1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng{nbsp}{revision}?
77 LTTng{nbsp}{revision} bears the name _Lafontaine_. This modern
78 https://en.wikipedia.org/wiki/saison[saison] from the
79 https://oshlag.com/[Oshlag] microbrewery is a refreshing--zesty--rice
80 beer with hints of fruit and spices. Some even say it makes for a great
81 https://en.wikipedia.org/wiki/Somaek[Somaek] when mixed with
82 Chamisul Soju, not that we've tried!
84 New features and changes in LTTng{nbsp}{revision}:
86 * Just like you can typically perform
87 https://en.wikipedia.org/wiki/Log_rotation[log rotation], you can
88 now <<session-rotation,_rotate_ a tracing session>>, that
89 is, according to man:lttng-rotate(1), archive the current trace
90 chunk (all the tracing session's trace data since the last rotation
91 or since its inception) so that LTTng does not manage it anymore.
93 Once LTTng archives a trace chunk, you are free to read it, modify it,
94 move it, or remove it.
96 You can rotate a tracing session immediately or set a rotation schedule
97 to automate rotations.
99 * When you <<enabling-disabling-events,create an event rule>>, the
100 filter expression syntax now supports the following new operators:
104 ** `<<` (bitwise left shift)
105 ** `>>` (bitwise right shift)
111 The syntax also supports array indexing with the usual square brackets:
114 regs[3][1] & 0xff7 == 0x240
117 There are peculiarities for both the new operators and the array
118 indexing brackets, like a custom precedence table and implicit casting.
119 See man:lttng-enable-event(1) to get all the details about the filter
122 * You can now dynamically instrument any application's or library's
123 function entry by symbol name thanks to the new
124 opt:lttng-enable-event(1):--userspace-probe option of
125 the `lttng enable-event` command:
129 $ lttng enable-event --kernel \
130 --userspace-probe=/usr/lib/libc.so.6:malloc libc_malloc
133 The option also supports tracing existing
134 https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
135 Statically Defined Tracing] (USDT) probe (DTrace-style marker). For
136 example, given the following probe:
140 DTRACE_PROBE2("server", "accept-request", request_id, ip_addr);
143 You can trace this probe with:
146 $ lttng enable-event --kernel \
147 --userspace-probe=sdt:/path/to/server:server:accept-request \
148 server_accept_request
151 This feature makes use of Linux's
152 https://www.kernel.org/doc/Documentation/trace/uprobetracer.txt[uprobe]
153 mechanism, therefore you must use the `--userspace-probe`
154 instrumentation option with the opt:lttng-enable-event(1):--kernel
157 NOTE: As of LTTng{nbsp}{revision}, LTTng does not record function
158 parameters with the opt:lttng-enable-event(1):--userspace-probe option.
160 * Two new <<adding-context,context>> fields are available for Linux
161 kernel <<channel,channels>>:
164 ** `callstack-kernel`
168 Thanks to those, you can record the Linux kernel and user call stacks
169 when a kernel event occurs. For example:
173 $ lttng enable-event --kernel --syscall read
174 $ lttng add-context --kernel --type=callstack-kernel --type=callstack-user
177 When an man:open(2) system call occurs, LTTng attaches the kernel and
178 user call stacks to the recorded event.
180 NOTE: LTTng cannot always sample the user space call stack reliably.
181 For instance, LTTng cannot sample the call stack of user applications
182 and libraries compiled with the
183 https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html[`-fomit-frame-pointer`]
184 option. In such a case, the tracing is not affected, but the sampled
185 user space call stack may only contain the user call stack's topmost
188 * User applications and libraries instrumented with
189 <<lttng-ust,LTTng-UST>> can now safely unload (man:dlclose(3)) a
191 <<building-tracepoint-providers-and-user-application,tracepoint
194 * The <<lttng-relayd,relay daemon>> is more efficient and presents fewer
195 connectivity issues, especially when a large number of targets send
196 trace data to a given relay daemon.
198 * LTTng-UST uses https://github.com/numactl/numactl[libnuma]
199 when available to allocate <<def-sub-buffer,sub-buffers>>, making them
201 https://en.wikipedia.org/wiki/Non-uniform_memory_access[NUMA] node.
203 This change makes the tracer more efficient on NUMA systems.
209 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
210 generation_ is a modern toolkit for tracing Linux systems and
211 applications. So your first question might be:
218 As the history of software engineering progressed and led to what
219 we now take for granted--complex, numerous and
220 interdependent software applications running in parallel on
221 sophisticated operating systems like Linux--the authors of such
222 components, software developers, began feeling a natural
223 urge to have tools that would ensure the robustness and good performance
224 of their masterpieces.
226 One major achievement in this field is, inarguably, the
227 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
228 an essential tool for developers to find and fix bugs. But even the best
229 debugger won't help make your software run faster, and nowadays, faster
230 software means either more work done by the same hardware, or cheaper
231 hardware for the same work.
233 A _profiler_ is often the tool of choice to identify performance
234 bottlenecks. Profiling is suitable to identify _where_ performance is
235 lost in a given software. The profiler outputs a profile, a statistical
236 summary of observed events, which you may use to discover which
237 functions took the most time to execute. However, a profiler won't
238 report _why_ some identified functions are the bottleneck. Bottlenecks
239 might only occur when specific conditions are met, conditions that are
240 sometimes impossible to capture by a statistical profiler, or impossible
241 to reproduce with an application altered by the overhead of an
242 event-based profiler. For a thorough investigation of software
243 performance issues, a history of execution is essential, with the
244 recorded values of variables and context fields you choose, and
245 with as little influence as possible on the instrumented software. This
246 is where tracing comes in handy.
248 _Tracing_ is a technique used to understand what goes on in a running
249 software system. The software used for tracing is called a _tracer_,
250 which is conceptually similar to a tape recorder. When recording,
251 specific instrumentation points placed in the software source code
252 generate events that are saved on a giant tape: a _trace_ file. You
253 can trace user applications and the operating system at the same time,
254 opening the possibility of resolving a wide range of problems that would
255 otherwise be extremely challenging.
257 Tracing is often compared to _logging_. However, tracers and loggers are
258 two different tools, serving two different purposes. Tracers are
259 designed to record much lower-level events that occur much more
260 frequently than log messages, often in the range of thousands per
261 second, with very little execution overhead. Logging is more appropriate
262 for a very high-level analysis of less frequent events: user accesses,
263 exceptional conditions (errors and warnings, for example), database
264 transactions, instant messaging communications, and such. Simply put,
265 logging is one of the many use cases that can be satisfied with tracing.
267 The list of recorded events inside a trace file can be read manually
268 like a log file for the maximum level of detail, but it is generally
269 much more interesting to perform application-specific analyses to
270 produce reduced statistics and graphs that are useful to resolve a
271 given problem. Trace viewers and analyzers are specialized tools
274 In the end, this is what LTTng is: a powerful, open source set of
275 tools to trace the Linux kernel and user applications at the same time.
276 LTTng is composed of several components actively maintained and
277 developed by its link:/community/#where[community].
280 [[lttng-alternatives]]
281 === Alternatives to noch:{LTTng}
283 Excluding proprietary solutions, a few competing software tracers
286 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
287 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
288 user scripts and is responsible for loading code into the
289 Linux kernel for further execution and collecting the outputted data.
290 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
291 subsystem in the Linux kernel in which a virtual machine can execute
292 programs passed from the user space to the kernel. You can attach
293 such programs to tracepoints and kprobes thanks to a system call, and
294 they can output data to the user space when executed thanks to
295 different mechanisms (pipe, VM register values, and eBPF maps, to name
297 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
298 is the de facto function tracer of the Linux kernel. Its user
299 interface is a set of special files in sysfs.
300 * https://perf.wiki.kernel.org/[perf] is
301 a performance analyzing tool for Linux which supports hardware
302 performance counters, tracepoints, as well as other counters and
303 types of probes. perf's controlling utility is the cmd:perf command
305 * http://linux.die.net/man/1/strace[strace]
306 is a command-line utility which records system calls made by a
307 user process, as well as signal deliveries and changes of process
308 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
309 to fulfill its function.
310 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
311 analyze Linux kernel events. You write scripts, or _chisels_ in
312 sysdig's jargon, in Lua and sysdig executes them while it traces the
313 system or afterwards. sysdig's interface is the cmd:sysdig
314 command-line tool as well as the curses-based cmd:csysdig tool.
315 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
316 user space tracer which uses custom user scripts to produce plain text
317 traces. SystemTap converts the scripts to the C language, and then
318 compiles them as Linux kernel modules which are loaded to produce
319 trace data. SystemTap's primary user interface is the cmd:stap
322 The main distinctive features of LTTng is that it produces correlated
323 kernel and user space traces, as well as doing so with the lowest
324 overhead amongst other solutions. It produces trace files in the
325 http://diamon.org/ctf[CTF] format, a file format optimized
326 for the production and analyses of multi-gigabyte data.
328 LTTng is the result of more than 10{nbsp}years of active open source
329 development by a community of passionate developers.
330 LTTng{nbsp}{revision} is currently available on major desktop and server
333 The main interface for tracing control is a single command-line tool
334 named cmd:lttng. The latter can create several tracing sessions, enable
335 and disable events on the fly, filter events efficiently with custom
336 user expressions, start and stop tracing, and much more. LTTng can
337 record the traces on the file system or send them over the network, and
338 keep them totally or partially. You can view the traces once tracing
339 becomes inactive or in real-time.
341 <<installing-lttng,Install LTTng now>> and
342 <<getting-started,start tracing>>!
348 **LTTng** is a set of software <<plumbing,components>> which interact to
349 <<instrumenting,instrument>> the Linux kernel and user applications, and
350 to <<controlling-tracing,control tracing>> (start and stop
351 tracing, enable and disable event rules, and the rest). Those
352 components are bundled into the following packages:
355 Libraries and command-line interface to control tracing.
358 Linux kernel modules to instrument and trace the kernel.
361 Libraries and Java/Python packages to instrument and trace user
364 Most distributions mark the LTTng-modules and LTTng-UST packages as
365 optional when installing LTTng-tools (which is always required). In the
366 following sections, we always provide the steps to install all three,
369 * You only need to install LTTng-modules if you intend to trace the
371 * You only need to install LTTng-UST if you intend to trace user
376 As of 22 October 2019, LTTng{nbsp}{revision} is not available
377 as distribution packages, except for <<arch-linux,Arch Linux>>.
379 You can <<building-from-source,build LTTng{nbsp}{revision} from source>>
380 to install and use it.
387 LTTng-UST{nbsp}{revision} is available in Arch Linux's _community_
388 repository, while LTTng-tools{nbsp}{revision} and
389 LTTng-modules{nbsp}{revision} are available in the
390 https://aur.archlinux.org/[AUR].
392 To install LTTng{nbsp}{revision} on Arch Linux, using
393 https://github.com/actionless/pikaur[pikaur] for the AUR packages:
395 . Install the main LTTng{nbsp}{revision} packages:
400 # pacman -Sy lttng-ust
401 $ pikaur -Sy lttng-tools
402 $ pikaur -Sy lttng-modules
406 . **If you need to instrument and trace <<python-application,Python
407 applications>>**, install the LTTng-UST Python agent:
412 # pacman -Sy python-lttngust
413 # pacman -Sy python2-lttngust
418 [[building-from-source]]
419 === Build from source
421 To build and install LTTng{nbsp}{revision} from source:
423 . Using your distribution's package manager, or from source, install
424 the following dependencies of LTTng-tools and LTTng-UST:
427 * https://sourceforge.net/projects/libuuid/[libuuid]
428 * http://directory.fsf.org/wiki/Popt[popt]
429 * http://liburcu.org/[Userspace RCU]
430 * http://www.xmlsoft.org/[libxml2]
431 * **Optional**: https://github.com/numactl/numactl[numactl]
434 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
440 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
441 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
442 cd lttng-modules-2.11.* &&
444 sudo make modules_install &&
449 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
455 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.11.tar.bz2 &&
456 tar -xf lttng-ust-latest-2.11.tar.bz2 &&
457 cd lttng-ust-2.11.* &&
465 Add `--disable-numa` to `./configure` if you don't have
466 https://github.com/numactl/numactl[numactl].
470 .Java and Python application tracing
472 If you need to instrument and trace <<java-application,Java
473 applications>>, pass the `--enable-java-agent-jul`,
474 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
475 `configure` script, depending on which Java logging framework you use.
477 If you need to instrument and trace <<python-application,Python
478 applications>>, pass the `--enable-python-agent` option to the
479 `configure` script. You can set the `PYTHON` environment variable to the
480 path to the Python interpreter for which to install the LTTng-UST Python
488 By default, LTTng-UST libraries are installed to
489 dir:{/usr/local/lib}, which is the de facto directory in which to
490 keep self-compiled and third-party libraries.
492 When <<building-tracepoint-providers-and-user-application,linking an
493 instrumented user application with `liblttng-ust`>>:
495 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
497 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
498 man:gcc(1), man:g++(1), or man:clang(1).
502 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
508 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
509 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
510 cd lttng-tools-2.11.* &&
518 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
519 previous steps automatically for a given version of LTTng and confine
520 the installed files in a specific directory. This can be useful to test
521 LTTng without installing it on your system.
527 This is a short guide to get started quickly with LTTng kernel and user
530 Before you follow this guide, make sure to <<installing-lttng,install>>
533 This tutorial walks you through the steps to:
535 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
536 . <<tracing-your-own-user-application,Trace a user application>> written
538 . <<viewing-and-analyzing-your-traces,View and analyze the
542 [[tracing-the-linux-kernel]]
543 === Trace the Linux kernel
545 The following command lines start with the `#` prompt because you need
546 root privileges to trace the Linux kernel. You can also trace the kernel
547 as a regular user if your Unix user is a member of the
548 <<tracing-group,tracing group>>.
550 . Create a <<tracing-session,tracing session>> which writes its traces
551 to dir:{/tmp/my-kernel-trace}:
556 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
560 . List the available kernel tracepoints and system calls:
565 # lttng list --kernel
566 # lttng list --kernel --syscall
570 . Create <<event,event rules>> which match the desired instrumentation
571 point names, for example the `sched_switch` and `sched_process_fork`
572 tracepoints, and the man:open(2) and man:close(2) system calls:
577 # lttng enable-event --kernel sched_switch,sched_process_fork
578 # lttng enable-event --kernel --syscall open,close
582 You can also create an event rule which matches _all_ the Linux kernel
583 tracepoints (this will generate a lot of data when tracing):
588 # lttng enable-event --kernel --all
592 . <<basic-tracing-session-control,Start tracing>>:
601 . Do some operation on your system for a few seconds. For example,
602 load a website, or list the files of a directory.
603 . <<creating-destroying-tracing-sessions,Destroy>> the current
613 The man:lttng-destroy(1) command does not destroy the trace data; it
614 only destroys the state of the tracing session.
616 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
617 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
618 session>>). You need to stop tracing to make LTTng flush the remaining
619 trace data and make the trace readable.
621 . For the sake of this example, make the recorded trace accessible to
627 # chown -R $(whoami) /tmp/my-kernel-trace
631 See <<viewing-and-analyzing-your-traces,View and analyze the
632 recorded events>> to view the recorded events.
635 [[tracing-your-own-user-application]]
636 === Trace a user application
638 This section steps you through a simple example to trace a
639 _Hello world_ program written in C.
641 To create the traceable user application:
643 . Create the tracepoint provider header file, which defines the
644 tracepoints and the events they can generate:
650 #undef TRACEPOINT_PROVIDER
651 #define TRACEPOINT_PROVIDER hello_world
653 #undef TRACEPOINT_INCLUDE
654 #define TRACEPOINT_INCLUDE "./hello-tp.h"
656 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
659 #include <lttng/tracepoint.h>
669 ctf_string(my_string_field, my_string_arg)
670 ctf_integer(int, my_integer_field, my_integer_arg)
674 #endif /* _HELLO_TP_H */
676 #include <lttng/tracepoint-event.h>
680 . Create the tracepoint provider package source file:
686 #define TRACEPOINT_CREATE_PROBES
687 #define TRACEPOINT_DEFINE
689 #include "hello-tp.h"
693 . Build the tracepoint provider package:
698 $ gcc -c -I. hello-tp.c
702 . Create the _Hello World_ application source file:
709 #include "hello-tp.h"
711 int main(int argc, char *argv[])
715 puts("Hello, World!\nPress Enter to continue...");
718 * The following getchar() call is only placed here for the purpose
719 * of this demonstration, to pause the application in order for
720 * you to have time to list its tracepoints. It is not
726 * A tracepoint() call.
728 * Arguments, as defined in hello-tp.h:
730 * 1. Tracepoint provider name (required)
731 * 2. Tracepoint name (required)
732 * 3. my_integer_arg (first user-defined argument)
733 * 4. my_string_arg (second user-defined argument)
735 * Notice the tracepoint provider and tracepoint names are
736 * NOT strings: they are in fact parts of variables that the
737 * macros in hello-tp.h create.
739 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
741 for (x = 0; x < argc; ++x) {
742 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
745 puts("Quitting now!");
746 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
753 . Build the application:
762 . Link the application with the tracepoint provider package,
763 `liblttng-ust`, and `libdl`:
768 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
772 Here's the whole build process:
775 .User space tracing tutorial's build steps.
776 image::ust-flow.png[]
778 To trace the user application:
780 . Run the application with a few arguments:
785 $ ./hello world and beyond
794 Press Enter to continue...
798 . Start an LTTng <<lttng-sessiond,session daemon>>:
803 $ lttng-sessiond --daemonize
807 Note that a session daemon might already be running, for example as
808 a service that the distribution's service manager started.
810 . List the available user space tracepoints:
815 $ lttng list --userspace
819 You see the `hello_world:my_first_tracepoint` tracepoint listed
820 under the `./hello` process.
822 . Create a <<tracing-session,tracing session>>:
827 $ lttng create my-user-space-session
831 . Create an <<event,event rule>> which matches the
832 `hello_world:my_first_tracepoint` event name:
837 $ lttng enable-event --userspace hello_world:my_first_tracepoint
841 . <<basic-tracing-session-control,Start tracing>>:
850 . Go back to the running `hello` application and press Enter. The
851 program executes all `tracepoint()` instrumentation points and exits.
852 . <<creating-destroying-tracing-sessions,Destroy>> the current
862 The man:lttng-destroy(1) command does not destroy the trace data; it
863 only destroys the state of the tracing session.
865 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
866 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
867 session>>). You need to stop tracing to make LTTng flush the remaining
868 trace data and make the trace readable.
870 By default, LTTng saves the traces in
871 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
872 where +__name__+ is the tracing session name. The
873 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
875 See <<viewing-and-analyzing-your-traces,View and analyze the
876 recorded events>> to view the recorded events.
879 [[viewing-and-analyzing-your-traces]]
880 === View and analyze the recorded events
882 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
883 kernel>> and <<tracing-your-own-user-application,Trace a user
884 application>> tutorials, you can inspect the recorded events.
886 There are many tools you can use to read LTTng traces:
888 * **cmd:babeltrace** is a command-line utility which converts trace
889 formats; it supports the format that LTTng produces, CTF, as well as a
890 basic text output which can be ++grep++ed. The cmd:babeltrace command
891 is part of the http://diamon.org/babeltrace[Babeltrace] project.
892 * Babeltrace also includes
893 **https://www.python.org/[Python{nbsp}3] bindings** so
894 that you can easily open and read an LTTng trace with your own script,
895 benefiting from the power of Python.
896 * http://tracecompass.org/[**Trace Compass**]
897 is a graphical user interface for viewing and analyzing any type of
898 logs or traces, including LTTng's.
899 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
900 project which includes many high-level analyses of LTTng kernel
901 traces, like scheduling statistics, interrupt frequency distribution,
902 top CPU usage, and more.
904 NOTE: This section assumes that LTTng saved the traces it recorded
905 during the previous tutorials to their default location, in the
906 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
907 environment variable defaults to `$HOME` if not set.
910 [[viewing-and-analyzing-your-traces-bt]]
911 ==== Use the cmd:babeltrace command-line tool
913 The simplest way to list all the recorded events of a trace is to pass
914 its path to cmd:babeltrace with no options:
918 $ babeltrace ~/lttng-traces/my-user-space-session*
921 cmd:babeltrace finds all traces recursively within the given path and
922 prints all their events, merging them in chronological order.
924 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
929 $ babeltrace /tmp/my-kernel-trace | grep _switch
932 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
933 count the recorded events:
937 $ babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
941 [[viewing-and-analyzing-your-traces-bt-python]]
942 ==== Use the Babeltrace{nbsp}1 Python bindings
944 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
945 is useful to isolate events by simple matching using man:grep(1) and
946 similar utilities. However, more elaborate filters, such as keeping only
947 event records with a field value falling within a specific range, are
948 not trivial to write using a shell. Moreover, reductions and even the
949 most basic computations involving multiple event records are virtually
950 impossible to implement.
952 Fortunately, Babeltrace{nbsp}1 ships with Python{nbsp}3 bindings which
953 makes it easy to read the event records of an LTTng trace sequentially
954 and compute the desired information.
