1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/warning-not-maintained.txt[]
13 include::../common/welcome.txt[]
16 include::../common/audience.txt[]
20 === What's in this documentation?
22 The LTTng Documentation is divided into the following sections:
24 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
25 rudiments of software tracing and the rationale behind the
28 You can skip this section if you’re familiar with software tracing and
29 with the LTTng project.
31 * **<<installing-lttng,Installation>>** describes the steps to
32 install the LTTng packages on common Linux distributions and from
35 You can skip this section if you already properly installed LTTng on
38 * **<<getting-started,Quick start>>** is a concise guide to
39 getting started quickly with LTTng kernel and user space tracing.
41 We recommend this section if you're new to LTTng or to software tracing
44 You can skip this section if you're not new to LTTng.
46 * **<<core-concepts,Core concepts>>** explains the concepts at
49 It's a good idea to become familiar with the core concepts
50 before attempting to use the toolkit.
52 * **<<plumbing,Components of LTTng>>** describes the various components
53 of the LTTng machinery, like the daemons, the libraries, and the
54 command-line interface.
55 * **<<instrumenting,Instrumentation>>** shows different ways to
56 instrument user applications and the Linux kernel.
58 Instrumenting source code is essential to provide a meaningful
61 You can skip this section if you do not have a programming background.
63 * **<<controlling-tracing,Tracing control>>** is divided into topics
64 which demonstrate how to use the vast array of features that
65 LTTng{nbsp}{revision} offers.
66 * **<<reference,Reference>>** contains reference tables.
67 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
68 to LTTng or to the field of software tracing.
71 include::../common/convention.txt[]
74 include::../common/acknowledgements.txt[]
78 == What's new in LTTng {revision}?
80 LTTng{nbsp}{revision} bears the name _KeKriek_. From
81 http://brasseriedunham.com/[Brasserie Dunham], the _**KeKriek**_ is a
82 sour mashed golden wheat ale fermented with local sour cherries from
83 Tougas orchards. Fresh sweet cherry notes with some tartness, lively
84 carbonation with a dry finish.
86 New features and changes in LTTng{nbsp}{revision}:
88 * **Tracing control**:
89 ** You can put more than one wildcard special character (`*`), and not
90 only at the end, when you <<enabling-disabling-events,create an event
91 rule>>, in both the instrumentation point name and the literal
93 link:/man/1/lttng-enable-event/v{revision}/#doc-filter-syntax[filter expressions]:
98 # lttng enable-event --kernel 'x86_*_local_timer_*' \
99 --filter='name == "*a*b*c*d*e" && count >= 23'
106 $ lttng enable-event --userspace '*_my_org:*msg*'
110 ** New trigger and notification API for
111 <<liblttng-ctl-lttng,`liblttng-ctl`>>. This new subsystem allows you
112 to register triggers which emit a notification when a given
113 condition is satisfied. As of LTTng{nbsp}{revision}, only
114 <<channel,channel>> buffer usage conditions are available.
115 Documentation is available in the
116 https://github.com/lttng/lttng-tools/tree/stable-{revision}/include/lttng[`liblttng-ctl`
118 <<notif-trigger-api,Get notified when a channel's buffer usage is too
121 ** You can now embed the whole textual LTTng-tools man pages into the
122 executables at build time with the `--enable-embedded-help`
123 configuration option. Thanks to this option, you don't need the
124 http://www.methods.co.nz/asciidoc/[AsciiDoc] and
125 https://directory.fsf.org/wiki/Xmlto[xmlto] tools at build time, and
126 a manual pager at run time, to get access to this documentation.
128 * **User space tracing**:
129 ** New blocking mode: an LTTng-UST tracepoint can now block until
130 <<channel,sub-buffer>> space is available instead of discarding event
131 records in <<channel-overwrite-mode-vs-discard-mode,discard mode>>.
132 With this feature, you can be sure that no event records are
133 discarded during your application's execution at the expense of
136 For example, the following command lines create a user space tracing
137 channel with an infinite blocking timeout and run an application
138 instrumented with LTTng-UST which is explicitly allowed to block:
144 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
145 $ lttng enable-event --userspace --channel=blocking-channel --all
147 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
151 See the complete <<blocking-timeout-example,blocking timeout example>>.
153 * **Linux kernel tracing**:
154 ** Linux 4.10, 4.11, and 4.12 support.
155 ** The thread state dump events recorded by LTTng-modules now contain
156 the task's CPU identifier. This improves the precision of the
157 scheduler model for analyses.
158 ** Extended man:socketpair(2) system call tracing data.
164 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
165 generation_ is a modern toolkit for tracing Linux systems and
166 applications. So your first question might be:
173 As the history of software engineering progressed and led to what
174 we now take for granted--complex, numerous and
175 interdependent software applications running in parallel on
176 sophisticated operating systems like Linux--the authors of such
177 components, software developers, began feeling a natural
178 urge to have tools that would ensure the robustness and good performance
179 of their masterpieces.
181 One major achievement in this field is, inarguably, the
182 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
183 an essential tool for developers to find and fix bugs. But even the best
184 debugger won't help make your software run faster, and nowadays, faster
185 software means either more work done by the same hardware, or cheaper
186 hardware for the same work.
188 A _profiler_ is often the tool of choice to identify performance
189 bottlenecks. Profiling is suitable to identify _where_ performance is
190 lost in a given software. The profiler outputs a profile, a statistical
191 summary of observed events, which you may use to discover which
192 functions took the most time to execute. However, a profiler won't
193 report _why_ some identified functions are the bottleneck. Bottlenecks
194 might only occur when specific conditions are met, conditions that are
195 sometimes impossible to capture by a statistical profiler, or impossible
196 to reproduce with an application altered by the overhead of an
197 event-based profiler. For a thorough investigation of software
198 performance issues, a history of execution is essential, with the
199 recorded values of variables and context fields you choose, and
200 with as little influence as possible on the instrumented software. This
201 is where tracing comes in handy.
203 _Tracing_ is a technique used to understand what goes on in a running
204 software system. The software used for tracing is called a _tracer_,
205 which is conceptually similar to a tape recorder. When recording,
206 specific instrumentation points placed in the software source code
207 generate events that are saved on a giant tape: a _trace_ file. You
208 can trace user applications and the operating system at the same time,
209 opening the possibility of resolving a wide range of problems that would
210 otherwise be extremely challenging.
212 Tracing is often compared to _logging_. However, tracers and loggers are
213 two different tools, serving two different purposes. Tracers are
214 designed to record much lower-level events that occur much more
215 frequently than log messages, often in the range of thousands per
216 second, with very little execution overhead. Logging is more appropriate
217 for a very high-level analysis of less frequent events: user accesses,
218 exceptional conditions (errors and warnings, for example), database
219 transactions, instant messaging communications, and such. Simply put,
220 logging is one of the many use cases that can be satisfied with tracing.
222 The list of recorded events inside a trace file can be read manually
223 like a log file for the maximum level of detail, but it is generally
224 much more interesting to perform application-specific analyses to
225 produce reduced statistics and graphs that are useful to resolve a
226 given problem. Trace viewers and analyzers are specialized tools
229 In the end, this is what LTTng is: a powerful, open source set of
230 tools to trace the Linux kernel and user applications at the same time.
231 LTTng is composed of several components actively maintained and
232 developed by its link:/community/#where[community].
235 [[lttng-alternatives]]
236 === Alternatives to noch:{LTTng}
238 Excluding proprietary solutions, a few competing software tracers
241 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
242 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
243 user scripts and is responsible for loading code into the
244 Linux kernel for further execution and collecting the outputted data.
245 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
246 subsystem in the Linux kernel in which a virtual machine can execute
247 programs passed from the user space to the kernel. You can attach
248 such programs to tracepoints and KProbes thanks to a system call, and
249 they can output data to the user space when executed thanks to
250 different mechanisms (pipe, VM register values, and eBPF maps, to name
252 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
253 is the de facto function tracer of the Linux kernel. Its user
254 interface is a set of special files in sysfs.
255 * https://perf.wiki.kernel.org/[perf] is
256 a performance analyzing tool for Linux which supports hardware
257 performance counters, tracepoints, as well as other counters and
258 types of probes. perf's controlling utility is the cmd:perf command
260 * http://linux.die.net/man/1/strace[strace]
261 is a command-line utility which records system calls made by a
262 user process, as well as signal deliveries and changes of process
263 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
264 to fulfill its function.
