Debian Codegraph @

doc/options.rst

raw
1 Running Flent
2 =============
3
4 When run, flent must be supplied either (a) a test name and one or more host
5 names to connect to, or (b) one or more input files containing data from
6 previous runs to post-process.
7
8 Test names, hostnames and input file names can all be specified as unqualified
9 arguments, and flent will do its best to guess which is which. For each
10 argument, if it is an existing file, it is assumed to be an input file, if it is
11 the name of an existing test configuration it’s assumed to be a test name, and
12 if neither of those are true, it is assumed to be a host name. The :option:`-i`
13 and :option:`-H` switches can be used to explicitly specify the interpretation
14 of an argument.
15
16 Invocation
17 ----------
18
19 **flent** [*options*\ ] *<host\|test\|input file* ...\ *>*
20
21
22
23 General options
24 ---------------
25
26 .. option:: -o OUTPUT, --output=OUTPUT
27
28 File to write processed output to (default standard out).
29
30 .. option:: -D DATA_DIR, --data-dir=DATA_DIR
31
32 Directory to store data files in. Defaults to the current directory.
33
34 .. option:: -i INPUT, --input=INPUT
35
36 File to read input from (instead of running tests). Input files can also be
37 specified as unqualified arguments without using the :option:`-i` switch.
38
39 .. option:: -f FORMAT, --format=FORMAT
40
41 Select output format (plot, csv, org\_table, stats). Default is no processed
42 output (just writes the JSON data file).
43
44 .. option:: -p PLOT, --plot=PLOT
45
46 Select which plot to output for the given test (implies :option:`-f` plot). Use the
47 :option:`--list-plots` option to see available plots.
48
49 .. option:: -t TITLE, --title-extra=TITLE
50
51 Text to add to plot title and data file name.
52
53 .. option:: -n NOTE, --note=NOTE
54
55 Add arbitrary text as a note to be stored in the JSON data file (under the
56 *NOTE* key in the metadata object).
57
58 .. option:: -r RCFILE, --rcfile=RCFILE
59
60 Load configuration data from *RCFILE* (default *~/.flentrc*). See section
61 below for information on the rc file format.
62
63 .. option:: -x, --extended-metadata
64
65 Collect extended metadata and store it with the data file. May include
66 details of your machine you don’t want to distribute; see the section on the
67 data format below.
68
69 .. option:: --remote-metadata=HOSTNAME
70
71 Collect extended metadata from a remote host. *HOSTNAME* is passed verbatim
72 to ssh, so can include hosts specified in ~/.ssh/config. Note that gathering
73 the data can take some time, since it involves executing several remote
74 commands. This option can be specified multiple times and implies
75 :option:`--extended-metadata`.
76
77 .. option:: --gui
78
79 Run the flent GUI. All other options are used as defaults in the GUI, but can
80 be changed once it is running. The GUI can also be started by running the
81 :command:`flent-gui` binary. For more information on the GUI, see the
82 :doc:`gui` section.
83
84 .. option:: --new-gui-instance
85
86 Start a new GUI instance. Otherwise, flent will try to connect to an already
87 running GUI instance and have that load any new data files specified as
88 arguments. Implies :option:`--gui` when passed on the command line, but not when
89 set in the rc file. Note that when multiple GUI instances are running, there
90 is no guarantee as to which instance will get a subsequent open request (if
91 run again without :option:`--new-gui-instance`).
92
93 .. option:: --gui-no-defer
94
95 Normally, the GUI defers redrawing plots until they are needed to avoid
96 redrawing all open plots every time an option changes. This switch turns off
97 that optimisation in favour of always redrawing everything straight away.
98 This is useful when loading a bunch of plots from the command line and then
99 wanting to flip through them without drawing delay.
100
101 .. option:: -b BATCH_NAME, --batch-name=BATCH_NAME
102
103 Run test batch BATCH\_NAME (must be specified in a batch file loaded by the
104 :option:`--batch-file` option). Can be supplied multiple times.
