | Job file format | |
309 | --------------- | |
310 | ||
311 | As previously described, fio accepts one or more job files describing what it is | |
312 | supposed to do. The job file format is the classic ini file, where the names | |
c60ebc45 | 313 | enclosed in [] brackets define the job name. You are free to use any ASCII name |
f80dba8d MT |
314 | you want, except *global* which has special meaning. Following the job name is |
315 | a sequence of zero or more parameters, one per line, that define the behavior of | |
316 | the job. If the first character in a line is a ';' or a '#', the entire line is | |
317 | discarded as a comment. | |
318 | ||
319 | A *global* section sets defaults for the jobs described in that file. A job may | |
320 | override a *global* section parameter, and a job file may even have several | |
321 | *global* sections if so desired. A job is only affected by a *global* section | |
322 | residing above it. | |
323 | ||
f50fbdda TK |
324 | The :option:`--cmdhelp` option also lists all options. If used with a `command` |
325 | argument, :option:`--cmdhelp` will detail the given `command`. | |
f80dba8d MT |
326 | |
327 | See the `examples/` directory for inspiration on how to write job files. Note | |
328 | the copyright and license requirements currently apply to `examples/` files. | |
329 | ||
330 | So let's look at a really simple job file that defines two processes, each | |
331 | randomly reading from a 128MiB file: | |
332 | ||
333 | .. code-block:: ini | |
334 | ||
335 | ; -- start job file -- | |
336 | [global] | |
337 | rw=randread | |
338 | size=128m | |
339 | ||
340 | [job1] | |
341 | ||
342 | [job2] | |
343 | ||
344 | ; -- end job file -- | |
345 | ||
346 | As you can see, the job file sections themselves are empty as all the described | |
347 | parameters are shared. As no :option:`filename` option is given, fio makes up a | |
348 | `filename` for each of the jobs as it sees fit. On the command line, this job | |
349 | would look as follows:: | |
350 | ||
351 | $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2 | |
352 | ||
353 | ||
354 | Let's look at an example that has a number of processes writing randomly to | |
355 | files: | |
356 | ||
357 | .. code-block:: ini | |
358 | ||
359 | ; -- start job file -- | |
360 | [random-writers] | |
361 | ioengine=libaio | |
362 | iodepth=4 | |
363 | rw=randwrite | |
364 | bs=32k | |
365 | direct=0 | |
366 | size=64m | |
367 | numjobs=4 | |
368 | ; -- end job file -- | |
369 | ||
370 | Here we have no *global* section, as we only have one job defined anyway. We | |
371 | want to use async I/O here, with a depth of 4 for each file. We also increased | |
372 | the buffer size used to 32KiB and define numjobs to 4 to fork 4 identical | |
373 | jobs. The result is 4 processes each randomly writing to their own 64MiB | |
374 | file. Instead of using the above job file, you could have given the parameters | |
375 | on the command line. For this case, you would specify:: | |
376 | ||
377 | $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4 | |
378 | ||
379 | When fio is utilized as a basis of any reasonably large test suite, it might be | |
380 | desirable to share a set of standardized settings across multiple job files. | |
381 | Instead of copy/pasting such settings, any section may pull in an external | |
382 | :file:`filename.fio` file with *include filename* directive, as in the following | |
383 | example:: | |
384 | ||
385 | ; -- start job file including.fio -- | |
386 | [global] | |
387 | filename=/tmp/test | |
388 | filesize=1m | |
389 | include glob-include.fio | |
390 | ||
391 | [test] | |
392 | rw=randread | |
393 | bs=4k | |
394 | time_based=1 | |
395 | runtime=10 | |
396 | include test-include.fio | |
397 | ; -- end job file including.fio -- | |
398 | ||
399 | .. code-block:: ini | |
400 | ||
401 | ; -- start job file glob-include.fio -- | |
402 | thread=1 | |
403 | group_reporting=1 | |
404 | ; -- end job file glob-include.fio -- | |
405 | ||
406 | .. code-block:: ini | |
407 | ||
408 | ; -- start job file test-include.fio -- | |
409 | ioengine=libaio | |
410 | iodepth=4 | |
411 | ; -- end job file test-include.fio -- | |
412 | ||
413 | Settings pulled into a section apply to that section only (except *global* | |
414 | section). Include directives may be nested in that any included file may contain | |
415 | further include directive(s). Include files may not contain [] sections. | |
416 | ||
417 | ||
418 | Environment variables | |
419 | ~~~~~~~~~~~~~~~~~~~~~ | |
420 | ||
421 | Fio also supports environment variable expansion in job files. Any sub-string of | |
422 | the form ``${VARNAME}`` as part of an option value (in other words, on the right | |
423 | of the '='), will be expanded to the value of the environment variable called | |
424 | `VARNAME`. If no such environment variable is defined, or `VARNAME` is the | |
425 | empty string, the empty string will be substituted. | |
426 | ||
427 | As an example, let's look at a sample fio invocation and job file:: | |
428 | ||
429 | $ SIZE=64m NUMJOBS=4 fio jobfile.fio | |
430 | ||
431 | .. code-block:: ini | |
432 | ||
433 | ; -- start job file -- | |
434 | [random-writers] | |
435 | rw=randwrite | |
436 | size=${SIZE} | |
437 | numjobs=${NUMJOBS} | |
438 | ; -- end job file -- | |
439 | ||
440 | This will expand to the following equivalent job file at runtime: | |
441 | ||
442 | .. code-block:: ini | |
443 | ||
444 | ; -- start job file -- | |
445 | [random-writers] | |
446 | rw=randwrite | |
447 | size=64m | |
448 | numjobs=4 | |
449 | ; -- end job file -- | |
450 | ||
451 | Fio ships with a few example job files, you can also look there for inspiration. | |
452 | ||
453 | Reserved keywords | |
454 | ~~~~~~~~~~~~~~~~~ | |
455 | ||
456 | Additionally, fio has a set of reserved keywords that will be replaced | |
457 | internally with the appropriate value. Those keywords are: | |
458 | ||
459 | **$pagesize** | |
460 | ||
461 | The architecture page size of the running system. | |
462 | ||
463 | **$mb_memory** | |
464 | ||
465 | Megabytes of total memory in the system. | |
466 | ||
467 | **$ncpus** | |
468 | ||
469 | Number of online available CPUs. | |
470 | ||
471 | These can be used on the command line or in the job file, and will be | |
472 | automatically substituted with the current system values when the job is | |
473 | run. Simple math is also supported on these keywords, so you can perform actions | |
474 | like:: | |
475 | ||
b034c0dd | 476 | size=8*$mb_memory |
f80dba8d MT |
477 | |
478 | and get that properly expanded to 8 times the size of memory in the machine. | |
479 | ||
480 | ||
481 | Job file parameters | |
482 | ------------------- | |
483 | ||
484 | This section describes in details each parameter associated with a job. Some | |
485 | parameters take an option of a given type, such as an integer or a | |
486 | string. Anywhere a numeric value is required, an arithmetic expression may be | |
487 | used, provided it is surrounded by parentheses. Supported operators are: | |
488 | ||
489 | - addition (+) | |
490 | - subtraction (-) | |
491 | - multiplication (*) | |
492 | - division (/) | |
493 | - modulus (%) | |
494 | - exponentiation (^) | |
495 | ||
496 | For time values in expressions, units are microseconds by default. This is | |
497 | different than for time values not in expressions (not enclosed in | |
498 | parentheses). The following types are used: | |
499 | ||
500 | ||
501 | Parameter types | |
502 | ~~~~~~~~~~~~~~~ | |
503 | ||
504 | **str** | |
b034c0dd | 505 | String: A sequence of alphanumeric characters. |
f80dba8d MT |
506 | |
507 | **time** | |
008d0feb SW |
508 | Integer with possible time suffix. Without a unit value is interpreted as |
509 | seconds unless otherwise specified. Accepts a suffix of 'd' for days, 'h' for | |
510 | hours, 'm' for minutes, 's' for seconds, 'ms' (or 'msec') for milliseconds and | |
511 | 'us' (or 'usec') for microseconds. For example, use 10m for 10 minutes. | |
f80dba8d MT |
512 | |
513 | .. _int: | |
514 | ||
515 | **int** | |
516 | Integer. A whole number value, which may contain an integer prefix | |
517 | and an integer suffix: | |
518 | ||
b034c0dd | 519 | [*integer prefix*] **number** [*integer suffix*] |
f80dba8d MT |
520 | |
521 | The optional *integer prefix* specifies the number's base. The default | |
522 | is decimal. *0x* specifies hexadecimal. | |
523 | ||
524 | The optional *integer suffix* specifies the number's units, and includes an | |
525 | optional unit prefix and an optional unit. For quantities of data, the | |
947e0fe0 SW |
526 | default unit is bytes. For quantities of time, the default unit is seconds |
527 | unless otherwise specified. | |
f80dba8d | 528 | |
9207a0cb | 529 | With :option:`kb_base`\=1000, fio follows international standards for unit |
f80dba8d MT |
530 | prefixes. To specify power-of-10 decimal values defined in the |
531 | International System of Units (SI): | |
532 | ||
eccce61a TK |
533 | * *K* -- means kilo (K) or 1000 |
534 | * *M* -- means mega (M) or 1000**2 | |
535 | * *G* -- means giga (G) or 1000**3 | |
536 | * *T* -- means tera (T) or 1000**4 | |
537 | * *P* -- means peta (P) or 1000**5 | |
f80dba8d MT |
538 | |
539 | To specify power-of-2 binary values defined in IEC 80000-13: | |
540 | ||
eccce61a TK |
541 | * *Ki* -- means kibi (Ki) or 1024 |
542 | * *Mi* -- means mebi (Mi) or 1024**2 | |
543 | * *Gi* -- means gibi (Gi) or 1024**3 | |
544 | * *Ti* -- means tebi (Ti) or 1024**4 | |
545 | * *Pi* -- means pebi (Pi) or 1024**5 | |
f80dba8d | 546 | |
193aaf6a G |
547 | For Zone Block Device Mode: |
548 | * *z* -- means Zone | |
549 | ||
9207a0cb | 550 | With :option:`kb_base`\=1024 (the default), the unit prefixes are opposite |
f80dba8d MT |
551 | from those specified in the SI and IEC 80000-13 standards to provide |
552 | compatibility with old scripts. For example, 4k means 4096. | |
553 | ||
554 | For quantities of data, an optional unit of 'B' may be included | |
b8f7e412 | 555 | (e.g., 'kB' is the same as 'k'). |
f80dba8d MT |
556 | |
557 | The *integer suffix* is not case sensitive (e.g., m/mi mean mebi/mega, | |
558 | not milli). 'b' and 'B' both mean byte, not bit. | |
559 | ||
9207a0cb | 560 | Examples with :option:`kb_base`\=1000: |
f80dba8d MT |
561 | |
562 | * *4 KiB*: 4096, 4096b, 4096B, 4ki, 4kib, 4kiB, 4Ki, 4KiB | |
563 | * *1 MiB*: 1048576, 1mi, 1024ki | |
564 | * *1 MB*: 1000000, 1m, 1000k | |
565 | * *1 TiB*: 1099511627776, 1ti, 1024gi, 1048576mi | |
566 | * *1 TB*: 1000000000, 1t, 1000m, 1000000k | |
567 | ||
9207a0cb | 568 | Examples with :option:`kb_base`\=1024 (default): |
f80dba8d MT |
569 | |
570 | * *4 KiB*: 4096, 4096b, 4096B, 4k, 4kb, 4kB, 4K, 4KB | |
571 | * *1 MiB*: 1048576, 1m, 1024k | |
572 | * *1 MB*: 1000000, 1mi, 1000ki | |
573 | * *1 TiB*: 1099511627776, 1t, 1024g, 1048576m | |
574 | * *1 TB*: 1000000000, 1ti, 1000mi, 1000000ki | |
575 | ||
576 | To specify times (units are not case sensitive): | |
577 | ||
578 | * *D* -- means days | |
579 | * *H* -- means hours | |
4502cb42 | 580 | * *M* -- means minutes |
f80dba8d MT |
581 | * *s* -- or sec means seconds (default) |
582 | * *ms* -- or *msec* means milliseconds | |
583 | * *us* -- or *usec* means microseconds | |
584 | ||
585 | If the option accepts an upper and lower range, use a colon ':' or | |
586 | minus '-' to separate such values. See :ref:`irange <irange>`. | |
4502cb42 SW |
587 | If the lower value specified happens to be larger than the upper value |
588 | the two values are swapped. | |
f80dba8d MT |
589 | |
590 | .. _bool: | |
591 | ||
592 | **bool** | |
593 | Boolean. Usually parsed as an integer, however only defined for | |
594 | true and false (1 and 0). | |
595 | ||
596 | .. _irange: | |
597 | ||
598 | **irange** | |
599 | Integer range with suffix. Allows value range to be given, such as | |
c60ebc45 | 600 | 1024-4096. A colon may also be used as the separator, e.g. 1k:4k. If the |
f80dba8d MT |
601 | option allows two sets of ranges, they can be specified with a ',' or '/' |
602 | delimiter: 1k-4k/8k-32k. Also see :ref:`int <int>`. | |
603 | ||
604 | **float_list** | |
605 | A list of floating point numbers, separated by a ':' character. | |
606 | ||
f5c3bcf2 TK |
607 | With the above in mind, here follows the complete list of fio job parameters. |
608 | ||
f80dba8d MT |
609 | |
610 | Units | |
611 | ~~~~~ | |
612 | ||
613 | .. option:: kb_base=int | |
614 | ||
615 | Select the interpretation of unit prefixes in input parameters. | |
616 | ||
617 | **1000** | |
618 | Inputs comply with IEC 80000-13 and the International | |
619 | System of Units (SI). Use: | |
620 | ||
621 | - power-of-2 values with IEC prefixes (e.g., KiB) | |
622 | - power-of-10 values with SI prefixes (e.g., kB) | |
623 | ||
624 | **1024** | |
625 | Compatibility mode (default). To avoid breaking old scripts: | |
626 | ||
627 | - power-of-2 values with SI prefixes | |
628 | - power-of-10 values with IEC prefixes | |
629 | ||
630 | See :option:`bs` for more details on input parameters. | |
631 | ||
632 | Outputs always use correct prefixes. Most outputs include both | |
633 | side-by-side, like:: | |
634 | ||
635 | bw=2383.3kB/s (2327.4KiB/s) | |
636 | ||
637 | If only one value is reported, then kb_base selects the one to use: | |
638 | ||
639 | **1000** -- SI prefixes | |
640 | ||
641 | **1024** -- IEC prefixes | |
642 | ||
643 | .. option:: unit_base=int | |
644 | ||
645 | Base unit for reporting. Allowed values are: | |
646 | ||
647 | **0** | |
648 | Use auto-detection (default). | |
649 | **8** | |
650 | Byte based. | |
651 | **1** | |
652 | Bit based. | |
653 | ||
654 | ||
f80dba8d MT |
655 | Job description |
656 | ~~~~~~~~~~~~~~~ | |
657 | ||
658 | .. option:: name=str | |
659 | ||
660 | ASCII name of the job. This may be used to override the name printed by fio | |
661 | for this job. Otherwise the job name is used. On the command line this | |
662 | parameter has the special purpose of also signaling the start of a new job. | |
663 | ||
664 | .. option:: description=str | |
665 | ||
666 | Text description of the job. Doesn't do anything except dump this text | |
667 | description when this job is run. It's not parsed. | |
668 | ||
669 | .. option:: loops=int | |
670 | ||
671 | Run the specified number of iterations of this job. Used to repeat the same | |
672 | workload a given number of times. Defaults to 1. | |
673 | ||
674 | .. option:: numjobs=int | |
675 | ||
79591fa9 TK |
676 | Create the specified number of clones of this job. Each clone of job |
677 | is spawned as an independent thread or process. May be used to setup a | |
f80dba8d MT |
678 | larger number of threads/processes doing the same thing. Each thread is |
679 | reported separately; to see statistics for all clones as a whole, use | |
680 | :option:`group_reporting` in conjunction with :option:`new_group`. | |
a47b697c | 681 | See :option:`--max-jobs`. Default: 1. |
f80dba8d MT |
682 | |
683 | ||
684 | Time related parameters | |
685 | ~~~~~~~~~~~~~~~~~~~~~~~ | |
686 | ||
687 | .. option:: runtime=time | |
688 | ||
c7927863 H |
689 | Limit runtime. The test will run until it completes the configured I/O |
690 | workload or until it has run for this specified amount of time, whichever | |
691 | occurs first. It can be quite hard to determine for how long a specified | |
692 | job will run, so this parameter is handy to cap the total runtime to a | |
693 | given time. When the unit is omitted, the value is interpreted in | |
694 | seconds. | |
f80dba8d MT |
695 | |
696 | .. option:: time_based | |
697 | ||
698 | If set, fio will run for the duration of the :option:`runtime` specified | |
699 | even if the file(s) are completely read or written. It will simply loop over | |
700 | the same workload as many times as the :option:`runtime` allows. | |
701 | ||
a881438b | 702 | .. option:: startdelay=irange(time) |
f80dba8d | 703 | |
947e0fe0 SW |
704 | Delay the start of job for the specified amount of time. Can be a single |
705 | value or a range. When given as a range, each thread will choose a value | |
706 | randomly from within the range. Value is in seconds if a unit is omitted. | |
f80dba8d MT |
707 | |
708 | .. option:: ramp_time=time | |
709 | ||
710 | If set, fio will run the specified workload for this amount of time before | |
711 | logging any performance numbers. Useful for letting performance settle | |
712 | before logging results, thus minimizing the runtime required for stable | |
713 | results. Note that the ``ramp_time`` is considered lead in time for a job, | |
714 | thus it will increase the total runtime if a special timeout or | |
f75ede1d SW |
715 | :option:`runtime` is specified. When the unit is omitted, the value is |
716 | given in seconds. | |
f80dba8d MT |
717 | |
718 | .. option:: clocksource=str | |
719 | ||
720 | Use the given clocksource as the base of timing. The supported options are: | |
721 | ||
722 | **gettimeofday** | |
723 | :manpage:`gettimeofday(2)` | |
724 | ||
725 | **clock_gettime** | |
726 | :manpage:`clock_gettime(2)` | |
727 | ||
728 | **cpu** | |
729 | Internal CPU clock source | |
730 | ||
731 | cpu is the preferred clocksource if it is reliable, as it is very fast (and | |
732 | fio is heavy on time calls). Fio will automatically use this clocksource if | |
733 | it's supported and considered reliable on the system it is running on, | |
734 | unless another clocksource is specifically set. For x86/x86-64 CPUs, this | |
735 | means supporting TSC Invariant. | |
736 | ||
737 | .. option:: gtod_reduce=bool | |
738 | ||
739 | Enable all of the :manpage:`gettimeofday(2)` reducing options | |
f75ede1d | 740 | (:option:`disable_clat`, :option:`disable_slat`, :option:`disable_bw_measurement`) plus |
f80dba8d MT |
741 | reduce precision of the timeout somewhat to really shrink the |
742 | :manpage:`gettimeofday(2)` call count. With this option enabled, we only do | |
743 | about 0.4% of the :manpage:`gettimeofday(2)` calls we would have done if all | |
744 | time keeping was enabled. | |
745 | ||
746 | .. option:: gtod_cpu=int | |
747 | ||
748 | Sometimes it's cheaper to dedicate a single thread of execution to just | |
749 | getting the current time. Fio (and databases, for instance) are very | |
750 | intensive on :manpage:`gettimeofday(2)` calls. With this option, you can set | |
751 | one CPU aside for doing nothing but logging current time to a shared memory | |
752 | location. Then the other threads/processes that run I/O workloads need only | |
753 | copy that segment, instead of entering the kernel with a | |
754 | :manpage:`gettimeofday(2)` call. The CPU set aside for doing these time | |
755 | calls will be excluded from other uses. Fio will manually clear it from the | |
756 | CPU mask of other jobs. | |
757 | ||
12d325ca | 758 | .. option:: job_start_clock_id=int |
97ef7f3a VF |
759 | |
760 | The clock_id passed to the call to `clock_gettime` used to record | |
761 | job_start in the `json` output format. Default is 0, or CLOCK_REALTIME. | |
12d325ca | 762 | |
f80dba8d MT |
763 | |
764 | Target file/device | |
765 | ~~~~~~~~~~~~~~~~~~ | |
766 | ||
767 | .. option:: directory=str | |
768 | ||
769 | Prefix filenames with this directory. Used to place files in a different | |
770 | location than :file:`./`. You can specify a number of directories by | |
771 | separating the names with a ':' character. These directories will be | |
02dd2689 | 772 | assigned equally distributed to job clones created by :option:`numjobs` as |
f80dba8d MT |
773 | long as they are using generated filenames. If specific `filename(s)` are |
774 | set fio will use the first listed directory, and thereby matching the | |
f4401bf8 SW |
775 | `filename` semantic (which generates a file for each clone if not |
776 | specified, but lets all clones use the same file if set). | |
f80dba8d | 777 | |
3b803fe1 SW |
778 | See the :option:`filename` option for information on how to escape "``:``" |
779 | characters within the directory path itself. | |
f80dba8d | 780 | |
f4401bf8 SW |
781 | Note: To control the directory fio will use for internal state files |
782 | use :option:`--aux-path`. | |
783 | ||
f80dba8d MT |
784 | .. option:: filename=str |
785 | ||
786 | Fio normally makes up a `filename` based on the job name, thread number, and | |
02dd2689 SW |
787 | file number (see :option:`filename_format`). If you want to share files |
788 | between threads in a job or several | |
79591fa9 TK |
789 | jobs with fixed file paths, specify a `filename` for each of them to override |
790 | the default. If the ioengine is file based, you can specify a number of files | |
791 | by separating the names with a ':' colon. So if you wanted a job to open | |
792 | :file:`/dev/sda` and :file:`/dev/sdb` as the two working files, you would use | |
793 | ``filename=/dev/sda:/dev/sdb``. This also means that whenever this option is | |
794 | specified, :option:`nrfiles` is ignored. The size of regular files specified | |
02dd2689 | 795 | by this option will be :option:`size` divided by number of files unless an |
79591fa9 TK |
796 | explicit size is specified by :option:`filesize`. |
797 | ||
3b803fe1 | 798 | Each colon in the wanted path must be escaped with a ``\`` |
02dd2689 SW |
799 | character. For instance, if the path is :file:`/dev/dsk/foo@3,0:c` then you |
800 | would use ``filename=/dev/dsk/foo@3,0\:c`` and if the path is | |
3b803fe1 | 801 | :file:`F:\\filename` then you would use ``filename=F\:\filename``. |
02dd2689 | 802 | |
f80dba8d MT |
803 | On Windows, disk devices are accessed as :file:`\\\\.\\PhysicalDrive0` for |
804 | the first device, :file:`\\\\.\\PhysicalDrive1` for the second etc. | |
a8ce8eb3 | 805 | Note: Windows and FreeBSD (refer to geom(4)) prevent write access to areas |
02dd2689 SW |
806 | of the disk containing in-use data (e.g. filesystems). |
807 | ||
808 | The filename "`-`" is a reserved name, meaning *stdin* or *stdout*. Which | |
809 | of the two depends on the read/write direction set. | |
f80dba8d MT |
810 | |
811 | .. option:: filename_format=str | |
812 | ||
813 | If sharing multiple files between jobs, it is usually necessary to have fio | |
814 | generate the exact names that you want. By default, fio will name a file | |
815 | based on the default file format specification of | |
816 | :file:`jobname.jobnumber.filenumber`. With this option, that can be | |
817 | customized. Fio will recognize and replace the following keywords in this | |
818 | string: | |
819 | ||
820 | **$jobname** | |
821 | The name of the worker thread or process. | |
8d53c5f8 TG |
822 | **$clientuid** |
823 | IP of the fio process when using client/server mode. | |
f80dba8d MT |
824 | **$jobnum** |
825 | The incremental number of the worker thread or process. | |
826 | **$filenum** | |
827 | The incremental number of the file for that worker thread or | |
828 | process. | |
829 | ||
830 | To have dependent jobs share a set of files, this option can be set to have | |
831 | fio generate filenames that are shared between the two. For instance, if | |
832 | :file:`testfiles.$filenum` is specified, file number 4 for any job will be | |
833 | named :file:`testfiles.4`. The default of :file:`$jobname.$jobnum.$filenum` | |
834 | will be used if no other format specifier is given. | |
835 | ||
645943c0 JB |
836 | If you specify a path then the directories will be created up to the |
837 | main directory for the file. So for example if you specify | |
838 | ``filename_format=a/b/c/$jobnum`` then the directories a/b/c will be | |
839 | created before the file setup part of the job. If you specify | |
840 | :option:`directory` then the path will be relative that directory, | |
841 | otherwise it is treated as the absolute path. | |
842 | ||
f80dba8d MT |
843 | .. option:: unique_filename=bool |
844 | ||
845 | To avoid collisions between networked clients, fio defaults to prefixing any | |
846 | generated filenames (with a directory specified) with the source of the | |
847 | client connecting. To disable this behavior, set this option to 0. | |
848 | ||
849 | .. option:: opendir=str | |
850 | ||
8ce9c400 VF |
851 | Recursively open any files below directory `str`. This accepts only a |
852 | single directory and unlike related options, colons appearing in the | |
853 | path must not be escaped. | |
f80dba8d MT |
854 | |
855 | .. option:: lockfile=str | |
856 | ||
857 | Fio defaults to not locking any files before it does I/O to them. If a file | |
858 | or file descriptor is shared, fio can serialize I/O to that file to make the | |
859 | end result consistent. This is usual for emulating real workloads that share | |
860 | files. The lock modes are: | |
861 | ||
862 | **none** | |
863 | No locking. The default. | |
864 | **exclusive** | |
865 | Only one thread or process may do I/O at a time, excluding all | |
866 | others. | |
867 | **readwrite** | |
868 | Read-write locking on the file. Many readers may | |
869 | access the file at the same time, but writes get exclusive access. | |
870 | ||
871 | .. option:: nrfiles=int | |
872 | ||
79591fa9 TK |
873 | Number of files to use for this job. Defaults to 1. The size of files |
874 | will be :option:`size` divided by this unless explicit size is specified by | |
875 | :option:`filesize`. Files are created for each thread separately, and each | |
876 | file will have a file number within its name by default, as explained in | |
877 | :option:`filename` section. | |
878 | ||
f80dba8d MT |
879 | |
880 | .. option:: openfiles=int | |
881 | ||
882 | Number of files to keep open at the same time. Defaults to the same as | |
883 | :option:`nrfiles`, can be set smaller to limit the number simultaneous | |
884 | opens. | |
885 | ||
886 | .. option:: file_service_type=str | |
887 | ||
888 | Defines how fio decides which file from a job to service next. The following | |
889 | types are defined: | |
890 | ||
891 | **random** | |
892 | Choose a file at random. | |
893 | ||
894 | **roundrobin** | |
895 | Round robin over opened files. This is the default. | |
896 | ||
897 | **sequential** | |
898 | Finish one file before moving on to the next. Multiple files can | |
f50fbdda | 899 | still be open depending on :option:`openfiles`. |
f80dba8d MT |
900 | |
901 | **zipf** | |
c60ebc45 | 902 | Use a *Zipf* distribution to decide what file to access. |
f80dba8d MT |
903 | |
904 | **pareto** | |
c60ebc45 | 905 | Use a *Pareto* distribution to decide what file to access. |
f80dba8d | 906 | |
dd3503d3 | 907 | **normal** |
c60ebc45 | 908 | Use a *Gaussian* (normal) distribution to decide what file to |
f80dba8d MT |
909 | access. |
910 | ||
dd3503d3 SW |
911 | **gauss** |
912 | Alias for normal. | |
913 | ||
f80dba8d MT |
914 | For *random*, *roundrobin*, and *sequential*, a postfix can be appended to |
915 | tell fio how many I/Os to issue before switching to a new file. For example, | |
916 | specifying ``file_service_type=random:8`` would cause fio to issue | |
917 | 8 I/Os before selecting a new file at random. For the non-uniform | |
918 | distributions, a floating point postfix can be given to influence how the | |
919 | distribution is skewed. See :option:`random_distribution` for a description | |
920 | of how that would work. | |
921 | ||
922 | .. option:: ioscheduler=str | |
923 | ||
924 | Attempt to switch the device hosting the file to the specified I/O scheduler | |
925 | before running. | |
926 | ||
927 | .. option:: create_serialize=bool | |
928 | ||
929 | If true, serialize the file creation for the jobs. This may be handy to | |
930 | avoid interleaving of data files, which may greatly depend on the filesystem | |
a47b697c | 931 | used and even the number of processors in the system. Default: true. |
f80dba8d MT |
932 | |
933 | .. option:: create_fsync=bool | |
934 | ||
22413915 | 935 | :manpage:`fsync(2)` the data file after creation. This is the default. |
f80dba8d MT |
936 | |
937 | .. option:: create_on_open=bool | |
938 | ||
730bd7d9 SW |
939 | If true, don't pre-create files but allow the job's open() to create a file |
940 | when it's time to do I/O. Default: false -- pre-create all necessary files | |
941 | when the job starts. | |
f80dba8d MT |
942 | |
943 | .. option:: create_only=bool | |
944 | ||
945 | If true, fio will only run the setup phase of the job. If files need to be | |
4502cb42 | 946 | laid out or updated on disk, only that will be done -- the actual job contents |
a47b697c | 947 | are not executed. Default: false. |
f80dba8d MT |
948 | |
949 | .. option:: allow_file_create=bool | |
950 | ||
730bd7d9 SW |
951 | If true, fio is permitted to create files as part of its workload. If this |
952 | option is false, then fio will error out if | |
f80dba8d MT |
953 | the files it needs to use don't already exist. Default: true. |
954 | ||
955 | .. option:: allow_mounted_write=bool | |
956 | ||
c60ebc45 | 957 | If this isn't set, fio will abort jobs that are destructive (e.g. that write) |
f80dba8d MT |
958 | to what appears to be a mounted device or partition. This should help catch |
959 | creating inadvertently destructive tests, not realizing that the test will | |
b1db0375 TK |
960 | destroy data on the mounted file system. Note that some platforms don't allow |
961 | writing against a mounted device regardless of this option. Default: false. | |
f80dba8d MT |
962 | |
963 | .. option:: pre_read=bool | |
964 | ||
965 | If this is given, files will be pre-read into memory before starting the | |
966 | given I/O operation. This will also clear the :option:`invalidate` flag, | |
967 | since it is pointless to pre-read and then drop the cache. This will only | |
968 | work for I/O engines that are seek-able, since they allow you to read the | |
a47b697c SW |
969 | same data multiple times. Thus it will not work on non-seekable I/O engines |
970 | (e.g. network, splice). Default: false. | |
f80dba8d MT |
971 | |
972 | .. option:: unlink=bool | |
973 | ||
3fe133a5 | 974 | Unlink (delete) the job files when done. Not the default, as repeated runs of that |
a47b697c SW |
975 | job would then waste time recreating the file set again and again. Default: |
976 | false. | |
f80dba8d MT |
977 | |
978 | .. option:: unlink_each_loop=bool | |
979 | ||
3fe133a5 | 980 | Unlink (delete) job files after each iteration or loop. Default: false. |
f80dba8d | 981 | |
7b865a2f BVA |
982 | .. option:: zonemode=str |
983 | ||
984 | Accepted values are: | |
985 | ||
986 | **none** | |
b8dd9750 | 987 | The :option:`zonerange`, :option:`zonesize`, |
4eef23f6 | 988 | :option:`zonecapacity` and :option:`zoneskip` |
b8dd9750 | 989 | parameters are ignored. |
7b865a2f BVA |
990 | **strided** |
991 | I/O happens in a single zone until | |
992 | :option:`zonesize` bytes have been transferred. | |
993 | After that number of bytes has been | |
994 | transferred processing of the next zone | |
4eef23f6 | 995 | starts. :option:`zonecapacity` is ignored. |
7b865a2f BVA |
996 | **zbd** |
997 | Zoned block device mode. I/O happens | |
998 | sequentially in each zone, even if random I/O | |
999 | has been selected. Random I/O happens across | |
1000 | all zones instead of being restricted to a | |
1001 | single zone. The :option:`zoneskip` parameter | |
1002 | is ignored. :option:`zonerange` and | |
1003 | :option:`zonesize` must be identical. | |
2455851d SK |
1004 | Trim is handled using a zone reset operation. |
1005 | Trim only considers non-empty sequential write | |
1006 | required and sequential write preferred zones. | |
7b865a2f | 1007 | |
5faddc64 | 1008 | .. option:: zonerange=int |
f80dba8d | 1009 | |
7b865a2f BVA |
1010 | Size of a single zone. See also :option:`zonesize` and |
