1 .TH "LTTNG" "1" "February 9, 2012" "" ""
4 lttng \(em LTTng 2.0 tracer control command line tool
10 lttng [OPTIONS] <COMMAND>
15 The LTTng project aims at providing highly efficient tracing tools for Linux.
16 It's tracers help tracking down performance issues and debugging problems
17 involving multiple concurrent processes and threads. Tracing across multiple
18 systems is also possible.
20 The \fBlttng\fP command line tool from the lttng-tools package is used to control
21 both kernel and user-space tracing. Every interactions with the tracer should
22 be done by this tool or by the liblttng-ctl provided with the lttng-tools
25 LTTng uses a session daemon (lttng-sessiond(8)), acting as a tracing registry,
26 which permits you to interact with multiple tracers (kernel and user-space)
27 inside the same container, a tracing session. Traces can be gathered from the
28 kernel and/or instrumented applications (lttng-ust(3)). Aggregating and reading
29 those traces is done using the babeltrace(1) text viewer.
31 In order to trace the kernel, the session daemon needs to be running as root.
32 LTTng provides the use of a \fBtracing group\fP (default: tracing). Whomever is
33 in that group can interact with the root session daemon and thus trace the
34 kernel. Session daemons can co-exist meaning that you can have a session daemon
35 running as Alice that can be used to trace her applications along side with a
36 root daemon or even a Bob daemon. We highly recommend to start the session
37 daemon at boot time for stable and long term tracing.
39 Every user-space applications instrumented with lttng-ust(3), will
40 automatically register to the session daemon. This feature gives you the
41 ability to list available traceable applications and tracepoints on a per user
42 basis. (See \fBlist\fP command).
46 This program follow the usual GNU command line syntax with long options starting with
47 two dashes. Below is a summary of the available options.
52 Show summary of possible options and commands.
54 .BR "\-v, \-\-verbose"
56 Three levels of verbosity are available which are triggered by putting additional v to
57 the option (\-vv or \-vvv)
60 Suppress all messages (even errors).
62 .BR "\-g, \-\-group NAME"
63 Set unix tracing group name. (default: tracing)
65 .BR "\-n, \-\-no-sessiond"
66 Don't automatically spawn a session daemon.
68 .BR "\-\-sessiond\-path"
69 Set session daemon full binary path.
71 .BR "\-\-list\-options"
72 Simple listing of lttng options.
74 .BR "\-\-list\-commands"
75 Simple listing of lttng commands.
81 Add context to event(s) and/or channel(s).
83 A context is basically extra information appended to a channel or event. For
84 instance, you could ask the tracer to add the PID information within the
85 "sched_switch" kernel event. You can also add performance monitoring unit
86 counters (perf PMU) using the perf kernel API).
88 For example, this command will add the context information 'prio' and two perf
89 counters (hardware branch misses and cache misses), to all events in the trace
92 # lttng add-context \-k \-t prio \-t perf:branch-misses \-t perf:cache-misses
94 Please take a look at the help (\-h/\-\-help) for a detailed list of available
97 If no channel and no event is given (\-c/\-e), the context is added to all
98 channels (which applies automatically to all events in that channel). Otherwise
99 the context will be added only to the channel (\-c) and/or event (\-e) indicated.
101 If \fB\-s, \-\-session\fP is omitted, the session name is taken from the .lttngrc
109 Show summary of possible options and commands.
110 \-s, \-\-session NAME
111 Apply on session name.
112 \-c, \-\-channel NAME
113 Apply on channel name.
117 Apply for the kernel tracer
119 Apply for the user-space tracer
121 Context type. You can repeat this option on the command line. Please
122 use "lttng add-context \-h" to list all available types.
127 .IP "\fBcalibrate\fP"
129 Quantify LTTng overhead
131 The LTTng calibrate command can be used to find out the combined average
132 overhead of the LTTng tracer and the instrumentation mechanisms used. This
133 overhead can be calibrated in terms of time or using any of the PMU performance
134 counter available on the system.
136 For now, the only calibration implemented is that of the kernel function
137 instrumentation (kretprobes).
139 * Calibrate kernel function instrumentation
141 Let's use an example to show this calibration. We use an i7 processor with 4
142 general-purpose PMU registers. This information is available by issuing dmesg,
143 looking for "generic registers".
145 This sequence of commands will gather a trace executing a kretprobe hooked on
146 an empty function, gathering PMU counters LLC (Last Level Cache) misses
147 information (see lttng add-context \-\-help to see the list of available PMU
150 # lttng create calibrate-function
151 # lttng enable-event calibrate \-\-kernel \-\-function lttng_calibrate_kretprobe
152 # lttng add-context \-\-kernel \-t perf:LLC-load-misses \-t perf:LLC-store-misses \\
153 \-t perf:LLC-prefetch-misses
155 # for a in $(seq 1 10); do \\
156 lttng calibrate \-\-kernel \-\-function;
159 # babeltrace $(ls \-1drt ~/lttng-traces/calibrate-function-* | tail \-n 1)
161 The output from babeltrace can be saved to a text file and opened in a
162 spreadsheet (e.g. oocalc) to focus on the per-PMU counter delta between
163 consecutive "calibrate_entry" and "calibrate_return" events. Note that these
164 counters are per-CPU, so scheduling events would need to be present to account
165 for migration between CPU. Therefore, for calibration purposes, only events
166 staying on the same CPU must be considered.