956 The following script accepts an LTTng Linux kernel trace path as its
957 first argument and prints the short names of the top five running
958 processes on CPU{nbsp}0 during the whole trace:
963 from collections import Counter
969 if len(sys.argv) != 2:
970 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
971 print(msg, file=sys.stderr)
974 # A trace collection contains one or more traces
975 col = babeltrace.TraceCollection()
977 # Add the trace provided by the user (LTTng traces always have
979 if col.add_trace(sys.argv[1], 'ctf') is None:
980 raise RuntimeError('Cannot add trace')
982 # This counter dict contains execution times:
984 # task command name -> total execution time (ns)
985 exec_times = Counter()
987 # This contains the last `sched_switch` timestamp
991 for event in col.events:
992 # Keep only `sched_switch` events
993 if event.name != 'sched_switch':
996 # Keep only events which happened on CPU 0
997 if event['cpu_id'] != 0:
1001 cur_ts = event.timestamp
1007 # Previous task command (short) name
1008 prev_comm = event['prev_comm']
1010 # Initialize entry in our dict if not yet done
1011 if prev_comm not in exec_times:
1012 exec_times[prev_comm] = 0
1014 # Compute previous command execution time
1015 diff = cur_ts - last_ts
1017 # Update execution time of this command
1018 exec_times[prev_comm] += diff
1020 # Update last timestamp
1024 for name, ns in exec_times.most_common(5):
1026 print('{:20}{} s'.format(name, s))
1031 if __name__ == '__main__':
1032 sys.exit(0 if top5proc() else 1)
1039 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1045 swapper/0 48.607245889 s
1046 chromium 7.192738188 s
1047 pavucontrol 0.709894415 s
1048 Compositor 0.660867933 s
1049 Xorg.bin 0.616753786 s
1052 Note that `swapper/0` is the "idle" process of CPU{nbsp}0 on Linux;
1053 since we weren't using the CPU that much when tracing, its first
1054 position in the list makes sense.
1058 == [[understanding-lttng]]Core concepts
1060 From a user's perspective, the LTTng system is built on a few concepts,
1061 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1062 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1063 Understanding how those objects relate to eachother is key in mastering
1066 The core concepts are:
1068 * <<tracing-session,Tracing session>>
1069 * <<domain,Tracing domain>>
1070 * <<channel,Channel and ring buffer>>
1071 * <<"event","Instrumentation point, event rule, event, and event record">>
1077 A _tracing session_ is a stateful dialogue between you and
1078 a <<lttng-sessiond,session daemon>>. You can
1079 <<creating-destroying-tracing-sessions,create a new tracing
1080 session>> with the `lttng create` command.
1082 Anything that you do when you control LTTng tracers happens within a
1083 tracing session. In particular, a tracing session:
1086 * Has its own set of trace files.
1087 * Has its own state of activity (started or stopped).
1088 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1090 * Has its own <<channel,channels>> to which are associated their own
1091 <<event,event rules>>.
1094 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1095 image::concepts.png[]
1097 Those attributes and objects are completely isolated between different
1100 A tracing session is analogous to a cash machine session:
1101 the operations you do on the banking system through the cash machine do
1102 not alter the data of other users of the same system. In the case of
1103 the cash machine, a session lasts as long as your bank card is inside.
1104 In the case of LTTng, a tracing session lasts from the `lttng create`
1105 command to the `lttng destroy` command.
1108 .Each Unix user has its own set of tracing sessions.
1109 image::many-sessions.png[]
1112 [[tracing-session-mode]]
1113 ==== Tracing session mode
1115 LTTng can send the generated trace data to different locations. The
1116 _tracing session mode_ dictates where to send it. The following modes
1117 are available in LTTng{nbsp}{revision}:
1120 LTTng writes the traces to the file system of the machine it traces
1123 Network streaming mode::
1124 LTTng sends the traces over the network to a
1125 <<lttng-relayd,relay daemon>> running on a remote system.
1128 LTTng does not write the traces by default. Instead, you can request
1129 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1130 tracing session's current sub-buffers, and to write it to the
1131 target's file system or to send it over the network to a
1132 <<lttng-relayd,relay daemon>> running on a remote system.
1135 This mode is similar to the network streaming mode, but a live
1136 trace viewer can connect to the distant relay daemon to
1137 <<lttng-live,view event records as LTTng generates them>>.
1143 A _tracing domain_ is a namespace for event sources. A tracing domain
1144 has its own properties and features.
1146 There are currently five available tracing domains:
1150 * `java.util.logging` (JUL)
1154 You must specify a tracing domain when using some commands to avoid
1155 ambiguity. For example, since all the domains support named tracepoints
1156 as event sources (instrumentation points that you manually insert in the
1157 source code), you need to specify a tracing domain when
1158 <<enabling-disabling-events,creating an event rule>> because all the
1159 tracing domains could have tracepoints with the same names.
1161 You can create <<channel,channels>> in the Linux kernel and user space
1162 tracing domains. The other tracing domains have a single default
1167 === Channel and ring buffer
1169 A _channel_ is an object which is responsible for a set of ring buffers.
1170 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1171 tracer emits an event, it can record it to one or more
1172 sub-buffers. The attributes of a channel determine what to do when
1173 there's no space left for a new event record because all sub-buffers
1174 are full, where to send a full sub-buffer, and other behaviours.
1176 A channel is always associated to a <<domain,tracing domain>>. The
1177 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1178 a default channel which you cannot configure.
1180 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1181 an event, it records it to the sub-buffers of all
1182 the enabled channels with a satisfied event rule, as long as those
1183 channels are part of active <<tracing-session,tracing sessions>>.
1186 [[channel-buffering-schemes]]
1187 ==== Per-user vs. per-process buffering schemes
1189 A channel has at least one ring buffer _per CPU_. LTTng always
1190 records an event to the ring buffer associated to the CPU on which it
1193 Two _buffering schemes_ are available when you
1194 <<enabling-disabling-channels,create a channel>> in the
1195 user space <<domain,tracing domain>>:
1197 Per-user buffering::
1198 Allocate one set of ring buffers--one per CPU--shared by all the
1199 instrumented processes of each Unix user.
1203 .Per-user buffering scheme.
1204 image::per-user-buffering.png[]
1207 Per-process buffering::
1208 Allocate one set of ring buffers--one per CPU--for each
1209 instrumented process.
1213 .Per-process buffering scheme.
1214 image::per-process-buffering.png[]
1217 The per-process buffering scheme tends to consume more memory than the
1218 per-user option because systems generally have more instrumented
1219 processes than Unix users running instrumented processes. However, the
1220 per-process buffering scheme ensures that one process having a high
1221 event throughput won't fill all the shared sub-buffers of the same
1224 The Linux kernel tracing domain has only one available buffering scheme
1225 which is to allocate a single set of ring buffers for the whole system.
1226 This scheme is similar to the per-user option, but with a single, global
1227 user "running" the kernel.
1230 [[channel-overwrite-mode-vs-discard-mode]]
1231 ==== Overwrite vs. discard event record loss modes
1233 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1234 arc in the following animations) of a specific channel's ring buffer.
1235 When there's no space left in a sub-buffer, the tracer marks it as
1236 consumable (red) and another, empty sub-buffer starts receiving the
1237 following event records. A <<lttng-consumerd,consumer daemon>>
1238 eventually consumes the marked sub-buffer (returns to white).
1241 [role="docsvg-channel-subbuf-anim"]
1246 In an ideal world, sub-buffers are consumed faster than they are filled,
1247 as it is the case in the previous animation. In the real world,
1248 however, all sub-buffers can be full at some point, leaving no space to
1249 record the following events.
1251 By default, LTTng-modules and LTTng-UST are _non-blocking_ tracers: when
1252 no empty sub-buffer is available, it is acceptable to lose event records
1253 when the alternative would be to cause substantial delays in the
1254 instrumented application's execution. LTTng privileges performance over
1255 integrity; it aims at perturbing the target system as little as possible
1256 in order to make tracing of subtle race conditions and rare interrupt
1259 Since LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST, supports
1260 a _blocking mode_. See the <<blocking-timeout-example,blocking timeout
1261 example>> to learn how to use the blocking mode.
1263 When it comes to losing event records because no empty sub-buffer is
1264 available, or because the <<opt-blocking-timeout,blocking timeout>> is
1265 reached, the channel's _event record loss mode_ determines what to do.
1266 The available event record loss modes are:
1269 Drop the newest event records until a the tracer releases a
1272 This is the only available mode when you specify a
1273 <<opt-blocking-timeout,blocking timeout>>.
1276 Clear the sub-buffer containing the oldest event records and start
1277 writing the newest event records there.
1279 This mode is sometimes called _flight recorder mode_ because it's
1281 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1282 always keep a fixed amount of the latest data.
1284 Which mechanism you should choose depends on your context: prioritize
1285 the newest or the oldest event records in the ring buffer?
1287 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1288 as soon as a there's no space left for a new event record, whereas in
1289 discard mode, the tracer only discards the event record that doesn't
1292 In discard mode, LTTng increments a count of lost event records when an
1293 event record is lost and saves this count to the trace. In overwrite
1294 mode, since LTTng{nbsp}2.8, LTTng increments a count of lost sub-buffers
1295 when a sub-buffer is lost and saves this count to the trace. In this
1296 mode, LTTng does not write to the trace the exact number of lost event
1297 records in those lost sub-buffers. Trace analyses can use the trace's
1298 saved discarded event record and sub-buffer counts to decide whether or
1299 not to perform the analyses even if trace data is known to be missing.
1301 There are a few ways to decrease your probability of losing event
1303 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1304 how you can fine-tune the sub-buffer count and size of a channel to
1305 virtually stop losing event records, though at the cost of greater
1309 [[channel-subbuf-size-vs-subbuf-count]]
1310 ==== Sub-buffer count and size
1312 When you <<enabling-disabling-channels,create a channel>>, you can
1313 set its number of sub-buffers and their size.
1315 Note that there is noticeable CPU overhead introduced when
1316 switching sub-buffers (marking a full one as consumable and switching
1317 to an empty one for the following events to be recorded). Knowing this,
1318 the following list presents a few practical situations along with how
1319 to configure the sub-buffer count and size for them:
1321 * **High event throughput**: In general, prefer bigger sub-buffers to
1322 lower the risk of losing event records.
1324 Having bigger sub-buffers also ensures a lower
1325 <<channel-switch-timer,sub-buffer switching frequency>>.
1327 The number of sub-buffers is only meaningful if you create the channel
1328 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1329 other sub-buffers are left unaltered.
1331 * **Low event throughput**: In general, prefer smaller sub-buffers
1332 since the risk of losing event records is low.
1334 Because events occur less frequently, the sub-buffer switching frequency
1335 should remain low and thus the tracer's overhead should not be a
1338 * **Low memory system**: If your target system has a low memory
1339 limit, prefer fewer first, then smaller sub-buffers.
1341 Even if the system is limited in memory, you want to keep the
1342 sub-buffers as big as possible to avoid a high sub-buffer switching
1345 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1346 which means event data is very compact. For example, the average
1347 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1348 sub-buffer size of 1{nbsp}MiB is considered big.
1350 The previous situations highlight the major trade-off between a few big
1351 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1352 frequency vs. how much data is lost in overwrite mode. Assuming a
1353 constant event throughput and using the overwrite mode, the two
1354 following configurations have the same ring buffer total size:
1357 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1362 * **Two sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1363 switching frequency, but if a sub-buffer overwrite happens, half of
1364 the event records so far (4{nbsp}MiB) are definitely lost.
1365 * **Eight sub-buffers of 1{nbsp}MiB each**: Expect four times the tracer's
1366 overhead as the previous configuration, but if a sub-buffer
1367 overwrite happens, only the eighth of event records so far are
1370 In discard mode, the sub-buffers count parameter is pointless: use two
1371 sub-buffers and set their size according to the requirements of your
1375 [[channel-switch-timer]]
1376 ==== Switch timer period
1378 The _switch timer period_ is an important configurable attribute of
1379 a channel to ensure periodic sub-buffer flushing.
1381 When the _switch timer_ expires, a sub-buffer switch happens. You can
1382 set the switch timer period attribute when you
1383 <<enabling-disabling-channels,create a channel>> to ensure that LTTng
1384 consumes and commits trace data to trace files or to a distant relay
1385 daemon periodically in case of a low event throughput.
1388 [role="docsvg-channel-switch-timer"]
1393 This attribute is also convenient when you use big sub-buffers to cope
1394 with a sporadic high event throughput, even if the throughput is
1398 [[channel-read-timer]]
1399 ==== Read timer period
1401 By default, the LTTng tracers use a notification mechanism to signal a
1402 full sub-buffer so that a consumer daemon can consume it. When such
1403 notifications must be avoided, for example in real-time applications,
1404 you can use the channel's _read timer_ instead. When the read timer
1405 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1406 consumable sub-buffers.
1409 [[tracefile-rotation]]
1410 ==== Trace file count and size
1412 By default, trace files can grow as large as needed. You can set the
1413 maximum size of each trace file that a channel writes when you
1414 <<enabling-disabling-channels,create a channel>>. When the size of
1415 a trace file reaches the channel's fixed maximum size, LTTng creates
1416 another file to contain the next event records. LTTng appends a file
1417 count to each trace file name in this case.
1419 If you set the trace file size attribute when you create a channel, the
1420 maximum number of trace files that LTTng creates is _unlimited_ by
1421 default. To limit them, you can also set a maximum number of trace
1422 files. When the number of trace files reaches the channel's fixed
1423 maximum count, the oldest trace file is overwritten. This mechanism is
1424 called _trace file rotation_.
1428 Even if you don't limit the trace file count, you cannot assume that
1429 LTTng doesn't manage any trace file.
1431 In other words, there is no safe way to know if LTTng still holds a
1432 given trace file open with the trace file rotation feature.
1434 The only way to obtain an unmanaged, self-contained LTTng trace before
1435 you <<creating-destroying-tracing-sessions,destroy>> the tracing session
1436 is with the <<session-rotation,tracing session rotation>> feature
1437 (available since LTTng{nbsp}2.11).
1442 === Instrumentation point, event rule, event, and event record
1444 An _event rule_ is a set of conditions which must be **all** satisfied
1445 for LTTng to record an occuring event.
1447 You set the conditions when you <<enabling-disabling-events,create
1450 You always attach an event rule to <<channel,channel>> when you create
1453 When an event passes the conditions of an event rule, LTTng records it
1454 in one of the attached channel's sub-buffers.
1456 The available conditions, as of LTTng{nbsp}{revision}, are:
1458 * The event rule _is enabled_.
1459 * The instrumentation point's type _is{nbsp}T_.
1460 * The instrumentation point's name (sometimes called _event name_)
1461 _matches{nbsp}N_, but _is not{nbsp}E_.
1462 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1463 _is exactly{nbsp}L_.
1464 * The fields of the event's payload _satisfy_ a filter
1465 expression{nbsp}__F__.
1467 As you can see, all the conditions but the dynamic filter are related to
1468 the event rule's status or to the instrumentation point, not to the
1469 occurring events. This is why, without a filter, checking if an event
1470 passes an event rule is not a dynamic task: when you create or modify an
1471 event rule, all the tracers of its tracing domain enable or disable the
1472 instrumentation points themselves once. This is possible because the
1473 attributes of an instrumentation point (type, name, and log level) are
1474 defined statically. In other words, without a dynamic filter, the tracer
1475 _does not evaluate_ the arguments of an instrumentation point unless it
1476 matches an enabled event rule.
1478 Note that, for LTTng to record an event, the <<channel,channel>> to
1479 which a matching event rule is attached must also be enabled, and the
1480 <<tracing-session,tracing session>> owning this channel must be active
1484 .Logical path from an instrumentation point to an event record.
1485 image::event-rule.png[]
1487 .Event, event record, or event rule?
1489 With so many similar terms, it's easy to get confused.
1491 An **event** is the consequence of the execution of an _instrumentation
1492 point_, like a tracepoint that you manually place in some source code,
1493 or a Linux kernel kprobe. An event is said to _occur_ at a specific
1494 time. Different actions can be taken upon the occurrence of an event,
1495 like record the event's payload to a buffer.
1497 An **event record** is the representation of an event in a sub-buffer. A
1498 tracer is responsible for capturing the payload of an event, current
1499 context variables, the event's ID, and the event's timestamp. LTTng
1500 can append this sub-buffer to a trace file.
1502 An **event rule** is a set of conditions which must _all_ be satisfied
1503 for LTTng to record an occuring event. Events still occur without
1504 satisfying event rules, but LTTng does not record them.
1509 == Components of noch:{LTTng}
1511 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1512 to call LTTng a simple _tool_ since it is composed of multiple
1513 interacting components. This section describes those components,
1514 explains their respective roles, and shows how they connect together to
1515 form the LTTng ecosystem.
1517 The following diagram shows how the most important components of LTTng
1518 interact with user applications, the Linux kernel, and you:
1521 .Control and trace data paths between LTTng components.
1522 image::plumbing.png[]
1524 The LTTng project incorporates:
1526 * **LTTng-tools**: Libraries and command-line interface to
1527 control tracing sessions.
1528 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1529 ** <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
1530 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1531 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1532 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1533 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1535 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1536 headers to instrument and trace any native user application.
1537 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1538 *** `liblttng-ust-libc-wrapper`
1539 *** `liblttng-ust-pthread-wrapper`
1540 *** `liblttng-ust-cyg-profile`
1541 *** `liblttng-ust-cyg-profile-fast`
1542 *** `liblttng-ust-dl`
1543 ** User space tracepoint provider source files generator command-line
1544 tool (man:lttng-gen-tp(1)).
1545 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1546 Java applications using `java.util.logging` or
1547 Apache log4j{nbsp}1.2 logging.
1548 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1549 Python applications using the standard `logging` package.
1550 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1552 ** LTTng kernel tracer module.
1553 ** Tracing ring buffer kernel modules.
1554 ** Probe kernel modules.
1555 ** LTTng logger kernel module.
1559 === Tracing control command-line interface
1562 .The tracing control command-line interface.
1563 image::plumbing-lttng-cli.png[]
1565 The _man:lttng(1) command-line tool_ is the standard user interface to
1566 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1567 is part of LTTng-tools.
1569 The cmd:lttng tool is linked with
1570 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1571 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1573 The cmd:lttng tool has a Git-like interface:
1577 $ lttng <GENERAL OPTIONS> <COMMAND> <COMMAND OPTIONS>
1580 The <<controlling-tracing,Tracing control>> section explores the
1581 available features of LTTng using the cmd:lttng tool.
1584 [[liblttng-ctl-lttng]]
1585 === Tracing control library
1588 .The tracing control library.
1589 image::plumbing-liblttng-ctl.png[]
1591 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1592 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1593 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1595 The <<lttng-cli,cmd:lttng command-line tool>>
1596 is linked with `liblttng-ctl`.
1598 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1603 #include <lttng/lttng.h>
1606 Some objects are referenced by name (C string), such as tracing
1607 sessions, but most of them require to create a handle first using
1608 `lttng_create_handle()`.
1610 The best available developer documentation for `liblttng-ctl` is, as of
1611 LTTng{nbsp}{revision}, its installed header files. Every function and
1612 structure is thoroughly documented.
1616 === User space tracing library
1619 .The user space tracing library.
1620 image::plumbing-liblttng-ust.png[]
1622 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1623 is the LTTng user space tracer. It receives commands from a
1624 <<lttng-sessiond,session daemon>>, for example to
1625 enable and disable specific instrumentation points, and writes event
1626 records to ring buffers shared with a
1627 <<lttng-consumerd,consumer daemon>>.
1628 `liblttng-ust` is part of LTTng-UST.
1630 Public C header files are installed beside `liblttng-ust` to
1631 instrument any <<c-application,C or $$C++$$ application>>.
1633 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1634 packages, use their own library providing tracepoints which is
1635 linked with `liblttng-ust`.
1637 An application or library does not have to initialize `liblttng-ust`
1638 manually: its constructor does the necessary tasks to properly register
1639 to a session daemon. The initialization phase also enables the
1640 instrumentation points matching the <<event,event rules>> that you
1644 [[lttng-ust-agents]]
1645 === User space tracing agents
1648 .The user space tracing agents.
1649 image::plumbing-lttng-ust-agents.png[]
1651 The _LTTng-UST Java and Python agents_ are regular Java and Python
1652 packages which add LTTng tracing capabilities to the
1653 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1655 In the case of Java, the
1656 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1657 core logging facilities] and
1658 https://logging.apache.org/log4j/1.2/[Apache log4j{nbsp}1.2] are supported.
1659 Note that Apache Log4{nbsp}2 is not supported.
1661 In the case of Python, the standard
1662 https://docs.python.org/3/library/logging.html[`logging`] package
1663 is supported. Both Python{nbsp}2 and Python{nbsp}3 modules can import the
1664 LTTng-UST Python agent package.
1666 The applications using the LTTng-UST agents are in the
1667 `java.util.logging` (JUL),
1668 log4j, and Python <<domain,tracing domains>>.
1670 Both agents use the same mechanism to trace the log statements. When an
1671 agent initializes, it creates a log handler that attaches to the root
1672 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1673 When the application executes a log statement, the root logger passes it
1674 to the agent's log handler. The agent's log handler calls a native
1675 function in a tracepoint provider package shared library linked with
1676 <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1677 other fields, like its logger name and its log level. This native
1678 function contains a user space instrumentation point, hence tracing the
1681 The log level condition of an
1682 <<event,event rule>> is considered when tracing
1683 a Java or a Python application, and it's compatible with the standard
1684 JUL, log4j, and Python log levels.
1688 === LTTng kernel modules
1691 .The LTTng kernel modules.
