Commit | Line | Data |
---|---|---|
d60e92d1 AC |
1 | .TH fio 1 "September 2007" "User Manual" |
2 | .SH NAME | |
3 | fio \- flexible I/O tester | |
4 | .SH SYNOPSIS | |
5 | .B fio | |
6 | [\fIoptions\fR] [\fIjobfile\fR]... | |
7 | .SH DESCRIPTION | |
8 | .B fio | |
9 | is a tool that will spawn a number of threads or processes doing a | |
10 | particular type of I/O action as specified by the user. | |
11 | The typical use of fio is to write a job file matching the I/O load | |
12 | one wants to simulate. | |
13 | .SH OPTIONS | |
14 | .TP | |
15 | .BI \-\-output \fR=\fPfilename | |
16 | Write output to \fIfilename\fR. | |
17 | .TP | |
18 | .BI \-\-timeout \fR=\fPtimeout | |
19 | Limit run time to \fItimeout\fR seconds. | |
20 | .TP | |
21 | .B \-\-latency\-log | |
22 | Generate per-job latency logs. | |
23 | .TP | |
24 | .B \-\-bandwidth\-log | |
25 | Generate per-job bandwidth logs. | |
26 | .TP | |
27 | .B \-\-minimal | |
d1429b5c | 28 | Print statistics in a terse, semicolon-delimited format. |
d60e92d1 AC |
29 | .TP |
30 | .BI \-\-showcmd \fR=\fPjobfile | |
31 | Convert \fIjobfile\fR to a set of command-line options. | |
32 | .TP | |
33 | .B \-\-readonly | |
34 | Enable read-only safety checks. | |
35 | .TP | |
36 | .BI \-\-eta \fR=\fPwhen | |
37 | Specifies when real-time ETA estimate should be printed. \fIwhen\fR may | |
38 | be one of `always', `never' or `auto'. | |
39 | .TP | |
c0a5d35e AC |
40 | .BI \-\-section \fR=\fPsec |
41 | Only run section \fIsec\fR from job file. | |
42 | .TP | |
d60e92d1 AC |
43 | .BI \-\-cmdhelp \fR=\fPcommand |
44 | Print help information for \fIcommand\fR. May be `all' for all commands. | |
45 | .TP | |
9183788d JA |
46 | .BI \-\-debug \fR=\fPtype |
47 | Enable verbose tracing of various fio actions. May be `all' for all types | |
5982a925 | 48 | or individual types separated by a comma (eg \-\-debug=io,file). `help' will |
9183788d JA |
49 | list all available tracing options. |
50 | .TP | |
d60e92d1 AC |
51 | .B \-\-help |
52 | Display usage information and exit. | |
53 | .TP | |
54 | .B \-\-version | |
55 | Display version information and exit. | |
56 | .SH "JOB FILE FORMAT" | |
57 | Job files are in `ini' format. They consist of one or more | |
58 | job definitions, which begin with a job name in square brackets and | |
59 | extend to the next job name. The job name can be any ASCII string | |
60 | except `global', which has a special meaning. Following the job name is | |
61 | a sequence of zero or more parameters, one per line, that define the | |
62 | behavior of the job. Any line starting with a `;' or `#' character is | |
d1429b5c | 63 | considered a comment and ignored. |
d9956b64 AC |
64 | .P |
65 | If \fIjobfile\fR is specified as `-', the job file will be read from | |
66 | standard input. | |
d60e92d1 AC |
67 | .SS "Global Section" |
68 | The global section contains default parameters for jobs specified in the | |
69 | job file. A job is only affected by global sections residing above it, | |
70 | and there may be any number of global sections. Specific job definitions | |
71 | may override any parameter set in global sections. | |
72 | .SH "JOB PARAMETERS" | |
73 | .SS Types | |
74 | Some parameters may take arguments of a specific type. The types used are: | |
75 | .TP | |
76 | .I str | |
77 | String: a sequence of alphanumeric characters. | |
78 | .TP | |
79 | .I int | |
d60e92d1 | 80 | SI integer: a whole number, possibly containing a suffix denoting the base unit |
b09da8fa JA |
81 | of the value. Accepted suffixes are `k', 'M', 'G', 'T', and 'P', denoting |
82 | kilo (1024), mega (1024^2), giga (1024^3), tera (1024^4), and peta (1024^5) | |
83 | respectively. The suffix is not case sensitive. If prefixed with '0x', the | |
5982a925 MS |
84 | value is assumed to be base 16 (hexadecimal). A suffix may include a trailing 'b', |
85 | for instance 'kb' is identical to 'k'. You can specify a base 10 value | |
57fc29fa JA |
86 | by using 'KiB', 'MiB', 'GiB', etc. This is useful for disk drives where |
87 | values are often given in base 10 values. Specifying '30GiB' will get you | |
88 | 30*1000^3 bytes. | |
d60e92d1 AC |
89 | .TP |
90 | .I bool | |
91 | Boolean: a true or false value. `0' denotes false, `1' denotes true. | |
92 | .TP | |
93 | .I irange | |
94 | Integer range: a range of integers specified in the format | |
d1429b5c AC |
95 | \fIlower\fR:\fIupper\fR or \fIlower\fR\-\fIupper\fR. \fIlower\fR and |
96 | \fIupper\fR may contain a suffix as described above. If an option allows two | |
97 | sets of ranges, they are separated with a `,' or `/' character. For example: | |
98 | `8\-8k/8M\-4G'. | |
83349190 YH |
99 | .TP |
100 | .I float_list | |
101 | List of floating numbers: A list of floating numbers, separated by | |
102 | a ':' charcater. | |
d60e92d1 AC |
103 | .SS "Parameter List" |
104 | .TP | |
105 | .BI name \fR=\fPstr | |
d9956b64 | 106 | May be used to override the job name. On the command line, this parameter |
d60e92d1 AC |
107 | has the special purpose of signalling the start of a new job. |
108 | .TP | |
109 | .BI description \fR=\fPstr | |
110 | Human-readable description of the job. It is printed when the job is run, but | |
111 | otherwise has no special purpose. | |
112 | .TP | |
113 | .BI directory \fR=\fPstr | |
114 | Prefix filenames with this directory. Used to place files in a location other | |
115 | than `./'. | |
116 | .TP | |
117 | .BI filename \fR=\fPstr | |
118 | .B fio | |
119 | normally makes up a file name based on the job name, thread number, and file | |
d1429b5c AC |
120 | number. If you want to share files between threads in a job or several jobs, |
121 | specify a \fIfilename\fR for each of them to override the default. If the I/O | |
122 | engine used is `net', \fIfilename\fR is the host and port to connect to in the | |
123 | format \fIhost\fR/\fIport\fR. If the I/O engine is file-based, you can specify | |
124 | a number of files by separating the names with a `:' character. `\-' is a | |
125 | reserved name, meaning stdin or stdout, depending on the read/write direction | |
126 | set. | |
d60e92d1 | 127 | .TP |
3ce9dcaf JA |
128 | .BI lockfile \fR=\fPstr |
129 | Fio defaults to not locking any files before it does IO to them. If a file or | |
130 | file descriptor is shared, fio can serialize IO to that file to make the end | |
131 | result consistent. This is usual for emulating real workloads that share files. | |
132 | The lock modes are: | |
133 | .RS | |
134 | .RS | |
135 | .TP | |
136 | .B none | |
137 | No locking. This is the default. | |
138 | .TP | |
139 | .B exclusive | |
140 | Only one thread or process may do IO at the time, excluding all others. | |
141 | .TP | |
142 | .B readwrite | |
143 | Read-write locking on the file. Many readers may access the file at the same | |
144 | time, but writes get exclusive access. | |
145 | .RE | |
146 | .P | |
147 | The option may be post-fixed with a lock batch number. If set, then each | |