1011 | :option:`zoneskip`. | |
f80dba8d | 1012 | |
5faddc64 | 1013 | .. option:: zonesize=int |
f80dba8d | 1014 | |
7b865a2f BVA |
1015 | For :option:`zonemode` =strided, this is the number of bytes to |
1016 | transfer before skipping :option:`zoneskip` bytes. If this parameter | |
1017 | is smaller than :option:`zonerange` then only a fraction of each zone | |
1018 | with :option:`zonerange` bytes will be accessed. If this parameter is | |
1019 | larger than :option:`zonerange` then each zone will be accessed | |
1020 | multiple times before skipping to the next zone. | |
1021 | ||
1022 | For :option:`zonemode` =zbd, this is the size of a single zone. The | |
1023 | :option:`zonerange` parameter is ignored in this mode. | |
f80dba8d | 1024 | |
b8dd9750 HH |
1025 | |
1026 | .. option:: zonecapacity=int | |
1027 | ||
1028 | For :option:`zonemode` =zbd, this defines the capacity of a single zone, | |
1029 | which is the accessible area starting from the zone start address. | |
1030 | This parameter only applies when using :option:`zonemode` =zbd in | |
1031 | combination with regular block devices. If not specified it defaults to | |
1032 | the zone size. If the target device is a zoned block device, the zone | |
1033 | capacity is obtained from the device information and this option is | |
1034 | ignored. | |
1035 | ||
f80dba8d MT |
1036 | .. option:: zoneskip=int |
1037 | ||
7b865a2f BVA |
1038 | For :option:`zonemode` =strided, the number of bytes to skip after |
1039 | :option:`zonesize` bytes of data have been transferred. This parameter | |
1040 | must be zero for :option:`zonemode` =zbd. | |
f80dba8d | 1041 | |
bfbdd35b BVA |
1042 | .. option:: read_beyond_wp=bool |
1043 | ||
1044 | This parameter applies to :option:`zonemode` =zbd only. | |
1045 | ||
1046 | Zoned block devices are block devices that consist of multiple zones. | |
1047 | Each zone has a type, e.g. conventional or sequential. A conventional | |
1048 | zone can be written at any offset that is a multiple of the block | |
1049 | size. Sequential zones must be written sequentially. The position at | |
1050 | which a write must occur is called the write pointer. A zoned block | |
1051 | device can be either drive managed, host managed or host aware. For | |
1052 | host managed devices the host must ensure that writes happen | |
1053 | sequentially. Fio recognizes host managed devices and serializes | |
1054 | writes to sequential zones for these devices. | |
1055 | ||
1056 | If a read occurs in a sequential zone beyond the write pointer then | |
1057 | the zoned block device will complete the read without reading any data | |
1058 | from the storage medium. Since such reads lead to unrealistically high | |
1059 | bandwidth and IOPS numbers fio only reads beyond the write pointer if | |
1060 | explicitly told to do so. Default: false. | |
1061 | ||
59b07544 BVA |
1062 | .. option:: max_open_zones=int |
1063 | ||
23a846b3 SK |
1064 | When a zone of a zoned block device is partially written (i.e. not all |
1065 | sectors of the zone have been written), the zone is in one of three | |
1066 | conditions: 'implicit open', 'explicit open' or 'closed'. Zoned block | |
1067 | devices may have a limit called 'max_open_zones' (same name as the | |
1068 | parameter) on the total number of zones that can simultaneously be in | |
1069 | the 'implicit open' or 'explicit open' conditions. Zoned block devices | |
1070 | may have another limit called 'max_active_zones', on the total number of | |
1071 | zones that can simultaneously be in the three conditions. The | |
1072 | :option:`max_open_zones` parameter limits the number of zones to which | |
1073 | write commands are issued by all fio jobs, that is, limits the number of | |
1074 | zones that will be in the conditions. When the device has the | |
1075 | max_open_zones limit and does not have the max_active_zones limit, the | |
1076 | :option:`max_open_zones` parameter limits the number of zones in the two | |
1077 | open conditions up to the limit. In this case, fio includes zones in the | |
1078 | two open conditions to the write target zones at fio start. When the | |
1079 | device has both the max_open_zones and the max_active_zones limits, the | |
1080 | :option:`max_open_zones` parameter limits the number of zones in the | |
1081 | three conditions up to the limit. In this case, fio includes zones in | |
1082 | the three conditions to the write target zones at fio start. | |
1083 | ||
1084 | This parameter is relevant only if the :option:`zonemode` =zbd is used. | |
1085 | The default value is always equal to the max_open_zones limit of the | |
1086 | target zoned block device and a value higher than this limit cannot be | |
1087 | specified by users unless the option :option:`ignore_zone_limits` is | |
1088 | specified. When :option:`ignore_zone_limits` is specified or the target | |
1089 | device does not have the max_open_zones limit, :option:`max_open_zones` | |
1090 | can specify 0 to disable any limit on the number of zones that can be | |
1091 | simultaneously written to by all jobs. | |
59b07544 | 1092 | |
3b78a972 AK |
1093 | .. option:: job_max_open_zones=int |
1094 | ||
1c97d909 SK |
1095 | In the same manner as :option:`max_open_zones`, limit the number of open |
1096 | zones per fio job, that is, the number of zones that a single job can | |
1097 | simultaneously write to. A value of zero indicates no limit. | |
1098 | Default: zero. | |
3b78a972 | 1099 | |
12324d56 | 1100 | .. option:: ignore_zone_limits=bool |
a3a6f105 | 1101 | |
12324d56 DLM |
1102 | If this option is used, fio will ignore the maximum number of open |
1103 | zones limit of the zoned block device in use, thus allowing the | |
1104 | option :option:`max_open_zones` value to be larger than the device | |
1105 | reported limit. Default: false. | |
1106 | ||
a7c2b6fc BVA |
1107 | .. option:: zone_reset_threshold=float |
1108 | ||
d65625eb SK |
1109 | A number between zero and one that indicates the ratio of written bytes |
1110 | in the zones with write pointers in the IO range to the size of the IO | |
1111 | range. When current ratio is above this ratio, zones are reset | |
b3e9bd03 SK |
1112 | periodically as :option:`zone_reset_frequency` specifies. If there are |
1113 | multiple jobs when using this option, the IO range for all write jobs | |
1114 | has to be the same. | |
a7c2b6fc BVA |
1115 | |
1116 | .. option:: zone_reset_frequency=float | |
1117 | ||
1118 | A number between zero and one that indicates how often a zone reset | |
1119 | should be issued if the zone reset threshold has been exceeded. A zone | |
1120 | reset is submitted after each (1 / zone_reset_frequency) write | |
1121 | requests. This and the previous parameter can be used to simulate | |
1122 | garbage collection activity. | |
1123 | ||
f80dba8d MT |
1124 | |
1125 | I/O type | |
1126 | ~~~~~~~~ | |
1127 | ||
1128 | .. option:: direct=bool | |
1129 | ||
1130 | If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that | |
8e889110 | 1131 | OpenBSD and ZFS on Solaris don't support direct I/O. On Windows the synchronous |
f80dba8d MT |
1132 | ioengines don't support direct I/O. Default: false. |
1133 | ||
f80dba8d MT |
1134 | .. option:: buffered=bool |
1135 | ||
1136 | If value is true, use buffered I/O. This is the opposite of the | |
1137 | :option:`direct` option. Defaults to true. | |
1138 | ||
1139 | .. option:: readwrite=str, rw=str | |
1140 | ||
1141 | Type of I/O pattern. Accepted values are: | |
1142 | ||
1143 | **read** | |
1144 | Sequential reads. | |
1145 | **write** | |
1146 | Sequential writes. | |
1147 | **trim** | |
3740cfc8 VF |
1148 | Sequential trims (Linux block devices and SCSI |
1149 | character devices only). | |
f80dba8d MT |
1150 | **randread** |
1151 | Random reads. | |
2831be97 SW |
1152 | **randwrite** |
1153 | Random writes. | |
f80dba8d | 1154 | **randtrim** |
3740cfc8 VF |
1155 | Random trims (Linux block devices and SCSI |
1156 | character devices only). | |
f80dba8d MT |
1157 | **rw,readwrite** |
1158 | Sequential mixed reads and writes. | |
1159 | **randrw** | |
1160 | Random mixed reads and writes. | |
1161 | **trimwrite** | |
1162 | Sequential trim+write sequences. Blocks will be trimmed first, | |
08996af4 AK |
1163 | then the same blocks will be written to. So if ``io_size=64K`` |
1164 | is specified, Fio will trim a total of 64K bytes and also | |
1165 | write 64K bytes on the same trimmed blocks. This behaviour | |
1166 | will be consistent with ``number_ios`` or other Fio options | |
1167 | limiting the total bytes or number of I/O's. | |
c16dc793 JA |
1168 | **randtrimwrite** |
1169 | Like trimwrite, but uses random offsets rather | |
1170 | than sequential writes. | |
f80dba8d MT |
1171 | |
1172 | Fio defaults to read if the option is not specified. For the mixed I/O | |
1173 | types, the default is to split them 50/50. For certain types of I/O the | |
730bd7d9 SW |
1174 | result may still be skewed a bit, since the speed may be different. |
1175 | ||
1176 | It is possible to specify the number of I/Os to do before getting a new | |
1177 | offset by appending ``:<nr>`` to the end of the string given. For a | |
f80dba8d MT |
1178 | random read, it would look like ``rw=randread:8`` for passing in an offset |
1179 | modifier with a value of 8. If the suffix is used with a sequential I/O | |
730bd7d9 SW |
1180 | pattern, then the *<nr>* value specified will be **added** to the generated |
1181 | offset for each I/O turning sequential I/O into sequential I/O with holes. | |
1182 | For instance, using ``rw=write:4k`` will skip 4k for every write. Also see | |
1183 | the :option:`rw_sequencer` option. | |
f80dba8d MT |
1184 | |
1185 | .. option:: rw_sequencer=str | |
1186 | ||
1187 | If an offset modifier is given by appending a number to the ``rw=<str>`` | |
1188 | line, then this option controls how that number modifies the I/O offset | |
1189 | being generated. Accepted values are: | |
1190 | ||
1191 | **sequential** | |
1192 | Generate sequential offset. | |
1193 | **identical** | |
1194 | Generate the same offset. | |
1195 | ||
1196 | ``sequential`` is only useful for random I/O, where fio would normally | |
537e0d23 AK |
1197 | generate a new random offset for every I/O. If you append e.g. 8 to |
1198 | randread, i.e. ``rw=randread:8`` you would get a new random offset for | |
1199 | every 8 I/Os. The result would be a sequence of 8 sequential offsets | |
1200 | with a random starting point. However this behavior may change if a | |
1201 | sequential I/O reaches end of the file. As sequential I/O is already | |
1202 | sequential, setting ``sequential`` for that would not result in any | |
1203 | difference. ``identical`` behaves in a similar fashion, except it sends | |
1204 | the same offset 8 number of times before generating a new offset. | |
1205 | ||
1206 | Example #1:: | |
1207 | ||
1208 | rw=randread:8 | |
1209 | rw_sequencer=sequential | |
1210 | bs=4k | |
1211 | ||
1212 | The generated sequence of offsets will look like this: | |
1213 | 4k, 8k, 12k, 16k, 20k, 24k, 28k, 32k, 92k, 96k, 100k, 104k, 108k, | |
1214 | 112k, 116k, 120k, 48k, 52k ... | |
1215 | ||
1216 | Example #2:: | |
1217 | ||
1218 | rw=randread:8 | |
1219 | rw_sequencer=identical | |
1220 | bs=4k | |
1221 | ||
1222 | The generated sequence of offsets will look like this: | |
1223 | 4k, 4k, 4k, 4k, 4k, 4k, 4k, 4k, 92k, 92k, 92k, 92k, 92k, 92k, 92k, 92k, | |
1224 | 48k, 48k, 48k ... | |
f80dba8d | 1225 | |
5cb8a8cd | 1226 | .. option:: unified_rw_reporting=str |
f80dba8d MT |
1227 | |
1228 | Fio normally reports statistics on a per data direction basis, meaning that | |
5cb8a8cd BP |
1229 | reads, writes, and trims are accounted and reported separately. This option |
1230 | determines whether fio reports the results normally, summed together, or as | |
1231 | both options. | |
1232 | Accepted values are: | |
1233 | ||
1234 | **none** | |
1235 | Normal statistics reporting. | |
1236 | ||
1237 | **mixed** | |
1238 | Statistics are summed per data direction and reported together. | |
1239 | ||
1240 | **both** | |
1241 | Statistics are reported normally, followed by the mixed statistics. | |
1242 | ||
1243 | **0** | |
1244 | Backward-compatible alias for **none**. | |
1245 | ||
1246 | **1** | |
1247 | Backward-compatible alias for **mixed**. | |
9326926b | 1248 | |
5cb8a8cd BP |
1249 | **2** |
1250 | Alias for **both**. | |
f80dba8d MT |
1251 | |
1252 | .. option:: randrepeat=bool | |
1253 | ||
7624d589 VF |
1254 | Seed all random number generators in a predictable way so the pattern |
1255 | is repeatable across runs. Default: true. | |
f80dba8d MT |
1256 | |
1257 | .. option:: allrandrepeat=bool | |
1258 | ||
7624d589 | 1259 | Alias for :option:`randrepeat`. Default: true. |
f80dba8d MT |
1260 | |
1261 | .. option:: randseed=int | |
1262 | ||
1263 | Seed the random number generators based on this seed value, to be able to | |
1264 | control what sequence of output is being generated. If not set, the random | |
1265 | sequence depends on the :option:`randrepeat` setting. | |
1266 | ||
1267 | .. option:: fallocate=str | |
1268 | ||
1269 | Whether pre-allocation is performed when laying down files. | |
1270 | Accepted values are: | |
1271 | ||
1272 | **none** | |
1273 | Do not pre-allocate space. | |
1274 | ||
2c3e17be SW |
1275 | **native** |
1276 | Use a platform's native pre-allocation call but fall back to | |
1277 | **none** behavior if it fails/is not implemented. | |
1278 | ||
f80dba8d MT |
1279 | **posix** |
1280 | Pre-allocate via :manpage:`posix_fallocate(3)`. | |
1281 | ||
1282 | **keep** | |
1283 | Pre-allocate via :manpage:`fallocate(2)` with | |
1284 | FALLOC_FL_KEEP_SIZE set. | |
1285 | ||
38ca5f03 TV |
1286 | **truncate** |
1287 | Extend file to final size via :manpage:`ftruncate(2)` | |
1288 | instead of allocating. | |
1289 | ||
f80dba8d MT |
1290 | **0** |
1291 | Backward-compatible alias for **none**. | |
1292 | ||
1293 | **1** | |
1294 | Backward-compatible alias for **posix**. | |
1295 | ||
1296 | May not be available on all supported platforms. **keep** is only available | |
2c3e17be SW |
1297 | on Linux. If using ZFS on Solaris this cannot be set to **posix** |
1298 | because ZFS doesn't support pre-allocation. Default: **native** if any | |
38ca5f03 TV |
1299 | pre-allocation methods except **truncate** are available, **none** if not. |
1300 | ||
1301 | Note that using **truncate** on Windows will interact surprisingly | |
1302 | with non-sequential write patterns. When writing to a file that has | |
1303 | been extended by setting the end-of-file information, Windows will | |
1304 | backfill the unwritten portion of the file up to that offset with | |
1305 | zeroes before issuing the new write. This means that a single small | |
1306 | write to the end of an extended file will stall until the entire | |
1307 | file has been filled with zeroes. | |
f80dba8d MT |
1308 | |
1309 | .. option:: fadvise_hint=str | |
1310 | ||
c712c97a JA |
1311 | Use :manpage:`posix_fadvise(2)` or :manpage:`posix_fadvise(2)` to |
1312 | advise the kernel on what I/O patterns are likely to be issued. | |
1313 | Accepted values are: | |
f80dba8d MT |
1314 | |
1315 | **0** | |
1316 | Backwards-compatible hint for "no hint". | |
1317 | ||
1318 | **1** | |
1319 | Backwards compatible hint for "advise with fio workload type". This | |
1320 | uses **FADV_RANDOM** for a random workload, and **FADV_SEQUENTIAL** | |
1321 | for a sequential workload. | |
1322 | ||
1323 | **sequential** | |
1324 | Advise using **FADV_SEQUENTIAL**. | |
1325 | ||
1326 | **random** | |
1327 | Advise using **FADV_RANDOM**. | |
1328 | ||
109aad50 YX |
1329 | **noreuse** |
1330 | Advise using **FADV_NOREUSE**. This may be a no-op on older Linux | |
1331 | kernels. Since Linux 6.3, it provides a hint to the LRU algorithm. | |
1332 | See the :manpage:`posix_fadvise(2)` man page. | |
1333 | ||
8f4b9f24 | 1334 | .. option:: write_hint=str |
f80dba8d | 1335 | |
8f4b9f24 JA |
1336 | Use :manpage:`fcntl(2)` to advise the kernel what life time to expect |
1337 | from a write. Only supported on Linux, as of version 4.13. Accepted | |
1338 | values are: | |
1339 | ||
1340 | **none** | |
1341 | No particular life time associated with this file. | |
1342 | ||
1343 | **short** | |
1344 | Data written to this file has a short life time. | |
1345 | ||
1346 | **medium** | |
1347 | Data written to this file has a medium life time. | |
1348 | ||
1349 | **long** | |
1350 | Data written to this file has a long life time. | |
1351 | ||
1352 | **extreme** | |
1353 | Data written to this file has a very long life time. | |
1354 | ||
1355 | The values are all relative to each other, and no absolute meaning | |
1356 | should be associated with them. | |
f80dba8d MT |
1357 | |
1358 | .. option:: offset=int | |
1359 | ||
82dbb8cb | 1360 | Start I/O at the provided offset in the file, given as either a fixed size in |
193aaf6a | 1361 | bytes, zones or a percentage. If a percentage is given, the generated offset will be |
83c8b093 JF |
1362 | aligned to the minimum ``blocksize`` or to the value of ``offset_align`` if |
1363 | provided. Data before the given offset will not be touched. This | |
89978a6b BW |
1364 | effectively caps the file size at `real_size - offset`. Can be combined with |
1365 | :option:`size` to constrain the start and end range of the I/O workload. | |
44bb1142 | 1366 | A percentage can be specified by a number between 1 and 100 followed by '%', |
adcc0730 | 1367 | for example, ``offset=20%`` to specify 20%. In ZBD mode, value can be set as |
193aaf6a | 1368 | number of zones using 'z'. |
f80dba8d | 1369 | |
83c8b093 JF |
1370 | .. option:: offset_align=int |
1371 | ||
1372 | If set to non-zero value, the byte offset generated by a percentage ``offset`` | |
1373 | is aligned upwards to this value. Defaults to 0 meaning that a percentage | |
1374 | offset is aligned to the minimum block size. | |
1375 | ||
f80dba8d MT |
1376 | .. option:: offset_increment=int |
1377 | ||
1378 | If this is provided, then the real offset becomes `offset + offset_increment | |
1379 | * thread_number`, where the thread number is a counter that starts at 0 and | |
1380 | is incremented for each sub-job (i.e. when :option:`numjobs` option is | |
1381 | specified). This option is useful if there are several jobs which are | |
1382 | intended to operate on a file in parallel disjoint segments, with even | |
0b288ba1 VF |
1383 | spacing between the starting points. Percentages can be used for this option. |
1384 | If a percentage is given, the generated offset will be aligned to the minimum | |
193aaf6a G |
1385 | ``blocksize`` or to the value of ``offset_align`` if provided. In ZBD mode, value can |
1386 | also be set as number of zones using 'z'. | |
f80dba8d MT |
1387 | |
1388 | .. option:: number_ios=int | |
1389 | ||
c60ebc45 | 1390 | Fio will normally perform I/Os until it has exhausted the size of the region |
f80dba8d MT |
1391 | set by :option:`size`, or if it exhaust the allocated time (or hits an error |
1392 | condition). With this setting, the range/size can be set independently of | |
c60ebc45 | 1393 | the number of I/Os to perform. When fio reaches this number, it will exit |
f80dba8d MT |
1394 | normally and report status. Note that this does not extend the amount of I/O |
1395 | that will be done, it will only stop fio if this condition is met before | |
1396 | other end-of-job criteria. | |
1397 | ||
1398 | .. option:: fsync=int | |
1399 | ||
730bd7d9 SW |
1400 | If writing to a file, issue an :manpage:`fsync(2)` (or its equivalent) of |
1401 | the dirty data for every number of blocks given. For example, if you give 32 | |
1402 | as a parameter, fio will sync the file after every 32 writes issued. If fio is | |
1403 | using non-buffered I/O, we may not sync the file. The exception is the sg | |
1404 | I/O engine, which synchronizes the disk cache anyway. Defaults to 0, which | |
1405 | means fio does not periodically issue and wait for a sync to complete. Also | |
1406 | see :option:`end_fsync` and :option:`fsync_on_close`. | |
f80dba8d MT |
1407 | |
1408 | .. option:: fdatasync=int | |
1409 | ||
1410 | Like :option:`fsync` but uses :manpage:`fdatasync(2)` to only sync data and | |
2550c71f | 1411 | not metadata blocks. In Windows, DragonFlyBSD or OSX there is no |
730bd7d9 SW |
1412 | :manpage:`fdatasync(2)` so this falls back to using :manpage:`fsync(2)`. |
1413 | Defaults to 0, which means fio does not periodically issue and wait for a | |
1414 | data-only sync to complete. | |
f80dba8d MT |
1415 | |
1416 | .. option:: write_barrier=int | |
1417 | ||
2831be97 | 1418 | Make every `N-th` write a barrier write. |
f80dba8d | 1419 | |
f50fbdda | 1420 | .. option:: sync_file_range=str:int |
f80dba8d | 1421 | |
f50fbdda | 1422 | Use :manpage:`sync_file_range(2)` for every `int` number of write |
f80dba8d MT |
1423 | operations. Fio will track range of writes that have happened since the last |
1424 | :manpage:`sync_file_range(2)` call. `str` can currently be one or more of: | |
1425 | ||
1426 | **wait_before** | |
1427 | SYNC_FILE_RANGE_WAIT_BEFORE | |
1428 | **write** | |
1429 | SYNC_FILE_RANGE_WRITE | |
1430 | **wait_after** | |
1431 | SYNC_FILE_RANGE_WAIT_AFTER | |
1432 | ||
1433 | So if you do ``sync_file_range=wait_before,write:8``, fio would use | |
1434 | ``SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE`` for every 8 | |
1435 | writes. Also see the :manpage:`sync_file_range(2)` man page. This option is | |
1436 | Linux specific. | |
1437 | ||
1438 | .. option:: overwrite=bool | |
1439 | ||
1440 | If true, writes to a file will always overwrite existing data. If the file | |
1441 | doesn't already exist, it will be created before the write phase begins. If | |
1442 | the file exists and is large enough for the specified write phase, nothing | |
a47b697c | 1443 | will be done. Default: false. |
f80dba8d MT |
1444 | |
1445 | .. option:: end_fsync=bool | |
1446 | ||
a47b697c SW |
1447 | If true, :manpage:`fsync(2)` file contents when a write stage has completed. |
1448 | Default: false. | |
f80dba8d MT |
1449 | |
1450 | .. option:: fsync_on_close=bool | |
1451 | ||
1452 | If true, fio will :manpage:`fsync(2)` a dirty file on close. This differs | |
a47b697c SW |
1453 | from :option:`end_fsync` in that it will happen on every file close, not |
1454 | just at the end of the job. Default: false. | |
f80dba8d MT |
1455 | |
1456 | .. option:: rwmixread=int | |
1457 | ||
1458 | Percentage of a mixed workload that should be reads. Default: 50. | |
1459 | ||
1460 | .. option:: rwmixwrite=int | |
1461 | ||
1462 | Percentage of a mixed workload that should be writes. If both | |
1463 | :option:`rwmixread` and :option:`rwmixwrite` is given and the values do not | |
1464 | add up to 100%, the latter of the two will be used to override the | |
1465 | first. This may interfere with a given rate setting, if fio is asked to | |
1466 | limit reads or writes to a certain rate. If that is the case, then the | |
1467 | distribution may be skewed. Default: 50. | |
1468 | ||
a87c90fd | 1469 | .. option:: random_distribution=str:float[:float][,str:float][,str:float] |
f80dba8d MT |
1470 | |
1471 | By default, fio will use a completely uniform random distribution when asked | |
1472 | to perform random I/O. Sometimes it is useful to skew the distribution in | |
1473 | specific ways, ensuring that some parts of the data is more hot than others. | |
1474 | fio includes the following distribution models: | |
1475 | ||
1476 | **random** | |
1477 | Uniform random distribution | |
1478 | ||
1479 | **zipf** | |
1480 | Zipf distribution | |
1481 | ||
1482 | **pareto** | |
1483 | Pareto distribution | |
1484 | ||
b2f4b559 | 1485 | **normal** |
c60ebc45 | 1486 | Normal (Gaussian) distribution |
f80dba8d MT |
1487 | |
1488 | **zoned** | |
1489 | Zoned random distribution | |
1490 | ||
59466396 JA |
1491 | **zoned_abs** |
1492 | Zone absolute random distribution | |
1493 | ||
f80dba8d | 1494 | When using a **zipf** or **pareto** distribution, an input value is also |
f50fbdda | 1495 | needed to define the access pattern. For **zipf**, this is the `Zipf |
c60ebc45 | 1496 | theta`. For **pareto**, it's the `Pareto power`. Fio includes a test |
f50fbdda | 1497 | program, :command:`fio-genzipf`, that can be used visualize what the given input |
f80dba8d MT |
1498 | values will yield in terms of hit rates. If you wanted to use **zipf** with |
1499 | a `theta` of 1.2, you would use ``random_distribution=zipf:1.2`` as the | |
1500 | option. If a non-uniform model is used, fio will disable use of the random | |
b2f4b559 SW |
1501 | map. For the **normal** distribution, a normal (Gaussian) deviation is |
1502 | supplied as a value between 0 and 100. | |
f80dba8d | 1503 | |
a87c90fd | 1504 | The second, optional float is allowed for **pareto**, **zipf** and **normal** distributions. |
12efafa3 | 1505 | It allows one to set base of distribution in non-default place, giving more control |
a87c90fd AK |
1506 | over most probable outcome. This value is in range [0-1] which maps linearly to |
1507 | range of possible random values. | |
1508 | Defaults are: random for **pareto** and **zipf**, and 0.5 for **normal**. | |
1509 | If you wanted to use **zipf** with a `theta` of 1.2 centered on 1/4 of allowed value range, | |
fc002f14 | 1510 | you would use ``random_distribution=zipf:1.2:0.25``. |
a87c90fd | 1511 | |
f80dba8d MT |
1512 | For a **zoned** distribution, fio supports specifying percentages of I/O |
1513 | access that should fall within what range of the file or device. For | |
1514 | example, given a criteria of: | |
1515 | ||
f50fbdda TK |
1516 | * 60% of accesses should be to the first 10% |
1517 | * 30% of accesses should be to the next 20% | |
1518 | * 8% of accesses should be to the next 30% | |
1519 | * 2% of accesses should be to the next 40% | |
f80dba8d MT |
1520 | |
1521 | we can define that through zoning of the random accesses. For the above | |
1522 | example, the user would do:: | |
1523 | ||
1524 | random_distribution=zoned:60/10:30/20:8/30:2/40 | |
1525 | ||
59466396 JA |
1526 | A **zoned_abs** distribution works exactly like the **zoned**, except |
1527 | that it takes absolute sizes. For example, let's say you wanted to | |
1528 | define access according to the following criteria: | |
1529 | ||
1530 | * 60% of accesses should be to the first 20G | |
1531 | * 30% of accesses should be to the next 100G | |
1532 | * 10% of accesses should be to the next 500G | |
1533 | ||
1534 | we can define an absolute zoning distribution with: | |
1535 | ||
1536 | random_distribution=zoned_abs=60/20G:30/100G:10/500g | |
1537 | ||
6a16ece8 JA |
1538 | For both **zoned** and **zoned_abs**, fio supports defining up to |
1539 | 256 separate zones. | |
1540 | ||
59466396 JA |
1541 | Similarly to how :option:`bssplit` works for setting ranges and |
1542 | percentages of block sizes. Like :option:`bssplit`, it's possible to | |
1543 | specify separate zones for reads, writes, and trims. If just one set | |
1544 | is given, it'll apply to all of them. This goes for both **zoned** | |
1545 | **zoned_abs** distributions. | |
f80dba8d MT |
1546 | |
1547 | .. option:: percentage_random=int[,int][,int] | |
1548 | ||
1549 | For a random workload, set how big a percentage should be random. This | |
1550 | defaults to 100%, in which case the workload is fully random. It can be set | |
1551 | from anywhere from 0 to 100. Setting it to 0 would make the workload fully | |
1552 | sequential. Any setting in between will result in a random mix of sequential | |
1553 | and random I/O, at the given percentages. Comma-separated values may be | |
1554 | specified for reads, writes, and trims as described in :option:`blocksize`. | |
1555 | ||
1556 | .. option:: norandommap | |
1557 | ||
1558 | Normally fio will cover every block of the file when doing random I/O. If | |
1559 | this option is given, fio will just get a new random offset without looking | |
1560 | at past I/O history. This means that some blocks may not be read or written, | |
1561 | and that some blocks may be read/written more than once. If this option is | |
1562 | used with :option:`verify` and multiple blocksizes (via :option:`bsrange`), | |
1563 | only intact blocks are verified, i.e., partially-overwritten blocks are | |
47e6a6e5 SW |
1564 | ignored. With an async I/O engine and an I/O depth > 1, it is possible for |
1565 | the same block to be overwritten, which can cause verification errors. Either | |
1566 | do not use norandommap in this case, or also use the lfsr random generator. | |
f80dba8d MT |
1567 | |
1568 | .. option:: softrandommap=bool | |
1569 | ||
1570 | See :option:`norandommap`. If fio runs with the random block map enabled and | |
1571 | it fails to allocate the map, if this option is set it will continue without | |
1572 | a random block map. As coverage will not be as complete as with random maps, | |
1573 | this option is disabled by default. | |
1574 | ||
1575 | .. option:: random_generator=str | |
1576 | ||
f50fbdda | 1577 | Fio supports the following engines for generating I/O offsets for random I/O: |
f80dba8d MT |
1578 | |
1579 | **tausworthe** | |
f50fbdda | 1580 | Strong 2^88 cycle random number generator. |
f80dba8d | 1581 | **lfsr** |
f50fbdda | 1582 | Linear feedback shift register generator. |
f80dba8d | 1583 | **tausworthe64** |
f50fbdda | 1584 | Strong 64-bit 2^258 cycle random number generator. |
f80dba8d MT |
1585 | |
1586 | **tausworthe** is a strong random number generator, but it requires tracking | |
1587 | on the side if we want to ensure that blocks are only read or written | |
f50fbdda | 1588 | once. **lfsr** guarantees that we never generate the same offset twice, and |
f80dba8d | 1589 | it's also less computationally expensive. It's not a true random generator, |
f50fbdda | 1590 | however, though for I/O purposes it's typically good enough. **lfsr** only |
f80dba8d MT |
1591 | works with single block sizes, not with workloads that use multiple block |
1592 | sizes. If used with such a workload, fio may read or write some blocks | |
1593 | multiple times. The default value is **tausworthe**, unless the required | |
1594 | space exceeds 2^32 blocks. If it does, then **tausworthe64** is | |
1595 | selected automatically. | |
1596 | ||
1597 | ||
1598 | Block size | |
1599 | ~~~~~~~~~~ | |
1600 | ||
1601 | .. option:: blocksize=int[,int][,int], bs=int[,int][,int] | |
1602 | ||
1603 | The block size in bytes used for I/O units. Default: 4096. A single value | |
1604 | applies to reads, writes, and trims. Comma-separated values may be | |
1605 | specified for reads, writes, and trims. A value not terminated in a comma | |
1606 | applies to subsequent types. | |
1607 | ||
1608 | Examples: | |
1609 | ||
1610 | **bs=256k** | |
1611 | means 256k for reads, writes and trims. | |
1612 | ||
1613 | **bs=8k,32k** | |
1614 | means 8k for reads, 32k for writes and trims. | |
1615 | ||
1616 | **bs=8k,32k,** | |
1617 | means 8k for reads, 32k for writes, and default for trims. | |
1618 | ||
1619 | **bs=,8k** | |
1620 | means default for reads, 8k for writes and trims. | |
1621 | ||
1622 | **bs=,8k,** | |
b443ae44 | 1623 | means default for reads, 8k for writes, and default for trims. |
f80dba8d MT |
1624 | |
1625 | .. option:: blocksize_range=irange[,irange][,irange], bsrange=irange[,irange][,irange] | |
1626 | ||
1627 | A range of block sizes in bytes for I/O units. The issued I/O unit will | |
1628 | always be a multiple of the minimum size, unless | |
1629 | :option:`blocksize_unaligned` is set. | |
1630 | ||
