1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
4 v2.11, 12 November 2018
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] (SDT) probe (DTrace-style marker). For example,
136 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 * The <<lttng-relayd,relay daemon>> is more efficient and presents fewer
189 connectivity issues, especially when a large number of targets send
190 trace data to a given relay daemon.
192 * User applications and libraries instrumented with
193 <<lttng-ust,LTTng-UST>> can now safely unload (man:dlclose(3)) a
195 <<building-tracepoint-providers-and-user-application,tracepoint
202 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
203 generation_ is a modern toolkit for tracing Linux systems and
204 applications. So your first question might be:
211 As the history of software engineering progressed and led to what
212 we now take for granted--complex, numerous and
213 interdependent software applications running in parallel on
214 sophisticated operating systems like Linux--the authors of such
215 components, software developers, began feeling a natural
216 urge to have tools that would ensure the robustness and good performance
217 of their masterpieces.
219 One major achievement in this field is, inarguably, the
220 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
221 an essential tool for developers to find and fix bugs. But even the best
222 debugger won't help make your software run faster, and nowadays, faster
223 software means either more work done by the same hardware, or cheaper
224 hardware for the same work.
226 A _profiler_ is often the tool of choice to identify performance
227 bottlenecks. Profiling is suitable to identify _where_ performance is
228 lost in a given software. The profiler outputs a profile, a statistical
229 summary of observed events, which you may use to discover which
230 functions took the most time to execute. However, a profiler won't
231 report _why_ some identified functions are the bottleneck. Bottlenecks
232 might only occur when specific conditions are met, conditions that are
233 sometimes impossible to capture by a statistical profiler, or impossible
234 to reproduce with an application altered by the overhead of an
235 event-based profiler. For a thorough investigation of software
236 performance issues, a history of execution is essential, with the
237 recorded values of variables and context fields you choose, and
238 with as little influence as possible on the instrumented software. This
239 is where tracing comes in handy.
241 _Tracing_ is a technique used to understand what goes on in a running
242 software system. The software used for tracing is called a _tracer_,
243 which is conceptually similar to a tape recorder. When recording,
244 specific instrumentation points placed in the software source code
245 generate events that are saved on a giant tape: a _trace_ file. You
246 can trace user applications and the operating system at the same time,
247 opening the possibility of resolving a wide range of problems that would
248 otherwise be extremely challenging.
250 Tracing is often compared to _logging_. However, tracers and loggers are
251 two different tools, serving two different purposes. Tracers are
252 designed to record much lower-level events that occur much more
253 frequently than log messages, often in the range of thousands per
254 second, with very little execution overhead. Logging is more appropriate
255 for a very high-level analysis of less frequent events: user accesses,
256 exceptional conditions (errors and warnings, for example), database
257 transactions, instant messaging communications, and such. Simply put,
258 logging is one of the many use cases that can be satisfied with tracing.
260 The list of recorded events inside a trace file can be read manually
261 like a log file for the maximum level of detail, but it is generally
262 much more interesting to perform application-specific analyses to
263 produce reduced statistics and graphs that are useful to resolve a
264 given problem. Trace viewers and analyzers are specialized tools
267 In the end, this is what LTTng is: a powerful, open source set of
268 tools to trace the Linux kernel and user applications at the same time.
269 LTTng is composed of several components actively maintained and
270 developed by its link:/community/#where[community].
273 [[lttng-alternatives]]
274 === Alternatives to noch:{LTTng}
276 Excluding proprietary solutions, a few competing software tracers
279 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
280 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
281 user scripts and is responsible for loading code into the
282 Linux kernel for further execution and collecting the outputted data.
283 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
284 subsystem in the Linux kernel in which a virtual machine can execute
285 programs passed from the user space to the kernel. You can attach
286 such programs to tracepoints and kprobes thanks to a system call, and
287 they can output data to the user space when executed thanks to
288 different mechanisms (pipe, VM register values, and eBPF maps, to name
290 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
291 is the de facto function tracer of the Linux kernel. Its user
292 interface is a set of special files in sysfs.
293 * https://perf.wiki.kernel.org/[perf] is
294 a performance analyzing tool for Linux which supports hardware
295 performance counters, tracepoints, as well as other counters and
296 types of probes. perf's controlling utility is the cmd:perf command
298 * http://linux.die.net/man/1/strace[strace]
299 is a command-line utility which records system calls made by a
300 user process, as well as signal deliveries and changes of process
301 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
302 to fulfill its function.
303 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
304 analyze Linux kernel events. You write scripts, or _chisels_ in
305 sysdig's jargon, in Lua and sysdig executes them while it traces the
306 system or afterwards. sysdig's interface is the cmd:sysdig
307 command-line tool as well as the curses-based cmd:csysdig tool.
308 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
309 user space tracer which uses custom user scripts to produce plain text
310 traces. SystemTap converts the scripts to the C language, and then
311 compiles them as Linux kernel modules which are loaded to produce
312 trace data. SystemTap's primary user interface is the cmd:stap
315 The main distinctive features of LTTng is that it produces correlated
316 kernel and user space traces, as well as doing so with the lowest
317 overhead amongst other solutions. It produces trace files in the
318 http://diamon.org/ctf[CTF] format, a file format optimized
319 for the production and analyses of multi-gigabyte data.
321 LTTng is the result of more than 10{nbsp}years of active open source
322 development by a community of passionate developers.
323 LTTng{nbsp}{revision} is currently available on major desktop and server
326 The main interface for tracing control is a single command-line tool
327 named cmd:lttng. The latter can create several tracing sessions, enable
328 and disable events on the fly, filter events efficiently with custom
329 user expressions, start and stop tracing, and much more. LTTng can
330 record the traces on the file system or send them over the network, and
331 keep them totally or partially. You can view the traces once tracing
332 becomes inactive or in real-time.
334 <<installing-lttng,Install LTTng now>> and
335 <<getting-started,start tracing>>!
341 **LTTng** is a set of software <<plumbing,components>> which interact to
342 <<instrumenting,instrument>> the Linux kernel and user applications, and
343 to <<controlling-tracing,control tracing>> (start and stop
344 tracing, enable and disable event rules, and the rest). Those
345 components are bundled into the following packages:
347 * **LTTng-tools**: Libraries and command-line interface to
349 * **LTTng-modules**: Linux kernel modules to instrument and
351 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
352 trace user applications.
354 Most distributions mark the LTTng-modules and LTTng-UST packages as
355 optional when installing LTTng-tools (which is always required). In the
356 following sections, we always provide the steps to install all three,
359 * You only need to install LTTng-modules if you intend to trace the
361 * You only need to install LTTng-UST if you intend to trace user
365 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 12 November 2018.
367 |Distribution |Available in releases |Alternatives
369 |https://www.ubuntu.com/[Ubuntu]
370 |Ubuntu{nbsp}14.04 _Trusty Tahr_, Ubuntu{nbsp}16.04 _Xenial Xerus_,
371 and Ubuntu{nbsp}18.04 _Bionic Beaver_:
372 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
373 |<<building-from-source,Build LTTng{nbsp}{revision} from source>>.
375 |https://getfedora.org/[Fedora]
378 link:/docs/v2.10#doc-fedora[LTTng{nbsp}2.10 for Fedora{nbsp}27,
379 Fedora{nbsp}28, and Fedora{nbsp}29].
381 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
383 |https://www.debian.org/[Debian]
385 |link:/docs/v2.10#doc-debian[LTTng{nbsp}2.10 for Debian "buster" (testing)].
387 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
389 |https://www.archlinux.org/[Arch Linux]
391 |<<building-from-source,Build LTTng{nbsp}{revision} from source>>.
393 |https://alpinelinux.org/[Alpine Linux]
396 link:/docs/v2.10#doc-alpine-linux[LTTng{nbsp}2.10 for
397 Alpine Linux{nbsp}3.7 and Alpine Linux{nbsp}3.8].
399 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
401 |https://www.opensuse.org/[openSUSE]
403 |link:/docs/v2.10#doc-opensuse[LTTng{nbsp}2.10 for openSUSE Leap{nbsp}15.0].
405 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
407 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
408 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
411 |https://buildroot.org/[Buildroot]
414 link:/docs/v2.10#doc-buildroot[LTTng{nbsp}2.10 for Buildroot{nbsp}2018.02,
415 Buildroot{nbsp}2018.05, Buildroot{nbsp}2018.08, and Buildroot{nbsp}2018.11].
417 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
419 |http://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
420 https://www.yoctoproject.org/[Yocto]
422 |<<building-from-source,Build LTTng{nbsp}{revision} from source>>.
427 === Ubuntu: noch:{LTTng} Stable {revision} PPA
429 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
430 Stable{nbsp}{revision} PPA] offers the latest stable
431 LTTng{nbsp}{revision} packages for:
433 * Ubuntu{nbsp}14.04 _Trusty Tahr_
434 * Ubuntu{nbsp}16.04 _Xenial Xerus_
435 * Ubuntu{nbsp}18.04 _Bionic Beaver_
437 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
439 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
445 # apt-add-repository ppa:lttng/stable-2.11
450 . Install the main LTTng{nbsp}{revision} packages:
455 # apt-get install lttng-tools
456 # apt-get install lttng-modules-dkms
457 # apt-get install liblttng-ust-dev
461 . **If you need to instrument and trace
462 <<java-application,Java applications>>**, install the LTTng-UST
468 # apt-get install liblttng-ust-agent-java
472 . **If you need to instrument and trace
473 <<python-application,Python{nbsp}3 applications>>**, install the
474 LTTng-UST Python agent:
479 # apt-get install python3-lttngust
484 [[enterprise-distributions]]
485 === RHEL, SUSE, and other enterprise distributions
487 To install LTTng on enterprise Linux distributions, such as Red Hat
488 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
489 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
492 [[building-from-source]]
493 === Build from source
495 To build and install LTTng{nbsp}{revision} from source:
497 . Using your distribution's package manager, or from source, install
498 the following dependencies of LTTng-tools and LTTng-UST:
501 * https://sourceforge.net/projects/libuuid/[libuuid]
502 * http://directory.fsf.org/wiki/Popt[popt]
503 * http://liburcu.org/[Userspace RCU]
504 * http://www.xmlsoft.org/[libxml2]
505 * **Optional**: https://github.com/numactl/numactl[numactl]
508 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
514 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
515 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
516 cd lttng-modules-2.11.* &&
518 sudo make modules_install &&
523 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
529 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.11.tar.bz2 &&
530 tar -xf lttng-ust-latest-2.11.tar.bz2 &&
531 cd lttng-ust-2.11.* &&
539 Add `--disable-numa` to `./configure` if you don't have
540 https://github.com/numactl/numactl[numactl].
544 .Java and Python application tracing
546 If you need to instrument and trace <<java-application,Java
547 applications>>, pass the `--enable-java-agent-jul`,
548 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
549 `configure` script, depending on which Java logging framework you use.
551 If you need to instrument and trace <<python-application,Python
552 applications>>, pass the `--enable-python-agent` option to the
553 `configure` script. You can set the `PYTHON` environment variable to the
554 path to the Python interpreter for which to install the LTTng-UST Python
562 By default, LTTng-UST libraries are installed to
563 dir:{/usr/local/lib}, which is the de facto directory in which to
564 keep self-compiled and third-party libraries.
566 When <<building-tracepoint-providers-and-user-application,linking an
567 instrumented user application with `liblttng-ust`>>:
569 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
571 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
572 man:gcc(1), man:g++(1), or man:clang(1).
576 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
582 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
583 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
584 cd lttng-tools-2.11.* &&
592 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
593 previous steps automatically for a given version of LTTng and confine
594 the installed files in a specific directory. This can be useful to test
595 LTTng without installing it on your system.
601 This is a short guide to get started quickly with LTTng kernel and user
604 Before you follow this guide, make sure to <<installing-lttng,install>>
607 This tutorial walks you through the steps to:
609 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
610 . <<tracing-your-own-user-application,Trace a user application>> written
612 . <<viewing-and-analyzing-your-traces,View and analyze the
616 [[tracing-the-linux-kernel]]
617 === Trace the Linux kernel
619 The following command lines start with the `#` prompt because you need
620 root privileges to trace the Linux kernel. You can also trace the kernel
621 as a regular user if your Unix user is a member of the
622 <<tracing-group,tracing group>>.
624 . Create a <<tracing-session,tracing session>> which writes its traces
625 to dir:{/tmp/my-kernel-trace}:
630 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
634 . List the available kernel tracepoints and system calls:
639 # lttng list --kernel
640 # lttng list --kernel --syscall
644 . Create <<event,event rules>> which match the desired instrumentation
645 point names, for example the `sched_switch` and `sched_process_fork`
646 tracepoints, and the man:open(2) and man:close(2) system calls:
651 # lttng enable-event --kernel sched_switch,sched_process_fork
652 # lttng enable-event --kernel --syscall open,close
656 You can also create an event rule which matches _all_ the Linux kernel
657 tracepoints (this will generate a lot of data when tracing):
662 # lttng enable-event --kernel --all
666 . <<basic-tracing-session-control,Start tracing>>:
675 . Do some operation on your system for a few seconds. For example,
676 load a website, or list the files of a directory.
677 . <<basic-tracing-session-control,Stop tracing>> and destroy the
688 The man:lttng-destroy(1) command does not destroy the trace data; it
689 only destroys the state of the tracing session.
691 . For the sake of this example, make the recorded trace accessible to
697 # chown -R $(whoami) /tmp/my-kernel-trace
701 See <<viewing-and-analyzing-your-traces,View and analyze the
702 recorded events>> to view the recorded events.
705 [[tracing-your-own-user-application]]
706 === Trace a user application
708 This section steps you through a simple example to trace a
709 _Hello world_ program written in C.
711 To create the traceable user application:
713 . Create the tracepoint provider header file, which defines the
714 tracepoints and the events they can generate:
720 #undef TRACEPOINT_PROVIDER
721 #define TRACEPOINT_PROVIDER hello_world
723 #undef TRACEPOINT_INCLUDE
724 #define TRACEPOINT_INCLUDE "./hello-tp.h"
726 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
729 #include <lttng/tracepoint.h>
739 ctf_string(my_string_field, my_string_arg)
740 ctf_integer(int, my_integer_field, my_integer_arg)
744 #endif /* _HELLO_TP_H */
746 #include <lttng/tracepoint-event.h>
750 . Create the tracepoint provider package source file:
756 #define TRACEPOINT_CREATE_PROBES
757 #define TRACEPOINT_DEFINE
759 #include "hello-tp.h"
763 . Build the tracepoint provider package:
768 $ gcc -c -I. hello-tp.c
772 . Create the _Hello World_ application source file:
779 #include "hello-tp.h"
781 int main(int argc, char *argv[])
785 puts("Hello, World!\nPress Enter to continue...");
788 * The following getchar() call is only placed here for the purpose
789 * of this demonstration, to pause the application in order for
790 * you to have time to list its tracepoints. It is not
796 * A tracepoint() call.
798 * Arguments, as defined in hello-tp.h:
800 * 1. Tracepoint provider name (required)
801 * 2. Tracepoint name (required)
802 * 3. my_integer_arg (first user-defined argument)
803 * 4. my_string_arg (second user-defined argument)
805 * Notice the tracepoint provider and tracepoint names are
806 * NOT strings: they are in fact parts of variables that the
807 * macros in hello-tp.h create.
809 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
811 for (x = 0; x < argc; ++x) {
812 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
815 puts("Quitting now!");
816 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
823 . Build the application:
832 . Link the application with the tracepoint provider package,
833 `liblttng-ust`, and `libdl`:
838 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
842 Here's the whole build process:
845 .User space tracing tutorial's build steps.
846 image::ust-flow.png[]
848 To trace the user application:
850 . Run the application with a few arguments:
855 $ ./hello world and beyond
864 Press Enter to continue...
868 . Start an LTTng <<lttng-sessiond,session daemon>>:
873 $ lttng-sessiond --daemonize
877 Note that a session daemon might already be running, for example as
878 a service that the distribution's service manager started.
880 . List the available user space tracepoints:
885 $ lttng list --userspace
889 You see the `hello_world:my_first_tracepoint` tracepoint listed
890 under the `./hello` process.
892 . Create a <<tracing-session,tracing session>>:
897 $ lttng create my-user-space-session
901 . Create an <<event,event rule>> which matches the
902 `hello_world:my_first_tracepoint` event name:
907 $ lttng enable-event --userspace hello_world:my_first_tracepoint
911 . <<basic-tracing-session-control,Start tracing>>:
920 . Go back to the running `hello` application and press Enter. The
921 program executes all `tracepoint()` instrumentation points and exits.
922 . <<basic-tracing-session-control,Stop tracing>> and destroy the
933 The man:lttng-destroy(1) command does not destroy the trace data; it
934 only destroys the state of the tracing session.
936 By default, LTTng saves the traces in
937 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
938 where +__name__+ is the tracing session name. The
939 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
941 See <<viewing-and-analyzing-your-traces,View and analyze the
942 recorded events>> to view the recorded events.
945 [[viewing-and-analyzing-your-traces]]
946 === View and analyze the recorded events
948 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
949 kernel>> and <<tracing-your-own-user-application,Trace a user
950 application>> tutorials, you can inspect the recorded events.
952 Many tools are available to read LTTng traces:
954 * **cmd:babeltrace** is a command-line utility which converts trace
955 formats; it supports the format that LTTng produces, CTF, as well as a
956 basic text output which can be ++grep++ed. The cmd:babeltrace command
957 is part of the http://diamon.org/babeltrace[Babeltrace] project.
958 * Babeltrace also includes
959 **https://www.python.org/[Python] bindings** so
960 that you can easily open and read an LTTng trace with your own script,
961 benefiting from the power of Python.
962 * http://tracecompass.org/[**Trace Compass**]
963 is a graphical user interface for viewing and analyzing any type of
964 logs or traces, including LTTng's.
965 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
966 project which includes many high-level analyses of LTTng kernel
967 traces, like scheduling statistics, interrupt frequency distribution,
968 top CPU usage, and more.
970 NOTE: This section assumes that LTTng saved the traces it recorded
971 during the previous tutorials to their default location, in the
972 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
973 environment variable defaults to `$HOME` if not set.
976 [[viewing-and-analyzing-your-traces-bt]]
977 ==== Use the cmd:babeltrace command-line tool
979 The simplest way to list all the recorded events of a trace is to pass
980 its path to cmd:babeltrace with no options:
984 $ babeltrace ~/lttng-traces/my-user-space-session*
987 cmd:babeltrace finds all traces recursively within the given path and
988 prints all their events, merging them in chronological order.
990 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
995 $ babeltrace /tmp/my-kernel-trace | grep _switch
998 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
999 count the recorded events:
1003 $ babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
1007 [[viewing-and-analyzing-your-traces-bt-python]]
1008 ==== Use the Babeltrace Python bindings
1010 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1011 is useful to isolate events by simple matching using man:grep(1) and
1012 similar utilities. However, more elaborate filters, such as keeping only
1013 event records with a field value falling within a specific range, are
1014 not trivial to write using a shell. Moreover, reductions and even the
1015 most basic computations involving multiple event records are virtually
1016 impossible to implement.
1018 Fortunately, Babeltrace ships with Python{nbsp}3 bindings which makes it
1019 easy to read the event records of an LTTng trace sequentially and
1020 compute the desired information.
1022 The following script accepts an LTTng Linux kernel trace path as its
1023 first argument and prints the short names of the top five running
1024 processes on CPU{nbsp}0 during the whole trace:
1029 from collections import Counter
1035 if len(sys.argv) != 2:
1036 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1037 print(msg, file=sys.stderr)
1040 # A trace collection contains one or more traces
1041 col = babeltrace.TraceCollection()
1043 # Add the trace provided by the user (LTTng traces always have
1045 if col.add_trace(sys.argv[1], 'ctf') is None:
1046 raise RuntimeError('Cannot add trace')
1048 # This counter dict contains execution times:
1050 # task command name -> total execution time (ns)
1051 exec_times = Counter()
1053 # This contains the last `sched_switch` timestamp
1057 for event in col.events:
1058 # Keep only `sched_switch` events
1059 if event.name != 'sched_switch':
1062 # Keep only events which happened on CPU 0
1063 if event['cpu_id'] != 0:
1067 cur_ts = event.timestamp
1073 # Previous task command (short) name
1074 prev_comm = event['prev_comm']
1076 # Initialize entry in our dict if not yet done
1077 if prev_comm not in exec_times:
1078 exec_times[prev_comm] = 0
1080 # Compute previous command execution time
1081 diff = cur_ts - last_ts
1083 # Update execution time of this command
1084 exec_times[prev_comm] += diff
1086 # Update last timestamp
1090 for name, ns in exec_times.most_common(5):
1092 print('{:20}{} s'.format(name, s))
1097 if __name__ == '__main__':
1098 sys.exit(0 if top5proc() else 1)
1105 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1111 swapper/0 48.607245889 s
1112 chromium 7.192738188 s
1113 pavucontrol 0.709894415 s
1114 Compositor 0.660867933 s
1115 Xorg.bin 0.616753786 s
1118 Note that `swapper/0` is the "idle" process of CPU{nbsp}0 on Linux;
1119 since we weren't using the CPU that much when tracing, its first
1120 position in the list makes sense.