148 | thread/process may do that amount of IOs to the file before giving up the lock. | |
149 | Since lock acquisition is expensive, batching the lock/unlocks will speed up IO. | |
150 | .RE | |
151 | .P | |
d60e92d1 AC |
152 | .BI opendir \fR=\fPstr |
153 | Recursively open any files below directory \fIstr\fR. | |
154 | .TP | |
155 | .BI readwrite \fR=\fPstr "\fR,\fP rw" \fR=\fPstr | |
156 | Type of I/O pattern. Accepted values are: | |
157 | .RS | |
158 | .RS | |
159 | .TP | |
160 | .B read | |
d1429b5c | 161 | Sequential reads. |
d60e92d1 AC |
162 | .TP |
163 | .B write | |
d1429b5c | 164 | Sequential writes. |
d60e92d1 AC |
165 | .TP |
166 | .B randread | |
d1429b5c | 167 | Random reads. |
d60e92d1 AC |
168 | .TP |
169 | .B randwrite | |
d1429b5c | 170 | Random writes. |
d60e92d1 AC |
171 | .TP |
172 | .B rw | |
d1429b5c | 173 | Mixed sequential reads and writes. |
d60e92d1 AC |
174 | .TP |
175 | .B randrw | |
d1429b5c | 176 | Mixed random reads and writes. |
d60e92d1 AC |
177 | .RE |
178 | .P | |
38dad62d JA |
179 | For mixed I/O, the default split is 50/50. For certain types of io the result |
180 | may still be skewed a bit, since the speed may be different. It is possible to | |
181 | specify a number of IO's to do before getting a new offset, this is one by | |
182 | appending a `:\fI<nr>\fR to the end of the string given. For a random read, it | |
183 | would look like \fBrw=randread:8\fR for passing in an offset modifier with a | |
059b0802 JA |
184 | value of 8. If the postfix is used with a sequential IO pattern, then the value |
185 | specified will be added to the generated offset for each IO. For instance, | |
186 | using \fBrw=write:4k\fR will skip 4k for every write. It turns sequential IO | |
187 | into sequential IO with holes. See the \fBrw_sequencer\fR option. | |
d60e92d1 AC |
188 | .RE |
189 | .TP | |
38dad62d JA |
190 | .BI rw_sequencer \fR=\fPstr |
191 | If an offset modifier is given by appending a number to the \fBrw=<str>\fR line, | |
192 | then this option controls how that number modifies the IO offset being | |
193 | generated. Accepted values are: | |
194 | .RS | |
195 | .RS | |
196 | .TP | |
197 | .B sequential | |
198 | Generate sequential offset | |
199 | .TP | |
200 | .B identical | |
201 | Generate the same offset | |
202 | .RE | |
203 | .P | |
204 | \fBsequential\fR is only useful for random IO, where fio would normally | |
205 | generate a new random offset for every IO. If you append eg 8 to randread, you | |
206 | would get a new random offset for every 8 IO's. The result would be a seek for | |
207 | only every 8 IO's, instead of for every IO. Use \fBrw=randread:8\fR to specify | |
208 | that. As sequential IO is already sequential, setting \fBsequential\fR for that | |
209 | would not result in any differences. \fBidentical\fR behaves in a similar | |
210 | fashion, except it sends the same offset 8 number of times before generating a | |
211 | new offset. | |
212 | .RE | |
213 | .P | |
214 | .TP | |
90fef2d1 JA |
215 | .BI kb_base \fR=\fPint |
216 | The base unit for a kilobyte. The defacto base is 2^10, 1024. Storage | |
217 | manufacturers like to use 10^3 or 1000 as a base ten unit instead, for obvious | |
218 | reasons. Allow values are 1024 or 1000, with 1024 being the default. | |
219 | .TP | |
d60e92d1 AC |
220 | .BI randrepeat \fR=\fPbool |
221 | Seed the random number generator in a predictable way so results are repeatable | |
d1429b5c | 222 | across runs. Default: true. |
d60e92d1 | 223 | .TP |
2615cc4b JA |
224 | .BI use_os_rand \fR=\fPbool |
225 | Fio can either use the random generator supplied by the OS to generator random | |
226 | offsets, or it can use it's own internal generator (based on Tausworthe). | |
227 | Default is to use the internal generator, which is often of better quality and | |
228 | faster. Default: false. | |
229 | .TP | |
a596f047 EG |
230 | .BI fallocate \fR=\fPstr |
231 | Whether pre-allocation is performed when laying down files. Accepted values | |
232 | are: | |
233 | .RS | |
234 | .RS | |
235 | .TP | |
236 | .B none | |
237 | Do not pre-allocate space. | |
238 | .TP | |
239 | .B posix | |
240 | Pre-allocate via posix_fallocate(). | |
241 | .TP | |
242 | .B keep | |
243 | Pre-allocate via fallocate() with FALLOC_FL_KEEP_SIZE set. | |
244 | .TP | |
245 | .B 0 | |
246 | Backward-compatible alias for 'none'. | |
247 | .TP | |
248 | .B 1 | |
249 | Backward-compatible alias for 'posix'. | |
250 | .RE | |
251 | .P | |
252 | May not be available on all supported platforms. 'keep' is only | |
253 | available on Linux. If using ZFS on Solaris this must be set to 'none' | |
254 | because ZFS doesn't support it. Default: 'posix'. | |
255 | .RE | |
7bc8c2cf | 256 | .TP |
d60e92d1 | 257 | .BI fadvise_hint \fR=\fPbool |
d1429b5c AC |
258 | Disable use of \fIposix_fadvise\fR\|(2) to advise the kernel what I/O patterns |
259 | are likely to be issued. Default: true. | |
d60e92d1 | 260 | .TP |
f7fa2653 | 261 | .BI size \fR=\fPint |
d60e92d1 AC |
262 | Total size of I/O for this job. \fBfio\fR will run until this many bytes have |
263 | been transfered, unless limited by other options (\fBruntime\fR, for instance). | |
d7c8be03 | 264 | Unless \fBnrfiles\fR and \fBfilesize\fR options are given, this amount will be |
d6667268 JA |
265 | divided between the available files for the job. If not set, fio will use the |
266 | full size of the given files or devices. If the the files do not exist, size | |
7bb59102 JA |
267 | must be given. It is also possible to give size as a percentage between 1 and |
268 | 100. If size=20% is given, fio will use 20% of the full size of the given files | |
269 | or devices. | |
d60e92d1 | 270 | .TP |
74586c1e | 271 | .BI fill_device \fR=\fPbool "\fR,\fB fill_fs" \fR=\fPbool |
3ce9dcaf JA |
272 | Sets size to something really large and waits for ENOSPC (no space left on |
273 | device) as the terminating condition. Only makes sense with sequential write. | |
274 | For a read workload, the mount point will be filled first then IO started on | |
4f12432e JA |
275 | the result. This option doesn't make sense if operating on a raw device node, |
276 | since the size of that is already known by the file system. Additionally, | |
277 | writing beyond end-of-device will not return ENOSPC there. | |
3ce9dcaf | 278 | .TP |
d60e92d1 AC |
279 | .BI filesize \fR=\fPirange |
280 | Individual file sizes. May be a range, in which case \fBfio\fR will select sizes | |
d1429b5c AC |
281 | for files at random within the given range, limited to \fBsize\fR in total (if |
282 | that is given). If \fBfilesize\fR is not specified, each created file is the | |
283 | same size. | |
d60e92d1 | 284 | .TP |
f7fa2653 | 285 | .BI blocksize \fR=\fPint[,int] "\fR,\fB bs" \fR=\fPint[,int] |
d60e92d1 | 286 | Block size for I/O units. Default: 4k. Values for reads and writes can be |
656ebab7 | 287 | specified separately in the format \fIread\fR,\fIwrite\fR, either of |
d60e92d1 AC |
288 | which may be empty to leave that value at its default. |
289 | .TP | |