1692 image::plumbing-lttng-modules.png[]
1694 The _LTTng kernel modules_ are a set of Linux kernel modules
1695 which implement the kernel tracer of the LTTng project. The LTTng
1696 kernel modules are part of LTTng-modules.
1698 The LTTng kernel modules include:
1700 * A set of _probe_ modules.
1702 Each module attaches to a specific subsystem
1703 of the Linux kernel using its tracepoint instrument points. There are
1704 also modules to attach to the entry and return points of the Linux
1705 system call functions.
1707 * _Ring buffer_ modules.
1709 A ring buffer implementation is provided as kernel modules. The LTTng
1710 kernel tracer writes to the ring buffer; a
1711 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1713 * The _LTTng kernel tracer_ module.
1714 * The _LTTng logger_ module.
1716 The LTTng logger module implements the special path:{/proc/lttng-logger}
1717 file so that any executable can generate LTTng events by opening and
1718 writing to this file.
1720 See <<proc-lttng-logger-abi,LTTng logger>>.
1722 Generally, you do not have to load the LTTng kernel modules manually
1723 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1724 daemon>> loads the necessary modules when starting. If you have extra
1725 probe modules, you can specify to load them to the session daemon on
1728 The LTTng kernel modules are installed in
1729 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1730 the kernel release (see `uname --kernel-release`).
1737 .The session daemon.
1738 image::plumbing-sessiond.png[]
1740 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1741 managing tracing sessions and for controlling the various components of
1742 LTTng. The session daemon is part of LTTng-tools.
1744 The session daemon sends control requests to and receives control
1747 * The <<lttng-ust,user space tracing library>>.
1749 Any instance of the user space tracing library first registers to
1750 a session daemon. Then, the session daemon can send requests to
1751 this instance, such as:
1754 ** Get the list of tracepoints.
1755 ** Share an <<event,event rule>> so that the user space tracing library
1756 can enable or disable tracepoints. Amongst the possible conditions
1757 of an event rule is a filter expression which `liblttng-ust` evalutes
1758 when an event occurs.
1759 ** Share <<channel,channel>> attributes and ring buffer locations.
1762 The session daemon and the user space tracing library use a Unix
1763 domain socket for their communication.
1765 * The <<lttng-ust-agents,user space tracing agents>>.
1767 Any instance of a user space tracing agent first registers to
1768 a session daemon. Then, the session daemon can send requests to
1769 this instance, such as:
1772 ** Get the list of loggers.
1773 ** Enable or disable a specific logger.
1776 The session daemon and the user space tracing agent use a TCP connection
1777 for their communication.
1779 * The <<lttng-modules,LTTng kernel tracer>>.
1780 * The <<lttng-consumerd,consumer daemon>>.
1782 The session daemon sends requests to the consumer daemon to instruct
1783 it where to send the trace data streams, amongst other information.
1785 * The <<lttng-relayd,relay daemon>>.
1787 The session daemon receives commands from the
1788 <<liblttng-ctl-lttng,tracing control library>>.
1790 The root session daemon loads the appropriate
1791 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1792 a <<lttng-consumerd,consumer daemon>> as soon as you create
1793 an <<event,event rule>>.
1795 The session daemon does not send and receive trace data: this is the
1796 role of the <<lttng-consumerd,consumer daemon>> and
1797 <<lttng-relayd,relay daemon>>. It does, however, generate the
1798 http://diamon.org/ctf/[CTF] metadata stream.
1800 Each Unix user can have its own session daemon instance. The
1801 tracing sessions which different session daemons manage are completely
1804 The root user's session daemon is the only one which is
1805 allowed to control the LTTng kernel tracer, and its spawned consumer
1806 daemon is the only one which is allowed to consume trace data from the
1807 LTTng kernel tracer. Note, however, that any Unix user which is a member
1808 of the <<tracing-group,tracing group>> is allowed
1809 to create <<channel,channels>> in the
1810 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1813 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1814 session daemon when using its `create` command if none is currently
1815 running. You can also start the session daemon manually.
1822 .The consumer daemon.
1823 image::plumbing-consumerd.png[]
1825 The _consumer daemon_, cmd:lttng-consumerd, is a daemon which shares
1826 ring buffers with user applications or with the LTTng kernel modules to
1827 collect trace data and send it to some location (on disk or to a
1828 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
1829 is part of LTTng-tools.
1831 You do not start a consumer daemon manually: a consumer daemon is always
1832 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
1833 <<event,event rule>>, that is, before you start tracing. When you kill
1834 its owner session daemon, the consumer daemon also exits because it is
1835 the session daemon's child process. Command-line options of
1836 man:lttng-sessiond(8) target the consumer daemon process.
1838 There are up to two running consumer daemons per Unix user, whereas only
1839 one session daemon can run per user. This is because each process can be
1840 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
1841 and 64-bit processes, it is more efficient to have separate
1842 corresponding 32-bit and 64-bit consumer daemons. The root user is an
1843 exception: it can have up to _three_ running consumer daemons: 32-bit
1844 and 64-bit instances for its user applications, and one more
1845 reserved for collecting kernel trace data.
1853 image::plumbing-relayd.png[]
1855 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
1856 between remote session and consumer daemons, local trace files, and a
1857 remote live trace viewer. The relay daemon is part of LTTng-tools.
1859 The main purpose of the relay daemon is to implement a receiver of
1860 <<sending-trace-data-over-the-network,trace data over the network>>.
1861 This is useful when the target system does not have much file system
1862 space to record trace files locally.
1864 The relay daemon is also a server to which a
1865 <<lttng-live,live trace viewer>> can
1866 connect. The live trace viewer sends requests to the relay daemon to
1867 receive trace data as the target system emits events. The
1868 communication protocol is named _LTTng live_; it is used over TCP
1871 Note that you can start the relay daemon on the target system directly.
1872 This is the setup of choice when the use case is to view events as
1873 the target system emits them without the need of a remote system.
1877 == [[using-lttng]]Instrumentation
1879 There are many examples of tracing and monitoring in our everyday life:
1881 * You have access to real-time and historical weather reports and
1882 forecasts thanks to weather stations installed around the country.
1883 * You know your heart is safe thanks to an electrocardiogram.
1884 * You make sure not to drive your car too fast and to have enough fuel
1885 to reach your destination thanks to gauges visible on your dashboard.
1887 All the previous examples have something in common: they rely on
1888 **instruments**. Without the electrodes attached to the surface of your
1889 body's skin, cardiac monitoring is futile.
1891 LTTng, as a tracer, is no different from those real life examples. If
1892 you're about to trace a software system or, in other words, record its
1893 history of execution, you better have **instrumentation points** in the
1894 subject you're tracing, that is, the actual software.
1896 Various ways were developed to instrument a piece of software for LTTng
1897 tracing. The most straightforward one is to manually place
1898 instrumentation points, called _tracepoints_, in the software's source
1899 code. It is also possible to add instrumentation points dynamically in
1900 the Linux kernel <<domain,tracing domain>>.
1902 If you're only interested in tracing the Linux kernel, your
1903 instrumentation needs are probably already covered by LTTng's built-in
1904 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
1905 user application which is already instrumented for LTTng tracing.
1906 In such cases, you can skip this whole section and read the topics of
1907 the <<controlling-tracing,Tracing control>> section.
1909 Many methods are available to instrument a piece of software for LTTng
1912 * <<c-application,User space instrumentation for C and $$C++$$
1914 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
1915 * <<java-application,User space Java agent>>.
1916 * <<python-application,User space Python agent>>.
1917 * <<proc-lttng-logger-abi,LTTng logger>>.
1918 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
1922 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
1924 The procedure to instrument a C or $$C++$$ user application with
1925 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
1927 . <<tracepoint-provider,Create the source files of a tracepoint provider
1929 . <<probing-the-application-source-code,Add tracepoints to
1930 the application's source code>>.
1931 . <<building-tracepoint-providers-and-user-application,Build and link
1932 a tracepoint provider package and the user application>>.
1934 If you need quick, man:printf(3)-like instrumentation, you can skip
1935 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
1938 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
1939 instrument a user application with `liblttng-ust`.
1942 [[tracepoint-provider]]
1943 ==== Create the source files of a tracepoint provider package
1945 A _tracepoint provider_ is a set of compiled functions which provide
1946 **tracepoints** to an application, the type of instrumentation point
1947 supported by LTTng-UST. Those functions can emit events with
1948 user-defined fields and serialize those events as event records to one
1949 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
1950 macro, which you <<probing-the-application-source-code,insert in a user
1951 application's source code>>, calls those functions.
1953 A _tracepoint provider package_ is an object file (`.o`) or a shared
1954 library (`.so`) which contains one or more tracepoint providers.
1955 Its source files are:
1957 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
1958 * A <<tpp-source,tracepoint provider package source>> (`.c`).
1960 A tracepoint provider package is dynamically linked with `liblttng-ust`,
1961 the LTTng user space tracer, at run time.
1964 .User application linked with `liblttng-ust` and containing a tracepoint provider.
1965 image::ust-app.png[]
1967 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
1968 skip creating and using a tracepoint provider and use
1969 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
1973 ===== Create a tracepoint provider header file template
1975 A _tracepoint provider header file_ contains the tracepoint
1976 definitions of a tracepoint provider.
1978 To create a tracepoint provider header file:
1980 . Start from this template:
1984 .Tracepoint provider header file template (`.h` file extension).
1986 #undef TRACEPOINT_PROVIDER
1987 #define TRACEPOINT_PROVIDER provider_name
1989 #undef TRACEPOINT_INCLUDE
1990 #define TRACEPOINT_INCLUDE "./tp.h"
1992 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
1995 #include <lttng/tracepoint.h>
1998 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
1999 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2004 #include <lttng/tracepoint-event.h>
2010 * `provider_name` with the name of your tracepoint provider.
2011 * `"tp.h"` with the name of your tracepoint provider header file.
2013 . Below the `#include <lttng/tracepoint.h>` line, put your
2014 <<defining-tracepoints,tracepoint definitions>>.
2016 Your tracepoint provider name must be unique amongst all the possible
2017 tracepoint provider names used on the same target system. We
2018 suggest to include the name of your project or company in the name,
2019 for example, `org_lttng_my_project_tpp`.
2021 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2022 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2023 write are the <<defining-tracepoints,tracepoint definitions>>.
2026 [[defining-tracepoints]]
2027 ===== Create a tracepoint definition
2029 A _tracepoint definition_ defines, for a given tracepoint:
2031 * Its **input arguments**. They are the macro parameters that the
2032 `tracepoint()` macro accepts for this particular tracepoint
2033 in the user application's source code.
2034 * Its **output event fields**. They are the sources of event fields
2035 that form the payload of any event that the execution of the
2036 `tracepoint()` macro emits for this particular tracepoint.
2038 You can create a tracepoint definition by using the
2039 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2041 <<tpp-header,tracepoint provider header file template>>.
2043 The syntax of the `TRACEPOINT_EVENT()` macro is:
2046 .`TRACEPOINT_EVENT()` macro syntax.
2049 /* Tracepoint provider name */
2052 /* Tracepoint name */
2055 /* Input arguments */
2060 /* Output event fields */
2069 * `provider_name` with your tracepoint provider name.
2070 * `tracepoint_name` with your tracepoint name.
2071 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2072 * `fields` with the <<tpp-def-output-fields,output event field>>
2075 This tracepoint emits events named `provider_name:tracepoint_name`.
2078 .Event name's length limitation
2080 The concatenation of the tracepoint provider name and the
2081 tracepoint name must not exceed **254{nbsp}characters**. If it does, the
2082 instrumented application compiles and runs, but LTTng throws multiple
2083 warnings and you could experience serious issues.
2086 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2089 .`TP_ARGS()` macro syntax.
2098 * `type` with the C type of the argument.
2099 * `arg_name` with the argument name.
2101 You can repeat `type` and `arg_name` up to 10{nbsp}times to have
2102 more than one argument.
2104 .`TP_ARGS()` usage with three arguments.
2116 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2117 tracepoint definition with no input arguments.
2119 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2120 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2121 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2122 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2125 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2126 C expression that the tracer evalutes at the `tracepoint()` macro site
2127 in the application's source code. This expression provides a field's
2128 source of data. The argument expression can include input argument names
2129 listed in the `TP_ARGS()` macro.
2131 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2132 must be unique within a given tracepoint definition.
2134 Here's a complete tracepoint definition example:
2136 .Tracepoint definition.
2138 The following tracepoint definition defines a tracepoint which takes
2139 three input arguments and has four output event fields.
2143 #include "my-custom-structure.h"
2149 const struct my_custom_structure*, my_custom_structure,
2154 ctf_string(query_field, query)
2155 ctf_float(double, ratio_field, ratio)
2156 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2157 ctf_integer(int, send_size, my_custom_structure->send_size)
2162 You can refer to this tracepoint definition with the `tracepoint()`
2163 macro in your application's source code like this:
2167 tracepoint(my_provider, my_tracepoint,
2168 my_structure, some_ratio, the_query);
2172 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2173 if they satisfy an enabled <<event,event rule>>.
2176 [[using-tracepoint-classes]]
2177 ===== Use a tracepoint class
2179 A _tracepoint class_ is a class of tracepoints which share the same
2180 output event field definitions. A _tracepoint instance_ is one
2181 instance of such a defined tracepoint class, with its own tracepoint
2184 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2185 shorthand which defines both a tracepoint class and a tracepoint
2186 instance at the same time.
2188 When you build a tracepoint provider package, the C or $$C++$$ compiler
2189 creates one serialization function for each **tracepoint class**. A
2190 serialization function is responsible for serializing the event fields
2191 of a tracepoint to a sub-buffer when tracing.
2193 For various performance reasons, when your situation requires multiple
2194 tracepoint definitions with different names, but with the same event
2195 fields, we recommend that you manually create a tracepoint class
2196 and instantiate as many tracepoint instances as needed. One positive
2197 effect of such a design, amongst other advantages, is that all
2198 tracepoint instances of the same tracepoint class reuse the same
2199 serialization function, thus reducing
2200 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2202 .Use a tracepoint class and tracepoint instances.
2204 Consider the following three tracepoint definitions:
2216 ctf_integer(int, userid, userid)
2217 ctf_integer(size_t, len, len)
2229 ctf_integer(int, userid, userid)
2230 ctf_integer(size_t, len, len)
2242 ctf_integer(int, userid, userid)
2243 ctf_integer(size_t, len, len)
2248 In this case, we create three tracepoint classes, with one implicit
2249 tracepoint instance for each of them: `get_account`, `get_settings`, and
2250 `get_transaction`. However, they all share the same event field names
2251 and types. Hence three identical, yet independent serialization
2252 functions are created when you build the tracepoint provider package.
2254 A better design choice is to define a single tracepoint class and three
2255 tracepoint instances:
2259 /* The tracepoint class */
2260 TRACEPOINT_EVENT_CLASS(
2261 /* Tracepoint provider name */
2264 /* Tracepoint class name */
2267 /* Input arguments */
2273 /* Output event fields */
2275 ctf_integer(int, userid, userid)
2276 ctf_integer(size_t, len, len)
2280 /* The tracepoint instances */
2281 TRACEPOINT_EVENT_INSTANCE(
2282 /* Tracepoint provider name */
2285 /* Tracepoint class name */
2288 /* Tracepoint name */
2291 /* Input arguments */
2297 TRACEPOINT_EVENT_INSTANCE(
2306 TRACEPOINT_EVENT_INSTANCE(
2319 [[assigning-log-levels]]
2320 ===== Assign a log level to a tracepoint definition
2322 You can assign an optional _log level_ to a
2323 <<defining-tracepoints,tracepoint definition>>.
2325 Assigning different levels of severity to tracepoint definitions can
2326 be useful: when you <<enabling-disabling-events,create an event rule>>,
2327 you can target tracepoints having a log level as severe as a specific
2330 The concept of LTTng-UST log levels is similar to the levels found
2331 in typical logging frameworks:
2333 * In a logging framework, the log level is given by the function
2334 or method name you use at the log statement site: `debug()`,
2335 `info()`, `warn()`, `error()`, and so on.
2336 * In LTTng-UST, you statically assign the log level to a tracepoint
2337 definition; any `tracepoint()` macro invocation which refers to
2338 this definition has this log level.
2340 You can assign a log level to a tracepoint definition with the
2341 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2342 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2343 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2346 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2349 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2351 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2356 * `provider_name` with the tracepoint provider name.
2357 * `tracepoint_name` with the tracepoint name.
2358 * `log_level` with the log level to assign to the tracepoint
2359 definition named `tracepoint_name` in the `provider_name`
2360 tracepoint provider.
2362 See man:lttng-ust(3) for a list of available log level names.
2364 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2368 /* Tracepoint definition */
2377 ctf_integer(int, userid, userid)
2378 ctf_integer(size_t, len, len)
2382 /* Log level assignment */
2383 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2389 ===== Create a tracepoint provider package source file
2391 A _tracepoint provider package source file_ is a C source file which
2392 includes a <<tpp-header,tracepoint provider header file>> to expand its
2393 macros into event serialization and other functions.
2395 You can always use the following tracepoint provider package source
2399 .Tracepoint provider package source file template.
2401 #define TRACEPOINT_CREATE_PROBES
2406 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2407 header file>> name. You may also include more than one tracepoint
2408 provider header file here to create a tracepoint provider package
2409 holding more than one tracepoint providers.
2412 [[probing-the-application-source-code]]
2413 ==== Add tracepoints to an application's source code
2415 Once you <<tpp-header,create a tracepoint provider header file>>, you
2416 can use the `tracepoint()` macro in your application's
2417 source code to insert the tracepoints that this header
2418 <<defining-tracepoints,defines>>.
2420 The `tracepoint()` macro takes at least two parameters: the tracepoint
2421 provider name and the tracepoint name. The corresponding tracepoint
2422 definition defines the other parameters.
2424 .`tracepoint()` usage.
2426 The following <<defining-tracepoints,tracepoint definition>> defines a
2427 tracepoint which takes two input arguments and has two output event
2431 .Tracepoint provider header file.
2433 #include "my-custom-structure.h"
2440 const char*, cmd_name
2443 ctf_string(cmd_name, cmd_name)
2444 ctf_integer(int, number_of_args, argc)
2449 You can refer to this tracepoint definition with the `tracepoint()`
2450 macro in your application's source code like this:
2453 .Application's source file.
2457 int main(int argc, char* argv[])
2459 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2465 Note how the application's source code includes
2466 the tracepoint provider header file containing the tracepoint
2467 definitions to use, path:{tp.h}.
2470 .`tracepoint()` usage with a complex tracepoint definition.
2472 Consider this complex tracepoint definition, where multiple event
2473 fields refer to the same input arguments in their argument expression
2477 .Tracepoint provider header file.
2479 /* For `struct stat` */
2480 #include <sys/types.h>
2481 #include <sys/stat.h>
2493 ctf_integer(int, my_constant_field, 23 + 17)
2494 ctf_integer(int, my_int_arg_field, my_int_arg)
2495 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2496 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2497 my_str_arg[2] + my_str_arg[3])
2498 ctf_string(my_str_arg_field, my_str_arg)
2499 ctf_integer_hex(off_t, size_field, st->st_size)
2500 ctf_float(double, size_dbl_field, (double) st->st_size)
2501 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2502 size_t, strlen(my_str_arg) / 2)
2507 You can refer to this tracepoint definition with the `tracepoint()`
2508 macro in your application's source code like this:
2511 .Application's source file.
2513 #define TRACEPOINT_DEFINE
2520 stat("/etc/fstab", &s);
2521 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2527 If you look at the event record that LTTng writes when tracing this
2528 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2529 it should look like this:
2531 .Event record fields
2533 |Field's name |Field's value
2534 |`my_constant_field` |40
2535 |`my_int_arg_field` |23
2536 |`my_int_arg_field2` |529
2538 |`my_str_arg_field` |`Hello, World!`
2539 |`size_field` |0x12d
2540 |`size_dbl_field` |301.0
2541 |`half_my_str_arg_field` |`Hello,`
2545 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2546 compute--they use the call stack, for example. To avoid this
2547 computation when the tracepoint is disabled, you can use the
2548 `tracepoint_enabled()` and `do_tracepoint()` macros.
2550 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2554 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2556 tracepoint_enabled(provider_name, tracepoint_name)
2557 do_tracepoint(provider_name, tracepoint_name, ...)
2562 * `provider_name` with the tracepoint provider name.
2563 * `tracepoint_name` with the tracepoint name.
2565 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2566 `tracepoint_name` from the provider named `provider_name` is enabled
2569 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2570 if the tracepoint is enabled. Using `tracepoint()` with
2571 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2572 the `tracepoint_enabled()` check, thus a race condition is
2573 possible in this situation:
2576 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2578 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2579 stuff = prepare_stuff();
2582 tracepoint(my_provider, my_tracepoint, stuff);
2585 If the tracepoint is enabled after the condition, then `stuff` is not
2586 prepared: the emitted event will either contain wrong data, or the whole
2587 application could crash (segmentation fault, for example).
2589 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2590 `STAP_PROBEV()` call. If you need it, you must emit
2594 [[building-tracepoint-providers-and-user-application]]
2595 ==== Build and link a tracepoint provider package and an application
2597 Once you have one or more <<tpp-header,tracepoint provider header
2598 files>> and a <<tpp-source,tracepoint provider package source file>>,
2599 you can create the tracepoint provider package by compiling its source
2600 file. From here, multiple build and run scenarios are possible. The
2601 following table shows common application and library configurations
2602 along with the required command lines to achieve them.