105
106 .. option:: -B BATCH_FILE, --batch-file=BATCH_FILE
107
108 Load batch file BATCH_FILE. Can be specified multiple times, in which case
109 the files will be combined (with identically-named sections being overridden
110 by later files). See appropriate section below for an explanation of the
111 batch file format.
112
113 .. option:: --batch-override=key=value
114
115 Override parameter ’key’ in the batch config and set it to ’value’. The key
116 name will be case folded to lower case. Can be specified multiple times.
117
118 .. option:: --batch-dry-run
119
120 Dry batch run. Prints what would be done, but doesn’t actually run any tests.
121
122 .. option:: --batch-verbose
123
124 Be verbose during batch run: Print all commands executed.
125
126 .. option:: --batch-no-shuffle
127
128 Do not randomise the order of test runs within each batch.
129
130 .. option:: --batch-repetitions=REPETITIONS
131
132 Shorthand for :option:`--batch-override` ``'repetitions=REPETITIONS’``.
133
134 .. option:: --batch-title=TITLE
135
136 Shorthand for :option:`--batch-override` ``'batch_title=TITLE’``.
137
138 .. option:: --batch-resume=DIR
139
140 Try to resume a previously interrupted batch run. The argument is the
141 top-level output directory from the previous run.
142
143 This will attempt to find a data file in the resume directory and load the
144 BATCH_TIME from the previous run from that and continue. The assumption is
145 that the output directory and filenames are generated from the batch time, so
146 that they will match with the previous run when the same time is used. Then,
147 tests for which data files already exist will be skipped on this run. If the
148 rest of the batch invocation is different from the one being resumed, results
149 may not be what you want.
150
151 There's a check to ensure that the generated output path is a subdirectory of
152 the resume directory, and the whole run will be aborted if it isn't.
153
154 Test configuration options
155 --------------------------
156
157 These options affect the behaviour of the test being run and have no effect when
158 parsing input files.
159
160 .. option :: -H HOST, --host=HOST
161
162 Host to connect to for tests. For tests that support it, multiple hosts can
163 be specified by supplying this option multiple times. Hosts can also be
164 specified as unqualified arguments; this parameter guarantees that the
165 argument be interpreted as a host name (rather than being subject to
166 auto-detection between input files, hostnames and test names).
167
168 .. option:: --local-bind=IP
169
170 Local hostname or IP address to bind to (for test tools that support this).
171 Can be specified multiple times for tests that connect to more than one host;
172 if it is, it must be specified as many times as there are hosts.
173
174 .. option:: --remote-host=idx=HOSTNAME
175
176 A remote hostname to connect to when starting a test. The idx is the runner
177 index, which is assigned sequentially to each *runner* (and so it is *not*
178 the same as the sequence of hostnames). Look for the 'IDX' key in SERIES_META
179 for a test get the idx used here, but note that the idx assignment depends on
180 the exact arguments to the test.
181
182 This works by simply prepending 'ssh HOSTNAME' to the runner command, so it
183 relies on the same binaries being in the same places on both machines, and
184 won't work for all runners.
185
186 This option can be specified multiple times to have multiple runners run on
187 remote hosts.
188
189 .. option:: -l LENGTH, --length=LENGTH
190
191 Base test length (some tests may add some time to this).
192
193 .. option:: -s STEP_SIZE, --step-size=STEP_SIZE
194
195 Measurement data point step size.
196
197 .. option:: -d DELAY, --delay=DELAY
198
199 Number of seconds to delay parts of test (such as bandwidth loaders).
200
201 .. option:: -4, --ipv4
202
203 Use IPv4 for tests (some tests may ignore this).
204
205 .. option:: -6, --ipv6
206
207 Use IPv6 for tests (some tests may ignore this).
208
209 .. option:: --socket-timeout=SOCKET_TIMEOUT
210
211 Socket timeout (in seconds) used for UDP delay measurement, to prevent stalls
212 on packet loss. Only enabled if the installed netperf version is detected to
213 support this (requires SVN version of netperf).
214
215 For the default value, see the output of flent :option:`-h`. The value of
216 this parameter is an implicit upper bound on how long a round-trip time that
217 can be measured. As such you may need to adjust it if you are experiencing
218 latency above the default value. Set to 0 to disable.