1631 | Comma-separated ranges may be specified for reads, writes, and trims as | |
1632 | described in :option:`blocksize`. | |
1633 | ||
9b699fb1 | 1634 | Example: ``bsrange=1k-4k,2k-8k`` also the ':' delimiter ``bsrange=1k:4k,2k:8k``. |
f80dba8d MT |
1635 | |
1636 | .. option:: bssplit=str[,str][,str] | |
1637 | ||
6a16ece8 JA |
1638 | Sometimes you want even finer grained control of the block sizes |
1639 | issued, not just an even split between them. This option allows you to | |
1640 | weight various block sizes, so that you are able to define a specific | |
1641 | amount of block sizes issued. The format for this option is:: | |
f80dba8d MT |
1642 | |
1643 | bssplit=blocksize/percentage:blocksize/percentage | |
1644 | ||
6a16ece8 JA |
1645 | for as many block sizes as needed. So if you want to define a workload |
1646 | that has 50% 64k blocks, 10% 4k blocks, and 40% 32k blocks, you would | |
1647 | write:: | |
f80dba8d MT |
1648 | |
1649 | bssplit=4k/10:64k/50:32k/40 | |
1650 | ||
6a16ece8 JA |
1651 | Ordering does not matter. If the percentage is left blank, fio will |
1652 | fill in the remaining values evenly. So a bssplit option like this one:: | |
f80dba8d MT |
1653 | |
1654 | bssplit=4k/50:1k/:32k/ | |
1655 | ||
6a16ece8 JA |
1656 | would have 50% 4k ios, and 25% 1k and 32k ios. The percentages always |
1657 | add up to 100, if bssplit is given a range that adds up to more, it | |
1658 | will error out. | |
f80dba8d MT |
1659 | |
1660 | Comma-separated values may be specified for reads, writes, and trims as | |
1661 | described in :option:`blocksize`. | |
1662 | ||
6a16ece8 JA |
1663 | If you want a workload that has 50% 2k reads and 50% 4k reads, while |
1664 | having 90% 4k writes and 10% 8k writes, you would specify:: | |
f80dba8d | 1665 | |
cf04b906 | 1666 | bssplit=2k/50:4k/50,4k/90:8k/10 |
f80dba8d | 1667 | |
6a16ece8 JA |
1668 | Fio supports defining up to 64 different weights for each data |
1669 | direction. | |
1670 | ||
f80dba8d MT |
1671 | .. option:: blocksize_unaligned, bs_unaligned |
1672 | ||
1673 | If set, fio will issue I/O units with any size within | |
1674 | :option:`blocksize_range`, not just multiples of the minimum size. This | |
1675 | typically won't work with direct I/O, as that normally requires sector | |
1676 | alignment. | |
1677 | ||
589e88b7 | 1678 | .. option:: bs_is_seq_rand=bool |
f80dba8d MT |
1679 | |
1680 | If this option is set, fio will use the normal read,write blocksize settings | |
1681 | as sequential,random blocksize settings instead. Any random read or write | |
1682 | will use the WRITE blocksize settings, and any sequential read or write will | |
1683 | use the READ blocksize settings. | |
1684 | ||
1685 | .. option:: blockalign=int[,int][,int], ba=int[,int][,int] | |
1686 | ||
1687 | Boundary to which fio will align random I/O units. Default: | |
1688 | :option:`blocksize`. Minimum alignment is typically 512b for using direct | |
1689 | I/O, though it usually depends on the hardware block size. This option is | |
1690 | mutually exclusive with using a random map for files, so it will turn off | |
1691 | that option. Comma-separated values may be specified for reads, writes, and | |
1692 | trims as described in :option:`blocksize`. | |
1693 | ||
1694 | ||
1695 | Buffers and memory | |
1696 | ~~~~~~~~~~~~~~~~~~ | |
1697 | ||
1698 | .. option:: zero_buffers | |
1699 | ||
1700 | Initialize buffers with all zeros. Default: fill buffers with random data. | |
1701 | ||
1702 | .. option:: refill_buffers | |
1703 | ||
1704 | If this option is given, fio will refill the I/O buffers on every | |
72592780 SW |
1705 | submit. Only makes sense if :option:`zero_buffers` isn't specified, |
1706 | naturally. Defaults to being unset i.e., the buffer is only filled at | |
1707 | init time and the data in it is reused when possible but if any of | |
1708 | :option:`verify`, :option:`buffer_compress_percentage` or | |
1709 | :option:`dedupe_percentage` are enabled then `refill_buffers` is also | |
1710 | automatically enabled. | |
f80dba8d MT |
1711 | |
1712 | .. option:: scramble_buffers=bool | |
1713 | ||
1714 | If :option:`refill_buffers` is too costly and the target is using data | |
1715 | deduplication, then setting this option will slightly modify the I/O buffer | |
1716 | contents to defeat normal de-dupe attempts. This is not enough to defeat | |
1717 | more clever block compression attempts, but it will stop naive dedupe of | |
1718 | blocks. Default: true. | |
1719 | ||
1720 | .. option:: buffer_compress_percentage=int | |
1721 | ||
72592780 SW |
1722 | If this is set, then fio will attempt to provide I/O buffer content |
1723 | (on WRITEs) that compresses to the specified level. Fio does this by | |
1724 | providing a mix of random data followed by fixed pattern data. The | |
1725 | fixed pattern is either zeros, or the pattern specified by | |
1726 | :option:`buffer_pattern`. If the `buffer_pattern` option is used, it | |
1727 | might skew the compression ratio slightly. Setting | |
1728 | `buffer_compress_percentage` to a value other than 100 will also | |
1729 | enable :option:`refill_buffers` in order to reduce the likelihood that | |
1730 | adjacent blocks are so similar that they over compress when seen | |
1731 | together. See :option:`buffer_compress_chunk` for how to set a finer or | |
1732 | coarser granularity for the random/fixed data region. Defaults to unset | |
1733 | i.e., buffer data will not adhere to any compression level. | |
f80dba8d MT |
1734 | |
1735 | .. option:: buffer_compress_chunk=int | |
1736 | ||
72592780 SW |
1737 | This setting allows fio to manage how big the random/fixed data region |
1738 | is when using :option:`buffer_compress_percentage`. When | |
1739 | `buffer_compress_chunk` is set to some non-zero value smaller than the | |
1740 | block size, fio can repeat the random/fixed region throughout the I/O | |
1741 | buffer at the specified interval (which particularly useful when | |
1742 | bigger block sizes are used for a job). When set to 0, fio will use a | |
1743 | chunk size that matches the block size resulting in a single | |
1744 | random/fixed region within the I/O buffer. Defaults to 512. When the | |
1745 | unit is omitted, the value is interpreted in bytes. | |
f80dba8d MT |
1746 | |
1747 | .. option:: buffer_pattern=str | |
1748 | ||
a1554f65 SB |
1749 | If set, fio will fill the I/O buffers with this pattern or with the contents |
1750 | of a file. If not set, the contents of I/O buffers are defined by the other | |
1751 | options related to buffer contents. The setting can be any pattern of bytes, | |
1752 | and can be prefixed with 0x for hex values. It may also be a string, where | |
1753 | the string must then be wrapped with ``""``. Or it may also be a filename, | |
1754 | where the filename must be wrapped with ``''`` in which case the file is | |
1755 | opened and read. Note that not all the file contents will be read if that | |
1756 | would cause the buffers to overflow. So, for example:: | |
1757 | ||
1758 | buffer_pattern='filename' | |
1759 | ||
1760 | or:: | |
f80dba8d MT |
1761 | |
1762 | buffer_pattern="abcd" | |
1763 | ||
1764 | or:: | |
1765 | ||
1766 | buffer_pattern=-12 | |
1767 | ||
1768 | or:: | |
1769 | ||
1770 | buffer_pattern=0xdeadface | |
1771 | ||
1772 | Also you can combine everything together in any order:: | |
1773 | ||
a1554f65 | 1774 | buffer_pattern=0xdeadface"abcd"-12'filename' |
f80dba8d MT |
1775 | |
1776 | .. option:: dedupe_percentage=int | |
1777 | ||
1778 | If set, fio will generate this percentage of identical buffers when | |
1779 | writing. These buffers will be naturally dedupable. The contents of the | |
1780 | buffers depend on what other buffer compression settings have been set. It's | |
1781 | possible to have the individual buffers either fully compressible, or not at | |
72592780 SW |
1782 | all -- this option only controls the distribution of unique buffers. Setting |
1783 | this option will also enable :option:`refill_buffers` to prevent every buffer | |
1784 | being identical. | |
f80dba8d | 1785 | |
0d71aa98 BD |
1786 | .. option:: dedupe_mode=str |
1787 | ||
1788 | If ``dedupe_percentage=<int>`` is given, then this option controls how fio | |
1789 | generates the dedupe buffers. | |
1790 | ||
1791 | **repeat** | |
1792 | Generate dedupe buffers by repeating previous writes | |
1793 | **working_set** | |
1794 | Generate dedupe buffers from working set | |
1795 | ||
1796 | ``repeat`` is the default option for fio. Dedupe buffers are generated | |
1797 | by repeating previous unique write. | |
1798 | ||
1799 | ``working_set`` is a more realistic workload. | |
1800 | With ``working_set``, ``dedupe_working_set_percentage=<int>`` should be provided. | |
1801 | Given that, fio will use the initial unique write buffers as its working set. | |
1802 | Upon deciding to dedupe, fio will randomly choose a buffer from the working set. | |
1803 | Note that by using ``working_set`` the dedupe percentage will converge | |
1804 | to the desired over time while ``repeat`` maintains the desired percentage | |
1805 | throughout the job. | |
1806 | ||
1807 | .. option:: dedupe_working_set_percentage=int | |
1808 | ||
1809 | If ``dedupe_mode=<str>`` is set to ``working_set``, then this controls | |
1810 | the percentage of size of the file or device used as the buffers | |
1811 | fio will choose to generate the dedupe buffers from | |
1812 | ||
1813 | Note that size needs to be explicitly provided and only 1 file per | |
1814 | job is supported | |
1815 | ||
c49cfc76 BD |
1816 | .. option:: dedupe_global=bool |
1817 | ||
1818 | This controls whether the deduplication buffers will be shared amongst | |
1819 | all jobs that have this option set. The buffers are spread evenly between | |
1820 | participating jobs. | |
1821 | ||
f80dba8d MT |
1822 | .. option:: invalidate=bool |
1823 | ||
730bd7d9 SW |
1824 | Invalidate the buffer/page cache parts of the files to be used prior to |
1825 | starting I/O if the platform and file type support it. Defaults to true. | |
21c1b29e TK |
1826 | This will be ignored if :option:`pre_read` is also specified for the |
1827 | same job. | |
f80dba8d | 1828 | |
eb9f8d7f AF |
1829 | .. option:: sync=str |
1830 | ||
1831 | Whether, and what type, of synchronous I/O to use for writes. The allowed | |
1832 | values are: | |
1833 | ||
1834 | **none** | |
1835 | Do not use synchronous IO, the default. | |
1836 | ||
1837 | **0** | |
1838 | Same as **none**. | |
1839 | ||
1840 | **sync** | |
1841 | Use synchronous file IO. For the majority of I/O engines, | |
1842 | this means using O_SYNC. | |
1843 | ||
1844 | **1** | |
1845 | Same as **sync**. | |
1846 | ||
1847 | **dsync** | |
1848 | Use synchronous data IO. For the majority of I/O engines, | |
1849 | this means using O_DSYNC. | |
f80dba8d | 1850 | |
f80dba8d MT |
1851 | |
1852 | .. option:: iomem=str, mem=str | |
1853 | ||
1854 | Fio can use various types of memory as the I/O unit buffer. The allowed | |
1855 | values are: | |
1856 | ||
1857 | **malloc** | |
1858 | Use memory from :manpage:`malloc(3)` as the buffers. Default memory | |
1859 | type. | |
1860 | ||
1861 | **shm** | |
1862 | Use shared memory as the buffers. Allocated through | |
1863 | :manpage:`shmget(2)`. | |
1864 | ||
1865 | **shmhuge** | |
1866 | Same as shm, but use huge pages as backing. | |
1867 | ||
1868 | **mmap** | |
22413915 | 1869 | Use :manpage:`mmap(2)` to allocate buffers. May either be anonymous memory, or can |
f80dba8d MT |
1870 | be file backed if a filename is given after the option. The format |
1871 | is `mem=mmap:/path/to/file`. | |
1872 | ||
1873 | **mmaphuge** | |
1874 | Use a memory mapped huge file as the buffer backing. Append filename | |
1875 | after mmaphuge, ala `mem=mmaphuge:/hugetlbfs/file`. | |
1876 | ||
1877 | **mmapshared** | |
1878 | Same as mmap, but use a MMAP_SHARED mapping. | |
1879 | ||
03553853 YR |
1880 | **cudamalloc** |
1881 | Use GPU memory as the buffers for GPUDirect RDMA benchmark. | |
f50fbdda | 1882 | The :option:`ioengine` must be `rdma`. |
03553853 | 1883 | |
f80dba8d MT |
1884 | The area allocated is a function of the maximum allowed bs size for the job, |
1885 | multiplied by the I/O depth given. Note that for **shmhuge** and | |
1886 | **mmaphuge** to work, the system must have free huge pages allocated. This | |
1887 | can normally be checked and set by reading/writing | |
1888 | :file:`/proc/sys/vm/nr_hugepages` on a Linux system. Fio assumes a huge page | |
cb8dcafa VF |
1889 | is 2 or 4MiB in size depending on the platform. So to calculate the |
1890 | number of huge pages you need for a given job file, add up the I/O | |
1891 | depth of all jobs (normally one unless :option:`iodepth` is used) and | |
1892 | multiply by the maximum bs set. Then divide that number by the huge | |
1893 | page size. You can see the size of the huge pages in | |
1894 | :file:`/proc/meminfo`. If no huge pages are allocated by having a | |
1895 | non-zero number in `nr_hugepages`, using **mmaphuge** or **shmhuge** | |
1896 | will fail. Also see :option:`hugepage-size`. | |
f80dba8d MT |
1897 | |
1898 | **mmaphuge** also needs to have hugetlbfs mounted and the file location | |
1899 | should point there. So if it's mounted in :file:`/huge`, you would use | |
1900 | `mem=mmaphuge:/huge/somefile`. | |
1901 | ||
f50fbdda | 1902 | .. option:: iomem_align=int, mem_align=int |
f80dba8d MT |
1903 | |
1904 | This indicates the memory alignment of the I/O memory buffers. Note that | |
1905 | the given alignment is applied to the first I/O unit buffer, if using | |
1906 | :option:`iodepth` the alignment of the following buffers are given by the | |
1907 | :option:`bs` used. In other words, if using a :option:`bs` that is a | |
1908 | multiple of the page sized in the system, all buffers will be aligned to | |
1909 | this value. If using a :option:`bs` that is not page aligned, the alignment | |
1910 | of subsequent I/O memory buffers is the sum of the :option:`iomem_align` and | |
1911 | :option:`bs` used. | |
1912 | ||
1913 | .. option:: hugepage-size=int | |
1914 | ||
cb8dcafa VF |
1915 | Defines the size of a huge page. Must at least be equal to the system |
1916 | setting, see :file:`/proc/meminfo` and | |
1917 | :file:`/sys/kernel/mm/hugepages/`. Defaults to 2 or 4MiB depending on | |
1918 | the platform. Should probably always be a multiple of megabytes, so | |
1919 | using ``hugepage-size=Xm`` is the preferred way to set this to avoid | |
1920 | setting a non-pow-2 bad value. | |
f80dba8d MT |
1921 | |
1922 | .. option:: lockmem=int | |
1923 | ||
1924 | Pin the specified amount of memory with :manpage:`mlock(2)`. Can be used to | |
1925 | simulate a smaller amount of memory. The amount specified is per worker. | |
1926 | ||
1927 | ||
1928 | I/O size | |
1929 | ~~~~~~~~ | |
1930 | ||
1931 | .. option:: size=int | |
1932 | ||
79591fa9 | 1933 | The total size of file I/O for each thread of this job. Fio will run until |
942d66c8 AK |
1934 | this many bytes has been transferred, unless runtime is altered by other means |
1935 | such as (1) :option:`runtime`, (2) :option:`io_size` (3) :option:`number_ios`, | |
1936 | (4) gaps/holes while doing I/O's such as ``rw=read:16K``, or (5) sequential | |
1937 | I/O reaching end of the file which is possible when :option:`percentage_random` | |
1938 | is less than 100. | |
79591fa9 TK |
1939 | Fio will divide this size between the available files determined by options |
1940 | such as :option:`nrfiles`, :option:`filename`, unless :option:`filesize` is | |
1941 | specified by the job. If the result of division happens to be 0, the size is | |
c4aa2d08 | 1942 | set to the physical size of the given files or devices if they exist. |
79591fa9 | 1943 | If this option is not specified, fio will use the full size of the given |
f80dba8d MT |
1944 | files or devices. If the files do not exist, size must be given. It is also |
1945 | possible to give size as a percentage between 1 and 100. If ``size=20%`` is | |
adcc0730 | 1946 | given, fio will use 20% of the full size of the given files or devices. |
193aaf6a | 1947 | In ZBD mode, value can also be set as number of zones using 'z'. |
9d25d068 SW |
1948 | Can be combined with :option:`offset` to constrain the start and end range |
1949 | that I/O will be done within. | |
f80dba8d MT |
1950 | |
1951 | .. option:: io_size=int, io_limit=int | |
1952 | ||
1953 | Normally fio operates within the region set by :option:`size`, which means | |
1954 | that the :option:`size` option sets both the region and size of I/O to be | |
1955 | performed. Sometimes that is not what you want. With this option, it is | |
1956 | possible to define just the amount of I/O that fio should do. For instance, | |
1957 | if :option:`size` is set to 20GiB and :option:`io_size` is set to 5GiB, fio | |
1958 | will perform I/O within the first 20GiB but exit when 5GiB have been | |
1959 | done. The opposite is also possible -- if :option:`size` is set to 20GiB, | |
1960 | and :option:`io_size` is set to 40GiB, then fio will do 40GiB of I/O within | |
1961 | the 0..20GiB region. | |
1962 | ||
7fdd97ca | 1963 | .. option:: filesize=irange(int) |
f80dba8d | 1964 | |
2a929257 NR |
1965 | Individual file sizes. May be a range, in which case fio will select sizes for |
1966 | files at random within the given range. If not given, each created file is the | |
1967 | same size. This option overrides :option:`size` in terms of file size, i.e. if | |
1968 | :option:`filesize` is specified then :option:`size` becomes merely the default | |
1969 | for :option:`io_size` and has no effect at all if :option:`io_size` is set | |
1970 | explicitly. | |
f80dba8d MT |
1971 | |
1972 | .. option:: file_append=bool | |
1973 | ||
1974 | Perform I/O after the end of the file. Normally fio will operate within the | |
1975 | size of a file. If this option is set, then fio will append to the file | |
1976 | instead. This has identical behavior to setting :option:`offset` to the size | |
1977 | of a file. This option is ignored on non-regular files. | |
1978 | ||
1979 | .. option:: fill_device=bool, fill_fs=bool | |
1980 | ||
1981 | Sets size to something really large and waits for ENOSPC (no space left on | |
418f5399 MB |
1982 | device) or EDQUOT (disk quota exceeded) |
1983 | as the terminating condition. Only makes sense with sequential | |
f80dba8d MT |
1984 | write. For a read workload, the mount point will be filled first then I/O |
1985 | started on the result. This option doesn't make sense if operating on a raw | |
1986 | device node, since the size of that is already known by the file system. | |
1987 | Additionally, writing beyond end-of-device will not return ENOSPC there. | |
1988 | ||
1989 | ||
1990 | I/O engine | |
1991 | ~~~~~~~~~~ | |
1992 | ||
1993 | .. option:: ioengine=str | |
1994 | ||
2cd920c1 | 1995 | fio supports 2 kinds of performance measurement: I/O and file/directory operation. |
1996 | ||
1997 | I/O engines define how the job issues I/O to the file. The following types are defined: | |
f80dba8d MT |
1998 | |
1999 | **sync** | |
2000 | Basic :manpage:`read(2)` or :manpage:`write(2)` | |
2001 | I/O. :manpage:`lseek(2)` is used to position the I/O location. | |
54227e6b | 2002 | See :option:`fsync` and :option:`fdatasync` for syncing write I/Os. |
f80dba8d MT |
2003 | |
2004 | **psync** | |
2005 | Basic :manpage:`pread(2)` or :manpage:`pwrite(2)` I/O. Default on | |
2006 | all supported operating systems except for Windows. | |
2007 | ||
2008 | **vsync** | |
2009 | Basic :manpage:`readv(2)` or :manpage:`writev(2)` I/O. Will emulate | |
c60ebc45 | 2010 | queuing by coalescing adjacent I/Os into a single submission. |
f80dba8d MT |
2011 | |
2012 | **pvsync** | |
2013 | Basic :manpage:`preadv(2)` or :manpage:`pwritev(2)` I/O. | |
2014 | ||
2015 | **pvsync2** | |
2016 | Basic :manpage:`preadv2(2)` or :manpage:`pwritev2(2)` I/O. | |
2017 | ||
029b42ac JA |
2018 | **io_uring** |
2019 | Fast Linux native asynchronous I/O. Supports async IO | |
2020 | for both direct and buffered IO. | |
2021 | This engine defines engine specific options. | |
2022 | ||
3716f9f1 AK |
2023 | **io_uring_cmd** |
2024 | Fast Linux native asynchronous I/O for pass through commands. | |
2025 | This engine defines engine specific options. | |
2026 | ||
f80dba8d MT |
2027 | **libaio** |
2028 | Linux native asynchronous I/O. Note that Linux may only support | |
22413915 | 2029 | queued behavior with non-buffered I/O (set ``direct=1`` or |
f80dba8d MT |
2030 | ``buffered=0``). |
2031 | This engine defines engine specific options. | |
2032 | ||
2033 | **posixaio** | |
2034 | POSIX asynchronous I/O using :manpage:`aio_read(3)` and | |
2035 | :manpage:`aio_write(3)`. | |
2036 | ||
2037 | **solarisaio** | |
2038 | Solaris native asynchronous I/O. | |
2039 | ||
2040 | **windowsaio** | |
2041 | Windows native asynchronous I/O. Default on Windows. | |
2042 | ||
2043 | **mmap** | |
2044 | File is memory mapped with :manpage:`mmap(2)` and data copied | |
2045 | to/from using :manpage:`memcpy(3)`. | |
2046 | ||
2047 | **splice** | |
2048 | :manpage:`splice(2)` is used to transfer the data and | |
2049 | :manpage:`vmsplice(2)` to transfer data from user space to the | |
2050 | kernel. | |
2051 | ||
2052 | **sg** | |
2053 | SCSI generic sg v3 I/O. May either be synchronous using the SG_IO | |
2054 | ioctl, or if the target is an sg character device we use | |
2055 | :manpage:`read(2)` and :manpage:`write(2)` for asynchronous | |
f50fbdda | 2056 | I/O. Requires :option:`filename` option to specify either block or |
3740cfc8 | 2057 | character devices. This engine supports trim operations. |
52b81b7c | 2058 | The sg engine includes engine specific options. |
f80dba8d | 2059 | |
2455851d SK |
2060 | **libzbc** |
2061 | Read, write, trim and ZBC/ZAC operations to a zoned | |
2062 | block device using libzbc library. The target can be | |
2063 | either an SG character device or a block device file. | |
2064 | ||
f80dba8d MT |
2065 | **null** |
2066 | Doesn't transfer any data, just pretends to. This is mainly used to | |
2067 | exercise fio itself and for debugging/testing purposes. | |
2068 | ||
2069 | **net** | |
2070 | Transfer over the network to given ``host:port``. Depending on the | |
2071 | :option:`protocol` used, the :option:`hostname`, :option:`port`, | |
2072 | :option:`listen` and :option:`filename` options are used to specify | |
2073 | what sort of connection to make, while the :option:`protocol` option | |
2074 | determines which protocol will be used. This engine defines engine | |
2075 | specific options. | |
2076 | ||
2077 | **netsplice** | |
2078 | Like **net**, but uses :manpage:`splice(2)` and | |
2079 | :manpage:`vmsplice(2)` to map data and send/receive. | |
2080 | This engine defines engine specific options. | |
2081 | ||
2082 | **cpuio** | |
2083 | Doesn't transfer any data, but burns CPU cycles according to the | |
9de473a8 EV |
2084 | :option:`cpuload`, :option:`cpuchunks` and :option:`cpumode` options. |
2085 | Setting :option:`cpuload`\=85 will cause that job to do nothing but burn 85% | |
71aa48eb | 2086 | of the CPU. In case of SMP machines, use :option:`numjobs`\=<nr_of_cpu> |
f50fbdda | 2087 | to get desired CPU usage, as the cpuload only loads a |
f80dba8d MT |
2088 | single CPU at the desired rate. A job never finishes unless there is |
2089 | at least one non-cpuio job. | |
9de473a8 EV |
2090 | Setting :option:`cpumode`\=qsort replace the default noop instructions loop |
2091 | by a qsort algorithm to consume more energy. | |
f80dba8d | 2092 | |
f80dba8d MT |
2093 | **rdma** |
2094 | The RDMA I/O engine supports both RDMA memory semantics | |
2095 | (RDMA_WRITE/RDMA_READ) and channel semantics (Send/Recv) for the | |
609ac152 SB |
2096 | InfiniBand, RoCE and iWARP protocols. This engine defines engine |
2097 | specific options. | |
f80dba8d MT |
2098 | |
2099 | **falloc** | |
2100 | I/O engine that does regular fallocate to simulate data transfer as | |
2101 | fio ioengine. | |
2102 | ||
2103 | DDIR_READ | |
2104 | does fallocate(,mode = FALLOC_FL_KEEP_SIZE,). | |
2105 | ||
2106 | DDIR_WRITE | |
2107 | does fallocate(,mode = 0). | |
2108 | ||
2109 | DDIR_TRIM | |
2110 | does fallocate(,mode = FALLOC_FL_KEEP_SIZE|FALLOC_FL_PUNCH_HOLE). | |
2111 | ||
761cd093 SW |
2112 | **ftruncate** |
2113 | I/O engine that sends :manpage:`ftruncate(2)` operations in response | |
2114 | to write (DDIR_WRITE) events. Each ftruncate issued sets the file's | |
f50fbdda | 2115 | size to the current block offset. :option:`blocksize` is ignored. |
761cd093 | 2116 | |
f80dba8d MT |
2117 | **e4defrag** |
2118 | I/O engine that does regular EXT4_IOC_MOVE_EXT ioctls to simulate | |
2119 | defragment activity in request to DDIR_WRITE event. | |
2120 | ||
f3f96717 IF |
2121 | **rados** |
2122 | I/O engine supporting direct access to Ceph Reliable Autonomic | |
2123 | Distributed Object Store (RADOS) via librados. This ioengine | |
2124 | defines engine specific options. | |
2125 | ||
f80dba8d MT |
2126 | **rbd** |
2127 | I/O engine supporting direct access to Ceph Rados Block Devices | |
2128 | (RBD) via librbd without the need to use the kernel rbd driver. This | |
2129 | ioengine defines engine specific options. | |
2130 | ||
c2f6a13d LMB |
2131 | **http** |
2132 | I/O engine supporting GET/PUT requests over HTTP(S) with libcurl to | |
2133 | a WebDAV or S3 endpoint. This ioengine defines engine specific options. | |
2134 | ||
2135 | This engine only supports direct IO of iodepth=1; you need to scale this | |
2136 | via numjobs. blocksize defines the size of the objects to be created. | |
2137 | ||
2138 | TRIM is translated to object deletion. | |
2139 | ||
f80dba8d | 2140 | **gfapi** |
ac8ca2af SW |
2141 | Using GlusterFS libgfapi sync interface to direct access to |
2142 | GlusterFS volumes without having to go through FUSE. This ioengine | |
f80dba8d MT |
2143 | defines engine specific options. |
2144 | ||
2145 | **gfapi_async** | |
ac8ca2af SW |
2146 | Using GlusterFS libgfapi async interface to direct access to |
2147 | GlusterFS volumes without having to go through FUSE. This ioengine | |
f80dba8d MT |
2148 | defines engine specific options. |
2149 | ||
2150 | **libhdfs** | |
f50fbdda | 2151 | Read and write through Hadoop (HDFS). The :option:`filename` option |
f80dba8d MT |
2152 | is used to specify host,port of the hdfs name-node to connect. This |
2153 | engine interprets offsets a little differently. In HDFS, files once | |
e25c0c91 SW |
2154 | created cannot be modified so random writes are not possible. To |
2155 | imitate this the libhdfs engine expects a bunch of small files to be | |
2156 | created over HDFS and will randomly pick a file from them | |
2157 | based on the offset generated by fio backend (see the example | |
f80dba8d | 2158 | job file to create such files, use ``rw=write`` option). Please |
e25c0c91 SW |
2159 | note, it may be necessary to set environment variables to work |
2160 | with HDFS/libhdfs properly. Each job uses its own connection to | |
f80dba8d MT |
2161 | HDFS. |
2162 | ||
2163 | **mtd** | |
2164 | Read, write and erase an MTD character device (e.g., | |
2165 | :file:`/dev/mtd0`). Discards are treated as erases. Depending on the | |
2166 | underlying device type, the I/O may have to go in a certain pattern, | |
2167 | e.g., on NAND, writing sequentially to erase blocks and discarding | |
c298ee71 | 2168 | before overwriting. The `trimwrite` mode works well for this |
f80dba8d MT |
2169 | constraint. |
2170 | ||
f80dba8d MT |
2171 | **dev-dax** |
2172 | Read and write using device DAX to a persistent memory device (e.g., | |
363a5f65 | 2173 | /dev/dax0.0) through the PMDK libpmem library. |
f80dba8d MT |
2174 | |
2175 | **external** | |
2176 | Prefix to specify loading an external I/O engine object file. Append | |
c60ebc45 | 2177 | the engine filename, e.g. ``ioengine=external:/tmp/foo.o`` to load |
d243fd6d TK |
2178 | ioengine :file:`foo.o` in :file:`/tmp`. The path can be either |
2179 | absolute or relative. See :file:`engines/skeleton_external.c` for | |
2180 | details of writing an external I/O engine. | |
f80dba8d | 2181 | |
ae0db592 TI |
2182 | **libpmem** |
2183 | Read and write using mmap I/O to a file on a filesystem | |
363a5f65 | 2184 | mounted with DAX on a persistent memory device through the PMDK |
ae0db592 TI |
2185 | libpmem library. |
2186 | ||
a40e7a59 GB |
2187 | **ime_psync** |
2188 | Synchronous read and write using DDN's Infinite Memory Engine (IME). | |
2189 | This engine is very basic and issues calls to IME whenever an IO is | |
2190 | queued. | |
2191 | ||
2192 | **ime_psyncv** | |
2193 | Synchronous read and write using DDN's Infinite Memory Engine (IME). | |
2194 | This engine uses iovecs and will try to stack as much IOs as possible | |
2195 | (if the IOs are "contiguous" and the IO depth is not exceeded) | |
2196 | before issuing a call to IME. | |
2197 | ||
2198 | **ime_aio** | |
2199 | Asynchronous read and write using DDN's Infinite Memory Engine (IME). | |
2200 | This engine will try to stack as much IOs as possible by creating | |
2201 | requests for IME. FIO will then decide when to commit these requests. | |
68522f38 | 2202 | |
247ef2aa KZ |
2203 | **libiscsi** |
2204 | Read and write iscsi lun with libiscsi. | |
68522f38 | 2205 | |
d643a1e2 | 2206 | **nbd** |
f2d6de5d | 2207 | Read and write a Network Block Device (NBD). |
a40e7a59 | 2208 | |
10756b2c BS |
2209 | **libcufile** |
2210 | I/O engine supporting libcufile synchronous access to nvidia-fs and a | |
2211 | GPUDirect Storage-supported filesystem. This engine performs | |
2212 | I/O without transferring buffers between user-space and the kernel, | |
2213 | unless :option:`verify` is set or :option:`cuda_io` is `posix`. | |
2214 | :option:`iomem` must not be `cudamalloc`. This ioengine defines | |
2215 | engine specific options. | |
68522f38 | 2216 | |
c363fdd7 JL |
2217 | **dfs** |
2218 | I/O engine supporting asynchronous read and write operations to the | |
2219 | DAOS File System (DFS) via libdfs. | |
10756b2c | 2220 | |
9326926b TG |
2221 | **nfs** |
2222 | I/O engine supporting asynchronous read and write operations to | |
2223 | NFS filesystems from userspace via libnfs. This is useful for | |
2224 | achieving higher concurrency and thus throughput than is possible | |
2225 | via kernel NFS. | |
2226 | ||
b50590bc EV |
2227 | **exec** |
2228 | Execute 3rd party tools. Could be used to perform monitoring during jobs runtime. | |
2229 | ||
454154e6 AK |
2230 | **xnvme** |
2231 | I/O engine using the xNVMe C API, for NVMe devices. The xnvme engine provides | |
2232 | flexibility to access GNU/Linux Kernel NVMe driver via libaio, IOCTLs, io_uring, | |
2233 | the SPDK NVMe driver, or your own custom NVMe driver. The xnvme engine includes | |
2234 | engine specific options. (See https://xnvme.io). | |
2235 | ||
a601337a AF |
2236 | **libblkio** |
2237 | Use the libblkio library | |
2238 | (https://gitlab.com/libblkio/libblkio). The specific | |
2239 | *driver* to use must be set using | |
ef9b6f2f AF |
2240 | :option:`libblkio_driver`. If |
2241 | :option:`mem`/:option:`iomem` is not specified, memory | |
2242 | allocation is delegated to libblkio (and so is | |
3afc2d8a AF |
2243 | guaranteed to work with the selected *driver*). One |
2244 | libblkio instance is used per process, so all jobs | |
2245 | setting option :option:`thread` will share a single | |
2246 | instance (with one queue per thread) and must specify | |
2247 | compatible options. Note that some drivers don't allow | |
2248 | several instances to access the same device or file | |