2604 In the following diagrams, we use the following file names:
2607 Executable application.
2610 Application's object file.
2613 Tracepoint provider package object file.
2616 Tracepoint provider package archive file.
2619 Tracepoint provider package shared object file.
2622 User library object file.
2625 User library shared object file.
2627 We use the following symbols in the diagrams of table below:
2630 .Symbols used in the build scenario diagrams.
2631 image::ust-sit-symbols.png[]
2633 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2634 variable in the following instructions.
2636 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2637 .Common tracepoint provider package scenarios.
2639 |Scenario |Instructions
2642 The instrumented application is statically linked with
2643 the tracepoint provider package object.
2645 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2648 include::../common/ust-sit-step-tp-o.txt[]
2650 To build the instrumented application:
2652 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2657 #define TRACEPOINT_DEFINE
2661 . Compile the application source file:
2670 . Build the application:
2675 $ gcc -o app app.o tpp.o -llttng-ust -ldl
2679 To run the instrumented application:
2681 * Start the application:
2691 The instrumented application is statically linked with the
2692 tracepoint provider package archive file.
2694 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2697 To create the tracepoint provider package archive file:
2699 . Compile the <<tpp-source,tracepoint provider package source file>>:
2708 . Create the tracepoint provider package archive file:
2713 $ ar rcs tpp.a tpp.o
2717 To build the instrumented application:
2719 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2724 #define TRACEPOINT_DEFINE
2728 . Compile the application source file:
2737 . Build the application:
2742 $ gcc -o app app.o tpp.a -llttng-ust -ldl
2746 To run the instrumented application:
2748 * Start the application:
2758 The instrumented application is linked with the tracepoint provider
2759 package shared object.
2761 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2764 include::../common/ust-sit-step-tp-so.txt[]
2766 To build the instrumented application:
2768 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2773 #define TRACEPOINT_DEFINE
2777 . Compile the application source file:
2786 . Build the application:
2791 $ gcc -o app app.o -ldl -L. -ltpp
2795 To run the instrumented application:
2797 * Start the application:
2807 The tracepoint provider package shared object is preloaded before the
2808 instrumented application starts.
2810 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2813 include::../common/ust-sit-step-tp-so.txt[]
2815 To build the instrumented application:
2817 . In path:{app.c}, before including path:{tpp.h}, add the
2823 #define TRACEPOINT_DEFINE
2824 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2828 . Compile the application source file:
2837 . Build the application:
2842 $ gcc -o app app.o -ldl
2846 To run the instrumented application with tracing support:
2848 * Preload the tracepoint provider package shared object and
2849 start the application:
2854 $ LD_PRELOAD=./libtpp.so ./app
2858 To run the instrumented application without tracing support:
2860 * Start the application:
2870 The instrumented application dynamically loads the tracepoint provider
2871 package shared object.
2873 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
2876 include::../common/ust-sit-step-tp-so.txt[]
2878 To build the instrumented application:
2880 . In path:{app.c}, before including path:{tpp.h}, add the
2886 #define TRACEPOINT_DEFINE
2887 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2891 . Compile the application source file:
2900 . Build the application:
2905 $ gcc -o app app.o -ldl
2909 To run the instrumented application:
2911 * Start the application:
2921 The application is linked with the instrumented user library.
2923 The instrumented user library is statically linked with the tracepoint
2924 provider package object file.
2926 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
2929 include::../common/ust-sit-step-tp-o-fpic.txt[]
2931 To build the instrumented user library:
2933 . In path:{emon.c}, before including path:{tpp.h}, add the
2939 #define TRACEPOINT_DEFINE
2943 . Compile the user library source file:
2948 $ gcc -I. -fpic -c emon.c
2952 . Build the user library shared object:
2957 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
2961 To build the application:
2963 . Compile the application source file:
2972 . Build the application:
2977 $ gcc -o app app.o -L. -lemon
2981 To run the application:
2983 * Start the application:
2993 The application is linked with the instrumented user library.
2995 The instrumented user library is linked with the tracepoint provider
2996 package shared object.
2998 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3001 include::../common/ust-sit-step-tp-so.txt[]
3003 To build the instrumented user library:
3005 . In path:{emon.c}, before including path:{tpp.h}, add the
3011 #define TRACEPOINT_DEFINE
3015 . Compile the user library source file:
3020 $ gcc -I. -fpic -c emon.c
3024 . Build the user library shared object:
3029 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3033 To build the application:
3035 . Compile the application source file:
3044 . Build the application:
3049 $ gcc -o app app.o -L. -lemon
3053 To run the application:
3055 * Start the application:
3065 The tracepoint provider package shared object is preloaded before the
3068 The application is linked with the instrumented user library.
3070 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3073 include::../common/ust-sit-step-tp-so.txt[]
3075 To build the instrumented user library:
3077 . In path:{emon.c}, before including path:{tpp.h}, add the
3083 #define TRACEPOINT_DEFINE
3084 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3088 . Compile the user library source file:
3093 $ gcc -I. -fpic -c emon.c
3097 . Build the user library shared object:
3102 $ gcc -shared -o libemon.so emon.o -ldl
3106 To build the application:
3108 . Compile the application source file:
3117 . Build the application:
3122 $ gcc -o app app.o -L. -lemon
3126 To run the application with tracing support:
3128 * Preload the tracepoint provider package shared object and
3129 start the application:
3134 $ LD_PRELOAD=./libtpp.so ./app
3138 To run the application without tracing support:
3140 * Start the application:
3150 The application is linked with the instrumented user library.
3152 The instrumented user library dynamically loads the tracepoint provider
3153 package shared object.
3155 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3158 include::../common/ust-sit-step-tp-so.txt[]
3160 To build the instrumented user library:
3162 . In path:{emon.c}, before including path:{tpp.h}, add the
3168 #define TRACEPOINT_DEFINE
3169 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3173 . Compile the user library source file:
3178 $ gcc -I. -fpic -c emon.c
3182 . Build the user library shared object:
3187 $ gcc -shared -o libemon.so emon.o -ldl
3191 To build the application:
3193 . Compile the application source file:
3202 . Build the application:
3207 $ gcc -o app app.o -L. -lemon
3211 To run the application:
3213 * Start the application:
3223 The application dynamically loads the instrumented user library.
3225 The instrumented user library is linked with the tracepoint provider
3226 package shared object.
3228 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3231 include::../common/ust-sit-step-tp-so.txt[]
3233 To build the instrumented user library:
3235 . In path:{emon.c}, before including path:{tpp.h}, add the
3241 #define TRACEPOINT_DEFINE
3245 . Compile the user library source file:
3250 $ gcc -I. -fpic -c emon.c
3254 . Build the user library shared object:
3259 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3263 To build the application:
3265 . Compile the application source file:
3274 . Build the application:
3279 $ gcc -o app app.o -ldl -L. -lemon
3283 To run the application:
3285 * Start the application:
3295 The application dynamically loads the instrumented user library.
3297 The instrumented user library dynamically loads the tracepoint provider
3298 package shared object.
3300 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3303 include::../common/ust-sit-step-tp-so.txt[]
3305 To build the instrumented user library:
3307 . In path:{emon.c}, before including path:{tpp.h}, add the
3313 #define TRACEPOINT_DEFINE
3314 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3318 . Compile the user library source file:
3323 $ gcc -I. -fpic -c emon.c
3327 . Build the user library shared object:
3332 $ gcc -shared -o libemon.so emon.o -ldl
3336 To build the application:
3338 . Compile the application source file:
3347 . Build the application:
3352 $ gcc -o app app.o -ldl -L. -lemon
3356 To run the application:
3358 * Start the application:
3368 The tracepoint provider package shared object is preloaded before the
3371 The application dynamically loads the instrumented user library.
3373 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3376 include::../common/ust-sit-step-tp-so.txt[]
3378 To build the instrumented user library:
3380 . In path:{emon.c}, before including path:{tpp.h}, add the
3386 #define TRACEPOINT_DEFINE
3387 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3391 . Compile the user library source file:
3396 $ gcc -I. -fpic -c emon.c
3400 . Build the user library shared object:
3405 $ gcc -shared -o libemon.so emon.o -ldl
3409 To build the application:
3411 . Compile the application source file:
3420 . Build the application:
3425 $ gcc -o app app.o -L. -lemon
3429 To run the application with tracing support:
3431 * Preload the tracepoint provider package shared object and
3432 start the application:
3437 $ LD_PRELOAD=./libtpp.so ./app
3441 To run the application without tracing support:
3443 * Start the application:
3453 The application is statically linked with the tracepoint provider
3454 package object file.
3456 The application is linked with the instrumented user library.
3458 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3461 include::../common/ust-sit-step-tp-o.txt[]
3463 To build the instrumented user library:
3465 . In path:{emon.c}, before including path:{tpp.h}, add the
3471 #define TRACEPOINT_DEFINE
3475 . Compile the user library source file:
3480 $ gcc -I. -fpic -c emon.c
3484 . Build the user library shared object:
3489 $ gcc -shared -o libemon.so emon.o
3493 To build the application:
3495 . Compile the application source file:
3504 . Build the application:
3509 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3513 To run the instrumented application:
3515 * Start the application:
3525 The application is statically linked with the tracepoint provider
3526 package object file.
3528 The application dynamically loads the instrumented user library.
3530 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3533 include::../common/ust-sit-step-tp-o.txt[]
3535 To build the application:
3537 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3542 #define TRACEPOINT_DEFINE
3546 . Compile the application source file:
3555 . Build the application:
3560 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
3565 The `--export-dynamic` option passed to the linker is necessary for the
3566 dynamically loaded library to ``see'' the tracepoint symbols defined in
3569 To build the instrumented user library:
3571 . Compile the user library source file:
3576 $ gcc -I. -fpic -c emon.c
3580 . Build the user library shared object:
3585 $ gcc -shared -o libemon.so emon.o
3589 To run the application:
3591 * Start the application:
3602 [[using-lttng-ust-with-daemons]]
3603 ===== Use noch:{LTTng-UST} with daemons
3605 If your instrumented application calls man:fork(2), man:clone(2),
3606 or BSD's man:rfork(2), without a following man:exec(3)-family
3607 system call, you must preload the path:{liblttng-ust-fork.so} shared
3608 object when you start the application.
3612 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
3615 If your tracepoint provider package is
3616 a shared library which you also preload, you must put both
3617 shared objects in env:LD_PRELOAD:
3621 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3627 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3629 If your instrumented application closes one or more file descriptors
3630 which it did not open itself, you must preload the
3631 path:{liblttng-ust-fd.so} shared object when you start the application:
3635 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
3638 Typical use cases include closing all the file descriptors after
3639 man:fork(2) or man:rfork(2) and buggy applications doing
3643 [[lttng-ust-pkg-config]]
3644 ===== Use noch:{pkg-config}
3646 On some distributions, LTTng-UST ships with a
3647 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3648 metadata file. If this is your case, then you can use cmd:pkg-config to
3649 build an application on the command line:
3653 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3657 [[instrumenting-32-bit-app-on-64-bit-system]]
3658 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3660 In order to trace a 32-bit application running on a 64-bit system,
3661 LTTng must use a dedicated 32-bit
3662 <<lttng-consumerd,consumer daemon>>.
3664 The following steps show how to build and install a 32-bit consumer
3665 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3666 build and install the 32-bit LTTng-UST libraries, and how to build and
3667 link an instrumented 32-bit application in that context.
3669 To build a 32-bit instrumented application for a 64-bit target system,
3670 assuming you have a fresh target system with no installed Userspace RCU
3673 . Download, build, and install a 32-bit version of Userspace RCU:
3678 $ cd $(mktemp -d) &&
3679 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3680 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3681 cd userspace-rcu-0.9.* &&
3682 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3684 sudo make install &&
3689 . Using your distribution's package manager, or from source, install
3690 the following 32-bit versions of the following dependencies of
3691 LTTng-tools and LTTng-UST:
3694 * https://sourceforge.net/projects/libuuid/[libuuid]
3695 * http://directory.fsf.org/wiki/Popt[popt]
3696 * http://www.xmlsoft.org/[libxml2]
3699 . Download, build, and install a 32-bit version of the latest
3700 LTTng-UST{nbsp}{revision}:
3705 $ cd $(mktemp -d) &&
3706 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.11.tar.bz2 &&
3707 tar -xf lttng-ust-latest-2.11.tar.bz2 &&
3708 cd lttng-ust-2.11.* &&
3709 ./configure --libdir=/usr/local/lib32 \
3710 CFLAGS=-m32 CXXFLAGS=-m32 \
3711 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3713 sudo make install &&
3720 Depending on your distribution,
3721 32-bit libraries could be installed at a different location than
3722 `/usr/lib32`. For example, Debian is known to install
3723 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3725 In this case, make sure to set `LDFLAGS` to all the
3726 relevant 32-bit library paths, for example:
3730 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3734 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3735 the 32-bit consumer daemon:
3740 $ cd $(mktemp -d) &&
3741 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
3742 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
3743 cd lttng-tools-2.11.* &&
3744 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3745 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
3746 --disable-bin-lttng --disable-bin-lttng-crash \
3747 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
3749 cd src/bin/lttng-consumerd &&
3750 sudo make install &&
3755 . From your distribution or from source,
3756 <<installing-lttng,install>> the 64-bit versions of
3757 LTTng-UST and Userspace RCU.
3758 . Download, build, and install the 64-bit version of the
3759 latest LTTng-tools{nbsp}{revision}:
3764 $ cd $(mktemp -d) &&
3765 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
3766 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
3767 cd lttng-tools-2.11.* &&
3768 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3769 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3771 sudo make install &&
3776 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3777 when linking your 32-bit application:
3780 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3781 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3784 For example, let's rebuild the quick start example in
3785 <<tracing-your-own-user-application,Trace a user application>> as an
3786 instrumented 32-bit application:
3791 $ gcc -m32 -c -I. hello-tp.c
3792 $ gcc -m32 -c hello.c
3793 $ gcc -m32 -o hello hello.o hello-tp.o \
3794 -L/usr/lib32 -L/usr/local/lib32 \
3795 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3800 No special action is required to execute the 32-bit application and
3801 to trace it: use the command-line man:lttng(1) tool as usual.
3808 man:tracef(3) is a small LTTng-UST API designed for quick,
3809 man:printf(3)-like instrumentation without the burden of
3810 <<tracepoint-provider,creating>> and
3811 <<building-tracepoint-providers-and-user-application,building>>
3812 a tracepoint provider package.
3814 To use `tracef()` in your application:
3816 . In the C or C++ source files where you need to use `tracef()`,
3817 include `<lttng/tracef.h>`:
3822 #include <lttng/tracef.h>
3826 . In the application's source code, use `tracef()` like you would use
3834 tracef("my message: %d (%s)", my_integer, my_string);
3840 . Link your application with `liblttng-ust`:
3845 $ gcc -o app app.c -llttng-ust
3849 To trace the events that `tracef()` calls emit:
3851 * <<enabling-disabling-events,Create an event rule>> which matches the
3852 `lttng_ust_tracef:*` event name:
3857 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
3862 .Limitations of `tracef()`
3864 The `tracef()` utility function was developed to make user space tracing
3865 super simple, albeit with notable disadvantages compared to
3866 <<defining-tracepoints,user-defined tracepoints>>:
3868 * All the emitted events have the same tracepoint provider and
3869 tracepoint names, respectively `lttng_ust_tracef` and `event`.
3870 * There is no static type checking.
3871 * The only event record field you actually get, named `msg`, is a string
3872 potentially containing the values you passed to `tracef()`
3873 using your own format string. This also means that you cannot filter
3874 events with a custom expression at run time because there are no
3876 * Since `tracef()` uses the C standard library's man:vasprintf(3)
3877 function behind the scenes to format the strings at run time, its
3878 expected performance is lower than with user-defined tracepoints,
3879 which do not require a conversion to a string.
3881 Taking this into consideration, `tracef()` is useful for some quick
3882 prototyping and debugging, but you should not consider it for any
3883 permanent and serious applicative instrumentation.
3889 ==== Use `tracelog()`
3891 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
3892 the difference that it accepts an additional log level parameter.
3894 The goal of `tracelog()` is to ease the migration from logging to
3897 To use `tracelog()` in your application:
3899 . In the C or C++ source files where you need to use `tracelog()`,
3900 include `<lttng/tracelog.h>`:
3905 #include <lttng/tracelog.h>
3909 . In the application's source code, use `tracelog()` like you would use
3910 man:printf(3), except for the first parameter which is the log
3918 tracelog(TRACE_WARNING, "my message: %d (%s)",
3919 my_integer, my_string);
3925 See man:lttng-ust(3) for a list of available log level names.
3927 . Link your application with `liblttng-ust`:
3932 $ gcc -o app app.c -llttng-ust
3936 To trace the events that `tracelog()` calls emit with a log level
3937 _as severe as_ a specific log level:
3939 * <<enabling-disabling-events,Create an event rule>> which matches the
3940 `lttng_ust_tracelog:*` event name and a minimum level
3946 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
3947 --loglevel=TRACE_WARNING
3951 To trace the events that `tracelog()` calls emit with a
3952 _specific log level_:
3954 * Create an event rule which matches the `lttng_ust_tracelog:*`
3955 event name and a specific log level:
3960 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
3961 --loglevel-only=TRACE_INFO
3966 [[prebuilt-ust-helpers]]
3967 === Prebuilt user space tracing helpers
3969 The LTTng-UST package provides a few helpers in the form or preloadable
3970 shared objects which automatically instrument system functions and
3973 The helper shared objects are normally found in dir:{/usr/lib}. If you
3974 built LTTng-UST <<building-from-source,from source>>, they are probably
3975 located in dir:{/usr/local/lib}.
3977 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
3980 path:{liblttng-ust-libc-wrapper.so}::
3981 path:{liblttng-ust-pthread-wrapper.so}::
3982 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
3983 memory and POSIX threads function tracing>>.
3985 path:{liblttng-ust-cyg-profile.so}::
3986 path:{liblttng-ust-cyg-profile-fast.so}::
3987 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
3989 path:{liblttng-ust-dl.so}::
3990 <<liblttng-ust-dl,Dynamic linker tracing>>.
3992 To use a user space tracing helper with any user application:
3994 * Preload the helper shared object when you start the application:
3999 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4003 You can preload more than one helper:
4008 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4014 [[liblttng-ust-libc-pthread-wrapper]]
4015 ==== Instrument C standard library memory and POSIX threads functions
4017 The path:{liblttng-ust-libc-wrapper.so} and
4018 path:{liblttng-ust-pthread-wrapper.so} helpers
4019 add instrumentation to some C standard library and POSIX
4023 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4025 |TP provider name |TP name |Instrumented function
4027 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4028 |`calloc` |man:calloc(3)
4029 |`realloc` |man:realloc(3)
4030 |`free` |man:free(3)
4031 |`memalign` |man:memalign(3)
4032 |`posix_memalign` |man:posix_memalign(3)
4036 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4038 |TP provider name |TP name |Instrumented function
4040 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4041 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4042 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4043 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4046 When you preload the shared object, it replaces the functions listed
4047 in the previous tables by wrappers which contain tracepoints and call
4048 the replaced functions.
4051 [[liblttng-ust-cyg-profile]]
4052 ==== Instrument function entry and exit
4054 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4055 to the entry and exit points of functions.
4057 man:gcc(1) and man:clang(1) have an option named
4058 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4059 which generates instrumentation calls for entry and exit to functions.
4060 The LTTng-UST function tracing helpers,
4061 path:{liblttng-ust-cyg-profile.so} and
4062 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4063 to add tracepoints to the two generated functions (which contain
4064 `cyg_profile` in their names, hence the helper's name).
4066 To use the LTTng-UST function tracing helper, the source files to
4067 instrument must be built using the `-finstrument-functions` compiler
4070 There are two versions of the LTTng-UST function tracing helper:
4072 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4073 that you should only use when it can be _guaranteed_ that the
4074 complete event stream is recorded without any lost event record.
4075 Any kind of duplicate information is left out.
4077 Assuming no event record is lost, having only the function addresses on
4078 entry is enough to create a call graph, since an event record always
4079 contains the ID of the CPU that generated it.
4081 You can use a tool like man:addr2line(1) to convert function addresses
4082 back to source file names and line numbers.
4084 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4085 which also works in use cases where event records might get discarded or
4086 not recorded from application startup.
4087 In these cases, the trace analyzer needs more information to be
4088 able to reconstruct the program flow.
4090 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4091 points of this helper.
4093 All the tracepoints that this helper provides have the
4094 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4096 TIP: It's sometimes a good idea to limit the number of source files that
4097 you compile with the `-finstrument-functions` option to prevent LTTng
4098 from writing an excessive amount of trace data at run time. When using
4099 man:gcc(1), you can use the
4100 `-finstrument-functions-exclude-function-list` option to avoid
4101 instrument entries and exits of specific function names.
4106 ==== Instrument the dynamic linker
4108 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4109 man:dlopen(3) and man:dlclose(3) function calls.
4111 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4116 [[java-application]]
4117 === User space Java agent
4119 You can instrument any Java application which uses one of the following
4122 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4123 (JUL) core logging facilities.
4124 * http://logging.apache.org/log4j/1.2/[**Apache log4j{nbsp}1.2**], since
4125 LTTng{nbsp}2.6. Note that Apache Log4j{nbsp}2 is not supported.
4128 .LTTng-UST Java agent imported by a Java application.