219
220 .. option:: --send-size=SEND_SIZE
221
222 Send size (in bytes) used for TCP tests. Netperf uses the socket buffer size
223 by default, which if too large can cause spikes in the throughput results.
224 Lowering this value will increase CPU usage but also improves the fidelity
225 of the throughput results without having to decrease the socket buffer size.
226
227 Can be specified multiple times, with each value corresponding to a stream
228 of a test. If only specified once, the same value will be applied to all
229 flows.
230
231 .. option:: --test-parameter=key=value
232
233 Arbitrary test parameter in key=value format. Key will be case folded to
234 lower case. The values are stored with the results metadata, and so can be
235 used for storing arbitrary information relevant for a particular test run.
236
237 In addition to serving as simple metadata, the test parameters can also
238 affect the behaviour of some test configurations. See the :doc:`tests`
239 section for information on these.
240
241 This option can be specified multiple times to set multiple test parameters.
242
243 .. option:: --swap-up-down
244
245 Switch upstream and downstream directions for data transfer. This means that
246 ’upload’ will become ’download’ and vice versa. Works by exchanging netperf
247 ``TCP_MAERTS`` and ``TCP_STREAM`` parameters, so only works for tests that employ
248 these as their data transfer, and only for the TCP streams.
249
250 .. option:: --socket-stats
251
252 Parse socket stats during test. This will capture and parse socket
253 statistics for all TCP upload flows during a test, adding TCP cwnd and RTT
254 values to the test data. Requires the 'ss' utility to be present on the
255 system, and can fail if there are too many simultaneous upload flows; which
256 is why this option is not enabled by default.
257
258 .. option:: --marking-name
259
260 Define a new symbolic name that can be used when specifying flow markings
261 using the 'markings' test parameter. This can be used to make it easier to
262 specify custom diffserv markings on flows by using symbolic names for each
263 marking value instead of the hex codes. Values specified here will be used
264 in addition to the common values (listed below), and cannot override the
265 built-in names. Names will be case-folded when matching.
266
267 The list of symbolic markings natively supported, along with their hex
268 expansions, are::
269
270 AF11: 0x28 CS0: 0x00
271 AF12: 0x30 CS1: 0x20
272 AF13: 0x38 CS2: 0x40
273 AF21: 0x48 CS3: 0x60
274 AF22: 0x50 CS4: 0x80
275 AF23: 0x58 CS5: 0xa0
276 AF31: 0x68 CS6: 0xc0
277 AF32: 0x70 CS7: 0xe0
278 AF33: 0x78 EF: 0xb8
279 AF41: 0x88
280 AF42: 0x90
281 AF43: 0x98
282
283 Note that the hexadecimal values denote the value of the full ToS byte
284 (including the two ECN bits), so they need to be right-shifted by two bits
285 to get the corresponding diffserv code points.
286
287
288 Plot configuration options
289 --------------------------
290
291 These options are used to configure the appearance of plot output and only make
292 sense combined with :option:`-f` *plot*.
293
294 .. option:: --label-x=LABEL
295 .. option:: --label-y=LABEL
296
297 Override the figure axis labels. Can be specified twice, corresponding to
298 figures with multiple axes.
299
300 .. option:: -I, --invert-latency-y
301
302 Invert latency data series axis (typically the Y-axis), making plots show
303 ’better' values upwards.
304
305 .. option:: -z, --zero-y
306
307 Always start Y axis of plot at zero, instead of autoscaling the axis.
308 Autoscaling is still enabled for the upper bound. This also disables log
309 scale if enabled.
310
311 .. option:: --log-scale={log2,log10}
312
313 Use the specified logarithmic scale on plots.
314
315 .. option:: --norm-factor=FACTOR
316
317 Data normalisation factor. Divide all data points by this value. Can be
318 specified multiple times, in which case each value corresponds to a data
319 series.
320
321 .. option:: --bounds-x=BOUNDS
322 .. option:: --bounds-y=BOUNDS
323
324 Specify bounds of the plot axes. If specifying one number, that will become
325 the upper bound. Specify two numbers separated by a comma to specify both
326 upper and lower bounds. To specify just the lower bound, add a comma
327 afterwards. Can be specified twice, corresponding to figures with multiple
328 axes.