9183788d | 290 | .BI blocksize_range \fR=\fPirange[,irange] "\fR,\fB bsrange" \fR=\fPirange[,irange] |
d1429b5c AC |
291 | Specify a range of I/O block sizes. The issued I/O unit will always be a |
292 | multiple of the minimum size, unless \fBblocksize_unaligned\fR is set. Applies | |
9183788d | 293 | to both reads and writes if only one range is given, but can be specified |
656ebab7 | 294 | separately with a comma seperating the values. Example: bsrange=1k-4k,2k-8k. |
9183788d JA |
295 | Also (see \fBblocksize\fR). |
296 | .TP | |
297 | .BI bssplit \fR=\fPstr | |
298 | This option allows even finer grained control of the block sizes issued, | |
299 | not just even splits between them. With this option, you can weight various | |
300 | block sizes for exact control of the issued IO for a job that has mixed | |
301 | block sizes. The format of the option is bssplit=blocksize/percentage, | |
5982a925 | 302 | optionally adding as many definitions as needed separated by a colon. |
9183788d | 303 | Example: bssplit=4k/10:64k/50:32k/40 would issue 50% 64k blocks, 10% 4k |
c83cdd3e JA |
304 | blocks and 40% 32k blocks. \fBbssplit\fR also supports giving separate |
305 | splits to reads and writes. The format is identical to what the | |
306 | \fBbs\fR option accepts, the read and write parts are separated with a | |
307 | comma. | |
d60e92d1 AC |
308 | .TP |
309 | .B blocksize_unaligned\fR,\fP bs_unaligned | |
d1429b5c AC |
310 | If set, any size in \fBblocksize_range\fR may be used. This typically won't |
311 | work with direct I/O, as that normally requires sector alignment. | |
d60e92d1 | 312 | .TP |
2b7a01d0 | 313 | .BI blockalign \fR=\fPint[,int] "\fR,\fB ba" \fR=\fPint[,int] |
639ce0f3 MS |
314 | At what boundary to align random IO offsets. Defaults to the same as 'blocksize' |
315 | the minimum blocksize given. Minimum alignment is typically 512b | |
2b7a01d0 JA |
316 | for using direct IO, though it usually depends on the hardware block size. |
317 | This option is mutually exclusive with using a random map for files, so it | |
318 | will turn off that option. | |
43602667 | 319 | .TP |
d60e92d1 AC |
320 | .B zero_buffers |
321 | Initialise buffers with all zeros. Default: fill buffers with random data. | |
322 | .TP | |
901bb994 JA |
323 | .B refill_buffers |
324 | If this option is given, fio will refill the IO buffers on every submit. The | |
325 | default is to only fill it at init time and reuse that data. Only makes sense | |
326 | if zero_buffers isn't specified, naturally. If data verification is enabled, | |
327 | refill_buffers is also automatically enabled. | |
328 | .TP | |
d60e92d1 AC |
329 | .BI nrfiles \fR=\fPint |
330 | Number of files to use for this job. Default: 1. | |
331 | .TP | |
332 | .BI openfiles \fR=\fPint | |
333 | Number of files to keep open at the same time. Default: \fBnrfiles\fR. | |
334 | .TP | |
335 | .BI file_service_type \fR=\fPstr | |
336 | Defines how files to service are selected. The following types are defined: | |
337 | .RS | |
338 | .RS | |
339 | .TP | |
340 | .B random | |
341 | Choose a file at random | |
342 | .TP | |
343 | .B roundrobin | |
344 | Round robin over open files (default). | |
6b7f6851 JA |
345 | .B sequential |
346 | Do each file in the set sequentially. | |
d60e92d1 AC |
347 | .RE |
348 | .P | |
349 | The number of I/Os to issue before switching a new file can be specified by | |
350 | appending `:\fIint\fR' to the service type. | |
351 | .RE | |
352 | .TP | |
353 | .BI ioengine \fR=\fPstr | |
354 | Defines how the job issues I/O. The following types are defined: | |
355 | .RS | |
356 | .RS | |
357 | .TP | |
358 | .B sync | |
359 | Basic \fIread\fR\|(2) or \fIwrite\fR\|(2) I/O. \fIfseek\fR\|(2) is used to | |
360 | position the I/O location. | |
361 | .TP | |
a31041ea | 362 | .B psync |
363 | Basic \fIpread\fR\|(2) or \fIpwrite\fR\|(2) I/O. | |
364 | .TP | |
9183788d JA |
365 | .B vsync |
366 | Basic \fIreadv\fR\|(2) or \fIwritev\fR\|(2) I/O. Will emulate queuing by | |
367 | coalescing adjacents IOs into a single submission. | |
368 | .TP | |
d60e92d1 | 369 | .B libaio |
c44b1ff5 JA |
370 | Linux native asynchronous I/O. This engine also has a sub-option, |
371 | \fBuserspace_reap\fR. To set it, use \fBioengine=libaio:userspace_reap\fR. | |
372 | Normally, with the libaio engine in use, fio will use the | |
373 | \fIio_getevents\fR\|(3) system call to reap newly returned events. With this | |
374 | flag turned on, the AIO ring will be read directly from user-space to reap | |
375 | events. The reaping mode is only enabled when polling for a minimum of \fB0\fR | |
376 | events (eg when \fBiodepth_batch_complete=0\fR). | |
d60e92d1 AC |
377 | .TP |
378 | .B posixaio | |
03e20d68 BC |
379 | POSIX asynchronous I/O using \fIaio_read\fR\|(3) and \fIaio_write\fR\|(3). |
380 | .TP | |
381 | .B solarisaio | |
382 | Solaris native asynchronous I/O. | |
383 | .TP | |
384 | .B windowsaio | |
385 | Windows native asynchronous I/O. | |
d60e92d1 AC |
386 | .TP |
387 | .B mmap | |
d1429b5c AC |
388 | File is memory mapped with \fImmap\fR\|(2) and data copied using |
389 | \fImemcpy\fR\|(3). | |
d60e92d1 AC |
390 | .TP |
391 | .B splice | |
d1429b5c AC |
392 | \fIsplice\fR\|(2) is used to transfer the data and \fIvmsplice\fR\|(2) to |
393 | transfer data from user-space to the kernel. | |
d60e92d1 AC |
394 | .TP |
395 | .B syslet-rw | |
396 | Use the syslet system calls to make regular read/write asynchronous. | |
397 | .TP | |
398 | .B sg | |
399 | SCSI generic sg v3 I/O. May be either synchronous using the SG_IO ioctl, or if | |
d1429b5c AC |
400 | the target is an sg character device, we use \fIread\fR\|(2) and |
401 | \fIwrite\fR\|(2) for asynchronous I/O. | |
d60e92d1 AC |
402 | .TP |
403 | .B null | |
404 | Doesn't transfer any data, just pretends to. Mainly used to exercise \fBfio\fR | |
405 | itself and for debugging and testing purposes. | |
406 | .TP | |
407 | .B net | |
408 | Transfer over the network. \fBfilename\fR must be set appropriately to | |
409 | `\fIhost\fR/\fIport\fR' regardless of data direction. If receiving, only the | |
410 | \fIport\fR argument is used. | |
411 | .TP | |
412 | .B netsplice | |
413 | Like \fBnet\fR, but uses \fIsplice\fR\|(2) and \fIvmsplice\fR\|(2) to map data | |
414 | and send/receive. | |
415 | .TP | |
53aec0a4 | 416 | .B cpuio |
d60e92d1 AC |
417 | Doesn't transfer any data, but burns CPU cycles according to \fBcpuload\fR and |
418 | \fBcpucycles\fR parameters. | |
419 | .TP | |
420 | .B guasi | |
421 | The GUASI I/O engine is the Generic Userspace Asynchronous Syscall Interface | |
422 | approach to asycnronous I/O. | |
d1429b5c AC |
423 | .br |
424 | See <http://www.xmailserver.org/guasi\-lib.html>. | |
d60e92d1 | 425 | .TP |
21b8aee8 | 426 | .B rdma |
85286c5c BVA |
427 | The RDMA I/O engine supports both RDMA memory semantics (RDMA_WRITE/RDMA_READ) |
428 | and channel semantics (Send/Recv) for the InfiniBand, RoCE and iWARP protocols. | |
21b8aee8 | 429 | .TP |
d60e92d1 AC |