1124 == [[understanding-lttng]]Core concepts
1126 From a user's perspective, the LTTng system is built on a few concepts,
1127 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1128 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1129 Understanding how those objects relate to eachother is key in mastering
1132 The core concepts are:
1134 * <<tracing-session,Tracing session>>
1135 * <<domain,Tracing domain>>
1136 * <<channel,Channel and ring buffer>>
1137 * <<"event","Instrumentation point, event rule, event, and event record">>
1143 A _tracing session_ is a stateful dialogue between you and
1144 a <<lttng-sessiond,session daemon>>. You can
1145 <<creating-destroying-tracing-sessions,create a new tracing
1146 session>> with the `lttng create` command.
1148 Anything that you do when you control LTTng tracers happens within a
1149 tracing session. In particular, a tracing session:
1152 * Has its own set of trace files.
1153 * Has its own state of activity (started or stopped).
1154 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1156 * Has its own <<channel,channels>> to which are associated their own
1157 <<event,event rules>>.
1160 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1161 image::concepts.png[]
1163 Those attributes and objects are completely isolated between different
1166 A tracing session is analogous to a cash machine session:
1167 the operations you do on the banking system through the cash machine do
1168 not alter the data of other users of the same system. In the case of
1169 the cash machine, a session lasts as long as your bank card is inside.
1170 In the case of LTTng, a tracing session lasts from the `lttng create`
1171 command to the `lttng destroy` command.
1174 .Each Unix user has its own set of tracing sessions.
1175 image::many-sessions.png[]
1178 [[tracing-session-mode]]
1179 ==== Tracing session mode
1181 LTTng can send the generated trace data to different locations. The
1182 _tracing session mode_ dictates where to send it. The following modes
1183 are available in LTTng{nbsp}{revision}:
1186 LTTng writes the traces to the file system of the machine it traces
1189 Network streaming mode::
1190 LTTng sends the traces over the network to a
1191 <<lttng-relayd,relay daemon>> running on a remote system.
1194 LTTng does not write the traces by default. Instead, you can request
1195 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1196 tracing session's current sub-buffers, and to write it to the
1197 target's file system or to send it over the network to a
1198 <<lttng-relayd,relay daemon>> running on a remote system.
1201 This mode is similar to the network streaming mode, but a live
1202 trace viewer can connect to the distant relay daemon to
1203 <<lttng-live,view event records as LTTng generates them>>.
1209 A _tracing domain_ is a namespace for event sources. A tracing domain
1210 has its own properties and features.
1212 There are currently five available tracing domains:
1216 * `java.util.logging` (JUL)
1220 You must specify a tracing domain when using some commands to avoid
1221 ambiguity. For example, since all the domains support named tracepoints
1222 as event sources (instrumentation points that you manually insert in the
1223 source code), you need to specify a tracing domain when
1224 <<enabling-disabling-events,creating an event rule>> because all the
1225 tracing domains could have tracepoints with the same names.
1227 Some features are reserved to specific tracing domains. Dynamic function
1228 entry and return instrumentation points, for example, are currently only
1229 supported in the Linux kernel tracing domain, but support for other
1230 tracing domains could be added in the future.
1232 You can create <<channel,channels>> in the Linux kernel and user space
1233 tracing domains. The other tracing domains have a single default
1238 === Channel and ring buffer
1240 A _channel_ is an object which is responsible for a set of ring buffers.
1241 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1242 tracer emits an event, it can record it to one or more
1243 sub-buffers. The attributes of a channel determine what to do when
1244 there's no space left for a new event record because all sub-buffers
1245 are full, where to send a full sub-buffer, and other behaviours.
1247 A channel is always associated to a <<domain,tracing domain>>. The
1248 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1249 a default channel which you cannot configure.
1251 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1252 an event, it records it to the sub-buffers of all
1253 the enabled channels with a satisfied event rule, as long as those
1254 channels are part of active <<tracing-session,tracing sessions>>.
1257 [[channel-buffering-schemes]]
1258 ==== Per-user vs. per-process buffering schemes
1260 A channel has at least one ring buffer _per CPU_. LTTng always
1261 records an event to the ring buffer associated to the CPU on which it
1264 Two _buffering schemes_ are available when you
1265 <<enabling-disabling-channels,create a channel>> in the
1266 user space <<domain,tracing domain>>:
1268 Per-user buffering::
1269 Allocate one set of ring buffers--one per CPU--shared by all the
1270 instrumented processes of each Unix user.
1274 .Per-user buffering scheme.
1275 image::per-user-buffering.png[]
1278 Per-process buffering::
1279 Allocate one set of ring buffers--one per CPU--for each
1280 instrumented process.
1284 .Per-process buffering scheme.
1285 image::per-process-buffering.png[]
1288 The per-process buffering scheme tends to consume more memory than the
1289 per-user option because systems generally have more instrumented
1290 processes than Unix users running instrumented processes. However, the
1291 per-process buffering scheme ensures that one process having a high
1292 event throughput won't fill all the shared sub-buffers of the same
1295 The Linux kernel tracing domain has only one available buffering scheme
1296 which is to allocate a single set of ring buffers for the whole system.
1297 This scheme is similar to the per-user option, but with a single, global
1298 user "running" the kernel.
1301 [[channel-overwrite-mode-vs-discard-mode]]
1302 ==== Overwrite vs. discard event record loss modes
1304 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1305 arc in the following animations) of a specific channel's ring buffer.
1306 When there's no space left in a sub-buffer, the tracer marks it as
1307 consumable (red) and another, empty sub-buffer starts receiving the
1308 following event records. A <<lttng-consumerd,consumer daemon>>
1309 eventually consumes the marked sub-buffer (returns to white).
1312 [role="docsvg-channel-subbuf-anim"]
1317 In an ideal world, sub-buffers are consumed faster than they are filled,
1318 as it is the case in the previous animation. In the real world,
1319 however, all sub-buffers can be full at some point, leaving no space to
1320 record the following events.
1322 By default, LTTng-modules and LTTng-UST are _non-blocking_ tracers: when
1323 no empty sub-buffer is available, it is acceptable to lose event records
1324 when the alternative would be to cause substantial delays in the
1325 instrumented application's execution. LTTng privileges performance over
1326 integrity; it aims at perturbing the target system as little as possible
1327 in order to make tracing of subtle race conditions and rare interrupt
1330 Since LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST, supports
1331 a _blocking mode_. See the <<blocking-timeout-example,blocking timeout
1332 example>> to learn how to use the blocking mode.
1334 When it comes to losing event records because no empty sub-buffer is
1335 available, or because the <<opt-blocking-timeout,blocking timeout>> is
1336 reached, the channel's _event record loss mode_ determines what to do.
1337 The available event record loss modes are:
1340 Drop the newest event records until a the tracer releases a
1343 This is the only available mode when you specify a
1344 <<opt-blocking-timeout,blocking timeout>>.
1347 Clear the sub-buffer containing the oldest event records and start
1348 writing the newest event records there.
1350 This mode is sometimes called _flight recorder mode_ because it's
1352 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1353 always keep a fixed amount of the latest data.
1355 Which mechanism you should choose depends on your context: prioritize
1356 the newest or the oldest event records in the ring buffer?
1358 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1359 as soon as a there's no space left for a new event record, whereas in
1360 discard mode, the tracer only discards the event record that doesn't
1363 In discard mode, LTTng increments a count of lost event records when an
1364 event record is lost and saves this count to the trace. In overwrite
1365 mode, since LTTng{nbsp}2.8, LTTng increments a count of lost sub-buffers
1366 when a sub-buffer is lost and saves this count to the trace. In this
1367 mode, LTTng does not write to the trace the exact number of lost event
1368 records in those lost sub-buffers. Trace analyses can use the trace's
1369 saved discarded event record and sub-buffer counts to decide whether or
1370 not to perform the analyses even if trace data is known to be missing.
1372 There are a few ways to decrease your probability of losing event
1374 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1375 how you can fine-tune the sub-buffer count and size of a channel to
1376 virtually stop losing event records, though at the cost of greater
1380 [[channel-subbuf-size-vs-subbuf-count]]
1381 ==== Sub-buffer count and size
1383 When you <<enabling-disabling-channels,create a channel>>, you can
1384 set its number of sub-buffers and their size.
1386 Note that there is noticeable CPU overhead introduced when
1387 switching sub-buffers (marking a full one as consumable and switching
1388 to an empty one for the following events to be recorded). Knowing this,
1389 the following list presents a few practical situations along with how
1390 to configure the sub-buffer count and size for them:
1392 * **High event throughput**: In general, prefer bigger sub-buffers to
1393 lower the risk of losing event records.
1395 Having bigger sub-buffers also ensures a lower
1396 <<channel-switch-timer,sub-buffer switching frequency>>.
1398 The number of sub-buffers is only meaningful if you create the channel
1399 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1400 other sub-buffers are left unaltered.
1402 * **Low event throughput**: In general, prefer smaller sub-buffers
1403 since the risk of losing event records is low.
1405 Because events occur less frequently, the sub-buffer switching frequency
1406 should remain low and thus the tracer's overhead should not be a
1409 * **Low memory system**: If your target system has a low memory
1410 limit, prefer fewer first, then smaller sub-buffers.
1412 Even if the system is limited in memory, you want to keep the
1413 sub-buffers as big as possible to avoid a high sub-buffer switching
1416 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1417 which means event data is very compact. For example, the average
1418 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1419 sub-buffer size of 1{nbsp}MiB is considered big.
1421 The previous situations highlight the major trade-off between a few big
1422 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1423 frequency vs. how much data is lost in overwrite mode. Assuming a
1424 constant event throughput and using the overwrite mode, the two
1425 following configurations have the same ring buffer total size:
1428 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1433 * **Two sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1434 switching frequency, but if a sub-buffer overwrite happens, half of
1435 the event records so far (4{nbsp}MiB) are definitely lost.
1436 * **Eight sub-buffers of 1{nbsp}MiB each**: Expect four times the tracer's
1437 overhead as the previous configuration, but if a sub-buffer
1438 overwrite happens, only the eighth of event records so far are
1441 In discard mode, the sub-buffers count parameter is pointless: use two
1442 sub-buffers and set their size according to the requirements of your
1446 [[channel-switch-timer]]
1447 ==== Switch timer period
1449 The _switch timer period_ is an important configurable attribute of
1450 a channel to ensure periodic sub-buffer flushing.
1452 When the _switch timer_ expires, a sub-buffer switch happens. You can
1453 set the switch timer period attribute when you
1454 <<enabling-disabling-channels,create a channel>> to ensure that LTTng
1455 consumes and commits trace data to trace files or to a distant relay
1456 daemon periodically in case of a low event throughput.
1459 [role="docsvg-channel-switch-timer"]
1464 This attribute is also convenient when you use big sub-buffers to cope
1465 with a sporadic high event throughput, even if the throughput is
1469 [[channel-read-timer]]
1470 ==== Read timer period
1472 By default, the LTTng tracers use a notification mechanism to signal a
1473 full sub-buffer so that a consumer daemon can consume it. When such
1474 notifications must be avoided, for example in real-time applications,
1475 you can use the channel's _read timer_ instead. When the read timer
1476 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1477 consumable sub-buffers.
1480 [[tracefile-rotation]]
1481 ==== Trace file count and size
1483 By default, trace files can grow as large as needed. You can set the
1484 maximum size of each trace file that a channel writes when you
1485 <<enabling-disabling-channels,create a channel>>. When the size of
1486 a trace file reaches the channel's fixed maximum size, LTTng creates
1487 another file to contain the next event records. LTTng appends a file
1488 count to each trace file name in this case.
1490 If you set the trace file size attribute when you create a channel, the
1491 maximum number of trace files that LTTng creates is _unlimited_ by
1492 default. To limit them, you can also set a maximum number of trace
1493 files. When the number of trace files reaches the channel's fixed
1494 maximum count, the oldest trace file is overwritten. This mechanism is
1495 called _trace file rotation_.
1497 Even if you don't limit the trace file count, you cannot assume that
1498 LTTng doesn't manage any trace file. In other words, there is no safe
1499 way to know if LTTng still holds a given trace file open with the trace
1500 file rotation feature. The only way to obtain an unmanaged,
1501 self-contained LTTng trace before you
1502 <<creating-destroying-tracing-sessions,destroy>> the tracing session is
1503 with the <<session-rotation,tracing session rotation>> feature
1504 (available since LTTng{nbsp}2.11).
1508 === Instrumentation point, event rule, event, and event record
1510 An _event rule_ is a set of conditions which must be **all** satisfied
1511 for LTTng to record an occuring event.
1513 You set the conditions when you <<enabling-disabling-events,create
1516 You always attach an event rule to <<channel,channel>> when you create
1519 When an event passes the conditions of an event rule, LTTng records it
1520 in one of the attached channel's sub-buffers.
1522 The available conditions, as of LTTng{nbsp}{revision}, are:
1524 * The event rule _is enabled_.
1525 * The instrumentation point's type _is{nbsp}T_.
1526 * The instrumentation point's name (sometimes called _event name_)
1527 _matches{nbsp}N_, but _is not{nbsp}E_.
1528 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1529 _is exactly{nbsp}L_.
1530 * The fields of the event's payload _satisfy_ a filter
1531 expression{nbsp}__F__.
1533 As you can see, all the conditions but the dynamic filter are related to
1534 the event rule's status or to the instrumentation point, not to the
1535 occurring events. This is why, without a filter, checking if an event
1536 passes an event rule is not a dynamic task: when you create or modify an
1537 event rule, all the tracers of its tracing domain enable or disable the
1538 instrumentation points themselves once. This is possible because the
1539 attributes of an instrumentation point (type, name, and log level) are
1540 defined statically. In other words, without a dynamic filter, the tracer
1541 _does not evaluate_ the arguments of an instrumentation point unless it
1542 matches an enabled event rule.
1544 Note that, for LTTng to record an event, the <<channel,channel>> to
1545 which a matching event rule is attached must also be enabled, and the
1546 <<tracing-session,tracing session>> owning this channel must be active
1550 .Logical path from an instrumentation point to an event record.
1551 image::event-rule.png[]
1553 .Event, event record, or event rule?
1555 With so many similar terms, it's easy to get confused.
1557 An **event** is the consequence of the execution of an _instrumentation
1558 point_, like a tracepoint that you manually place in some source code,
1559 or a Linux kernel kprobe. An event is said to _occur_ at a specific
1560 time. Different actions can be taken upon the occurrence of an event,
1561 like record the event's payload to a buffer.
1563 An **event record** is the representation of an event in a sub-buffer. A
1564 tracer is responsible for capturing the payload of an event, current
1565 context variables, the event's ID, and the event's timestamp. LTTng
1566 can append this sub-buffer to a trace file.
1568 An **event rule** is a set of conditions which must _all_ be satisfied
1569 for LTTng to record an occuring event. Events still occur without
1570 satisfying event rules, but LTTng does not record them.
1575 == Components of noch:{LTTng}
1577 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1578 to call LTTng a simple _tool_ since it is composed of multiple
1579 interacting components. This section describes those components,
1580 explains their respective roles, and shows how they connect together to
1581 form the LTTng ecosystem.
1583 The following diagram shows how the most important components of LTTng
1584 interact with user applications, the Linux kernel, and you:
1587 .Control and trace data paths between LTTng components.
1588 image::plumbing.png[]
1590 The LTTng project incorporates:
1592 * **LTTng-tools**: Libraries and command-line interface to
1593 control tracing sessions.
1594 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1595 ** <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
1596 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1597 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1598 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1599 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1601 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1602 headers to instrument and trace any native user application.
1603 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1604 *** `liblttng-ust-libc-wrapper`
1605 *** `liblttng-ust-pthread-wrapper`
1606 *** `liblttng-ust-cyg-profile`
1607 *** `liblttng-ust-cyg-profile-fast`
1608 *** `liblttng-ust-dl`
1609 ** User space tracepoint provider source files generator command-line
1610 tool (man:lttng-gen-tp(1)).
1611 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1612 Java applications using `java.util.logging` or
1613 Apache log4j{nbsp}1.2 logging.
1614 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1615 Python applications using the standard `logging` package.
1616 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1618 ** LTTng kernel tracer module.
1619 ** Tracing ring buffer kernel modules.
1620 ** Probe kernel modules.
1621 ** LTTng logger kernel module.
1625 === Tracing control command-line interface
1628 .The tracing control command-line interface.
1629 image::plumbing-lttng-cli.png[]
1631 The _man:lttng(1) command-line tool_ is the standard user interface to
1632 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1633 is part of LTTng-tools.
1635 The cmd:lttng tool is linked with
1636 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1637 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1639 The cmd:lttng tool has a Git-like interface:
1643 $ lttng <GENERAL OPTIONS> <COMMAND> <COMMAND OPTIONS>
1646 The <<controlling-tracing,Tracing control>> section explores the
1647 available features of LTTng using the cmd:lttng tool.
1650 [[liblttng-ctl-lttng]]
1651 === Tracing control library
1654 .The tracing control library.
1655 image::plumbing-liblttng-ctl.png[]
1657 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1658 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1659 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1661 The <<lttng-cli,cmd:lttng command-line tool>>
1662 is linked with `liblttng-ctl`.
1664 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1669 #include <lttng/lttng.h>
1672 Some objects are referenced by name (C string), such as tracing
1673 sessions, but most of them require to create a handle first using
1674 `lttng_create_handle()`.
1676 The best available developer documentation for `liblttng-ctl` is, as of
1677 LTTng{nbsp}{revision}, its installed header files. Every function and
1678 structure is thoroughly documented.
1682 === User space tracing library
1685 .The user space tracing library.
1686 image::plumbing-liblttng-ust.png[]
1688 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1689 is the LTTng user space tracer. It receives commands from a
1690 <<lttng-sessiond,session daemon>>, for example to
1691 enable and disable specific instrumentation points, and writes event
1692 records to ring buffers shared with a
1693 <<lttng-consumerd,consumer daemon>>.
1694 `liblttng-ust` is part of LTTng-UST.
1696 Public C header files are installed beside `liblttng-ust` to
1697 instrument any <<c-application,C or $$C++$$ application>>.
1699 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1700 packages, use their own library providing tracepoints which is
1701 linked with `liblttng-ust`.
1703 An application or library does not have to initialize `liblttng-ust`
1704 manually: its constructor does the necessary tasks to properly register
1705 to a session daemon. The initialization phase also enables the
1706 instrumentation points matching the <<event,event rules>> that you
1710 [[lttng-ust-agents]]
1711 === User space tracing agents
1714 .The user space tracing agents.
1715 image::plumbing-lttng-ust-agents.png[]
1717 The _LTTng-UST Java and Python agents_ are regular Java and Python
1718 packages which add LTTng tracing capabilities to the
1719 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1721 In the case of Java, the
1722 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1723 core logging facilities] and
1724 https://logging.apache.org/log4j/1.2/[Apache log4j{nbsp}1.2] are supported.
1725 Note that Apache Log4{nbsp}2 is not supported.
1727 In the case of Python, the standard
1728 https://docs.python.org/3/library/logging.html[`logging`] package
1729 is supported. Both Python{nbsp}2 and Python{nbsp}3 modules can import the
1730 LTTng-UST Python agent package.
1732 The applications using the LTTng-UST agents are in the
1733 `java.util.logging` (JUL),
1734 log4j, and Python <<domain,tracing domains>>.
1736 Both agents use the same mechanism to trace the log statements. When an
1737 agent initializes, it creates a log handler that attaches to the root
1738 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1739 When the application executes a log statement, the root logger passes it
1740 to the agent's log handler. The agent's log handler calls a native
1741 function in a tracepoint provider package shared library linked with
1742 <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1743 other fields, like its logger name and its log level. This native
1744 function contains a user space instrumentation point, hence tracing the
1747 The log level condition of an
1748 <<event,event rule>> is considered when tracing
1749 a Java or a Python application, and it's compatible with the standard
1750 JUL, log4j, and Python log levels.
1754 === LTTng kernel modules
1757 .The LTTng kernel modules.
1758 image::plumbing-lttng-modules.png[]
1760 The _LTTng kernel modules_ are a set of Linux kernel modules
1761 which implement the kernel tracer of the LTTng project. The LTTng
1762 kernel modules are part of LTTng-modules.
1764 The LTTng kernel modules include:
1766 * A set of _probe_ modules.
1768 Each module attaches to a specific subsystem
1769 of the Linux kernel using its tracepoint instrument points. There are
1770 also modules to attach to the entry and return points of the Linux
1771 system call functions.
1773 * _Ring buffer_ modules.