2249 | simultaneously, but allow it for threads. | |
a601337a | 2250 | |
2cd920c1 | 2251 | File/directory operation engines define how the job operates file or directory. The |
2252 | following types are defined: | |
2253 | ||
2254 | **filecreate** | |
2255 | Simply create the files and do no I/O to them. You still need to | |
2256 | set `filesize` so that all the accounting still occurs, but no | |
2257 | actual I/O will be done other than creating the file. | |
2258 | Example job file: filecreate-ioengine.fio. | |
2259 | ||
2260 | **filestat** | |
2261 | Simply do stat() and do no I/O to the file. You need to set 'filesize' | |
2262 | and 'nrfiles', so that files will be created. | |
2263 | This engine is to measure file lookup and meta data access. | |
2264 | Example job file: filestat-ioengine.fio. | |
2265 | ||
2266 | **filedelete** | |
2267 | Simply delete the files by unlink() and do no I/O to them. You need to set 'filesize' | |
2268 | and 'nrfiles', so that the files will be created. | |
2269 | This engine is to measure file delete. | |
2270 | Example job file: filedelete-ioengine.fio. | |
2271 | ||
2272 | **dircreate** | |
2273 | Simply create the directories and do no I/O to them. You still need to | |
2274 | set `filesize` so that all the accounting still occurs, but no | |
2275 | actual I/O will be done other than creating the directories. | |
2276 | Example job file: dircreate-ioengine.fio. | |
2277 | ||
2278 | **dirstat** | |
2279 | Simply do stat() and do no I/O to the directories. You need to set 'filesize' | |
2280 | and 'nrfiles', so that directories will be created. | |
2281 | This engine is to measure directory lookup and meta data access. | |
2282 | Example job file: dirstat-ioengine.fio. | |
2283 | ||
2284 | **dirdelete** | |
2285 | Simply delete the directories by rmdir() and do no I/O to them. You need to set 'filesize' | |
2286 | and 'nrfiles', so that the directories will be created. | |
2287 | This engine is to measure directory delete. | |
2288 | Example job file: dirdelete-ioengine.fio. | |
2289 | ||
2290 | For file and directory operation engines, there is no I/O throughput, then the | |
2291 | statistics data in report have different meanings. The meaningful output indexes are: 'iops' and 'clat'. | |
2292 | 'bw' is meaningless. Refer to section: "Interpreting the output" for more details. | |
2293 | ||
2294 | ||
f80dba8d MT |
2295 | I/O engine specific parameters |
2296 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
2297 | ||
2298 | In addition, there are some parameters which are only valid when a specific | |
f50fbdda TK |
2299 | :option:`ioengine` is in use. These are used identically to normal parameters, |
2300 | with the caveat that when used on the command line, they must come after the | |
f80dba8d MT |
2301 | :option:`ioengine` that defines them is selected. |
2302 | ||
e9f6567a DLM |
2303 | .. option:: cmdprio_percentage=int[,int] : [io_uring] [libaio] |
2304 | ||
2305 | Set the percentage of I/O that will be issued with the highest priority. | |
2306 | Default: 0. A single value applies to reads and writes. Comma-separated | |
acf2e2d9 | 2307 | values may be specified for reads and writes. For this option to be |
68522f38 VF |
2308 | effective, NCQ priority must be supported and enabled, and the :option:`direct` |
2309 | option must be set. fio must also be run as the root user. Unlike | |
bebf1407 NC |
2310 | slat/clat/lat stats, which can be tracked and reported independently, per |
2311 | priority stats only track and report a single type of latency. By default, | |
2312 | completion latency (clat) will be reported, if :option:`lat_percentiles` is | |
2313 | set, total latency (lat) will be reported. | |
029b42ac | 2314 | |
12f9d54a DLM |
2315 | .. option:: cmdprio_class=int[,int] : [io_uring] [libaio] |
2316 | ||
2317 | Set the I/O priority class to use for I/Os that must be issued with | |
a48f0cc7 DLM |
2318 | a priority when :option:`cmdprio_percentage` or |
2319 | :option:`cmdprio_bssplit` is set. If not specified when | |
2320 | :option:`cmdprio_percentage` or :option:`cmdprio_bssplit` is set, | |
2321 | this defaults to the highest priority class. A single value applies | |
2322 | to reads and writes. Comma-separated values may be specified for | |
2323 | reads and writes. See :manpage:`ionice(1)`. See also the | |
2324 | :option:`prioclass` option. | |
12f9d54a | 2325 | |
79012fec DLM |
2326 | .. option:: cmdprio_hint=int[,int] : [io_uring] [libaio] |
2327 | ||
2328 | Set the I/O priority hint to use for I/Os that must be issued with | |
2329 | a priority when :option:`cmdprio_percentage` or | |
2330 | :option:`cmdprio_bssplit` is set. If not specified when | |
2331 | :option:`cmdprio_percentage` or :option:`cmdprio_bssplit` is set, | |
2332 | this defaults to 0 (no hint). A single value applies to reads and | |
2333 | writes. Comma-separated values may be specified for reads and writes. | |
2334 | See also the :option:`priohint` option. | |
2335 | ||
12f9d54a DLM |
2336 | .. option:: cmdprio=int[,int] : [io_uring] [libaio] |
2337 | ||
2338 | Set the I/O priority value to use for I/Os that must be issued with | |
a48f0cc7 DLM |
2339 | a priority when :option:`cmdprio_percentage` or |
2340 | :option:`cmdprio_bssplit` is set. If not specified when | |
2341 | :option:`cmdprio_percentage` or :option:`cmdprio_bssplit` is set, | |
2342 | this defaults to 0. | |
12f9d54a DLM |
2343 | Linux limits us to a positive value between 0 and 7, with 0 being the |
2344 | highest. A single value applies to reads and writes. Comma-separated | |
2345 | values may be specified for reads and writes. See :manpage:`ionice(1)`. | |
2346 | Refer to an appropriate manpage for other operating systems since | |
2347 | meaning of priority may differ. See also the :option:`prio` option. | |
2348 | ||
a48f0cc7 | 2349 | .. option:: cmdprio_bssplit=str[,str] : [io_uring] [libaio] |
68522f38 | 2350 | |
a48f0cc7 DLM |
2351 | To get a finer control over I/O priority, this option allows |
2352 | specifying the percentage of IOs that must have a priority set | |
2353 | depending on the block size of the IO. This option is useful only | |
2354 | when used together with the :option:`bssplit` option, that is, | |
2355 | multiple different block sizes are used for reads and writes. | |
f0547200 NC |
2356 | |
2357 | The first accepted format for this option is the same as the format of | |
2358 | the :option:`bssplit` option: | |
2359 | ||
2360 | cmdprio_bssplit=blocksize/percentage:blocksize/percentage | |
2361 | ||
79012fec DLM |
2362 | In this case, each entry will use the priority class, priority hint |
2363 | and priority level defined by the options :option:`cmdprio_class`, | |
2364 | :option:`cmdprio` and :option:`cmdprio_hint` respectively. | |
f0547200 NC |
2365 | |
2366 | The second accepted format for this option is: | |
2367 | ||
2368 | cmdprio_bssplit=blocksize/percentage/class/level:blocksize/percentage/class/level | |
2369 | ||
2370 | In this case, the priority class and priority level is defined inside | |
2371 | each entry. In comparison with the first accepted format, the second | |
2372 | accepted format does not restrict all entries to have the same priority | |
2373 | class and priority level. | |
2374 | ||
79012fec DLM |
2375 | The third accepted format for this option is: |
2376 | ||
2377 | cmdprio_bssplit=blocksize/percentage/class/level/hint:... | |
2378 | ||
85ccc10a MS |
2379 | This is an extension of the second accepted format that allows one to |
2380 | also specify a priority hint. | |
79012fec DLM |
2381 | |
2382 | For all formats, only the read and write data directions are supported, | |
f0547200 NC |
2383 | values for trim IOs are ignored. This option is mutually exclusive with |
2384 | the :option:`cmdprio_percentage` option. | |
a48f0cc7 | 2385 | |
3716f9f1 AK |
2386 | .. option:: fixedbufs : [io_uring] [io_uring_cmd] |
2387 | ||
2388 | If fio is asked to do direct IO, then Linux will map pages for each | |
2389 | IO call, and release them when IO is done. If this option is set, the | |
2390 | pages are pre-mapped before IO is started. This eliminates the need to | |
2391 | map and release for each IO. This is more efficient, and reduces the | |
2392 | IO latency as well. | |
2393 | ||
d6f936d1 | 2394 | .. option:: nonvectored=int : [io_uring] [io_uring_cmd] |
029b42ac | 2395 | |
3716f9f1 AK |
2396 | With this option, fio will use non-vectored read/write commands, where |
2397 | address must contain the address directly. Default is -1. | |
b2a432bf | 2398 | |
3716f9f1 AK |
2399 | .. option:: force_async=int : [io_uring] [io_uring_cmd] |
2400 | ||
2401 | Normal operation for io_uring is to try and issue an sqe as | |
2402 | non-blocking first, and if that fails, execute it in an async manner. | |
2403 | With this option set to N, then every N request fio will ask sqe to | |
2404 | be issued in an async manner. Default is 0. | |
2405 | ||
2406 | .. option:: registerfiles : [io_uring] [io_uring_cmd] | |
2c870598 | 2407 | |
5ffd5626 JA |
2408 | With this option, fio registers the set of files being used with the |
2409 | kernel. This avoids the overhead of managing file counts in the kernel, | |
2410 | making the submission and completion part more lightweight. Required | |
2411 | for the below :option:`sqthread_poll` option. | |
2412 | ||
3716f9f1 | 2413 | .. option:: sqthread_poll : [io_uring] [io_uring_cmd] [xnvme] |
029b42ac JA |
2414 | |
2415 | Normally fio will submit IO by issuing a system call to notify the | |
2416 | kernel of available items in the SQ ring. If this option is set, the | |
2417 | act of submitting IO will be done by a polling thread in the kernel. | |
2418 | This frees up cycles for fio, at the cost of using more CPU in the | |
72044c66 AK |
2419 | system. As submission is just the time it takes to fill in the sqe |
2420 | entries and any syscall required to wake up the idle kernel thread, | |
2421 | fio will not report submission latencies. | |
029b42ac | 2422 | |
d6f936d1 | 2423 | .. option:: sqthread_poll_cpu=int : [io_uring] [io_uring_cmd] |
029b42ac JA |
2424 | |
2425 | When :option:`sqthread_poll` is set, this option provides a way to | |
2426 | define which CPU should be used for the polling thread. | |
2427 | ||
3716f9f1 AK |
2428 | .. option:: cmd_type=str : [io_uring_cmd] |
2429 | ||
2430 | Specifies the type of uring passthrough command to be used. Supported | |
2431 | value is nvme. Default is nvme. | |
2432 | ||
8253a66b VF |
2433 | .. option:: hipri |
2434 | ||
3716f9f1 | 2435 | [io_uring] [io_uring_cmd] [xnvme] |
8253a66b VF |
2436 | |
2437 | If this option is set, fio will attempt to use polled IO completions. | |
2438 | Normal IO completions generate interrupts to signal the completion of | |
2439 | IO, polled completions do not. Hence they are require active reaping | |
2440 | by the application. The benefits are more efficient IO for high IOPS | |
2441 | scenarios, and lower latencies for low queue depth IO. | |
2442 | ||
a870d6ff AF |
2443 | [libblkio] |
2444 | ||
b158577d | 2445 | Use poll queues. This is incompatible with |
b1bd09b5 AF |
2446 | :option:`libblkio_wait_mode=eventfd <libblkio_wait_mode>` and |
2447 | :option:`libblkio_force_enable_completion_eventfd`. | |
a870d6ff | 2448 | |
8253a66b VF |
2449 | [pvsync2] |
2450 | ||
2451 | Set RWF_HIPRI on I/O, indicating to the kernel that it's of higher priority | |
2452 | than normal. | |
2453 | ||
8253a66b VF |
2454 | [sg] |
2455 | ||
2456 | If this option is set, fio will attempt to use polled IO completions. | |
2457 | This will have a similar effect as (io_uring)hipri. Only SCSI READ and | |
2458 | WRITE commands will have the SGV4_FLAG_HIPRI set (not UNMAP (trim) nor | |
2459 | VERIFY). Older versions of the Linux sg driver that do not support | |
2460 | hipri will simply ignore this flag and do normal IO. The Linux SCSI | |
2461 | Low Level Driver (LLD) that "owns" the device also needs to support | |
2462 | hipri (also known as iopoll and mq_poll). The MegaRAID driver is an | |
2463 | example of a SCSI LLD. Default: clear (0) which does normal | |
2464 | (interrupted based) IO. | |
2465 | ||
f80dba8d MT |
2466 | .. option:: userspace_reap : [libaio] |
2467 | ||
2468 | Normally, with the libaio engine in use, fio will use the | |
2469 | :manpage:`io_getevents(2)` system call to reap newly returned events. With | |
2470 | this flag turned on, the AIO ring will be read directly from user-space to | |
2471 | reap events. The reaping mode is only enabled when polling for a minimum of | |
c60ebc45 | 2472 | 0 events (e.g. when :option:`iodepth_batch_complete` `=0`). |
f80dba8d | 2473 | |
a0679ce5 SB |
2474 | .. option:: hipri_percentage : [pvsync2] |
2475 | ||
f50fbdda | 2476 | When hipri is set this determines the probability of a pvsync2 I/O being high |
a0679ce5 SB |
2477 | priority. The default is 100%. |
2478 | ||
d6f936d1 | 2479 | .. option:: nowait=bool : [pvsync2] [libaio] [io_uring] [io_uring_cmd] |
7d42e66e KK |
2480 | |
2481 | By default if a request cannot be executed immediately (e.g. resource starvation, | |
2482 | waiting on locks) it is queued and the initiating process will be blocked until | |
2483 | the required resource becomes free. | |
2484 | ||
2485 | This option sets the RWF_NOWAIT flag (supported from the 4.14 Linux kernel) and | |
2486 | the call will return instantly with EAGAIN or a partial result rather than waiting. | |
2487 | ||
2488 | It is useful to also use ignore_error=EAGAIN when using this option. | |
2489 | ||
2490 | Note: glibc 2.27, 2.28 have a bug in syscall wrappers preadv2, pwritev2. | |
2491 | They return EOPNOTSUP instead of EAGAIN. | |
2492 | ||
2493 | For cached I/O, using this option usually means a request operates only with | |
2494 | cached data. Currently the RWF_NOWAIT flag does not supported for cached write. | |
2495 | ||
2496 | For direct I/O, requests will only succeed if cache invalidation isn't required, | |
2497 | file blocks are fully allocated and the disk request could be issued immediately. | |
2498 | ||
e5f3b613 | 2499 | .. option:: fdp=bool : [io_uring_cmd] [xnvme] |
a7e8aae0 KB |
2500 | |
2501 | Enable Flexible Data Placement mode for write commands. | |
2502 | ||
e5f3b613 | 2503 | .. option:: fdp_pli_select=str : [io_uring_cmd] [xnvme] |
d3e310c5 AK |
2504 | |
2505 | Defines how fio decides which placement ID to use next. The following | |
2506 | types are defined: | |
2507 | ||
2508 | **random** | |
2509 | Choose a placement ID at random (uniform). | |
2510 | ||
2511 | **roundrobin** | |
2512 | Round robin over available placement IDs. This is the | |
2513 | default. | |
2514 | ||
2515 | The available placement ID index/indices is defined by the option | |
2516 | :option:`fdp_pli`. | |
2517 | ||
e5f3b613 | 2518 | .. option:: fdp_pli=str : [io_uring_cmd] [xnvme] |
a7e8aae0 KB |
2519 | |
2520 | Select which Placement ID Index/Indicies this job is allowed to use for | |
2521 | writes. By default, the job will cycle through all available Placement | |
2522 | IDs, so use this to isolate these identifiers to specific jobs. If you | |
2523 | want fio to use placement identifier only at indices 0, 2 and 5 specify | |
2524 | ``fdp_pli=0,2,5``. | |
2525 | ||
be5514e3 | 2526 | .. option:: md_per_io_size=int : [io_uring_cmd] [xnvme] |
2d6451c9 AK |
2527 | |
2528 | Size in bytes for separate metadata buffer per IO. Default: 0. | |
2529 | ||
90ec1ecc | 2530 | .. option:: pi_act=int : [io_uring_cmd] [xnvme] |
3ee8311a AK |
2531 | |
2532 | Action to take when nvme namespace is formatted with protection | |
2533 | information. If this is set to 1 and namespace is formatted with | |
2534 | metadata size equal to protection information size, fio won't use | |
2535 | separate metadata buffer or extended logical block. If this is set to | |
2536 | 1 and namespace is formatted with metadata size greater than protection | |
2537 | information size, fio will not generate or verify the protection | |
2538 | information portion of metadata for write or read case respectively. | |
2539 | If this is set to 0, fio generates protection information for | |
2540 | write case and verifies for read case. Default: 1. | |
2541 | ||
a4614bff AK |
2542 | For 16 bit CRC generation fio will use isa-l if available otherwise |
2543 | it will use the default slower generator. | |
2544 | (see: https://github.com/intel/isa-l) | |
2545 | ||
90ec1ecc | 2546 | .. option:: pi_chk=str[,str][,str] : [io_uring_cmd] [xnvme] |
3ee8311a AK |
2547 | |
2548 | Controls the protection information check. This can take one or more | |
2549 | of these values. Default: none. | |
2550 | ||
2551 | **GUARD** | |
2552 | Enables protection information checking of guard field. | |
2553 | **REFTAG** | |
2554 | Enables protection information checking of logical block | |
2555 | reference tag field. | |
2556 | **APPTAG** | |
2557 | Enables protection information checking of application tag field. | |
2558 | ||
90ec1ecc | 2559 | .. option:: apptag=int : [io_uring_cmd] [xnvme] |
3ee8311a AK |
2560 | |
2561 | Specifies logical block application tag value, if namespace is | |
2562 | formatted to use end to end protection information. Default: 0x1234. | |
2563 | ||
90ec1ecc | 2564 | .. option:: apptag_mask=int : [io_uring_cmd] [xnvme] |
3ee8311a AK |
2565 | |
2566 | Specifies logical block application tag mask value, if namespace is | |
2567 | formatted to use end to end protection information. Default: 0xffff. | |
2568 | ||
b3251e31 AK |
2569 | .. option:: num_range=int : [io_uring_cmd] |
2570 | ||
2571 | For trim command this will be the number of ranges to trim per I/O | |
2572 | request. The number of logical blocks per range is determined by the | |
2573 | :option:`bs` option which should be a multiple of logical block size. | |
2574 | This cannot be used with read or write. Note that setting this | |
2575 | option > 1, :option:`log_offset` will not be able to log all the | |
2576 | offsets. Default: 1. | |
2577 | ||
f80dba8d MT |
2578 | .. option:: cpuload=int : [cpuio] |
2579 | ||
da19cdb4 TK |
2580 | Attempt to use the specified percentage of CPU cycles. This is a mandatory |
2581 | option when using cpuio I/O engine. | |
f80dba8d MT |
2582 | |
2583 | .. option:: cpuchunks=int : [cpuio] | |
2584 | ||
2585 | Split the load into cycles of the given time. In microseconds. | |
2586 | ||
8a7bf04c VF |
2587 | .. option:: cpumode=str : [cpuio] |
2588 | ||
2589 | Specify how to stress the CPU. It can take these two values: | |
2590 | ||
2591 | **noop** | |
2592 | This is the default where the CPU executes noop instructions. | |
2593 | **qsort** | |
2594 | Replace the default noop instructions loop with a qsort algorithm to | |
2595 | consume more energy. | |
2596 | ||
f80dba8d MT |
2597 | .. option:: exit_on_io_done=bool : [cpuio] |
2598 | ||
2599 | Detect when I/O threads are done, then exit. | |
2600 | ||
f80dba8d MT |
2601 | .. option:: namenode=str : [libhdfs] |
2602 | ||
22413915 | 2603 | The hostname or IP address of a HDFS cluster namenode to contact. |
f80dba8d MT |
2604 | |
2605 | .. option:: port=int | |
2606 | ||
f50fbdda TK |
2607 | [libhdfs] |
2608 | ||
2609 | The listening port of the HFDS cluster namenode. | |
2610 | ||
f80dba8d MT |
2611 | [netsplice], [net] |
2612 | ||
2613 | The TCP or UDP port to bind to or connect to. If this is used with | |
2614 | :option:`numjobs` to spawn multiple instances of the same job type, then | |
2615 | this will be the starting port number since fio will use a range of | |
2616 | ports. | |
2617 | ||
e4c4625f | 2618 | [rdma], [librpma_*] |
609ac152 SB |
2619 | |
2620 | The port to use for RDMA-CM communication. This should be the same value | |
2621 | on the client and the server side. | |
2622 | ||
2623 | .. option:: hostname=str : [netsplice] [net] [rdma] | |
f80dba8d | 2624 | |
609ac152 SB |
2625 | The hostname or IP address to use for TCP, UDP or RDMA-CM based I/O. If the job |
2626 | is a TCP listener or UDP reader, the hostname is not used and must be omitted | |
f50fbdda | 2627 | unless it is a valid UDP multicast address. |
f80dba8d | 2628 | |
e4c4625f JM |
2629 | .. option:: serverip=str : [librpma_*] |
2630 | ||
2631 | The IP address to be used for RDMA-CM based I/O. | |
2632 | ||
2633 | .. option:: direct_write_to_pmem=bool : [librpma_*] | |
2634 | ||
2635 | Set to 1 only when Direct Write to PMem from the remote host is possible. | |
2636 | Otherwise, set to 0. | |
2637 | ||
6a229978 OS |
2638 | .. option:: busy_wait_polling=bool : [librpma_*_server] |
2639 | ||
2640 | Set to 0 to wait for completion instead of busy-wait polling completion. | |
2641 | Default: 1. | |
2642 | ||
f80dba8d MT |
2643 | .. option:: interface=str : [netsplice] [net] |
2644 | ||
2645 | The IP address of the network interface used to send or receive UDP | |
2646 | multicast. | |
2647 | ||
2648 | .. option:: ttl=int : [netsplice] [net] | |
2649 | ||
2650 | Time-to-live value for outgoing UDP multicast packets. Default: 1. | |
2651 | ||
2652 | .. option:: nodelay=bool : [netsplice] [net] | |
2653 | ||
2654 | Set TCP_NODELAY on TCP connections. | |
2655 | ||
f50fbdda | 2656 | .. option:: protocol=str, proto=str : [netsplice] [net] |
f80dba8d MT |
2657 | |
2658 | The network protocol to use. Accepted values are: | |
2659 | ||
2660 | **tcp** | |
2661 | Transmission control protocol. | |
2662 | **tcpv6** | |
2663 | Transmission control protocol V6. | |
2664 | **udp** | |
2665 | User datagram protocol. | |
2666 | **udpv6** | |
2667 | User datagram protocol V6. | |
2668 | **unix** | |
2669 | UNIX domain socket. | |
80cc242a MP |
2670 | **vsock** |
2671 | VSOCK protocol. | |
f80dba8d | 2672 | |
80cc242a MP |
2673 | When the protocol is TCP, UDP or VSOCK, the port must also be given, as well as the |
2674 | hostname if the job is a TCP or VSOCK listener or UDP reader. For unix sockets, the | |
f50fbdda | 2675 | normal :option:`filename` option should be used and the port is invalid. |
80cc242a | 2676 | When the protocol is VSOCK, the :option:`hostname` is the CID of the remote VM. |
f80dba8d | 2677 | |
e9184ec1 | 2678 | .. option:: listen : [netsplice] [net] |
f80dba8d MT |
2679 | |
2680 | For TCP network connections, tell fio to listen for incoming connections | |
2681 | rather than initiating an outgoing connection. The :option:`hostname` must | |
2682 | be omitted if this option is used. | |
2683 | ||
e9184ec1 | 2684 | .. option:: pingpong : [netsplice] [net] |
f80dba8d MT |
2685 | |
2686 | Normally a network writer will just continue writing data, and a network | |
2687 | reader will just consume packages. If ``pingpong=1`` is set, a writer will | |
2688 | send its normal payload to the reader, then wait for the reader to send the | |
2689 | same payload back. This allows fio to measure network latencies. The | |
2690 | submission and completion latencies then measure local time spent sending or | |
2691 | receiving, and the completion latency measures how long it took for the | |
2692 | other end to receive and send back. For UDP multicast traffic | |
2693 | ``pingpong=1`` should only be set for a single reader when multiple readers | |
2694 | are listening to the same address. | |
2695 | ||
e9184ec1 | 2696 | .. option:: window_size : [netsplice] [net] |
f80dba8d MT |
2697 | |
2698 | Set the desired socket buffer size for the connection. | |
2699 | ||
e9184ec1 | 2700 | .. option:: mss : [netsplice] [net] |
f80dba8d MT |
2701 | |
2702 | Set the TCP maximum segment size (TCP_MAXSEG). | |
2703 | ||
2704 | .. option:: donorname=str : [e4defrag] | |
2705 | ||
730bd7d9 | 2706 | File will be used as a block donor (swap extents between files). |
f80dba8d MT |
2707 | |
2708 | .. option:: inplace=int : [e4defrag] | |
2709 | ||
2710 | Configure donor file blocks allocation strategy: | |
2711 | ||
2712 | **0** | |
2713 | Default. Preallocate donor's file on init. | |
2714 | **1** | |
2b455dbf | 2715 | Allocate space immediately inside defragment event, and free right |
f80dba8d MT |
2716 | after event. |
2717 | ||
f3f96717 | 2718 | .. option:: clustername=str : [rbd,rados] |
f80dba8d MT |
2719 | |
2720 | Specifies the name of the Ceph cluster. | |
2721 | ||
2722 | .. option:: rbdname=str : [rbd] | |
2723 | ||
2724 | Specifies the name of the RBD. | |
2725 | ||
f3f96717 | 2726 | .. option:: clientname=str : [rbd,rados] |
f80dba8d MT |
2727 | |
2728 | Specifies the username (without the 'client.' prefix) used to access the | |
2729 | Ceph cluster. If the *clustername* is specified, the *clientname* shall be | |
2730 | the full *type.id* string. If no type. prefix is given, fio will add | |
2731 | 'client.' by default. | |
2732 | ||
873db854 | 2733 | .. option:: conf=str : [rados] |
2734 | ||
2735 | Specifies the configuration path of ceph cluster, so conf file does not | |
2736 | have to be /etc/ceph/ceph.conf. | |
2737 | ||
f3f96717 IF |
2738 | .. option:: busy_poll=bool : [rbd,rados] |
2739 | ||
2740 | Poll store instead of waiting for completion. Usually this provides better | |
2741 | throughput at cost of higher(up to 100%) CPU utilization. | |
2742 | ||
2b728756 AK |
2743 | .. option:: touch_objects=bool : [rados] |
2744 | ||
2745 | During initialization, touch (create if do not exist) all objects (files). | |
2746 | Touching all objects affects ceph caches and likely impacts test results. | |
2747 | Enabled by default. | |
2748 | ||
68522f38 VF |
2749 | .. option:: pool=str : |
2750 | ||
2751 | [rbd,rados] | |
2752 | ||
2753 | Specifies the name of the Ceph pool containing RBD or RADOS data. | |
2754 | ||
2755 | [dfs] | |
2756 | ||
2757 | Specify the label or UUID of the DAOS pool to connect to. | |
2758 | ||
2759 | .. option:: cont=str : [dfs] | |
2760 | ||
2761 | Specify the label or UUID of the DAOS container to open. | |
2762 | ||
19d8e50a VF |
2763 | .. option:: chunk_size=int |
2764 | ||
2765 | [dfs] | |
68522f38 | 2766 | |
ffe1d11f | 2767 | Specify a different chunk size (in bytes) for the dfs file. |
68522f38 VF |
2768 | Use DAOS container's chunk size by default. |
2769 | ||
19d8e50a VF |
2770 | [libhdfs] |
2771 | ||
2772 | The size of the chunk to use for each file. | |
2773 | ||
68522f38 VF |
2774 | .. option:: object_class=str : [dfs] |
2775 | ||
ffe1d11f | 2776 | Specify a different object class for the dfs file. |
68522f38 VF |
2777 | Use DAOS container's object class by default. |
2778 | ||
f80dba8d MT |
2779 | .. option:: skip_bad=bool : [mtd] |
2780 | ||
2781 | Skip operations against known bad blocks. | |
2782 | ||
2783 | .. option:: hdfsdirectory : [libhdfs] | |
2784 | ||
2785 | libhdfs will create chunk in this HDFS directory. | |
2786 | ||
609ac152 SB |
2787 | .. option:: verb=str : [rdma] |
2788 | ||
2789 | The RDMA verb to use on this side of the RDMA ioengine connection. Valid | |
2790 | values are write, read, send and recv. These correspond to the equivalent | |
2791 | RDMA verbs (e.g. write = rdma_write etc.). Note that this only needs to be | |
2792 | specified on the client side of the connection. See the examples folder. | |
2793 | ||
2794 | .. option:: bindname=str : [rdma] | |
2795 | ||
2796 | The name to use to bind the local RDMA-CM connection to a local RDMA device. | |
2797 | This could be a hostname or an IPv4 or IPv6 address. On the server side this | |
2798 | will be passed into the rdma_bind_addr() function and on the client site it | |
2799 | will be used in the rdma_resolve_add() function. This can be useful when | |
2800 | multiple paths exist between the client and the server or in certain loopback | |
2801 | configurations. | |
f80dba8d | 2802 | |
93a13ba5 | 2803 | .. option:: stat_type=str : [filestat] |
c446eff0 | 2804 | |
93a13ba5 TK |
2805 | Specify stat system call type to measure lookup/getattr performance. |
2806 | Default is **stat** for :manpage:`stat(2)`. | |
c446eff0 | 2807 | |
52b81b7c KD |
2808 | .. option:: readfua=bool : [sg] |
2809 | ||
2810 | With readfua option set to 1, read operations include | |
2811 | the force unit access (fua) flag. Default is 0. | |
2812 | ||
2813 | .. option:: writefua=bool : [sg] | |
2814 | ||
2815 | With writefua option set to 1, write operations include | |
2816 | the force unit access (fua) flag. Default is 0. | |
2817 | ||
2c3a9150 | 2818 | .. option:: sg_write_mode=str : [sg] |
3740cfc8 | 2819 | |
5b7565b0 | 2820 | Specify the type of write commands to issue. This option can take ten values: |
2c3a9150 VF |
2821 | |
2822 | **write** | |
2823 | This is the default where write opcodes are issued as usual. | |
eadf3260 | 2824 | **write_and_verify** |
2c3a9150 VF |
2825 | Issue WRITE AND VERIFY commands. The BYTCHK bit is set to 0. This |
2826 | directs the device to carry out a medium verification with no data | |
2827 | comparison. The writefua option is ignored with this selection. | |
eadf3260 VF |
2828 | **verify** |
2829 | This option is deprecated. Use write_and_verify instead. | |
2830 | **write_same** | |
2c3a9150 VF |
2831 | Issue WRITE SAME commands. This transfers a single block to the device |
2832 | and writes this same block of data to a contiguous sequence of LBAs | |
2833 | beginning at the specified offset. fio's block size parameter specifies | |
2834 | the amount of data written with each command. However, the amount of data | |
2835 | actually transferred to the device is equal to the device's block | |
2836 | (sector) size. For a device with 512 byte sectors, blocksize=8k will | |
2837 | write 16 sectors with each command. fio will still generate 8k of data | |
2838 | for each command but only the first 512 bytes will be used and | |
2839 | transferred to the device. The writefua option is ignored with this | |
2840 | selection. | |
eadf3260 VF |
2841 | **same** |
2842 | This option is deprecated. Use write_same instead. | |
91e13ff5 VF |
2843 | **write_same_ndob** |
2844 | Issue WRITE SAME(16) commands as above but with the No Data Output | |
2845 | Buffer (NDOB) bit set. No data will be transferred to the device with | |
2846 | this bit set. Data written will be a pre-determined pattern such as | |
2847 | all zeroes. | |
71efbed6 VF |
2848 | **write_stream** |
2849 | Issue WRITE STREAM(16) commands. Use the **stream_id** option to specify | |
2850 | the stream identifier. | |
e8ab121c VF |
2851 | **verify_bytchk_00** |
2852 | Issue VERIFY commands with BYTCHK set to 00. This directs the | |
2853 | device to carry out a medium verification with no data comparison. | |
2854 | **verify_bytchk_01** | |
2855 | Issue VERIFY commands with BYTCHK set to 01. This directs the device to | |
2856 | compare the data on the device with the data transferred to the device. | |
2857 | **verify_bytchk_11** | |
2858 | Issue VERIFY commands with BYTCHK set to 11. This transfers a | |
2859 | single block to the device and compares the contents of this block with the | |
2860 | data on the device beginning at the specified offset. fio's block size | |
2861 | parameter specifies the total amount of data compared with this command. | |
2862 | However, only one block (sector) worth of data is transferred to the device. | |
2863 | This is similar to the WRITE SAME command except that data is compared instead | |
2864 | of written. | |
52b81b7c | 2865 | |
71efbed6 VF |
2866 | .. option:: stream_id=int : [sg] |
2867 | ||
2868 | Set the stream identifier for WRITE STREAM commands. If this is set to 0 (which is not | |
2869 | a valid stream identifier) fio will open a stream and then close it when done. Default | |
2870 | is 0. | |
2871 | ||
c2f6a13d LMB |
2872 | .. option:: http_host=str : [http] |
2873 | ||
2874 | Hostname to connect to. For S3, this could be the bucket hostname. | |
2875 | Default is **localhost** | |
2876 | ||
2877 | .. option:: http_user=str : [http] | |