430 | .B external |
431 | Loads an external I/O engine object file. Append the engine filename as | |
432 | `:\fIenginepath\fR'. | |
433 | .RE | |
434 | .RE | |
435 | .TP | |
436 | .BI iodepth \fR=\fPint | |
8489dae4 SK |
437 | Number of I/O units to keep in flight against the file. Note that increasing |
438 | iodepth beyond 1 will not affect synchronous ioengines (except for small | |
ee72ca09 JA |
439 | degress when verify_async is in use). Even async engines my impose OS |
440 | restrictions causing the desired depth not to be achieved. This may happen on | |
441 | Linux when using libaio and not setting \fBdirect\fR=1, since buffered IO is | |
442 | not async on that OS. Keep an eye on the IO depth distribution in the | |
443 | fio output to verify that the achieved depth is as expected. Default: 1. | |
d60e92d1 AC |
444 | .TP |
445 | .BI iodepth_batch \fR=\fPint | |
446 | Number of I/Os to submit at once. Default: \fBiodepth\fR. | |
447 | .TP | |
3ce9dcaf JA |
448 | .BI iodepth_batch_complete \fR=\fPint |
449 | This defines how many pieces of IO to retrieve at once. It defaults to 1 which | |
450 | means that we'll ask for a minimum of 1 IO in the retrieval process from the | |
451 | kernel. The IO retrieval will go on until we hit the limit set by | |
452 | \fBiodepth_low\fR. If this variable is set to 0, then fio will always check for | |
453 | completed events before queuing more IO. This helps reduce IO latency, at the | |
454 | cost of more retrieval system calls. | |
455 | .TP | |
d60e92d1 AC |
456 | .BI iodepth_low \fR=\fPint |
457 | Low watermark indicating when to start filling the queue again. Default: | |
458 | \fBiodepth\fR. | |
459 | .TP | |
460 | .BI direct \fR=\fPbool | |
461 | If true, use non-buffered I/O (usually O_DIRECT). Default: false. | |
462 | .TP | |
463 | .BI buffered \fR=\fPbool | |
464 | If true, use buffered I/O. This is the opposite of the \fBdirect\fR parameter. | |
465 | Default: true. | |
466 | .TP | |
f7fa2653 | 467 | .BI offset \fR=\fPint |
d60e92d1 AC |
468 | Offset in the file to start I/O. Data before the offset will not be touched. |
469 | .TP | |
470 | .BI fsync \fR=\fPint | |
d1429b5c AC |
471 | How many I/Os to perform before issuing an \fBfsync\fR\|(2) of dirty data. If |
472 | 0, don't sync. Default: 0. | |
d60e92d1 | 473 | .TP |
5f9099ea JA |
474 | .BI fdatasync \fR=\fPint |
475 | Like \fBfsync\fR, but uses \fBfdatasync\fR\|(2) instead to only sync the | |
476 | data parts of the file. Default: 0. | |
477 | .TP | |
e76b1da4 JA |
478 | .BI sync_file_range \fR=\fPstr:int |
479 | Use sync_file_range() for every \fRval\fP number of write operations. Fio will | |
480 | track range of writes that have happened since the last sync_file_range() call. | |
481 | \fRstr\fP can currently be one or more of: | |
482 | .RS | |
483 | .TP | |
484 | .B wait_before | |
485 | SYNC_FILE_RANGE_WAIT_BEFORE | |
486 | .TP | |
487 | .B write | |
488 | SYNC_FILE_RANGE_WRITE | |
489 | .TP | |
490 | .B wait_after | |
491 | SYNC_FILE_RANGE_WRITE | |
492 | .TP | |
493 | .RE | |
494 | .P | |
495 | So if you do sync_file_range=wait_before,write:8, fio would use | |
496 | \fBSYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE\fP for every 8 writes. | |
497 | Also see the sync_file_range(2) man page. This option is Linux specific. | |
498 | .TP | |
d60e92d1 | 499 | .BI overwrite \fR=\fPbool |
d1429b5c | 500 | If writing, setup the file first and do overwrites. Default: false. |
d60e92d1 AC |
501 | .TP |
502 | .BI end_fsync \fR=\fPbool | |
d1429b5c | 503 | Sync file contents when job exits. Default: false. |
d60e92d1 AC |
504 | .TP |
505 | .BI fsync_on_close \fR=\fPbool | |
506 | If true, sync file contents on close. This differs from \fBend_fsync\fR in that | |
d1429b5c | 507 | it will happen on every close, not just at the end of the job. Default: false. |
d60e92d1 AC |
508 | .TP |
509 | .BI rwmixcycle \fR=\fPint | |
510 | How many milliseconds before switching between reads and writes for a mixed | |
511 | workload. Default: 500ms. | |
512 | .TP | |
513 | .BI rwmixread \fR=\fPint | |
514 | Percentage of a mixed workload that should be reads. Default: 50. | |
515 | .TP | |
516 | .BI rwmixwrite \fR=\fPint | |
d1429b5c | 517 | Percentage of a mixed workload that should be writes. If \fBrwmixread\fR and |
c35dd7a6 JA |
518 | \fBrwmixwrite\fR are given and do not sum to 100%, the latter of the two |
519 | overrides the first. This may interfere with a given rate setting, if fio is | |
520 | asked to limit reads or writes to a certain rate. If that is the case, then | |
521 | the distribution may be skewed. Default: 50. | |
d60e92d1 AC |
522 | .TP |
523 | .B norandommap | |
524 | Normally \fBfio\fR will cover every block of the file when doing random I/O. If | |
525 | this parameter is given, a new offset will be chosen without looking at past | |
526 | I/O history. This parameter is mutually exclusive with \fBverify\fR. | |
527 | .TP | |
744492c9 | 528 | .BI softrandommap \fR=\fPbool |
3ce9dcaf JA |
529 | See \fBnorandommap\fR. If fio runs with the random block map enabled and it |
530 | fails to allocate the map, if this option is set it will continue without a | |
531 | random block map. As coverage will not be as complete as with random maps, this | |
532 | option is disabled by default. | |
533 | .TP | |
d60e92d1 AC |
534 | .BI nice \fR=\fPint |
535 | Run job with given nice value. See \fInice\fR\|(2). | |
536 | .TP | |
537 | .BI prio \fR=\fPint | |
538 | Set I/O priority value of this job between 0 (highest) and 7 (lowest). See | |
539 | \fIionice\fR\|(1). | |
540 | .TP | |
541 | .BI prioclass \fR=\fPint | |
542 | Set I/O priority class. See \fIionice\fR\|(1). | |
543 | .TP | |
544 | .BI thinktime \fR=\fPint | |
545 | Stall job for given number of microseconds between issuing I/Os. | |
546 | .TP | |
547 | .BI thinktime_spin \fR=\fPint | |
548 | Pretend to spend CPU time for given number of microseconds, sleeping the rest | |
549 | of the time specified by \fBthinktime\fR. Only valid if \fBthinktime\fR is set. | |
550 | .TP | |
551 | .BI thinktime_blocks \fR=\fPint | |
552 | Number of blocks to issue before waiting \fBthinktime\fR microseconds. | |
553 | Default: 1. | |
554 | .TP | |
555 | .BI rate \fR=\fPint | |
c35dd7a6 JA |
556 | Cap bandwidth used by this job. The number is in bytes/sec, the normal postfix |
557 | rules apply. You can use \fBrate\fR=500k to limit reads and writes to 500k each, | |
558 | or you can specify read and writes separately. Using \fBrate\fR=1m,500k would | |
559 | limit reads to 1MB/sec and writes to 500KB/sec. Capping only reads or writes | |
560 | can be done with \fBrate\fR=,500k or \fBrate\fR=500k,. The former will only | |
561 | limit writes (to 500KB/sec), the latter will only limit reads. | |
d60e92d1 AC |
562 | .TP |
563 | .BI ratemin \fR=\fPint | |
564 | Tell \fBfio\fR to do whatever it can to maintain at least the given bandwidth. | |
c35dd7a6 JA |
565 | Failing to meet this requirement will cause the job to exit. The same format |