1775 A ring buffer implementation is provided as kernel modules. The LTTng
1776 kernel tracer writes to the ring buffer; a
1777 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1779 * The _LTTng kernel tracer_ module.
1780 * The _LTTng logger_ module.
1782 The LTTng logger module implements the special path:{/proc/lttng-logger}
1783 file so that any executable can generate LTTng events by opening and
1784 writing to this file.
1786 See <<proc-lttng-logger-abi,LTTng logger>>.
1788 Generally, you do not have to load the LTTng kernel modules manually
1789 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1790 daemon>> loads the necessary modules when starting. If you have extra
1791 probe modules, you can specify to load them to the session daemon on
1794 The LTTng kernel modules are installed in
1795 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1796 the kernel release (see `uname --kernel-release`).
1803 .The session daemon.
1804 image::plumbing-sessiond.png[]
1806 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1807 managing tracing sessions and for controlling the various components of
1808 LTTng. The session daemon is part of LTTng-tools.
1810 The session daemon sends control requests to and receives control
1813 * The <<lttng-ust,user space tracing library>>.
1815 Any instance of the user space tracing library first registers to
1816 a session daemon. Then, the session daemon can send requests to
1817 this instance, such as:
1820 ** Get the list of tracepoints.
1821 ** Share an <<event,event rule>> so that the user space tracing library
1822 can enable or disable tracepoints. Amongst the possible conditions
1823 of an event rule is a filter expression which `liblttng-ust` evalutes
1824 when an event occurs.
1825 ** Share <<channel,channel>> attributes and ring buffer locations.
1828 The session daemon and the user space tracing library use a Unix
1829 domain socket for their communication.
1831 * The <<lttng-ust-agents,user space tracing agents>>.
1833 Any instance of a user space tracing agent first registers to
1834 a session daemon. Then, the session daemon can send requests to
1835 this instance, such as:
1838 ** Get the list of loggers.
1839 ** Enable or disable a specific logger.
1842 The session daemon and the user space tracing agent use a TCP connection
1843 for their communication.
1845 * The <<lttng-modules,LTTng kernel tracer>>.
1846 * The <<lttng-consumerd,consumer daemon>>.
1848 The session daemon sends requests to the consumer daemon to instruct
1849 it where to send the trace data streams, amongst other information.
1851 * The <<lttng-relayd,relay daemon>>.
1853 The session daemon receives commands from the
1854 <<liblttng-ctl-lttng,tracing control library>>.
1856 The root session daemon loads the appropriate
1857 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1858 a <<lttng-consumerd,consumer daemon>> as soon as you create
1859 an <<event,event rule>>.
1861 The session daemon does not send and receive trace data: this is the
1862 role of the <<lttng-consumerd,consumer daemon>> and
1863 <<lttng-relayd,relay daemon>>. It does, however, generate the
1864 http://diamon.org/ctf/[CTF] metadata stream.
1866 Each Unix user can have its own session daemon instance. The
1867 tracing sessions which different session daemons manage are completely
1870 The root user's session daemon is the only one which is
1871 allowed to control the LTTng kernel tracer, and its spawned consumer
1872 daemon is the only one which is allowed to consume trace data from the
1873 LTTng kernel tracer. Note, however, that any Unix user which is a member
1874 of the <<tracing-group,tracing group>> is allowed
1875 to create <<channel,channels>> in the
1876 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1879 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1880 session daemon when using its `create` command if none is currently
1881 running. You can also start the session daemon manually.
1888 .The consumer daemon.
1889 image::plumbing-consumerd.png[]
1891 The _consumer daemon_, cmd:lttng-consumerd, is a daemon which shares
1892 ring buffers with user applications or with the LTTng kernel modules to
1893 collect trace data and send it to some location (on disk or to a
1894 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
1895 is part of LTTng-tools.
1897 You do not start a consumer daemon manually: a consumer daemon is always
1898 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
1899 <<event,event rule>>, that is, before you start tracing. When you kill
1900 its owner session daemon, the consumer daemon also exits because it is
1901 the session daemon's child process. Command-line options of
1902 man:lttng-sessiond(8) target the consumer daemon process.
1904 There are up to two running consumer daemons per Unix user, whereas only
1905 one session daemon can run per user. This is because each process can be
1906 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
1907 and 64-bit processes, it is more efficient to have separate
1908 corresponding 32-bit and 64-bit consumer daemons. The root user is an
1909 exception: it can have up to _three_ running consumer daemons: 32-bit
1910 and 64-bit instances for its user applications, and one more
1911 reserved for collecting kernel trace data.
1919 image::plumbing-relayd.png[]
1921 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
1922 between remote session and consumer daemons, local trace files, and a
1923 remote live trace viewer. The relay daemon is part of LTTng-tools.
1925 The main purpose of the relay daemon is to implement a receiver of
1926 <<sending-trace-data-over-the-network,trace data over the network>>.
1927 This is useful when the target system does not have much file system
1928 space to record trace files locally.
1930 The relay daemon is also a server to which a
1931 <<lttng-live,live trace viewer>> can
1932 connect. The live trace viewer sends requests to the relay daemon to
1933 receive trace data as the target system emits events. The
1934 communication protocol is named _LTTng live_; it is used over TCP
1937 Note that you can start the relay daemon on the target system directly.
1938 This is the setup of choice when the use case is to view events as
1939 the target system emits them without the need of a remote system.
1943 == [[using-lttng]]Instrumentation
1945 There are many examples of tracing and monitoring in our everyday life:
1947 * You have access to real-time and historical weather reports and
1948 forecasts thanks to weather stations installed around the country.
1949 * You know your heart is safe thanks to an electrocardiogram.
1950 * You make sure not to drive your car too fast and to have enough fuel
1951 to reach your destination thanks to gauges visible on your dashboard.
1953 All the previous examples have something in common: they rely on
1954 **instruments**. Without the electrodes attached to the surface of your
1955 body's skin, cardiac monitoring is futile.
1957 LTTng, as a tracer, is no different from those real life examples. If
1958 you're about to trace a software system or, in other words, record its
1959 history of execution, you better have **instrumentation points** in the
1960 subject you're tracing, that is, the actual software.
1962 Various ways were developed to instrument a piece of software for LTTng
1963 tracing. The most straightforward one is to manually place
1964 instrumentation points, called _tracepoints_, in the software's source
1965 code. It is also possible to add instrumentation points dynamically in
1966 the Linux kernel <<domain,tracing domain>>.
1968 If you're only interested in tracing the Linux kernel, your
1969 instrumentation needs are probably already covered by LTTng's built-in
1970 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
1971 user application which is already instrumented for LTTng tracing.
1972 In such cases, you can skip this whole section and read the topics of
1973 the <<controlling-tracing,Tracing control>> section.
1975 Many methods are available to instrument a piece of software for LTTng
1978 * <<c-application,User space instrumentation for C and $$C++$$
1980 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
1981 * <<java-application,User space Java agent>>.
1982 * <<python-application,User space Python agent>>.
1983 * <<proc-lttng-logger-abi,LTTng logger>>.
1984 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
1988 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
1990 The procedure to instrument a C or $$C++$$ user application with
1991 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
1993 . <<tracepoint-provider,Create the source files of a tracepoint provider
1995 . <<probing-the-application-source-code,Add tracepoints to
1996 the application's source code>>.
1997 . <<building-tracepoint-providers-and-user-application,Build and link
1998 a tracepoint provider package and the user application>>.
2000 If you need quick, man:printf(3)-like instrumentation, you can skip
2001 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2004 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2005 instrument a user application with `liblttng-ust`.
2008 [[tracepoint-provider]]
2009 ==== Create the source files of a tracepoint provider package
2011 A _tracepoint provider_ is a set of compiled functions which provide
2012 **tracepoints** to an application, the type of instrumentation point
2013 supported by LTTng-UST. Those functions can emit events with
2014 user-defined fields and serialize those events as event records to one
2015 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2016 macro, which you <<probing-the-application-source-code,insert in a user
2017 application's source code>>, calls those functions.
2019 A _tracepoint provider package_ is an object file (`.o`) or a shared
2020 library (`.so`) which contains one or more tracepoint providers.
2021 Its source files are:
2023 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2024 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2026 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2027 the LTTng user space tracer, at run time.
2030 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2031 image::ust-app.png[]
2033 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2034 skip creating and using a tracepoint provider and use
2035 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2039 ===== Create a tracepoint provider header file template
2041 A _tracepoint provider header file_ contains the tracepoint
2042 definitions of a tracepoint provider.
2044 To create a tracepoint provider header file:
2046 . Start from this template:
2050 .Tracepoint provider header file template (`.h` file extension).
2052 #undef TRACEPOINT_PROVIDER
2053 #define TRACEPOINT_PROVIDER provider_name
2055 #undef TRACEPOINT_INCLUDE
2056 #define TRACEPOINT_INCLUDE "./tp.h"
2058 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2061 #include <lttng/tracepoint.h>
2064 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2065 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2070 #include <lttng/tracepoint-event.h>
2076 * `provider_name` with the name of your tracepoint provider.
2077 * `"tp.h"` with the name of your tracepoint provider header file.
2079 . Below the `#include <lttng/tracepoint.h>` line, put your
2080 <<defining-tracepoints,tracepoint definitions>>.
2082 Your tracepoint provider name must be unique amongst all the possible
2083 tracepoint provider names used on the same target system. We
2084 suggest to include the name of your project or company in the name,
2085 for example, `org_lttng_my_project_tpp`.
2087 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2088 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2089 write are the <<defining-tracepoints,tracepoint definitions>>.
2092 [[defining-tracepoints]]
2093 ===== Create a tracepoint definition
2095 A _tracepoint definition_ defines, for a given tracepoint:
2097 * Its **input arguments**. They are the macro parameters that the
2098 `tracepoint()` macro accepts for this particular tracepoint
2099 in the user application's source code.
2100 * Its **output event fields**. They are the sources of event fields
2101 that form the payload of any event that the execution of the
2102 `tracepoint()` macro emits for this particular tracepoint.
2104 You can create a tracepoint definition by using the
2105 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2107 <<tpp-header,tracepoint provider header file template>>.
2109 The syntax of the `TRACEPOINT_EVENT()` macro is:
2112 .`TRACEPOINT_EVENT()` macro syntax.
2115 /* Tracepoint provider name */
2118 /* Tracepoint name */
2121 /* Input arguments */
2126 /* Output event fields */
2135 * `provider_name` with your tracepoint provider name.
2136 * `tracepoint_name` with your tracepoint name.
2137 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2138 * `fields` with the <<tpp-def-output-fields,output event field>>
2141 This tracepoint emits events named `provider_name:tracepoint_name`.
2144 .Event name's length limitation
2146 The concatenation of the tracepoint provider name and the
2147 tracepoint name must not exceed **254{nbsp}characters**. If it does, the
2148 instrumented application compiles and runs, but LTTng throws multiple
2149 warnings and you could experience serious issues.
2152 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2155 .`TP_ARGS()` macro syntax.
2164 * `type` with the C type of the argument.
2165 * `arg_name` with the argument name.
2167 You can repeat `type` and `arg_name` up to 10{nbsp}times to have
2168 more than one argument.
2170 .`TP_ARGS()` usage with three arguments.
2182 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2183 tracepoint definition with no input arguments.
2185 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2186 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2187 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2188 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2191 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2192 C expression that the tracer evalutes at the `tracepoint()` macro site
2193 in the application's source code. This expression provides a field's
2194 source of data. The argument expression can include input argument names
2195 listed in the `TP_ARGS()` macro.
2197 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2198 must be unique within a given tracepoint definition.
2200 Here's a complete tracepoint definition example:
2202 .Tracepoint definition.
2204 The following tracepoint definition defines a tracepoint which takes
2205 three input arguments and has four output event fields.
2209 #include "my-custom-structure.h"
2215 const struct my_custom_structure*, my_custom_structure,
2220 ctf_string(query_field, query)
2221 ctf_float(double, ratio_field, ratio)
2222 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2223 ctf_integer(int, send_size, my_custom_structure->send_size)
2228 You can refer to this tracepoint definition with the `tracepoint()`
2229 macro in your application's source code like this:
2233 tracepoint(my_provider, my_tracepoint,
2234 my_structure, some_ratio, the_query);
2238 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2239 if they satisfy an enabled <<event,event rule>>.
2242 [[using-tracepoint-classes]]
2243 ===== Use a tracepoint class
2245 A _tracepoint class_ is a class of tracepoints which share the same
2246 output event field definitions. A _tracepoint instance_ is one
2247 instance of such a defined tracepoint class, with its own tracepoint
2250 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2251 shorthand which defines both a tracepoint class and a tracepoint
2252 instance at the same time.
2254 When you build a tracepoint provider package, the C or $$C++$$ compiler
2255 creates one serialization function for each **tracepoint class**. A
2256 serialization function is responsible for serializing the event fields
2257 of a tracepoint to a sub-buffer when tracing.
2259 For various performance reasons, when your situation requires multiple
2260 tracepoint definitions with different names, but with the same event
2261 fields, we recommend that you manually create a tracepoint class
2262 and instantiate as many tracepoint instances as needed. One positive
2263 effect of such a design, amongst other advantages, is that all
2264 tracepoint instances of the same tracepoint class reuse the same
2265 serialization function, thus reducing
2266 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2268 .Use a tracepoint class and tracepoint instances.
2270 Consider the following three tracepoint definitions:
2282 ctf_integer(int, userid, userid)
2283 ctf_integer(size_t, len, len)
2295 ctf_integer(int, userid, userid)
2296 ctf_integer(size_t, len, len)
2308 ctf_integer(int, userid, userid)
2309 ctf_integer(size_t, len, len)
2314 In this case, we create three tracepoint classes, with one implicit
2315 tracepoint instance for each of them: `get_account`, `get_settings`, and
2316 `get_transaction`. However, they all share the same event field names
2317 and types. Hence three identical, yet independent serialization
2318 functions are created when you build the tracepoint provider package.
2320 A better design choice is to define a single tracepoint class and three
2321 tracepoint instances:
2325 /* The tracepoint class */
2326 TRACEPOINT_EVENT_CLASS(
2327 /* Tracepoint provider name */
2330 /* Tracepoint class name */
2333 /* Input arguments */
2339 /* Output event fields */
2341 ctf_integer(int, userid, userid)
2342 ctf_integer(size_t, len, len)
2346 /* The tracepoint instances */
2347 TRACEPOINT_EVENT_INSTANCE(
2348 /* Tracepoint provider name */
2351 /* Tracepoint class name */
2354 /* Tracepoint name */
2357 /* Input arguments */
2363 TRACEPOINT_EVENT_INSTANCE(
2372 TRACEPOINT_EVENT_INSTANCE(
2385 [[assigning-log-levels]]
2386 ===== Assign a log level to a tracepoint definition
2388 You can assign an optional _log level_ to a
2389 <<defining-tracepoints,tracepoint definition>>.
2391 Assigning different levels of severity to tracepoint definitions can
2392 be useful: when you <<enabling-disabling-events,create an event rule>>,
2393 you can target tracepoints having a log level as severe as a specific
2396 The concept of LTTng-UST log levels is similar to the levels found
2397 in typical logging frameworks:
2399 * In a logging framework, the log level is given by the function
2400 or method name you use at the log statement site: `debug()`,
2401 `info()`, `warn()`, `error()`, and so on.
2402 * In LTTng-UST, you statically assign the log level to a tracepoint
2403 definition; any `tracepoint()` macro invocation which refers to
2404 this definition has this log level.
2406 You can assign a log level to a tracepoint definition with the
2407 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2408 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2409 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2412 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2415 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2417 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2422 * `provider_name` with the tracepoint provider name.
2423 * `tracepoint_name` with the tracepoint name.
2424 * `log_level` with the log level to assign to the tracepoint
2425 definition named `tracepoint_name` in the `provider_name`
2426 tracepoint provider.
2428 See man:lttng-ust(3) for a list of available log level names.
2430 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2434 /* Tracepoint definition */
2443 ctf_integer(int, userid, userid)
2444 ctf_integer(size_t, len, len)
2448 /* Log level assignment */
2449 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2455 ===== Create a tracepoint provider package source file
2457 A _tracepoint provider package source file_ is a C source file which
2458 includes a <<tpp-header,tracepoint provider header file>> to expand its
2459 macros into event serialization and other functions.
2461 You can always use the following tracepoint provider package source
2465 .Tracepoint provider package source file template.
2467 #define TRACEPOINT_CREATE_PROBES
2472 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2473 header file>> name. You may also include more than one tracepoint
2474 provider header file here to create a tracepoint provider package
2475 holding more than one tracepoint providers.
2478 [[probing-the-application-source-code]]
2479 ==== Add tracepoints to an application's source code
2481 Once you <<tpp-header,create a tracepoint provider header file>>, you
2482 can use the `tracepoint()` macro in your application's
2483 source code to insert the tracepoints that this header
2484 <<defining-tracepoints,defines>>.
2486 The `tracepoint()` macro takes at least two parameters: the tracepoint
2487 provider name and the tracepoint name. The corresponding tracepoint
2488 definition defines the other parameters.
2490 .`tracepoint()` usage.
2492 The following <<defining-tracepoints,tracepoint definition>> defines a
2493 tracepoint which takes two input arguments and has two output event
2497 .Tracepoint provider header file.
2499 #include "my-custom-structure.h"
2506 const char*, cmd_name
2509 ctf_string(cmd_name, cmd_name)
2510 ctf_integer(int, number_of_args, argc)
2515 You can refer to this tracepoint definition with the `tracepoint()`
2516 macro in your application's source code like this:
2519 .Application's source file.
2523 int main(int argc, char* argv[])
2525 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2531 Note how the application's source code includes
2532 the tracepoint provider header file containing the tracepoint
2533 definitions to use, path:{tp.h}.
2536 .`tracepoint()` usage with a complex tracepoint definition.
2538 Consider this complex tracepoint definition, where multiple event
2539 fields refer to the same input arguments in their argument expression
2543 .Tracepoint provider header file.
2545 /* For `struct stat` */
2546 #include <sys/types.h>
2547 #include <sys/stat.h>
2559 ctf_integer(int, my_constant_field, 23 + 17)
2560 ctf_integer(int, my_int_arg_field, my_int_arg)
2561 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2562 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2563 my_str_arg[2] + my_str_arg[3])
2564 ctf_string(my_str_arg_field, my_str_arg)
2565 ctf_integer_hex(off_t, size_field, st->st_size)
2566 ctf_float(double, size_dbl_field, (double) st->st_size)
2567 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2568 size_t, strlen(my_str_arg) / 2)
2573 You can refer to this tracepoint definition with the `tracepoint()`
2574 macro in your application's source code like this:
2577 .Application's source file.
2579 #define TRACEPOINT_DEFINE
2586 stat("/etc/fstab", &s);
2587 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2593 If you look at the event record that LTTng writes when tracing this
2594 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2595 it should look like this:
2597 .Event record fields
2599 |Field's name |Field's value
2600 |`my_constant_field` |40
2601 |`my_int_arg_field` |23
2602 |`my_int_arg_field2` |529
2604 |`my_str_arg_field` |`Hello, World!`
2605 |`size_field` |0x12d
2606 |`size_dbl_field` |301.0
2607 |`half_my_str_arg_field` |`Hello,`
2611 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2612 compute--they use the call stack, for example. To avoid this
2613 computation when the tracepoint is disabled, you can use the
2614 `tracepoint_enabled()` and `do_tracepoint()` macros.
2616 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2620 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2622 tracepoint_enabled(provider_name, tracepoint_name)
2623 do_tracepoint(provider_name, tracepoint_name, ...)
2628 * `provider_name` with the tracepoint provider name.
2629 * `tracepoint_name` with the tracepoint name.
2631 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2632 `tracepoint_name` from the provider named `provider_name` is enabled
2635 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2636 if the tracepoint is enabled. Using `tracepoint()` with
2637 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2638 the `tracepoint_enabled()` check, thus a race condition is
2639 possible in this situation:
2642 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2644 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2645 stuff = prepare_stuff();
2648 tracepoint(my_provider, my_tracepoint, stuff);
2651 If the tracepoint is enabled after the condition, then `stuff` is not
2652 prepared: the emitted event will either contain wrong data, or the whole
2653 application could crash (segmentation fault, for example).
2655 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2656 `STAP_PROBEV()` call. If you need it, you must emit
2660 [[building-tracepoint-providers-and-user-application]]
2661 ==== Build and link a tracepoint provider package and an application
2663 Once you have one or more <<tpp-header,tracepoint provider header
2664 files>> and a <<tpp-source,tracepoint provider package source file>>,
2665 you can create the tracepoint provider package by compiling its source
2666 file. From here, multiple build and run scenarios are possible. The
2667 following table shows common application and library configurations
2668 along with the required command lines to achieve them.