265 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
266 analyze Linux kernel events. You write scripts, or _chisels_ in
267 sysdig's jargon, in Lua and sysdig executes them while the system is
268 being traced or afterwards. sysdig's interface is the cmd:sysdig
269 command-line tool as well as the curses-based cmd:csysdig tool.
270 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
271 user space tracer which uses custom user scripts to produce plain text
272 traces. SystemTap converts the scripts to the C language, and then
273 compiles them as Linux kernel modules which are loaded to produce
274 trace data. SystemTap's primary user interface is the cmd:stap
277 The main distinctive features of LTTng is that it produces correlated
278 kernel and user space traces, as well as doing so with the lowest
279 overhead amongst other solutions. It produces trace files in the
280 http://diamon.org/ctf[CTF] format, a file format optimized
281 for the production and analyses of multi-gigabyte data.
283 LTTng is the result of more than 10 years of active open source
284 development by a community of passionate developers.
285 LTTng{nbsp}{revision} is currently available on major desktop and server
288 The main interface for tracing control is a single command-line tool
289 named cmd:lttng. The latter can create several tracing sessions, enable
290 and disable events on the fly, filter events efficiently with custom
291 user expressions, start and stop tracing, and much more. LTTng can
292 record the traces on the file system or send them over the network, and
293 keep them totally or partially. You can view the traces once tracing
294 becomes inactive or in real-time.
296 <<installing-lttng,Install LTTng now>> and
297 <<getting-started,start tracing>>!
303 **LTTng** is a set of software <<plumbing,components>> which interact to
304 <<instrumenting,instrument>> the Linux kernel and user applications, and
305 to <<controlling-tracing,control tracing>> (start and stop
306 tracing, enable and disable event rules, and the rest). Those
307 components are bundled into the following packages:
309 * **LTTng-tools**: Libraries and command-line interface to
311 * **LTTng-modules**: Linux kernel modules to instrument and
313 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
314 trace user applications.
316 Most distributions mark the LTTng-modules and LTTng-UST packages as
317 optional when installing LTTng-tools (which is always required). In the
318 following sections, we always provide the steps to install all three,
321 * You only need to install LTTng-modules if you intend to trace the
323 * You only need to install LTTng-UST if you intend to trace user
327 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 18 October 2019.
329 |Distribution |Available in releases
331 |https://www.ubuntu.com/[Ubuntu]
332 |xref:ubuntu[Ubuntu{nbsp}18.04 _Bionic Beaver_,
333 Ubuntu{nbsp}19.04 _Disco Dingo_, and
334 Ubuntu{nbsp}19.10 _Eoan Ermine_].
336 Ubuntu{nbsp}16.04 _Xenial Xerus_:
337 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
339 |https://getfedora.org/[Fedora]
340 |xref:fedora[Fedora{nbsp}29, Fedora{nbsp}30, Fedora{nbsp}31,
343 |https://www.debian.org/[Debian]
344 |<<debian,Debian "buster" (stable) and Debian "bullseye" (testing)>>.
346 |https://alpinelinux.org/[Alpine Linux]
347 |xref:alpine-linux[Alpine Linux{nbsp}3.7, Alpine Linux{nbsp}3.8,
348 Alpine Linux{nbsp}3.9, and Alpine Linux{nbsp}3.10].
350 |https://www.opensuse.org/[openSUSE]
351 |<<opensuse,openSUSE Leap{nbsp}15.1>>.
353 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
354 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
356 |https://buildroot.org/[Buildroot]
357 |xref:buildroot[Buildroot{nbsp}2018.02, Buildroot{nbsp}2018.05,
358 Buildroot{nbsp}2018.08, Buildroot{nbsp}2018.11, Buildroot{nbsp}2019.02,
359 Buildroot{nbsp}2018.05, Buildroot{nbsp}2018.08, and
360 Buildroot{nbsp}2018.11].
362 |http://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
363 https://www.yoctoproject.org/[Yocto]
364 |<<oe-yocto,Yocto Project{nbsp}2.7 _Warrior_ and
365 Yocto Project{nbsp}3.0 _Zeus_>>.
370 === [[ubuntu-official-repositories]]Ubuntu
372 LTTng{nbsp}{revision} is available on:
374 * Ubuntu{nbsp}18.04 _Bionic Beaver_
375 * Ubuntu{nbsp}19.04 _Disco Dingo_
376 * Ubuntu{nbsp}19.10 _Eoan Ermine_
378 For other releases of Ubuntu, <<ubuntu-ppa,use the LTTng
379 Stable{nbsp}{revision} PPA>>.
381 To install LTTng{nbsp}{revision} on Ubuntu{nbsp}18.04 _Bionic Beaver_,
382 Ubuntu{nbsp}19.04 _Disco Dingo_, or
383 Ubuntu{nbsp}19.10 _Eoan Ermine_:
385 . Install the main LTTng{nbsp}{revision} packages:
390 # apt-get install lttng-tools
391 # apt-get install lttng-modules-dkms
392 # apt-get install liblttng-ust-dev
396 . **If you need to instrument and trace
397 <<java-application,Java applications>>**, install the LTTng-UST
403 # apt-get install liblttng-ust-agent-java
407 . **If you need to instrument and trace
408 <<python-application,Python{nbsp}3 applications>>**, install the
409 LTTng-UST Python agent:
414 # apt-get install python3-lttngust
420 ==== noch:{LTTng} Stable {revision} PPA
422 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
423 Stable{nbsp}{revision} PPA] offers the latest stable
424 LTTng{nbsp}{revision} packages for Ubuntu{nbsp}18.04 _Bionic Beaver_.
426 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
428 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
434 # apt-add-repository ppa:lttng/stable-2.10
439 . Install the main LTTng{nbsp}{revision} packages:
444 # apt-get install lttng-tools
445 # apt-get install lttng-modules-dkms
446 # apt-get install liblttng-ust-dev
450 . **If you need to instrument and trace
451 <<java-application,Java applications>>**, install the LTTng-UST
457 # apt-get install liblttng-ust-agent-java
461 . **If you need to instrument and trace
462 <<python-application,Python{nbsp}3 applications>>**, install the
463 LTTng-UST Python agent:
468 # apt-get install python3-lttngust
476 To install LTTng{nbsp}{revision} on Fedora{nbsp}29, Fedora{nbsp}30,
477 Fedora{nbsp}31, or Fedora{nbsp}32:
479 . Install the LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision}
485 # yum install lttng-tools
486 # yum install lttng-ust
490 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
496 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.10.tar.bz2 &&
497 tar -xf lttng-modules-latest-2.10.tar.bz2 &&
498 cd lttng-modules-2.10.* &&
500 sudo make modules_install &&
506 .Java and Python application instrumentation and tracing
508 If you need to instrument and trace <<java-application,Java
509 applications>> on Fedora, you need to build and install
510 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
511 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
512 `--enable-java-agent-all` options to the `configure` script, depending
513 on which Java logging framework you use.
515 If you need to instrument and trace <<python-application,Python
516 applications>> on Fedora, you need to build and install
517 LTTng-UST{nbsp}{revision} from source and pass the
518 `--enable-python-agent` option to the `configure` script.
525 To install LTTng{nbsp}{revision} on Debian "buster" (stable) or
526 Debian "bullseye" (testing):
528 . Install the main LTTng{nbsp}{revision} packages:
533 # apt-get install lttng-modules-dkms
534 # apt-get install liblttng-ust-dev
535 # apt-get install lttng-tools
539 . **If you need to instrument and trace <<java-application,Java
540 applications>>**, install the LTTng-UST Java agent:
545 # apt-get install liblttng-ust-agent-java
549 . **If you need to instrument and trace <<python-application,Python
550 applications>>**, install the LTTng-UST Python agent:
555 # apt-get install python3-lttngust
563 To install LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} on
564 Alpine Linux{nbsp}3.7, Alpine Linux{nbsp}3.8, Alpine Linux{nbsp}3.9, or
565 Alpine Linux{nbsp}3.10:
567 . Add the LTTng packages:
572 # apk add lttng-tools
573 # apk add lttng-ust-dev
577 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
583 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.10.tar.bz2 &&
584 tar -xf lttng-modules-latest-2.10.tar.bz2 &&
585 cd lttng-modules-2.10.* &&
587 sudo make modules_install &&
596 To install LTTng{nbsp}{revision} on openSUSE Leap{nbsp}15.1:
598 * Install the main LTTng{nbsp}{revision} packages:
603 sudo zypper install lttng-tools
604 sudo zypper install lttng-modules
605 sudo zypper install lttng-ust-devel
610 .Java and Python application instrumentation and tracing
612 If you need to instrument and trace <<java-application,Java
613 applications>> on openSUSE, you need to build and install
614 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
615 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
616 `--enable-java-agent-all` options to the `configure` script, depending
617 on which Java logging framework you use.