566 | as \fBrate\fR is used for read vs write separation. | |
d60e92d1 AC |
567 | .TP |
568 | .BI rate_iops \fR=\fPint | |
c35dd7a6 JA |
569 | Cap the bandwidth to this number of IOPS. Basically the same as rate, just |
570 | specified independently of bandwidth. The same format as \fBrate\fR is used for | |
571 | read vs write seperation. If \fBblocksize\fR is a range, the smallest block | |
572 | size is used as the metric. | |
d60e92d1 AC |
573 | .TP |
574 | .BI rate_iops_min \fR=\fPint | |
c35dd7a6 JA |
575 | If this rate of I/O is not met, the job will exit. The same format as \fBrate\fR |
576 | is used for read vs write seperation. | |
d60e92d1 AC |
577 | .TP |
578 | .BI ratecycle \fR=\fPint | |
579 | Average bandwidth for \fBrate\fR and \fBratemin\fR over this number of | |
580 | milliseconds. Default: 1000ms. | |
581 | .TP | |
582 | .BI cpumask \fR=\fPint | |
583 | Set CPU affinity for this job. \fIint\fR is a bitmask of allowed CPUs the job | |
584 | may run on. See \fBsched_setaffinity\fR\|(2). | |
585 | .TP | |
586 | .BI cpus_allowed \fR=\fPstr | |
587 | Same as \fBcpumask\fR, but allows a comma-delimited list of CPU numbers. | |
588 | .TP | |
589 | .BI startdelay \fR=\fPint | |
590 | Delay start of job for the specified number of seconds. | |
591 | .TP | |
592 | .BI runtime \fR=\fPint | |
593 | Terminate processing after the specified number of seconds. | |
594 | .TP | |
595 | .B time_based | |
596 | If given, run for the specified \fBruntime\fR duration even if the files are | |
597 | completely read or written. The same workload will be repeated as many times | |
598 | as \fBruntime\fR allows. | |
599 | .TP | |
901bb994 JA |
600 | .BI ramp_time \fR=\fPint |
601 | If set, fio will run the specified workload for this amount of time before | |
602 | logging any performance numbers. Useful for letting performance settle before | |
603 | logging results, thus minimizing the runtime required for stable results. Note | |
c35dd7a6 JA |
604 | that the \fBramp_time\fR is considered lead in time for a job, thus it will |
605 | increase the total runtime if a special timeout or runtime is specified. | |
901bb994 | 606 | .TP |
d60e92d1 AC |
607 | .BI invalidate \fR=\fPbool |
608 | Invalidate buffer-cache for the file prior to starting I/O. Default: true. | |
609 | .TP | |
610 | .BI sync \fR=\fPbool | |
611 | Use synchronous I/O for buffered writes. For the majority of I/O engines, | |
d1429b5c | 612 | this means using O_SYNC. Default: false. |
d60e92d1 AC |
613 | .TP |
614 | .BI iomem \fR=\fPstr "\fR,\fP mem" \fR=\fPstr | |
615 | Allocation method for I/O unit buffer. Allowed values are: | |
616 | .RS | |
617 | .RS | |
618 | .TP | |
619 | .B malloc | |
620 | Allocate memory with \fImalloc\fR\|(3). | |
621 | .TP | |
622 | .B shm | |
623 | Use shared memory buffers allocated through \fIshmget\fR\|(2). | |
624 | .TP | |
625 | .B shmhuge | |
626 | Same as \fBshm\fR, but use huge pages as backing. | |
627 | .TP | |
628 | .B mmap | |
629 | Use \fImmap\fR\|(2) for allocation. Uses anonymous memory unless a filename | |
630 | is given after the option in the format `:\fIfile\fR'. | |
631 | .TP | |
632 | .B mmaphuge | |
633 | Same as \fBmmap\fR, but use huge files as backing. | |
634 | .RE | |
635 | .P | |
636 | The amount of memory allocated is the maximum allowed \fBblocksize\fR for the | |
637 | job multiplied by \fBiodepth\fR. For \fBshmhuge\fR or \fBmmaphuge\fR to work, | |
638 | the system must have free huge pages allocated. \fBmmaphuge\fR also needs to | |
2e266ba6 JA |
639 | have hugetlbfs mounted, and \fIfile\fR must point there. At least on Linux, |
640 | huge pages must be manually allocated. See \fB/proc/sys/vm/nr_hugehages\fR | |
641 | and the documentation for that. Normally you just need to echo an appropriate | |
642 | number, eg echoing 8 will ensure that the OS has 8 huge pages ready for | |
643 | use. | |
d60e92d1 AC |
644 | .RE |
645 | .TP | |
d392365e | 646 | .BI iomem_align \fR=\fPint "\fR,\fP mem_align" \fR=\fPint |
d529ee19 JA |
647 | This indiciates the memory alignment of the IO memory buffers. Note that the |
648 | given alignment is applied to the first IO unit buffer, if using \fBiodepth\fR | |
649 | the alignment of the following buffers are given by the \fBbs\fR used. In | |
650 | other words, if using a \fBbs\fR that is a multiple of the page sized in the | |
651 | system, all buffers will be aligned to this value. If using a \fBbs\fR that | |
652 | is not page aligned, the alignment of subsequent IO memory buffers is the | |
653 | sum of the \fBiomem_align\fR and \fBbs\fR used. | |
654 | .TP | |
f7fa2653 | 655 | .BI hugepage\-size \fR=\fPint |
d60e92d1 | 656 | Defines the size of a huge page. Must be at least equal to the system setting. |
b22989b9 | 657 | Should be a multiple of 1MB. Default: 4MB. |
d60e92d1 AC |
658 | .TP |
659 | .B exitall | |
660 | Terminate all jobs when one finishes. Default: wait for each job to finish. | |
661 | .TP | |
662 | .BI bwavgtime \fR=\fPint | |
663 | Average bandwidth calculations over the given time in milliseconds. Default: | |
664 | 500ms. | |
665 | .TP | |
666 | .BI create_serialize \fR=\fPbool | |
d1429b5c | 667 | If true, serialize file creation for the jobs. Default: true. |
d60e92d1 AC |
668 | .TP |
669 | .BI create_fsync \fR=\fPbool | |
670 | \fIfsync\fR\|(2) data file after creation. Default: true. | |
671 | .TP | |
6b7f6851 JA |
672 | .BI create_on_open \fR=\fPbool |
673 | If true, the files are not created until they are opened for IO by the job. | |
674 | .TP | |
e9f48479 JA |
675 | .BI pre_read \fR=\fPbool |
676 | If this is given, files will be pre-read into memory before starting the given | |
677 | IO operation. This will also clear the \fR \fBinvalidate\fR flag, since it is | |
9c0d2241 JA |
678 | pointless to pre-read and then drop the cache. This will only work for IO |
679 | engines that are seekable, since they allow you to read the same data | |
680 | multiple times. Thus it will not work on eg network or splice IO. | |
e9f48479 | 681 | .TP |
d60e92d1 AC |
682 | .BI unlink \fR=\fPbool |
683 | Unlink job files when done. Default: false. | |
684 | .TP | |
685 | .BI loops \fR=\fPint | |
686 | Specifies the number of iterations (runs of the same workload) of this job. | |
687 | Default: 1. | |
688 | .TP | |
689 | .BI do_verify \fR=\fPbool | |
690 | Run the verify phase after a write phase. Only valid if \fBverify\fR is set. | |
691 | Default: true. | |
692 | .TP | |
693 | .BI verify \fR=\fPstr | |
694 | Method of verifying file contents after each iteration of the job. Allowed | |
695 | values are: | |
696 | .RS | |
697 | .RS | |
698 | .TP | |
b892dc08 | 699 | .B md5 crc16 crc32 crc32c crc32c-intel crc64 crc7 sha256 sha512 sha1 |
0539d758 JA |
700 | Store appropriate checksum in the header of each block. crc32c-intel is |
701 | hardware accelerated SSE4.2 driven, falls back to regular crc32c if | |
702 | not supported by the system. | |
d60e92d1 AC |
703 | .TP |
704 | .B meta | |