2670 In the following diagrams, we use the following file names:
2673 Executable application.
2676 Application's object file.
2679 Tracepoint provider package object file.
2682 Tracepoint provider package archive file.
2685 Tracepoint provider package shared object file.
2688 User library object file.
2691 User library shared object file.
2693 We use the following symbols in the diagrams of table below:
2696 .Symbols used in the build scenario diagrams.
2697 image::ust-sit-symbols.png[]
2699 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2700 variable in the following instructions.
2702 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2703 .Common tracepoint provider package scenarios.
2705 |Scenario |Instructions
2708 The instrumented application is statically linked with
2709 the tracepoint provider package object.
2711 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2714 include::../common/ust-sit-step-tp-o.txt[]
2716 To build the instrumented application:
2718 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2723 #define TRACEPOINT_DEFINE
2727 . Compile the application source file:
2736 . Build the application:
2741 $ gcc -o app app.o tpp.o -llttng-ust -ldl
2745 To run the instrumented application:
2747 * Start the application:
2757 The instrumented application is statically linked with the
2758 tracepoint provider package archive file.
2760 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2763 To create the tracepoint provider package archive file:
2765 . Compile the <<tpp-source,tracepoint provider package source file>>:
2774 . Create the tracepoint provider package archive file:
2779 $ ar rcs tpp.a tpp.o
2783 To build the instrumented application:
2785 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2790 #define TRACEPOINT_DEFINE
2794 . Compile the application source file:
2803 . Build the application:
2808 $ gcc -o app app.o tpp.a -llttng-ust -ldl
2812 To run the instrumented application:
2814 * Start the application:
2824 The instrumented application is linked with the tracepoint provider
2825 package shared object.
2827 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2830 include::../common/ust-sit-step-tp-so.txt[]
2832 To build the instrumented application:
2834 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2839 #define TRACEPOINT_DEFINE
2843 . Compile the application source file:
2852 . Build the application:
2857 $ gcc -o app app.o -ldl -L. -ltpp
2861 To run the instrumented application:
2863 * Start the application:
2873 The tracepoint provider package shared object is preloaded before the
2874 instrumented application starts.
2876 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2879 include::../common/ust-sit-step-tp-so.txt[]
2881 To build the instrumented application:
2883 . In path:{app.c}, before including path:{tpp.h}, add the
2889 #define TRACEPOINT_DEFINE
2890 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2894 . Compile the application source file:
2903 . Build the application:
2908 $ gcc -o app app.o -ldl
2912 To run the instrumented application with tracing support:
2914 * Preload the tracepoint provider package shared object and
2915 start the application:
2920 $ LD_PRELOAD=./libtpp.so ./app
2924 To run the instrumented application without tracing support:
2926 * Start the application:
2936 The instrumented application dynamically loads the tracepoint provider
2937 package shared object.
2939 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
2942 include::../common/ust-sit-step-tp-so.txt[]
2944 To build the instrumented application:
2946 . In path:{app.c}, before including path:{tpp.h}, add the
2952 #define TRACEPOINT_DEFINE
2953 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2957 . Compile the application source file:
2966 . Build the application:
2971 $ gcc -o app app.o -ldl
2975 To run the instrumented application:
2977 * Start the application:
2987 The application is linked with the instrumented user library.
2989 The instrumented user library is statically linked with the tracepoint
2990 provider package object file.
2992 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
2995 include::../common/ust-sit-step-tp-o-fpic.txt[]
2997 To build the instrumented user library:
2999 . In path:{emon.c}, before including path:{tpp.h}, add the
3005 #define TRACEPOINT_DEFINE
3009 . Compile the user library source file:
3014 $ gcc -I. -fpic -c emon.c
3018 . Build the user library shared object:
3023 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3027 To build the application:
3029 . Compile the application source file:
3038 . Build the application:
3043 $ gcc -o app app.o -L. -lemon
3047 To run the application:
3049 * Start the application:
3059 The application is linked with the instrumented user library.
3061 The instrumented user library is linked with the tracepoint provider
3062 package shared object.
3064 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3067 include::../common/ust-sit-step-tp-so.txt[]
3069 To build the instrumented user library:
3071 . In path:{emon.c}, before including path:{tpp.h}, add the
3077 #define TRACEPOINT_DEFINE
3081 . Compile the user library source file:
3086 $ gcc -I. -fpic -c emon.c
3090 . Build the user library shared object:
3095 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3099 To build the application:
3101 . Compile the application source file:
3110 . Build the application:
3115 $ gcc -o app app.o -L. -lemon
3119 To run the application:
3121 * Start the application:
3131 The tracepoint provider package shared object is preloaded before the
3134 The application is linked with the instrumented user library.
3136 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3139 include::../common/ust-sit-step-tp-so.txt[]
3141 To build the instrumented user library:
3143 . In path:{emon.c}, before including path:{tpp.h}, add the
3149 #define TRACEPOINT_DEFINE
3150 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3154 . Compile the user library source file:
3159 $ gcc -I. -fpic -c emon.c
3163 . Build the user library shared object:
3168 $ gcc -shared -o libemon.so emon.o -ldl
3172 To build the application:
3174 . Compile the application source file:
3183 . Build the application:
3188 $ gcc -o app app.o -L. -lemon
3192 To run the application with tracing support:
3194 * Preload the tracepoint provider package shared object and
3195 start the application:
3200 $ LD_PRELOAD=./libtpp.so ./app
3204 To run the application without tracing support:
3206 * Start the application:
3216 The application is linked with the instrumented user library.
3218 The instrumented user library dynamically loads the tracepoint provider
3219 package shared object.
3221 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3224 include::../common/ust-sit-step-tp-so.txt[]
3226 To build the instrumented user library:
3228 . In path:{emon.c}, before including path:{tpp.h}, add the
3234 #define TRACEPOINT_DEFINE
3235 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3239 . Compile the user library source file:
3244 $ gcc -I. -fpic -c emon.c
3248 . Build the user library shared object:
3253 $ gcc -shared -o libemon.so emon.o -ldl
3257 To build the application:
3259 . Compile the application source file:
3268 . Build the application:
3273 $ gcc -o app app.o -L. -lemon
3277 To run the application:
3279 * Start the application:
3289 The application dynamically loads the instrumented user library.
3291 The instrumented user library is linked with the tracepoint provider
3292 package shared object.
3294 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3297 include::../common/ust-sit-step-tp-so.txt[]
3299 To build the instrumented user library:
3301 . In path:{emon.c}, before including path:{tpp.h}, add the
3307 #define TRACEPOINT_DEFINE
3311 . Compile the user library source file:
3316 $ gcc -I. -fpic -c emon.c
3320 . Build the user library shared object:
3325 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3329 To build the application:
3331 . Compile the application source file:
3340 . Build the application:
3345 $ gcc -o app app.o -ldl -L. -lemon
3349 To run the application:
3351 * Start the application:
3361 The application dynamically loads the instrumented user library.
3363 The instrumented user library dynamically loads the tracepoint provider
3364 package shared object.
3366 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3369 include::../common/ust-sit-step-tp-so.txt[]
3371 To build the instrumented user library:
3373 . In path:{emon.c}, before including path:{tpp.h}, add the
3379 #define TRACEPOINT_DEFINE
3380 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3384 . Compile the user library source file:
3389 $ gcc -I. -fpic -c emon.c
3393 . Build the user library shared object:
3398 $ gcc -shared -o libemon.so emon.o -ldl
3402 To build the application:
3404 . Compile the application source file:
3413 . Build the application:
3418 $ gcc -o app app.o -ldl -L. -lemon
3422 To run the application:
3424 * Start the application:
3434 The tracepoint provider package shared object is preloaded before the
3437 The application dynamically loads the instrumented user library.
3439 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3442 include::../common/ust-sit-step-tp-so.txt[]
3444 To build the instrumented user library:
3446 . In path:{emon.c}, before including path:{tpp.h}, add the
3452 #define TRACEPOINT_DEFINE
3453 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3457 . Compile the user library source file:
3462 $ gcc -I. -fpic -c emon.c
3466 . Build the user library shared object:
3471 $ gcc -shared -o libemon.so emon.o -ldl
3475 To build the application:
3477 . Compile the application source file:
3486 . Build the application:
3491 $ gcc -o app app.o -L. -lemon
3495 To run the application with tracing support:
3497 * Preload the tracepoint provider package shared object and
3498 start the application:
3503 $ LD_PRELOAD=./libtpp.so ./app
3507 To run the application without tracing support:
3509 * Start the application:
3519 The application is statically linked with the tracepoint provider
3520 package object file.
3522 The application is linked with the instrumented user library.
3524 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3527 include::../common/ust-sit-step-tp-o.txt[]
3529 To build the instrumented user library:
3531 . In path:{emon.c}, before including path:{tpp.h}, add the
3537 #define TRACEPOINT_DEFINE
3541 . Compile the user library source file:
3546 $ gcc -I. -fpic -c emon.c
3550 . Build the user library shared object:
3555 $ gcc -shared -o libemon.so emon.o
3559 To build the application:
3561 . Compile the application source file:
3570 . Build the application:
3575 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3579 To run the instrumented application:
3581 * Start the application:
3591 The application is statically linked with the tracepoint provider
3592 package object file.
3594 The application dynamically loads the instrumented user library.
3596 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3599 include::../common/ust-sit-step-tp-o.txt[]
3601 To build the application:
3603 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3608 #define TRACEPOINT_DEFINE
3612 . Compile the application source file:
3621 . Build the application:
3626 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
3631 The `--export-dynamic` option passed to the linker is necessary for the
3632 dynamically loaded library to ``see'' the tracepoint symbols defined in
3635 To build the instrumented user library:
3637 . Compile the user library source file:
3642 $ gcc -I. -fpic -c emon.c
3646 . Build the user library shared object:
3651 $ gcc -shared -o libemon.so emon.o
3655 To run the application:
3657 * Start the application:
3668 [[using-lttng-ust-with-daemons]]
3669 ===== Use noch:{LTTng-UST} with daemons
3671 If your instrumented application calls man:fork(2), man:clone(2),
3672 or BSD's man:rfork(2), without a following man:exec(3)-family
3673 system call, you must preload the path:{liblttng-ust-fork.so} shared
3674 object when you start the application.
3678 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
3681 If your tracepoint provider package is
3682 a shared library which you also preload, you must put both
3683 shared objects in env:LD_PRELOAD:
3687 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3693 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3695 If your instrumented application closes one or more file descriptors
3696 which it did not open itself, you must preload the
3697 path:{liblttng-ust-fd.so} shared object when you start the application:
3701 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
3704 Typical use cases include closing all the file descriptors after
3705 man:fork(2) or man:rfork(2) and buggy applications doing
3709 [[lttng-ust-pkg-config]]
3710 ===== Use noch:{pkg-config}
3712 On some distributions, LTTng-UST ships with a
3713 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3714 metadata file. If this is your case, then you can use cmd:pkg-config to
3715 build an application on the command line:
3719 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3723 [[instrumenting-32-bit-app-on-64-bit-system]]
3724 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3726 In order to trace a 32-bit application running on a 64-bit system,
3727 LTTng must use a dedicated 32-bit
3728 <<lttng-consumerd,consumer daemon>>.
3730 The following steps show how to build and install a 32-bit consumer
3731 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3732 build and install the 32-bit LTTng-UST libraries, and how to build and
3733 link an instrumented 32-bit application in that context.
3735 To build a 32-bit instrumented application for a 64-bit target system,
3736 assuming you have a fresh target system with no installed Userspace RCU
3739 . Download, build, and install a 32-bit version of Userspace RCU:
3744 $ cd $(mktemp -d) &&
3745 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3746 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3747 cd userspace-rcu-0.9.* &&
3748 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3750 sudo make install &&
3755 . Using your distribution's package manager, or from source, install
3756 the following 32-bit versions of the following dependencies of
3757 LTTng-tools and LTTng-UST:
3760 * https://sourceforge.net/projects/libuuid/[libuuid]
3761 * http://directory.fsf.org/wiki/Popt[popt]
3762 * http://www.xmlsoft.org/[libxml2]
3765 . Download, build, and install a 32-bit version of the latest
3766 LTTng-UST{nbsp}{revision}:
3771 $ cd $(mktemp -d) &&
3772 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.11.tar.bz2 &&
3773 tar -xf lttng-ust-latest-2.11.tar.bz2 &&
3774 cd lttng-ust-2.11.* &&
3775 ./configure --libdir=/usr/local/lib32 \
3776 CFLAGS=-m32 CXXFLAGS=-m32 \
3777 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3779 sudo make install &&
3786 Depending on your distribution,
3787 32-bit libraries could be installed at a different location than
3788 `/usr/lib32`. For example, Debian is known to install
3789 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3791 In this case, make sure to set `LDFLAGS` to all the
3792 relevant 32-bit library paths, for example:
3796 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3800 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3801 the 32-bit consumer daemon:
3806 $ cd $(mktemp -d) &&
3807 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
3808 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
3809 cd lttng-tools-2.11.* &&
3810 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3811 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
3812 --disable-bin-lttng --disable-bin-lttng-crash \
3813 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
3815 cd src/bin/lttng-consumerd &&
3816 sudo make install &&
3821 . From your distribution or from source,
3822 <<installing-lttng,install>> the 64-bit versions of
3823 LTTng-UST and Userspace RCU.
3824 . Download, build, and install the 64-bit version of the
3825 latest LTTng-tools{nbsp}{revision}:
3830 $ cd $(mktemp -d) &&
3831 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
3832 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
3833 cd lttng-tools-2.11.* &&
3834 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3835 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3837 sudo make install &&
3842 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3843 when linking your 32-bit application:
3846 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3847 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3850 For example, let's rebuild the quick start example in
3851 <<tracing-your-own-user-application,Trace a user application>> as an
3852 instrumented 32-bit application:
3857 $ gcc -m32 -c -I. hello-tp.c
3858 $ gcc -m32 -c hello.c
3859 $ gcc -m32 -o hello hello.o hello-tp.o \
3860 -L/usr/lib32 -L/usr/local/lib32 \
3861 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3866 No special action is required to execute the 32-bit application and
3867 to trace it: use the command-line man:lttng(1) tool as usual.
3874 man:tracef(3) is a small LTTng-UST API designed for quick,
3875 man:printf(3)-like instrumentation without the burden of
3876 <<tracepoint-provider,creating>> and
3877 <<building-tracepoint-providers-and-user-application,building>>
3878 a tracepoint provider package.
3880 To use `tracef()` in your application:
3882 . In the C or C++ source files where you need to use `tracef()`,
3883 include `<lttng/tracef.h>`:
3888 #include <lttng/tracef.h>
3892 . In the application's source code, use `tracef()` like you would use
3900 tracef("my message: %d (%s)", my_integer, my_string);
3906 . Link your application with `liblttng-ust`:
3911 $ gcc -o app app.c -llttng-ust
3915 To trace the events that `tracef()` calls emit:
3917 * <<enabling-disabling-events,Create an event rule>> which matches the
3918 `lttng_ust_tracef:*` event name:
3923 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
3928 .Limitations of `tracef()`
3930 The `tracef()` utility function was developed to make user space tracing
3931 super simple, albeit with notable disadvantages compared to
3932 <<defining-tracepoints,user-defined tracepoints>>:
3934 * All the emitted events have the same tracepoint provider and
3935 tracepoint names, respectively `lttng_ust_tracef` and `event`.
3936 * There is no static type checking.
3937 * The only event record field you actually get, named `msg`, is a string
3938 potentially containing the values you passed to `tracef()`
3939 using your own format string. This also means that you cannot filter
3940 events with a custom expression at run time because there are no
3942 * Since `tracef()` uses the C standard library's man:vasprintf(3)
3943 function behind the scenes to format the strings at run time, its
3944 expected performance is lower than with user-defined tracepoints,
3945 which do not require a conversion to a string.
3947 Taking this into consideration, `tracef()` is useful for some quick
3948 prototyping and debugging, but you should not consider it for any
3949 permanent and serious applicative instrumentation.
3955 ==== Use `tracelog()`
3957 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
3958 the difference that it accepts an additional log level parameter.
3960 The goal of `tracelog()` is to ease the migration from logging to
3963 To use `tracelog()` in your application:
3965 . In the C or C++ source files where you need to use `tracelog()`,
3966 include `<lttng/tracelog.h>`:
3971 #include <lttng/tracelog.h>
3975 . In the application's source code, use `tracelog()` like you would use
3976 man:printf(3), except for the first parameter which is the log
3984 tracelog(TRACE_WARNING, "my message: %d (%s)",
3985 my_integer, my_string);
3991 See man:lttng-ust(3) for a list of available log level names.
3993 . Link your application with `liblttng-ust`:
3998 $ gcc -o app app.c -llttng-ust
4002 To trace the events that `tracelog()` calls emit with a log level
4003 _as severe as_ a specific log level:
4005 * <<enabling-disabling-events,Create an event rule>> which matches the
4006 `lttng_ust_tracelog:*` event name and a minimum level
4012 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4013 --loglevel=TRACE_WARNING
4017 To trace the events that `tracelog()` calls emit with a
4018 _specific log level_:
4020 * Create an event rule which matches the `lttng_ust_tracelog:*`
4021 event name and a specific log level:
4026 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4027 --loglevel-only=TRACE_INFO
4032 [[prebuilt-ust-helpers]]
4033 === Prebuilt user space tracing helpers
4035 The LTTng-UST package provides a few helpers in the form or preloadable
4036 shared objects which automatically instrument system functions and
4039 The helper shared objects are normally found in dir:{/usr/lib}. If you
4040 built LTTng-UST <<building-from-source,from source>>, they are probably
4041 located in dir:{/usr/local/lib}.
4043 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4046 path:{liblttng-ust-libc-wrapper.so}::
4047 path:{liblttng-ust-pthread-wrapper.so}::
4048 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4049 memory and POSIX threads function tracing>>.
4051 path:{liblttng-ust-cyg-profile.so}::
4052 path:{liblttng-ust-cyg-profile-fast.so}::
4053 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4055 path:{liblttng-ust-dl.so}::
4056 <<liblttng-ust-dl,Dynamic linker tracing>>.
4058 To use a user space tracing helper with any user application:
4060 * Preload the helper shared object when you start the application:
4065 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4069 You can preload more than one helper:
4074 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4080 [[liblttng-ust-libc-pthread-wrapper]]
4081 ==== Instrument C standard library memory and POSIX threads functions
4083 The path:{liblttng-ust-libc-wrapper.so} and
4084 path:{liblttng-ust-pthread-wrapper.so} helpers
4085 add instrumentation to some C standard library and POSIX
4089 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4091 |TP provider name |TP name |Instrumented function
4093 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4094 |`calloc` |man:calloc(3)
4095 |`realloc` |man:realloc(3)
4096 |`free` |man:free(3)
4097 |`memalign` |man:memalign(3)
4098 |`posix_memalign` |man:posix_memalign(3)
4102 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4104 |TP provider name |TP name |Instrumented function
4106 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4107 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4108 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4109 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4112 When you preload the shared object, it replaces the functions listed
4113 in the previous tables by wrappers which contain tracepoints and call
4114 the replaced functions.
4117 [[liblttng-ust-cyg-profile]]
4118 ==== Instrument function entry and exit
4120 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4121 to the entry and exit points of functions.
4123 man:gcc(1) and man:clang(1) have an option named
4124 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4125 which generates instrumentation calls for entry and exit to functions.
4126 The LTTng-UST function tracing helpers,
4127 path:{liblttng-ust-cyg-profile.so} and
4128 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4129 to add tracepoints to the two generated functions (which contain
4130 `cyg_profile` in their names, hence the helper's name).
4132 To use the LTTng-UST function tracing helper, the source files to
4133 instrument must be built using the `-finstrument-functions` compiler
4136 There are two versions of the LTTng-UST function tracing helper:
4138 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4139 that you should only use when it can be _guaranteed_ that the
4140 complete event stream is recorded without any lost event record.
4141 Any kind of duplicate information is left out.
4143 Assuming no event record is lost, having only the function addresses on
4144 entry is enough to create a call graph, since an event record always
4145 contains the ID of the CPU that generated it.
4147 You can use a tool like man:addr2line(1) to convert function addresses
4148 back to source file names and line numbers.
4150 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4151 which also works in use cases where event records might get discarded or
4152 not recorded from application startup.
4153 In these cases, the trace analyzer needs more information to be
4154 able to reconstruct the program flow.
4156 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4157 points of this helper.
4159 All the tracepoints that this helper provides have the
4160 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4162 TIP: It's sometimes a good idea to limit the number of source files that
4163 you compile with the `-finstrument-functions` option to prevent LTTng
4164 from writing an excessive amount of trace data at run time. When using
4165 man:gcc(1), you can use the
4166 `-finstrument-functions-exclude-function-list` option to avoid
4167 instrument entries and exits of specific function names.
4172 ==== Instrument the dynamic linker
4174 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4175 man:dlopen(3) and man:dlclose(3) function calls.