168 The average result, for the i7, on 10 samples:
171 perf_LLC_load_misses: 5.0 0.577
172 perf_LLC_store_misses: 1.6 0.516
173 perf_LLC_prefetch_misses: 9.0 14.742
175 As we can notice, the load and store misses are relatively stable across runs
176 (their standard deviation is relatively low) compared to the prefetch misses.
177 We can conclude from this information that LLC load and store misses can be
178 accounted for quite precisely, but prefetches within a function seems to behave
179 too erratically (not much causality link between the code executed and the CPU
180 prefetch activity) to be accounted for.
187 Show summary of possible options and commands.
189 Apply for the kernel tracer
191 Apply for the user-space tracer
193 Dynamic function entry/return probe (default)
198 .IP "\fBcreate\fP [OPTIONS] [NAME]
200 Create tracing session.
202 A tracing session contains channel(s) which contains event(s). It is domain
203 agnostic meaning that you can enable channels and events for either the
204 user-space tracer and/or the kernel tracer. It acts like a container
205 aggregating multiple tracing sources.
207 On creation, a \fB.lttngrc\fP file is created in your $HOME directory
208 containing the current session name. If NAME is omitted, a session name is
209 automatically created having this form: 'auto-yyyymmdd-hhmmss'.
211 If no \fB\-o, \-\-output\fP is specified, the traces will be written in
219 Show summary of possible options and commands.
221 Simple listing of options
223 Specify output path for traces
228 .IP "\fBdestroy\fP [OPTIONS] [NAME]"
230 Teardown tracing session
232 Free memory on the session daemon and tracer side. It's gone!
234 If NAME is omitted, the session name is taken from the .lttngrc file.
241 Show summary of possible options and commands.
245 Simple listing of options
250 .IP "\fBenable-channel\fP NAME[,NAME2,...] [-k|-u] [OPTIONS]"
252 Enable tracing channel
254 To enable event, you must first enable a channel which contains event(s).
256 If \fB\-s, \-\-session\fP is omitted, the session name is taken from the .lttngrc
266 Simple listing of options
268 Apply on session name
270 Apply to the kernel tracer
272 Apply to the user-space tracer
275 Discard event when subbuffers are full (default)
277 Flight recorder mode : overwrites events when subbuffers are full
279 Subbuffer size in bytes (default: 4096, kernel default: 262144)
281 Number of subbuffers (default: 4)
282 Needs to be a power of 2 for kernel and ust tracers
284 Switch subbuffer timer interval in usec (default: 0)
285 Needs to be a power of 2 for kernel and ust tracers
287 Read timer interval in usec (default: 200)
292 .IP "\fBenable-event\fP NAME[,NAME2,...] [-k|-u] [OPTIONS]"
296 A tracing event is always assigned to a channel. If \fB\-c, \-\-channel\fP is
297 omitted, a default channel named '\fBchannel0\fP' is created and the event is
298 added to it. For the user-space tracer, using \fB\-a, \-\-all\fP is the same as
299 using the wildcard "*".
301 If \fB\-s, \-\-session\fP is omitted, the session name is taken from the .lttngrc
309 Show summary of possible options and commands.
311 Simple listing of options
313 Apply on session name
315 Apply on channel name
317 Enable all tracepoints and syscalls
319 Apply for the kernel tracer
321 Apply for the user-space tracer
324 Tracepoint event (default)
325 - userspace tracer supports wildcards at end of string. Don't forget to
326 quote to deal with bash expansion.
332 \-\-probe [addr | symbol | symbol+offset]
333 Dynamic probe. Addr and offset can be octal (0NNN...), decimal (NNN...)
334 or hexadecimal (0xNNN...)
335 \-\-function [addr | symbol | symbol+offset]
336 Dynamic function entry/return probe. Addr and offset can be octal
337 (0NNN...), decimal (NNN...) or hexadecimal (0xNNN...)
340 Enabling syscalls tracing (kernel tracer), you will not be able to disable them
341 with disable-event. This is a known limitation. You can disable the entire
342 channel to do the trick.
343 \-\-filter 'expression'
344 Set a filter on a newly enabled event.
345 Filter expression on event fields event recording
346 depends on evaluation. Only specify on first activation
347 of a given event within a session. Filter only allowed
348 when enabling events within a session before tracing is
352 .IP "\fBdisable-channel\fP NAME[,NAME2,...] [\-k|\-u] [OPTIONS]"
354 Disable tracing channel
356 Disabling a channel makes all event(s) in that channel to stop tracing. You can
357 enable it back by calling \fBlttng enable-channel NAME\fP again.