705 | Write extra information about each I/O (timestamp, block number, etc.). The | |
996093bb | 706 | block number is verified. See \fBverify_pattern\fR as well. |
d60e92d1 AC |
707 | .TP |
708 | .B null | |
709 | Pretend to verify. Used for testing internals. | |
710 | .RE | |
b892dc08 JA |
711 | |
712 | This option can be used for repeated burn-in tests of a system to make sure | |
713 | that the written data is also correctly read back. If the data direction given | |
714 | is a read or random read, fio will assume that it should verify a previously | |
715 | written file. If the data direction includes any form of write, the verify will | |
716 | be of the newly written data. | |
d60e92d1 AC |
717 | .RE |
718 | .TP | |
719 | .BI verify_sort \fR=\fPbool | |
720 | If true, written verify blocks are sorted if \fBfio\fR deems it to be faster to | |
721 | read them back in a sorted manner. Default: true. | |
722 | .TP | |
f7fa2653 | 723 | .BI verify_offset \fR=\fPint |
d60e92d1 | 724 | Swap the verification header with data somewhere else in the block before |
d1429b5c | 725 | writing. It is swapped back before verifying. |
d60e92d1 | 726 | .TP |
f7fa2653 | 727 | .BI verify_interval \fR=\fPint |
d60e92d1 AC |
728 | Write the verification header for this number of bytes, which should divide |
729 | \fBblocksize\fR. Default: \fBblocksize\fR. | |
730 | .TP | |
996093bb JA |
731 | .BI verify_pattern \fR=\fPstr |
732 | If set, fio will fill the io buffers with this pattern. Fio defaults to filling | |
733 | with totally random bytes, but sometimes it's interesting to fill with a known | |
734 | pattern for io verification purposes. Depending on the width of the pattern, | |
735 | fio will fill 1/2/3/4 bytes of the buffer at the time(it can be either a | |
736 | decimal or a hex number). The verify_pattern if larger than a 32-bit quantity | |
737 | has to be a hex number that starts with either "0x" or "0X". Use with | |
738 | \fBverify\fP=meta. | |
739 | .TP | |
d60e92d1 AC |
740 | .BI verify_fatal \fR=\fPbool |
741 | If true, exit the job on the first observed verification failure. Default: | |
742 | false. | |
743 | .TP | |
b463e936 JA |
744 | .BI verify_dump \fR=\fPbool |
745 | If set, dump the contents of both the original data block and the data block we | |
746 | read off disk to files. This allows later analysis to inspect just what kind of | |
747 | data corruption occurred. On by default. | |
748 | .TP | |
e8462bd8 JA |
749 | .BI verify_async \fR=\fPint |
750 | Fio will normally verify IO inline from the submitting thread. This option | |
751 | takes an integer describing how many async offload threads to create for IO | |
752 | verification instead, causing fio to offload the duty of verifying IO contents | |
c85c324c JA |
753 | to one or more separate threads. If using this offload option, even sync IO |
754 | engines can benefit from using an \fBiodepth\fR setting higher than 1, as it | |
755 | allows them to have IO in flight while verifies are running. | |
e8462bd8 JA |
756 | .TP |
757 | .BI verify_async_cpus \fR=\fPstr | |
758 | Tell fio to set the given CPU affinity on the async IO verification threads. | |
759 | See \fBcpus_allowed\fP for the format used. | |
760 | .TP | |
6f87418f JA |
761 | .BI verify_backlog \fR=\fPint |
762 | Fio will normally verify the written contents of a job that utilizes verify | |
763 | once that job has completed. In other words, everything is written then | |
764 | everything is read back and verified. You may want to verify continually | |
765 | instead for a variety of reasons. Fio stores the meta data associated with an | |
766 | IO block in memory, so for large verify workloads, quite a bit of memory would | |
092f707f DN |
767 | be used up holding this meta data. If this option is enabled, fio will write |
768 | only N blocks before verifying these blocks. | |
6f87418f JA |
769 | .TP |
770 | .BI verify_backlog_batch \fR=\fPint | |
771 | Control how many blocks fio will verify if verify_backlog is set. If not set, | |
772 | will default to the value of \fBverify_backlog\fR (meaning the entire queue is | |
092f707f DN |
773 | read back and verified). If \fBverify_backlog_batch\fR is less than |
774 | \fBverify_backlog\fR then not all blocks will be verified, if | |
775 | \fBverify_backlog_batch\fR is larger than \fBverify_backlog\fR, some blocks | |
776 | will be verified more than once. | |
6f87418f | 777 | .TP |
d392365e | 778 | .B stonewall "\fR,\fP wait_for_previous" |
5982a925 | 779 | Wait for preceding jobs in the job file to exit before starting this one. |
d60e92d1 AC |
780 | \fBstonewall\fR implies \fBnew_group\fR. |
781 | .TP | |
782 | .B new_group | |
783 | Start a new reporting group. If not given, all jobs in a file will be part | |
784 | of the same reporting group, unless separated by a stonewall. | |
785 | .TP | |
786 | .BI numjobs \fR=\fPint | |
787 | Number of clones (processes/threads performing the same workload) of this job. | |
788 | Default: 1. | |
789 | .TP | |
790 | .B group_reporting | |
791 | If set, display per-group reports instead of per-job when \fBnumjobs\fR is | |
792 | specified. | |
793 | .TP | |
794 | .B thread | |
795 | Use threads created with \fBpthread_create\fR\|(3) instead of processes created | |
796 | with \fBfork\fR\|(2). | |
797 | .TP | |
f7fa2653 | 798 | .BI zonesize \fR=\fPint |
d60e92d1 AC |
799 | Divide file into zones of the specified size in bytes. See \fBzoneskip\fR. |
800 | .TP | |
f7fa2653 | 801 | .BI zoneskip \fR=\fPint |
d1429b5c | 802 | Skip the specified number of bytes when \fBzonesize\fR bytes of data have been |
d60e92d1 AC |
803 | read. |
804 | .TP | |
805 | .BI write_iolog \fR=\fPstr | |
5b42a488 SH |
806 | Write the issued I/O patterns to the specified file. Specify a separate file |
807 | for each job, otherwise the iologs will be interspersed and the file may be | |
808 | corrupt. | |
d60e92d1 AC |
809 | .TP |
810 | .BI read_iolog \fR=\fPstr | |
811 | Replay the I/O patterns contained in the specified file generated by | |
812 | \fBwrite_iolog\fR, or may be a \fBblktrace\fR binary file. | |
813 | .TP | |
64bbb865 DN |
814 | .BI replay_no_stall \fR=\fPint |
815 | While replaying I/O patterns using \fBread_iolog\fR the default behavior | |
816 | attempts to respect timing information between I/Os. Enabling | |
817 | \fBreplay_no_stall\fR causes I/Os to be replayed as fast as possible while | |
818 | still respecting ordering. | |
819 | .TP | |
d1c46c04 DN |
820 | .BI replay_redirect \fR=\fPstr |
821 | While replaying I/O patterns using \fBread_iolog\fR the default behavior | |
822 | is to replay the IOPS onto the major/minor device that each IOP was recorded | |
823 | from. Setting \fBreplay_redirect\fR causes all IOPS to be replayed onto the | |
824 | single specified device regardless of the device it was recorded from. | |
825 | .TP | |
901bb994 JA |
826 | .B write_bw_log \fR=\fPstr |
827 | If given, write a bandwidth log of the jobs in this job file. Can be used to | |