4129 image::java-app.png[]
4131 Note that the methods described below are new in LTTng{nbsp}{revision}.
4132 Previous LTTng versions use another technique.
4134 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4135 and https://ci.lttng.org/[continuous integration], thus this version is
4136 directly supported. However, the LTTng-UST Java agent is also tested
4137 with OpenJDK{nbsp}7.
4142 ==== Use the LTTng-UST Java agent for `java.util.logging`
4144 To use the LTTng-UST Java agent in a Java application which uses
4145 `java.util.logging` (JUL):
4147 . In the Java application's source code, import the LTTng-UST
4148 log handler package for `java.util.logging`:
4153 import org.lttng.ust.agent.jul.LttngLogHandler;
4157 . Create an LTTng-UST JUL log handler:
4162 Handler lttngUstLogHandler = new LttngLogHandler();
4166 . Add this handler to the JUL loggers which should emit LTTng events:
4171 Logger myLogger = Logger.getLogger("some-logger");
4173 myLogger.addHandler(lttngUstLogHandler);
4177 . Use `java.util.logging` log statements and configuration as usual.
4178 The loggers with an attached LTTng-UST log handler can emit
4181 . Before exiting the application, remove the LTTng-UST log handler from
4182 the loggers attached to it and call its `close()` method:
4187 myLogger.removeHandler(lttngUstLogHandler);
4188 lttngUstLogHandler.close();
4192 This is not strictly necessary, but it is recommended for a clean
4193 disposal of the handler's resources.
4195 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4196 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4198 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4199 path] when you build the Java application.
4201 The JAR files are typically located in dir:{/usr/share/java}.
4203 IMPORTANT: The LTTng-UST Java agent must be
4204 <<installing-lttng,installed>> for the logging framework your
4207 .Use the LTTng-UST Java agent for `java.util.logging`.
4212 import java.io.IOException;
4213 import java.util.logging.Handler;
4214 import java.util.logging.Logger;
4215 import org.lttng.ust.agent.jul.LttngLogHandler;
4219 private static final int answer = 42;
4221 public static void main(String[] argv) throws Exception
4224 Logger logger = Logger.getLogger("jello");
4226 // Create an LTTng-UST log handler
4227 Handler lttngUstLogHandler = new LttngLogHandler();
4229 // Add the LTTng-UST log handler to our logger
4230 logger.addHandler(lttngUstLogHandler);
4233 logger.info("some info");
4234 logger.warning("some warning");
4236 logger.finer("finer information; the answer is " + answer);
4238 logger.severe("error!");
4240 // Not mandatory, but cleaner
4241 logger.removeHandler(lttngUstLogHandler);
4242 lttngUstLogHandler.close();
4251 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4254 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4255 <<enabling-disabling-events,create an event rule>> matching the
4256 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4261 $ lttng enable-event --jul jello
4265 Run the compiled class:
4269 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4272 <<basic-tracing-session-control,Stop tracing>> and inspect the
4282 In the resulting trace, an <<event,event record>> generated by a Java
4283 application using `java.util.logging` is named `lttng_jul:event` and
4284 has the following fields:
4287 Log record's message.
4293 Name of the class in which the log statement was executed.
4296 Name of the method in which the log statement was executed.
4299 Logging time (timestamp in milliseconds).
4302 Log level integer value.
4305 ID of the thread in which the log statement was executed.
4307 You can use the opt:lttng-enable-event(1):--loglevel or
4308 opt:lttng-enable-event(1):--loglevel-only option of the
4309 man:lttng-enable-event(1) command to target a range of JUL log levels
4310 or a specific JUL log level.
4315 ==== Use the LTTng-UST Java agent for Apache log4j
4317 To use the LTTng-UST Java agent in a Java application which uses
4318 Apache log4j{nbsp}1.2:
4320 . In the Java application's source code, import the LTTng-UST
4321 log appender package for Apache log4j:
4326 import org.lttng.ust.agent.log4j.LttngLogAppender;
4330 . Create an LTTng-UST log4j log appender:
4335 Appender lttngUstLogAppender = new LttngLogAppender();
4339 . Add this appender to the log4j loggers which should emit LTTng events:
4344 Logger myLogger = Logger.getLogger("some-logger");
4346 myLogger.addAppender(lttngUstLogAppender);
4350 . Use Apache log4j log statements and configuration as usual. The
4351 loggers with an attached LTTng-UST log appender can emit LTTng events.
4353 . Before exiting the application, remove the LTTng-UST log appender from
4354 the loggers attached to it and call its `close()` method:
4359 myLogger.removeAppender(lttngUstLogAppender);
4360 lttngUstLogAppender.close();
4364 This is not strictly necessary, but it is recommended for a clean
4365 disposal of the appender's resources.
4367 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4368 files, path:{lttng-ust-agent-common.jar} and
4369 path:{lttng-ust-agent-log4j.jar}, in the
4370 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4371 path] when you build the Java application.
4373 The JAR files are typically located in dir:{/usr/share/java}.
4375 IMPORTANT: The LTTng-UST Java agent must be
4376 <<installing-lttng,installed>> for the logging framework your
4379 .Use the LTTng-UST Java agent for Apache log4j.
4384 import org.apache.log4j.Appender;
4385 import org.apache.log4j.Logger;
4386 import org.lttng.ust.agent.log4j.LttngLogAppender;
4390 private static final int answer = 42;
4392 public static void main(String[] argv) throws Exception
4395 Logger logger = Logger.getLogger("jello");
4397 // Create an LTTng-UST log appender
4398 Appender lttngUstLogAppender = new LttngLogAppender();
4400 // Add the LTTng-UST log appender to our logger
4401 logger.addAppender(lttngUstLogAppender);
4404 logger.info("some info");
4405 logger.warn("some warning");
4407 logger.debug("debug information; the answer is " + answer);
4409 logger.fatal("error!");
4411 // Not mandatory, but cleaner
4412 logger.removeAppender(lttngUstLogAppender);
4413 lttngUstLogAppender.close();
4419 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4424 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4427 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4428 <<enabling-disabling-events,create an event rule>> matching the
4429 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4434 $ lttng enable-event --log4j jello
4438 Run the compiled class:
4442 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4445 <<basic-tracing-session-control,Stop tracing>> and inspect the
4455 In the resulting trace, an <<event,event record>> generated by a Java
4456 application using log4j is named `lttng_log4j:event` and
4457 has the following fields:
4460 Log record's message.
4466 Name of the class in which the log statement was executed.
4469 Name of the method in which the log statement was executed.
4472 Name of the file in which the executed log statement is located.
4475 Line number at which the log statement was executed.
4481 Log level integer value.
4484 Name of the Java thread in which the log statement was executed.
4486 You can use the opt:lttng-enable-event(1):--loglevel or
4487 opt:lttng-enable-event(1):--loglevel-only option of the
4488 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4489 or a specific log4j log level.
4493 [[java-application-context]]
4494 ==== Provide application-specific context fields in a Java application
4496 A Java application-specific context field is a piece of state provided
4497 by the application which <<adding-context,you can add>>, using the
4498 man:lttng-add-context(1) command, to each <<event,event record>>
4499 produced by the log statements of this application.
4501 For example, a given object might have a current request ID variable.
4502 You can create a context information retriever for this object and
4503 assign a name to this current request ID. You can then, using the
4504 man:lttng-add-context(1) command, add this context field by name to
4505 the JUL or log4j <<channel,channel>>.
4507 To provide application-specific context fields in a Java application:
4509 . In the Java application's source code, import the LTTng-UST
4510 Java agent context classes and interfaces:
4515 import org.lttng.ust.agent.context.ContextInfoManager;
4516 import org.lttng.ust.agent.context.IContextInfoRetriever;
4520 . Create a context information retriever class, that is, a class which
4521 implements the `IContextInfoRetriever` interface:
4526 class MyContextInfoRetriever implements IContextInfoRetriever
4529 public Object retrieveContextInfo(String key)
4531 if (key.equals("intCtx")) {
4533 } else if (key.equals("strContext")) {
4534 return "context value!";
4543 This `retrieveContextInfo()` method is the only member of the
4544 `IContextInfoRetriever` interface. Its role is to return the current
4545 value of a state by name to create a context field. The names of the
4546 context fields and which state variables they return depends on your
4549 All primitive types and objects are supported as context fields.
4550 When `retrieveContextInfo()` returns an object, the context field
4551 serializer calls its `toString()` method to add a string field to
4552 event records. The method can also return `null`, which means that
4553 no context field is available for the required name.
4555 . Register an instance of your context information retriever class to
4556 the context information manager singleton:
4561 IContextInfoRetriever cir = new MyContextInfoRetriever();
4562 ContextInfoManager cim = ContextInfoManager.getInstance();
4563 cim.registerContextInfoRetriever("retrieverName", cir);
4567 . Before exiting the application, remove your context information
4568 retriever from the context information manager singleton:
4573 ContextInfoManager cim = ContextInfoManager.getInstance();
4574 cim.unregisterContextInfoRetriever("retrieverName");
4578 This is not strictly necessary, but it is recommended for a clean
4579 disposal of some manager's resources.
4581 . Build your Java application with LTTng-UST Java agent support as
4582 usual, following the procedure for either the <<jul,JUL>> or
4583 <<log4j,Apache log4j>> framework.
4586 .Provide application-specific context fields in a Java application.
4591 import java.util.logging.Handler;
4592 import java.util.logging.Logger;
4593 import org.lttng.ust.agent.jul.LttngLogHandler;
4594 import org.lttng.ust.agent.context.ContextInfoManager;
4595 import org.lttng.ust.agent.context.IContextInfoRetriever;
4599 // Our context information retriever class
4600 private static class MyContextInfoRetriever
4601 implements IContextInfoRetriever
4604 public Object retrieveContextInfo(String key) {
4605 if (key.equals("intCtx")) {
4607 } else if (key.equals("strContext")) {
4608 return "context value!";
4615 private static final int answer = 42;
4617 public static void main(String args[]) throws Exception
4619 // Get the context information manager instance
4620 ContextInfoManager cim = ContextInfoManager.getInstance();
4622 // Create and register our context information retriever
4623 IContextInfoRetriever cir = new MyContextInfoRetriever();
4624 cim.registerContextInfoRetriever("myRetriever", cir);
4627 Logger logger = Logger.getLogger("jello");
4629 // Create an LTTng-UST log handler
4630 Handler lttngUstLogHandler = new LttngLogHandler();
4632 // Add the LTTng-UST log handler to our logger
4633 logger.addHandler(lttngUstLogHandler);
4636 logger.info("some info");
4637 logger.warning("some warning");
4639 logger.finer("finer information; the answer is " + answer);
4641 logger.severe("error!");
4643 // Not mandatory, but cleaner
4644 logger.removeHandler(lttngUstLogHandler);
4645 lttngUstLogHandler.close();
4646 cim.unregisterContextInfoRetriever("myRetriever");
4655 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4658 <<creating-destroying-tracing-sessions,Create a tracing session>>
4659 and <<enabling-disabling-events,create an event rule>> matching the
4665 $ lttng enable-event --jul jello
4668 <<adding-context,Add the application-specific context fields>> to the
4673 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
4674 $ lttng add-context --jul --type='$app.myRetriever:strContext'
4677 <<basic-tracing-session-control,Start tracing>>:
4684 Run the compiled class:
4688 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4691 <<basic-tracing-session-control,Stop tracing>> and inspect the
4703 [[python-application]]
4704 === User space Python agent
4706 You can instrument a Python{nbsp}2 or Python{nbsp}3 application which
4708 https://docs.python.org/3/library/logging.html[`logging`] package.
4710 Each log statement emits an LTTng event once the
4711 application module imports the
4712 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4715 .A Python application importing the LTTng-UST Python agent.
4716 image::python-app.png[]
4718 To use the LTTng-UST Python agent:
4720 . In the Python application's source code, import the LTTng-UST Python
4730 The LTTng-UST Python agent automatically adds its logging handler to the
4731 root logger at import time.
4733 Any log statement that the application executes before this import does
4734 not emit an LTTng event.
4736 IMPORTANT: The LTTng-UST Python agent must be
4737 <<installing-lttng,installed>>.
4739 . Use log statements and logging configuration as usual.
4740 Since the LTTng-UST Python agent adds a handler to the _root_
4741 logger, you can trace any log statement from any logger.
4743 .Use the LTTng-UST Python agent.
4754 logging.basicConfig()
4755 logger = logging.getLogger('my-logger')
4758 logger.debug('debug message')
4759 logger.info('info message')
4760 logger.warn('warn message')
4761 logger.error('error message')
4762 logger.critical('critical message')
4766 if __name__ == '__main__':
4770 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4771 logging handler which prints to the standard error stream, is not
4772 strictly required for LTTng-UST tracing to work, but in versions of
4773 Python preceding{nbsp}3.2, you could see a warning message which indicates
4774 that no handler exists for the logger `my-logger`.
4776 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4777 <<enabling-disabling-events,create an event rule>> matching the
4778 `my-logger` Python logger, and <<basic-tracing-session-control,start
4784 $ lttng enable-event --python my-logger
4788 Run the Python script:
4795 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4805 In the resulting trace, an <<event,event record>> generated by a Python
4806 application is named `lttng_python:event` and has the following fields:
4809 Logging time (string).
4812 Log record's message.
4818 Name of the function in which the log statement was executed.
4821 Line number at which the log statement was executed.
4824 Log level integer value.
4827 ID of the Python thread in which the log statement was executed.
4830 Name of the Python thread in which the log statement was executed.
4832 You can use the opt:lttng-enable-event(1):--loglevel or
4833 opt:lttng-enable-event(1):--loglevel-only option of the
4834 man:lttng-enable-event(1) command to target a range of Python log levels
4835 or a specific Python log level.
4837 When an application imports the LTTng-UST Python agent, the agent tries
4838 to register to a <<lttng-sessiond,session daemon>>. Note that you must
4839 <<start-sessiond,start the session daemon>> _before_ you run the Python
4840 application. If a session daemon is found, the agent tries to register
4841 to it during five seconds, after which the application continues
4842 without LTTng tracing support. You can override this timeout value with
4843 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
4846 If the session daemon stops while a Python application with an imported
4847 LTTng-UST Python agent runs, the agent retries to connect and to
4848 register to a session daemon every three seconds. You can override this
4849 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
4854 [[proc-lttng-logger-abi]]
4857 The `lttng-tracer` Linux kernel module, part of
4858 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
4859 path:{/proc/lttng-logger} when it's loaded. Any application can write
4860 text data to this file to emit an LTTng event.
4863 .An application writes to the LTTng logger file to emit an LTTng event.
4864 image::lttng-logger.png[]
4866 The LTTng logger is the quickest method--not the most efficient,
4867 however--to add instrumentation to an application. It is designed
4868 mostly to instrument shell scripts:
4872 $ echo "Some message, some $variable" > /proc/lttng-logger
4875 Any event that the LTTng logger emits is named `lttng_logger` and
4876 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
4877 other instrumentation points in the kernel tracing domain, **any Unix
4878 user** can <<enabling-disabling-events,create an event rule>> which
4879 matches its event name, not only the root user or users in the
4880 <<tracing-group,tracing group>>.
4882 To use the LTTng logger:
4884 * From any application, write text data to the path:{/proc/lttng-logger}
4887 The `msg` field of `lttng_logger` event records contains the
4890 NOTE: The maximum message length of an LTTng logger event is
4891 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
4892 than one event to contain the remaining data.
4894 You should not use the LTTng logger to trace a user application which
4895 can be instrumented in a more efficient way, namely:
4897 * <<c-application,C and $$C++$$ applications>>.
4898 * <<java-application,Java applications>>.
4899 * <<python-application,Python applications>>.
4901 .Use the LTTng logger.
4906 echo 'Hello, World!' > /proc/lttng-logger
4908 df --human-readable --print-type / > /proc/lttng-logger
4911 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4912 <<enabling-disabling-events,create an event rule>> matching the
4913 `lttng_logger` Linux kernel tracepoint, and
4914 <<basic-tracing-session-control,start tracing>>:
4919 $ lttng enable-event --kernel lttng_logger
4923 Run the Bash script:
4930 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4941 [[instrumenting-linux-kernel]]
4942 === LTTng kernel tracepoints
4944 NOTE: This section shows how to _add_ instrumentation points to the
4945 Linux kernel. The kernel's subsystems are already thoroughly
4946 instrumented at strategic places for LTTng when you
4947 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
4951 There are two methods to instrument the Linux kernel:
4953 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
4954 tracepoint which uses the `TRACE_EVENT()` API.
4956 Choose this if you want to instrumentation a Linux kernel tree with an
4957 instrumentation point compatible with ftrace, perf, and SystemTap.
4959 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
4960 instrument an out-of-tree kernel module.
4962 Choose this if you don't need ftrace, perf, or SystemTap support.
4966 [[linux-add-lttng-layer]]
4967 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
4969 This section shows how to add an LTTng layer to existing ftrace
4970 instrumentation using the `TRACE_EVENT()` API.
4972 This section does not document the `TRACE_EVENT()` macro. You can
4973 read the following articles to learn more about this API:
4975 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part{nbsp}1)]
4976 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part{nbsp}2)]
4977 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part{nbsp}3)]
4979 The following procedure assumes that your ftrace tracepoints are
4980 correctly defined in their own header and that they are created in
4981 one source file using the `CREATE_TRACE_POINTS` definition.
4983 To add an LTTng layer over an existing ftrace tracepoint:
4985 . Make sure the following kernel configuration options are
4991 * `CONFIG_HIGH_RES_TIMERS`
4992 * `CONFIG_TRACEPOINTS`
4995 . Build the Linux source tree with your custom ftrace tracepoints.
4996 . Boot the resulting Linux image on your target system.
4998 Confirm that the tracepoints exist by looking for their names in the
4999 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5000 is your subsystem's name.
5002 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5007 $ cd $(mktemp -d) &&
5008 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
5009 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
5010 cd lttng-modules-2.11.*
5014 . In dir:{instrumentation/events/lttng-module}, relative to the root
5015 of the LTTng-modules source tree, create a header file named
5016 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5017 LTTng-modules tracepoint definitions using the LTTng-modules
5020 Start with this template:
5024 .path:{instrumentation/events/lttng-module/my_subsys.h}
5027 #define TRACE_SYSTEM my_subsys
5029 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5030 #define _LTTNG_MY_SUBSYS_H
5032 #include "../../../probes/lttng-tracepoint-event.h"
5033 #include <linux/tracepoint.h>
5035 LTTNG_TRACEPOINT_EVENT(
5037 * Format is identical to TRACE_EVENT()'s version for the three
5038 * following macro parameters:
5041 TP_PROTO(int my_int, const char *my_string),
5042 TP_ARGS(my_int, my_string),
5044 /* LTTng-modules specific macros */
5046 ctf_integer(int, my_int_field, my_int)
5047 ctf_string(my_bar_field, my_bar)
5051 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5053 #include "../../../probes/define_trace.h"
5057 The entries in the `TP_FIELDS()` section are the list of fields for the
5058 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5059 ftrace's `TRACE_EVENT()` macro.
5061 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5062 complete description of the available `ctf_*()` macros.
5064 . Create the LTTng-modules probe's kernel module C source file,
5065 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5070 .path:{probes/lttng-probe-my-subsys.c}
5072 #include <linux/module.h>
5073 #include "../lttng-tracer.h"
5076 * Build-time verification of mismatch between mainline
5077 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5078 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5080 #include <trace/events/my_subsys.h>
5082 /* Create LTTng tracepoint probes */
5083 #define LTTNG_PACKAGE_BUILD
5084 #define CREATE_TRACE_POINTS
5085 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5087 #include "../instrumentation/events/lttng-module/my_subsys.h"
5089 MODULE_LICENSE("GPL and additional rights");
5090 MODULE_AUTHOR("Your name <your-email>");
5091 MODULE_DESCRIPTION("LTTng my_subsys probes");
5092 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5093 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5094 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5095 LTTNG_MODULES_EXTRAVERSION);
5099 . Edit path:{probes/KBuild} and add your new kernel module object
5100 next to the existing ones:
5104 .path:{probes/KBuild}
5108 obj-m += lttng-probe-module.o
5109 obj-m += lttng-probe-power.o
5111 obj-m += lttng-probe-my-subsys.o
5117 . Build and install the LTTng kernel modules:
5122 $ make KERNELDIR=/path/to/linux
5123 # make modules_install && depmod -a
5127 Replace `/path/to/linux` with the path to the Linux source tree where
5128 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5130 Note that you can also use the
5131 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5132 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5133 C code that need to be executed before the event fields are recorded.
5135 The best way to learn how to use the previous LTTng-modules macros is to
5136 inspect the existing LTTng-modules tracepoint definitions in the
5137 dir:{instrumentation/events/lttng-module} header files. Compare them
5138 with the Linux kernel mainline versions in the
5139 dir:{include/trace/events} directory of the Linux source tree.
5143 [[lttng-tracepoint-event-code]]
5144 ===== Use custom C code to access the data for tracepoint fields
5146 Although we recommended to always use the
5147 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5148 the arguments and fields of an LTTng-modules tracepoint when possible,
5149 sometimes you need a more complex process to access the data that the
5150 tracer records as event record fields. In other words, you need local
5151 variables and multiple C{nbsp}statements instead of simple
5152 argument-based expressions that you pass to the
5153 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5155 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5156 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5157 a block of C{nbsp}code to be executed before LTTng records the fields.