619 If you need to instrument and trace <<python-application,Python
620 applications>> on openSUSE, you need to build and install
621 LTTng-UST{nbsp}{revision} from source and pass the
622 `--enable-python-agent` option to the `configure` script.
626 [[enterprise-distributions]]
627 === RHEL, SUSE, and other enterprise distributions
629 To install LTTng on enterprise Linux distributions, such as Red Hat
630 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
631 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
637 To install LTTng{nbsp}{revision} on Buildroot{nbsp}2018.02,
638 Buildroot{nbsp}2018.05, Buildroot{nbsp}2018.08,
639 Buildroot{nbsp}2018.11, Buildroot{nbsp}2019.02,
640 Buildroot{nbsp}2019.05, Buildroot{nbsp}2019.08, or
641 Buildroot{nbsp}2019.11:
643 . Launch the Buildroot configuration tool:
652 . In **Kernel**, check **Linux kernel**.
653 . In **Toolchain**, check **Enable WCHAR support**.
654 . In **Target packages**{nbsp}→ **Debugging, profiling and benchmark**,
655 check **lttng-modules** and **lttng-tools**.
656 . In **Target packages**{nbsp}→ **Libraries**{nbsp}→
657 **Other**, check **lttng-libust**.
661 === OpenEmbedded and Yocto
663 LTTng{nbsp}{revision} recipes are available in the
664 http://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
665 layer for Yocto Project{nbsp}2.7 _Warrior_ and
666 Yocto Project{nbsp}3.0 _Zeus_ under the following names:
672 With BitBake, the simplest way to include LTTng recipes in your target
673 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
676 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
681 . Select a machine and an image recipe.
682 . Click **Edit image recipe**.
683 . Under the **All recipes** tab, search for **lttng**.
684 . Check the desired LTTng recipes.
687 .Java and Python application instrumentation and tracing
689 If you need to instrument and trace <<java-application,Java
690 applications>> on Yocto/OpenEmbedded, you need to build and install
691 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
692 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
693 `--enable-java-agent-all` options to the `configure` script, depending
694 on which Java logging framework you use.
696 If you need to instrument and trace <<python-application,Python
697 applications>> on Yocto/OpenEmbedded, you need to build and install
698 LTTng-UST{nbsp}{revision} from source and pass the
699 `--enable-python-agent` option to the `configure` script.
703 [[building-from-source]]
704 === Build from source
706 To build and install LTTng{nbsp}{revision} from source:
708 . Using your distribution's package manager, or from source, install
709 the following dependencies of LTTng-tools and LTTng-UST:
712 * https://sourceforge.net/projects/libuuid/[libuuid]
713 * http://directory.fsf.org/wiki/Popt[popt]
714 * http://liburcu.org/[Userspace RCU]
715 * http://www.xmlsoft.org/[libxml2]
718 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
724 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.10.tar.bz2 &&
725 tar -xf lttng-modules-latest-2.10.tar.bz2 &&
726 cd lttng-modules-2.10.* &&
728 sudo make modules_install &&
733 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
739 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.10.tar.bz2 &&
740 tar -xf lttng-ust-latest-2.10.tar.bz2 &&
741 cd lttng-ust-2.10.* &&
751 .Java and Python application tracing
753 If you need to instrument and trace <<java-application,Java
754 applications>>, pass the `--enable-java-agent-jul`,
755 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
756 `configure` script, depending on which Java logging framework you use.
758 If you need to instrument and trace <<python-application,Python
759 applications>>, pass the `--enable-python-agent` option to the
760 `configure` script. You can set the `PYTHON` environment variable to the
761 path to the Python interpreter for which to install the LTTng-UST Python
769 By default, LTTng-UST libraries are installed to
770 dir:{/usr/local/lib}, which is the de facto directory in which to
771 keep self-compiled and third-party libraries.
773 When <<building-tracepoint-providers-and-user-application,linking an
774 instrumented user application with `liblttng-ust`>>:
776 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
778 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
779 man:gcc(1), man:g++(1), or man:clang(1).
783 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
789 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.10.tar.bz2 &&
790 tar -xf lttng-tools-latest-2.10.tar.bz2 &&
791 cd lttng-tools-2.10.* &&
799 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
800 previous steps automatically for a given version of LTTng and confine
801 the installed files in a specific directory. This can be useful to test
802 LTTng without installing it on your system.
808 This is a short guide to get started quickly with LTTng kernel and user
811 Before you follow this guide, make sure to <<installing-lttng,install>>
814 This tutorial walks you through the steps to:
816 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
817 . <<tracing-your-own-user-application,Trace a user application>> written
819 . <<viewing-and-analyzing-your-traces,View and analyze the
823 [[tracing-the-linux-kernel]]
824 === Trace the Linux kernel
826 The following command lines start with the `#` prompt because you need
827 root privileges to trace the Linux kernel. You can also trace the kernel
828 as a regular user if your Unix user is a member of the
829 <<tracing-group,tracing group>>.
831 . Create a <<tracing-session,tracing session>> which writes its traces
832 to dir:{/tmp/my-kernel-trace}:
837 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
841 . List the available kernel tracepoints and system calls:
846 # lttng list --kernel
847 # lttng list --kernel --syscall
851 . Create <<event,event rules>> which match the desired instrumentation
852 point names, for example the `sched_switch` and `sched_process_fork`
853 tracepoints, and the man:open(2) and man:close(2) system calls:
858 # lttng enable-event --kernel sched_switch,sched_process_fork
859 # lttng enable-event --kernel --syscall open,close
863 You can also create an event rule which matches _all_ the Linux kernel
864 tracepoints (this will generate a lot of data when tracing):
869 # lttng enable-event --kernel --all
873 . <<basic-tracing-session-control,Start tracing>>:
882 . Do some operation on your system for a few seconds. For example,
883 load a website, or list the files of a directory.
884 . <<creating-destroying-tracing-sessions,Destroy>> the current
894 The man:lttng-destroy(1) command does not destroy the trace data; it
895 only destroys the state of the tracing session.
897 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
898 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
899 session>>). You need to stop tracing to make LTTng flush the remaining
900 trace data and make the trace readable.
902 . For the sake of this example, make the recorded trace accessible to
908 # chown -R $(whoami) /tmp/my-kernel-trace
912 See <<viewing-and-analyzing-your-traces,View and analyze the
913 recorded events>> to view the recorded events.
916 [[tracing-your-own-user-application]]
917 === Trace a user application
919 This section steps you through a simple example to trace a
920 _Hello world_ program written in C.
922 To create the traceable user application:
924 . Create the tracepoint provider header file, which defines the
925 tracepoints and the events they can generate:
931 #undef TRACEPOINT_PROVIDER
932 #define TRACEPOINT_PROVIDER hello_world
934 #undef TRACEPOINT_INCLUDE
935 #define TRACEPOINT_INCLUDE "./hello-tp.h"
937 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
940 #include <lttng/tracepoint.h>
950 ctf_string(my_string_field, my_string_arg)
951 ctf_integer(int, my_integer_field, my_integer_arg)
955 #endif /* _HELLO_TP_H */
957 #include <lttng/tracepoint-event.h>
961 . Create the tracepoint provider package source file:
967 #define TRACEPOINT_CREATE_PROBES
968 #define TRACEPOINT_DEFINE
970 #include "hello-tp.h"
974 . Build the tracepoint provider package:
979 $ gcc -c -I. hello-tp.c
983 . Create the _Hello World_ application source file:
990 #include "hello-tp.h"
992 int main(int argc, char *argv[])
996 puts("Hello, World!\nPress Enter to continue...");
999 * The following getchar() call is only placed here for the purpose
1000 * of this demonstration, to pause the application in order for
1001 * you to have time to list its tracepoints. It is not
1007 * A tracepoint() call.
1009 * Arguments, as defined in hello-tp.h:
1011 * 1. Tracepoint provider name (required)
1012 * 2. Tracepoint name (required)
1013 * 3. my_integer_arg (first user-defined argument)
1014 * 4. my_string_arg (second user-defined argument)
1016 * Notice the tracepoint provider and tracepoint names are
1017 * NOT strings: they are in fact parts of variables that the
1018 * macros in hello-tp.h create.