2878 | ||
2879 | Username for HTTP authentication. | |
2880 | ||
2881 | .. option:: http_pass=str : [http] | |
2882 | ||
2883 | Password for HTTP authentication. | |
2884 | ||
09fd2966 | 2885 | .. option:: https=str : [http] |
c2f6a13d | 2886 | |
09fd2966 LMB |
2887 | Enable HTTPS instead of http. *on* enables HTTPS; *insecure* |
2888 | will enable HTTPS, but disable SSL peer verification (use with | |
2889 | caution!). Default is **off** | |
c2f6a13d | 2890 | |
09fd2966 | 2891 | .. option:: http_mode=str : [http] |
c2f6a13d | 2892 | |
09fd2966 LMB |
2893 | Which HTTP access mode to use: *webdav*, *swift*, or *s3*. |
2894 | Default is **webdav** | |
c2f6a13d LMB |
2895 | |
2896 | .. option:: http_s3_region=str : [http] | |
2897 | ||
2898 | The S3 region/zone string. | |
2899 | Default is **us-east-1** | |
2900 | ||
2901 | .. option:: http_s3_key=str : [http] | |
2902 | ||
2903 | The S3 secret key. | |
2904 | ||
2905 | .. option:: http_s3_keyid=str : [http] | |
2906 | ||
2907 | The S3 key/access id. | |
2908 | ||
a2084df0 FH |
2909 | .. option:: http_s3_sse_customer_key=str : [http] |
2910 | ||
2911 | The encryption customer key in SSE server side. | |
2912 | ||
2913 | .. option:: http_s3_sse_customer_algorithm=str : [http] | |
2914 | ||
2915 | The encryption customer algorithm in SSE server side. | |
2916 | Default is **AES256** | |
2917 | ||
2918 | .. option:: http_s3_storage_class=str : [http] | |
2919 | ||
2920 | Which storage class to access. User-customizable settings. | |
2921 | Default is **STANDARD** | |
2922 | ||
09fd2966 LMB |
2923 | .. option:: http_swift_auth_token=str : [http] |
2924 | ||
2925 | The Swift auth token. See the example configuration file on how | |
2926 | to retrieve this. | |
2927 | ||
c2f6a13d LMB |
2928 | .. option:: http_verbose=int : [http] |
2929 | ||
2930 | Enable verbose requests from libcurl. Useful for debugging. 1 | |
2931 | turns on verbose logging from libcurl, 2 additionally enables | |
2932 | HTTP IO tracing. Default is **0** | |
2933 | ||
f2d6de5d RJ |
2934 | .. option:: uri=str : [nbd] |
2935 | ||
2936 | Specify the NBD URI of the server to test. The string | |
2937 | is a standard NBD URI | |
2938 | (see https://github.com/NetworkBlockDevice/nbd/tree/master/doc). | |
2939 | Example URIs: nbd://localhost:10809 | |
2940 | nbd+unix:///?socket=/tmp/socket | |
2941 | nbds://tlshost/exportname | |
2942 | ||
10756b2c BS |
2943 | .. option:: gpu_dev_ids=str : [libcufile] |
2944 | ||
2945 | Specify the GPU IDs to use with CUDA. This is a colon-separated list of | |
2946 | int. GPUs are assigned to workers roundrobin. Default is 0. | |
2947 | ||
2948 | .. option:: cuda_io=str : [libcufile] | |
2949 | ||
2950 | Specify the type of I/O to use with CUDA. Default is **cufile**. | |
2951 | ||
2952 | **cufile** | |
2953 | Use libcufile and nvidia-fs. This option performs I/O directly | |
2954 | between a GPUDirect Storage filesystem and GPU buffers, | |
2955 | avoiding use of a bounce buffer. If :option:`verify` is set, | |
2956 | cudaMemcpy is used to copy verificaton data between RAM and GPU. | |
2957 | Verification data is copied from RAM to GPU before a write | |
2958 | and from GPU to RAM after a read. :option:`direct` must be 1. | |
2959 | **posix** | |
2960 | Use POSIX to perform I/O with a RAM buffer, and use cudaMemcpy | |
2961 | to transfer data between RAM and the GPUs. Data is copied from | |
2962 | GPU to RAM before a write and copied from RAM to GPU after a | |
2963 | read. :option:`verify` does not affect use of cudaMemcpy. | |
2964 | ||
9326926b TG |
2965 | .. option:: nfs_url=str : [nfs] |
2966 | ||
2967 | URL in libnfs format, eg nfs://<server|ipv4|ipv6>/path[?arg=val[&arg=val]*] | |
2968 | Refer to the libnfs README for more details. | |
2969 | ||
b50590bc EV |
2970 | .. option:: program=str : [exec] |
2971 | ||
2972 | Specify the program to execute. | |
2973 | ||
2974 | .. option:: arguments=str : [exec] | |
2975 | ||
2976 | Specify arguments to pass to program. | |
2977 | Some special variables can be expanded to pass fio's job details to the program. | |
2978 | ||
2979 | **%r** | |
2980 | Replaced by the duration of the job in seconds. | |
2981 | **%n** | |
2982 | Replaced by the name of the job. | |
2983 | ||
2984 | .. option:: grace_time=int : [exec] | |
2985 | ||
2986 | Specify the time between the SIGTERM and SIGKILL signals. Default is 1 second. | |
2987 | ||
81c7079c | 2988 | .. option:: std_redirect=bool : [exec] |
b50590bc EV |
2989 | |
2990 | If set, stdout and stderr streams are redirected to files named from the job name. Default is true. | |
2991 | ||
454154e6 AK |
2992 | .. option:: xnvme_async=str : [xnvme] |
2993 | ||
2994 | Select the xnvme async command interface. This can take these values. | |
2995 | ||
2996 | **emu** | |
4deb92f9 AK |
2997 | This is default and use to emulate asynchronous I/O by using a |
2998 | single thread to create a queue pair on top of a synchronous | |
2999 | I/O interface using the NVMe driver IOCTL. | |
454154e6 | 3000 | **thrpool** |
4deb92f9 AK |
3001 | Emulate an asynchronous I/O interface with a pool of userspace |
3002 | threads on top of a synchronous I/O interface using the NVMe | |
3003 | driver IOCTL. By default four threads are used. | |
454154e6 | 3004 | **io_uring** |
4deb92f9 AK |
3005 | Linux native asynchronous I/O interface which supports both |
3006 | direct and buffered I/O. | |
3007 | **io_uring_cmd** | |
3008 | Fast Linux native asynchronous I/O interface for NVMe pass | |
3009 | through commands. This only works with NVMe character device | |
3010 | (/dev/ngXnY). | |
454154e6 AK |
3011 | **libaio** |
3012 | Use Linux aio for Asynchronous I/O. | |
3013 | **posix** | |
4deb92f9 AK |
3014 | Use the posix asynchronous I/O interface to perform one or |
3015 | more I/O operations asynchronously. | |
203a4c7c AK |
3016 | **vfio** |
3017 | Use the user-space VFIO-based backend, implemented using | |
3018 | libvfn instead of SPDK. | |
454154e6 | 3019 | **nil** |
4deb92f9 AK |
3020 | Do not transfer any data; just pretend to. This is mainly used |
3021 | for introspective performance evaluation. | |
454154e6 AK |
3022 | |
3023 | .. option:: xnvme_sync=str : [xnvme] | |
3024 | ||
3025 | Select the xnvme synchronous command interface. This can take these values. | |
3026 | ||
3027 | **nvme** | |
4deb92f9 AK |
3028 | This is default and uses Linux NVMe Driver ioctl() for |
3029 | synchronous I/O. | |
454154e6 | 3030 | **psync** |
4deb92f9 AK |
3031 | This supports regular as well as vectored pread() and pwrite() |
3032 | commands. | |
3033 | **block** | |
3034 | This is the same as psync except that it also supports zone | |
3035 | management commands using Linux block layer IOCTLs. | |
454154e6 AK |
3036 | |
3037 | .. option:: xnvme_admin=str : [xnvme] | |
3038 | ||
3039 | Select the xnvme admin command interface. This can take these values. | |
3040 | ||
3041 | **nvme** | |
4deb92f9 AK |
3042 | This is default and uses linux NVMe Driver ioctl() for admin |
3043 | commands. | |
454154e6 AK |
3044 | **block** |
3045 | Use Linux Block Layer ioctl() and sysfs for admin commands. | |
454154e6 AK |
3046 | |
3047 | .. option:: xnvme_dev_nsid=int : [xnvme] | |
3048 | ||
203a4c7c | 3049 | xnvme namespace identifier for userspace NVMe driver, SPDK or vfio. |
454154e6 | 3050 | |
efbafe2a AK |
3051 | .. option:: xnvme_dev_subnqn=str : [xnvme] |
3052 | ||
3053 | Sets the subsystem NQN for fabrics. This is for xNVMe to utilize a | |
3054 | fabrics target with multiple systems. | |
3055 | ||
c945074c AK |
3056 | .. option:: xnvme_mem=str : [xnvme] |
3057 | ||
3058 | Select the xnvme memory backend. This can take these values. | |
3059 | ||
3060 | **posix** | |
3061 | This is the default posix memory backend for linux NVMe driver. | |
3062 | **hugepage** | |
3063 | Use hugepages, instead of existing posix memory backend. The | |
3064 | memory backend uses hugetlbfs. This require users to allocate | |
85ccc10a | 3065 | hugepages, mount hugetlbfs and set an environment variable for |
c945074c AK |
3066 | XNVME_HUGETLB_PATH. |
3067 | **spdk** | |
3068 | Uses SPDK's memory allocator. | |
3069 | **vfio** | |
3070 | Uses libvfn's memory allocator. This also specifies the use | |
3071 | of libvfn backend instead of SPDK. | |
3072 | ||
454154e6 AK |
3073 | .. option:: xnvme_iovec=int : [xnvme] |
3074 | ||
3075 | If this option is set. xnvme will use vectored read/write commands. | |
3076 | ||
a601337a AF |
3077 | .. option:: libblkio_driver=str : [libblkio] |
3078 | ||
3079 | The libblkio *driver* to use. Different drivers access devices through | |
3080 | different underlying interfaces. Available drivers depend on the | |
3081 | libblkio version in use and are listed at | |
3082 | https://libblkio.gitlab.io/libblkio/blkio.html#drivers | |
3083 | ||
13fffdfb AF |
3084 | .. option:: libblkio_path=str : [libblkio] |
3085 | ||
3086 | Sets the value of the driver-specific "path" property before connecting | |
3087 | the libblkio instance, which identifies the target device or file on | |
3088 | which to perform I/O. Its exact semantics are driver-dependent and not | |
3089 | all drivers may support it; see | |
3090 | https://libblkio.gitlab.io/libblkio/blkio.html#drivers | |
3091 | ||
a601337a AF |
3092 | .. option:: libblkio_pre_connect_props=str : [libblkio] |
3093 | ||
13fffdfb AF |
3094 | A colon-separated list of additional libblkio properties to be set after |
3095 | creating but before connecting the libblkio instance. Each property must | |
3096 | have the format ``<name>=<value>``. Colons can be escaped as ``\:``. | |
3097 | These are set after the engine sets any other properties, so those can | |
85ccc10a | 3098 | be overridden. Available properties depend on the libblkio version in use |
a601337a AF |
3099 | and are listed at |
3100 | https://libblkio.gitlab.io/libblkio/blkio.html#properties | |
3101 | ||
13fffdfb AF |
3102 | .. option:: libblkio_num_entries=int : [libblkio] |
3103 | ||
3104 | Sets the value of the driver-specific "num-entries" property before | |
3105 | starting the libblkio instance. Its exact semantics are driver-dependent | |
3106 | and not all drivers may support it; see | |
3107 | https://libblkio.gitlab.io/libblkio/blkio.html#drivers | |
3108 | ||
3109 | .. option:: libblkio_queue_size=int : [libblkio] | |
3110 | ||
3111 | Sets the value of the driver-specific "queue-size" property before | |
3112 | starting the libblkio instance. Its exact semantics are driver-dependent | |
3113 | and not all drivers may support it; see | |
3114 | https://libblkio.gitlab.io/libblkio/blkio.html#drivers | |
3115 | ||
a601337a AF |
3116 | .. option:: libblkio_pre_start_props=str : [libblkio] |
3117 | ||
13fffdfb AF |
3118 | A colon-separated list of additional libblkio properties to be set after |
3119 | connecting but before starting the libblkio instance. Each property must | |
3120 | have the format ``<name>=<value>``. Colons can be escaped as ``\:``. | |
3121 | These are set after the engine sets any other properties, so those can | |
85ccc10a | 3122 | be overridden. Available properties depend on the libblkio version in use |
a601337a AF |
3123 | and are listed at |
3124 | https://libblkio.gitlab.io/libblkio/blkio.html#properties | |
3125 | ||
6dd4291c AF |
3126 | .. option:: libblkio_vectored : [libblkio] |
3127 | ||
3128 | Submit vectored read and write requests. | |
3129 | ||
464981ff AF |
3130 | .. option:: libblkio_write_zeroes_on_trim : [libblkio] |
3131 | ||
3132 | Submit trims as "write zeroes" requests instead of discard requests. | |
3133 | ||
b158577d AF |
3134 | .. option:: libblkio_wait_mode=str : [libblkio] |
3135 | ||
3136 | How to wait for completions: | |
3137 | ||
3138 | **block** (default) | |
3139 | Use a blocking call to ``blkioq_do_io()``. | |
3140 | **eventfd** | |
3141 | Use a blocking call to ``read()`` on the completion eventfd. | |
3142 | **loop** | |
3143 | Use a busy loop with a non-blocking call to ``blkioq_do_io()``. | |
3144 | ||
b1bd09b5 AF |
3145 | .. option:: libblkio_force_enable_completion_eventfd : [libblkio] |
3146 | ||
3147 | Enable the queue's completion eventfd even when unused. This may impact | |
3148 | performance. The default is to enable it only if | |
3149 | :option:`libblkio_wait_mode=eventfd <libblkio_wait_mode>`. | |
3150 | ||
a64fd9c7 VF |
3151 | .. option:: no_completion_thread : [windowsaio] |
3152 | ||
3153 | Avoid using a separate thread for completion polling. | |
3154 | ||
f80dba8d MT |
3155 | I/O depth |
3156 | ~~~~~~~~~ | |
3157 | ||
3158 | .. option:: iodepth=int | |
3159 | ||
3160 | Number of I/O units to keep in flight against the file. Note that | |
3161 | increasing *iodepth* beyond 1 will not affect synchronous ioengines (except | |
c60ebc45 | 3162 | for small degrees when :option:`verify_async` is in use). Even async |
f80dba8d MT |
3163 | engines may impose OS restrictions causing the desired depth not to be |
3164 | achieved. This may happen on Linux when using libaio and not setting | |
9207a0cb | 3165 | :option:`direct`\=1, since buffered I/O is not async on that OS. Keep an |
f80dba8d MT |
3166 | eye on the I/O depth distribution in the fio output to verify that the |
3167 | achieved depth is as expected. Default: 1. | |
3168 | ||
3169 | .. option:: iodepth_batch_submit=int, iodepth_batch=int | |
3170 | ||
3171 | This defines how many pieces of I/O to submit at once. It defaults to 1 | |
3172 | which means that we submit each I/O as soon as it is available, but can be | |
3173 | raised to submit bigger batches of I/O at the time. If it is set to 0 the | |
3174 | :option:`iodepth` value will be used. | |
3175 | ||
3176 | .. option:: iodepth_batch_complete_min=int, iodepth_batch_complete=int | |
3177 | ||
3178 | This defines how many pieces of I/O to retrieve at once. It defaults to 1 | |
3179 | which means that we'll ask for a minimum of 1 I/O in the retrieval process | |
3180 | from the kernel. The I/O retrieval will go on until we hit the limit set by | |
3181 | :option:`iodepth_low`. If this variable is set to 0, then fio will always | |
3182 | check for completed events before queuing more I/O. This helps reduce I/O | |
3183 | latency, at the cost of more retrieval system calls. | |
3184 | ||
3185 | .. option:: iodepth_batch_complete_max=int | |
3186 | ||
3187 | This defines maximum pieces of I/O to retrieve at once. This variable should | |
9207a0cb | 3188 | be used along with :option:`iodepth_batch_complete_min`\=int variable, |
f80dba8d | 3189 | specifying the range of min and max amount of I/O which should be |
730bd7d9 | 3190 | retrieved. By default it is equal to the :option:`iodepth_batch_complete_min` |
f80dba8d MT |
3191 | value. |
3192 | ||
3193 | Example #1:: | |
3194 | ||
3195 | iodepth_batch_complete_min=1 | |
3196 | iodepth_batch_complete_max=<iodepth> | |
3197 | ||
3198 | which means that we will retrieve at least 1 I/O and up to the whole | |
3199 | submitted queue depth. If none of I/O has been completed yet, we will wait. | |
3200 | ||
3201 | Example #2:: | |
3202 | ||
3203 | iodepth_batch_complete_min=0 | |
3204 | iodepth_batch_complete_max=<iodepth> | |
3205 | ||
3206 | which means that we can retrieve up to the whole submitted queue depth, but | |
3207 | if none of I/O has been completed yet, we will NOT wait and immediately exit | |
3208 | the system call. In this example we simply do polling. | |
3209 | ||
3210 | .. option:: iodepth_low=int | |
3211 | ||
3212 | The low water mark indicating when to start filling the queue | |
3213 | again. Defaults to the same as :option:`iodepth`, meaning that fio will | |
3214 | attempt to keep the queue full at all times. If :option:`iodepth` is set to | |
c60ebc45 | 3215 | e.g. 16 and *iodepth_low* is set to 4, then after fio has filled the queue of |
f80dba8d MT |
3216 | 16 requests, it will let the depth drain down to 4 before starting to fill |
3217 | it again. | |
3218 | ||
997b5680 SW |
3219 | .. option:: serialize_overlap=bool |
3220 | ||
3221 | Serialize in-flight I/Os that might otherwise cause or suffer from data races. | |
3222 | When two or more I/Os are submitted simultaneously, there is no guarantee that | |
3223 | the I/Os will be processed or completed in the submitted order. Further, if | |
3224 | two or more of those I/Os are writes, any overlapping region between them can | |
3225 | become indeterminate/undefined on certain storage. These issues can cause | |
3226 | verification to fail erratically when at least one of the racing I/Os is | |
3227 | changing data and the overlapping region has a non-zero size. Setting | |
3228 | ``serialize_overlap`` tells fio to avoid provoking this behavior by explicitly | |
3229 | serializing in-flight I/Os that have a non-zero overlap. Note that setting | |
ee21ebee | 3230 | this option can reduce both performance and the :option:`iodepth` achieved. |
3d6a6f04 VF |
3231 | |
3232 | This option only applies to I/Os issued for a single job except when it is | |
a02ec45a | 3233 | enabled along with :option:`io_submit_mode`\=offload. In offload mode, fio |
3d6a6f04 | 3234 | will check for overlap among all I/Os submitted by offload jobs with :option:`serialize_overlap` |
307f2246 | 3235 | enabled. |
3d6a6f04 VF |
3236 | |
3237 | Default: false. | |
997b5680 | 3238 | |
f80dba8d MT |
3239 | .. option:: io_submit_mode=str |
3240 | ||
3241 | This option controls how fio submits the I/O to the I/O engine. The default | |
3242 | is `inline`, which means that the fio job threads submit and reap I/O | |
3243 | directly. If set to `offload`, the job threads will offload I/O submission | |
3244 | to a dedicated pool of I/O threads. This requires some coordination and thus | |
3245 | has a bit of extra overhead, especially for lower queue depth I/O where it | |
3246 | can increase latencies. The benefit is that fio can manage submission rates | |
3247 | independently of the device completion rates. This avoids skewed latency | |
730bd7d9 | 3248 | reporting if I/O gets backed up on the device side (the coordinated omission |
abfd235a JA |
3249 | problem). Note that this option cannot reliably be used with async IO |
3250 | engines. | |
f80dba8d MT |
3251 | |
3252 | ||
3253 | I/O rate | |
3254 | ~~~~~~~~ | |
3255 | ||
0d8cc753 CL |
3256 | .. option:: thinkcycles=int |
3257 | ||
3258 | Stall the job for the specified number of cycles after an I/O has completed before | |
3259 | issuing the next. May be used to simulate processing being done by an application. | |
3260 | This is not taken into account for the time to be waited on for :option:`thinktime`. | |
3261 | Might not have any effect on some platforms, this can be checked by trying a setting | |
3262 | a high enough amount of thinkcycles. | |
3263 | ||
a881438b | 3264 | .. option:: thinktime=time |
f80dba8d | 3265 | |
f75ede1d SW |
3266 | Stall the job for the specified period of time after an I/O has completed before issuing the |
3267 | next. May be used to simulate processing being done by an application. | |
947e0fe0 | 3268 | When the unit is omitted, the value is interpreted in microseconds. See |
f7942acd | 3269 | :option:`thinktime_blocks`, :option:`thinktime_iotime` and :option:`thinktime_spin`. |
f80dba8d | 3270 | |
a881438b | 3271 | .. option:: thinktime_spin=time |
f80dba8d MT |
3272 | |
3273 | Only valid if :option:`thinktime` is set - pretend to spend CPU time doing | |
3274 | something with the data received, before falling back to sleeping for the | |
f75ede1d | 3275 | rest of the period specified by :option:`thinktime`. When the unit is |
947e0fe0 | 3276 | omitted, the value is interpreted in microseconds. |
f80dba8d MT |
3277 | |
3278 | .. option:: thinktime_blocks=int | |
3279 | ||
3280 | Only valid if :option:`thinktime` is set - control how many blocks to issue, | |
f50fbdda TK |
3281 | before waiting :option:`thinktime` usecs. If not set, defaults to 1 which will make |
3282 | fio wait :option:`thinktime` usecs after every block. This effectively makes any | |
f80dba8d | 3283 | queue depth setting redundant, since no more than 1 I/O will be queued |
f50fbdda | 3284 | before we have to complete it and do our :option:`thinktime`. In other words, this |
f80dba8d | 3285 | setting effectively caps the queue depth if the latter is larger. |
71bfa161 | 3286 | |
33f42c20 HQ |
3287 | .. option:: thinktime_blocks_type=str |
3288 | ||
3289 | Only valid if :option:`thinktime` is set - control how :option:`thinktime_blocks` | |
3290 | triggers. The default is `complete`, which triggers thinktime when fio completes | |
3291 | :option:`thinktime_blocks` blocks. If this is set to `issue`, then the trigger happens | |
3292 | at the issue side. | |
3293 | ||
f7942acd SK |
3294 | .. option:: thinktime_iotime=time |
3295 | ||
3296 | Only valid if :option:`thinktime` is set - control :option:`thinktime` | |
3297 | interval by time. The :option:`thinktime` stall is repeated after IOs | |
3298 | are executed for :option:`thinktime_iotime`. For example, | |
3299 | ``--thinktime_iotime=9s --thinktime=1s`` repeat 10-second cycle with IOs | |
3300 | for 9 seconds and stall for 1 second. When the unit is omitted, | |
3301 | :option:`thinktime_iotime` is interpreted as a number of seconds. If | |
3302 | this option is used together with :option:`thinktime_blocks`, the | |
3303 | :option:`thinktime` stall is repeated after :option:`thinktime_iotime` | |
3304 | or after :option:`thinktime_blocks` IOs, whichever happens first. | |
3305 | ||
f80dba8d | 3306 | .. option:: rate=int[,int][,int] |
71bfa161 | 3307 | |
f80dba8d MT |
3308 | Cap the bandwidth used by this job. The number is in bytes/sec, the normal |
3309 | suffix rules apply. Comma-separated values may be specified for reads, | |
3310 | writes, and trims as described in :option:`blocksize`. | |
71bfa161 | 3311 | |
b25b3464 SW |
3312 | For example, using `rate=1m,500k` would limit reads to 1MiB/sec and writes to |
3313 | 500KiB/sec. Capping only reads or writes can be done with `rate=,500k` or | |
3314 | `rate=500k,` where the former will only limit writes (to 500KiB/sec) and the | |
3315 | latter will only limit reads. | |
3316 | ||
f80dba8d | 3317 | .. option:: rate_min=int[,int][,int] |
71bfa161 | 3318 | |
f80dba8d MT |
3319 | Tell fio to do whatever it can to maintain at least this bandwidth. Failing |
3320 | to meet this requirement will cause the job to exit. Comma-separated values | |
3321 | may be specified for reads, writes, and trims as described in | |
3322 | :option:`blocksize`. | |
71bfa161 | 3323 | |
f80dba8d | 3324 | .. option:: rate_iops=int[,int][,int] |
71bfa161 | 3325 | |
f80dba8d MT |
3326 | Cap the bandwidth to this number of IOPS. Basically the same as |
3327 | :option:`rate`, just specified independently of bandwidth. If the job is | |
3328 | given a block size range instead of a fixed value, the smallest block size | |
3329 | is used as the metric. Comma-separated values may be specified for reads, | |
3330 | writes, and trims as described in :option:`blocksize`. | |
71bfa161 | 3331 | |
f80dba8d | 3332 | .. option:: rate_iops_min=int[,int][,int] |
71bfa161 | 3333 | |
f80dba8d MT |
3334 | If fio doesn't meet this rate of I/O, it will cause the job to exit. |
3335 | Comma-separated values may be specified for reads, writes, and trims as | |
3336 | described in :option:`blocksize`. | |
71bfa161 | 3337 | |
f80dba8d | 3338 | .. option:: rate_process=str |
66c098b8 | 3339 | |
f80dba8d MT |
3340 | This option controls how fio manages rated I/O submissions. The default is |
3341 | `linear`, which submits I/O in a linear fashion with fixed delays between | |
c60ebc45 | 3342 | I/Os that gets adjusted based on I/O completion rates. If this is set to |
f80dba8d MT |
3343 | `poisson`, fio will submit I/O based on a more real world random request |
3344 | flow, known as the Poisson process | |
3345 | (https://en.wikipedia.org/wiki/Poisson_point_process). The lambda will be | |
3346 | 10^6 / IOPS for the given workload. | |
71bfa161 | 3347 | |
1a9bf814 JA |
3348 | .. option:: rate_ignore_thinktime=bool |
3349 | ||
3350 | By default, fio will attempt to catch up to the specified rate setting, | |
3351 | if any kind of thinktime setting was used. If this option is set, then | |
3352 | fio will ignore the thinktime and continue doing IO at the specified | |
3353 | rate, instead of entering a catch-up mode after thinktime is done. | |
3354 | ||
afb34fb1 VF |
3355 | .. option:: rate_cycle=int |
3356 | ||
7d6c99e9 VF |
3357 | Average bandwidth for :option:`rate_min` and :option:`rate_iops_min` |
3358 | over this number of milliseconds. Defaults to 1000. | |
afb34fb1 | 3359 | |
71bfa161 | 3360 | |
f80dba8d MT |
3361 | I/O latency |
3362 | ~~~~~~~~~~~ | |
71bfa161 | 3363 | |
a881438b | 3364 | .. option:: latency_target=time |
71bfa161 | 3365 | |
f80dba8d | 3366 | If set, fio will attempt to find the max performance point that the given |
f75ede1d | 3367 | workload will run at while maintaining a latency below this target. When |
947e0fe0 | 3368 | the unit is omitted, the value is interpreted in microseconds. See |
f75ede1d | 3369 | :option:`latency_window` and :option:`latency_percentile`. |
71bfa161 | 3370 | |
a881438b | 3371 | .. option:: latency_window=time |
71bfa161 | 3372 | |
f80dba8d | 3373 | Used with :option:`latency_target` to specify the sample window that the job |
f75ede1d | 3374 | is run at varying queue depths to test the performance. When the unit is |
947e0fe0 | 3375 | omitted, the value is interpreted in microseconds. |
b4692828 | 3376 | |
f80dba8d | 3377 | .. option:: latency_percentile=float |
71bfa161 | 3378 | |
c60ebc45 | 3379 | The percentage of I/Os that must fall within the criteria specified by |
f80dba8d | 3380 | :option:`latency_target` and :option:`latency_window`. If not set, this |
c60ebc45 | 3381 | defaults to 100.0, meaning that all I/Os must be equal or below to the value |
f80dba8d | 3382 | set by :option:`latency_target`. |
71bfa161 | 3383 | |
e1bcd541 SL |
3384 | .. option:: latency_run=bool |
3385 | ||
3386 | Used with :option:`latency_target`. If false (default), fio will find | |
3387 | the highest queue depth that meets :option:`latency_target` and exit. If | |
3388 | true, fio will continue running and try to meet :option:`latency_target` | |
3389 | by adjusting queue depth. | |
3390 | ||
f7cf63bf | 3391 | .. option:: max_latency=time[,time][,time] |
71bfa161 | 3392 | |
f75ede1d | 3393 | If set, fio will exit the job with an ETIMEDOUT error if it exceeds this |
947e0fe0 | 3394 | maximum latency. When the unit is omitted, the value is interpreted in |
f7cf63bf VR |
3395 | microseconds. Comma-separated values may be specified for reads, writes, |
3396 | and trims as described in :option:`blocksize`. | |
71bfa161 | 3397 | |
71bfa161 | 3398 | |
f80dba8d MT |
3399 | I/O replay |
3400 | ~~~~~~~~~~ | |
71bfa161 | 3401 | |
f80dba8d | 3402 | .. option:: write_iolog=str |
c2b1e753 | 3403 | |
f80dba8d MT |
3404 | Write the issued I/O patterns to the specified file. See |
3405 | :option:`read_iolog`. Specify a separate file for each job, otherwise the | |
02a36caa VF |
3406 | iologs will be interspersed and the file may be corrupt. This file will |
3407 | be opened in append mode. | |
c2b1e753 | 3408 | |
f80dba8d | 3409 | .. option:: read_iolog=str |
71bfa161 | 3410 | |
22413915 | 3411 | Open an iolog with the specified filename and replay the I/O patterns it |
f80dba8d MT |
3412 | contains. This can be used to store a workload and replay it sometime |
3413 | later. The iolog given may also be a blktrace binary file, which allows fio | |
3414 | to replay a workload captured by :command:`blktrace`. See | |
3415 | :manpage:`blktrace(8)` for how to capture such logging data. For blktrace | |
3416 | replay, the file needs to be turned into a blkparse binary data file first | |
3417 | (``blkparse <device> -o /dev/null -d file_for_fio.bin``). | |
78439a18 JA |
3418 | You can specify a number of files by separating the names with a ':' |
3419 | character. See the :option:`filename` option for information on how to | |
3b803fe1 | 3420 | escape ':' characters within the file names. These files will |
78439a18 | 3421 | be sequentially assigned to job clones created by :option:`numjobs`. |
d19c04d1 | 3422 | '-' is a reserved name, meaning read from stdin, notably if |
3423 | :option:`filename` is set to '-' which means stdin as well, then | |
3424 | this flag can't be set to '-'. | |
71bfa161 | 3425 | |
77be374d AK |
3426 | .. option:: read_iolog_chunked=bool |
3427 | ||
3428 | Determines how iolog is read. If false(default) entire :option:`read_iolog` | |
3429 | will be read at once. If selected true, input from iolog will be read | |
3430 | gradually. Useful when iolog is very large, or it is generated. | |
3431 | ||
b9921d1a DZ |
3432 | .. option:: merge_blktrace_file=str |
3433 | ||
3434 | When specified, rather than replaying the logs passed to :option:`read_iolog`, | |
3435 | the logs go through a merge phase which aggregates them into a single | |
3436 | blktrace. The resulting file is then passed on as the :option:`read_iolog` | |
3437 | parameter. The intention here is to make the order of events consistent. | |
3438 | This limits the influence of the scheduler compared to replaying multiple | |
3439 | blktraces via concurrent jobs. | |
3440 | ||
87a48ada DZ |
3441 | .. option:: merge_blktrace_scalars=float_list |
3442 | ||
3443 | This is a percentage based option that is index paired with the list of | |
3444 | files passed to :option:`read_iolog`. When merging is performed, scale | |
3445 | the time of each event by the corresponding amount. For example, | |
3446 | ``--merge_blktrace_scalars="50:100"`` runs the first trace in halftime | |
3447 | and the second trace in realtime. This knob is separately tunable from | |
3448 | :option:`replay_time_scale` which scales the trace during runtime and | |
3449 | does not change the output of the merge unlike this option. | |
3450 | ||
55bfd8c8 DZ |
3451 | .. option:: merge_blktrace_iters=float_list |
3452 | ||
3453 | This is a whole number option that is index paired with the list of files | |
3454 | passed to :option:`read_iolog`. When merging is performed, run each trace | |
3455 | for the specified number of iterations. For example, | |
3456 | ``--merge_blktrace_iters="2:1"`` runs the first trace for two iterations | |
3457 | and the second trace for one iteration. | |
3458 | ||
589e88b7 | 3459 | .. option:: replay_no_stall=bool |
71bfa161 | 3460 | |
f80dba8d | 3461 | When replaying I/O with :option:`read_iolog` the default behavior is to |
22413915 | 3462 | attempt to respect the timestamps within the log and replay them with the |
f80dba8d MT |
3463 | appropriate delay between IOPS. By setting this variable fio will not |
3464 | respect the timestamps and attempt to replay them as fast as possible while | |
3465 | still respecting ordering. The result is the same I/O pattern to a given | |
3466 | device, but different timings. | |
71bfa161 | 3467 | |
6dd7fa77 JA |
3468 | .. option:: replay_time_scale=int |
3469 | ||
3470 | When replaying I/O with :option:`read_iolog`, fio will honor the | |
3471 | original timing in the trace. With this option, it's possible to scale | |
3472 | the time. It's a percentage option, if set to 50 it means run at 50% | |
3473 | the original IO rate in the trace. If set to 200, run at twice the | |
3474 | original IO rate. Defaults to 100. | |
3475 | ||
f80dba8d | 3476 | .. option:: replay_redirect=str |
b4692828 | 3477 | |
f80dba8d MT |
3478 | While replaying I/O patterns using :option:`read_iolog` the default behavior |
3479 | is to replay the IOPS onto the major/minor device that each IOP was recorded | |