5158 The structure of this macro is:
5161 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5163 LTTNG_TRACEPOINT_EVENT_CODE(
5165 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5166 * version for the following three macro parameters:
5169 TP_PROTO(int my_int, const char *my_string),
5170 TP_ARGS(my_int, my_string),
5172 /* Declarations of custom local variables */
5175 unsigned long b = 0;
5176 const char *name = "(undefined)";
5177 struct my_struct *my_struct;
5181 * Custom code which uses both tracepoint arguments
5182 * (in TP_ARGS()) and local variables (in TP_locvar()).
5184 * Local variables are actually members of a structure pointed
5185 * to by the special variable tp_locvar.
5189 tp_locvar->a = my_int + 17;
5190 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5191 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5192 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5193 put_my_struct(tp_locvar->my_struct);
5202 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5203 * version for this, except that tp_locvar members can be
5204 * used in the argument expression parameters of
5205 * the ctf_*() macros.
5208 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5209 ctf_integer(int, my_struct_a, tp_locvar->a)
5210 ctf_string(my_string_field, my_string)
5211 ctf_string(my_struct_name, tp_locvar->name)
5216 IMPORTANT: The C code defined in `TP_code()` must not have any side
5217 effects when executed. In particular, the code must not allocate
5218 memory or get resources without deallocating this memory or putting
5219 those resources afterwards.
5222 [[instrumenting-linux-kernel-tracing]]
5223 ==== Load and unload a custom probe kernel module
5225 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5226 kernel module>> in the kernel before it can emit LTTng events.
5228 To load the default probe kernel modules and a custom probe kernel
5231 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5232 probe modules to load when starting a root <<lttng-sessiond,session
5236 .Load the `my_subsys`, `usb`, and the default probe modules.
5240 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5245 You only need to pass the subsystem name, not the whole kernel module
5248 To load _only_ a given custom probe kernel module:
5250 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5251 modules to load when starting a root session daemon:
5254 .Load only the `my_subsys` and `usb` probe modules.
5258 # lttng-sessiond --kmod-probes=my_subsys,usb
5263 To confirm that a probe module is loaded:
5270 $ lsmod | grep lttng_probe_usb
5274 To unload the loaded probe modules:
5276 * Kill the session daemon with `SIGTERM`:
5281 # pkill lttng-sessiond
5285 You can also use man:modprobe(8)'s `--remove` option if the session
5286 daemon terminates abnormally.
5289 [[controlling-tracing]]
5292 Once an application or a Linux kernel is
5293 <<instrumenting,instrumented>> for LTTng tracing,
5296 This section is divided in topics on how to use the various
5297 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5298 command-line tool>>, to _control_ the LTTng daemons and tracers.
5300 NOTE: In the following subsections, we refer to an man:lttng(1) command
5301 using its man page name. For example, instead of _Run the `create`
5302 command to..._, we use _Run the man:lttng-create(1) command to..._.
5306 === Start a session daemon
5308 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5309 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5312 You will see the following error when you run a command while no session
5316 Error: No session daemon is available
5319 The only command that automatically runs a session daemon is
5320 man:lttng-create(1), which you use to
5321 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5322 this is most of the time the first operation that you do, sometimes it's
5323 not. Some examples are:
5325 * <<list-instrumentation-points,List the available instrumentation points>>.
5326 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5328 [[tracing-group]] Each Unix user must have its own running session
5329 daemon to trace user applications. The session daemon that the root user
5330 starts is the only one allowed to control the LTTng kernel tracer. Users
5331 that are part of the _tracing group_ can control the root session
5332 daemon. The default tracing group name is `tracing`; you can set it to
5333 something else with the opt:lttng-sessiond(8):--group option when you
5334 start the root session daemon.
5336 To start a user session daemon:
5338 * Run man:lttng-sessiond(8):
5343 $ lttng-sessiond --daemonize
5347 To start the root session daemon:
5349 * Run man:lttng-sessiond(8) as the root user:
5354 # lttng-sessiond --daemonize
5358 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5359 start the session daemon in foreground.
5361 To stop a session daemon, use man:kill(1) on its process ID (standard
5364 Note that some Linux distributions could manage the LTTng session daemon
5365 as a service. In this case, you should use the service manager to
5366 start, restart, and stop session daemons.
5369 [[creating-destroying-tracing-sessions]]
5370 === Create and destroy a tracing session
5372 Almost all the LTTng control operations happen in the scope of
5373 a <<tracing-session,tracing session>>, which is the dialogue between the
5374 <<lttng-sessiond,session daemon>> and you.
5376 To create a tracing session with a generated name:
5378 * Use the man:lttng-create(1) command:
5387 The created tracing session's name is `auto` followed by the
5390 To create a tracing session with a specific name:
5392 * Use the optional argument of the man:lttng-create(1) command:
5397 $ lttng create my-session
5401 Replace `my-session` with the specific tracing session name.
5403 LTTng appends the creation date to the created tracing session's name.
5405 LTTng writes the traces of a tracing session in
5406 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5407 name of the tracing session. Note that the env:LTTNG_HOME environment
5408 variable defaults to `$HOME` if not set.
5410 To output LTTng traces to a non-default location:
5412 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5417 $ lttng create my-session --output=/tmp/some-directory
5421 You may create as many tracing sessions as you wish.
5423 To list all the existing tracing sessions for your Unix user:
5425 * Use the man:lttng-list(1) command:
5434 When you create a tracing session, it is set as the _current tracing
5435 session_. The following man:lttng(1) commands operate on the current
5436 tracing session when you don't specify one:
5438 [role="list-3-cols"]
5439 * man:lttng-add-context(1)
5440 * man:lttng-destroy(1)
5441 * man:lttng-disable-channel(1)
5442 * man:lttng-disable-event(1)
5443 * man:lttng-disable-rotation(1)
5444 * man:lttng-enable-channel(1)
5445 * man:lttng-enable-event(1)
5446 * man:lttng-enable-rotation(1)
5448 * man:lttng-regenerate(1)
5449 * man:lttng-rotate(1)
5451 * man:lttng-snapshot(1)
5452 * man:lttng-start(1)
5453 * man:lttng-status(1)
5455 * man:lttng-track(1)
5456 * man:lttng-untrack(1)
5459 To change the current tracing session:
5461 * Use the man:lttng-set-session(1) command:
5466 $ lttng set-session new-session
5470 Replace `new-session` by the name of the new current tracing session.
5472 When you are done tracing in a given tracing session, you can destroy
5473 it. This operation frees the resources taken by the tracing session
5474 to destroy; it does not destroy the trace data that LTTng wrote for
5475 this tracing session.
5477 To destroy the current tracing session:
5479 * Use the man:lttng-destroy(1) command:
5488 The man:lttng-destroy(1) command also runs the man:lttng-stop(1)
5489 command implicitly (see <<basic-tracing-session-control,Start and stop a
5490 tracing session>>). You need to stop tracing to make LTTng flush the
5491 remaining trace data and make the trace readable.
5494 [[list-instrumentation-points]]
5495 === List the available instrumentation points
5497 The <<lttng-sessiond,session daemon>> can query the running instrumented
5498 user applications and the Linux kernel to get a list of available
5499 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5500 they are tracepoints and system calls. For the user space tracing
5501 domain, they are tracepoints. For the other tracing domains, they are
5504 To list the available instrumentation points:
5506 * Use the man:lttng-list(1) command with the requested tracing domain's
5510 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5511 must be a root user, or it must be a member of the
5512 <<tracing-group,tracing group>>).
5513 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5514 kernel system calls (your Unix user must be a root user, or it must be
5515 a member of the tracing group).
5516 * opt:lttng-list(1):--userspace: user space tracepoints.
5517 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5518 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5519 * opt:lttng-list(1):--python: Python loggers.
5522 .List the available user space tracepoints.
5526 $ lttng list --userspace
5530 .List the available Linux kernel system call tracepoints.
5534 $ lttng list --kernel --syscall
5539 [[enabling-disabling-events]]
5540 === Create and enable an event rule
5542 Once you <<creating-destroying-tracing-sessions,create a tracing
5543 session>>, you can create <<event,event rules>> with the
5544 man:lttng-enable-event(1) command.
5546 You specify each condition with a command-line option. The available
5547 condition arguments are shown in the following table.
5549 [role="growable",cols="asciidoc,asciidoc,default"]
5550 .Condition command-line arguments for the man:lttng-enable-event(1) command.
5552 |Argument |Description |Applicable tracing domains
5558 . +--probe=__ADDR__+
5559 . +--function=__ADDR__+
5560 . +--userspace-probe=__PATH__:__SYMBOL__+
5561 . +--userspace-probe=sdt:__PATH__:__PROVIDER__:__NAME__+
5564 Instead of using the default _tracepoint_ instrumentation type, use:
5566 . A Linux system call (entry and exit).
5567 . A Linux https://lwn.net/Articles/132196/[kprobe] (symbol or address).
5568 . The entry and return points of a Linux function (symbol or address).
5569 . The entry point of a user application or library function (path to
5570 application/library and symbol).
5571 . A https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
5572 Statically Defined Tracing] (USDT) probe (path to application/library,
5573 provider and probe names).
5577 |First positional argument.
5580 Tracepoint or system call name.
5582 With the opt:lttng-enable-event(1):--probe,
5583 opt:lttng-enable-event(1):--function, and
5584 opt:lttng-enable-event(1):--userspace-probe options, this is a custom
5585 name given to the event rule. With the JUL, log4j, and Python domains,
5586 this is a logger name.
5588 With a tracepoint, logger, or system call name, you can use the special
5589 `*` globbing character to match anything (for example, `sched_*`,
5597 . +--loglevel=__LEVEL__+
5598 . +--loglevel-only=__LEVEL__+
5601 . Match only tracepoints or log statements with a logging level at
5602 least as severe as +__LEVEL__+.
5603 . Match only tracepoints or log statements with a logging level
5604 equal to +__LEVEL__+.
5606 See man:lttng-enable-event(1) for the list of available logging level
5609 |User space, JUL, log4j, and Python.
5611 |+--exclude=__EXCLUSIONS__+
5614 When you use a `*` character at the end of the tracepoint or logger
5615 name (first positional argument), exclude the specific names in the
5616 comma-delimited list +__EXCLUSIONS__+.
5619 User space, JUL, log4j, and Python.
5621 |+--filter=__EXPR__+
5624 Match only events which satisfy the expression +__EXPR__+.
5626 See man:lttng-enable-event(1) to learn more about the syntax of a
5633 You attach an event rule to a <<channel,channel>> on creation. If you do
5634 not specify the channel with the opt:lttng-enable-event(1):--channel
5635 option, and if the event rule to create is the first in its
5636 <<domain,tracing domain>> for a given tracing session, then LTTng
5637 creates a _default channel_ for you. This default channel is reused in
5638 subsequent invocations of the man:lttng-enable-event(1) command for the
5639 same tracing domain.
5641 An event rule is always enabled at creation time.
5643 The following examples show how you can combine the previous
5644 command-line options to create simple to more complex event rules.
5646 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5650 $ lttng enable-event --kernel sched_switch
5654 .Create an event rule matching four Linux kernel system calls (default channel).
5658 $ lttng enable-event --kernel --syscall open,write,read,close
5662 .Create event rules matching tracepoints with filter expressions (default channel).
5666 $ lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5671 $ lttng enable-event --kernel --all \
5672 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5677 $ lttng enable-event --jul my_logger \
5678 --filter='$app.retriever:cur_msg_id > 3'
5681 IMPORTANT: Make sure to always quote the filter string when you
5682 use man:lttng(1) from a shell.
5685 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5689 $ lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5692 IMPORTANT: Make sure to always quote the wildcard character when you
5693 use man:lttng(1) from a shell.
5696 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5700 $ lttng enable-event --python my-app.'*' \
5701 --exclude='my-app.module,my-app.hello'
5705 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5709 $ lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5713 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5717 $ lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5721 .Create an event rule matching the `malloc` function entry in path:{/usr/lib/libc.so.6}:
5725 $ lttng enable-event --kernel --userspace-probe=/usr/lib/libc.so.6:malloc \
5730 .Create an event rule matching the `server`/`accept_request` https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[USDT probe] in path:{/usr/bin/serv}:
5734 $ lttng enable-event --kernel --userspace-probe=sdt:serv:server:accept_request \
5735 server_accept_request
5739 The event rules of a given channel form a whitelist: as soon as an
5740 emitted event passes one of them, LTTng can record the event. For
5741 example, an event named `my_app:my_tracepoint` emitted from a user space
5742 tracepoint with a `TRACE_ERROR` log level passes both of the following
5747 $ lttng enable-event --userspace my_app:my_tracepoint
5748 $ lttng enable-event --userspace my_app:my_tracepoint \
5749 --loglevel=TRACE_INFO
5752 The second event rule is redundant: the first one includes
5756 [[disable-event-rule]]
5757 === Disable an event rule
5759 To disable an event rule that you <<enabling-disabling-events,created>>
5760 previously, use the man:lttng-disable-event(1) command. This command
5761 disables _all_ the event rules (of a given tracing domain and channel)
5762 which match an instrumentation point. The other conditions are not
5763 supported as of LTTng{nbsp}{revision}.
5765 The LTTng tracer does not record an emitted event which passes
5766 a _disabled_ event rule.
5768 .Disable an event rule matching a Python logger (default channel).
5772 $ lttng disable-event --python my-logger
5776 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5780 $ lttng disable-event --jul '*'
5784 .Disable _all_ the event rules of the default channel.
5786 The opt:lttng-disable-event(1):--all-events option is not, like the
5787 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5788 equivalent of the event name `*` (wildcard): it disables _all_ the event
5789 rules of a given channel.
5793 $ lttng disable-event --jul --all-events
5797 NOTE: You cannot delete an event rule once you create it.
5801 === Get the status of a tracing session
5803 To get the status of the current tracing session, that is, its
5804 parameters, its channels, event rules, and their attributes:
5806 * Use the man:lttng-status(1) command:
5816 To get the status of any tracing session:
5818 * Use the man:lttng-list(1) command with the tracing session's name:
5823 $ lttng list my-session
5827 Replace `my-session` with the desired tracing session's name.
5830 [[basic-tracing-session-control]]
5831 === Start and stop a tracing session
5833 Once you <<creating-destroying-tracing-sessions,create a tracing
5835 <<enabling-disabling-events,create one or more event rules>>,
5836 you can start and stop the tracers for this tracing session.
5838 To start tracing in the current tracing session:
5840 * Use the man:lttng-start(1) command:
5849 LTTng is very flexible: you can launch user applications before
5850 or after the you start the tracers. The tracers only record the events
5851 if they pass enabled event rules and if they occur while the tracers are
5854 To stop tracing in the current tracing session:
5856 * Use the man:lttng-stop(1) command:
5865 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
5866 records>> or lost sub-buffers since the last time you ran
5867 man:lttng-start(1), warnings are printed when you run the
5868 man:lttng-stop(1) command.
5870 IMPORTANT: You need to stop tracing to make LTTng flush the remaining
5871 trace data and make the trace readable. Note that the
5872 man:lttng-destroy(1) command (see
5873 <<creating-destroying-tracing-sessions,Create and destroy a tracing
5874 session>>) also runs the man:lttng-stop(1) command implicitly.
5877 [[enabling-disabling-channels]]
5878 === Create a channel
5880 Once you create a tracing session, you can create a <<channel,channel>>
5881 with the man:lttng-enable-channel(1) command.
5883 Note that LTTng automatically creates a default channel when, for a
5884 given <<domain,tracing domain>>, no channels exist and you
5885 <<enabling-disabling-events,create>> the first event rule. This default
5886 channel is named `channel0` and its attributes are set to reasonable
5887 values. Therefore, you only need to create a channel when you need
5888 non-default attributes.
5890 You specify each non-default channel attribute with a command-line
5891 option when you use the man:lttng-enable-channel(1) command. The
5892 available command-line options are:
5894 [role="growable",cols="asciidoc,asciidoc"]
5895 .Command-line options for the man:lttng-enable-channel(1) command.
5897 |Option |Description
5903 <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> instead
5904 of the default _discard_ mode.
5906 |`--buffers-pid` (user space tracing domain only)
5909 Use the per-process <<channel-buffering-schemes,buffering scheme>>
5910 instead of the default per-user buffering scheme.
5912 |+--subbuf-size=__SIZE__+
5915 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
5916 either for each Unix user (default), or for each instrumented process.
5918 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5920 |+--num-subbuf=__COUNT__+
5923 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
5924 for each Unix user (default), or for each instrumented process.
5926 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5928 |+--tracefile-size=__SIZE__+
5931 Set the maximum size of each trace file that this channel writes within
5932 a stream to +__SIZE__+ bytes instead of no maximum.
5934 See <<tracefile-rotation,Trace file count and size>>.
5936 |+--tracefile-count=__COUNT__+
5939 Limit the number of trace files that this channel creates to
5940 +__COUNT__+ channels instead of no limit.
5942 See <<tracefile-rotation,Trace file count and size>>.
5944 |+--switch-timer=__PERIODUS__+
5947 Set the <<channel-switch-timer,switch timer period>>
5948 to +__PERIODUS__+{nbsp}µs.
5950 |+--read-timer=__PERIODUS__+
5953 Set the <<channel-read-timer,read timer period>>
5954 to +__PERIODUS__+{nbsp}µs.
5956 |[[opt-blocking-timeout]]+--blocking-timeout=__TIMEOUTUS__+
5959 Set the timeout of user space applications which load LTTng-UST
5960 in blocking mode to +__TIMEOUTUS__+:
5963 Never block (non-blocking mode).
5966 Block forever until space is available in a sub-buffer to record
5969 __n__, a positive value::
5970 Wait for at most __n__ µs when trying to write into a sub-buffer.
5972 Note that, for this option to have any effect on an instrumented
5973 user space application, you need to run the application with a set
5974 env:LTTNG_UST_ALLOW_BLOCKING environment variable.
5976 |+--output=__TYPE__+ (Linux kernel tracing domain only)
5979 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
5983 You can only create a channel in the Linux kernel and user space
5984 <<domain,tracing domains>>: other tracing domains have their own channel
5985 created on the fly when <<enabling-disabling-events,creating event
5990 Because of a current LTTng limitation, you must create all channels
5991 _before_ you <<basic-tracing-session-control,start tracing>> in a given
5992 tracing session, that is, before the first time you run
5995 Since LTTng automatically creates a default channel when you use the
5996 man:lttng-enable-event(1) command with a specific tracing domain, you
5997 cannot, for example, create a Linux kernel event rule, start tracing,
5998 and then create a user space event rule, because no user space channel
5999 exists yet and it's too late to create one.
6001 For this reason, make sure to configure your channels properly
6002 before starting the tracers for the first time!
6005 The following examples show how you can combine the previous
6006 command-line options to create simple to more complex channels.
6008 .Create a Linux kernel channel with default attributes.
6012 $ lttng enable-channel --kernel my-channel
6016 .Create a user space channel with four sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6020 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6021 --buffers-pid my-channel
6025 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6027 <<creating-destroying-tracing-sessions,Create a tracing-session>>,
6028 create the channel, <<enabling-disabling-events,create an event rule>>,
6029 and <<basic-tracing-session-control,start tracing>>:
6034 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
6035 $ lttng enable-event --userspace --channel=blocking-channel --all
6039 Run an application instrumented with LTTng-UST and allow it to block:
6043 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6047 .Create a Linux kernel channel which rotates eight trace files of 4{nbsp}MiB each for each stream
6051 $ lttng enable-channel --kernel --tracefile-count=8 \
6052 --tracefile-size=4194304 my-channel
6056 .Create a user space channel in overwrite (or _flight recorder_) mode.
6060 $ lttng enable-channel --userspace --overwrite my-channel
6064 You can <<enabling-disabling-events,create>> the same event rule in
6065 two different channels:
6069 $ lttng enable-event --userspace --channel=my-channel app:tp
6070 $ lttng enable-event --userspace --channel=other-channel app:tp
6073 If both channels are enabled, when a tracepoint named `app:tp` is
6074 reached, LTTng records two events, one for each channel.
6078 === Disable a channel
6080 To disable a specific channel that you <<enabling-disabling-channels,created>>
6081 previously, use the man:lttng-disable-channel(1) command.
6083 .Disable a specific Linux kernel channel.
6087 $ lttng disable-channel --kernel my-channel
6091 The state of a channel precedes the individual states of event rules
6092 attached to it: event rules which belong to a disabled channel, even if
6093 they are enabled, are also considered disabled.
6097 === Add context fields to a channel
6099 Event record fields in trace files provide important information about
6100 events that occured previously, but sometimes some external context may
6101 help you solve a problem faster. Examples of context fields are:
6103 * The **process ID**, **thread ID**, **process name**, and
6104 **process priority** of the thread in which the event occurs.
6105 * The **hostname** of the system on which the event occurs.
6106 * The Linux kernel and user call stacks (since
6107 LTTng{nbsp}{revision}).
6108 * The current values of many possible **performance counters** using
6110 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6112 ** Branch instructions, misses, and loads.
6114 * Any context defined at the application level (supported for the
6115 JUL and log4j <<domain,tracing domains>>).
6117 To get the full list of available context fields, see
6118 `lttng add-context --list`. Some context fields are reserved for a
6119 specific <<domain,tracing domain>> (Linux kernel or user space).
6121 You add context fields to <<channel,channels>>. All the events
6122 that a channel with added context fields records contain those fields.
6124 To add context fields to one or all the channels of a given tracing
6127 * Use the man:lttng-add-context(1) command.