1020 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
1022 for (x = 0; x < argc; ++x) {
1023 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
1026 puts("Quitting now!");
1027 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
1034 . Build the application:
1043 . Link the application with the tracepoint provider package,
1044 `liblttng-ust`, and `libdl`:
1049 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
1053 Here's the whole build process:
1056 .User space tracing tutorial's build steps.
1057 image::ust-flow.png[]
1059 To trace the user application:
1061 . Run the application with a few arguments:
1066 $ ./hello world and beyond
1075 Press Enter to continue...
1079 . Start an LTTng <<lttng-sessiond,session daemon>>:
1084 $ lttng-sessiond --daemonize
1088 Note that a session daemon might already be running, for example as
1089 a service that the distribution's service manager started.
1091 . List the available user space tracepoints:
1096 $ lttng list --userspace
1100 You see the `hello_world:my_first_tracepoint` tracepoint listed
1101 under the `./hello` process.
1103 . Create a <<tracing-session,tracing session>>:
1108 $ lttng create my-user-space-session
1112 . Create an <<event,event rule>> which matches the
1113 `hello_world:my_first_tracepoint` event name:
1118 $ lttng enable-event --userspace hello_world:my_first_tracepoint
1122 . <<basic-tracing-session-control,Start tracing>>:
1131 . Go back to the running `hello` application and press Enter. The
1132 program executes all `tracepoint()` instrumentation points and exits.
1133 . <<creating-destroying-tracing-sessions,Destroy>> the current
1143 The man:lttng-destroy(1) command does not destroy the trace data; it
1144 only destroys the state of the tracing session.
1146 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
1147 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
1148 session>>). You need to stop tracing to make LTTng flush the remaining
1149 trace data and make the trace readable.
1151 By default, LTTng saves the traces in
1152 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1153 where +__name__+ is the tracing session name. The
1154 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1156 See <<viewing-and-analyzing-your-traces,View and analyze the
1157 recorded events>> to view the recorded events.
1160 [[viewing-and-analyzing-your-traces]]
1161 === View and analyze the recorded events
1163 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1164 kernel>> and <<tracing-your-own-user-application,Trace a user
1165 application>> tutorials, you can inspect the recorded events.
1167 Many tools are available to read LTTng traces:
1169 * **cmd:babeltrace** is a command-line utility which converts trace
1170 formats; it supports the format that LTTng produces, CTF, as well as a
1171 basic text output which can be ++grep++ed. The cmd:babeltrace command
1172 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1173 * Babeltrace also includes
1174 **https://www.python.org/[Python] bindings** so
1175 that you can easily open and read an LTTng trace with your own script,
1176 benefiting from the power of Python.
1177 * http://tracecompass.org/[**Trace Compass**]
1178 is a graphical user interface for viewing and analyzing any type of
1179 logs or traces, including LTTng's.
1180 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1181 project which includes many high-level analyses of LTTng kernel
1182 traces, like scheduling statistics, interrupt frequency distribution,
1183 top CPU usage, and more.
1185 NOTE: This section assumes that the traces recorded during the previous
1186 tutorials were saved to their default location, in the
1187 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
1188 environment variable defaults to `$HOME` if not set.
1191 [[viewing-and-analyzing-your-traces-bt]]
1192 ==== Use the cmd:babeltrace command-line tool
1194 The simplest way to list all the recorded events of a trace is to pass
1195 its path to cmd:babeltrace with no options:
1199 $ babeltrace ~/lttng-traces/my-user-space-session*
1202 cmd:babeltrace finds all traces recursively within the given path and
1203 prints all their events, merging them in chronological order.
1205 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1210 $ babeltrace /tmp/my-kernel-trace | grep _switch
1213 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1214 count the recorded events:
1218 $ babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
1222 [[viewing-and-analyzing-your-traces-bt-python]]
1223 ==== Use the Babeltrace Python bindings
1225 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1226 is useful to isolate events by simple matching using man:grep(1) and
1227 similar utilities. However, more elaborate filters, such as keeping only
1228 event records with a field value falling within a specific range, are
1229 not trivial to write using a shell. Moreover, reductions and even the
1230 most basic computations involving multiple event records are virtually
1231 impossible to implement.
1233 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1234 to read the event records of an LTTng trace sequentially and compute the
1235 desired information.
1237 The following script accepts an LTTng Linux kernel trace path as its
1238 first argument and prints the short names of the top 5 running processes
1239 on CPU 0 during the whole trace:
1244 from collections import Counter
1250 if len(sys.argv) != 2:
1251 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1252 print(msg, file=sys.stderr)
1255 # A trace collection contains one or more traces
1256 col = babeltrace.TraceCollection()
1258 # Add the trace provided by the user (LTTng traces always have
1260 if col.add_trace(sys.argv[1], 'ctf') is None:
1261 raise RuntimeError('Cannot add trace')
1263 # This counter dict contains execution times:
1265 # task command name -> total execution time (ns)
1266 exec_times = Counter()
1268 # This contains the last `sched_switch` timestamp
1272 for event in col.events:
1273 # Keep only `sched_switch` events
1274 if event.name != 'sched_switch':
1277 # Keep only events which happened on CPU 0
1278 if event['cpu_id'] != 0:
1282 cur_ts = event.timestamp
1288 # Previous task command (short) name
1289 prev_comm = event['prev_comm']
1291 # Initialize entry in our dict if not yet done
1292 if prev_comm not in exec_times:
1293 exec_times[prev_comm] = 0
1295 # Compute previous command execution time
1296 diff = cur_ts - last_ts
1298 # Update execution time of this command
1299 exec_times[prev_comm] += diff
1301 # Update last timestamp
1305 for name, ns in exec_times.most_common(5):
1307 print('{:20}{} s'.format(name, s))
1312 if __name__ == '__main__':
1313 sys.exit(0 if top5proc() else 1)
1320 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1326 swapper/0 48.607245889 s
1327 chromium 7.192738188 s
1328 pavucontrol 0.709894415 s
1329 Compositor 0.660867933 s
1330 Xorg.bin 0.616753786 s
1333 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1334 weren't using the CPU that much when tracing, its first position in the
1339 == [[understanding-lttng]]Core concepts
1341 From a user's perspective, the LTTng system is built on a few concepts,
1342 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1343 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1344 Understanding how those objects relate to eachother is key in mastering
1347 The core concepts are:
1349 * <<tracing-session,Tracing session>>
1350 * <<domain,Tracing domain>>
1351 * <<channel,Channel and ring buffer>>
1352 * <<"event","Instrumentation point, event rule, event, and event record">>
1358 A _tracing session_ is a stateful dialogue between you and
1359 a <<lttng-sessiond,session daemon>>. You can
1360 <<creating-destroying-tracing-sessions,create a new tracing
1361 session>> with the `lttng create` command.
1363 Anything that you do when you control LTTng tracers happens within a
1364 tracing session. In particular, a tracing session:
1367 * Has its own set of trace files.
1368 * Has its own state of activity (started or stopped).
1369 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1371 * Has its own <<channel,channels>> which have their own
1372 <<event,event rules>>.
1375 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1376 image::concepts.png[]
1378 Those attributes and objects are completely isolated between different
1381 A tracing session is analogous to a cash machine session:
1382 the operations you do on the banking system through the cash machine do
1383 not alter the data of other users of the same system. In the case of
1384 the cash machine, a session lasts as long as your bank card is inside.
1385 In the case of LTTng, a tracing session lasts from the `lttng create`
1386 command to the `lttng destroy` command.
1389 .Each Unix user has its own set of tracing sessions.
1390 image::many-sessions.png[]
1393 [[tracing-session-mode]]
1394 ==== Tracing session mode
1396 LTTng can send the generated trace data to different locations. The
1397 _tracing session mode_ dictates where to send it. The following modes
1398 are available in LTTng{nbsp}{revision}:
1401 LTTng writes the traces to the file system of the machine being traced
1404 Network streaming mode::
1405 LTTng sends the traces over the network to a
1406 <<lttng-relayd,relay daemon>> running on a remote system.
1409 LTTng does not write the traces by default. Instead, you can request
1410 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1411 current tracing buffers, and to write it to the target's file system
1412 or to send it over the network to a <<lttng-relayd,relay daemon>>
1413 running on a remote system.