329
330 .. option:: -S, --scale-mode
331
332 Treat file names (except for the first one) passed as unqualified arguments
333 as if passed as :option:`--scale-data` (default as if passed as
334 :option:`--input`).
335
336 .. option:: --concatenate
337
338 Concatenate multiple result sets into one data series. This means that each
339 data file will have its time axis shifted by the preceding series duration
340 and appended to the first data set specified. Only works for data sets from
341 the same test, obviously.
342
343 .. option:: --absolute-time
344
345 Plot data points with absolute UNIX time on the x-axis. This requires the
346 absolute starting time for the test run to be stored in the data file, and so
347 it won’t work with data files that predates this feature.
348
349 .. option:: --subplot-combine
350
351 When plotting multiple data series, plot each one on a separate subplot
352 instead of combining them into one plot. This mode is not supported for all
353 plot types, and only works when :option:`--scale-mode` is disabled.
354
355 .. option:: --skip-missing-series
356
357 Skip missing series entirely from bar plots, instead of leaving an empty
358 space for it.
359
360 .. option:: --no-print-n
361
362 Do not print the number of data points on combined plots. When using plot
363 types that combines results from several test runs, the number of data series
364 in each combined data point is normally added after the series name, (n=X)
365 for X data series. This option turns that off.
366
367 .. option:: --no-annotation
368
369 Exclude annotation with hostnames, time and test length from plots.
370
371 .. option:: --figure-note=NOTE, --fig-note=NOTE
372
373 Add a note (arbitrary text) to the bottom-left of the figure.
374
375 .. option:: --no-title
376
377 Exclude title from plots.
378
379 .. option:: --override-title=TITLE
380
381 Override plot title with this string. Completely discards the configured
382 title (from the test configuration), as well as the title stored in the data
383 set, and replaces it with the value supplied here. This is useful to override
384 the plot title *at the time of plotting*, for instance to add a title to an
385 aggregate plot from several data series. When this parameter is specified,
386 :option:`--no-title` has no effect.
387
388 .. option:: --no-labels
389
390 Hides tick labels from box and bar plots.
391
392 .. option:: --no-markers
393
394 Don’t use line markers to differentiate data series on plots.
395
396 .. option:: --no-legend
397
398 Exclude legend from plots.
399
400 .. option:: --horizontal-legend
401
402 Place a horizontal legend below the plot instead of a vertical one next to
403 it. Doesn't always work well if there are too many items in the legend.
404
405 .. option:: --legend-title=LEGEND_TITLE
406
407 Override legend title on plot.
408
409 .. option:: --legend-placement=LEGEND_PLACEMENT
410
411 Control legend placement. Enabling this option will place the legend inside
412 the plot at the specified location. Can be one of 'best', 'upper right',
413 'upper left', 'lower left', 'lower right', 'right', 'center left', 'center
414 right', 'lower center', 'upper center' or 'center'.
415
416 .. option:: --legend-columns=LEGEND_COLUMNS
417 Set the number of columns in the legend.
418
419 .. option:: --reverse-legend
420
421 Reverse the order of items in the legend. This can be useful to make the
422 legend order match the data series in some cases.
423
424 .. option:: --filter-legend
425
426 Filter legend labels by removing the longest common substring from all
427 entries. This is not particularly smart, so use with care.
428
429 .. option:: --replace-legend=src=dest
430
431 Replace 'src' with 'dst' in legends. Can be specified multiple times.
432
433 .. option:: --filter-regexp=REGEXP
434
435 Filter the plot legend by the supplied regular expression (removing any text
436 from the legend that matches the expression). Can be specified multiple
437 times, in which case the filters will be applied in the order specified. Note
438 that for combining several plot results, the regular expression is also
439 applied before the grouping logic, meaning that a too wide filter can mess up
440 the grouping.
441
442 .. option:: --override-label=LABEL
443
444 Override dataset label. Can be specified multiple times when multiple
445 datasets are being plotted, in which case the order of labels corresponds to
446 the order of datasets.