4177 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4182 [[java-application]]
4183 === User space Java agent
4185 You can instrument any Java application which uses one of the following
4188 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4189 (JUL) core logging facilities.
4190 * http://logging.apache.org/log4j/1.2/[**Apache log4j{nbsp}1.2**], since
4191 LTTng{nbsp}2.6. Note that Apache Log4j{nbsp}2 is not supported.
4194 .LTTng-UST Java agent imported by a Java application.
4195 image::java-app.png[]
4197 Note that the methods described below are new in LTTng{nbsp}{revision}.
4198 Previous LTTng versions use another technique.
4200 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4201 and https://ci.lttng.org/[continuous integration], thus this version is
4202 directly supported. However, the LTTng-UST Java agent is also tested
4203 with OpenJDK{nbsp}7.
4208 ==== Use the LTTng-UST Java agent for `java.util.logging`
4210 To use the LTTng-UST Java agent in a Java application which uses
4211 `java.util.logging` (JUL):
4213 . In the Java application's source code, import the LTTng-UST
4214 log handler package for `java.util.logging`:
4219 import org.lttng.ust.agent.jul.LttngLogHandler;
4223 . Create an LTTng-UST JUL log handler:
4228 Handler lttngUstLogHandler = new LttngLogHandler();
4232 . Add this handler to the JUL loggers which should emit LTTng events:
4237 Logger myLogger = Logger.getLogger("some-logger");
4239 myLogger.addHandler(lttngUstLogHandler);
4243 . Use `java.util.logging` log statements and configuration as usual.
4244 The loggers with an attached LTTng-UST log handler can emit
4247 . Before exiting the application, remove the LTTng-UST log handler from
4248 the loggers attached to it and call its `close()` method:
4253 myLogger.removeHandler(lttngUstLogHandler);
4254 lttngUstLogHandler.close();
4258 This is not strictly necessary, but it is recommended for a clean
4259 disposal of the handler's resources.
4261 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4262 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4264 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4265 path] when you build the Java application.
4267 The JAR files are typically located in dir:{/usr/share/java}.
4269 IMPORTANT: The LTTng-UST Java agent must be
4270 <<installing-lttng,installed>> for the logging framework your
4273 .Use the LTTng-UST Java agent for `java.util.logging`.
4278 import java.io.IOException;
4279 import java.util.logging.Handler;
4280 import java.util.logging.Logger;
4281 import org.lttng.ust.agent.jul.LttngLogHandler;
4285 private static final int answer = 42;
4287 public static void main(String[] argv) throws Exception
4290 Logger logger = Logger.getLogger("jello");
4292 // Create an LTTng-UST log handler
4293 Handler lttngUstLogHandler = new LttngLogHandler();
4295 // Add the LTTng-UST log handler to our logger
4296 logger.addHandler(lttngUstLogHandler);
4299 logger.info("some info");
4300 logger.warning("some warning");
4302 logger.finer("finer information; the answer is " + answer);
4304 logger.severe("error!");
4306 // Not mandatory, but cleaner
4307 logger.removeHandler(lttngUstLogHandler);
4308 lttngUstLogHandler.close();
4317 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4320 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4321 <<enabling-disabling-events,create an event rule>> matching the
4322 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4327 $ lttng enable-event --jul jello
4331 Run the compiled class:
4335 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4338 <<basic-tracing-session-control,Stop tracing>> and inspect the
4348 In the resulting trace, an <<event,event record>> generated by a Java
4349 application using `java.util.logging` is named `lttng_jul:event` and
4350 has the following fields:
4353 Log record's message.
4359 Name of the class in which the log statement was executed.
4362 Name of the method in which the log statement was executed.
4365 Logging time (timestamp in milliseconds).
4368 Log level integer value.
4371 ID of the thread in which the log statement was executed.
4373 You can use the opt:lttng-enable-event(1):--loglevel or
4374 opt:lttng-enable-event(1):--loglevel-only option of the
4375 man:lttng-enable-event(1) command to target a range of JUL log levels
4376 or a specific JUL log level.
4381 ==== Use the LTTng-UST Java agent for Apache log4j
4383 To use the LTTng-UST Java agent in a Java application which uses
4384 Apache log4j{nbsp}1.2:
4386 . In the Java application's source code, import the LTTng-UST
4387 log appender package for Apache log4j:
4392 import org.lttng.ust.agent.log4j.LttngLogAppender;
4396 . Create an LTTng-UST log4j log appender:
4401 Appender lttngUstLogAppender = new LttngLogAppender();
4405 . Add this appender to the log4j loggers which should emit LTTng events:
4410 Logger myLogger = Logger.getLogger("some-logger");
4412 myLogger.addAppender(lttngUstLogAppender);
4416 . Use Apache log4j log statements and configuration as usual. The
4417 loggers with an attached LTTng-UST log appender can emit LTTng events.
4419 . Before exiting the application, remove the LTTng-UST log appender from
4420 the loggers attached to it and call its `close()` method:
4425 myLogger.removeAppender(lttngUstLogAppender);
4426 lttngUstLogAppender.close();
4430 This is not strictly necessary, but it is recommended for a clean
4431 disposal of the appender's resources.
4433 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4434 files, path:{lttng-ust-agent-common.jar} and
4435 path:{lttng-ust-agent-log4j.jar}, in the
4436 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4437 path] when you build the Java application.
4439 The JAR files are typically located in dir:{/usr/share/java}.
4441 IMPORTANT: The LTTng-UST Java agent must be
4442 <<installing-lttng,installed>> for the logging framework your
4445 .Use the LTTng-UST Java agent for Apache log4j.
4450 import org.apache.log4j.Appender;
4451 import org.apache.log4j.Logger;
4452 import org.lttng.ust.agent.log4j.LttngLogAppender;
4456 private static final int answer = 42;
4458 public static void main(String[] argv) throws Exception
4461 Logger logger = Logger.getLogger("jello");
4463 // Create an LTTng-UST log appender
4464 Appender lttngUstLogAppender = new LttngLogAppender();
4466 // Add the LTTng-UST log appender to our logger
4467 logger.addAppender(lttngUstLogAppender);
4470 logger.info("some info");
4471 logger.warn("some warning");
4473 logger.debug("debug information; the answer is " + answer);
4475 logger.fatal("error!");
4477 // Not mandatory, but cleaner
4478 logger.removeAppender(lttngUstLogAppender);
4479 lttngUstLogAppender.close();
4485 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4490 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4493 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4494 <<enabling-disabling-events,create an event rule>> matching the
4495 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4500 $ lttng enable-event --log4j jello
4504 Run the compiled class:
4508 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4511 <<basic-tracing-session-control,Stop tracing>> and inspect the
4521 In the resulting trace, an <<event,event record>> generated by a Java
4522 application using log4j is named `lttng_log4j:event` and
4523 has the following fields:
4526 Log record's message.
4532 Name of the class in which the log statement was executed.
4535 Name of the method in which the log statement was executed.
4538 Name of the file in which the executed log statement is located.
4541 Line number at which the log statement was executed.
4547 Log level integer value.
4550 Name of the Java thread in which the log statement was executed.
4552 You can use the opt:lttng-enable-event(1):--loglevel or
4553 opt:lttng-enable-event(1):--loglevel-only option of the
4554 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4555 or a specific log4j log level.
4559 [[java-application-context]]
4560 ==== Provide application-specific context fields in a Java application
4562 A Java application-specific context field is a piece of state provided
4563 by the application which <<adding-context,you can add>>, using the
4564 man:lttng-add-context(1) command, to each <<event,event record>>
4565 produced by the log statements of this application.
4567 For example, a given object might have a current request ID variable.
4568 You can create a context information retriever for this object and
4569 assign a name to this current request ID. You can then, using the
4570 man:lttng-add-context(1) command, add this context field by name to
4571 the JUL or log4j <<channel,channel>>.
4573 To provide application-specific context fields in a Java application:
4575 . In the Java application's source code, import the LTTng-UST
4576 Java agent context classes and interfaces:
4581 import org.lttng.ust.agent.context.ContextInfoManager;
4582 import org.lttng.ust.agent.context.IContextInfoRetriever;
4586 . Create a context information retriever class, that is, a class which
4587 implements the `IContextInfoRetriever` interface:
4592 class MyContextInfoRetriever implements IContextInfoRetriever
4595 public Object retrieveContextInfo(String key)
4597 if (key.equals("intCtx")) {
4599 } else if (key.equals("strContext")) {
4600 return "context value!";
4609 This `retrieveContextInfo()` method is the only member of the
4610 `IContextInfoRetriever` interface. Its role is to return the current
4611 value of a state by name to create a context field. The names of the
4612 context fields and which state variables they return depends on your
4615 All primitive types and objects are supported as context fields.
4616 When `retrieveContextInfo()` returns an object, the context field
4617 serializer calls its `toString()` method to add a string field to
4618 event records. The method can also return `null`, which means that
4619 no context field is available for the required name.
4621 . Register an instance of your context information retriever class to
4622 the context information manager singleton:
4627 IContextInfoRetriever cir = new MyContextInfoRetriever();
4628 ContextInfoManager cim = ContextInfoManager.getInstance();
4629 cim.registerContextInfoRetriever("retrieverName", cir);
4633 . Before exiting the application, remove your context information
4634 retriever from the context information manager singleton:
4639 ContextInfoManager cim = ContextInfoManager.getInstance();
4640 cim.unregisterContextInfoRetriever("retrieverName");
4644 This is not strictly necessary, but it is recommended for a clean
4645 disposal of some manager's resources.
4647 . Build your Java application with LTTng-UST Java agent support as
4648 usual, following the procedure for either the <<jul,JUL>> or
4649 <<log4j,Apache log4j>> framework.
4652 .Provide application-specific context fields in a Java application.
4657 import java.util.logging.Handler;
4658 import java.util.logging.Logger;
4659 import org.lttng.ust.agent.jul.LttngLogHandler;
4660 import org.lttng.ust.agent.context.ContextInfoManager;
4661 import org.lttng.ust.agent.context.IContextInfoRetriever;
4665 // Our context information retriever class
4666 private static class MyContextInfoRetriever
4667 implements IContextInfoRetriever
4670 public Object retrieveContextInfo(String key) {
4671 if (key.equals("intCtx")) {
4673 } else if (key.equals("strContext")) {
4674 return "context value!";
4681 private static final int answer = 42;
4683 public static void main(String args[]) throws Exception
4685 // Get the context information manager instance
4686 ContextInfoManager cim = ContextInfoManager.getInstance();
4688 // Create and register our context information retriever
4689 IContextInfoRetriever cir = new MyContextInfoRetriever();
4690 cim.registerContextInfoRetriever("myRetriever", cir);
4693 Logger logger = Logger.getLogger("jello");
4695 // Create an LTTng-UST log handler
4696 Handler lttngUstLogHandler = new LttngLogHandler();
4698 // Add the LTTng-UST log handler to our logger
4699 logger.addHandler(lttngUstLogHandler);
4702 logger.info("some info");
4703 logger.warning("some warning");
4705 logger.finer("finer information; the answer is " + answer);
4707 logger.severe("error!");
4709 // Not mandatory, but cleaner
4710 logger.removeHandler(lttngUstLogHandler);
4711 lttngUstLogHandler.close();
4712 cim.unregisterContextInfoRetriever("myRetriever");
4721 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4724 <<creating-destroying-tracing-sessions,Create a tracing session>>
4725 and <<enabling-disabling-events,create an event rule>> matching the
4731 $ lttng enable-event --jul jello
4734 <<adding-context,Add the application-specific context fields>> to the
4739 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
4740 $ lttng add-context --jul --type='$app.myRetriever:strContext'
4743 <<basic-tracing-session-control,Start tracing>>:
4750 Run the compiled class:
4754 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4757 <<basic-tracing-session-control,Stop tracing>> and inspect the
4769 [[python-application]]
4770 === User space Python agent
4772 You can instrument a Python{nbsp}2 or Python{nbsp}3 application which
4774 https://docs.python.org/3/library/logging.html[`logging`] package.
4776 Each log statement emits an LTTng event once the
4777 application module imports the
4778 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4781 .A Python application importing the LTTng-UST Python agent.
4782 image::python-app.png[]
4784 To use the LTTng-UST Python agent:
4786 . In the Python application's source code, import the LTTng-UST Python
4796 The LTTng-UST Python agent automatically adds its logging handler to the
4797 root logger at import time.
4799 Any log statement that the application executes before this import does
4800 not emit an LTTng event.
4802 IMPORTANT: The LTTng-UST Python agent must be
4803 <<installing-lttng,installed>>.
4805 . Use log statements and logging configuration as usual.
4806 Since the LTTng-UST Python agent adds a handler to the _root_
4807 logger, you can trace any log statement from any logger.
4809 .Use the LTTng-UST Python agent.
4820 logging.basicConfig()
4821 logger = logging.getLogger('my-logger')
4824 logger.debug('debug message')
4825 logger.info('info message')
4826 logger.warn('warn message')
4827 logger.error('error message')
4828 logger.critical('critical message')
4832 if __name__ == '__main__':
4836 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4837 logging handler which prints to the standard error stream, is not
4838 strictly required for LTTng-UST tracing to work, but in versions of
4839 Python preceding{nbsp}3.2, you could see a warning message which indicates
4840 that no handler exists for the logger `my-logger`.
4842 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4843 <<enabling-disabling-events,create an event rule>> matching the
4844 `my-logger` Python logger, and <<basic-tracing-session-control,start
4850 $ lttng enable-event --python my-logger
4854 Run the Python script:
4861 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4871 In the resulting trace, an <<event,event record>> generated by a Python
4872 application is named `lttng_python:event` and has the following fields:
4875 Logging time (string).
4878 Log record's message.
4884 Name of the function in which the log statement was executed.
4887 Line number at which the log statement was executed.
4890 Log level integer value.
4893 ID of the Python thread in which the log statement was executed.
4896 Name of the Python thread in which the log statement was executed.
4898 You can use the opt:lttng-enable-event(1):--loglevel or
4899 opt:lttng-enable-event(1):--loglevel-only option of the
4900 man:lttng-enable-event(1) command to target a range of Python log levels
4901 or a specific Python log level.
4903 When an application imports the LTTng-UST Python agent, the agent tries
4904 to register to a <<lttng-sessiond,session daemon>>. Note that you must
4905 <<start-sessiond,start the session daemon>> _before_ you run the Python
4906 application. If a session daemon is found, the agent tries to register
4907 to it during five seconds, after which the application continues
4908 without LTTng tracing support. You can override this timeout value with
4909 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
4912 If the session daemon stops while a Python application with an imported
4913 LTTng-UST Python agent runs, the agent retries to connect and to
4914 register to a session daemon every three seconds. You can override this
4915 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
4920 [[proc-lttng-logger-abi]]
4923 The `lttng-tracer` Linux kernel module, part of
4924 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
4925 path:{/proc/lttng-logger} when it's loaded. Any application can write
4926 text data to this file to emit an LTTng event.
4929 .An application writes to the LTTng logger file to emit an LTTng event.
4930 image::lttng-logger.png[]
4932 The LTTng logger is the quickest method--not the most efficient,
4933 however--to add instrumentation to an application. It is designed
4934 mostly to instrument shell scripts:
4938 $ echo "Some message, some $variable" > /proc/lttng-logger
4941 Any event that the LTTng logger emits is named `lttng_logger` and
4942 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
4943 other instrumentation points in the kernel tracing domain, **any Unix
4944 user** can <<enabling-disabling-events,create an event rule>> which
4945 matches its event name, not only the root user or users in the
4946 <<tracing-group,tracing group>>.
4948 To use the LTTng logger:
4950 * From any application, write text data to the path:{/proc/lttng-logger}
4953 The `msg` field of `lttng_logger` event records contains the
4956 NOTE: The maximum message length of an LTTng logger event is
4957 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
4958 than one event to contain the remaining data.
4960 You should not use the LTTng logger to trace a user application which
4961 can be instrumented in a more efficient way, namely:
4963 * <<c-application,C and $$C++$$ applications>>.
4964 * <<java-application,Java applications>>.
4965 * <<python-application,Python applications>>.
4967 .Use the LTTng logger.
4972 echo 'Hello, World!' > /proc/lttng-logger
4974 df --human-readable --print-type / > /proc/lttng-logger
4977 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4978 <<enabling-disabling-events,create an event rule>> matching the
4979 `lttng_logger` Linux kernel tracepoint, and
4980 <<basic-tracing-session-control,start tracing>>:
4985 $ lttng enable-event --kernel lttng_logger
4989 Run the Bash script:
4996 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5007 [[instrumenting-linux-kernel]]
5008 === LTTng kernel tracepoints
5010 NOTE: This section shows how to _add_ instrumentation points to the
5011 Linux kernel. The kernel's subsystems are already thoroughly
5012 instrumented at strategic places for LTTng when you
5013 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5017 There are two methods to instrument the Linux kernel:
5019 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5020 tracepoint which uses the `TRACE_EVENT()` API.
5022 Choose this if you want to instrumentation a Linux kernel tree with an
5023 instrumentation point compatible with ftrace, perf, and SystemTap.
5025 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5026 instrument an out-of-tree kernel module.
5028 Choose this if you don't need ftrace, perf, or SystemTap support.
5032 [[linux-add-lttng-layer]]
5033 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5035 This section shows how to add an LTTng layer to existing ftrace
5036 instrumentation using the `TRACE_EVENT()` API.
5038 This section does not document the `TRACE_EVENT()` macro. You can
5039 read the following articles to learn more about this API:
5041 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part{nbsp}1)]
5042 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part{nbsp}2)]
5043 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part{nbsp}3)]
5045 The following procedure assumes that your ftrace tracepoints are
5046 correctly defined in their own header and that they are created in
5047 one source file using the `CREATE_TRACE_POINTS` definition.
5049 To add an LTTng layer over an existing ftrace tracepoint:
5051 . Make sure the following kernel configuration options are
5057 * `CONFIG_HIGH_RES_TIMERS`
5058 * `CONFIG_TRACEPOINTS`
5061 . Build the Linux source tree with your custom ftrace tracepoints.
5062 . Boot the resulting Linux image on your target system.
5064 Confirm that the tracepoints exist by looking for their names in the
5065 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5066 is your subsystem's name.
5068 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5073 $ cd $(mktemp -d) &&
5074 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
5075 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
5076 cd lttng-modules-2.11.*
5080 . In dir:{instrumentation/events/lttng-module}, relative to the root
5081 of the LTTng-modules source tree, create a header file named
5082 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5083 LTTng-modules tracepoint definitions using the LTTng-modules
5086 Start with this template:
5090 .path:{instrumentation/events/lttng-module/my_subsys.h}
5093 #define TRACE_SYSTEM my_subsys
5095 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5096 #define _LTTNG_MY_SUBSYS_H
5098 #include "../../../probes/lttng-tracepoint-event.h"
5099 #include <linux/tracepoint.h>
5101 LTTNG_TRACEPOINT_EVENT(
5103 * Format is identical to TRACE_EVENT()'s version for the three
5104 * following macro parameters:
5107 TP_PROTO(int my_int, const char *my_string),
5108 TP_ARGS(my_int, my_string),
5110 /* LTTng-modules specific macros */
5112 ctf_integer(int, my_int_field, my_int)
5113 ctf_string(my_bar_field, my_bar)
5117 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5119 #include "../../../probes/define_trace.h"
5123 The entries in the `TP_FIELDS()` section are the list of fields for the
5124 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5125 ftrace's `TRACE_EVENT()` macro.
5127 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5128 complete description of the available `ctf_*()` macros.
5130 . Create the LTTng-modules probe's kernel module C source file,
5131 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5136 .path:{probes/lttng-probe-my-subsys.c}
5138 #include <linux/module.h>
5139 #include "../lttng-tracer.h"
5142 * Build-time verification of mismatch between mainline
5143 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5144 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5146 #include <trace/events/my_subsys.h>
5148 /* Create LTTng tracepoint probes */
5149 #define LTTNG_PACKAGE_BUILD
5150 #define CREATE_TRACE_POINTS
5151 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5153 #include "../instrumentation/events/lttng-module/my_subsys.h"
5155 MODULE_LICENSE("GPL and additional rights");
5156 MODULE_AUTHOR("Your name <your-email>");
5157 MODULE_DESCRIPTION("LTTng my_subsys probes");
5158 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5159 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5160 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5161 LTTNG_MODULES_EXTRAVERSION);
5165 . Edit path:{probes/KBuild} and add your new kernel module object
5166 next to the existing ones:
5170 .path:{probes/KBuild}
5174 obj-m += lttng-probe-module.o
5175 obj-m += lttng-probe-power.o
5177 obj-m += lttng-probe-my-subsys.o
5183 . Build and install the LTTng kernel modules:
5188 $ make KERNELDIR=/path/to/linux
5189 # make modules_install && depmod -a
5193 Replace `/path/to/linux` with the path to the Linux source tree where
5194 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5196 Note that you can also use the
5197 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5198 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5199 C code that need to be executed before the event fields are recorded.