3480 | from. This is sometimes undesirable because on a different machine those | |
3481 | major/minor numbers can map to a different device. Changing hardware on the | |
3482 | same system can also result in a different major/minor mapping. | |
730bd7d9 | 3483 | ``replay_redirect`` causes all I/Os to be replayed onto the single specified |
f80dba8d | 3484 | device regardless of the device it was recorded |
9207a0cb | 3485 | from. i.e. :option:`replay_redirect`\= :file:`/dev/sdc` would cause all I/O |
f80dba8d MT |
3486 | in the blktrace or iolog to be replayed onto :file:`/dev/sdc`. This means |
3487 | multiple devices will be replayed onto a single device, if the trace | |
3488 | contains multiple devices. If you want multiple devices to be replayed | |
3489 | concurrently to multiple redirected devices you must blkparse your trace | |
3490 | into separate traces and replay them with independent fio invocations. | |
3491 | Unfortunately this also breaks the strict time ordering between multiple | |
3492 | device accesses. | |
71bfa161 | 3493 | |
f80dba8d | 3494 | .. option:: replay_align=int |
74929ac2 | 3495 | |
350a535d DZ |
3496 | Force alignment of the byte offsets in a trace to this value. The value |
3497 | must be a power of 2. | |
3c54bc46 | 3498 | |
f80dba8d | 3499 | .. option:: replay_scale=int |
3c54bc46 | 3500 | |
350a535d DZ |
3501 | Scale byte offsets down by this factor when replaying traces. Should most |
3502 | likely use :option:`replay_align` as well. | |
3c54bc46 | 3503 | |
38f68906 JA |
3504 | .. option:: replay_skip=str |
3505 | ||
3506 | Sometimes it's useful to skip certain IO types in a replay trace. | |
3507 | This could be, for instance, eliminating the writes in the trace. | |
3508 | Or not replaying the trims/discards, if you are redirecting to | |
3509 | a device that doesn't support them. This option takes a comma | |
3510 | separated list of read, write, trim, sync. | |
3511 | ||
3c54bc46 | 3512 | |
f80dba8d MT |
3513 | Threads, processes and job synchronization |
3514 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
3c54bc46 | 3515 | |
f80dba8d | 3516 | .. option:: thread |
3c54bc46 | 3517 | |
730bd7d9 SW |
3518 | Fio defaults to creating jobs by using fork, however if this option is |
3519 | given, fio will create jobs by using POSIX Threads' function | |
3520 | :manpage:`pthread_create(3)` to create threads instead. | |
71bfa161 | 3521 | |
f80dba8d | 3522 | .. option:: wait_for=str |
74929ac2 | 3523 | |
730bd7d9 SW |
3524 | If set, the current job won't be started until all workers of the specified |
3525 | waitee job are done. | |
74929ac2 | 3526 | |
f80dba8d MT |
3527 | ``wait_for`` operates on the job name basis, so there are a few |
3528 | limitations. First, the waitee must be defined prior to the waiter job | |
3529 | (meaning no forward references). Second, if a job is being referenced as a | |
3530 | waitee, it must have a unique name (no duplicate waitees). | |
74929ac2 | 3531 | |
f80dba8d | 3532 | .. option:: nice=int |
892a6ffc | 3533 | |
f80dba8d | 3534 | Run the job with the given nice value. See man :manpage:`nice(2)`. |
892a6ffc | 3535 | |
f80dba8d MT |
3536 | On Windows, values less than -15 set the process class to "High"; -1 through |
3537 | -15 set "Above Normal"; 1 through 15 "Below Normal"; and above 15 "Idle" | |
3538 | priority class. | |
74929ac2 | 3539 | |
f80dba8d | 3540 | .. option:: prio=int |
71bfa161 | 3541 | |
f80dba8d MT |
3542 | Set the I/O priority value of this job. Linux limits us to a positive value |
3543 | between 0 and 7, with 0 being the highest. See man | |
3544 | :manpage:`ionice(1)`. Refer to an appropriate manpage for other operating | |
b2a432bf | 3545 | systems since meaning of priority may differ. For per-command priority |
12f9d54a DLM |
3546 | setting, see I/O engine specific :option:`cmdprio_percentage` and |
3547 | :option:`cmdprio` options. | |
71bfa161 | 3548 | |
f80dba8d | 3549 | .. option:: prioclass=int |
d59aa780 | 3550 | |
b2a432bf | 3551 | Set the I/O priority class. See man :manpage:`ionice(1)`. For per-command |
12f9d54a DLM |
3552 | priority setting, see I/O engine specific :option:`cmdprio_percentage` |
3553 | and :option:`cmdprio_class` options. | |
d59aa780 | 3554 | |
860462da DLM |
3555 | .. option:: priohint=int |
3556 | ||
3557 | Set the I/O priority hint. This is only applicable to platforms that | |
3558 | support I/O priority classes and to devices with features controlled | |
3559 | through priority hints, e.g. block devices supporting command duration | |
3560 | limits, or CDL. CDL is a way to indicate the desired maximum latency | |
3561 | of I/Os so that the device can optimize its internal command scheduling | |
3562 | according to the latency limits indicated by the user. | |
3563 | ||
79012fec DLM |
3564 | For per-I/O priority hint setting, see the I/O engine specific |
3565 | :option:`cmdprio_hint` option. | |
3566 | ||
f80dba8d | 3567 | .. option:: cpus_allowed=str |
6d500c2e | 3568 | |
730bd7d9 | 3569 | Controls the same options as :option:`cpumask`, but accepts a textual |
b570e037 SW |
3570 | specification of the permitted CPUs instead and CPUs are indexed from 0. So |
3571 | to use CPUs 0 and 5 you would specify ``cpus_allowed=0,5``. This option also | |
3572 | allows a range of CPUs to be specified -- say you wanted a binding to CPUs | |
3573 | 0, 5, and 8 to 15, you would set ``cpus_allowed=0,5,8-15``. | |
3574 | ||
3575 | On Windows, when ``cpus_allowed`` is unset only CPUs from fio's current | |
3576 | processor group will be used and affinity settings are inherited from the | |
3577 | system. An fio build configured to target Windows 7 makes options that set | |
3578 | CPUs processor group aware and values will set both the processor group | |
3579 | and a CPU from within that group. For example, on a system where processor | |
3580 | group 0 has 40 CPUs and processor group 1 has 32 CPUs, ``cpus_allowed`` | |
3581 | values between 0 and 39 will bind CPUs from processor group 0 and | |
3582 | ``cpus_allowed`` values between 40 and 71 will bind CPUs from processor | |
3583 | group 1. When using ``cpus_allowed_policy=shared`` all CPUs specified by a | |
3584 | single ``cpus_allowed`` option must be from the same processor group. For | |
3585 | Windows fio builds not built for Windows 7, CPUs will only be selected from | |
3586 | (and be relative to) whatever processor group fio happens to be running in | |
3587 | and CPUs from other processor groups cannot be used. | |
6d500c2e | 3588 | |
f80dba8d | 3589 | .. option:: cpus_allowed_policy=str |
6d500c2e | 3590 | |
f80dba8d | 3591 | Set the policy of how fio distributes the CPUs specified by |
730bd7d9 | 3592 | :option:`cpus_allowed` or :option:`cpumask`. Two policies are supported: |
6d500c2e | 3593 | |
f80dba8d MT |
3594 | **shared** |
3595 | All jobs will share the CPU set specified. | |
3596 | **split** | |
3597 | Each job will get a unique CPU from the CPU set. | |
6d500c2e | 3598 | |
22413915 | 3599 | **shared** is the default behavior, if the option isn't specified. If |
b21fc93f | 3600 | **split** is specified, then fio will assign one cpu per job. If not |
f80dba8d MT |
3601 | enough CPUs are given for the jobs listed, then fio will roundrobin the CPUs |
3602 | in the set. | |
6d500c2e | 3603 | |
b570e037 SW |
3604 | .. option:: cpumask=int |
3605 | ||
3606 | Set the CPU affinity of this job. The parameter given is a bit mask of | |
3607 | allowed CPUs the job may run on. So if you want the allowed CPUs to be 1 | |
3608 | and 5, you would pass the decimal value of (1 << 1 | 1 << 5), or 34. See man | |
3609 | :manpage:`sched_setaffinity(2)`. This may not work on all supported | |
3610 | operating systems or kernel versions. This option doesn't work well for a | |
3611 | higher CPU count than what you can store in an integer mask, so it can only | |
3612 | control cpus 1-32. For boxes with larger CPU counts, use | |
3613 | :option:`cpus_allowed`. | |
3614 | ||
f80dba8d | 3615 | .. option:: numa_cpu_nodes=str |
6d500c2e | 3616 | |
f80dba8d MT |
3617 | Set this job running on specified NUMA nodes' CPUs. The arguments allow |
3618 | comma delimited list of cpu numbers, A-B ranges, or `all`. Note, to enable | |
ac8ca2af | 3619 | NUMA options support, fio must be built on a system with libnuma-dev(el) |
f80dba8d | 3620 | installed. |
61b9861d | 3621 | |
f80dba8d | 3622 | .. option:: numa_mem_policy=str |
61b9861d | 3623 | |
f80dba8d MT |
3624 | Set this job's memory policy and corresponding NUMA nodes. Format of the |
3625 | arguments:: | |
5c94b008 | 3626 | |
f80dba8d | 3627 | <mode>[:<nodelist>] |
ce35b1ec | 3628 | |
804c0839 | 3629 | ``mode`` is one of the following memory policies: ``default``, ``prefer``, |
730bd7d9 SW |
3630 | ``bind``, ``interleave`` or ``local``. For ``default`` and ``local`` memory |
3631 | policies, no node needs to be specified. For ``prefer``, only one node is | |
3632 | allowed. For ``bind`` and ``interleave`` the ``nodelist`` may be as | |
3633 | follows: a comma delimited list of numbers, A-B ranges, or `all`. | |
71bfa161 | 3634 | |
f80dba8d | 3635 | .. option:: cgroup=str |
390b1537 | 3636 | |
f80dba8d MT |
3637 | Add job to this control group. If it doesn't exist, it will be created. The |
3638 | system must have a mounted cgroup blkio mount point for this to work. If | |
3639 | your system doesn't have it mounted, you can do so with:: | |
5af1c6f3 | 3640 | |
f80dba8d | 3641 | # mount -t cgroup -o blkio none /cgroup |
5af1c6f3 | 3642 | |
f80dba8d | 3643 | .. option:: cgroup_weight=int |
5af1c6f3 | 3644 | |
f80dba8d MT |
3645 | Set the weight of the cgroup to this value. See the documentation that comes |
3646 | with the kernel, allowed values are in the range of 100..1000. | |
a086c257 | 3647 | |
f80dba8d | 3648 | .. option:: cgroup_nodelete=bool |
8c07860d | 3649 | |
f80dba8d MT |
3650 | Normally fio will delete the cgroups it has created after the job |
3651 | completion. To override this behavior and to leave cgroups around after the | |
3652 | job completion, set ``cgroup_nodelete=1``. This can be useful if one wants | |
3653 | to inspect various cgroup files after job completion. Default: false. | |
8c07860d | 3654 | |
f80dba8d | 3655 | .. option:: flow_id=int |
8c07860d | 3656 | |
f80dba8d MT |
3657 | The ID of the flow. If not specified, it defaults to being a global |
3658 | flow. See :option:`flow`. | |
1907dbc6 | 3659 | |
f80dba8d | 3660 | .. option:: flow=int |
71bfa161 | 3661 | |
73f168ea VF |
3662 | Weight in token-based flow control. If this value is used, then fio |
3663 | regulates the activity between two or more jobs sharing the same | |
3664 | flow_id. Fio attempts to keep each job activity proportional to other | |
3665 | jobs' activities in the same flow_id group, with respect to requested | |
3666 | weight per job. That is, if one job has `flow=3', another job has | |
3667 | `flow=2' and another with `flow=1`, then there will be a roughly 3:2:1 | |
3668 | ratio in how much one runs vs the others. | |
71bfa161 | 3669 | |
f80dba8d | 3670 | .. option:: flow_sleep=int |
82407585 | 3671 | |
d4e74fda DB |
3672 | The period of time, in microseconds, to wait after the flow counter |
3673 | has exceeded its proportion before retrying operations. | |
82407585 | 3674 | |
f80dba8d | 3675 | .. option:: stonewall, wait_for_previous |
82407585 | 3676 | |
f80dba8d MT |
3677 | Wait for preceding jobs in the job file to exit, before starting this |
3678 | one. Can be used to insert serialization points in the job file. A stone | |
3679 | wall also implies starting a new reporting group, see | |
3680 | :option:`group_reporting`. | |
3681 | ||
3682 | .. option:: exitall | |
3683 | ||
64402a8a HW |
3684 | By default, fio will continue running all other jobs when one job finishes. |
3685 | Sometimes this is not the desired action. Setting ``exitall`` will instead | |
3686 | make fio terminate all jobs in the same group, as soon as one job of that | |
3687 | group finishes. | |
3688 | ||
7fc3a553 | 3689 | .. option:: exit_what=str |
64402a8a HW |
3690 | |
3691 | By default, fio will continue running all other jobs when one job finishes. | |
7fc3a553 | 3692 | Sometimes this is not the desired action. Setting ``exitall`` will |
64402a8a | 3693 | instead make fio terminate all jobs in the same group. The option |
85ccc10a MS |
3694 | ``exit_what`` allows one to control which jobs get terminated when ``exitall`` |
3695 | is enabled. The default is ``group`` and does not change the behaviour of | |
64402a8a HW |
3696 | ``exitall``. The setting ``all`` terminates all jobs. The setting ``stonewall`` |
3697 | terminates all currently running jobs across all groups and continues execution | |
3698 | with the next stonewalled group. | |
f80dba8d MT |
3699 | |
3700 | .. option:: exec_prerun=str | |
3701 | ||
3702 | Before running this job, issue the command specified through | |
3703 | :manpage:`system(3)`. Output is redirected in a file called | |
3704 | :file:`jobname.prerun.txt`. | |
3705 | ||
3706 | .. option:: exec_postrun=str | |
3707 | ||
3708 | After the job completes, issue the command specified though | |
3709 | :manpage:`system(3)`. Output is redirected in a file called | |
3710 | :file:`jobname.postrun.txt`. | |
3711 | ||
3712 | .. option:: uid=int | |
3713 | ||
3714 | Instead of running as the invoking user, set the user ID to this value | |
3715 | before the thread/process does any work. | |
3716 | ||
3717 | .. option:: gid=int | |
3718 | ||
3719 | Set group ID, see :option:`uid`. | |
3720 | ||
3721 | ||
3722 | Verification | |
3723 | ~~~~~~~~~~~~ | |
3724 | ||
3725 | .. option:: verify_only | |
3726 | ||
3727 | Do not perform specified workload, only verify data still matches previous | |
3728 | invocation of this workload. This option allows one to check data multiple | |
3729 | times at a later date without overwriting it. This option makes sense only | |
3730 | for workloads that write data, and does not support workloads with the | |
3731 | :option:`time_based` option set. | |
3732 | ||
3733 | .. option:: do_verify=bool | |
3734 | ||
3735 | Run the verify phase after a write phase. Only valid if :option:`verify` is | |
3736 | set. Default: true. | |
3737 | ||
3738 | .. option:: verify=str | |
3739 | ||
3740 | If writing to a file, fio can verify the file contents after each iteration | |
3741 | of the job. Each verification method also implies verification of special | |
3742 | header, which is written to the beginning of each block. This header also | |
3743 | includes meta information, like offset of the block, block number, timestamp | |
3744 | when block was written, etc. :option:`verify` can be combined with | |
3745 | :option:`verify_pattern` option. The allowed values are: | |
3746 | ||
3747 | **md5** | |
3748 | Use an md5 sum of the data area and store it in the header of | |
3749 | each block. | |
3750 | ||
3751 | **crc64** | |
3752 | Use an experimental crc64 sum of the data area and store it in the | |
3753 | header of each block. | |
3754 | ||
3755 | **crc32c** | |
a5896300 SW |
3756 | Use a crc32c sum of the data area and store it in the header of |
3757 | each block. This will automatically use hardware acceleration | |
3758 | (e.g. SSE4.2 on an x86 or CRC crypto extensions on ARM64) but will | |
3759 | fall back to software crc32c if none is found. Generally the | |
804c0839 | 3760 | fastest checksum fio supports when hardware accelerated. |
f80dba8d MT |
3761 | |
3762 | **crc32c-intel** | |
a5896300 | 3763 | Synonym for crc32c. |
f80dba8d MT |
3764 | |
3765 | **crc32** | |
3766 | Use a crc32 sum of the data area and store it in the header of each | |
3767 | block. | |
3768 | ||
3769 | **crc16** | |
3770 | Use a crc16 sum of the data area and store it in the header of each | |
3771 | block. | |
3772 | ||
3773 | **crc7** | |
3774 | Use a crc7 sum of the data area and store it in the header of each | |
3775 | block. | |
3776 | ||
3777 | **xxhash** | |
3778 | Use xxhash as the checksum function. Generally the fastest software | |
3779 | checksum that fio supports. | |
3780 | ||
3781 | **sha512** | |
3782 | Use sha512 as the checksum function. | |
3783 | ||
3784 | **sha256** | |
3785 | Use sha256 as the checksum function. | |
3786 | ||
3787 | **sha1** | |
3788 | Use optimized sha1 as the checksum function. | |
82407585 | 3789 | |
ae3a5acc JA |
3790 | **sha3-224** |
3791 | Use optimized sha3-224 as the checksum function. | |
3792 | ||
3793 | **sha3-256** | |
3794 | Use optimized sha3-256 as the checksum function. | |
3795 | ||
3796 | **sha3-384** | |
3797 | Use optimized sha3-384 as the checksum function. | |
3798 | ||
3799 | **sha3-512** | |
3800 | Use optimized sha3-512 as the checksum function. | |
3801 | ||
f80dba8d MT |
3802 | **meta** |
3803 | This option is deprecated, since now meta information is included in | |
3804 | generic verification header and meta verification happens by | |
3805 | default. For detailed information see the description of the | |
3806 | :option:`verify` setting. This option is kept because of | |
3807 | compatibility's sake with old configurations. Do not use it. | |
3808 | ||
3809 | **pattern** | |
3810 | Verify a strict pattern. Normally fio includes a header with some | |
3811 | basic information and checksumming, but if this option is set, only | |
3812 | the specific pattern set with :option:`verify_pattern` is verified. | |
3813 | ||
3814 | **null** | |
3815 | Only pretend to verify. Useful for testing internals with | |
9207a0cb | 3816 | :option:`ioengine`\=null, not for much else. |
f80dba8d MT |
3817 | |
3818 | This option can be used for repeated burn-in tests of a system to make sure | |
3819 | that the written data is also correctly read back. If the data direction | |
3820 | given is a read or random read, fio will assume that it should verify a | |
3821 | previously written file. If the data direction includes any form of write, | |
3822 | the verify will be of the newly written data. | |
3823 | ||
47e6a6e5 SW |
3824 | To avoid false verification errors, do not use the norandommap option when |
3825 | verifying data with async I/O engines and I/O depths > 1. Or use the | |
3826 | norandommap and the lfsr random generator together to avoid writing to the | |
fc002f14 | 3827 | same offset with multiple outstanding I/Os. |
47e6a6e5 | 3828 | |
f80dba8d MT |
3829 | .. option:: verify_offset=int |
3830 | ||
3831 | Swap the verification header with data somewhere else in the block before | |
3832 | writing. It is swapped back before verifying. | |
3833 | ||
3834 | .. option:: verify_interval=int | |
3835 | ||
3836 | Write the verification header at a finer granularity than the | |
3837 | :option:`blocksize`. It will be written for chunks the size of | |
3838 | ``verify_interval``. :option:`blocksize` should divide this evenly. | |
3839 | ||
3840 | .. option:: verify_pattern=str | |
3841 | ||
3842 | If set, fio will fill the I/O buffers with this pattern. Fio defaults to | |
3843 | filling with totally random bytes, but sometimes it's interesting to fill | |
3844 | with a known pattern for I/O verification purposes. Depending on the width | |
730bd7d9 | 3845 | of the pattern, fio will fill 1/2/3/4 bytes of the buffer at the time (it can |
f80dba8d MT |
3846 | be either a decimal or a hex number). The ``verify_pattern`` if larger than |
3847 | a 32-bit quantity has to be a hex number that starts with either "0x" or | |
3848 | "0X". Use with :option:`verify`. Also, ``verify_pattern`` supports %o | |
3849 | format, which means that for each block offset will be written and then | |
3850 | verified back, e.g.:: | |
61b9861d RP |
3851 | |
3852 | verify_pattern=%o | |
3853 | ||
f80dba8d MT |
3854 | Or use combination of everything:: |
3855 | ||
61b9861d | 3856 | verify_pattern=0xff%o"abcd"-12 |
e28218f3 | 3857 | |
f80dba8d MT |
3858 | .. option:: verify_fatal=bool |
3859 | ||
3860 | Normally fio will keep checking the entire contents before quitting on a | |
3861 | block verification failure. If this option is set, fio will exit the job on | |
3862 | the first observed failure. Default: false. | |
3863 | ||
3864 | .. option:: verify_dump=bool | |
3865 | ||
3866 | If set, dump the contents of both the original data block and the data block | |
3867 | we read off disk to files. This allows later analysis to inspect just what | |
3868 | kind of data corruption occurred. Off by default. | |
3869 | ||
3870 | .. option:: verify_async=int | |
3871 | ||
3872 | Fio will normally verify I/O inline from the submitting thread. This option | |
3873 | takes an integer describing how many async offload threads to create for I/O | |
3874 | verification instead, causing fio to offload the duty of verifying I/O | |
3875 | contents to one or more separate threads. If using this offload option, even | |
3876 | sync I/O engines can benefit from using an :option:`iodepth` setting higher | |
3877 | than 1, as it allows them to have I/O in flight while verifies are running. | |
d7e6ea1c | 3878 | Defaults to 0 async threads, i.e. verification is not asynchronous. |
f80dba8d MT |
3879 | |
3880 | .. option:: verify_async_cpus=str | |
3881 | ||
3882 | Tell fio to set the given CPU affinity on the async I/O verification | |
3883 | threads. See :option:`cpus_allowed` for the format used. | |
3884 | ||
3885 | .. option:: verify_backlog=int | |
3886 | ||
3887 | Fio will normally verify the written contents of a job that utilizes verify | |
3888 | once that job has completed. In other words, everything is written then | |
3889 | everything is read back and verified. You may want to verify continually | |
3890 | instead for a variety of reasons. Fio stores the meta data associated with | |
3891 | an I/O block in memory, so for large verify workloads, quite a bit of memory | |
3892 | would be used up holding this meta data. If this option is enabled, fio will | |
3893 | write only N blocks before verifying these blocks. | |
3894 | ||
3895 | .. option:: verify_backlog_batch=int | |
3896 | ||
3897 | Control how many blocks fio will verify if :option:`verify_backlog` is | |
3898 | set. If not set, will default to the value of :option:`verify_backlog` | |
3899 | (meaning the entire queue is read back and verified). If | |
3900 | ``verify_backlog_batch`` is less than :option:`verify_backlog` then not all | |
3901 | blocks will be verified, if ``verify_backlog_batch`` is larger than | |
3902 | :option:`verify_backlog`, some blocks will be verified more than once. | |
3903 | ||
3904 | .. option:: verify_state_save=bool | |
3905 | ||
3906 | When a job exits during the write phase of a verify workload, save its | |
3907 | current state. This allows fio to replay up until that point, if the verify | |
3908 | state is loaded for the verify read phase. The format of the filename is, | |
3909 | roughly:: | |
3910 | ||
f50fbdda | 3911 | <type>-<jobname>-<jobindex>-verify.state. |
f80dba8d MT |
3912 | |
3913 | <type> is "local" for a local run, "sock" for a client/server socket | |
3914 | connection, and "ip" (192.168.0.1, for instance) for a networked | |
d7e6ea1c | 3915 | client/server connection. Defaults to true. |
f80dba8d MT |
3916 | |
3917 | .. option:: verify_state_load=bool | |
3918 | ||
3919 | If a verify termination trigger was used, fio stores the current write state | |
3920 | of each thread. This can be used at verification time so that fio knows how | |
3921 | far it should verify. Without this information, fio will run a full | |
a47b697c SW |
3922 | verification pass, according to the settings in the job file used. Default |
3923 | false. | |
f80dba8d | 3924 | |
899e057e VF |
3925 | .. option:: experimental_verify=bool |
3926 | ||
3927 | Enable experimental verification. Standard verify records I/O metadata | |
3928 | for later use during the verification phase. Experimental verify | |
3929 | instead resets the file after the write phase and then replays I/Os for | |
3930 | the verification phase. | |
3931 | ||
f80dba8d MT |
3932 | .. option:: trim_percentage=int |
3933 | ||
3934 | Number of verify blocks to discard/trim. | |
3935 | ||
3936 | .. option:: trim_verify_zero=bool | |
3937 | ||
22413915 | 3938 | Verify that trim/discarded blocks are returned as zeros. |
f80dba8d MT |
3939 | |
3940 | .. option:: trim_backlog=int | |
3941 | ||
5cfd1e9a | 3942 | Trim after this number of blocks are written. |
f80dba8d MT |
3943 | |
3944 | .. option:: trim_backlog_batch=int | |
3945 | ||
3946 | Trim this number of I/O blocks. | |
3947 | ||
f80dba8d MT |
3948 | Steady state |
3949 | ~~~~~~~~~~~~ | |
3950 | ||
3951 | .. option:: steadystate=str:float, ss=str:float | |
3952 | ||
3953 | Define the criterion and limit for assessing steady state performance. The | |
3954 | first parameter designates the criterion whereas the second parameter sets | |
3955 | the threshold. When the criterion falls below the threshold for the | |
3956 | specified duration, the job will stop. For example, `iops_slope:0.1%` will | |
3957 | direct fio to terminate the job when the least squares regression slope | |
3958 | falls below 0.1% of the mean IOPS. If :option:`group_reporting` is enabled | |
3959 | this will apply to all jobs in the group. Below is the list of available | |
3960 | steady state assessment criteria. All assessments are carried out using only | |
3961 | data from the rolling collection window. Threshold limits can be expressed | |
3962 | as a fixed value or as a percentage of the mean in the collection window. | |
3963 | ||
1cb049d9 VF |
3964 | When using this feature, most jobs should include the :option:`time_based` |
3965 | and :option:`runtime` options or the :option:`loops` option so that fio does not | |
3966 | stop running after it has covered the full size of the specified file(s) or device(s). | |
3967 | ||
f80dba8d MT |
3968 | **iops** |
3969 | Collect IOPS data. Stop the job if all individual IOPS measurements | |
3970 | are within the specified limit of the mean IOPS (e.g., ``iops:2`` | |
3971 | means that all individual IOPS values must be within 2 of the mean, | |
3972 | whereas ``iops:0.2%`` means that all individual IOPS values must be | |
3973 | within 0.2% of the mean IOPS to terminate the job). | |
3974 | ||
3975 | **iops_slope** | |
3976 | Collect IOPS data and calculate the least squares regression | |
3977 | slope. Stop the job if the slope falls below the specified limit. | |
3978 | ||
3979 | **bw** | |
3980 | Collect bandwidth data. Stop the job if all individual bandwidth | |
3981 | measurements are within the specified limit of the mean bandwidth. | |
3982 | ||
3983 | **bw_slope** | |
3984 | Collect bandwidth data and calculate the least squares regression | |
3985 | slope. Stop the job if the slope falls below the specified limit. | |
3986 | ||
3987 | .. option:: steadystate_duration=time, ss_dur=time | |
3988 | ||
51bbb1a1 VF |
3989 | A rolling window of this duration will be used to judge whether steady |
3990 | state has been reached. Data will be collected every | |
3991 | :option:`ss_interval`. The default is 0 which disables steady state | |
3992 | detection. When the unit is omitted, the value is interpreted in | |
3993 | seconds. | |
f80dba8d MT |
3994 | |
3995 | .. option:: steadystate_ramp_time=time, ss_ramp=time | |
3996 | ||
3997 | Allow the job to run for the specified duration before beginning data | |
3998 | collection for checking the steady state job termination criterion. The | |
947e0fe0 | 3999 | default is 0. When the unit is omitted, the value is interpreted in seconds. |
f80dba8d | 4000 | |
90e678ba CL |
4001 | .. option:: steadystate_check_interval=time, ss_interval=time |
4002 | ||
51bbb1a1 VF |
4003 | The values during the rolling window will be collected with a period of |
4004 | this value. If :option:`ss_interval` is 30s and :option:`ss_dur` is | |
4005 | 300s, 10 measurements will be taken. Default is 1s but that might not | |
4006 | converge, especially for slower devices, so set this accordingly. When | |
4007 | the unit is omitted, the value is interpreted in seconds. | |
90e678ba | 4008 | |
f80dba8d MT |
4009 | |
4010 | Measurements and reporting | |
4011 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
4012 | ||
4013 | .. option:: per_job_logs=bool | |
4014 | ||
05fce19c VF |
4015 | If set to true, fio generates bw/clat/iops logs with per job unique |
4016 | filenames. If set to false, jobs with identical names will share a log | |
4017 | filename. Note that when this option is set to false log files will be | |
4018 | opened in append mode and if log files already exist the previous | |
4019 | contents will not be overwritten. Default: true. | |
f80dba8d MT |
4020 | |
4021 | .. option:: group_reporting | |
4022 | ||
4023 | It may sometimes be interesting to display statistics for groups of jobs as | |
4024 | a whole instead of for each individual job. This is especially true if | |
4025 | :option:`numjobs` is used; looking at individual thread/process output | |
4026 | quickly becomes unwieldy. To see the final report per-group instead of | |
4027 | per-job, use :option:`group_reporting`. Jobs in a file will be part of the | |
4028 | same reporting group, unless if separated by a :option:`stonewall`, or by | |
4029 | using :option:`new_group`. | |
4030 | ||
96d28909 MP |
4031 | NOTE: When :option:`group_reporting` is used along with `json` output, |
4032 | there are certain per-job properties which can be different between jobs | |
4033 | but do not have a natural group-level equivalent. Examples include | |
4034 | `kb_base`, `unit_base`, `sig_figs`, `thread_number`, `pid`, and | |
4035 | `job_start`. For these properties, the values for the first job are | |
4036 | recorded for the group. | |
12d325ca | 4037 | |
f80dba8d MT |
4038 | .. option:: new_group |
4039 | ||
4040 | Start a new reporting group. See: :option:`group_reporting`. If not given, | |
4041 | all jobs in a file will be part of the same reporting group, unless | |
4042 | separated by a :option:`stonewall`. | |
4043 | ||
589e88b7 | 4044 | .. option:: stats=bool |
8243be59 JA |
4045 | |
4046 | By default, fio collects and shows final output results for all jobs | |
4047 | that run. If this option is set to 0, then fio will ignore it in | |
4048 | the final stat output. | |
4049 | ||
f80dba8d MT |
4050 | .. option:: write_bw_log=str |
4051 | ||
4052 | If given, write a bandwidth log for this job. Can be used to store data of | |
074f0817 | 4053 | the bandwidth of the jobs in their lifetime. |
f80dba8d | 4054 | |
074f0817 SW |
4055 | If no str argument is given, the default filename of |
4056 | :file:`jobname_type.x.log` is used. Even when the argument is given, fio | |
4057 | will still append the type of log. So if one specifies:: | |
4058 | ||
4059 | write_bw_log=foo | |
f80dba8d | 4060 | |
074f0817 SW |
4061 | The actual log name will be :file:`foo_bw.x.log` where `x` is the index |
4062 | of the job (`1..N`, where `N` is the number of jobs). If | |
4063 | :option:`per_job_logs` is false, then the filename will not include the | |
4064 | `.x` job index. | |
e3cedca7 | 4065 | |
074f0817 SW |
4066 | The included :command:`fio_generate_plots` script uses :command:`gnuplot` to turn these |
4067 | text files into nice graphs. See `Log File Formats`_ for how data is | |
4068 | structured within the file. | |
4069 | ||
4070 | .. option:: write_lat_log=str | |
e3cedca7 | 4071 | |
074f0817 | 4072 | Same as :option:`write_bw_log`, except this option creates I/O |
77b7e675 SW |
4073 | submission (e.g., :file:`name_slat.x.log`), completion (e.g., |
4074 | :file:`name_clat.x.log`), and total (e.g., :file:`name_lat.x.log`) | |
074f0817 SW |
4075 | latency files instead. See :option:`write_bw_log` for details about |