359 If \fB\-s, \-\-session\fP is omitted, the session name is taken from the .lttngrc
367 Show summary of possible options and commands.
369 Simple listing of options
371 Apply on session name
373 Apply for the kernel tracer
375 Apply for the user-space tracer
378 .IP "\fBdisable-event\fP NAME[,NAME2,...] [\-k|\-u] [OPTIONS]"
380 Disable tracing event
382 The event, once disabled, can be re-enabled by calling \fBlttng enable-event
385 If \fB\-s, \-\-session\fP is omitted, the session name is taken from the .lttngrc
393 Show summary of possible options and commands.
395 Simple listing of options
397 Apply on session name
399 Apply for the kernel tracer
401 Apply for the user-space tracer
404 .IP "\fBlist\fP [\-k|\-u] [SESSION [SESSION_OPTIONS]]"
406 List tracing session information.
408 With no arguments, it will list available tracing session(s).
410 With the session name, it will display the details of the session including
411 the trace file path, the associated channels and their state (activated
412 and deactivated), the activated events and more.
414 With \-k alone, it will list all available kernel events (except the system
416 With \-u alone, it will list all available user-space events from registered
417 applications. Here is an example of 'lttng list \-u':
419 PID: 7448 - Name: /tmp/lttng-ust/tests/hello/.libs/lt-hello
420 ust_tests_hello:tptest_sighandler (type: tracepoint)
421 ust_tests_hello:tptest (type: tracepoint)
423 You can now enable any event listed by using the name :
424 \fBust_tests_hello:tptest\fP.
431 Show summary of possible options and commands.
433 Simple listing of options
437 Select user-space domain.
440 \-c, \-\-channel NAME
441 List details of a channel
443 List available domain(s)
446 .IP "\fBset-session\fP NAME"
448 Set current session name
450 Will change the session name in the .lttngrc file.
457 Show summary of possible options and commands.
459 Simple listing of options
464 .IP "\fBstart\fP [OPTIONS] [NAME]"
468 It will start tracing for all tracers for a specific tracing session.
470 If NAME is omitted, the session name is taken from the .lttngrc file.
477 Show summary of possible options and commands.
479 Simple listing of options
484 .IP "\fBstop\fP [OPTIONS] [NAME]"
488 It will stop tracing for all tracers for a specific tracing session.
490 If NAME is omitted, the session name is taken from the .lttngrc file.
497 Show summary of possible options and commands.
499 Simple listing of options
506 Show version information
513 Show summary of possible options and commands.
515 Simple listing of options
520 .IP "\fBview\fP [SESSION_NAME] [OPTIONS]"
522 View traces of a tracing session
524 By default, the babeltrace viewer will be used for text viewing.
526 If SESSION_NAME is omitted, the session name is taken from the .lttngrc file.
536 Simple listing of options
537 \-t, \-\-trace-path PATH
538 Trace directory path for the viewer
540 Specify viewer and/or options to use
541 This will completely override the default viewers so
542 please make sure to specify the full command. The trace
543 directory path of the session will be appended at the end
565 No session found by the name given
568 Error in session creation
571 Error in application(s) listing
574 Session name already exists
577 Kernel tracer unavailable
583 Kernel channel exists
586 Kernel channel creation failed
589 Kernel channel not found
592 Kernel channel disable failed
595 Kernel channel enable failed
598 Kernel context failed
601 Kernel enable event failed
604 Kernel disable event failed
607 Kernel listing events failed
610 UST channel disable failed
613 UST channel enable failed
616 UST adding context failed
619 UST event enable failed
622 UST event disable failed
643 Tracing the kernel requires a root lttng-sessiond daemon and "tracing" group
647 Tracing already started
650 Tracing already stopped
653 No UST consumer detected
656 No Kernel consumer detected
659 Event already enabled with different loglevel
661 .SH "ENVIRONMENT VARIABLES"
664 Note that all command line options override environment variables.
668 .IP "LTTNG_SESSIOND_PATH"
669 Allows one to specify the full session daemon binary path to lttng command line
670 tool. You can also use \-\-sessiond-path option having the same effect.
674 babeltrace(1), lttng-ust(3), lttng-sessiond(8)
679 No show stopper bugs are known yet in this version.
681 If you encounter any issues or usability problem, please report it on our
682 mailing list <lttng-dev@lists.lttng.org> to help improve this project.
686 lttng is distributed under the GNU General Public License version 2. See the file
689 A Web site is available at http://lttng.org for more information on the LTTng
692 You can also find our git tree at http://git.lttng.org.
694 Mailing lists for support and development: <lttng-dev@lists.lttng.org>.
696 You can find us on IRC server irc.oftc.net (OFTC) in #lttng.
701 Thanks to Yannick Brosseau without whom this project would never have been so
702 lean and mean! Also thanks to the Ericsson teams working on tracing which
703 helped us greatly with detailed bug reports and unusual test cases.
705 Thanks to our beloved packager Alexandre Montplaisir-Goncalves (Ubuntu and PPA
706 maintainer) and Jon Bernard for our Debian packages.
708 Special thanks to Michel Dagenais and the DORSAL laboratory at Polytechnique de
709 Montreal for the LTTng journey.
714 lttng-tools was originally written by Mathieu Desnoyers, Julien Desfossez and
715 David Goulet. More people have since contributed to it. It is currently
716 maintained by David Goulet <dgoulet@efficios.com>.