5201 The best way to learn how to use the previous LTTng-modules macros is to
5202 inspect the existing LTTng-modules tracepoint definitions in the
5203 dir:{instrumentation/events/lttng-module} header files. Compare them
5204 with the Linux kernel mainline versions in the
5205 dir:{include/trace/events} directory of the Linux source tree.
5209 [[lttng-tracepoint-event-code]]
5210 ===== Use custom C code to access the data for tracepoint fields
5212 Although we recommended to always use the
5213 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5214 the arguments and fields of an LTTng-modules tracepoint when possible,
5215 sometimes you need a more complex process to access the data that the
5216 tracer records as event record fields. In other words, you need local
5217 variables and multiple C{nbsp}statements instead of simple
5218 argument-based expressions that you pass to the
5219 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5221 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5222 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5223 a block of C{nbsp}code to be executed before LTTng records the fields.
5224 The structure of this macro is:
5227 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5229 LTTNG_TRACEPOINT_EVENT_CODE(
5231 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5232 * version for the following three macro parameters:
5235 TP_PROTO(int my_int, const char *my_string),
5236 TP_ARGS(my_int, my_string),
5238 /* Declarations of custom local variables */
5241 unsigned long b = 0;
5242 const char *name = "(undefined)";
5243 struct my_struct *my_struct;
5247 * Custom code which uses both tracepoint arguments
5248 * (in TP_ARGS()) and local variables (in TP_locvar()).
5250 * Local variables are actually members of a structure pointed
5251 * to by the special variable tp_locvar.
5255 tp_locvar->a = my_int + 17;
5256 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5257 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5258 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5259 put_my_struct(tp_locvar->my_struct);
5268 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5269 * version for this, except that tp_locvar members can be
5270 * used in the argument expression parameters of
5271 * the ctf_*() macros.
5274 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5275 ctf_integer(int, my_struct_a, tp_locvar->a)
5276 ctf_string(my_string_field, my_string)
5277 ctf_string(my_struct_name, tp_locvar->name)
5282 IMPORTANT: The C code defined in `TP_code()` must not have any side
5283 effects when executed. In particular, the code must not allocate
5284 memory or get resources without deallocating this memory or putting
5285 those resources afterwards.
5288 [[instrumenting-linux-kernel-tracing]]
5289 ==== Load and unload a custom probe kernel module
5291 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5292 kernel module>> in the kernel before it can emit LTTng events.
5294 To load the default probe kernel modules and a custom probe kernel
5297 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5298 probe modules to load when starting a root <<lttng-sessiond,session
5302 .Load the `my_subsys`, `usb`, and the default probe modules.
5306 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5311 You only need to pass the subsystem name, not the whole kernel module
5314 To load _only_ a given custom probe kernel module:
5316 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5317 modules to load when starting a root session daemon:
5320 .Load only the `my_subsys` and `usb` probe modules.
5324 # lttng-sessiond --kmod-probes=my_subsys,usb
5329 To confirm that a probe module is loaded:
5336 $ lsmod | grep lttng_probe_usb
5340 To unload the loaded probe modules:
5342 * Kill the session daemon with `SIGTERM`:
5347 # pkill lttng-sessiond
5351 You can also use man:modprobe(8)'s `--remove` option if the session
5352 daemon terminates abnormally.
5355 [[controlling-tracing]]
5358 Once an application or a Linux kernel is
5359 <<instrumenting,instrumented>> for LTTng tracing,
5362 This section is divided in topics on how to use the various
5363 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5364 command-line tool>>, to _control_ the LTTng daemons and tracers.
5366 NOTE: In the following subsections, we refer to an man:lttng(1) command
5367 using its man page name. For example, instead of _Run the `create`
5368 command to..._, we use _Run the man:lttng-create(1) command to..._.
5372 === Start a session daemon
5374 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5375 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5378 You will see the following error when you run a command while no session
5382 Error: No session daemon is available
5385 The only command that automatically runs a session daemon is
5386 man:lttng-create(1), which you use to
5387 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5388 this is most of the time the first operation that you do, sometimes it's
5389 not. Some examples are:
5391 * <<list-instrumentation-points,List the available instrumentation points>>.
5392 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5394 [[tracing-group]] Each Unix user must have its own running session
5395 daemon to trace user applications. The session daemon that the root user
5396 starts is the only one allowed to control the LTTng kernel tracer. Users
5397 that are part of the _tracing group_ can control the root session
5398 daemon. The default tracing group name is `tracing`; you can set it to
5399 something else with the opt:lttng-sessiond(8):--group option when you
5400 start the root session daemon.
5402 To start a user session daemon:
5404 * Run man:lttng-sessiond(8):
5409 $ lttng-sessiond --daemonize
5413 To start the root session daemon:
5415 * Run man:lttng-sessiond(8) as the root user:
5420 # lttng-sessiond --daemonize
5424 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5425 start the session daemon in foreground.
5427 To stop a session daemon, use man:kill(1) on its process ID (standard
5430 Note that some Linux distributions could manage the LTTng session daemon
5431 as a service. In this case, you should use the service manager to
5432 start, restart, and stop session daemons.
5435 [[creating-destroying-tracing-sessions]]
5436 === Create and destroy a tracing session
5438 Almost all the LTTng control operations happen in the scope of
5439 a <<tracing-session,tracing session>>, which is the dialogue between the
5440 <<lttng-sessiond,session daemon>> and you.
5442 To create a tracing session with a generated name:
5444 * Use the man:lttng-create(1) command:
5453 The created tracing session's name is `auto` followed by the
5456 To create a tracing session with a specific name:
5458 * Use the optional argument of the man:lttng-create(1) command:
5463 $ lttng create my-session
5467 Replace `my-session` with the specific tracing session name.
5469 LTTng appends the creation date to the created tracing session's name.
5471 LTTng writes the traces of a tracing session in
5472 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5473 name of the tracing session. Note that the env:LTTNG_HOME environment
5474 variable defaults to `$HOME` if not set.
5476 To output LTTng traces to a non-default location:
5478 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5483 $ lttng create my-session --output=/tmp/some-directory
5487 You may create as many tracing sessions as you wish.
5489 To list all the existing tracing sessions for your Unix user:
5491 * Use the man:lttng-list(1) command:
5500 When you create a tracing session, it is set as the _current tracing
5501 session_. The following man:lttng(1) commands operate on the current
5502 tracing session when you don't specify one:
5504 [role="list-3-cols"]
5505 * man:lttng-add-context(1)
5506 * man:lttng-destroy(1)
5507 * man:lttng-disable-channel(1)
5508 * man:lttng-disable-event(1)
5509 * man:lttng-disable-rotation(1)
5510 * man:lttng-enable-channel(1)
5511 * man:lttng-enable-event(1)
5512 * man:lttng-enable-rotation(1)
5514 * man:lttng-regenerate(1)
5515 * man:lttng-rotate(1)
5517 * man:lttng-snapshot(1)
5518 * man:lttng-start(1)
5519 * man:lttng-status(1)
5521 * man:lttng-track(1)
5522 * man:lttng-untrack(1)
5525 To change the current tracing session:
5527 * Use the man:lttng-set-session(1) command:
5532 $ lttng set-session new-session
5536 Replace `new-session` by the name of the new current tracing session.
5538 When you are done tracing in a given tracing session, you can destroy
5539 it. This operation frees the resources taken by the tracing session
5540 to destroy; it does not destroy the trace data that LTTng wrote for
5541 this tracing session.
5543 To destroy the current tracing session:
5545 * Use the man:lttng-destroy(1) command:
5555 [[list-instrumentation-points]]
5556 === List the available instrumentation points
5558 The <<lttng-sessiond,session daemon>> can query the running instrumented
5559 user applications and the Linux kernel to get a list of available
5560 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5561 they are tracepoints and system calls. For the user space tracing
5562 domain, they are tracepoints. For the other tracing domains, they are
5565 To list the available instrumentation points:
5567 * Use the man:lttng-list(1) command with the requested tracing domain's
5571 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5572 must be a root user, or it must be a member of the
5573 <<tracing-group,tracing group>>).
5574 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5575 kernel system calls (your Unix user must be a root user, or it must be
5576 a member of the tracing group).
5577 * opt:lttng-list(1):--userspace: user space tracepoints.
5578 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5579 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5580 * opt:lttng-list(1):--python: Python loggers.
5583 .List the available user space tracepoints.
5587 $ lttng list --userspace
5591 .List the available Linux kernel system call tracepoints.
5595 $ lttng list --kernel --syscall
5600 [[enabling-disabling-events]]
5601 === Create and enable an event rule
5603 Once you <<creating-destroying-tracing-sessions,create a tracing
5604 session>>, you can create <<event,event rules>> with the
5605 man:lttng-enable-event(1) command.
5607 You specify each condition with a command-line option. The available
5608 condition arguments are shown in the following table.
5610 [role="growable",cols="asciidoc,asciidoc,default"]
5611 .Condition command-line arguments for the man:lttng-enable-event(1) command.
5613 |Argument |Description |Applicable tracing domains
5619 . +--probe=__ADDR__+
5620 . +--function=__ADDR__+
5621 . +--userspace-probe=__PATH__:__SYMBOL__+
5622 . +--userspace-probe=sdt:__PATH__:__PROVIDER__:__NAME__+
5625 Instead of using the default _tracepoint_ instrumentation type, use:
5627 . A Linux system call (entry and exit).
5628 . A Linux https://lwn.net/Articles/132196/[kprobe] (symbol or address).
5629 . The entry and return points of a Linux function (symbol or address).
5630 . The entry point of a user application or library function (path to
5631 application/library and symbol).
5632 . A https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
5633 Statically Defined Tracing] (SDT) probe (path to application/library,
5634 provider and probe names).
5638 |First positional argument.
5641 Tracepoint or system call name.
5643 With the opt:lttng-enable-event(1):--probe,
5644 opt:lttng-enable-event(1):--function, and
5645 opt:lttng-enable-event(1):--userspace-probe options, this is a custom
5646 name given to the event rule. With the JUL, log4j, and Python domains,
5647 this is a logger name.
5649 With a tracepoint, logger, or system call name, you can use the special
5650 `*` globbing character to match anything (for example, `sched_*`,
5658 . +--loglevel=__LEVEL__+
5659 . +--loglevel-only=__LEVEL__+
5662 . Match only tracepoints or log statements with a logging level at
5663 least as severe as +__LEVEL__+.
5664 . Match only tracepoints or log statements with a logging level
5665 equal to +__LEVEL__+.
5667 See man:lttng-enable-event(1) for the list of available logging level
5670 |User space, JUL, log4j, and Python.
5672 |+--exclude=__EXCLUSIONS__+
5675 When you use a `*` character at the end of the tracepoint or logger
5676 name (first positional argument), exclude the specific names in the
5677 comma-delimited list +__EXCLUSIONS__+.
5680 User space, JUL, log4j, and Python.
5682 |+--filter=__EXPR__+
5685 Match only events which satisfy the expression +__EXPR__+.
5687 See man:lttng-enable-event(1) to learn more about the syntax of a
5694 You attach an event rule to a <<channel,channel>> on creation. If you do
5695 not specify the channel with the opt:lttng-enable-event(1):--channel
5696 option, and if the event rule to create is the first in its
5697 <<domain,tracing domain>> for a given tracing session, then LTTng
5698 creates a _default channel_ for you. This default channel is reused in
5699 subsequent invocations of the man:lttng-enable-event(1) command for the
5700 same tracing domain.
5702 An event rule is always enabled at creation time.
5704 The following examples show how you can combine the previous
5705 command-line options to create simple to more complex event rules.
5707 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5711 $ lttng enable-event --kernel sched_switch
5715 .Create an event rule matching four Linux kernel system calls (default channel).
5719 $ lttng enable-event --kernel --syscall open,write,read,close
5723 .Create event rules matching tracepoints with filter expressions (default channel).
5727 $ lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5732 $ lttng enable-event --kernel --all \
5733 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5738 $ lttng enable-event --jul my_logger \
5739 --filter='$app.retriever:cur_msg_id > 3'
5742 IMPORTANT: Make sure to always quote the filter string when you
5743 use man:lttng(1) from a shell.
5746 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5750 $ lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5753 IMPORTANT: Make sure to always quote the wildcard character when you
5754 use man:lttng(1) from a shell.
5757 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5761 $ lttng enable-event --python my-app.'*' \
5762 --exclude='my-app.module,my-app.hello'
5766 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5770 $ lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5774 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5778 $ lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5782 .Create an event rule matching the `malloc` function entry in path:{/usr/lib/libc.so.6}:
5786 $ lttng enable-event --kernel --userspace-probe=/usr/lib/libc.so.6:malloc \
5791 .Create an event rule matching the `server`/`accept_request` https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SDT probe] in path:{/usr/bin/serv}:
5795 $ lttng enable-event --kernel --userspace-probe=sdt:serv:server:accept_request \
5796 server_accept_request
5800 The event rules of a given channel form a whitelist: as soon as an
5801 emitted event passes one of them, LTTng can record the event. For
5802 example, an event named `my_app:my_tracepoint` emitted from a user space
5803 tracepoint with a `TRACE_ERROR` log level passes both of the following
5808 $ lttng enable-event --userspace my_app:my_tracepoint
5809 $ lttng enable-event --userspace my_app:my_tracepoint \
5810 --loglevel=TRACE_INFO
5813 The second event rule is redundant: the first one includes
5817 [[disable-event-rule]]
5818 === Disable an event rule
5820 To disable an event rule that you <<enabling-disabling-events,created>>
5821 previously, use the man:lttng-disable-event(1) command. This command
5822 disables _all_ the event rules (of a given tracing domain and channel)
5823 which match an instrumentation point. The other conditions are not
5824 supported as of LTTng{nbsp}{revision}.
5826 The LTTng tracer does not record an emitted event which passes
5827 a _disabled_ event rule.
5829 .Disable an event rule matching a Python logger (default channel).
5833 $ lttng disable-event --python my-logger
5837 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5841 $ lttng disable-event --jul '*'
5845 .Disable _all_ the event rules of the default channel.
5847 The opt:lttng-disable-event(1):--all-events option is not, like the
5848 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5849 equivalent of the event name `*` (wildcard): it disables _all_ the event
5850 rules of a given channel.
5854 $ lttng disable-event --jul --all-events
5858 NOTE: You cannot delete an event rule once you create it.
5862 === Get the status of a tracing session
5864 To get the status of the current tracing session, that is, its
5865 parameters, its channels, event rules, and their attributes:
5867 * Use the man:lttng-status(1) command:
5877 To get the status of any tracing session:
5879 * Use the man:lttng-list(1) command with the tracing session's name:
5884 $ lttng list my-session
5888 Replace `my-session` with the desired tracing session's name.
5891 [[basic-tracing-session-control]]
5892 === Start and stop a tracing session
5894 Once you <<creating-destroying-tracing-sessions,create a tracing
5896 <<enabling-disabling-events,create one or more event rules>>,
5897 you can start and stop the tracers for this tracing session.
5899 To start tracing in the current tracing session:
5901 * Use the man:lttng-start(1) command:
5910 LTTng is very flexible: you can launch user applications before
5911 or after the you start the tracers. The tracers only record the events
5912 if they pass enabled event rules and if they occur while the tracers are
5915 To stop tracing in the current tracing session:
5917 * Use the man:lttng-stop(1) command:
5926 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
5927 records>> or lost sub-buffers since the last time you ran
5928 man:lttng-start(1), warnings are printed when you run the
5929 man:lttng-stop(1) command.
5932 [[enabling-disabling-channels]]
5933 === Create a channel
5935 Once you create a tracing session, you can create a <<channel,channel>>
5936 with the man:lttng-enable-channel(1) command.
5938 Note that LTTng automatically creates a default channel when, for a
5939 given <<domain,tracing domain>>, no channels exist and you
5940 <<enabling-disabling-events,create>> the first event rule. This default
5941 channel is named `channel0` and its attributes are set to reasonable
5942 values. Therefore, you only need to create a channel when you need
5943 non-default attributes.
5945 You specify each non-default channel attribute with a command-line
5946 option when you use the man:lttng-enable-channel(1) command. The
5947 available command-line options are:
5949 [role="growable",cols="asciidoc,asciidoc"]
5950 .Command-line options for the man:lttng-enable-channel(1) command.
5952 |Option |Description
5958 <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> instead
5959 of the default _discard_ mode.
5961 |`--buffers-pid` (user space tracing domain only)
5964 Use the per-process <<channel-buffering-schemes,buffering scheme>>
5965 instead of the default per-user buffering scheme.
5967 |+--subbuf-size=__SIZE__+
5970 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
5971 either for each Unix user (default), or for each instrumented process.
5973 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5975 |+--num-subbuf=__COUNT__+
5978 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
5979 for each Unix user (default), or for each instrumented process.
5981 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5983 |+--tracefile-size=__SIZE__+
5986 Set the maximum size of each trace file that this channel writes within
5987 a stream to +__SIZE__+ bytes instead of no maximum.
5989 See <<tracefile-rotation,Trace file count and size>>.
5991 |+--tracefile-count=__COUNT__+
5994 Limit the number of trace files that this channel creates to
5995 +__COUNT__+ channels instead of no limit.
5997 See <<tracefile-rotation,Trace file count and size>>.
5999 |+--switch-timer=__PERIODUS__+
6002 Set the <<channel-switch-timer,switch timer period>>
6003 to +__PERIODUS__+{nbsp}µs.
6005 |+--read-timer=__PERIODUS__+
6008 Set the <<channel-read-timer,read timer period>>
6009 to +__PERIODUS__+{nbsp}µs.
6011 |[[opt-blocking-timeout]]+--blocking-timeout=__TIMEOUTUS__+
6014 Set the timeout of user space applications which load LTTng-UST
6015 in blocking mode to +__TIMEOUTUS__+:
6018 Never block (non-blocking mode).
6021 Block forever until space is available in a sub-buffer to record
6024 __n__, a positive value::
6025 Wait for at most __n__ µs when trying to write into a sub-buffer.
6027 Note that, for this option to have any effect on an instrumented
6028 user space application, you need to run the application with a set
6029 env:LTTNG_UST_ALLOW_BLOCKING environment variable.
6031 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6034 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
6038 You can only create a channel in the Linux kernel and user space
6039 <<domain,tracing domains>>: other tracing domains have their own channel
6040 created on the fly when <<enabling-disabling-events,creating event
6045 Because of a current LTTng limitation, you must create all channels
6046 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6047 tracing session, that is, before the first time you run
6050 Since LTTng automatically creates a default channel when you use the
6051 man:lttng-enable-event(1) command with a specific tracing domain, you
6052 cannot, for example, create a Linux kernel event rule, start tracing,
6053 and then create a user space event rule, because no user space channel
6054 exists yet and it's too late to create one.
6056 For this reason, make sure to configure your channels properly
6057 before starting the tracers for the first time!
6060 The following examples show how you can combine the previous
6061 command-line options to create simple to more complex channels.
6063 .Create a Linux kernel channel with default attributes.
6067 $ lttng enable-channel --kernel my-channel
6071 .Create a user space channel with four sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6075 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6076 --buffers-pid my-channel
6080 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6082 <<creating-destroying-tracing-sessions,Create a tracing-session>>,
6083 create the channel, <<enabling-disabling-events,create an event rule>>,
6084 and <<basic-tracing-session-control,start tracing>>:
6089 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
6090 $ lttng enable-event --userspace --channel=blocking-channel --all
6094 Run an application instrumented with LTTng-UST and allow it to block:
6098 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6102 .Create a Linux kernel channel which rotates eight trace files of 4{nbsp}MiB each for each stream
6106 $ lttng enable-channel --kernel --tracefile-count=8 \
6107 --tracefile-size=4194304 my-channel
6111 .Create a user space channel in overwrite (or _flight recorder_) mode.
6115 $ lttng enable-channel --userspace --overwrite my-channel
6119 You can <<enabling-disabling-events,create>> the same event rule in
6120 two different channels:
6124 $ lttng enable-event --userspace --channel=my-channel app:tp
6125 $ lttng enable-event --userspace --channel=other-channel app:tp
6128 If both channels are enabled, when a tracepoint named `app:tp` is
6129 reached, LTTng records two events, one for each channel.
6133 === Disable a channel
6135 To disable a specific channel that you <<enabling-disabling-channels,created>>
6136 previously, use the man:lttng-disable-channel(1) command.
6138 .Disable a specific Linux kernel channel.
6142 $ lttng disable-channel --kernel my-channel
6146 The state of a channel precedes the individual states of event rules
6147 attached to it: event rules which belong to a disabled channel, even if
6148 they are enabled, are also considered disabled.
6152 === Add context fields to a channel
6154 Event record fields in trace files provide important information about
6155 events that occured previously, but sometimes some external context may
6156 help you solve a problem faster. Examples of context fields are:
6158 * The **process ID**, **thread ID**, **process name**, and
6159 **process priority** of the thread in which the event occurs.