6129 .Add context fields to all the channels of the current tracing session.
6131 The following command line adds the virtual process identifier and
6132 the per-thread CPU cycles count fields to all the user space channels
6133 of the current tracing session.
6137 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6141 .Add performance counter context fields by raw ID
6143 See man:lttng-add-context(1) for the exact format of the context field
6144 type, which is partly compatible with the format used in
6149 $ lttng add-context --userspace --type=perf:thread:raw:r0110:test
6150 $ lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6154 .Add context fields to a specific channel.
6156 The following command line adds the thread identifier and user call
6157 stack context fields to the Linux kernel channel named `my-channel` in
6158 the current tracing session.
6162 $ lttng add-context --kernel --channel=my-channel \
6163 --type=tid --type=callstack-user
6167 .Add an application-specific context field to a specific channel.
6169 The following command line adds the `cur_msg_id` context field of the
6170 `retriever` context retriever for all the instrumented
6171 <<java-application,Java applications>> recording <<event,event records>>
6172 in the channel named `my-channel`:
6176 $ lttng add-context --kernel --channel=my-channel \
6177 --type='$app:retriever:cur_msg_id'
6180 IMPORTANT: Make sure to always quote the `$` character when you
6181 use man:lttng-add-context(1) from a shell.
6184 NOTE: You cannot remove context fields from a channel once you add it.
6189 === Track process IDs
6191 It's often useful to allow only specific process IDs (PIDs) to emit
6192 events. For example, you may wish to record all the system calls made by
6193 a given process (Ă la http://linux.die.net/man/1/strace[strace]).
6195 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6196 purpose. Both commands operate on a whitelist of process IDs. You _add_
6197 entries to this whitelist with the man:lttng-track(1) command and remove
6198 entries with the man:lttng-untrack(1) command. Any process which has one
6199 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6200 an enabled <<event,event rule>>.
6202 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6203 process with a given tracked ID exit and another process be given this
6204 ID, then the latter would also be allowed to emit events.
6206 .Track and untrack process IDs.
6208 For the sake of the following example, assume the target system has
6209 16{nbsp}possible PIDs.
6212 <<creating-destroying-tracing-sessions,create a tracing session>>,
6213 the whitelist contains all the possible PIDs:
6216 .All PIDs are tracked.
6217 image::track-all.png[]
6219 When the whitelist is full and you use the man:lttng-track(1) command to
6220 specify some PIDs to track, LTTng first clears the whitelist, then it
6221 tracks the specific PIDs. After:
6225 $ lttng track --pid=3,4,7,10,13
6231 .PIDs 3, 4, 7, 10, and 13 are tracked.
6232 image::track-3-4-7-10-13.png[]
6234 You can add more PIDs to the whitelist afterwards:
6238 $ lttng track --pid=1,15,16
6244 .PIDs 1, 15, and 16 are added to the whitelist.
6245 image::track-1-3-4-7-10-13-15-16.png[]
6247 The man:lttng-untrack(1) command removes entries from the PID tracker's
6248 whitelist. Given the previous example, the following command:
6252 $ lttng untrack --pid=3,7,10,13
6255 leads to this whitelist:
6258 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6259 image::track-1-4-15-16.png[]
6261 LTTng can track all possible PIDs again using the
6262 opt:lttng-track(1):--all option:
6266 $ lttng track --pid --all
6269 The result is, again:
6272 .All PIDs are tracked.
6273 image::track-all.png[]
6276 .Track only specific PIDs
6278 A very typical use case with PID tracking is to start with an empty
6279 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6280 then add PIDs manually while tracers are active. You can accomplish this
6281 by using the opt:lttng-untrack(1):--all option of the
6282 man:lttng-untrack(1) command to clear the whitelist after you
6283 <<creating-destroying-tracing-sessions,create a tracing session>>:
6287 $ lttng untrack --pid --all
6293 .No PIDs are tracked.
6294 image::untrack-all.png[]
6296 If you trace with this whitelist configuration, the tracer records no
6297 events for this <<domain,tracing domain>> because no processes are
6298 tracked. You can use the man:lttng-track(1) command as usual to track
6299 specific PIDs, for example:
6303 $ lttng track --pid=6,11
6309 .PIDs 6 and 11 are tracked.
6310 image::track-6-11.png[]
6315 [[saving-loading-tracing-session]]
6316 === Save and load tracing session configurations
6318 Configuring a <<tracing-session,tracing session>> can be long. Some of
6319 the tasks involved are:
6321 * <<enabling-disabling-channels,Create channels>> with
6322 specific attributes.
6323 * <<adding-context,Add context fields>> to specific channels.
6324 * <<enabling-disabling-events,Create event rules>> with specific log
6325 level and filter conditions.
6327 If you use LTTng to solve real world problems, chances are you have to
6328 record events using the same tracing session setup over and over,
6329 modifying a few variables each time in your instrumented program
6330 or environment. To avoid constant tracing session reconfiguration,
6331 the man:lttng(1) command-line tool can save and load tracing session
6332 configurations to/from XML files.
6334 To save a given tracing session configuration:
6336 * Use the man:lttng-save(1) command:
6341 $ lttng save my-session
6345 Replace `my-session` with the name of the tracing session to save.
6347 LTTng saves tracing session configurations to
6348 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6349 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6350 the opt:lttng-save(1):--output-path option to change this destination
6353 LTTng saves all configuration parameters, for example:
6355 * The tracing session name.
6356 * The trace data output path.
6357 * The channels with their state and all their attributes.
6358 * The context fields you added to channels.
6359 * The event rules with their state, log level and filter conditions.
6361 To load a tracing session:
6363 * Use the man:lttng-load(1) command:
6368 $ lttng load my-session
6372 Replace `my-session` with the name of the tracing session to load.
6374 When LTTng loads a configuration, it restores your saved tracing session
6375 as if you just configured it manually.
6377 See man:lttng(1) for the complete list of command-line options. You
6378 can also save and load all many sessions at a time, and decide in which
6379 directory to output the XML files.
6382 [[sending-trace-data-over-the-network]]
6383 === Send trace data over the network
6385 LTTng can send the recorded trace data to a remote system over the
6386 network instead of writing it to the local file system.
6388 To send the trace data over the network:
6390 . On the _remote_ system (which can also be the target system),
6391 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6400 . On the _target_ system, create a tracing session configured to
6401 send trace data over the network:
6406 $ lttng create my-session --set-url=net://remote-system
6410 Replace `remote-system` by the host name or IP address of the
6411 remote system. See man:lttng-create(1) for the exact URL format.
6413 . On the target system, use the man:lttng(1) command-line tool as usual.
6414 When tracing is active, the target's consumer daemon sends sub-buffers
6415 to the relay daemon running on the remote system instead of flushing
6416 them to the local file system. The relay daemon writes the received
6417 packets to the local file system.
6419 The relay daemon writes trace files to
6420 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6421 +__hostname__+ is the host name of the target system and +__session__+
6422 is the tracing session name. Note that the env:LTTNG_HOME environment
6423 variable defaults to `$HOME` if not set. Use the
6424 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6425 trace files to another base directory.
6430 === View events as LTTng emits them (noch:{LTTng} live)
6432 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6433 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6434 display events as LTTng emits them on the target system while tracing is
6437 The relay daemon creates a _tee_: it forwards the trace data to both
6438 the local file system and to connected live viewers:
6441 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6446 . On the _target system_, create a <<tracing-session,tracing session>>
6452 $ lttng create my-session --live
6456 This spawns a local relay daemon.
6458 . Start the live viewer and configure it to connect to the relay
6459 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6464 $ babeltrace --input-format=lttng-live \
6465 net://localhost/host/hostname/my-session
6472 * `hostname` with the host name of the target system.
6473 * `my-session` with the name of the tracing session to view.
6476 . Configure the tracing session as usual with the man:lttng(1)
6477 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6479 You can list the available live tracing sessions with Babeltrace:
6483 $ babeltrace --input-format=lttng-live net://localhost
6486 You can start the relay daemon on another system. In this case, you need
6487 to specify the relay daemon's URL when you create the tracing session
6488 with the opt:lttng-create(1):--set-url option. You also need to replace
6489 `localhost` in the procedure above with the host name of the system on
6490 which the relay daemon is running.
6492 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6493 command-line options.
6497 [[taking-a-snapshot]]
6498 === Take a snapshot of the current sub-buffers of a tracing session
6500 The normal behavior of LTTng is to append full sub-buffers to growing
6501 trace data files. This is ideal to keep a full history of the events
6502 that occurred on the target system, but it can
6503 represent too much data in some situations. For example, you may wish
6504 to trace your application continuously until some critical situation
6505 happens, in which case you only need the latest few recorded
6506 events to perform the desired analysis, not multi-gigabyte trace files.
6508 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6509 current sub-buffers of a given <<tracing-session,tracing session>>.
6510 LTTng can write the snapshot to the local file system or send it over
6514 .A snapshot is a copy of the current sub-buffers, which are not cleared after the operation.
6515 image::snapshot.png[]
6517 If you wish to create unmanaged, self-contained, non-overlapping
6518 trace chunk archives instead of a simple copy of the current
6519 sub-buffers, see the <<session-rotation,tracing session rotation>>
6520 feature (available since LTTng{nbsp}2.11).
6524 . Create a tracing session in _snapshot mode_:
6529 $ lttng create my-session --snapshot
6533 The <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> of
6534 <<channel,channels>> created in this mode is automatically set to
6535 _overwrite_ (flight recorder mode).
6537 . Configure the tracing session as usual with the man:lttng(1)
6538 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6540 . **Optional**: When you need to take a snapshot,
6541 <<basic-tracing-session-control,stop tracing>>.
6543 You can take a snapshot when the tracers are active, but if you stop
6544 them first, you are sure that the data in the sub-buffers does not
6545 change before you actually take the snapshot.
6552 $ lttng snapshot record --name=my-first-snapshot
6556 LTTng writes the current sub-buffers of all the current tracing
6557 session's channels to trace files on the local file system. Those trace
6558 files have `my-first-snapshot` in their name.
6560 There is no difference between the format of a normal trace file and the
6561 format of a snapshot: viewers of LTTng traces also support LTTng
6564 By default, LTTng writes snapshot files to the path shown by
6565 `lttng snapshot list-output`. You can change this path or decide to send
6566 snapshots over the network using either:
6568 . An output path or URL that you specify when you
6569 <<creating-destroying-tracing-sessions,create the tracing session>>.
6570 . A snapshot output path or URL that you add using
6571 `lttng snapshot add-output`.
6572 . An output path or URL that you provide directly to the
6573 `lttng snapshot record` command.
6575 Method{nbsp}3 overrides method{nbsp}2, which overrides method 1. When
6576 you specify a URL, a relay daemon must listen on a remote system (see
6577 <<sending-trace-data-over-the-network,Send trace data over the
6582 [[session-rotation]]
6583 === Archive the current trace chunk (rotate a tracing session)
6585 The <<taking-a-snapshot,snapshot user guide>> shows how you can dump
6586 a tracing session's current sub-buffers to the file system or send them
6587 over the network. When you take a snapshot, LTTng does not clear the
6588 tracing session's ring buffers: if you take another snapshot immediately
6589 after, both snapshots could contain overlapping trace data.
6591 Inspired by https://en.wikipedia.org/wiki/Log_rotation[log rotation],
6592 _tracing session rotation_ is a feature which appends the content of the
6593 ring buffers to what's already on the file system or sent over the
6594 network since the tracing session's creation or since the last
6595 rotation, and then clears those ring buffers to avoid trace data
6598 What LTTng is about to write when performing a tracing session rotation
6599 is called the _current trace chunk_. When this current trace chunk is
6600 written to the file system or sent over the network, it becomes a _trace
6601 chunk archive_. Therefore, a tracing session rotation _archives_ the
6602 current trace chunk.
6605 .A tracing session rotation operation _archives_ the current trace chunk.
6606 image::rotation.png[]
6608 A trace chunk archive is a self-contained LTTng trace which LTTng
6609 doesn't manage anymore: you can read it, modify it, move it, or remove
6612 There are two methods to perform a tracing session rotation: immediately
6613 or with a rotation schedule.
6615 To perform an immediate tracing session rotation:
6617 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6618 in _normal mode_ or _network streaming mode_
6619 (only those two creation modes support tracing session rotation):
6624 $ lttng create my-session
6628 . <<enabling-disabling-events,Create one or more event rules>>
6629 and <<basic-tracing-session-control,start tracing>>:
6634 $ lttng enable-event --kernel sched_'*'
6639 . When needed, immediately rotate the current tracing session:
6648 The cmd:lttng-rotate command prints the path to the created trace
6649 chunk archive. See man:lttng-rotate(1) to learn about the format
6650 of trace chunk archive directory names.
6652 You can perform other immediate rotations while the tracing session is
6653 active. It is guaranteed that all the trace chunk archives do not
6654 contain overlapping trace data. You can also perform an immediate
6655 rotation once you have <<basic-tracing-session-control,stopped>> the
6658 . When you are done tracing,
6659 <<creating-destroying-tracing-sessions,destroy the current tracing
6669 The tracing session destruction operation creates one last trace
6670 chunk archive from the current trace chunk.
6672 A tracing session rotation schedule is a planned rotation which LTTng
6673 performs automatically based on one of the following conditions:
6675 * A timer with a configured period times out.
6677 * The total size of the flushed part of the current trace chunk
6678 becomes greater than or equal to a configured value.
6680 To schedule a tracing session rotation, set a _rotation schedule_:
6682 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6683 in _normal mode_ or _network streaming mode_
6684 (only those two creation modes support tracing session rotation):
6689 $ lttng create my-session
6693 . <<enabling-disabling-events,Create one or more event rules>>:
6698 $ lttng enable-event --kernel sched_'*'
6702 . Set a tracing session rotation schedule:
6707 $ lttng enable-rotation --timer=10s
6711 In this example, we set a rotation schedule so that LTTng performs a
6712 tracing session rotation every ten seconds.
6714 See man:lttng-enable-rotation(1) to learn more about other ways to set a
6717 . <<basic-tracing-session-control,Start tracing>>:
6726 LTTng performs tracing session rotations automatically while the tracing
6727 session is active thanks to the rotation schedule.
6729 . When you are done tracing,
6730 <<creating-destroying-tracing-sessions,destroy the current tracing
6740 The tracing session destruction operation creates one last trace chunk
6741 archive from the current trace chunk.
6743 You can use man:lttng-disable-rotation(1) to unset a tracing session
6746 NOTE: man:lttng-rotate(1) and man:lttng-enable-rotation(1) list
6747 limitations regarding those two commands.
6752 === Use the machine interface
6754 With any command of the man:lttng(1) command-line tool, you can set the
6755 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6756 XML machine interface output, for example:
6760 $ lttng --mi=xml enable-event --kernel --syscall open
6763 A schema definition (XSD) is
6764 https://github.com/lttng/lttng-tools/blob/stable-2.11/src/common/mi-lttng-3.0.xsd[available]
6765 to ease the integration with external tools as much as possible.
6769 [[metadata-regenerate]]
6770 === Regenerate the metadata of an LTTng trace
6772 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6773 data stream files and a metadata file. This metadata file contains,
6774 amongst other things, information about the offset of the clock sources
6775 used to timestamp <<event,event records>> when tracing.
6777 If, once a <<tracing-session,tracing session>> is
6778 <<basic-tracing-session-control,started>>, a major
6779 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6780 happens, the trace's clock offset also needs to be updated. You
6781 can use the `metadata` item of the man:lttng-regenerate(1) command
6784 The main use case of this command is to allow a system to boot with
6785 an incorrect wall time and trace it with LTTng before its wall time
6786 is corrected. Once the system is known to be in a state where its
6787 wall time is correct, it can run `lttng regenerate metadata`.
6789 To regenerate the metadata of an LTTng trace:
6791 * Use the `metadata` item of the man:lttng-regenerate(1) command:
6796 $ lttng regenerate metadata
6802 `lttng regenerate metadata` has the following limitations:
6804 * Tracing session <<creating-destroying-tracing-sessions,created>>
6806 * User space <<channel,channels>>, if any, are using
6807 <<channel-buffering-schemes,per-user buffering>>.
6812 [[regenerate-statedump]]
6813 === Regenerate the state dump of a tracing session
6815 The LTTng kernel and user space tracers generate state dump
6816 <<event,event records>> when the application starts or when you
6817 <<basic-tracing-session-control,start a tracing session>>. An analysis
6818 can use the state dump event records to set an initial state before it
6819 builds the rest of the state from the following event records.
6820 http://tracecompass.org/[Trace Compass] is a notable example of an
6821 application which uses the state dump of an LTTng trace.
6823 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
6824 state dump event records are not included in the snapshot because they
6825 were recorded to a sub-buffer that has been consumed or overwritten
6828 You can use the `lttng regenerate statedump` command to emit the state
6829 dump event records again.
6831 To regenerate the state dump of the current tracing session, provided
6832 create it in snapshot mode, before you take a snapshot:
6834 . Use the `statedump` item of the man:lttng-regenerate(1) command:
6839 $ lttng regenerate statedump
6843 . <<basic-tracing-session-control,Stop the tracing session>>:
6852 . <<taking-a-snapshot,Take a snapshot>>:
6857 $ lttng snapshot record --name=my-snapshot
6861 Depending on the event throughput, you should run steps 1 and 2
6862 as closely as possible.
6864 NOTE: To record the state dump events, you need to
6865 <<enabling-disabling-events,create event rules>> which enable them.
6866 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
6867 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
6871 [[persistent-memory-file-systems]]
6872 === Record trace data on persistent memory file systems
6874 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6875 (NVRAM) is random-access memory that retains its information when power
6876 is turned off (non-volatile). Systems with such memory can store data
6877 structures in RAM and retrieve them after a reboot, without flushing
6878 to typical _storage_.
6880 Linux supports NVRAM file systems thanks to either
6881 http://pramfs.sourceforge.net/[PRAMFS] or
6882 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6883 (requires Linux{nbsp}4.1+).
6885 This section does not describe how to operate such file systems;
6886 we assume that you have a working persistent memory file system.
6888 When you create a <<tracing-session,tracing session>>, you can specify
6889 the path of the shared memory holding the sub-buffers. If you specify a
6890 location on an NVRAM file system, then you can retrieve the latest
6891 recorded trace data when the system reboots after a crash.
6893 To record trace data on a persistent memory file system and retrieve the
6894 trace data after a system crash:
6896 . Create a tracing session with a sub-buffer shared memory path located
6897 on an NVRAM file system:
6902 $ lttng create my-session --shm-path=/path/to/shm
6906 . Configure the tracing session as usual with the man:lttng(1)
6907 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6909 . After a system crash, use the man:lttng-crash(1) command-line tool to
6910 view the trace data recorded on the NVRAM file system:
6915 $ lttng-crash /path/to/shm
6919 The binary layout of the ring buffer files is not exactly the same as
6920 the trace files layout. This is why you need to use man:lttng-crash(1)
6921 instead of your preferred trace viewer directly.
6923 To convert the ring buffer files to LTTng trace files:
6925 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6930 $ lttng-crash --extract=/path/to/trace /path/to/shm
6936 [[notif-trigger-api]]
6937 === Get notified when a channel's buffer usage is too high or too low
6939 With LTTng's $$C/C++$$ notification and trigger API, your user
6940 application can get notified when the buffer usage of one or more
6941 <<channel,channels>> becomes too low or too high. You can use this API
6942 and enable or disable <<event,event rules>> during tracing to avoid
6943 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>.
6945 .Have a user application get notified when an LTTng channel's buffer usage is too high.
6947 In this example, we create and build an application which gets notified
6948 when the buffer usage of a specific LTTng channel is higher than
6949 75{nbsp}%. We only print that it is the case in the example, but we
6950 could as well use the API of <<liblttng-ctl-lttng,`liblttng-ctl`>> to
6951 disable event rules when this happens.
6953 . Create the application's C source file:
6961 #include <lttng/domain.h>
6962 #include <lttng/action/action.h>
6963 #include <lttng/action/notify.h>
6964 #include <lttng/condition/condition.h>
6965 #include <lttng/condition/buffer-usage.h>
6966 #include <lttng/condition/evaluation.h>
6967 #include <lttng/notification/channel.h>
6968 #include <lttng/notification/notification.h>
6969 #include <lttng/trigger/trigger.h>
6970 #include <lttng/endpoint.h>
6972 int main(int argc, char *argv[])
6974 int exit_status = 0;
6975 struct lttng_notification_channel *notification_channel;
6976 struct lttng_condition *condition;
6977 struct lttng_action *action;
6978 struct lttng_trigger *trigger;
6979 const char *tracing_session_name;
6980 const char *channel_name;
6983 tracing_session_name = argv[1];
6984 channel_name = argv[2];
6987 * Create a notification channel. A notification channel
6988 * connects the user application to the LTTng session daemon.
6989 * This notification channel can be used to listen to various
6990 * types of notifications.
6992 notification_channel = lttng_notification_channel_create(
6993 lttng_session_daemon_notification_endpoint);
6996 * Create a "high buffer usage" condition. In this case, the
6997 * condition is reached when the buffer usage is greater than or
6998 * equal to 75 %. We create the condition for a specific tracing
6999 * session name, channel name, and for the user space tracing
7002 * The "low buffer usage" condition type also exists.