4076 | the filename format and `Log File Formats`_ for how data is structured | |
4077 | within the files. | |
be4ecfdf | 4078 | |
f80dba8d | 4079 | .. option:: write_hist_log=str |
06842027 | 4080 | |
074f0817 | 4081 | Same as :option:`write_bw_log` but writes an I/O completion latency |
77b7e675 | 4082 | histogram file (e.g., :file:`name_hist.x.log`) instead. Note that this |
074f0817 SW |
4083 | file will be empty unless :option:`log_hist_msec` has also been set. |
4084 | See :option:`write_bw_log` for details about the filename format and | |
4085 | `Log File Formats`_ for how data is structured within the file. | |
06842027 | 4086 | |
f80dba8d | 4087 | .. option:: write_iops_log=str |
06842027 | 4088 | |
074f0817 | 4089 | Same as :option:`write_bw_log`, but writes an IOPS file (e.g. |
15417073 SW |
4090 | :file:`name_iops.x.log`) instead. Because fio defaults to individual |
4091 | I/O logging, the value entry in the IOPS log will be 1 unless windowed | |
4092 | logging (see :option:`log_avg_msec`) has been enabled. See | |
4093 | :option:`write_bw_log` for details about the filename format and `Log | |
4094 | File Formats`_ for how data is structured within the file. | |
06842027 | 4095 | |
0a852a50 DLM |
4096 | .. option:: log_entries=int |
4097 | ||
4098 | By default, fio will log an entry in the iops, latency, or bw log for | |
4099 | every I/O that completes. The initial number of I/O log entries is 1024. | |
4100 | When the log entries are all used, new log entries are dynamically | |
4101 | allocated. This dynamic log entry allocation may negatively impact | |
4102 | time-related statistics such as I/O tail latencies (e.g. 99.9th percentile | |
4103 | completion latency). This option allows specifying a larger initial | |
4104 | number of log entries to avoid run-time allocations of new log entries, | |
4105 | resulting in more precise time-related I/O statistics. | |
4106 | Also see :option:`log_avg_msec`. Defaults to 1024. | |
4107 | ||
f80dba8d | 4108 | .. option:: log_avg_msec=int |
06842027 | 4109 | |
119b7ce8 VF |
4110 | By default, fio will log an entry in the iops, latency, or bw log for |
4111 | every I/O that completes. When writing to the disk log, that can | |
4112 | quickly grow to a very large size. Setting this option directs fio to | |
4113 | instead record an average over the specified duration for each log | |
4114 | entry, reducing the resolution of the log. When the job completes, fio | |
4115 | will flush any accumulated latency log data, so the final log interval | |
4116 | may not match the value specified by this option and there may even be | |
4117 | duplicate timestamps. See :option:`log_window_value` as well. Defaults | |
4118 | to 0, logging entries for each I/O. Also see `Log File Formats`_. | |
06842027 | 4119 | |
f80dba8d | 4120 | .. option:: log_hist_msec=int |
06842027 | 4121 | |
f80dba8d MT |
4122 | Same as :option:`log_avg_msec`, but logs entries for completion latency |
4123 | histograms. Computing latency percentiles from averages of intervals using | |
c60ebc45 | 4124 | :option:`log_avg_msec` is inaccurate. Setting this option makes fio log |
f80dba8d MT |
4125 | histogram entries over the specified period of time, reducing log sizes for |
4126 | high IOPS devices while retaining percentile accuracy. See | |
074f0817 SW |
4127 | :option:`log_hist_coarseness` and :option:`write_hist_log` as well. |
4128 | Defaults to 0, meaning histogram logging is disabled. | |
06842027 | 4129 | |
f80dba8d | 4130 | .. option:: log_hist_coarseness=int |
06842027 | 4131 | |
f80dba8d MT |
4132 | Integer ranging from 0 to 6, defining the coarseness of the resolution of |
4133 | the histogram logs enabled with :option:`log_hist_msec`. For each increment | |
4134 | in coarseness, fio outputs half as many bins. Defaults to 0, for which | |
074f0817 SW |
4135 | histogram logs contain 1216 latency bins. See :option:`write_hist_log` |
4136 | and `Log File Formats`_. | |
8b28bd41 | 4137 | |
f391405c | 4138 | .. option:: log_window_value=str, log_max_value=str |
66c098b8 | 4139 | |
065212b3 AK |
4140 | If :option:`log_avg_msec` is set, fio by default logs the average over that |
4141 | window. This option determines whether fio logs the average, maximum or | |
4142 | both the values over the window. This only affects the latency logging, | |
4143 | as both average and maximum values for iops or bw log will be same. | |
4144 | Accepted values are: | |
4145 | ||
4146 | **avg** | |
4147 | Log average value over the window. The default. | |
4148 | ||
4149 | **max** | |
4150 | Log maximum value in the window. | |
4151 | ||
4152 | **both** | |
4153 | Log both average and maximum value over the window. | |
4154 | ||
4155 | **0** | |
4156 | Backward-compatible alias for **avg**. | |
4157 | ||
4158 | **1** | |
4159 | Backward-compatible alias for **max**. | |
a696fa2a | 4160 | |
589e88b7 | 4161 | .. option:: log_offset=bool |
a696fa2a | 4162 | |
f80dba8d | 4163 | If this is set, the iolog options will include the byte offset for the I/O |
5a83478f SW |
4164 | entry as well as the other data values. Defaults to 0 meaning that |
4165 | offsets are not present in logs. Also see `Log File Formats`_. | |
71bfa161 | 4166 | |
f80dba8d | 4167 | .. option:: log_compression=int |
7de87099 | 4168 | |
f80dba8d MT |
4169 | If this is set, fio will compress the I/O logs as it goes, to keep the |
4170 | memory footprint lower. When a log reaches the specified size, that chunk is | |
4171 | removed and compressed in the background. Given that I/O logs are fairly | |
4172 | highly compressible, this yields a nice memory savings for longer runs. The | |
4173 | downside is that the compression will consume some background CPU cycles, so | |
4174 | it may impact the run. This, however, is also true if the logging ends up | |
4175 | consuming most of the system memory. So pick your poison. The I/O logs are | |
4176 | saved normally at the end of a run, by decompressing the chunks and storing | |
4177 | them in the specified log file. This feature depends on the availability of | |
4178 | zlib. | |
e0b0d892 | 4179 | |
f80dba8d | 4180 | .. option:: log_compression_cpus=str |
e0b0d892 | 4181 | |
f80dba8d MT |
4182 | Define the set of CPUs that are allowed to handle online log compression for |
4183 | the I/O jobs. This can provide better isolation between performance | |
0cf90a62 SW |
4184 | sensitive jobs, and background compression work. See |
4185 | :option:`cpus_allowed` for the format used. | |
9e684a49 | 4186 | |
f80dba8d | 4187 | .. option:: log_store_compressed=bool |
9e684a49 | 4188 | |
f80dba8d MT |
4189 | If set, fio will store the log files in a compressed format. They can be |
4190 | decompressed with fio, using the :option:`--inflate-log` command line | |
4191 | parameter. The files will be stored with a :file:`.fz` suffix. | |
9e684a49 | 4192 | |
f80dba8d | 4193 | .. option:: log_unix_epoch=bool |
9e684a49 | 4194 | |
d252275b | 4195 | Backwards compatible alias for log_alternate_epoch. |
f80dba8d | 4196 | |
d5b3cfd4 | 4197 | .. option:: log_alternate_epoch=bool |
4198 | ||
4199 | If set, fio will log timestamps based on the epoch used by the clock specified | |
4200 | in the log_alternate_epoch_clock_id option, to the log files produced by | |
4201 | enabling write_type_log for each log type, instead of the default zero-based | |
4202 | timestamps. | |
4203 | ||
4204 | .. option:: log_alternate_epoch_clock_id=int | |
4205 | ||
d252275b | 4206 | Specifies the clock_id to be used by clock_gettime to obtain the alternate |
4207 | epoch if log_alternate_epoch is true. Otherwise has no effect. Default | |
4208 | value is 0, or CLOCK_REALTIME. | |
d5b3cfd4 | 4209 | |
f80dba8d MT |
4210 | .. option:: block_error_percentiles=bool |
4211 | ||
4212 | If set, record errors in trim block-sized units from writes and trims and | |
4213 | output a histogram of how many trims it took to get to errors, and what kind | |
4214 | of error was encountered. | |
4215 | ||
4216 | .. option:: bwavgtime=int | |
4217 | ||
4218 | Average the calculated bandwidth over the given time. Value is specified in | |
4219 | milliseconds. If the job also does bandwidth logging through | |
4220 | :option:`write_bw_log`, then the minimum of this option and | |
4221 | :option:`log_avg_msec` will be used. Default: 500ms. | |
4222 | ||
4223 | .. option:: iopsavgtime=int | |
4224 | ||
4225 | Average the calculated IOPS over the given time. Value is specified in | |
4226 | milliseconds. If the job also does IOPS logging through | |
4227 | :option:`write_iops_log`, then the minimum of this option and | |
4228 | :option:`log_avg_msec` will be used. Default: 500ms. | |
4229 | ||
4230 | .. option:: disk_util=bool | |
4231 | ||
4232 | Generate disk utilization statistics, if the platform supports it. | |
4233 | Default: true. | |
4234 | ||
4235 | .. option:: disable_lat=bool | |
4236 | ||
4237 | Disable measurements of total latency numbers. Useful only for cutting back | |
4238 | the number of calls to :manpage:`gettimeofday(2)`, as that does impact | |
4239 | performance at really high IOPS rates. Note that to really get rid of a | |
4240 | large amount of these calls, this option must be used with | |
f75ede1d | 4241 | :option:`disable_slat` and :option:`disable_bw_measurement` as well. |
f80dba8d MT |
4242 | |
4243 | .. option:: disable_clat=bool | |
4244 | ||
4245 | Disable measurements of completion latency numbers. See | |
4246 | :option:`disable_lat`. | |
4247 | ||
4248 | .. option:: disable_slat=bool | |
4249 | ||
4250 | Disable measurements of submission latency numbers. See | |
f50fbdda | 4251 | :option:`disable_lat`. |
f80dba8d | 4252 | |
f75ede1d | 4253 | .. option:: disable_bw_measurement=bool, disable_bw=bool |
f80dba8d MT |
4254 | |
4255 | Disable measurements of throughput/bandwidth numbers. See | |
4256 | :option:`disable_lat`. | |
4257 | ||
dd39b9ce VF |
4258 | .. option:: slat_percentiles=bool |
4259 | ||
4260 | Report submission latency percentiles. Submission latency is not recorded | |
4261 | for synchronous ioengines. | |
4262 | ||
f80dba8d MT |
4263 | .. option:: clat_percentiles=bool |
4264 | ||
dd39b9ce | 4265 | Report completion latency percentiles. |
b599759b JA |
4266 | |
4267 | .. option:: lat_percentiles=bool | |
4268 | ||
dd39b9ce VF |
4269 | Report total latency percentiles. Total latency is the sum of submission |
4270 | latency and completion latency. | |
f80dba8d MT |
4271 | |
4272 | .. option:: percentile_list=float_list | |
4273 | ||
dd39b9ce VF |
4274 | Overwrite the default list of percentiles for latencies and the block error |
4275 | histogram. Each number is a floating point number in the range (0,100], and | |
4276 | the maximum length of the list is 20. Use ``:`` to separate the numbers. For | |
c32ba107 | 4277 | example, ``--percentile_list=99.5:99.9`` will cause fio to report the |
dd39b9ce VF |
4278 | latency durations below which 99.5% and 99.9% of the observed latencies fell, |
4279 | respectively. | |
f80dba8d | 4280 | |
e883cb35 JF |
4281 | .. option:: significant_figures=int |
4282 | ||
c32ba107 JA |
4283 | If using :option:`--output-format` of `normal`, set the significant |
4284 | figures to this value. Higher values will yield more precise IOPS and | |
4285 | throughput units, while lower values will round. Requires a minimum | |
4286 | value of 1 and a maximum value of 10. Defaults to 4. | |
e883cb35 | 4287 | |
f80dba8d MT |
4288 | |
4289 | Error handling | |
4290 | ~~~~~~~~~~~~~~ | |
4291 | ||
4292 | .. option:: exitall_on_error | |
4293 | ||
4294 | When one job finishes in error, terminate the rest. The default is to wait | |
4295 | for each job to finish. | |
4296 | ||
4297 | .. option:: continue_on_error=str | |
4298 | ||
4299 | Normally fio will exit the job on the first observed failure. If this option | |
4300 | is set, fio will continue the job when there is a 'non-fatal error' (EIO or | |
4301 | EILSEQ) until the runtime is exceeded or the I/O size specified is | |
4302 | completed. If this option is used, there are two more stats that are | |
4303 | appended, the total error count and the first error. The error field given | |
4304 | in the stats is the first error that was hit during the run. | |
4305 | ||
dc305989 KK |
4306 | Note: a write error from the device may go unnoticed by fio when using |
4307 | buffered IO, as the write() (or similar) system call merely dirties the | |
4308 | kernel pages, unless :option:`sync` or :option:`direct` is used. Device IO | |
4309 | errors occur when the dirty data is actually written out to disk. If fully | |
4310 | sync writes aren't desirable, :option:`fsync` or :option:`fdatasync` can be | |
4311 | used as well. This is specific to writes, as reads are always synchronous. | |
4312 | ||
f80dba8d MT |
4313 | The allowed values are: |
4314 | ||
4315 | **none** | |
4316 | Exit on any I/O or verify errors. | |
4317 | ||
4318 | **read** | |
4319 | Continue on read errors, exit on all others. | |
4320 | ||
4321 | **write** | |
4322 | Continue on write errors, exit on all others. | |
4323 | ||
4324 | **io** | |
4325 | Continue on any I/O error, exit on all others. | |
4326 | ||
4327 | **verify** | |
4328 | Continue on verify errors, exit on all others. | |
4329 | ||
4330 | **all** | |
4331 | Continue on all errors. | |
4332 | ||
4333 | **0** | |
4334 | Backward-compatible alias for 'none'. | |
4335 | ||
4336 | **1** | |
4337 | Backward-compatible alias for 'all'. | |
4338 | ||
4339 | .. option:: ignore_error=str | |
4340 | ||
4341 | Sometimes you want to ignore some errors during test in that case you can | |
a35ef7cb TK |
4342 | specify error list for each error type, instead of only being able to |
4343 | ignore the default 'non-fatal error' using :option:`continue_on_error`. | |
f80dba8d MT |
4344 | ``ignore_error=READ_ERR_LIST,WRITE_ERR_LIST,VERIFY_ERR_LIST`` errors for |
4345 | given error type is separated with ':'. Error may be symbol ('ENOSPC', | |
4346 | 'ENOMEM') or integer. Example:: | |
4347 | ||
4348 | ignore_error=EAGAIN,ENOSPC:122 | |
4349 | ||
4350 | This option will ignore EAGAIN from READ, and ENOSPC and 122(EDQUOT) from | |
a35ef7cb TK |
4351 | WRITE. This option works by overriding :option:`continue_on_error` with |
4352 | the list of errors for each error type if any. | |
f80dba8d MT |
4353 | |
4354 | .. option:: error_dump=bool | |
4355 | ||
4356 | If set dump every error even if it is non fatal, true by default. If | |
4357 | disabled only fatal error will be dumped. | |
4358 | ||
f75ede1d SW |
4359 | Running predefined workloads |
4360 | ---------------------------- | |
4361 | ||
4362 | Fio includes predefined profiles that mimic the I/O workloads generated by | |
4363 | other tools. | |
4364 | ||
4365 | .. option:: profile=str | |
4366 | ||
4367 | The predefined workload to run. Current profiles are: | |
4368 | ||
4369 | **tiobench** | |
4370 | Threaded I/O bench (tiotest/tiobench) like workload. | |
4371 | ||
4372 | **act** | |
4373 | Aerospike Certification Tool (ACT) like workload. | |
4374 | ||
4375 | To view a profile's additional options use :option:`--cmdhelp` after specifying | |
4376 | the profile. For example:: | |
4377 | ||
f50fbdda | 4378 | $ fio --profile=act --cmdhelp |
f75ede1d SW |
4379 | |
4380 | Act profile options | |
4381 | ~~~~~~~~~~~~~~~~~~~ | |
4382 | ||
4383 | .. option:: device-names=str | |
4384 | :noindex: | |
4385 | ||
4386 | Devices to use. | |
4387 | ||
4388 | .. option:: load=int | |
4389 | :noindex: | |
4390 | ||
4391 | ACT load multiplier. Default: 1. | |
4392 | ||
4393 | .. option:: test-duration=time | |
4394 | :noindex: | |
4395 | ||
947e0fe0 SW |
4396 | How long the entire test takes to run. When the unit is omitted, the value |
4397 | is given in seconds. Default: 24h. | |
f75ede1d SW |
4398 | |
4399 | .. option:: threads-per-queue=int | |
4400 | :noindex: | |
4401 | ||
f50fbdda | 4402 | Number of read I/O threads per device. Default: 8. |
f75ede1d SW |
4403 | |
4404 | .. option:: read-req-num-512-blocks=int | |
4405 | :noindex: | |
4406 | ||
4407 | Number of 512B blocks to read at the time. Default: 3. | |
4408 | ||
4409 | .. option:: large-block-op-kbytes=int | |
4410 | :noindex: | |
4411 | ||
4412 | Size of large block ops in KiB (writes). Default: 131072. | |
4413 | ||
4414 | .. option:: prep | |
4415 | :noindex: | |
4416 | ||
4417 | Set to run ACT prep phase. | |
4418 | ||
4419 | Tiobench profile options | |
4420 | ~~~~~~~~~~~~~~~~~~~~~~~~ | |
4421 | ||
4422 | .. option:: size=str | |
4423 | :noindex: | |
4424 | ||
f50fbdda | 4425 | Size in MiB. |
f75ede1d SW |
4426 | |
4427 | .. option:: block=int | |
4428 | :noindex: | |
4429 | ||
4430 | Block size in bytes. Default: 4096. | |
4431 | ||
4432 | .. option:: numruns=int | |
4433 | :noindex: | |
4434 | ||
4435 | Number of runs. | |
4436 | ||
4437 | .. option:: dir=str | |
4438 | :noindex: | |
4439 | ||
4440 | Test directory. | |
4441 | ||
4442 | .. option:: threads=int | |
4443 | :noindex: | |
4444 | ||
4445 | Number of threads. | |
f80dba8d MT |
4446 | |
4447 | Interpreting the output | |
4448 | ----------------------- | |
4449 | ||
36214730 SW |
4450 | .. |
4451 | Example output was based on the following: | |
4452 | TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --time_based \ | |
4453 | --rate=1256k --bs=14K --name=quick --runtime=1s --name=mixed \ | |
4454 | --runtime=2m --rw=rw | |
4455 | ||
f80dba8d MT |
4456 | Fio spits out a lot of output. While running, fio will display the status of the |
4457 | jobs created. An example of that would be:: | |
4458 | ||
9d25d068 | 4459 | Jobs: 1 (f=1): [_(1),M(1)][24.8%][r=20.5MiB/s,w=23.5MiB/s][r=82,w=94 IOPS][eta 01m:31s] |
f80dba8d | 4460 | |
36214730 SW |
4461 | The characters inside the first set of square brackets denote the current status of |
4462 | each thread. The first character is the first job defined in the job file, and so | |
4463 | forth. The possible values (in typical life cycle order) are: | |
f80dba8d MT |
4464 | |
4465 | +------+-----+-----------------------------------------------------------+ | |
4466 | | Idle | Run | | | |
4467 | +======+=====+===========================================================+ | |
4468 | | P | | Thread setup, but not started. | | |
4469 | +------+-----+-----------------------------------------------------------+ | |
4470 | | C | | Thread created. | | |
4471 | +------+-----+-----------------------------------------------------------+ | |
4472 | | I | | Thread initialized, waiting or generating necessary data. | | |
4473 | +------+-----+-----------------------------------------------------------+ | |
4474 | | | p | Thread running pre-reading file(s). | | |
4475 | +------+-----+-----------------------------------------------------------+ | |
36214730 SW |
4476 | | | / | Thread is in ramp period. | |
4477 | +------+-----+-----------------------------------------------------------+ | |
f80dba8d MT |
4478 | | | R | Running, doing sequential reads. | |
4479 | +------+-----+-----------------------------------------------------------+ | |
4480 | | | r | Running, doing random reads. | | |
4481 | +------+-----+-----------------------------------------------------------+ | |
4482 | | | W | Running, doing sequential writes. | | |
4483 | +------+-----+-----------------------------------------------------------+ | |
4484 | | | w | Running, doing random writes. | | |
4485 | +------+-----+-----------------------------------------------------------+ | |
4486 | | | M | Running, doing mixed sequential reads/writes. | | |
4487 | +------+-----+-----------------------------------------------------------+ | |
4488 | | | m | Running, doing mixed random reads/writes. | | |
4489 | +------+-----+-----------------------------------------------------------+ | |
36214730 SW |
4490 | | | D | Running, doing sequential trims. | |
4491 | +------+-----+-----------------------------------------------------------+ | |
4492 | | | d | Running, doing random trims. | | |
4493 | +------+-----+-----------------------------------------------------------+ | |
4494 | | | F | Running, currently waiting for :manpage:`fsync(2)`. | | |
f80dba8d MT |
4495 | +------+-----+-----------------------------------------------------------+ |
4496 | | | V | Running, doing verification of written data. | | |
4497 | +------+-----+-----------------------------------------------------------+ | |
36214730 SW |
4498 | | f | | Thread finishing. | |
4499 | +------+-----+-----------------------------------------------------------+ | |
f80dba8d MT |
4500 | | E | | Thread exited, not reaped by main thread yet. | |
4501 | +------+-----+-----------------------------------------------------------+ | |
36214730 | 4502 | | _ | | Thread reaped. | |
f80dba8d MT |
4503 | +------+-----+-----------------------------------------------------------+ |
4504 | | X | | Thread reaped, exited with an error. | | |
4505 | +------+-----+-----------------------------------------------------------+ | |
4506 | | K | | Thread reaped, exited due to signal. | | |
4507 | +------+-----+-----------------------------------------------------------+ | |
4508 | ||
36214730 SW |
4509 | .. |
4510 | Example output was based on the following: | |
4511 | TZ=UTC fio --iodepth=8 --ioengine=null --size=100M --runtime=58m \ | |
4512 | --time_based --rate=2512k --bs=256K --numjobs=10 \ | |
4513 | --name=readers --rw=read --name=writers --rw=write | |
4514 | ||
f80dba8d | 4515 | Fio will condense the thread string as not to take up more space on the command |
36214730 | 4516 | line than needed. For instance, if you have 10 readers and 10 writers running, |
f80dba8d MT |
4517 | the output would look like this:: |
4518 | ||
9d25d068 | 4519 | Jobs: 20 (f=20): [R(10),W(10)][4.0%][r=20.5MiB/s,w=23.5MiB/s][r=82,w=94 IOPS][eta 57m:36s] |
f80dba8d | 4520 | |
36214730 SW |
4521 | Note that the status string is displayed in order, so it's possible to tell which of |
4522 | the jobs are currently doing what. In the example above this means that jobs 1--10 | |
4523 | are readers and 11--20 are writers. | |
f80dba8d MT |
4524 | |
4525 | The other values are fairly self explanatory -- number of threads currently | |
36214730 SW |
4526 | running and doing I/O, the number of currently open files (f=), the estimated |
4527 | completion percentage, the rate of I/O since last check (read speed listed first, | |
f50fbdda TK |
4528 | then write speed and optionally trim speed) in terms of bandwidth and IOPS, |
4529 | and time to completion for the current running group. It's impossible to estimate | |
4530 | runtime of the following groups (if any). | |
36214730 SW |
4531 | |
4532 | .. | |
4533 | Example output was based on the following: | |
4534 | TZ=UTC fio --iodepth=16 --ioengine=posixaio --filename=/tmp/fiofile \ | |
4535 | --direct=1 --size=100M --time_based --runtime=50s --rate_iops=89 \ | |
4536 | --bs=7K --name=Client1 --rw=write | |
4537 | ||
4538 | When fio is done (or interrupted by :kbd:`Ctrl-C`), it will show the data for | |
4539 | each thread, group of threads, and disks in that order. For each overall thread (or | |
4540 | group) the output looks like:: | |
4541 | ||
4542 | Client1: (groupid=0, jobs=1): err= 0: pid=16109: Sat Jun 24 12:07:54 2017 | |
4543 | write: IOPS=88, BW=623KiB/s (638kB/s)(30.4MiB/50032msec) | |
4544 | slat (nsec): min=500, max=145500, avg=8318.00, stdev=4781.50 | |
4545 | clat (usec): min=170, max=78367, avg=4019.02, stdev=8293.31 | |
4546 | lat (usec): min=174, max=78375, avg=4027.34, stdev=8291.79 | |
4547 | clat percentiles (usec): | |
4548 | | 1.00th=[ 302], 5.00th=[ 326], 10.00th=[ 343], 20.00th=[ 363], | |
4549 | | 30.00th=[ 392], 40.00th=[ 404], 50.00th=[ 416], 60.00th=[ 445], | |
4550 | | 70.00th=[ 816], 80.00th=[ 6718], 90.00th=[12911], 95.00th=[21627], | |
4551 | | 99.00th=[43779], 99.50th=[51643], 99.90th=[68682], 99.95th=[72877], | |
4552 | | 99.99th=[78119] | |
4553 | bw ( KiB/s): min= 532, max= 686, per=0.10%, avg=622.87, stdev=24.82, samples= 100 | |
4554 | iops : min= 76, max= 98, avg=88.98, stdev= 3.54, samples= 100 | |
29092211 VF |
4555 | lat (usec) : 250=0.04%, 500=64.11%, 750=4.81%, 1000=2.79% |
4556 | lat (msec) : 2=4.16%, 4=1.84%, 10=4.90%, 20=11.33%, 50=5.37% | |
4557 | lat (msec) : 100=0.65% | |
36214730 SW |
4558 | cpu : usr=0.27%, sys=0.18%, ctx=12072, majf=0, minf=21 |
4559 | IO depths : 1=85.0%, 2=13.1%, 4=1.8%, 8=0.1%, 16=0.0%, 32=0.0%, >=64=0.0% | |
4560 | submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% | |
4561 | complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% | |
4562 | issued rwt: total=0,4450,0, short=0,0,0, dropped=0,0,0 | |
4563 | latency : target=0, window=0, percentile=100.00%, depth=8 | |
4564 | ||
4565 | The job name (or first job's name when using :option:`group_reporting`) is printed, | |
4566 | along with the group id, count of jobs being aggregated, last error id seen (which | |
4567 | is 0 when there are no errors), pid/tid of that thread and the time the job/group | |
4568 | completed. Below are the I/O statistics for each data direction performed (showing | |
4569 | writes in the example above). In the order listed, they denote: | |
4570 | ||
4571 | **read/write/trim** | |
4572 | The string before the colon shows the I/O direction the statistics | |
4573 | are for. **IOPS** is the average I/Os performed per second. **BW** | |
4574 | is the average bandwidth rate shown as: value in power of 2 format | |
4575 | (value in power of 10 format). The last two values show: (**total | |
4576 | I/O performed** in power of 2 format / **runtime** of that thread). | |
f80dba8d MT |
4577 | |
4578 | **slat** | |
36214730 SW |
4579 | Submission latency (**min** being the minimum, **max** being the |
4580 | maximum, **avg** being the average, **stdev** being the standard | |
13ddd98b VF |
4581 | deviation). This is the time from when fio initialized the I/O |
4582 | to submission. For synchronous ioengines this includes the time | |
4583 | up until just before the ioengine's queue function is called. | |
4584 | For asynchronous ioengines this includes the time up through the | |
4585 | completion of the ioengine's queue function (and commit function | |
4586 | if it is defined). For sync I/O this row is not displayed as the | |
4587 | slat is negligible. This value can be in nanoseconds, | |
4588 | microseconds or milliseconds --- fio will choose the most | |
4589 | appropriate base and print that (in the example above | |
4590 | nanoseconds was the best scale). Note: in :option:`--minimal` | |
4591 | mode latencies are always expressed in microseconds. | |
f80dba8d MT |
4592 | |
4593 | **clat** | |
4594 | Completion latency. Same names as slat, this denotes the time from | |
13ddd98b VF |
4595 | submission to completion of the I/O pieces. For sync I/O, this |
4596 | represents the time from when the I/O was submitted to the | |
4597 | operating system to when it was completed. For asynchronous | |
4598 | ioengines this is the time from when the ioengine's queue (and | |
4599 | commit if available) functions were completed to when the I/O's | |
4600 | completion was reaped by fio. | |
f80dba8d | 4601 | |
2cd920c1 | 4602 | For file and directory operation engines, **clat** denotes the time |
4603 | to complete one file or directory operation. | |
4604 | ||
4605 | **filecreate engine**:the time cost to create a new file | |
4606 | ||
4607 | **filestat engine**: the time cost to look up an existing file | |
4608 | ||
4609 | **filedelete engine**:the time cost to delete a file | |
4610 | ||
4611 | **dircreate engine**: the time cost to create a new directory | |
4612 | ||
4613 | **dirstat engine**: the time cost to look up an existing directory | |
4614 | ||
4615 | **dirdelete engine**: the time cost to delete a directory | |
4616 | ||
29092211 VF |
4617 | **lat** |
4618 | Total latency. Same names as slat and clat, this denotes the time from | |
4619 | when fio created the I/O unit to completion of the I/O operation. | |
13ddd98b | 4620 | It is the sum of submission and completion latency. |
29092211 | 4621 | |
f80dba8d | 4622 | **bw** |
f6f80750 VF |
4623 | Bandwidth statistics based on measurements from discrete |
4624 | intervals. Fio continuously monitors bytes transferred and I/O | |
4625 | operations completed. By default fio calculates bandwidth in | |
4626 | each half-second interval (see :option:`bwavgtime`) and reports | |
4627 | descriptive statistics for the measurements here. Same names as | |
4628 | the xlat stats, but also includes the number of samples taken | |
4629 | (**samples**) and an approximate percentage of total aggregate | |
4630 | bandwidth this thread received in its group (**per**). This | |
4631 | last value is only really useful if the threads in this group | |
4632 | are on the same disk, since they are then competing for disk | |
4633 | access. | |
36214730 | 4634 | |
2cd920c1 | 4635 | For file and directory operation engines, **bw** is meaningless. |
4636 | ||
36214730 | 4637 | **iops** |
f6f80750 VF |
4638 | IOPS statistics based on measurements from discrete intervals. |
4639 | For details see the description for bw above. See | |
4640 | :option:`iopsavgtime` to control the duration of the intervals. | |
4641 | Same values reported here as for bw except for percentage. | |
f80dba8d | 4642 | |
2cd920c1 | 4643 | For file and directory operation engines, **iops** is the most |
4644 | fundamental index to denote the performance. | |
4645 | It means how many files or directories can be operated per second. | |
4646 | ||
4647 | **filecreate engine**:number of files can be created per second | |
4648 | ||
4649 | **filestat engine**: number of files can be looked up per second | |
4650 | ||
4651 | **filedelete engine**:number of files can be deleted per second | |
4652 | ||
4653 | **dircreate engine**: number of directories can be created per second | |
4654 | ||
4655 | **dirstat engine**: number of directories can be looked up per second | |
4656 | ||
4657 | **dirdelete engine**: number of directories can be deleted per second | |
4658 | ||
29092211 VF |
4659 | **lat (nsec/usec/msec)** |
4660 | The distribution of I/O completion latencies. This is the time from when | |
4661 | I/O leaves fio and when it gets completed. Unlike the separate | |
4662 | read/write/trim sections above, the data here and in the remaining | |
4663 | sections apply to all I/Os for the reporting group. 250=0.04% means that | |
4664 | 0.04% of the I/Os completed in under 250us. 500=64.11% means that 64.11% | |
4665 | of the I/Os required 250 to 499us for completion. | |
4666 | ||
f80dba8d MT |
4667 | **cpu** |
4668 | CPU usage. User and system time, along with the number of context | |
4669 | switches this thread went through, usage of system and user time, and | |
4670 | finally the number of major and minor page faults. The CPU utilization | |
4671 | numbers are averages for the jobs in that reporting group, while the | |
23a8e176 | 4672 | context and fault counters are summed. |
f80dba8d MT |
4673 | |
4674 | **IO depths** | |
a2140525 SW |
4675 | The distribution of I/O depths over the job lifetime. The numbers are |
4676 | divided into powers of 2 and each entry covers depths from that value | |
4677 | up to those that are lower than the next entry -- e.g., 16= covers | |
4678 | depths from 16 to 31. Note that the range covered by a depth | |
4679 | distribution entry can be different to the range covered by the | |
4680 | equivalent submit/complete distribution entry. | |
f80dba8d MT |
4681 | |
4682 | **IO submit** | |
4683 | How many pieces of I/O were submitting in a single submit call. Each | |