6160 * The **hostname** of the system on which the event occurs.
6161 * The Linux kernel and user call stacks (since
6162 LTTng{nbsp}{revision}).
6163 * The current values of many possible **performance counters** using
6165 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6167 ** Branch instructions, misses, and loads.
6169 * Any context defined at the application level (supported for the
6170 JUL and log4j <<domain,tracing domains>>).
6172 To get the full list of available context fields, see
6173 `lttng add-context --list`. Some context fields are reserved for a
6174 specific <<domain,tracing domain>> (Linux kernel or user space).
6176 You add context fields to <<channel,channels>>. All the events
6177 that a channel with added context fields records contain those fields.
6179 To add context fields to one or all the channels of a given tracing
6182 * Use the man:lttng-add-context(1) command.
6184 .Add context fields to all the channels of the current tracing session.
6186 The following command line adds the virtual process identifier and
6187 the per-thread CPU cycles count fields to all the user space channels
6188 of the current tracing session.
6192 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6196 .Add performance counter context fields by raw ID
6198 See man:lttng-add-context(1) for the exact format of the context field
6199 type, which is partly compatible with the format used in
6204 $ lttng add-context --userspace --type=perf:thread:raw:r0110:test
6205 $ lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6209 .Add context fields to a specific channel.
6211 The following command line adds the thread identifier and user call
6212 stack context fields to the Linux kernel channel named `my-channel` in
6213 the current tracing session.
6217 $ lttng add-context --kernel --channel=my-channel \
6218 --type=tid --type=callstack-user
6222 .Add an application-specific context field to a specific channel.
6224 The following command line adds the `cur_msg_id` context field of the
6225 `retriever` context retriever for all the instrumented
6226 <<java-application,Java applications>> recording <<event,event records>>
6227 in the channel named `my-channel`:
6231 $ lttng add-context --kernel --channel=my-channel \
6232 --type='$app:retriever:cur_msg_id'
6235 IMPORTANT: Make sure to always quote the `$` character when you
6236 use man:lttng-add-context(1) from a shell.
6239 NOTE: You cannot remove context fields from a channel once you add it.
6244 === Track process IDs
6246 It's often useful to allow only specific process IDs (PIDs) to emit
6247 events. For example, you may wish to record all the system calls made by
6248 a given process (Ă la http://linux.die.net/man/1/strace[strace]).
6250 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6251 purpose. Both commands operate on a whitelist of process IDs. You _add_
6252 entries to this whitelist with the man:lttng-track(1) command and remove
6253 entries with the man:lttng-untrack(1) command. Any process which has one
6254 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6255 an enabled <<event,event rule>>.
6257 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6258 process with a given tracked ID exit and another process be given this
6259 ID, then the latter would also be allowed to emit events.
6261 .Track and untrack process IDs.
6263 For the sake of the following example, assume the target system has
6264 16{nbsp}possible PIDs.
6267 <<creating-destroying-tracing-sessions,create a tracing session>>,
6268 the whitelist contains all the possible PIDs:
6271 .All PIDs are tracked.
6272 image::track-all.png[]
6274 When the whitelist is full and you use the man:lttng-track(1) command to
6275 specify some PIDs to track, LTTng first clears the whitelist, then it
6276 tracks the specific PIDs. After:
6280 $ lttng track --pid=3,4,7,10,13
6286 .PIDs 3, 4, 7, 10, and 13 are tracked.
6287 image::track-3-4-7-10-13.png[]
6289 You can add more PIDs to the whitelist afterwards:
6293 $ lttng track --pid=1,15,16
6299 .PIDs 1, 15, and 16 are added to the whitelist.
6300 image::track-1-3-4-7-10-13-15-16.png[]
6302 The man:lttng-untrack(1) command removes entries from the PID tracker's
6303 whitelist. Given the previous example, the following command:
6307 $ lttng untrack --pid=3,7,10,13
6310 leads to this whitelist:
6313 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6314 image::track-1-4-15-16.png[]
6316 LTTng can track all possible PIDs again using the
6317 opt:lttng-track(1):--all option:
6321 $ lttng track --pid --all
6324 The result is, again:
6327 .All PIDs are tracked.
6328 image::track-all.png[]
6331 .Track only specific PIDs
6333 A very typical use case with PID tracking is to start with an empty
6334 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6335 then add PIDs manually while tracers are active. You can accomplish this
6336 by using the opt:lttng-untrack(1):--all option of the
6337 man:lttng-untrack(1) command to clear the whitelist after you
6338 <<creating-destroying-tracing-sessions,create a tracing session>>:
6342 $ lttng untrack --pid --all
6348 .No PIDs are tracked.
6349 image::untrack-all.png[]
6351 If you trace with this whitelist configuration, the tracer records no
6352 events for this <<domain,tracing domain>> because no processes are
6353 tracked. You can use the man:lttng-track(1) command as usual to track
6354 specific PIDs, for example:
6358 $ lttng track --pid=6,11
6364 .PIDs 6 and 11 are tracked.
6365 image::track-6-11.png[]
6370 [[saving-loading-tracing-session]]
6371 === Save and load tracing session configurations
6373 Configuring a <<tracing-session,tracing session>> can be long. Some of
6374 the tasks involved are:
6376 * <<enabling-disabling-channels,Create channels>> with
6377 specific attributes.
6378 * <<adding-context,Add context fields>> to specific channels.
6379 * <<enabling-disabling-events,Create event rules>> with specific log
6380 level and filter conditions.
6382 If you use LTTng to solve real world problems, chances are you have to
6383 record events using the same tracing session setup over and over,
6384 modifying a few variables each time in your instrumented program
6385 or environment. To avoid constant tracing session reconfiguration,
6386 the man:lttng(1) command-line tool can save and load tracing session
6387 configurations to/from XML files.
6389 To save a given tracing session configuration:
6391 * Use the man:lttng-save(1) command:
6396 $ lttng save my-session
6400 Replace `my-session` with the name of the tracing session to save.
6402 LTTng saves tracing session configurations to
6403 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6404 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6405 the opt:lttng-save(1):--output-path option to change this destination
6408 LTTng saves all configuration parameters, for example:
6410 * The tracing session name.
6411 * The trace data output path.
6412 * The channels with their state and all their attributes.
6413 * The context fields you added to channels.
6414 * The event rules with their state, log level and filter conditions.
6416 To load a tracing session:
6418 * Use the man:lttng-load(1) command:
6423 $ lttng load my-session
6427 Replace `my-session` with the name of the tracing session to load.
6429 When LTTng loads a configuration, it restores your saved tracing session
6430 as if you just configured it manually.
6432 See man:lttng(1) for the complete list of command-line options. You
6433 can also save and load all many sessions at a time, and decide in which
6434 directory to output the XML files.
6437 [[sending-trace-data-over-the-network]]
6438 === Send trace data over the network
6440 LTTng can send the recorded trace data to a remote system over the
6441 network instead of writing it to the local file system.
6443 To send the trace data over the network:
6445 . On the _remote_ system (which can also be the target system),
6446 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6455 . On the _target_ system, create a tracing session configured to
6456 send trace data over the network:
6461 $ lttng create my-session --set-url=net://remote-system
6465 Replace `remote-system` by the host name or IP address of the
6466 remote system. See man:lttng-create(1) for the exact URL format.
6468 . On the target system, use the man:lttng(1) command-line tool as usual.
6469 When tracing is active, the target's consumer daemon sends sub-buffers
6470 to the relay daemon running on the remote system instead of flushing
6471 them to the local file system. The relay daemon writes the received
6472 packets to the local file system.
6474 The relay daemon writes trace files to
6475 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6476 +__hostname__+ is the host name of the target system and +__session__+
6477 is the tracing session name. Note that the env:LTTNG_HOME environment
6478 variable defaults to `$HOME` if not set. Use the
6479 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6480 trace files to another base directory.
6485 === View events as LTTng emits them (noch:{LTTng} live)
6487 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6488 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6489 display events as LTTng emits them on the target system while tracing is
6492 The relay daemon creates a _tee_: it forwards the trace data to both
6493 the local file system and to connected live viewers:
6496 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6501 . On the _target system_, create a <<tracing-session,tracing session>>
6507 $ lttng create my-session --live
6511 This spawns a local relay daemon.
6513 . Start the live viewer and configure it to connect to the relay
6514 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6519 $ babeltrace --input-format=lttng-live \
6520 net://localhost/host/hostname/my-session
6527 * `hostname` with the host name of the target system.
6528 * `my-session` with the name of the tracing session to view.
6531 . Configure the tracing session as usual with the man:lttng(1)
6532 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6534 You can list the available live tracing sessions with Babeltrace:
6538 $ babeltrace --input-format=lttng-live net://localhost
6541 You can start the relay daemon on another system. In this case, you need
6542 to specify the relay daemon's URL when you create the tracing session
6543 with the opt:lttng-create(1):--set-url option. You also need to replace
6544 `localhost` in the procedure above with the host name of the system on
6545 which the relay daemon is running.
6547 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6548 command-line options.
6552 [[taking-a-snapshot]]
6553 === Take a snapshot of the current sub-buffers of a tracing session
6555 The normal behavior of LTTng is to append full sub-buffers to growing
6556 trace data files. This is ideal to keep a full history of the events
6557 that occurred on the target system, but it can
6558 represent too much data in some situations. For example, you may wish
6559 to trace your application continuously until some critical situation
6560 happens, in which case you only need the latest few recorded
6561 events to perform the desired analysis, not multi-gigabyte trace files.
6563 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6564 current sub-buffers of a given <<tracing-session,tracing session>>.
6565 LTTng can write the snapshot to the local file system or send it over
6569 .A snapshot is a copy of the current sub-buffers, which are not cleared after the operation.
6570 image::snapshot.png[]
6572 If you wish to create unmanaged, self-contained, non-overlapping
6573 trace chunk archives instead of a simple copy of the current
6574 sub-buffers, see the <<session-rotation,tracing session rotation>>
6575 feature (available since LTTng{nbsp}2.11).
6579 . Create a tracing session in _snapshot mode_:
6584 $ lttng create my-session --snapshot
6588 The <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> of
6589 <<channel,channels>> created in this mode is automatically set to
6590 _overwrite_ (flight recorder mode).
6592 . Configure the tracing session as usual with the man:lttng(1)
6593 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6595 . **Optional**: When you need to take a snapshot,
6596 <<basic-tracing-session-control,stop tracing>>.
6598 You can take a snapshot when the tracers are active, but if you stop
6599 them first, you are sure that the data in the sub-buffers does not
6600 change before you actually take the snapshot.
6607 $ lttng snapshot record --name=my-first-snapshot
6611 LTTng writes the current sub-buffers of all the current tracing
6612 session's channels to trace files on the local file system. Those trace
6613 files have `my-first-snapshot` in their name.
6615 There is no difference between the format of a normal trace file and the
6616 format of a snapshot: viewers of LTTng traces also support LTTng
6619 By default, LTTng writes snapshot files to the path shown by
6620 `lttng snapshot list-output`. You can change this path or decide to send
6621 snapshots over the network using either:
6623 . An output path or URL that you specify when you
6624 <<creating-destroying-tracing-sessions,create the tracing session>>.
6625 . A snapshot output path or URL that you add using
6626 `lttng snapshot add-output`.
6627 . An output path or URL that you provide directly to the
6628 `lttng snapshot record` command.
6630 Method{nbsp}3 overrides method{nbsp}2, which overrides method 1. When
6631 you specify a URL, a relay daemon must listen on a remote system (see
6632 <<sending-trace-data-over-the-network,Send trace data over the
6637 [[session-rotation]]
6638 === Archive the current trace chunk (rotate a tracing session)
6640 The <<taking-a-snapshot,snapshot user guide>> shows how you can dump
6641 a tracing session's current sub-buffers to the file system or send them
6642 over the network. When you take a snapshot, LTTng does not clear the
6643 tracing session's ring buffers: if you take another snapshot immediately
6644 after, both snapshots could contain overlapping trace data.
6646 Inspired by https://en.wikipedia.org/wiki/Log_rotation[log rotation],
6647 _tracing session rotation_ is a feature which appends the content of the
6648 ring buffers to what's already on the file system or sent over the
6649 network since the tracing session's creation or since the last
6650 rotation, and then clears those ring buffers to avoid trace data
6653 What LTTng is about to write when performing a tracing session rotation
6654 is called the _current trace chunk_. When this current trace chunk is
6655 written to the file system or sent over the network, it is called a
6656 _trace chunk archive_. Therefore, a tracing session rotation _archives_
6657 the current trace chunk.
6660 .A tracing session rotation operation _archives_ the current trace chunk.
6661 image::rotation.png[]
6663 A trace chunk archive is a self-contained LTTng trace and is not managed
6664 anymore by LTTng: you can read it, modify it, move it, or remove it.
6666 There are two methods to perform a tracing session rotation:
6667 immediately or automatically.
6669 To perform an immediate tracing session rotation:
6671 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6672 in _normal mode_ or _network streaming mode_
6673 (only those two creation modes support tracing session rotation):
6678 $ lttng create my-session
6682 . <<enabling-disabling-events,Create one or more event rules>>
6683 and <<basic-tracing-session-control,start tracing>>:
6688 $ lttng enable-event --kernel sched_'*'
6693 . When needed, immediately rotate the current tracing session:
6702 The cmd:lttng-rotate command prints the path to the created trace
6703 chunk archive. See man:lttng-rotate(1) to learn about the format
6704 of trace chunk archive directory names.
6706 You can perform other immediate rotations while the tracing session is
6707 active. It is guaranteed that all the trace chunk archives do not
6708 contain overlapping trace data. You can also perform an immediate
6709 rotation once the tracing session is
6710 <<basic-tracing-session-control,stopped>>.
6712 . When you are done tracing,
6713 <<creating-destroying-tracing-sessions,destroy the current tracing
6723 The tracing session destruction operation creates one last trace
6724 chunk archive from the current trace chunk.
6726 An automatic tracing session rotation is a rotation which LTTng
6727 performs automatically based on one of the following conditions:
6729 * A timer with a configured period times out.
6730 * The total size of the flushed part of the current trace chunk
6731 becomes greater than or equal to a configured value.
6733 To configure a future, automatic tracing session rotation, you need
6734 to set a _rotation schedule_.
6736 To set a rotation schedule:
6738 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6739 in _normal mode_ or _network streaming mode_
6740 (only those two creation modes support tracing session rotation):
6745 $ lttng create my-session
6749 . <<enabling-disabling-events,Create one or more event rules>>:
6754 $ lttng enable-event --kernel sched_'*'
6758 . Set a tracing session rotation schedule:
6763 $ lttng enable-rotation --timer=12s
6767 In this example, we set a rotation schedule so that LTTng performs a
6768 tracing session rotation every 12{nbsp}seconds.
6770 See man:lttng-enable-rotation(1) to learn more about other ways to set a
6773 . <<basic-tracing-session-control,Start tracing>>:
6782 LTTng performs tracing session rotations automatically while the tracing
6783 session is active thanks to the rotation schedule.
6785 . When you are done tracing,
6786 <<creating-destroying-tracing-sessions,destroy the current tracing
6796 The tracing session destruction operation creates one last trace chunk
6797 archive from the current trace chunk.
6799 You can use man:lttng-disable-rotation(1) to unset an a tracing session
6802 NOTE: man:lttng-rotate(1) and man:lttng-enable-rotation(1) list
6803 limitations regarding those two commands.
6808 === Use the machine interface
6810 With any command of the man:lttng(1) command-line tool, you can set the
6811 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6812 XML machine interface output, for example:
6816 $ lttng --mi=xml enable-event --kernel --syscall open
6819 A schema definition (XSD) is
6820 https://github.com/lttng/lttng-tools/blob/stable-2.11/src/common/mi-lttng-3.0.xsd[available]
6821 to ease the integration with external tools as much as possible.
6825 [[metadata-regenerate]]
6826 === Regenerate the metadata of an LTTng trace
6828 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6829 data stream files and a metadata file. This metadata file contains,
6830 amongst other things, information about the offset of the clock sources
6831 used to timestamp <<event,event records>> when tracing.
6833 If, once a <<tracing-session,tracing session>> is
6834 <<basic-tracing-session-control,started>>, a major
6835 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6836 happens, the trace's clock offset also needs to be updated. You
6837 can use the `metadata` item of the man:lttng-regenerate(1) command
6840 The main use case of this command is to allow a system to boot with
6841 an incorrect wall time and trace it with LTTng before its wall time
6842 is corrected. Once the system is known to be in a state where its
6843 wall time is correct, it can run `lttng regenerate metadata`.
6845 To regenerate the metadata of an LTTng trace:
6847 * Use the `metadata` item of the man:lttng-regenerate(1) command:
6852 $ lttng regenerate metadata
6858 `lttng regenerate metadata` has the following limitations:
6860 * Tracing session <<creating-destroying-tracing-sessions,created>>
6862 * User space <<channel,channels>>, if any, are using
6863 <<channel-buffering-schemes,per-user buffering>>.
6868 [[regenerate-statedump]]
6869 === Regenerate the state dump of a tracing session
6871 The LTTng kernel and user space tracers generate state dump
6872 <<event,event records>> when the application starts or when you
6873 <<basic-tracing-session-control,start a tracing session>>. An analysis
6874 can use the state dump event records to set an initial state before it
6875 builds the rest of the state from the following event records.
6876 http://tracecompass.org/[Trace Compass] is a notable example of an
6877 application which uses the state dump of an LTTng trace.
6879 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
6880 state dump event records are not included in the snapshot because they
6881 were recorded to a sub-buffer that has been consumed or overwritten
6884 You can use the `lttng regenerate statedump` command to emit the state
6885 dump event records again.
6887 To regenerate the state dump of the current tracing session, provided
6888 create it in snapshot mode, before you take a snapshot:
6890 . Use the `statedump` item of the man:lttng-regenerate(1) command:
6895 $ lttng regenerate statedump
6899 . <<basic-tracing-session-control,Stop the tracing session>>:
6908 . <<taking-a-snapshot,Take a snapshot>>:
6913 $ lttng snapshot record --name=my-snapshot
6917 Depending on the event throughput, you should run steps 1 and 2
6918 as closely as possible.
6920 NOTE: To record the state dump events, you need to
6921 <<enabling-disabling-events,create event rules>> which enable them.
6922 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
6923 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
6927 [[persistent-memory-file-systems]]
6928 === Record trace data on persistent memory file systems
6930 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6931 (NVRAM) is random-access memory that retains its information when power
6932 is turned off (non-volatile). Systems with such memory can store data
6933 structures in RAM and retrieve them after a reboot, without flushing
6934 to typical _storage_.
6936 Linux supports NVRAM file systems thanks to either
6937 http://pramfs.sourceforge.net/[PRAMFS] or
6938 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6939 (requires Linux{nbsp}4.1+).
6941 This section does not describe how to operate such file systems;
6942 we assume that you have a working persistent memory file system.
6944 When you create a <<tracing-session,tracing session>>, you can specify
6945 the path of the shared memory holding the sub-buffers. If you specify a
6946 location on an NVRAM file system, then you can retrieve the latest
6947 recorded trace data when the system reboots after a crash.
6949 To record trace data on a persistent memory file system and retrieve the
6950 trace data after a system crash:
6952 . Create a tracing session with a sub-buffer shared memory path located
6953 on an NVRAM file system:
6958 $ lttng create my-session --shm-path=/path/to/shm
6962 . Configure the tracing session as usual with the man:lttng(1)
6963 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6965 . After a system crash, use the man:lttng-crash(1) command-line tool to
6966 view the trace data recorded on the NVRAM file system:
6971 $ lttng-crash /path/to/shm
6975 The binary layout of the ring buffer files is not exactly the same as
6976 the trace files layout. This is why you need to use man:lttng-crash(1)
6977 instead of your preferred trace viewer directly.
6979 To convert the ring buffer files to LTTng trace files:
6981 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6986 $ lttng-crash --extract=/path/to/trace /path/to/shm
6992 [[notif-trigger-api]]
6993 === Get notified when a channel's buffer usage is too high or too low
6995 With LTTng's $$C/C++$$ notification and trigger API, your user
6996 application can get notified when the buffer usage of one or more
6997 <<channel,channels>> becomes too low or too high. You can use this API
6998 and enable or disable <<event,event rules>> during tracing to avoid
6999 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>.
7001 .Have a user application get notified when an LTTng channel's buffer usage is too high.
7003 In this example, we create and build an application which gets notified
7004 when the buffer usage of a specific LTTng channel is higher than
7005 75{nbsp}%. We only print that it is the case in the example, but we
7006 could as well use the API of <<liblttng-ctl-lttng,`liblttng-ctl`>> to
7007 disable event rules when this happens.