1416 This mode is similar to the network streaming mode, but a live
1417 trace viewer can connect to the distant relay daemon to
1418 <<lttng-live,view event records as LTTng generates them>> by
1425 A _tracing domain_ is a namespace for event sources. A tracing domain
1426 has its own properties and features.
1428 There are currently five available tracing domains:
1432 * `java.util.logging` (JUL)
1436 You must specify a tracing domain when using some commands to avoid
1437 ambiguity. For example, since all the domains support named tracepoints
1438 as event sources (instrumentation points that you manually insert in the
1439 source code), you need to specify a tracing domain when
1440 <<enabling-disabling-events,creating an event rule>> because all the
1441 tracing domains could have tracepoints with the same names.
1443 Some features are reserved to specific tracing domains. Dynamic function
1444 entry and return instrumentation points, for example, are currently only
1445 supported in the Linux kernel tracing domain, but support for other
1446 tracing domains could be added in the future.
1448 You can create <<channel,channels>> in the Linux kernel and user space
1449 tracing domains. The other tracing domains have a single default
1454 === Channel and ring buffer
1456 A _channel_ is an object which is responsible for a set of ring buffers.
1457 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1458 tracer emits an event, it can record it to one or more
1459 sub-buffers. The attributes of a channel determine what to do when
1460 there's no space left for a new event record because all sub-buffers
1461 are full, where to send a full sub-buffer, and other behaviours.
1463 A channel is always associated to a <<domain,tracing domain>>. The
1464 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1465 a default channel which you cannot configure.
1467 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1468 an event, it records it to the sub-buffers of all
1469 the enabled channels with a satisfied event rule, as long as those
1470 channels are part of active <<tracing-session,tracing sessions>>.
1473 [[channel-buffering-schemes]]
1474 ==== Per-user vs. per-process buffering schemes
1476 A channel has at least one ring buffer _per CPU_. LTTng always
1477 records an event to the ring buffer associated to the CPU on which it
1480 Two _buffering schemes_ are available when you
1481 <<enabling-disabling-channels,create a channel>> in the
1482 user space <<domain,tracing domain>>:
1484 Per-user buffering::
1485 Allocate one set of ring buffers--one per CPU--shared by all the
1486 instrumented processes of each Unix user.
1490 .Per-user buffering scheme.
1491 image::per-user-buffering.png[]
1494 Per-process buffering::
1495 Allocate one set of ring buffers--one per CPU--for each
1496 instrumented process.
1500 .Per-process buffering scheme.
1501 image::per-process-buffering.png[]
1504 The per-process buffering scheme tends to consume more memory than the
1505 per-user option because systems generally have more instrumented
1506 processes than Unix users running instrumented processes. However, the
1507 per-process buffering scheme ensures that one process having a high
1508 event throughput won't fill all the shared sub-buffers of the same
1511 The Linux kernel tracing domain has only one available buffering scheme
1512 which is to allocate a single set of ring buffers for the whole system.
1513 This scheme is similar to the per-user option, but with a single, global
1514 user "running" the kernel.
1517 [[channel-overwrite-mode-vs-discard-mode]]
1518 ==== Overwrite vs. discard event loss modes
1520 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1521 arc in the following animation) of a specific channel's ring buffer.
1522 When there's no space left in a sub-buffer, the tracer marks it as
1523 consumable (red) and another, empty sub-buffer starts receiving the
1524 following event records. A <<lttng-consumerd,consumer daemon>>
1525 eventually consumes the marked sub-buffer (returns to white).
1528 [role="docsvg-channel-subbuf-anim"]
1533 In an ideal world, sub-buffers are consumed faster than they are filled,
1534 as is the case in the previous animation. In the real world,
1535 however, all sub-buffers can be full at some point, leaving no space to
1536 record the following events.
1538 By default, LTTng-modules and LTTng-UST are _non-blocking_ tracers: when
1539 no empty sub-buffer is available, it is acceptable to lose event records
1540 when the alternative would be to cause substantial delays in the
1541 instrumented application's execution. LTTng privileges performance over
1542 integrity; it aims at perturbing the traced system as little as possible
1543 in order to make tracing of subtle race conditions and rare interrupt
1546 Starting from LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST,
1547 supports a _blocking mode_. See the <<blocking-timeout-example,blocking
1548 timeout example>> to learn how to use the blocking mode.
1550 When it comes to losing event records because no empty sub-buffer is
1551 available, or because the <<opt-blocking-timeout,blocking timeout>> is
1552 reached, the channel's _event loss mode_ determines what to do. The
1553 available event loss modes are:
1556 Drop the newest event records until a the tracer releases a
1559 This is the only available mode when you specify a
1560 <<opt-blocking-timeout,blocking timeout>>.
1563 Clear the sub-buffer containing the oldest event records and start
1564 writing the newest event records there.
1566 This mode is sometimes called _flight recorder mode_ because it's
1568 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1569 always keep a fixed amount of the latest data.
1571 Which mechanism you should choose depends on your context: prioritize
1572 the newest or the oldest event records in the ring buffer?
1574 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1575 as soon as a there's no space left for a new event record, whereas in
1576 discard mode, the tracer only discards the event record that doesn't
1579 In discard mode, LTTng increments a count of lost event records when an
1580 event record is lost and saves this count to the trace. In overwrite
1581 mode, since LTTng 2.8, LTTng increments a count of lost sub-buffers when
1582 a sub-buffer is lost and saves this count to the trace. In this mode,
1583 the exact number of lost event records in those lost sub-buffers is not
1584 saved to the trace. Trace analyses can use the trace's saved discarded
1585 event record and sub-buffer counts to decide whether or not to perform
1586 the analyses even if trace data is known to be missing.
1588 There are a few ways to decrease your probability of losing event
1590 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1591 how you can fine-tune the sub-buffer count and size of a channel to
1592 virtually stop losing event records, though at the cost of greater
1596 [[channel-subbuf-size-vs-subbuf-count]]
1597 ==== Sub-buffer count and size
1599 When you <<enabling-disabling-channels,create a channel>>, you can
1600 set its number of sub-buffers and their size.
1602 Note that there is noticeable CPU overhead introduced when
1603 switching sub-buffers (marking a full one as consumable and switching
1604 to an empty one for the following events to be recorded). Knowing this,
1605 the following list presents a few practical situations along with how
1606 to configure the sub-buffer count and size for them:
1608 * **High event throughput**: In general, prefer bigger sub-buffers to
1609 lower the risk of losing event records.
1611 Having bigger sub-buffers also ensures a lower
1612 <<channel-switch-timer,sub-buffer switching frequency>>.
1614 The number of sub-buffers is only meaningful if you create the channel
1615 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1616 other sub-buffers are left unaltered.
1618 * **Low event throughput**: In general, prefer smaller sub-buffers
1619 since the risk of losing event records is low.
1621 Because events occur less frequently, the sub-buffer switching frequency
1622 should remain low and thus the tracer's overhead should not be a
1625 * **Low memory system**: If your target system has a low memory
1626 limit, prefer fewer first, then smaller sub-buffers.
1628 Even if the system is limited in memory, you want to keep the
1629 sub-buffers as big as possible to avoid a high sub-buffer switching
1632 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1633 which means event data is very compact. For example, the average
1634 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1635 sub-buffer size of 1{nbsp}MiB is considered big.
1637 The previous situations highlight the major trade-off between a few big
1638 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1639 frequency vs. how much data is lost in overwrite mode. Assuming a
1640 constant event throughput and using the overwrite mode, the two
1641 following configurations have the same ring buffer total size:
1644 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1649 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1650 switching frequency, but if a sub-buffer overwrite happens, half of
1651 the event records so far (4{nbsp}MiB) are definitely lost.
1652 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1653 overhead as the previous configuration, but if a sub-buffer
1654 overwrite happens, only the eighth of event records so far are
1657 In discard mode, the sub-buffers count parameter is pointless: use two
1658 sub-buffers and set their size according to the requirements of your
1662 [[channel-switch-timer]]
1663 ==== Switch timer period
1665 The _switch timer period_ is an important configurable attribute of
1666 a channel to ensure periodic sub-buffer flushing.
1668 When the _switch timer_ expires, a sub-buffer switch happens. You can
1669 set the switch timer period attribute when you
1670 <<enabling-disabling-channels,create a channel>> to ensure that event
1671 data is consumed and committed to trace files or to a distant relay
1672 daemon periodically in case of a low event throughput.