c60ebc45 | 4684 | entry denotes that amount and below, until the previous entry -- e.g., |
a2140525 SW |
4685 | 16=100% means that we submitted anywhere between 9 to 16 I/Os per submit |
4686 | call. Note that the range covered by a submit distribution entry can | |
4687 | be different to the range covered by the equivalent depth distribution | |
4688 | entry. | |
f80dba8d MT |
4689 | |
4690 | **IO complete** | |
4691 | Like the above submit number, but for completions instead. | |
4692 | ||
36214730 SW |
4693 | **IO issued rwt** |
4694 | The number of read/write/trim requests issued, and how many of them were | |
4695 | short or dropped. | |
f80dba8d | 4696 | |
29092211 | 4697 | **IO latency** |
ee21ebee | 4698 | These values are for :option:`latency_target` and related options. When |
29092211 VF |
4699 | these options are engaged, this section describes the I/O depth required |
4700 | to meet the specified latency target. | |
71bfa161 | 4701 | |
36214730 SW |
4702 | .. |
4703 | Example output was based on the following: | |
4704 | TZ=UTC fio --ioengine=null --iodepth=2 --size=100M --numjobs=2 \ | |
4705 | --rate_process=poisson --io_limit=32M --name=read --bs=128k \ | |
4706 | --rate=11M --name=write --rw=write --bs=2k --rate=700k | |
4707 | ||
71bfa161 | 4708 | After each client has been listed, the group statistics are printed. They |
f80dba8d | 4709 | will look like this:: |
71bfa161 | 4710 | |
f80dba8d | 4711 | Run status group 0 (all jobs): |
36214730 SW |
4712 | READ: bw=20.9MiB/s (21.9MB/s), 10.4MiB/s-10.8MiB/s (10.9MB/s-11.3MB/s), io=64.0MiB (67.1MB), run=2973-3069msec |
4713 | WRITE: bw=1231KiB/s (1261kB/s), 616KiB/s-621KiB/s (630kB/s-636kB/s), io=64.0MiB (67.1MB), run=52747-53223msec | |
71bfa161 | 4714 | |
36214730 | 4715 | For each data direction it prints: |
71bfa161 | 4716 | |
36214730 SW |
4717 | **bw** |
4718 | Aggregate bandwidth of threads in this group followed by the | |
4719 | minimum and maximum bandwidth of all the threads in this group. | |
4720 | Values outside of brackets are power-of-2 format and those | |
4721 | within are the equivalent value in a power-of-10 format. | |
f80dba8d | 4722 | **io** |
36214730 SW |
4723 | Aggregate I/O performed of all threads in this group. The |
4724 | format is the same as bw. | |
4725 | **run** | |
4726 | The smallest and longest runtimes of the threads in this group. | |
71bfa161 | 4727 | |
f50fbdda | 4728 | And finally, the disk statistics are printed. This is Linux specific. They will look like this:: |
71bfa161 | 4729 | |
f80dba8d | 4730 | Disk stats (read/write): |
75cbc26d | 4731 | sda: ios=16398/16511, sectors=32321/65472, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00% |
71bfa161 JA |
4732 | |
4733 | Each value is printed for both reads and writes, with reads first. The | |
4734 | numbers denote: | |
4735 | ||
f80dba8d | 4736 | **ios** |
c60ebc45 | 4737 | Number of I/Os performed by all groups. |
75cbc26d BVA |
4738 | **sectors** |
4739 | Amount of data transferred in units of 512 bytes for all groups. | |
f80dba8d | 4740 | **merge** |
007c7be9 | 4741 | Number of merges performed by the I/O scheduler. |
f80dba8d MT |
4742 | **ticks** |
4743 | Number of ticks we kept the disk busy. | |
36214730 | 4744 | **in_queue** |
f80dba8d MT |
4745 | Total time spent in the disk queue. |
4746 | **util** | |
4747 | The disk utilization. A value of 100% means we kept the disk | |
71bfa161 JA |
4748 | busy constantly, 50% would be a disk idling half of the time. |
4749 | ||
f80dba8d MT |
4750 | It is also possible to get fio to dump the current output while it is running, |
4751 | without terminating the job. To do that, send fio the **USR1** signal. You can | |
4752 | also get regularly timed dumps by using the :option:`--status-interval` | |
4753 | parameter, or by creating a file in :file:`/tmp` named | |
4754 | :file:`fio-dump-status`. If fio sees this file, it will unlink it and dump the | |
4755 | current output status. | |
8423bd11 | 4756 | |
71bfa161 | 4757 | |
f80dba8d MT |
4758 | Terse output |
4759 | ------------ | |
71bfa161 | 4760 | |
f80dba8d MT |
4761 | For scripted usage where you typically want to generate tables or graphs of the |
4762 | results, fio can output the results in a semicolon separated format. The format | |
4763 | is one long line of values, such as:: | |
71bfa161 | 4764 | |
f80dba8d MT |
4765 | 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% |
4766 | A description of this job goes here. | |
562c2d2f | 4767 | |
4e757af1 VF |
4768 | The job description (if provided) follows on a second line for terse v2. |
4769 | It appears on the same line for other terse versions. | |
71bfa161 | 4770 | |
a7f77fa6 SW |
4771 | To enable terse output, use the :option:`--minimal` or |
4772 | :option:`--output-format`\=terse command line options. The | |
f80dba8d MT |
4773 | first value is the version of the terse output format. If the output has to be |
4774 | changed for some reason, this number will be incremented by 1 to signify that | |
4775 | change. | |
6820cb3b | 4776 | |
a2c95580 | 4777 | Split up, the format is as follows (comments in brackets denote when a |
007c7be9 | 4778 | field was introduced or whether it's specific to some terse version): |
71bfa161 | 4779 | |
f80dba8d MT |
4780 | :: |
4781 | ||
f50fbdda | 4782 | terse version, fio version [v3], jobname, groupid, error |
f80dba8d MT |
4783 | |
4784 | READ status:: | |
4785 | ||
4786 | Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec) | |
4787 | Submission latency: min, max, mean, stdev (usec) | |
4788 | Completion latency: min, max, mean, stdev (usec) | |
4789 | Completion latency percentiles: 20 fields (see below) | |
4790 | Total latency: min, max, mean, stdev (usec) | |
f50fbdda TK |
4791 | Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5] |
4792 | IOPS [v5]: min, max, mean, stdev, number of samples | |
f80dba8d MT |
4793 | |
4794 | WRITE status: | |
4795 | ||
4796 | :: | |
4797 | ||
4798 | Total IO (KiB), bandwidth (KiB/sec), IOPS, runtime (msec) | |
4799 | Submission latency: min, max, mean, stdev (usec) | |
247823cc | 4800 | Completion latency: min, max, mean, stdev (usec) |
f80dba8d MT |
4801 | Completion latency percentiles: 20 fields (see below) |
4802 | Total latency: min, max, mean, stdev (usec) | |
f50fbdda TK |
4803 | Bw (KiB/s): min, max, aggregate percentage of total, mean, stdev, number of samples [v5] |
4804 | IOPS [v5]: min, max, mean, stdev, number of samples | |
a2c95580 AH |
4805 | |
4806 | TRIM status [all but version 3]: | |
4807 | ||
f50fbdda | 4808 | Fields are similar to READ/WRITE status. |
f80dba8d MT |
4809 | |
4810 | CPU usage:: | |
4811 | ||
4812 | user, system, context switches, major faults, minor faults | |
4813 | ||
4814 | I/O depths:: | |
4815 | ||
4816 | <=1, 2, 4, 8, 16, 32, >=64 | |
4817 | ||
4818 | I/O latencies microseconds:: | |
4819 | ||
4820 | <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000 | |
4821 | ||
4822 | I/O latencies milliseconds:: | |
4823 | ||
4824 | <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000 | |
4825 | ||
a2c95580 | 4826 | Disk utilization [v3]:: |
f80dba8d | 4827 | |
f50fbdda TK |
4828 | disk name, read ios, write ios, read merges, write merges, read ticks, write ticks, |
4829 | time spent in queue, disk utilization percentage | |
f80dba8d MT |
4830 | |
4831 | Additional Info (dependent on continue_on_error, default off):: | |
4832 | ||
4833 | total # errors, first error code | |
4834 | ||
4835 | Additional Info (dependent on description being set):: | |
4836 | ||
4837 | Text description | |
4838 | ||
4839 | Completion latency percentiles can be a grouping of up to 20 sets, so for the | |
4840 | terse output fio writes all of them. Each field will look like this:: | |
1db92cb6 | 4841 | |
f50fbdda | 4842 | 1.00%=6112 |
1db92cb6 | 4843 | |
f80dba8d | 4844 | which is the Xth percentile, and the `usec` latency associated with it. |
1db92cb6 | 4845 | |
f50fbdda | 4846 | For `Disk utilization`, all disks used by fio are shown. So for each disk there |
f80dba8d | 4847 | will be a disk utilization section. |
f2f788dd | 4848 | |
2fc26c3d | 4849 | Below is a single line containing short names for each of the fields in the |
2831be97 | 4850 | minimal output v3, separated by semicolons:: |
2fc26c3d | 4851 | |
f95689d3 | 4852 | terse_version_3;fio_version;jobname;groupid;error;read_kb;read_bandwidth_kb;read_iops;read_runtime_ms;read_slat_min_us;read_slat_max_us;read_slat_mean_us;read_slat_dev_us;read_clat_min_us;read_clat_max_us;read_clat_mean_us;read_clat_dev_us;read_clat_pct01;read_clat_pct02;read_clat_pct03;read_clat_pct04;read_clat_pct05;read_clat_pct06;read_clat_pct07;read_clat_pct08;read_clat_pct09;read_clat_pct10;read_clat_pct11;read_clat_pct12;read_clat_pct13;read_clat_pct14;read_clat_pct15;read_clat_pct16;read_clat_pct17;read_clat_pct18;read_clat_pct19;read_clat_pct20;read_tlat_min_us;read_lat_max_us;read_lat_mean_us;read_lat_dev_us;read_bw_min_kb;read_bw_max_kb;read_bw_agg_pct;read_bw_mean_kb;read_bw_dev_kb;write_kb;write_bandwidth_kb;write_iops;write_runtime_ms;write_slat_min_us;write_slat_max_us;write_slat_mean_us;write_slat_dev_us;write_clat_min_us;write_clat_max_us;write_clat_mean_us;write_clat_dev_us;write_clat_pct01;write_clat_pct02;write_clat_pct03;write_clat_pct04;write_clat_pct05;write_clat_pct06;write_clat_pct07;write_clat_pct08;write_clat_pct09;write_clat_pct10;write_clat_pct11;write_clat_pct12;write_clat_pct13;write_clat_pct14;write_clat_pct15;write_clat_pct16;write_clat_pct17;write_clat_pct18;write_clat_pct19;write_clat_pct20;write_tlat_min_us;write_lat_max_us;write_lat_mean_us;write_lat_dev_us;write_bw_min_kb;write_bw_max_kb;write_bw_agg_pct;write_bw_mean_kb;write_bw_dev_kb;cpu_user;cpu_sys;cpu_csw;cpu_mjf;cpu_minf;iodepth_1;iodepth_2;iodepth_4;iodepth_8;iodepth_16;iodepth_32;iodepth_64;lat_2us;lat_4us;lat_10us;lat_20us;lat_50us;lat_100us;lat_250us;lat_500us;lat_750us;lat_1000us;lat_2ms;lat_4ms;lat_10ms;lat_20ms;lat_50ms;lat_100ms;lat_250ms;lat_500ms;lat_750ms;lat_1000ms;lat_2000ms;lat_over_2000ms;disk_name;disk_read_iops;disk_write_iops;disk_read_merges;disk_write_merges;disk_read_ticks;write_ticks;disk_queue_time;disk_util |
2fc26c3d | 4853 | |
4e757af1 VF |
4854 | In client/server mode terse output differs from what appears when jobs are run |
4855 | locally. Disk utilization data is omitted from the standard terse output and | |
4856 | for v3 and later appears on its own separate line at the end of each terse | |
4857 | reporting cycle. | |
4858 | ||
25c8b9d7 | 4859 | |
44c82dba VF |
4860 | JSON output |
4861 | ------------ | |
4862 | ||
4863 | The `json` output format is intended to be both human readable and convenient | |
4864 | for automated parsing. For the most part its sections mirror those of the | |
4865 | `normal` output. The `runtime` value is reported in msec and the `bw` value is | |
4866 | reported in 1024 bytes per second units. | |
4867 | ||
4868 | ||
d29c4a91 VF |
4869 | JSON+ output |
4870 | ------------ | |
4871 | ||
4872 | The `json+` output format is identical to the `json` output format except that it | |
4873 | adds a full dump of the completion latency bins. Each `bins` object contains a | |
4874 | set of (key, value) pairs where keys are latency durations and values count how | |
4875 | many I/Os had completion latencies of the corresponding duration. For example, | |
4876 | consider: | |
4877 | ||
4878 | "bins" : { "87552" : 1, "89600" : 1, "94720" : 1, "96768" : 1, "97792" : 1, "99840" : 1, "100864" : 2, "103936" : 6, "104960" : 534, "105984" : 5995, "107008" : 7529, ... } | |
4879 | ||
4880 | This data indicates that one I/O required 87,552ns to complete, two I/Os required | |
4881 | 100,864ns to complete, and 7529 I/Os required 107,008ns to complete. | |
4882 | ||
4883 | Also included with fio is a Python script `fio_jsonplus_clat2csv` that takes | |
4884 | json+ output and generates CSV-formatted latency data suitable for plotting. | |
4885 | ||
4886 | The latency durations actually represent the midpoints of latency intervals. | |
f50fbdda | 4887 | For details refer to :file:`stat.h`. |
d29c4a91 VF |
4888 | |
4889 | ||
f80dba8d MT |
4890 | Trace file format |
4891 | ----------------- | |
4892 | ||
4893 | There are two trace file format that you can encounter. The older (v1) format is | |
4894 | unsupported since version 1.20-rc3 (March 2008). It will still be described | |
25c8b9d7 PD |
4895 | below in case that you get an old trace and want to understand it. |
4896 | ||
4897 | In any case the trace is a simple text file with a single action per line. | |
4898 | ||
4899 | ||
f80dba8d MT |
4900 | Trace file format v1 |
4901 | ~~~~~~~~~~~~~~~~~~~~ | |
4902 | ||
4903 | Each line represents a single I/O action in the following format:: | |
4904 | ||
4905 | rw, offset, length | |
25c8b9d7 | 4906 | |
f50fbdda | 4907 | where `rw=0/1` for read/write, and the `offset` and `length` entries being in bytes. |
25c8b9d7 | 4908 | |
22413915 | 4909 | This format is not supported in fio versions >= 1.20-rc3. |
25c8b9d7 | 4910 | |
25c8b9d7 | 4911 | |
f80dba8d MT |
4912 | Trace file format v2 |
4913 | ~~~~~~~~~~~~~~~~~~~~ | |
25c8b9d7 | 4914 | |
f80dba8d | 4915 | The second version of the trace file format was added in fio version 1.17. It |
12efafa3 | 4916 | allows one to access more than one file per trace and has a bigger set of possible |
f80dba8d | 4917 | file actions. |
25c8b9d7 | 4918 | |
f80dba8d | 4919 | The first line of the trace file has to be:: |
25c8b9d7 | 4920 | |
f80dba8d | 4921 | fio version 2 iolog |
25c8b9d7 PD |
4922 | |
4923 | Following this can be lines in two different formats, which are described below. | |
4924 | ||
f80dba8d | 4925 | The file management format:: |
25c8b9d7 | 4926 | |
f80dba8d | 4927 | filename action |
25c8b9d7 | 4928 | |
f50fbdda | 4929 | The `filename` is given as an absolute path. The `action` can be one of these: |
25c8b9d7 | 4930 | |
f80dba8d | 4931 | **add** |
f50fbdda | 4932 | Add the given `filename` to the trace. |
f80dba8d | 4933 | **open** |
f50fbdda | 4934 | Open the file with the given `filename`. The `filename` has to have |
f80dba8d MT |
4935 | been added with the **add** action before. |
4936 | **close** | |
f50fbdda | 4937 | Close the file with the given `filename`. The file has to have been |
f80dba8d MT |
4938 | opened before. |
4939 | ||
4940 | ||
4941 | The file I/O action format:: | |
4942 | ||
4943 | filename action offset length | |
4944 | ||
4945 | The `filename` is given as an absolute path, and has to have been added and | |
4946 | opened before it can be used with this format. The `offset` and `length` are | |
4947 | given in bytes. The `action` can be one of these: | |
4948 | ||
4949 | **wait** | |
4950 | Wait for `offset` microseconds. Everything below 100 is discarded. | |
5c2c0db4 MG |
4951 | The time is relative to the previous `wait` statement. Note that |
4952 | action `wait` is not allowed as of version 3, as the same behavior | |
4953 | can be achieved using timestamps. | |
f80dba8d MT |
4954 | **read** |
4955 | Read `length` bytes beginning from `offset`. | |
4956 | **write** | |
4957 | Write `length` bytes beginning from `offset`. | |
4958 | **sync** | |
4959 | :manpage:`fsync(2)` the file. | |
4960 | **datasync** | |
4961 | :manpage:`fdatasync(2)` the file. | |
4962 | **trim** | |
4963 | Trim the given file from the given `offset` for `length` bytes. | |
4964 | ||
b9921d1a | 4965 | |
5c2c0db4 MG |
4966 | Trace file format v3 |
4967 | ~~~~~~~~~~~~~~~~~~~~ | |
4968 | ||
4969 | The third version of the trace file format was added in fio version 3.31. It | |
4970 | forces each action to have a timestamp associated with it. | |
4971 | ||
4972 | The first line of the trace file has to be:: | |
4973 | ||
4974 | fio version 3 iolog | |
4975 | ||
4976 | Following this can be lines in two different formats, which are described below. | |
4977 | ||
4978 | The file management format:: | |
4979 | ||
4980 | timestamp filename action | |
4981 | ||
4982 | The file I/O action format:: | |
4983 | ||
4984 | timestamp filename action offset length | |
4985 | ||
4986 | The `timestamp` is relative to the beginning of the run (ie starts at 0). The | |
4987 | `filename`, `action`, `offset` and `length` are identical to version 2, except | |
4988 | that version 3 does not allow the `wait` action. | |
4989 | ||
4990 | ||
b9921d1a DZ |
4991 | I/O Replay - Merging Traces |
4992 | --------------------------- | |
4993 | ||
4994 | Colocation is a common practice used to get the most out of a machine. | |
4995 | Knowing which workloads play nicely with each other and which ones don't is | |
4996 | a much harder task. While fio can replay workloads concurrently via multiple | |
4997 | jobs, it leaves some variability up to the scheduler making results harder to | |
4998 | reproduce. Merging is a way to make the order of events consistent. | |
4999 | ||
5000 | Merging is integrated into I/O replay and done when a | |
5001 | :option:`merge_blktrace_file` is specified. The list of files passed to | |
5002 | :option:`read_iolog` go through the merge process and output a single file | |
5003 | stored to the specified file. The output file is passed on as if it were the | |
5004 | only file passed to :option:`read_iolog`. An example would look like:: | |
5005 | ||
5006 | $ fio --read_iolog="<file1>:<file2>" --merge_blktrace_file="<output_file>" | |
5007 | ||
5008 | Creating only the merged file can be done by passing the command line argument | |
d443e3af | 5009 | :option:`--merge-blktrace-only`. |
b9921d1a | 5010 | |
87a48ada DZ |
5011 | Scaling traces can be done to see the relative impact of any particular trace |
5012 | being slowed down or sped up. :option:`merge_blktrace_scalars` takes in a colon | |
5013 | separated list of percentage scalars. It is index paired with the files passed | |
5014 | to :option:`read_iolog`. | |
5015 | ||
55bfd8c8 DZ |
5016 | With scaling, it may be desirable to match the running time of all traces. |
5017 | This can be done with :option:`merge_blktrace_iters`. It is index paired with | |
5018 | :option:`read_iolog` just like :option:`merge_blktrace_scalars`. | |
5019 | ||
5020 | In an example, given two traces, A and B, each 60s long. If we want to see | |
5021 | the impact of trace A issuing IOs twice as fast and repeat trace A over the | |
5022 | runtime of trace B, the following can be done:: | |
5023 | ||
5024 | $ fio --read_iolog="<trace_a>:"<trace_b>" --merge_blktrace_file"<output_file>" --merge_blktrace_scalars="50:100" --merge_blktrace_iters="2:1" | |
5025 | ||
5026 | This runs trace A at 2x the speed twice for approximately the same runtime as | |
5027 | a single run of trace B. | |
5028 | ||
b9921d1a | 5029 | |
f80dba8d MT |
5030 | CPU idleness profiling |
5031 | ---------------------- | |
5032 | ||
5033 | In some cases, we want to understand CPU overhead in a test. For example, we | |
5034 | test patches for the specific goodness of whether they reduce CPU usage. | |
5035 | Fio implements a balloon approach to create a thread per CPU that runs at idle | |
5036 | priority, meaning that it only runs when nobody else needs the cpu. | |
5037 | By measuring the amount of work completed by the thread, idleness of each CPU | |
5038 | can be derived accordingly. | |
5039 | ||
5040 | An unit work is defined as touching a full page of unsigned characters. Mean and | |
5041 | standard deviation of time to complete an unit work is reported in "unit work" | |
5042 | section. Options can be chosen to report detailed percpu idleness or overall | |
5043 | system idleness by aggregating percpu stats. | |
5044 | ||
5045 | ||
5046 | Verification and triggers | |
5047 | ------------------------- | |
5048 | ||
5049 | Fio is usually run in one of two ways, when data verification is done. The first | |
5050 | is a normal write job of some sort with verify enabled. When the write phase has | |
5051 | completed, fio switches to reads and verifies everything it wrote. The second | |
5052 | model is running just the write phase, and then later on running the same job | |
5053 | (but with reads instead of writes) to repeat the same I/O patterns and verify | |
5054 | the contents. Both of these methods depend on the write phase being completed, | |
5055 | as fio otherwise has no idea how much data was written. | |
5056 | ||
5057 | With verification triggers, fio supports dumping the current write state to | |
5058 | local files. Then a subsequent read verify workload can load this state and know | |
5059 | exactly where to stop. This is useful for testing cases where power is cut to a | |
5060 | server in a managed fashion, for instance. | |
99b9a85a JA |
5061 | |
5062 | A verification trigger consists of two things: | |
5063 | ||
f80dba8d MT |
5064 | 1) Storing the write state of each job. |
5065 | 2) Executing a trigger command. | |
99b9a85a | 5066 | |
f80dba8d MT |
5067 | The write state is relatively small, on the order of hundreds of bytes to single |
5068 | kilobytes. It contains information on the number of completions done, the last X | |
5069 | completions, etc. | |
99b9a85a | 5070 | |
f80dba8d MT |
5071 | A trigger is invoked either through creation ('touch') of a specified file in |
5072 | the system, or through a timeout setting. If fio is run with | |
9207a0cb | 5073 | :option:`--trigger-file`\= :file:`/tmp/trigger-file`, then it will continually |
f80dba8d MT |
5074 | check for the existence of :file:`/tmp/trigger-file`. When it sees this file, it |
5075 | will fire off the trigger (thus saving state, and executing the trigger | |
99b9a85a JA |
5076 | command). |
5077 | ||
f80dba8d MT |
5078 | For client/server runs, there's both a local and remote trigger. If fio is |
5079 | running as a server backend, it will send the job states back to the client for | |
5080 | safe storage, then execute the remote trigger, if specified. If a local trigger | |
5081 | is specified, the server will still send back the write state, but the client | |
5082 | will then execute the trigger. | |
99b9a85a | 5083 | |
f80dba8d MT |
5084 | Verification trigger example |
5085 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
99b9a85a | 5086 | |
f50fbdda TK |
5087 | Let's say we want to run a powercut test on the remote Linux machine 'server'. |
5088 | Our write workload is in :file:`write-test.fio`. We want to cut power to 'server' at | |
f80dba8d MT |
5089 | some point during the run, and we'll run this test from the safety or our local |
5090 | machine, 'localbox'. On the server, we'll start the fio backend normally:: | |
99b9a85a | 5091 | |
f80dba8d | 5092 | server# fio --server |
99b9a85a | 5093 | |
f80dba8d | 5094 | and on the client, we'll fire off the workload:: |
99b9a85a | 5095 | |
f80dba8d | 5096 | localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger-remote="bash -c \"echo b > /proc/sysrq-triger\"" |
99b9a85a | 5097 | |
f80dba8d | 5098 | We set :file:`/tmp/my-trigger` as the trigger file, and we tell fio to execute:: |
99b9a85a | 5099 | |
f80dba8d | 5100 | echo b > /proc/sysrq-trigger |
99b9a85a | 5101 | |
f80dba8d MT |
5102 | on the server once it has received the trigger and sent us the write state. This |
5103 | will work, but it's not **really** cutting power to the server, it's merely | |
5104 | abruptly rebooting it. If we have a remote way of cutting power to the server | |
5105 | through IPMI or similar, we could do that through a local trigger command | |
4502cb42 | 5106 | instead. Let's assume we have a script that does IPMI reboot of a given hostname, |
f80dba8d MT |
5107 | ipmi-reboot. On localbox, we could then have run fio with a local trigger |
5108 | instead:: | |
99b9a85a | 5109 | |
f80dba8d | 5110 | localbox$ fio --client=server --trigger-file=/tmp/my-trigger --trigger="ipmi-reboot server" |
99b9a85a | 5111 | |
f80dba8d MT |
5112 | For this case, fio would wait for the server to send us the write state, then |
5113 | execute ``ipmi-reboot server`` when that happened. | |
5114 | ||
5115 | Loading verify state | |
5116 | ~~~~~~~~~~~~~~~~~~~~ | |
5117 | ||
4502cb42 | 5118 | To load stored write state, a read verification job file must contain the |
f80dba8d | 5119 | :option:`verify_state_load` option. If that is set, fio will load the previously |
99b9a85a | 5120 | stored state. For a local fio run this is done by loading the files directly, |
f80dba8d MT |
5121 | and on a client/server run, the server backend will ask the client to send the |
5122 | files over and load them from there. | |
a3ae5b05 JA |
5123 | |
5124 | ||
f80dba8d MT |
5125 | Log File Formats |
5126 | ---------------- | |
a3ae5b05 JA |
5127 | |
5128 | Fio supports a variety of log file formats, for logging latencies, bandwidth, | |
5129 | and IOPS. The logs share a common format, which looks like this: | |
5130 | ||
5a83478f | 5131 | *time* (`msec`), *value*, *data direction*, *block size* (`bytes`), |
1a953d97 | 5132 | *offset* (`bytes`), *command priority* |
a3ae5b05 | 5133 | |
5a83478f | 5134 | *Time* for the log entry is always in milliseconds. The *value* logged depends |
a3ae5b05 JA |
5135 | on the type of log, it will be one of the following: |
5136 | ||
f80dba8d | 5137 | **Latency log** |
168bb587 | 5138 | Value is latency in nsecs |
f80dba8d MT |
5139 | **Bandwidth log** |
5140 | Value is in KiB/sec | |
5141 | **IOPS log** | |
5142 | Value is IOPS | |
5143 | ||
5144 | *Data direction* is one of the following: | |
5145 | ||
5146 | **0** | |
5147 | I/O is a READ | |
5148 | **1** | |
5149 | I/O is a WRITE | |
5150 | **2** | |
5151 | I/O is a TRIM | |
5152 | ||
15417073 SW |
5153 | The entry's *block size* is always in bytes. The *offset* is the position in bytes |
5154 | from the start of the file for that particular I/O. The logging of the offset can be | |
5a83478f | 5155 | toggled with :option:`log_offset`. |
f80dba8d | 5156 | |
1a953d97 PC |
5157 | *Command priority* is 0 for normal priority and 1 for high priority. This is controlled |
5158 | by the ioengine specific :option:`cmdprio_percentage`. | |
5159 | ||
15417073 | 5160 | Fio defaults to logging every individual I/O but when windowed logging is set |
065212b3 AK |
5161 | through :option:`log_avg_msec`, either the average (by default), the maximum |
5162 | (:option:`log_window_value` is set to max) *value* seen over the specified period | |
5163 | of time, or both the average *value* and maximum *value1* (:option:`log_window_value` | |
5164 | is set to both) is recorded. The log file format when both the values are reported | |
5165 | takes this form: | |
5166 | ||
5167 | *time* (`msec`), *value*, *value1*, *data direction*, *block size* (`bytes`), | |
5168 | *offset* (`bytes`), *command priority* | |
5169 | ||
5170 | ||
5171 | Each *data direction* seen within the window period will aggregate its values in a | |
5172 | separate row. Further, when using windowed logging the *block size* and *offset* | |
5173 | entries will always contain 0. | |
f80dba8d | 5174 | |
4e757af1 | 5175 | |
b8f7e412 | 5176 | Client/Server |
f80dba8d MT |
5177 | ------------- |
5178 | ||
5179 | Normally fio is invoked as a stand-alone application on the machine where the | |
6cf30ac0 SW |
5180 | I/O workload should be generated. However, the backend and frontend of fio can |
5181 | be run separately i.e., the fio server can generate an I/O workload on the "Device | |
5182 | Under Test" while being controlled by a client on another machine. | |
f80dba8d MT |
5183 | |
5184 | Start the server on the machine which has access to the storage DUT:: | |
5185 | ||
f50fbdda | 5186 | $ fio --server=args |
f80dba8d | 5187 | |
dbb257bb | 5188 | where `args` defines what fio listens to. The arguments are of the form |
f80dba8d MT |
5189 | ``type,hostname`` or ``IP,port``. *type* is either ``ip`` (or ip4) for TCP/IP |
5190 | v4, ``ip6`` for TCP/IP v6, or ``sock`` for a local unix domain socket. | |
5191 | *hostname* is either a hostname or IP address, and *port* is the port to listen | |
5192 | to (only valid for TCP/IP, not a local socket). Some examples: | |
5193 | ||
5194 | 1) ``fio --server`` | |
5195 | ||
5196 | Start a fio server, listening on all interfaces on the default port (8765). | |
5197 | ||
5198 | 2) ``fio --server=ip:hostname,4444`` | |
5199 | ||
5200 | Start a fio server, listening on IP belonging to hostname and on port 4444. | |
5201 | ||
5202 | 3) ``fio --server=ip6:::1,4444`` | |
5203 | ||
5204 | Start a fio server, listening on IPv6 localhost ::1 and on port 4444. | |
5205 | ||
5206 | 4) ``fio --server=,4444`` | |
5207 | ||
5208 | Start a fio server, listening on all interfaces on port 4444. | |
5209 | ||
5210 | 5) ``fio --server=1.2.3.4`` | |
5211 | ||
5212 | Start a fio server, listening on IP 1.2.3.4 on the default port. | |
5213 | ||
5214 | 6) ``fio --server=sock:/tmp/fio.sock`` | |
5215 | ||
dbb257bb | 5216 | Start a fio server, listening on the local socket :file:`/tmp/fio.sock`. |
f80dba8d MT |
5217 | |
5218 | Once a server is running, a "client" can connect to the fio server with:: | |
5219 | ||
5220 | fio <local-args> --client=<server> <remote-args> <job file(s)> | |
5221 | ||
5222 | where `local-args` are arguments for the client where it is running, `server` | |
5223 | is the connect string, and `remote-args` and `job file(s)` are sent to the | |
5224 | server. The `server` string follows the same format as it does on the server | |
5225 | side, to allow IP/hostname/socket and port strings. | |
5226 | ||
702e3f34 | 5227 | Note that all job options must be defined in job files when running fio as a |
5228 | client. Any job options specified in `remote-args` will be ignored. | |
5229 | ||
f80dba8d MT |
5230 | Fio can connect to multiple servers this way:: |
5231 | ||
5232 | fio --client=<server1> <job file(s)> --client=<server2> <job file(s)> | |
5233 | ||
5234 | If the job file is located on the fio server, then you can tell the server to | |
5235 | load a local file as well. This is done by using :option:`--remote-config` :: | |
5236 | ||
5237 | fio --client=server --remote-config /path/to/file.fio | |
5238 | ||
5239 | Then fio will open this local (to the server) job file instead of being passed | |
5240 | one from the client. | |
5241 | ||
5242 | If you have many servers (example: 100 VMs/containers), you can input a pathname | |
5243 | of a file containing host IPs/names as the parameter value for the | |
5244 | :option:`--client` option. For example, here is an example :file:`host.list` | |
5245 | file containing 2 hostnames:: | |
5246 | ||
5247 | host1.your.dns.domain | |
5248 | host2.your.dns.domain | |
5249 | ||
5250 | The fio command would then be:: | |
a3ae5b05 | 5251 | |
f80dba8d | 5252 | fio --client=host.list <job file(s)> |
a3ae5b05 | 5253 | |
f80dba8d MT |
5254 | In this mode, you cannot input server-specific parameters or job files -- all |
5255 | servers receive the same job file. | |
a3ae5b05 | 5256 | |
f80dba8d MT |
5257 | In order to let ``fio --client`` runs use a shared filesystem from multiple |
5258 | hosts, ``fio --client`` now prepends the IP address of the server to the | |
4502cb42 | 5259 | filename. For example, if fio is using the directory :file:`/mnt/nfs/fio` and is |
f80dba8d MT |
5260 | writing filename :file:`fileio.tmp`, with a :option:`--client` `hostfile` |
5261 | containing two hostnames ``h1`` and ``h2`` with IP addresses 192.168.10.120 and | |
5262 | 192.168.10.121, then fio will create two files:: | |
a3ae5b05 | 5263 | |
f80dba8d MT |
5264 | /mnt/nfs/fio/192.168.10.120.fileio.tmp |
5265 | /mnt/nfs/fio/192.168.10.121.fileio.tmp | |
4e757af1 VF |
5266 | |
5267 | Terse output in client/server mode will differ slightly from what is produced | |
5268 | when fio is run in stand-alone mode. See the terse output section for details. |