7009 . Create the application's C source file:
7017 #include <lttng/domain.h>
7018 #include <lttng/action/action.h>
7019 #include <lttng/action/notify.h>
7020 #include <lttng/condition/condition.h>
7021 #include <lttng/condition/buffer-usage.h>
7022 #include <lttng/condition/evaluation.h>
7023 #include <lttng/notification/channel.h>
7024 #include <lttng/notification/notification.h>
7025 #include <lttng/trigger/trigger.h>
7026 #include <lttng/endpoint.h>
7028 int main(int argc, char *argv[])
7030 int exit_status = 0;
7031 struct lttng_notification_channel *notification_channel;
7032 struct lttng_condition *condition;
7033 struct lttng_action *action;
7034 struct lttng_trigger *trigger;
7035 const char *tracing_session_name;
7036 const char *channel_name;
7039 tracing_session_name = argv[1];
7040 channel_name = argv[2];
7043 * Create a notification channel. A notification channel
7044 * connects the user application to the LTTng session daemon.
7045 * This notification channel can be used to listen to various
7046 * types of notifications.
7048 notification_channel = lttng_notification_channel_create(
7049 lttng_session_daemon_notification_endpoint);
7052 * Create a "high buffer usage" condition. In this case, the
7053 * condition is reached when the buffer usage is greater than or
7054 * equal to 75 %. We create the condition for a specific tracing
7055 * session name, channel name, and for the user space tracing
7058 * The "low buffer usage" condition type also exists.
7060 condition = lttng_condition_buffer_usage_high_create();
7061 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
7062 lttng_condition_buffer_usage_set_session_name(
7063 condition, tracing_session_name);
7064 lttng_condition_buffer_usage_set_channel_name(condition,
7066 lttng_condition_buffer_usage_set_domain_type(condition,
7070 * Create an action (get a notification) to take when the
7071 * condition created above is reached.
7073 action = lttng_action_notify_create();
7076 * Create a trigger. A trigger associates a condition to an
7077 * action: the action is executed when the condition is reached.
7079 trigger = lttng_trigger_create(condition, action);
7081 /* Register the trigger to LTTng. */
7082 lttng_register_trigger(trigger);
7085 * Now that we have registered a trigger, a notification will be
7086 * emitted everytime its condition is met. To receive this
7087 * notification, we must subscribe to notifications that match
7088 * the same condition.
7090 lttng_notification_channel_subscribe(notification_channel,
7094 * Notification loop. You can put this in a dedicated thread to
7095 * avoid blocking the main thread.
7098 struct lttng_notification *notification;
7099 enum lttng_notification_channel_status status;
7100 const struct lttng_evaluation *notification_evaluation;
7101 const struct lttng_condition *notification_condition;
7102 double buffer_usage;
7104 /* Receive the next notification. */
7105 status = lttng_notification_channel_get_next_notification(
7106 notification_channel, ¬ification);
7109 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
7111 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
7113 * The session daemon can drop notifications if
7114 * a monitoring application is not consuming the
7115 * notifications fast enough.
7118 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
7120 * The notification channel has been closed by the
7121 * session daemon. This is typically caused by a session
7122 * daemon shutting down.
7126 /* Unhandled conditions or errors. */
7132 * A notification provides, amongst other things:
7134 * * The condition that caused this notification to be
7136 * * The condition evaluation, which provides more
7137 * specific information on the evaluation of the
7140 * The condition evaluation provides the buffer usage
7141 * value at the moment the condition was reached.
7143 notification_condition = lttng_notification_get_condition(
7145 notification_evaluation = lttng_notification_get_evaluation(
7148 /* We're subscribed to only one condition. */
7149 assert(lttng_condition_get_type(notification_condition) ==
7150 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
7153 * Get the exact sampled buffer usage from the
7154 * condition evaluation.
7156 lttng_evaluation_buffer_usage_get_usage_ratio(
7157 notification_evaluation, &buffer_usage);
7160 * At this point, instead of printing a message, we
7161 * could do something to reduce the channel's buffer
7162 * usage, like disable specific events.
7164 printf("Buffer usage is %f %% in tracing session \"%s\", "
7165 "user space channel \"%s\".\n", buffer_usage * 100,
7166 tracing_session_name, channel_name);
7167 lttng_notification_destroy(notification);
7171 lttng_action_destroy(action);
7172 lttng_condition_destroy(condition);
7173 lttng_trigger_destroy(trigger);
7174 lttng_notification_channel_destroy(notification_channel);
7180 . Build the `notif-app` application, linking it to `liblttng-ctl`:
7185 $ gcc -o notif-app notif-app.c -llttng-ctl
7189 . <<creating-destroying-tracing-sessions,Create a tracing session>>,
7190 <<enabling-disabling-events,create an event rule>> matching all the
7191 user space tracepoints, and
7192 <<basic-tracing-session-control,start tracing>>:
7197 $ lttng create my-session
7198 $ lttng enable-event --userspace --all
7203 If you create the channel manually with the man:lttng-enable-channel(1)
7204 command, you can control how frequently are the current values of the
7205 channel's properties sampled to evaluate user conditions with the
7206 opt:lttng-enable-channel(1):--monitor-timer option.
7208 . Run the `notif-app` application. This program accepts the
7209 <<tracing-session,tracing session>> name and the user space channel
7210 name as its two first arguments. The channel which LTTng automatically
7211 creates with the man:lttng-enable-event(1) command above is named
7217 $ ./notif-app my-session channel0
7221 . In another terminal, run an application with a very high event
7222 throughput so that the 75{nbsp}% buffer usage condition is reached.
7224 In the first terminal, the application should print lines like this:
7227 Buffer usage is 81.45197 % in tracing session "my-session", user space
7231 If you don't see anything, try modifying the condition in
7232 path:{notif-app.c} to a lower value (0.1, for example), rebuilding it
7233 (step{nbsp}2) and running it again (step{nbsp}4).
7240 [[lttng-modules-ref]]
7241 === noch:{LTTng-modules}
7245 [[lttng-tracepoint-enum]]
7246 ==== `LTTNG_TRACEPOINT_ENUM()` usage
7248 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
7252 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
7257 * `name` with the name of the enumeration (C identifier, unique
7258 amongst all the defined enumerations).
7259 * `entries` with a list of enumeration entries.
7261 The available enumeration entry macros are:
7263 +ctf_enum_value(__name__, __value__)+::
7264 Entry named +__name__+ mapped to the integral value +__value__+.
7266 +ctf_enum_range(__name__, __begin__, __end__)+::
7267 Entry named +__name__+ mapped to the range of integral values between
7268 +__begin__+ (included) and +__end__+ (included).
7270 +ctf_enum_auto(__name__)+::
7271 Entry named +__name__+ mapped to the integral value following the
7272 last mapping's value.
7274 The last value of a `ctf_enum_value()` entry is its +__value__+
7277 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
7279 If `ctf_enum_auto()` is the first entry in the list, its integral
7282 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
7283 to use a defined enumeration as a tracepoint field.
7285 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
7289 LTTNG_TRACEPOINT_ENUM(
7292 ctf_enum_auto("AUTO: EXPECT 0")
7293 ctf_enum_value("VALUE: 23", 23)
7294 ctf_enum_value("VALUE: 27", 27)
7295 ctf_enum_auto("AUTO: EXPECT 28")
7296 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
7297 ctf_enum_auto("AUTO: EXPECT 304")
7305 [[lttng-modules-tp-fields]]
7306 ==== Tracepoint fields macros (for `TP_FIELDS()`)
7308 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
7309 tracepoint fields, which must be listed within `TP_FIELDS()` in
7310 `LTTNG_TRACEPOINT_EVENT()`, are:
7312 [role="func-desc growable",cols="asciidoc,asciidoc"]
7313 .Available macros to define LTTng-modules tracepoint fields
7315 |Macro |Description and parameters
7318 +ctf_integer(__t__, __n__, __e__)+
7320 +ctf_integer_nowrite(__t__, __n__, __e__)+
7322 +ctf_user_integer(__t__, __n__, __e__)+
7324 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
7326 Standard integer, displayed in base{nbsp}10.
7329 Integer C type (`int`, `long`, `size_t`, ...).
7335 Argument expression.
7338 +ctf_integer_hex(__t__, __n__, __e__)+
7340 +ctf_user_integer_hex(__t__, __n__, __e__)+
7342 Standard integer, displayed in base{nbsp}16.
7351 Argument expression.
7353 |+ctf_integer_oct(__t__, __n__, __e__)+
7355 Standard integer, displayed in base{nbsp}8.
7364 Argument expression.
7367 +ctf_integer_network(__t__, __n__, __e__)+
7369 +ctf_user_integer_network(__t__, __n__, __e__)+
7371 Integer in network byte order (big-endian), displayed in base{nbsp}10.
7380 Argument expression.
7383 +ctf_integer_network_hex(__t__, __n__, __e__)+
7385 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
7387 Integer in network byte order, displayed in base{nbsp}16.
7396 Argument expression.
7399 +ctf_enum(__N__, __t__, __n__, __e__)+
7401 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
7403 +ctf_user_enum(__N__, __t__, __n__, __e__)+
7405 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
7410 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
7413 Integer C type (`int`, `long`, `size_t`, ...).
7419 Argument expression.
7422 +ctf_string(__n__, __e__)+
7424 +ctf_string_nowrite(__n__, __e__)+
7426 +ctf_user_string(__n__, __e__)+
7428 +ctf_user_string_nowrite(__n__, __e__)+
7430 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7436 Argument expression.
7439 +ctf_array(__t__, __n__, __e__, __s__)+
7441 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7443 +ctf_user_array(__t__, __n__, __e__, __s__)+
7445 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7447 Statically-sized array of integers.
7450 Array element C type.
7456 Argument expression.
7462 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7464 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7466 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7468 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7470 Statically-sized array of bits.
7472 The type of +__e__+ must be an integer type. +__s__+ is the number
7473 of elements of such type in +__e__+, not the number of bits.
7476 Array element C type.
7482 Argument expression.
7488 +ctf_array_text(__t__, __n__, __e__, __s__)+
7490 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7492 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7494 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7496 Statically-sized array, printed as text.
7498 The string does not need to be null-terminated.
7501 Array element C type (always `char`).
7507 Argument expression.
7513 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7515 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7517 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7519 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7521 Dynamically-sized array of integers.
7523 The type of +__E__+ must be unsigned.
7526 Array element C type.
7532 Argument expression.
7535 Length expression C type.
7541 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7543 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7545 Dynamically-sized array of integers, displayed in base{nbsp}16.
7547 The type of +__E__+ must be unsigned.
7550 Array element C type.
7556 Argument expression.
7559 Length expression C type.
7564 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7566 Dynamically-sized array of integers in network byte order (big-endian),
7567 displayed in base{nbsp}10.
7569 The type of +__E__+ must be unsigned.
7572 Array element C type.
7578 Argument expression.
7581 Length expression C type.
7587 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7589 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7591 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7593 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7595 Dynamically-sized array of bits.
7597 The type of +__e__+ must be an integer type. +__s__+ is the number
7598 of elements of such type in +__e__+, not the number of bits.
7600 The type of +__E__+ must be unsigned.
7603 Array element C type.
7609 Argument expression.
7612 Length expression C type.
7618 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7620 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7622 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7624 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7626 Dynamically-sized array, displayed as text.
7628 The string does not need to be null-terminated.
7630 The type of +__E__+ must be unsigned.
7632 The behaviour is undefined if +__e__+ is `NULL`.
7635 Sequence element C type (always `char`).
7641 Argument expression.
7644 Length expression C type.
7650 Use the `_user` versions when the argument expression, `e`, is
7651 a user space address. In the cases of `ctf_user_integer*()` and
7652 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7655 The `_nowrite` versions omit themselves from the session trace, but are
7656 otherwise identical. This means the `_nowrite` fields won't be written
7657 in the recorded trace. Their primary purpose is to make some
7658 of the event context available to the
7659 <<enabling-disabling-events,event filters>> without having to
7660 commit the data to sub-buffers.
7666 Terms related to LTTng and to tracing in general:
7669 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7670 the cmd:babeltrace command, libraries, and Python bindings.
7672 [[def-buffering-scheme]]<<channel-buffering-schemes,buffering scheme>>::
7673 A layout of <<def-sub-buffer,sub-buffers>> applied to a given channel.
7675 <<channel,channel>>::
7676 An entity which is responsible for a set of
7677 <<def-ring-buffer,ring buffers>>.
7679 <<event,Event rules>> are always attached to a specific channel.
7682 A source of time for a <<def-tracer,tracer>>.
7684 <<lttng-consumerd,consumer daemon>>::
7685 A process which is responsible for consuming the full
7686 <<def-sub-buffer,sub-buffers>> and write them to a file system or
7687 send them over the network.
7689 [[def-current-trace-chunk]]current trace chunk::
7690 A <<def-trace-chunk,trace chunk>> which includes the current content
7691 of all the <<tracing-session,tracing session>>'s
7692 <<def-sub-buffer,sub-buffers>> and the stream files produced since the
7693 latest event amongst:
7695 * The creation of the tracing session.
7696 * The last <<session-rotation,tracing session rotation>>, if any.
7698 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event
7699 record loss mode in which the <<def-tracer,tracer>> _discards_ new
7700 event records when there's no
7701 <<def-sub-buffer,sub-buffer>> space left to store them.
7703 [[def-event]]event::
7704 The consequence of the execution of an
7705 <<def-instrumentation-point,instrumentation point>>, like a
7706 <<def-tracepoint,tracepoint>> that you manually place in some source
7707 code, or a Linux kernel kprobe.
7709 An event is said to _occur_ at a specific time. Different actions can
7710 be taken upon the occurrence of an event, like record the event's payload
7711 to a <<def-sub-buffer,sub-buffer>>.
7713 [[def-event-name]]event name::
7714 The name of an event, which is also the name of the event record.
7715 This is also called the _instrumentation point name_.
7717 [[def-event-record]]event record::
7718 A record, in a <<def-trace,trace>>, of the payload of an event
7721 [[def-event-record-loss-mode]]<<channel-overwrite-mode-vs-discard-mode,event record loss mode>>::
7722 The mechanism by which event records of a given <<channel,channel>>
7723 are lost (not recorded) when there is no <<def-sub-buffer,sub-buffer>>
7724 space left to store them.
7726 <<event,event rule>>::
7727 Set of conditions which must be satisfied for one or more occuring
7728 events to be recorded.
7730 `java.util.logging`::
7732 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7734 <<instrumenting,instrumentation>>::
7735 The use of LTTng probes to make a piece of software traceable.
7737 [[def-instrumentation-point]]instrumentation point::
7738 A point in the execution path of a piece of software that, when
7739 reached by this execution, can emit an <<def-event,event>>.
7741 instrumentation point name::
7742 See _<<def-event-name,event name>>_.
7745 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7746 developed by the Apache Software Foundation.
7749 Level of severity of a log statement or user space
7750 instrumentation point.
7753 The _Linux Trace Toolkit: next generation_ project.
7755 <<lttng-cli,cmd:lttng>>::
7756 A command-line tool provided by the LTTng-tools project which you
7757 can use to send and receive control messages to and from a
7761 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7762 which is a set of analyzing programs that are used to obtain a
7763 higher level view of an LTTng <<def-trace,trace>>.
7765 cmd:lttng-consumerd::
7766 The name of the consumer daemon program.
7769 A utility provided by the LTTng-tools project which can convert
7770 <<def-ring-buffer,ring buffer>> files (usually
7771 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7772 to <<def-trace,trace>> files.
7774 LTTng Documentation::
7777 <<lttng-live,LTTng live>>::
7778 A communication protocol between the <<lttng-relayd,relay daemon>> and
7779 live viewers which makes it possible to see <<def-event-record,event
7780 records>> "live", as they are received by the relay daemon.
7782 <<lttng-modules,LTTng-modules>>::
7783 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7784 which contains the Linux kernel modules to make the Linux kernel
7785 <<def-instrumentation-point,instrumentation points>> available for
7789 The name of the <<lttng-relayd,relay daemon>> program.
7791 cmd:lttng-sessiond::
7792 The name of the <<lttng-sessiond,session daemon>> program.
7795 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7796 contains the various programs and libraries used to
7797 <<controlling-tracing,control tracing>>.
7799 <<lttng-ust,LTTng-UST>>::
7800 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7801 contains libraries to instrument
7802 <<def-user-application,user applications>>.
7804 <<lttng-ust-agents,LTTng-UST Java agent>>::
7805 A Java package provided by the LTTng-UST project to allow the
7806 LTTng instrumentation of `java.util.logging` and Apache log4j{nbsp}1.2
7809 <<lttng-ust-agents,LTTng-UST Python agent>>::
7810 A Python package provided by the LTTng-UST project to allow the
7811 LTTng instrumentation of Python logging statements.
7813 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7814 The <<def-event-record-loss-mode,event record loss mode>> in which new
7815 <<def-event-record,event records>> _overwrite_ older event records
7816 when there's no <<def-sub-buffer,sub-buffer>> space left to store
7819 <<channel-buffering-schemes,per-process buffering>>::
7820 A <<def-buffering-scheme,buffering scheme>> in which each instrumented
7821 process has its own <<def-sub-buffer,sub-buffers>> for a given user
7822 space <<channel,channel>>.
7824 <<channel-buffering-schemes,per-user buffering>>::
7825 A <<def-buffering-scheme,buffering scheme>> in which all the processes
7826 of a Unix user share the same <<def-sub-buffer,sub-buffers>> for a
7827 given user space <<channel,channel>>.
7829 <<lttng-relayd,relay daemon>>::
7830 A process which is responsible for receiving the <<def-trace,trace>>
7831 data sent by a distant <<lttng-consumerd,consumer daemon>>.
7833 [[def-ring-buffer]]ring buffer::
7834 A set of <<def-sub-buffer,sub-buffers>>.
7837 See _<<def-tracing-session-rotation,tracing session rotation>>_.
7839 <<lttng-sessiond,session daemon>>::
7840 A process which receives control commands from you and orchestrates
7841 the <<def-tracer,tracers>> and various LTTng daemons.
7843 <<taking-a-snapshot,snapshot>>::
7844 A copy of the current data of all the <<def-sub-buffer,sub-buffers>>
7845 of a given <<tracing-session,tracing session>>, saved as
7846 <<def-trace,trace>> files.
7848 [[def-sub-buffer]]sub-buffer::
7849 One part of an LTTng <<def-ring-buffer,ring buffer>> which contains
7850 <<def-event-record,event records>>.
7853 The time information attached to an
7854 <<def-event,event>> when it is emitted.
7856 [[def-trace]]trace (_noun_)::
7859 * One http://diamon.org/ctf/[CTF] metadata stream file.
7860 * One or more CTF data stream files which are the concatenations of one
7861 or more flushed <<def-sub-buffer,sub-buffers>>.
7864 The action of recording the <<def-event,events>> emitted by an
7865 application or by a system, or to initiate such recording by
7866 controlling a tracer.
7868 [[def-trace-chunk]]trace chunk::
7869 A self-contained trace which is part of a <<tracing-session,tracing
7870 session>>. Each <<session-rotation, tracing session rotation>>
7871 produces a trace chunk archive.
7873 trace chunk archive::
7874 The result of a <<session-rotation, tracing session rotation>>. A
7875 trace chunk archive is not managed by LTTng, even if its containing
7876 <<tracing-session,tracing session>> is still active: you are free to
7877 read it, modify it, move it, or remove it.
7880 The http://tracecompass.org[Trace Compass] project and application.
7882 [[def-tracepoint]]tracepoint::
7883 An instrumentation point using the tracepoint mechanism of the Linux
7884 kernel or of LTTng-UST.
7886 tracepoint definition::
7887 The definition of a single tracepoint.
7890 The name of a tracepoint.
7892 tracepoint provider::
7893 A set of functions providing tracepoints to an instrumented user
7896 Not to be confused with a _tracepoint provider package_: many tracepoint
7897 providers can exist within a tracepoint provider package.
7899 tracepoint provider package::
7900 One or more tracepoint providers compiled as an
7901 https://en.wikipedia.org/wiki/Object_file[object file] or as
7902 a link:https://en.wikipedia.org/wiki/Library_(computing)#Shared_libraries[shared library].
7904 [[def-tracer]]tracer::
7905 A software which records emitted <<def-event,events>>.
7907 <<domain,tracing domain>>::
7908 A namespace for <<def-event,event>> sources.
7910 <<tracing-group,tracing group>>::
7911 The Unix group in which a Unix user can be to be allowed to trace the
7914 [[def-tracing-session-rotation]]<<tracing-session,tracing session>>::
7915 A stateful dialogue between you and a <<lttng-sessiond,session
7918 <<session-rotation,tracing session rotation>>::
7919 The action of archiving the
7920 <<def-current-trace-chunk,current trace chunk>> of a
7921 <<tracing-session,tracing session>>.
7923 [[def-user-application]]user application::
7924 An application running in user space, as opposed to a Linux kernel
7925 module, for example.