828 | store data of the bandwidth of the jobs in their lifetime. The included | |
829 | fio_generate_plots script uses gnuplot to turn these text files into nice | |
830 | graphs. See \fBwrite_log_log\fR for behaviour of given filename. For this | |
831 | option, the postfix is _bw.log. | |
d60e92d1 | 832 | .TP |
64bbb865 | 833 | .B write_lat_log \fR=\fPstr |
901bb994 JA |
834 | Same as \fBwrite_bw_log\fR, but writes I/O completion latencies. If no |
835 | filename is given with this option, the default filename of "jobname_type.log" | |
836 | is used. Even if the filename is given, fio will still append the type of log. | |
837 | .TP | |
02af0988 JA |
838 | .B disable_lat \fR=\fPbool |
839 | Disable measurements of total latency numbers. Useful only for cutting | |
901bb994 JA |
840 | back the number of calls to gettimeofday, as that does impact performance at |
841 | really high IOPS rates. Note that to really get rid of a large amount of these | |
842 | calls, this option must be used with disable_slat and disable_bw as well. | |
843 | .TP | |
02af0988 | 844 | .B disable_clat \fR=\fPbool |
c95f9daf | 845 | Disable measurements of completion latency numbers. See \fBdisable_lat\fR. |
02af0988 | 846 | .TP |
901bb994 | 847 | .B disable_slat \fR=\fPbool |
02af0988 | 848 | Disable measurements of submission latency numbers. See \fBdisable_lat\fR. |
901bb994 JA |
849 | .TP |
850 | .B disable_bw_measurement \fR=\fPbool | |
02af0988 | 851 | Disable measurements of throughput/bandwidth numbers. See \fBdisable_lat\fR. |
d60e92d1 | 852 | .TP |
f7fa2653 | 853 | .BI lockmem \fR=\fPint |
d60e92d1 AC |
854 | Pin the specified amount of memory with \fBmlock\fR\|(2). Can be used to |
855 | simulate a smaller amount of memory. | |
856 | .TP | |
857 | .BI exec_prerun \fR=\fPstr | |
858 | Before running the job, execute the specified command with \fBsystem\fR\|(3). | |
859 | .TP | |
860 | .BI exec_postrun \fR=\fPstr | |
861 | Same as \fBexec_prerun\fR, but the command is executed after the job completes. | |
862 | .TP | |
863 | .BI ioscheduler \fR=\fPstr | |
864 | Attempt to switch the device hosting the file to the specified I/O scheduler. | |
865 | .TP | |
866 | .BI cpuload \fR=\fPint | |
867 | If the job is a CPU cycle-eater, attempt to use the specified percentage of | |
868 | CPU cycles. | |
869 | .TP | |
870 | .BI cpuchunks \fR=\fPint | |
871 | If the job is a CPU cycle-eater, split the load into cycles of the | |
872 | given time in milliseconds. | |
873 | .TP | |
874 | .BI disk_util \fR=\fPbool | |
d1429b5c | 875 | Generate disk utilization statistics if the platform supports it. Default: true. |
901bb994 JA |
876 | .TP |
877 | .BI gtod_reduce \fR=\fPbool | |
878 | Enable all of the gettimeofday() reducing options (disable_clat, disable_slat, | |
879 | disable_bw) plus reduce precision of the timeout somewhat to really shrink the | |
880 | gettimeofday() call count. With this option enabled, we only do about 0.4% of | |
881 | the gtod() calls we would have done if all time keeping was enabled. | |
882 | .TP | |
883 | .BI gtod_cpu \fR=\fPint | |
884 | Sometimes it's cheaper to dedicate a single thread of execution to just getting | |
885 | the current time. Fio (and databases, for instance) are very intensive on | |
886 | gettimeofday() calls. With this option, you can set one CPU aside for doing | |
887 | nothing but logging current time to a shared memory location. Then the other | |
888 | threads/processes that run IO workloads need only copy that segment, instead of | |
889 | entering the kernel with a gettimeofday() call. The CPU set aside for doing | |
890 | these time calls will be excluded from other uses. Fio will manually clear it | |
891 | from the CPU mask of other jobs. | |
f2bba182 | 892 | .TP |
a696fa2a JA |
893 | .BI cgroup \fR=\fPstr |
894 | Add job to this control group. If it doesn't exist, it will be created. | |
6adb38a1 JA |
895 | The system must have a mounted cgroup blkio mount point for this to work. If |
896 | your system doesn't have it mounted, you can do so with: | |
897 | ||
5982a925 | 898 | # mount \-t cgroup \-o blkio none /cgroup |
a696fa2a JA |
899 | .TP |
900 | .BI cgroup_weight \fR=\fPint | |
901 | Set the weight of the cgroup to this value. See the documentation that comes | |
902 | with the kernel, allowed values are in the range of 100..1000. | |
e0b0d892 | 903 | .TP |
7de87099 VG |
904 | .BI cgroup_nodelete \fR=\fPbool |
905 | Normally fio will delete the cgroups it has created after the job completion. | |
906 | To override this behavior and to leave cgroups around after the job completion, | |
907 | set cgroup_nodelete=1. This can be useful if one wants to inspect various | |
908 | cgroup files after job completion. Default: false | |
909 | .TP | |
e0b0d892 JA |
910 | .BI uid \fR=\fPint |
911 | Instead of running as the invoking user, set the user ID to this value before | |
912 | the thread/process does any work. | |
913 | .TP | |
914 | .BI gid \fR=\fPint | |
915 | Set group ID, see \fBuid\fR. | |
83349190 YH |
916 | .TP |
917 | .BI clat_percentiles \fR=\fPbool | |
918 | Enable the reporting of percentiles of completion latencies. | |
919 | .TP | |
920 | .BI percentile_list \fR=\fPfloat_list | |
921 | Overwrite the default list of percentiles for completion | |
922 | latencies. Each number is a floating number in the range (0,100], and | |
923 | the maximum length of the list is 20. Use ':' to separate the | |
924 | numbers. For example, --percentile_list=99.5:99.9 will cause fio to | |
925 | report the values of completion latency below which 99.5% and 99.9% of | |
926 | the observed latencies fell, respectively. | |
d60e92d1 | 927 | .SH OUTPUT |
d1429b5c AC |
928 | While running, \fBfio\fR will display the status of the created jobs. For |
929 | example: | |
d60e92d1 | 930 | .RS |
d1429b5c | 931 | .P |
d60e92d1 AC |
932 | Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s] |
933 | .RE | |
934 | .P | |
d1429b5c AC |
935 | The characters in the first set of brackets denote the current status of each |
936 | threads. The possible values are: | |
937 | .P | |
938 | .PD 0 | |
d60e92d1 AC |
939 | .RS |
940 | .TP | |
941 | .B P | |
942 | Setup but not started. | |
943 | .TP | |
944 | .B C | |
945 | Thread created. | |
946 | .TP | |
947 | .B I | |
948 | Initialized, waiting. | |
949 | .TP | |
950 | .B R | |
951 | Running, doing sequential reads. | |
952 | .TP | |
953 | .B r | |
954 | Running, doing random reads. | |
955 | .TP | |
956 | .B W | |
957 | Running, doing sequential writes. | |
958 | .TP | |
959 | .B w | |
960 | Running, doing random writes. | |
961 | .TP | |
962 | .B M | |
963 | Running, doing mixed sequential reads/writes. | |
964 | .TP | |
965 | .B m | |
966 | Running, doing mixed random reads/writes. | |
967 | .TP | |
968 | .B F | |
969 | Running, currently waiting for \fBfsync\fR\|(2). | |
970 | .TP | |
971 | .B V | |
972 | Running, verifying written data. | |
973 | .TP | |
974 | .B E | |
975 | Exited, not reaped by main thread. | |
976 | .TP | |