1675 [role="docsvg-channel-switch-timer"]
1680 This attribute is also convenient when you use big sub-buffers to cope
1681 with a sporadic high event throughput, even if the throughput is
1685 [[channel-read-timer]]
1686 ==== Read timer period
1688 By default, the LTTng tracers use a notification mechanism to signal a
1689 full sub-buffer so that a consumer daemon can consume it. When such
1690 notifications must be avoided, for example in real-time applications,
1691 you can use the channel's _read timer_ instead. When the read timer
1692 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1693 consumable sub-buffers.
1696 [[tracefile-rotation]]
1697 ==== Trace file count and size
1699 By default, trace files can grow as large as needed. You can set the
1700 maximum size of each trace file that a channel writes when you
1701 <<enabling-disabling-channels,create a channel>>. When the size of
1702 a trace file reaches the channel's fixed maximum size, LTTng creates
1703 another file to contain the next event records. LTTng appends a file
1704 count to each trace file name in this case.
1706 If you set the trace file size attribute when you create a channel, the
1707 maximum number of trace files that LTTng creates is _unlimited_ by
1708 default. To limit them, you can also set a maximum number of trace
1709 files. When the number of trace files reaches the channel's fixed
1710 maximum count, the oldest trace file is overwritten. This mechanism is
1711 called _trace file rotation_.
1715 === Instrumentation point, event rule, event, and event record
1717 An _event rule_ is a set of conditions which must be **all** satisfied
1718 for LTTng to record an occuring event.
1720 You set the conditions when you <<enabling-disabling-events,create
1723 You always attach an event rule to <<channel,channel>> when you create
1726 When an event passes the conditions of an event rule, LTTng records it
1727 in one of the attached channel's sub-buffers.
1729 The available conditions, as of LTTng{nbsp}{revision}, are:
1731 * The event rule _is enabled_.
1732 * The instrumentation point's type _is{nbsp}T_.
1733 * The instrumentation point's name (sometimes called _event name_)
1734 _matches{nbsp}N_, but _is not{nbsp}E_.
1735 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1736 _is exactly{nbsp}L_.
1737 * The fields of the event's payload _satisfy_ a filter
1738 expression{nbsp}__F__.
1740 As you can see, all the conditions but the dynamic filter are related to
1741 the event rule's status or to the instrumentation point, not to the
1742 occurring events. This is why, without a filter, checking if an event
1743 passes an event rule is not a dynamic task: when you create or modify an
1744 event rule, all the tracers of its tracing domain enable or disable the
1745 instrumentation points themselves once. This is possible because the
1746 attributes of an instrumentation point (type, name, and log level) are
1747 defined statically. In other words, without a dynamic filter, the tracer
1748 _does not evaluate_ the arguments of an instrumentation point unless it
1749 matches an enabled event rule.
1751 Note that, for LTTng to record an event, the <<channel,channel>> to
1752 which a matching event rule is attached must also be enabled, and the
1753 tracing session owning this channel must be active.
1756 .Logical path from an instrumentation point to an event record.
1757 image::event-rule.png[]
1759 .Event, event record, or event rule?
1761 With so many similar terms, it's easy to get confused.
1763 An **event** is the consequence of the execution of an _instrumentation
1764 point_, like a tracepoint that you manually place in some source code,
1765 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1766 time. Different actions can be taken upon the occurrence of an event,
1767 like record the event's payload to a buffer.
1769 An **event record** is the representation of an event in a sub-buffer. A
1770 tracer is responsible for capturing the payload of an event, current
1771 context variables, the event's ID, and the event's timestamp. LTTng
1772 can append this sub-buffer to a trace file.
1774 An **event rule** is a set of conditions which must all be satisfied for
1775 LTTng to record an occuring event. Events still occur without
1776 satisfying event rules, but LTTng does not record them.
1781 == Components of noch:{LTTng}
1783 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1784 to call LTTng a simple _tool_ since it is composed of multiple
1785 interacting components. This section describes those components,
1786 explains their respective roles, and shows how they connect together to
1787 form the LTTng ecosystem.
1789 The following diagram shows how the most important components of LTTng
1790 interact with user applications, the Linux kernel, and you:
1793 .Control and trace data paths between LTTng components.
1794 image::plumbing.png[]
1796 The LTTng project incorporates:
1798 * **LTTng-tools**: Libraries and command-line interface to
1799 control tracing sessions.
1800 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1801 ** <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
1802 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1803 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1804 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1805 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1807 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1808 headers to instrument and trace any native user application.
1809 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1810 *** `liblttng-ust-libc-wrapper`
1811 *** `liblttng-ust-pthread-wrapper`
1812 *** `liblttng-ust-cyg-profile`
1813 *** `liblttng-ust-cyg-profile-fast`
1814 *** `liblttng-ust-dl`
1815 ** User space tracepoint provider source files generator command-line
1816 tool (man:lttng-gen-tp(1)).
1817 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1818 Java applications using `java.util.logging` or
1819 Apache log4j 1.2 logging.
1820 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1821 Python applications using the standard `logging` package.
1822 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1824 ** LTTng kernel tracer module.
1825 ** Tracing ring buffer kernel modules.
1826 ** Probe kernel modules.
1827 ** LTTng logger kernel module.
1831 === Tracing control command-line interface
1834 .The tracing control command-line interface.
1835 image::plumbing-lttng-cli.png[]
1837 The _man:lttng(1) command-line tool_ is the standard user interface to
1838 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1839 is part of LTTng-tools.
1841 The cmd:lttng tool is linked with
1842 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1843 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1845 The cmd:lttng tool has a Git-like interface:
1849 $ lttng <GENERAL OPTIONS> <COMMAND> <COMMAND OPTIONS>
1852 The <<controlling-tracing,Tracing control>> section explores the
1853 available features of LTTng using the cmd:lttng tool.
1856 [[liblttng-ctl-lttng]]
1857 === Tracing control library
1860 .The tracing control library.
1861 image::plumbing-liblttng-ctl.png[]
1863 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1864 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1865 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1867 The <<lttng-cli,cmd:lttng command-line tool>>
1868 is linked with `liblttng-ctl`.
1870 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1875 #include <lttng/lttng.h>
1878 Some objects are referenced by name (C string), such as tracing
1879 sessions, but most of them require to create a handle first using
1880 `lttng_create_handle()`.
1882 The best available developer documentation for `liblttng-ctl` is, as of
1883 LTTng{nbsp}{revision}, its installed header files. Every function and
1884 structure is thoroughly documented.
1888 === User space tracing library
1891 .The user space tracing library.
1892 image::plumbing-liblttng-ust.png[]
1894 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1895 is the LTTng user space tracer. It receives commands from a
1896 <<lttng-sessiond,session daemon>>, for example to
1897 enable and disable specific instrumentation points, and writes event
1898 records to ring buffers shared with a
1899 <<lttng-consumerd,consumer daemon>>.
1900 `liblttng-ust` is part of LTTng-UST.
1902 Public C header files are installed beside `liblttng-ust` to
1903 instrument any <<c-application,C or $$C++$$ application>>.
1905 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1906 packages, use their own library providing tracepoints which is
1907 linked with `liblttng-ust`.
1909 An application or library does not have to initialize `liblttng-ust`
1910 manually: its constructor does the necessary tasks to properly register
1911 to a session daemon. The initialization phase also enables the
1912 instrumentation points matching the <<event,event rules>> that you
1916 [[lttng-ust-agents]]
1917 === User space tracing agents
1920 .The user space tracing agents.
1921 image::plumbing-lttng-ust-agents.png[]
1923 The _LTTng-UST Java and Python agents_ are regular Java and Python
1924 packages which add LTTng tracing capabilities to the
1925 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1927 In the case of Java, the
1928 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1929 core logging facilities] and
1930 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1931 Note that Apache Log4{nbsp}2 is not supported.
1933 In the case of Python, the standard
1934 https://docs.python.org/3/library/logging.html[`logging`] package
1935 is supported. Both Python 2 and Python 3 modules can import the
1936 LTTng-UST Python agent package.
1938 The applications using the LTTng-UST agents are in the
1939 `java.util.logging` (JUL),
1940 log4j, and Python <<domain,tracing domains>>.