447
448 Like :option:`--override-title`, this is applied *at the time of plotting*.
449
450 .. option:: --filter-series=SERIES
451
452 Filter out specified series from plot. Can be specified multiple times.
453
454 .. option:: --split-group=LABEL
455
456 Split data sets into groups when creating box plots. Specify this option
457 multiple times to define the new groups; the value of each option is the
458 group name.
459
460 Say you're plotting nine datasets which are really testing two variables with
461 three values each. In this case, it can be useful to have the box plot of the
462 results be split into three parts (corresponding to the values of one
463 variable) with each three boxes in each of them (corresponding to the values
464 of the second variable). This option makes this possible; simply specify it
465 three times with the labels to be used for the three groups.
466
467 A constraint on this option is that the number of datasets being plotted must
468 be divisible by the number of groups.
469
470 .. option:: --colours=COLOURS
471
472 Comma-separated list of colours to be used for the plot colour cycle. Can be
473 specified in any format understood by matplotlib (including HTML hex values
474 prefixed with a #).
475
476 Yes, this option uses British spelling. No, American spelling is not
477 supported.
478
479 .. option:: --override-colour-mode=MODE
480
481 Override colour_mode attribute. This changes the way colours are
482 assigned to bar plots. The default is 'groups' which assigns a separate
483 colour to each group of data series. The alternative is 'series' which
484 assigns a separate colour to each series, repeating them for each data
485 group.
486
487 .. option:: --override-group-by=GROUP
488
489 Override the ``group_by`` setting for combination plots. This is useful to,
490 for instance, switch to splitting up combined data sets by batch run instead
491 of by file name.
492
493 .. option:: --combine-save-dir=DIRNAME
494
495 When doing a combination plot save the intermediate data to ``DIRNAME``. This
496 can then be used for subsequent plotting to avoid having to load all the
497 source data files again on each plot.
498
499 .. option:: --figure-width=FIG_WIDTH
500
501 Figure width in inches. Used when saving plots to file and for default size
502 of the interactive plot window.
503
504 .. option:: --figure-height=FIG_HEIGHT
505
506 Figure height in inches. Used when saving plots to file and for default size
507 of the interactive plot window.
508
509 .. option:: --figure-dpi=FIG_DPI
510
511 Figure DPI. Used when saving plots to raster format files.
512
513 .. option:: --fallback-layout
514
515 Use the fallback layout engine (tight_layout built in to matplotlib). Use
516 this if text is cut off on saved figures. The downside to the fallback engine
517 is that the size of the figure (as specified by :option:`--figure-width` and
518 :option:`--figure-height`) is no longer kept constant.)
519
520 .. option:: --no-matplotlibrc
521
522 Don’t load included matplotlibrc values. Use this if autodetection of custom
523 matplotlibrc fails and flent is inadvertently overriding rc values.
524
525 .. option:: --no-hover-highlight
526
527 Don't highlight data series on hover in interactive plot views. Use this if
528 redrawing is too slow, or the highlighting is undesired for other reasons.
529
530 .. option:: --scale-data=SCALE_DATA
531
532 Additional data files to consider when scaling the plot axes (for plotting
533 several plots with identical axes). Note, this displays only the first data
534 set, but with axis scaling taking into account the additional data sets. Can
535 be supplied multiple times; see also :option:`--scale-mode`.
536
537
538 Test tool-related options
539 -------------------------
540 .. option:: --control-host=HOST
541
542 Hostname for the test control connection (for test tools that support this).
543 Default: First hostname of test target.
544
545 When running tests that uses D-ITG as a test tool (such as the voip-\*
546 tests), this switch controls where flent will look for the D-ITG control
547 server (see section below on running tests with D-ITG). For Netperf-based
548 tests, this option is passed to Netperf to control where to point the control
549 connection. This is useful to, for instance, to run the control server
550 communication over a separate control network so as to not interfere with
551 test traffic.
552
553 There is also a per-flow setting for this for tests that connect to multiple
554 hosts; see the `control_hosts` test parameter in :doc:`tests`. If both are
555 set, the per-flow setting takes precedence for those tests that use it.