7004 condition = lttng_condition_buffer_usage_high_create();
7005 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
7006 lttng_condition_buffer_usage_set_session_name(
7007 condition, tracing_session_name);
7008 lttng_condition_buffer_usage_set_channel_name(condition,
7010 lttng_condition_buffer_usage_set_domain_type(condition,
7014 * Create an action (get a notification) to take when the
7015 * condition created above is reached.
7017 action = lttng_action_notify_create();
7020 * Create a trigger. A trigger associates a condition to an
7021 * action: the action is executed when the condition is reached.
7023 trigger = lttng_trigger_create(condition, action);
7025 /* Register the trigger to LTTng. */
7026 lttng_register_trigger(trigger);
7029 * Now that we have registered a trigger, a notification will be
7030 * emitted everytime its condition is met. To receive this
7031 * notification, we must subscribe to notifications that match
7032 * the same condition.
7034 lttng_notification_channel_subscribe(notification_channel,
7038 * Notification loop. You can put this in a dedicated thread to
7039 * avoid blocking the main thread.
7042 struct lttng_notification *notification;
7043 enum lttng_notification_channel_status status;
7044 const struct lttng_evaluation *notification_evaluation;
7045 const struct lttng_condition *notification_condition;
7046 double buffer_usage;
7048 /* Receive the next notification. */
7049 status = lttng_notification_channel_get_next_notification(
7050 notification_channel, ¬ification);
7053 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
7055 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
7057 * The session daemon can drop notifications if
7058 * a monitoring application is not consuming the
7059 * notifications fast enough.
7062 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
7064 * The notification channel has been closed by the
7065 * session daemon. This is typically caused by a session
7066 * daemon shutting down.
7070 /* Unhandled conditions or errors. */
7076 * A notification provides, amongst other things:
7078 * * The condition that caused this notification to be
7080 * * The condition evaluation, which provides more
7081 * specific information on the evaluation of the
7084 * The condition evaluation provides the buffer usage
7085 * value at the moment the condition was reached.
7087 notification_condition = lttng_notification_get_condition(
7089 notification_evaluation = lttng_notification_get_evaluation(
7092 /* We're subscribed to only one condition. */
7093 assert(lttng_condition_get_type(notification_condition) ==
7094 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
7097 * Get the exact sampled buffer usage from the
7098 * condition evaluation.
7100 lttng_evaluation_buffer_usage_get_usage_ratio(
7101 notification_evaluation, &buffer_usage);
7104 * At this point, instead of printing a message, we
7105 * could do something to reduce the channel's buffer
7106 * usage, like disable specific events.
7108 printf("Buffer usage is %f %% in tracing session \"%s\", "
7109 "user space channel \"%s\".\n", buffer_usage * 100,
7110 tracing_session_name, channel_name);
7111 lttng_notification_destroy(notification);
7115 lttng_action_destroy(action);
7116 lttng_condition_destroy(condition);
7117 lttng_trigger_destroy(trigger);
7118 lttng_notification_channel_destroy(notification_channel);
7124 . Build the `notif-app` application, linking it to `liblttng-ctl`:
7129 $ gcc -o notif-app notif-app.c -llttng-ctl
7133 . <<creating-destroying-tracing-sessions,Create a tracing session>>,
7134 <<enabling-disabling-events,create an event rule>> matching all the
7135 user space tracepoints, and
7136 <<basic-tracing-session-control,start tracing>>:
7141 $ lttng create my-session
7142 $ lttng enable-event --userspace --all
7147 If you create the channel manually with the man:lttng-enable-channel(1)
7148 command, you can control how frequently are the current values of the
7149 channel's properties sampled to evaluate user conditions with the
7150 opt:lttng-enable-channel(1):--monitor-timer option.
7152 . Run the `notif-app` application. This program accepts the
7153 <<tracing-session,tracing session>> name and the user space channel
7154 name as its two first arguments. The channel which LTTng automatically
7155 creates with the man:lttng-enable-event(1) command above is named
7161 $ ./notif-app my-session channel0
7165 . In another terminal, run an application with a very high event
7166 throughput so that the 75{nbsp}% buffer usage condition is reached.
7168 In the first terminal, the application should print lines like this:
7171 Buffer usage is 81.45197 % in tracing session "my-session", user space
7175 If you don't see anything, try modifying the condition in
7176 path:{notif-app.c} to a lower value (0.1, for example), rebuilding it
7177 (step{nbsp}2) and running it again (step{nbsp}4).
7184 [[lttng-modules-ref]]
7185 === noch:{LTTng-modules}
7189 [[lttng-tracepoint-enum]]
7190 ==== `LTTNG_TRACEPOINT_ENUM()` usage
7192 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
7196 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
7201 * `name` with the name of the enumeration (C identifier, unique
7202 amongst all the defined enumerations).
7203 * `entries` with a list of enumeration entries.
7205 The available enumeration entry macros are:
7207 +ctf_enum_value(__name__, __value__)+::
7208 Entry named +__name__+ mapped to the integral value +__value__+.
7210 +ctf_enum_range(__name__, __begin__, __end__)+::
7211 Entry named +__name__+ mapped to the range of integral values between
7212 +__begin__+ (included) and +__end__+ (included).
7214 +ctf_enum_auto(__name__)+::
7215 Entry named +__name__+ mapped to the integral value following the
7216 last mapping's value.
7218 The last value of a `ctf_enum_value()` entry is its +__value__+
7221 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
7223 If `ctf_enum_auto()` is the first entry in the list, its integral
7226 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
7227 to use a defined enumeration as a tracepoint field.
7229 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
7233 LTTNG_TRACEPOINT_ENUM(
7236 ctf_enum_auto("AUTO: EXPECT 0")
7237 ctf_enum_value("VALUE: 23", 23)
7238 ctf_enum_value("VALUE: 27", 27)
7239 ctf_enum_auto("AUTO: EXPECT 28")
7240 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
7241 ctf_enum_auto("AUTO: EXPECT 304")
7249 [[lttng-modules-tp-fields]]
7250 ==== Tracepoint fields macros (for `TP_FIELDS()`)
7252 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
7253 tracepoint fields, which must be listed within `TP_FIELDS()` in
7254 `LTTNG_TRACEPOINT_EVENT()`, are:
7256 [role="func-desc growable",cols="asciidoc,asciidoc"]
7257 .Available macros to define LTTng-modules tracepoint fields
7259 |Macro |Description and parameters
7262 +ctf_integer(__t__, __n__, __e__)+
7264 +ctf_integer_nowrite(__t__, __n__, __e__)+
7266 +ctf_user_integer(__t__, __n__, __e__)+
7268 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
7270 Standard integer, displayed in base{nbsp}10.
7273 Integer C type (`int`, `long`, `size_t`, ...).
7279 Argument expression.
7282 +ctf_integer_hex(__t__, __n__, __e__)+
7284 +ctf_user_integer_hex(__t__, __n__, __e__)+
7286 Standard integer, displayed in base{nbsp}16.
7295 Argument expression.
7297 |+ctf_integer_oct(__t__, __n__, __e__)+
7299 Standard integer, displayed in base{nbsp}8.
7308 Argument expression.
7311 +ctf_integer_network(__t__, __n__, __e__)+
7313 +ctf_user_integer_network(__t__, __n__, __e__)+
7315 Integer in network byte order (big-endian), displayed in base{nbsp}10.
7324 Argument expression.
7327 +ctf_integer_network_hex(__t__, __n__, __e__)+
7329 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
7331 Integer in network byte order, displayed in base{nbsp}16.
7340 Argument expression.
7343 +ctf_enum(__N__, __t__, __n__, __e__)+
7345 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
7347 +ctf_user_enum(__N__, __t__, __n__, __e__)+
7349 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
7354 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
7357 Integer C type (`int`, `long`, `size_t`, ...).
7363 Argument expression.
7366 +ctf_string(__n__, __e__)+
7368 +ctf_string_nowrite(__n__, __e__)+
7370 +ctf_user_string(__n__, __e__)+
7372 +ctf_user_string_nowrite(__n__, __e__)+
7374 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7380 Argument expression.
7383 +ctf_array(__t__, __n__, __e__, __s__)+
7385 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7387 +ctf_user_array(__t__, __n__, __e__, __s__)+
7389 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7391 Statically-sized array of integers.
7394 Array element C type.
7400 Argument expression.
7406 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7408 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7410 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7412 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7414 Statically-sized array of bits.
7416 The type of +__e__+ must be an integer type. +__s__+ is the number
7417 of elements of such type in +__e__+, not the number of bits.
7420 Array element C type.
7426 Argument expression.
7432 +ctf_array_text(__t__, __n__, __e__, __s__)+
7434 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7436 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7438 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7440 Statically-sized array, printed as text.
7442 The string does not need to be null-terminated.
7445 Array element C type (always `char`).
7451 Argument expression.
7457 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7459 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7461 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7463 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7465 Dynamically-sized array of integers.
7467 The type of +__E__+ must be unsigned.
7470 Array element C type.
7476 Argument expression.
7479 Length expression C type.
7485 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7487 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7489 Dynamically-sized array of integers, displayed in base{nbsp}16.
7491 The type of +__E__+ must be unsigned.
7494 Array element C type.
7500 Argument expression.
7503 Length expression C type.
7508 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7510 Dynamically-sized array of integers in network byte order (big-endian),
7511 displayed in base{nbsp}10.
7513 The type of +__E__+ must be unsigned.
7516 Array element C type.
7522 Argument expression.
7525 Length expression C type.
7531 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7533 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7535 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7537 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7539 Dynamically-sized array of bits.
7541 The type of +__e__+ must be an integer type. +__s__+ is the number
7542 of elements of such type in +__e__+, not the number of bits.
7544 The type of +__E__+ must be unsigned.
7547 Array element C type.
7553 Argument expression.
7556 Length expression C type.
7562 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7564 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7566 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7568 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7570 Dynamically-sized array, displayed as text.
7572 The string does not need to be null-terminated.
7574 The type of +__E__+ must be unsigned.
7576 The behaviour is undefined if +__e__+ is `NULL`.
7579 Sequence element C type (always `char`).
7585 Argument expression.
7588 Length expression C type.
7594 Use the `_user` versions when the argument expression, `e`, is
7595 a user space address. In the cases of `ctf_user_integer*()` and
7596 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7599 The `_nowrite` versions omit themselves from the session trace, but are
7600 otherwise identical. This means the `_nowrite` fields won't be written
7601 in the recorded trace. Their primary purpose is to make some
7602 of the event context available to the
7603 <<enabling-disabling-events,event filters>> without having to
7604 commit the data to sub-buffers.
7610 Terms related to LTTng and to tracing in general:
7613 The http://diamon.org/babeltrace[Babeltrace] project, which includes:
7615 * The cmd:babeltrace (Babeltrace{nbsp}1) or cmd:babeltrace2
7616 (Babeltrace{nbsp}2) command.
7617 * Libraries with a C{nbsp}API.
7618 * Python{nbsp}3 bindings.
7619 * Plugins (Babeltrace{nbsp}2).
7621 [[def-buffering-scheme]]<<channel-buffering-schemes,buffering scheme>>::
7622 A layout of <<def-sub-buffer,sub-buffers>> applied to a given channel.
7624 [[def-channel]]<<channel,channel>>::
7625 An entity which is responsible for a set of
7626 <<def-ring-buffer,ring buffers>>.
7628 <<def-event-rule,Event rules>> are always attached to a specific
7632 A source of time for a <<def-tracer,tracer>>.
7634 [[def-consumer-daemon]]<<lttng-consumerd,consumer daemon>>::
7635 A process which is responsible for consuming the full
7636 <<def-sub-buffer,sub-buffers>> and write them to a file system or
7637 send them over the network.
7639 [[def-current-trace-chunk]]current trace chunk::
7640 A <<def-trace-chunk,trace chunk>> which includes the current content
7641 of all the <<def-tracing-session-rotation,tracing session>>'s
7642 <<def-sub-buffer,sub-buffers>> and the stream files produced since the
7643 latest event amongst:
7645 * The creation of the <<def-tracing-session,tracing session>>.
7646 * The last tracing session rotation, if any.
7648 <<channel-overwrite-mode-vs-discard-mode,discard mode>>::
7649 The <<def-event-record-loss-mode,event record loss mode>> in which
7650 the <<def-tracer,tracer>> _discards_ new event records when there's no
7651 <<def-sub-buffer,sub-buffer>> space left to store them.
7653 [[def-event]]event::
7654 The consequence of the execution of an
7655 <<def-instrumentation-point,instrumentation point>>, like a
7656 <<def-tracepoint,tracepoint>> that you manually place in some source
7657 code, or a Linux kernel kprobe.
7659 An event is said to _occur_ at a specific time. <<def-lttng,LTTng>> can
7660 take various actions upon the occurrence of an event, like record the
7661 event's payload to a <<def-sub-buffer,sub-buffer>>.
7663 [[def-event-name]]event name::
7664 The name of an <<def-event,event>>, which is also the name of the
7665 <<def-event-record,event record>>.
7667 This is also called the _instrumentation point name_.
7669 [[def-event-record]]event record::
7670 A record, in a <<def-trace,trace>>, of the payload of an
7671 <<def-event,event>> which occured.
7673 [[def-event-record-loss-mode]]<<channel-overwrite-mode-vs-discard-mode,event record loss mode>>::
7674 The mechanism by which event records of a given
7675 <<def-channel,channel>> are lost (not recorded) when there is no
7676 <<def-sub-buffer,sub-buffer>> space left to store them.
7678 [[def-event-rule]]<<event,event rule>>::
7679 Set of conditions which must be satisfied for one or more occuring
7680 <<def-event,events>> to be recorded.
7682 `java.util.logging`::
7684 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7686 <<instrumenting,instrumentation>>::
7687 The use of <<def-lttng,LTTng>> probes to make a piece of software
7690 [[def-instrumentation-point]]instrumentation point::
7691 A point in the execution path of a piece of software that, when
7692 reached by this execution, can emit an <<def-event,event>>.
7694 instrumentation point name::
7695 See _<<def-event-name,event name>>_.
7698 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7699 developed by the Apache Software Foundation.
7702 Level of severity of a log statement or user space
7703 <<def-instrumentation-point,instrumentation point>>.
7705 [[def-lttng]]LTTng::
7706 The _Linux Trace Toolkit: next generation_ project.
7708 <<lttng-cli,cmd:lttng>>::
7709 A command-line tool provided by the <<def-lttng-tools,LTTng-tools>>
7710 project which you can use to send and receive control messages to and
7711 from a <<def-session-daemon,session daemon>>.
7714 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7715 which is a set of analyzing programs that you can use to obtain a
7716 higher level view of an <<def-lttng,LTTng>> <<def-trace,trace>>.
7718 cmd:lttng-consumerd::
7719 The name of the <<def-consumer-daemon,consumer daemon>> program.
7722 A utility provided by the <<def-lttng-tools,LTTng-tools>> project
7723 which can convert <<def-ring-buffer,ring buffer>> files (usually
7724 <<persistent-memory-file-systems,saved on a persistent memory file
7725 system>>) to <<def-trace,trace>> files.
7727 See man:lttng-crash(1).
7729 LTTng Documentation::
7732 <<lttng-live,LTTng live>>::
7733 A communication protocol between the <<lttng-relayd,relay daemon>> and
7734 live viewers which makes it possible to see <<def-event-record,event
7735 records>> "live", as they are received by the
7736 <<def-relay-daemon,relay daemon>>.
7738 <<lttng-modules,LTTng-modules>>::
7739 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7740 which contains the Linux kernel modules to make the Linux kernel
7741 <<def-instrumentation-point,instrumentation points>> available for
7742 <<def-lttng,LTTng>> tracing.
7745 The name of the <<def-relay-daemon,relay daemon>> program.
7747 cmd:lttng-sessiond::
7748 The name of the <<def-session-daemon,session daemon>> program.
7750 [[def-lttng-tools]]LTTng-tools::
7751 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7752 contains the various programs and libraries used to
7753 <<controlling-tracing,control tracing>>.
7755 [[def-lttng-ust]]<<lttng-ust,LTTng-UST>>::
7756 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7757 contains libraries to instrument
7758 <<def-user-application,user applications>>.
7760 <<lttng-ust-agents,LTTng-UST Java agent>>::
7761 A Java package provided by the <<def-lttng-ust,LTTng-UST>> project to
7762 allow the LTTng instrumentation of `java.util.logging` and Apache
7763 log4j{nbsp}1.2 logging statements.
7765 <<lttng-ust-agents,LTTng-UST Python agent>>::
7766 A Python package provided by the <<def-lttng-ust,LTTng-UST>> project
7767 to allow the <<def-lttng,LTTng>> instrumentation of Python logging
7770 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7771 The <<def-event-record-loss-mode,event record loss mode>> in which new
7772 <<def-event-record,event records>> _overwrite_ older event records
7773 when there's no <<def-sub-buffer,sub-buffer>> space left to store
7776 <<channel-buffering-schemes,per-process buffering>>::
7777 A <<def-buffering-scheme,buffering scheme>> in which each instrumented
7778 process has its own <<def-sub-buffer,sub-buffers>> for a given user
7779 space <<def-channel,channel>>.
7781 <<channel-buffering-schemes,per-user buffering>>::
7782 A <<def-buffering-scheme,buffering scheme>> in which all the processes
7783 of a Unix user share the same <<def-sub-buffer,sub-buffers>> for a
7784 given user space <<def-channel,channel>>.
7786 [[def-relay-daemon]]<<lttng-relayd,relay daemon>>::
7787 A process which is responsible for receiving the <<def-trace,trace>>
7788 data which a distant <<def-consumer-daemon,consumer daemon>> sends.
7790 [[def-ring-buffer]]ring buffer::
7791 A set of <<def-sub-buffer,sub-buffers>>.
7794 See _<<def-tracing-session-rotation,tracing session rotation>>_.
7796 [[def-session-daemon]]<<lttng-sessiond,session daemon>>::
7797 A process which receives control commands from you and orchestrates
7798 the <<def-tracer,tracers>> and various <<def-lttng,LTTng>> daemons.
7800 <<taking-a-snapshot,snapshot>>::
7801 A copy of the current data of all the <<def-sub-buffer,sub-buffers>>
7802 of a given <<def-tracing-session,tracing session>>, saved as
7803 <<def-trace,trace>> files.
7805 [[def-sub-buffer]]sub-buffer::
7806 One part of an <<def-lttng,LTTng>> <<def-ring-buffer,ring buffer>>
7807 which contains <<def-event-record,event records>>.
7810 The time information attached to an <<def-event,event>> when it is
7813 [[def-trace]]trace (_noun_)::
7816 * One http://diamon.org/ctf/[CTF] metadata stream file.
7817 * One or more CTF data stream files which are the concatenations of one
7818 or more flushed <<def-sub-buffer,sub-buffers>>.
7820 [[def-trace-verb]]trace (_verb_)::
7821 The action of recording the <<def-event,events>> emitted by an
7822 application or by a system, or to initiate such recording by
7823 controlling a <<def-tracer,tracer>>.
7825 [[def-trace-chunk]]trace chunk::
7826 A self-contained <<def-trace,trace>> which is part of a
7827 <<def-tracing-session,tracing session>>. Each
7828 <<def-tracing-session-rotation, tracing session rotation>> produces a
7829 <<def-trace-chunk-archive,trace chunk archive>>.
7831 [[def-trace-chunk-archive]]trace chunk archive::
7832 The result of a <<def-tracing-session-rotation, tracing session rotation>>.
7834 <<def-lttng,LTTng>> does not manage any trace chunk archive, even if its
7835 containing <<def-tracing-session,tracing session>> is still active: you
7836 are free to read it, modify it, move it, or remove it.
7839 The http://tracecompass.org[Trace Compass] project and application.
7841 [[def-tracepoint]]tracepoint::
7842 An instrumentation point using the tracepoint mechanism of the Linux
7843 kernel or of <<def-lttng-ust,LTTng-UST>>.
7845 tracepoint definition::
7846 The definition of a single <<def-tracepoint,tracepoint>>.
7849 The name of a <<def-tracepoint,tracepoint>>.
7851 [[def-tracepoint-provider]]tracepoint provider::
7852 A set of functions providing <<def-tracepoint,tracepoints>> to an
7853 instrumented <<def-user-application,user application>>.
7855 Not to be confused with a <<def-tracepoint-provider-package,tracepoint
7856 provider package>>: many tracepoint providers can exist within a
7857 tracepoint provider package.
7859 [[def-tracepoint-provider-package]]tracepoint provider package::
7860 One or more <<def-tracepoint-provider,tracepoint providers>> compiled
7861 as an https://en.wikipedia.org/wiki/Object_file[object file] or as a
7862 link:https://en.wikipedia.org/wiki/Library_(computing)#Shared_libraries[shared
7865 [[def-tracer]]tracer::
7866 A software which records emitted <<def-event,events>>.
7868 <<domain,tracing domain>>::
7869 A namespace for <<def-event,event>> sources.
7871 <<tracing-group,tracing group>>::
7872 The Unix group in which a Unix user can be to be allowed to
7873 <<def-trace-verb,trace>> the Linux kernel.
7875 [[def-tracing-session]]<<tracing-session,tracing session>>::
7876 A stateful dialogue between you and a <<lttng-sessiond,session daemon>>.
7878 [[def-tracing-session-rotation]]<<session-rotation,tracing session rotation>>::
7879 The action of archiving the
7880 <<def-current-trace-chunk,current trace chunk>> of a
7881 <<def-tracing-session,tracing session>>.
7883 [[def-user-application]]user application::
7884 An application running in user space, as opposed to a Linux kernel
7885 module, for example.