1942 Both agents use the same mechanism to trace the log statements. When an
1943 agent is initialized, it creates a log handler that attaches to the root
1944 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1945 When the application executes a log statement, it is passed to the
1946 agent's log handler by the root logger. The agent's log handler calls a
1947 native function in a tracepoint provider package shared library linked
1948 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1949 other fields, like its logger name and its log level. This native
1950 function contains a user space instrumentation point, hence tracing the
1953 The log level condition of an
1954 <<event,event rule>> is considered when tracing
1955 a Java or a Python application, and it's compatible with the standard
1956 JUL, log4j, and Python log levels.
1960 === LTTng kernel modules
1963 .The LTTng kernel modules.
1964 image::plumbing-lttng-modules.png[]
1966 The _LTTng kernel modules_ are a set of Linux kernel modules
1967 which implement the kernel tracer of the LTTng project. The LTTng
1968 kernel modules are part of LTTng-modules.
1970 The LTTng kernel modules include:
1972 * A set of _probe_ modules.
1974 Each module attaches to a specific subsystem
1975 of the Linux kernel using its tracepoint instrument points. There are
1976 also modules to attach to the entry and return points of the Linux
1977 system call functions.
1979 * _Ring buffer_ modules.
1981 A ring buffer implementation is provided as kernel modules. The LTTng
1982 kernel tracer writes to the ring buffer; a
1983 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1985 * The _LTTng kernel tracer_ module.
1986 * The _LTTng logger_ module.
1988 The LTTng logger module implements the special path:{/proc/lttng-logger}
1989 file so that any executable can generate LTTng events by opening and
1990 writing to this file.
1992 See <<proc-lttng-logger-abi,LTTng logger>>.
1994 Generally, you do not have to load the LTTng kernel modules manually
1995 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1996 daemon>> loads the necessary modules when starting. If you have extra
1997 probe modules, you can specify to load them to the session daemon on
2000 The LTTng kernel modules are installed in
2001 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
2002 the kernel release (see `uname --kernel-release`).
2009 .The session daemon.
2010 image::plumbing-sessiond.png[]
2012 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
2013 managing tracing sessions and for controlling the various components of
2014 LTTng. The session daemon is part of LTTng-tools.
2016 The session daemon sends control requests to and receives control
2019 * The <<lttng-ust,user space tracing library>>.
2021 Any instance of the user space tracing library first registers to
2022 a session daemon. Then, the session daemon can send requests to
2023 this instance, such as:
2026 ** Get the list of tracepoints.
2027 ** Share an <<event,event rule>> so that the user space tracing library
2028 can enable or disable tracepoints. Amongst the possible conditions
2029 of an event rule is a filter expression which `liblttng-ust` evalutes
2030 when an event occurs.
2031 ** Share <<channel,channel>> attributes and ring buffer locations.
2034 The session daemon and the user space tracing library use a Unix
2035 domain socket for their communication.
2037 * The <<lttng-ust-agents,user space tracing agents>>.
2039 Any instance of a user space tracing agent first registers to
2040 a session daemon. Then, the session daemon can send requests to
2041 this instance, such as:
2044 ** Get the list of loggers.
2045 ** Enable or disable a specific logger.
2048 The session daemon and the user space tracing agent use a TCP connection
2049 for their communication.
2051 * The <<lttng-modules,LTTng kernel tracer>>.
2052 * The <<lttng-consumerd,consumer daemon>>.
2054 The session daemon sends requests to the consumer daemon to instruct
2055 it where to send the trace data streams, amongst other information.
2057 * The <<lttng-relayd,relay daemon>>.
2059 The session daemon receives commands from the
2060 <<liblttng-ctl-lttng,tracing control library>>.
2062 The root session daemon loads the appropriate
2063 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
2064 a <<lttng-consumerd,consumer daemon>> as soon as you create
2065 an <<event,event rule>>.
2067 The session daemon does not send and receive trace data: this is the
2068 role of the <<lttng-consumerd,consumer daemon>> and
2069 <<lttng-relayd,relay daemon>>. It does, however, generate the
2070 http://diamon.org/ctf/[CTF] metadata stream.
2072 Each Unix user can have its own session daemon instance. The
2073 tracing sessions managed by different session daemons are completely
2076 The root user's session daemon is the only one which is
2077 allowed to control the LTTng kernel tracer, and its spawned consumer
2078 daemon is the only one which is allowed to consume trace data from the
2079 LTTng kernel tracer. Note, however, that any Unix user which is a member
2080 of the <<tracing-group,tracing group>> is allowed
2081 to create <<channel,channels>> in the
2082 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
2085 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
2086 session daemon when using its `create` command if none is currently
2087 running. You can also start the session daemon manually.
2094 .The consumer daemon.
2095 image::plumbing-consumerd.png[]
2097 The _consumer daemon_, cmd:lttng-consumerd, is a daemon which shares
2098 ring buffers with user applications or with the LTTng kernel modules to
2099 collect trace data and send it to some location (on disk or to a
2100 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2101 is part of LTTng-tools.
2103 You do not start a consumer daemon manually: a consumer daemon is always
2104 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2105 <<event,event rule>>, that is, before you start tracing. When you kill
2106 its owner session daemon, the consumer daemon also exits because it is
2107 the session daemon's child process. Command-line options of
2108 man:lttng-sessiond(8) target the consumer daemon process.
2110 There are up to two running consumer daemons per Unix user, whereas only
2111 one session daemon can run per user. This is because each process can be
2112 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2113 and 64-bit processes, it is more efficient to have separate
2114 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2115 exception: it can have up to _three_ running consumer daemons: 32-bit
2116 and 64-bit instances for its user applications, and one more
2117 reserved for collecting kernel trace data.
2125 image::plumbing-relayd.png[]
2127 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2128 between remote session and consumer daemons, local trace files, and a
2129 remote live trace viewer. The relay daemon is part of LTTng-tools.
2131 The main purpose of the relay daemon is to implement a receiver of
2132 <<sending-trace-data-over-the-network,trace data over the network>>.
2133 This is useful when the target system does not have much file system
2134 space to record trace files locally.
2136 The relay daemon is also a server to which a
2137 <<lttng-live,live trace viewer>> can
2138 connect. The live trace viewer sends requests to the relay daemon to
2139 receive trace data as the target system emits events. The
2140 communication protocol is named _LTTng live_; it is used over TCP
2143 Note that you can start the relay daemon on the target system directly.
2144 This is the setup of choice when the use case is to view events as
2145 the target system emits them without the need of a remote system.
2149 == [[using-lttng]]Instrumentation
2151 There are many examples of tracing and monitoring in our everyday life:
2153 * You have access to real-time and historical weather reports and
2154 forecasts thanks to weather stations installed around the country.
2155 * You know your heart is safe thanks to an electrocardiogram.
2156 * You make sure not to drive your car too fast and to have enough fuel
2157 to reach your destination thanks to gauges visible on your dashboard.
2159 All the previous examples have something in common: they rely on
2160 **instruments**. Without the electrodes attached to the surface of your
2161 body's skin, cardiac monitoring is futile.
2163 LTTng, as a tracer, is no different from those real life examples. If
2164 you're about to trace a software system or, in other words, record its
2165 history of execution, you better have **instrumentation points** in the
2166 subject you're tracing, that is, the actual software.
2168 Various ways were developed to instrument a piece of software for LTTng
2169 tracing. The most straightforward one is to manually place
2170 instrumentation points, called _tracepoints_, in the software's source
2171 code. It is also possible to add instrumentation points dynamically in
2172 the Linux kernel <<domain,tracing domain>>.
2174 If you're only interested in tracing the Linux kernel, your
2175 instrumentation needs are probably already covered by LTTng's built-in
2176 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2177 user application which is already instrumented for LTTng tracing.
2178 In such cases, you can skip this whole section and read the topics of
2179 the <<controlling-tracing,Tracing control>> section.
2181 Many methods are available to instrument a piece of software for LTTng
2184 * <<c-application,User space instrumentation for C and $$C++$$
2186 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2187 * <<java-application,User space Java agent>>.
2188 * <<python-application,User space Python agent>>.
2189 * <<proc-lttng-logger-abi,LTTng logger>>.
2190 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2194 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2196 The procedure to instrument a C or $$C++$$ user application with
2197 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2199 . <<tracepoint-provider,Create the source files of a tracepoint provider
2201 . <<probing-the-application-source-code,Add tracepoints to
2202 the application's source code>>.
2203 . <<building-tracepoint-providers-and-user-application,Build and link
2204 a tracepoint provider package and the user application>>.
2206 If you need quick, man:printf(3)-like instrumentation, you can skip
2207 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2210 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2211 instrument a user application with `liblttng-ust`.