556
557 .. option:: --control-local-bind=IP
558
559 Local hostname or IP to bind control connection to (for test tools that
560 support it; currently netperf). If not supplied, the value for
561 :option:`--local-bind` will be used. Note that if this value is passed but
562 :option:`--local-bind` is *not*, netperf will use the value specified here to
563 bind the data connections to as well.
564
565 .. option:: --netperf-control-port=PORT
566
567 Port for Netperf control server. Default: 12865.
568
569 .. option:: --ditg-control-port=PORT
570
571 Port for D-ITG control server. Default: 8000.
572
573 .. option:: --ditg-control-secret=SECRET
574
575 Secret for D-ITG control server authentication. Default: ’’.
576
577 .. option:: --http-getter-urllist=FILENAME
578
579 When running HTTP tests, the http-getter tool is used to fetch URLs (see
580 https://github.com/tohojo/http-getter). This option specifies the filename
581 containing the list of HTTP URLs to get. Can also be a URL, which will then
582 be downloaded as part of each test iteration. If not specified, this is set
583 to http://<hostname>/filelist.txt where <hostname> is the first test
584 hostname.
585
586 .. option:: --http-getter-dns-servers=DNS_SERVERS
587
588 DNS servers to use for http-getter lookups. Format is
589 host[:port][,host[:port]]... This option will only work if libcurl supports
590 it (needs to be built with the ares resolver). Default is none (use the
591 system resolver).
592
593 .. option:: --http-getter-timeout=MILLISECONDS
594
595 Timeout for HTTP connections. Default is to use the test length.
596
597 .. option:: --http-getter-workers=NUMBER
598
599 Number of workers to use for getting HTTP urls. Default is 4.
600
601 .. option:: --irtt-sampling-interval=MILLISECONDS
602
603 Override the sampling interval passed to irtt, in milliseconds. Can be used
604 to run irtt with a higher sampling frequency than the rest of the test. If
605 set, this will override the sampling interval for all instances of irtt used
606 in the test.
607
608 Misc and debugging options:
609 ---------------------------
610
611 .. option:: -L LOG_FILE, --log-file=LOG_FILE
612
613 Write debug log (test program output) to log file.
614
615 .. option:: --list-tests
616
617 List available tests and exit.
618
619 .. option:: --list-plots
620
621 List available plots for selected test and exit.
622
623 .. option:: -V, --version
624
625 Show Flent version information and exit.
626
627 .. option:: -v, --verbose
628
629 Enable verbose logging to console.
630
631 .. option:: -q, --quiet
632
633 Disable normal logging to console (and only log warnings and errors).
634
635 .. option:: --debug-error
636
637 Print full exception backtraces to console.
638
639 .. option:: -h, --help
640
641 Show usage help message and exit.
642
643
644 Signals
645 -------
646
647 Flent will abort what it is currently doing on receiving a **SIGINT** -- this
648 includes killing all runners, cleaning up temporary files and shutting down as
649 gracefully as possible. Runners are killed with **SIGTERM** in this mode, and
650 their output is discarded. If a batch run is in progress, the current test will
651 be interrupted in this way, and the rest of the batch run is aborted. Previously
652 completed tests and their results are not aborted. Post-commands marked as
653 ’essential’ will be run after the test is interrupted. Additionally, flent
654 converts **SIGTERM** into **SIGINT** internally and reacts accordingly.
655
656 Upon receiving a **SIGUSR1**, flent will try to gracefully abort the test it is
657 currently running, and parse the output of the runners to the extent that any
658 such output exists. That is, each runner will be killed by a **SIGINT**, which
659 will cause a graceful shutdown for at least ping and netperf (although netperf
660 running in *TCP_MAERTS* mode will bug out when interrupted like this, so
661 end-of-tests statistics will be missing). Flent will only react once to a
662 **SIGUSR1**, sending exactly one **SIGINT** to the active runners, then wait for
663 them to exit. This may take several seconds in the case of netperf. If the
664 runners for some reason fail to exit, flent will be stuck and will need to be
665 killed with **SIGINT**. If running in batch mode, **SIGUSR1** will only affect
666 the currently running test; subsequent tests will still be run.