977 | .B \- | |
978 | Exited, thread reaped. | |
979 | .RE | |
d1429b5c | 980 | .PD |
d60e92d1 AC |
981 | .P |
982 | The second set of brackets shows the estimated completion percentage of | |
983 | the current group. The third set shows the read and write I/O rate, | |
984 | respectively. Finally, the estimated run time of the job is displayed. | |
985 | .P | |
986 | When \fBfio\fR completes (or is interrupted by Ctrl-C), it will show data | |
987 | for each thread, each group of threads, and each disk, in that order. | |
988 | .P | |
989 | Per-thread statistics first show the threads client number, group-id, and | |
990 | error code. The remaining figures are as follows: | |
991 | .RS | |
d60e92d1 AC |
992 | .TP |
993 | .B io | |
994 | Number of megabytes of I/O performed. | |
995 | .TP | |
996 | .B bw | |
997 | Average data rate (bandwidth). | |
998 | .TP | |
999 | .B runt | |
1000 | Threads run time. | |
1001 | .TP | |
1002 | .B slat | |
1003 | Submission latency minimum, maximum, average and standard deviation. This is | |
1004 | the time it took to submit the I/O. | |
1005 | .TP | |
1006 | .B clat | |
1007 | Completion latency minimum, maximum, average and standard deviation. This | |
1008 | is the time between submission and completion. | |
1009 | .TP | |
1010 | .B bw | |
1011 | Bandwidth minimum, maximum, percentage of aggregate bandwidth received, average | |
1012 | and standard deviation. | |
1013 | .TP | |
1014 | .B cpu | |
1015 | CPU usage statistics. Includes user and system time, number of context switches | |
1016 | this thread went through and number of major and minor page faults. | |
1017 | .TP | |
1018 | .B IO depths | |
1019 | Distribution of I/O depths. Each depth includes everything less than (or equal) | |
1020 | to it, but greater than the previous depth. | |
1021 | .TP | |
1022 | .B IO issued | |
1023 | Number of read/write requests issued, and number of short read/write requests. | |
1024 | .TP | |
1025 | .B IO latencies | |
1026 | Distribution of I/O completion latencies. The numbers follow the same pattern | |
1027 | as \fBIO depths\fR. | |
1028 | .RE | |
d60e92d1 AC |
1029 | .P |
1030 | The group statistics show: | |
d1429b5c | 1031 | .PD 0 |
d60e92d1 AC |
1032 | .RS |
1033 | .TP | |
1034 | .B io | |
1035 | Number of megabytes I/O performed. | |
1036 | .TP | |
1037 | .B aggrb | |
1038 | Aggregate bandwidth of threads in the group. | |
1039 | .TP | |
1040 | .B minb | |
1041 | Minimum average bandwidth a thread saw. | |
1042 | .TP | |
1043 | .B maxb | |
1044 | Maximum average bandwidth a thread saw. | |
1045 | .TP | |
1046 | .B mint | |
d1429b5c | 1047 | Shortest runtime of threads in the group. |
d60e92d1 AC |
1048 | .TP |
1049 | .B maxt | |
1050 | Longest runtime of threads in the group. | |
1051 | .RE | |
d1429b5c | 1052 | .PD |
d60e92d1 AC |
1053 | .P |
1054 | Finally, disk statistics are printed with reads first: | |
d1429b5c | 1055 | .PD 0 |
d60e92d1 AC |
1056 | .RS |
1057 | .TP | |
1058 | .B ios | |
1059 | Number of I/Os performed by all groups. | |
1060 | .TP | |
1061 | .B merge | |
1062 | Number of merges in the I/O scheduler. | |
1063 | .TP | |
1064 | .B ticks | |
1065 | Number of ticks we kept the disk busy. | |
1066 | .TP | |
1067 | .B io_queue | |
1068 | Total time spent in the disk queue. | |
1069 | .TP | |
1070 | .B util | |
1071 | Disk utilization. | |
1072 | .RE | |
d1429b5c | 1073 | .PD |
d60e92d1 AC |
1074 | .SH TERSE OUTPUT |
1075 | If the \fB\-\-minimal\fR option is given, the results will be printed in a | |
562c2d2f DN |
1076 | semicolon-delimited format suitable for scripted use - a job description |
1077 | (if provided) follows on a new line. Note that the first | |
525c2bfa JA |
1078 | number in the line is the version number. If the output has to be changed |
1079 | for some reason, this number will be incremented by 1 to signify that | |
1080 | change. The fields are: | |
d60e92d1 AC |
1081 | .P |
1082 | .RS | |
525c2bfa | 1083 | .B version, jobname, groupid, error |
d60e92d1 AC |
1084 | .P |
1085 | Read status: | |
1086 | .RS | |
b22989b9 | 1087 | .B KB I/O, bandwidth \fR(KB/s)\fP, runtime \fR(ms)\fP |
d60e92d1 AC |
1088 | .P |
1089 | Submission latency: | |
1090 | .RS | |
1091 | .B min, max, mean, standard deviation | |
1092 | .RE | |
1093 | Completion latency: | |
1094 | .RS | |
1095 | .B min, max, mean, standard deviation | |
1096 | .RE | |
525c2bfa JA |
1097 | Total latency: |
1098 | .RS | |
1099 | .B min, max, mean, standard deviation | |
1100 | .RE | |
d60e92d1 AC |
1101 | Bandwidth: |
1102 | .RS | |
1103 | .B min, max, aggregate percentage of total, mean, standard deviation | |
1104 | .RE | |
1105 | .RE | |
1106 | .P | |
1107 | Write status: | |
1108 | .RS | |
b22989b9 | 1109 | .B KB I/O, bandwidth \fR(KB/s)\fP, runtime \fR(ms)\fP |
d60e92d1 AC |
1110 | .P |
1111 | Submission latency: | |
1112 | .RS | |
1113 | .B min, max, mean, standard deviation | |
1114 | .RE | |
1115 | Completion latency: | |
1116 | .RS | |
1117 | .B min, max, mean, standard deviation | |
1118 | .RE | |
525c2bfa JA |
1119 | Total latency: |
1120 | .RS | |
1121 | .B min, max, mean, standard deviation | |
1122 | .RE | |
d60e92d1 AC |
1123 | Bandwidth: |
1124 | .RS | |
1125 | .B min, max, aggregate percentage of total, mean, standard deviation | |
1126 | .RE | |
1127 | .RE | |
1128 | .P | |
d1429b5c | 1129 | CPU usage: |
d60e92d1 | 1130 | .RS |
bd2626f0 | 1131 | .B user, system, context switches, major page faults, minor page faults |
d60e92d1 AC |
1132 | .RE |
1133 | .P | |
1134 | IO depth distribution: | |
1135 | .RS | |
1136 | .B <=1, 2, 4, 8, 16, 32, >=64 | |
1137 | .RE | |
1138 | .P | |
562c2d2f | 1139 | IO latency distribution: |
d60e92d1 | 1140 | .RS |
562c2d2f DN |
1141 | Microseconds: |
1142 | .RS | |
1143 | .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000 | |
1144 | .RE | |
1145 | Milliseconds: | |
1146 | .RS | |
1147 | .B <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, 2000, >=2000 | |
1148 | .RE | |
1149 | .RE | |
1150 | .P | |
5982a925 | 1151 | Error Info (dependent on continue_on_error, default off): |
562c2d2f DN |
1152 | .RS |
1153 | .B total # errors, first error code | |
d60e92d1 AC |
1154 | .RE |
1155 | .P | |
562c2d2f | 1156 | .B text description (if provided in config - appears on newline) |
d60e92d1 AC |
1157 | .RE |
1158 | .SH AUTHORS | |
1159 | .B fio | |
aa58d252 JA |
1160 | was written by Jens Axboe <jens.axboe@oracle.com>, |
1161 | now Jens Axboe <jaxboe@fusionio.com>. | |
d1429b5c AC |
1162 | .br |
1163 | This man page was written by Aaron Carroll <aaronc@cse.unsw.edu.au> based | |
d60e92d1 AC |
1164 | on documentation by Jens Axboe. |
1165 | .SH "REPORTING BUGS" | |
482900c9 | 1166 | Report bugs to the \fBfio\fR mailing list <fio@vger.kernel.org>. |
d1429b5c | 1167 | See \fBREADME\fR. |
d60e92d1 | 1168 | .SH "SEE ALSO" |
d1429b5c AC |
1169 | For further documentation see \fBHOWTO\fR and \fBREADME\fR. |
1170 | .br | |
1171 | Sample jobfiles are available in the \fBexamples\fR directory. | |
d60e92d1 | 1172 |