2214 [[tracepoint-provider]]
2215 ==== Create the source files of a tracepoint provider package
2217 A _tracepoint provider_ is a set of compiled functions which provide
2218 **tracepoints** to an application, the type of instrumentation point
2219 supported by LTTng-UST. Those functions can emit events with
2220 user-defined fields and serialize those events as event records to one
2221 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2222 macro, which you <<probing-the-application-source-code,insert in a user
2223 application's source code>>, calls those functions.
2225 A _tracepoint provider package_ is an object file (`.o`) or a shared
2226 library (`.so`) which contains one or more tracepoint providers.
2227 Its source files are:
2229 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2230 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2232 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2233 the LTTng user space tracer, at run time.
2236 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2237 image::ust-app.png[]
2239 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2240 skip creating and using a tracepoint provider and use
2241 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2245 ===== Create a tracepoint provider header file template
2247 A _tracepoint provider header file_ contains the tracepoint
2248 definitions of a tracepoint provider.
2250 To create a tracepoint provider header file:
2252 . Start from this template:
2256 .Tracepoint provider header file template (`.h` file extension).
2258 #undef TRACEPOINT_PROVIDER
2259 #define TRACEPOINT_PROVIDER provider_name
2261 #undef TRACEPOINT_INCLUDE
2262 #define TRACEPOINT_INCLUDE "./tp.h"
2264 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2267 #include <lttng/tracepoint.h>
2270 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2271 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2276 #include <lttng/tracepoint-event.h>
2282 * `provider_name` with the name of your tracepoint provider.
2283 * `"tp.h"` with the name of your tracepoint provider header file.
2285 . Below the `#include <lttng/tracepoint.h>` line, put your
2286 <<defining-tracepoints,tracepoint definitions>>.
2288 Your tracepoint provider name must be unique amongst all the possible
2289 tracepoint provider names used on the same target system. We
2290 suggest to include the name of your project or company in the name,
2291 for example, `org_lttng_my_project_tpp`.
2293 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2294 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2295 write are the <<defining-tracepoints,tracepoint definitions>>.
2298 [[defining-tracepoints]]
2299 ===== Create a tracepoint definition
2301 A _tracepoint definition_ defines, for a given tracepoint:
2303 * Its **input arguments**. They are the macro parameters that the
2304 `tracepoint()` macro accepts for this particular tracepoint
2305 in the user application's source code.
2306 * Its **output event fields**. They are the sources of event fields
2307 that form the payload of any event that the execution of the
2308 `tracepoint()` macro emits for this particular tracepoint.
2310 You can create a tracepoint definition by using the
2311 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2313 <<tpp-header,tracepoint provider header file template>>.
2315 The syntax of the `TRACEPOINT_EVENT()` macro is:
2318 .`TRACEPOINT_EVENT()` macro syntax.
2321 /* Tracepoint provider name */
2324 /* Tracepoint name */
2327 /* Input arguments */
2332 /* Output event fields */
2341 * `provider_name` with your tracepoint provider name.
2342 * `tracepoint_name` with your tracepoint name.
2343 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2344 * `fields` with the <<tpp-def-output-fields,output event field>>
2347 This tracepoint emits events named `provider_name:tracepoint_name`.
2350 .Event name's length limitation
2352 The concatenation of the tracepoint provider name and the
2353 tracepoint name must not exceed **254 characters**. If it does, the
2354 instrumented application compiles and runs, but LTTng throws multiple
2355 warnings and you could experience serious issues.
2358 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2361 .`TP_ARGS()` macro syntax.
2370 * `type` with the C type of the argument.
2371 * `arg_name` with the argument name.
2373 You can repeat `type` and `arg_name` up to 10 times to have
2374 more than one argument.
2376 .`TP_ARGS()` usage with three arguments.
2388 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2389 tracepoint definition with no input arguments.
2391 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2392 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2393 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2394 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2397 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2398 C expression that the tracer evalutes at the `tracepoint()` macro site
2399 in the application's source code. This expression provides a field's
2400 source of data. The argument expression can include input argument names
2401 listed in the `TP_ARGS()` macro.
2403 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2404 must be unique within a given tracepoint definition.
2406 Here's a complete tracepoint definition example:
2408 .Tracepoint definition.
2410 The following tracepoint definition defines a tracepoint which takes
2411 three input arguments and has four output event fields.
2415 #include "my-custom-structure.h"
2421 const struct my_custom_structure*, my_custom_structure,
2426 ctf_string(query_field, query)
2427 ctf_float(double, ratio_field, ratio)
2428 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2429 ctf_integer(int, send_size, my_custom_structure->send_size)
2434 You can refer to this tracepoint definition with the `tracepoint()`
2435 macro in your application's source code like this:
2439 tracepoint(my_provider, my_tracepoint,
2440 my_structure, some_ratio, the_query);
2444 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2445 if they satisfy an enabled <<event,event rule>>.
2448 [[using-tracepoint-classes]]
2449 ===== Use a tracepoint class
2451 A _tracepoint class_ is a class of tracepoints which share the same
2452 output event field definitions. A _tracepoint instance_ is one
2453 instance of such a defined tracepoint class, with its own tracepoint
2456 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2457 shorthand which defines both a tracepoint class and a tracepoint
2458 instance at the same time.
2460 When you build a tracepoint provider package, the C or $$C++$$ compiler
2461 creates one serialization function for each **tracepoint class**. A
2462 serialization function is responsible for serializing the event fields
2463 of a tracepoint to a sub-buffer when tracing.
2465 For various performance reasons, when your situation requires multiple
2466 tracepoint definitions with different names, but with the same event
2467 fields, we recommend that you manually create a tracepoint class
2468 and instantiate as many tracepoint instances as needed. One positive
2469 effect of such a design, amongst other advantages, is that all
2470 tracepoint instances of the same tracepoint class reuse the same
2471 serialization function, thus reducing
2472 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2474 .Use a tracepoint class and tracepoint instances.
2476 Consider the following three tracepoint definitions:
2488 ctf_integer(int, userid, userid)
2489 ctf_integer(size_t, len, len)
2501 ctf_integer(int, userid, userid)
2502 ctf_integer(size_t, len, len)
2514 ctf_integer(int, userid, userid)
2515 ctf_integer(size_t, len, len)
2520 In this case, we create three tracepoint classes, with one implicit
2521 tracepoint instance for each of them: `get_account`, `get_settings`, and
2522 `get_transaction`. However, they all share the same event field names
2523 and types. Hence three identical, yet independent serialization
2524 functions are created when you build the tracepoint provider package.
2526 A better design choice is to define a single tracepoint class and three
2527 tracepoint instances:
2531 /* The tracepoint class */
2532 TRACEPOINT_EVENT_CLASS(
2533 /* Tracepoint provider name */
2536 /* Tracepoint class name */
2539 /* Input arguments */
2545 /* Output event fields */
2547 ctf_integer(int, userid, userid)
2548 ctf_integer(size_t, len, len)
2552 /* The tracepoint instances */
2553 TRACEPOINT_EVENT_INSTANCE(
2554 /* Tracepoint provider name */
2557 /* Tracepoint class name */
2560 /* Tracepoint name */
2563 /* Input arguments */
2569 TRACEPOINT_EVENT_INSTANCE(
2578 TRACEPOINT_EVENT_INSTANCE(
2591 [[assigning-log-levels]]
2592 ===== Assign a log level to a tracepoint definition
2594 You can assign an optional _log level_ to a
2595 <<defining-tracepoints,tracepoint definition>>.
2597 Assigning different levels of severity to tracepoint definitions can
2598 be useful: when you <<enabling-disabling-events,create an event rule>>,
2599 you can target tracepoints having a log level as severe as a specific
2602 The concept of LTTng-UST log levels is similar to the levels found
2603 in typical logging frameworks:
2605 * In a logging framework, the log level is given by the function
2606 or method name you use at the log statement site: `debug()`,
2607 `info()`, `warn()`, `error()`, and so on.
2608 * In LTTng-UST, you statically assign the log level to a tracepoint
2609 definition; any `tracepoint()` macro invocation which refers to
2610 this definition has this log level.
2612 You can assign a log level to a tracepoint definition with the
2613 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2614 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2615 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2618 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2621 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2623 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2628 * `provider_name` with the tracepoint provider name.
2629 * `tracepoint_name` with the tracepoint name.
2630 * `log_level` with the log level to assign to the tracepoint
2631 definition named `tracepoint_name` in the `provider_name`
2632 tracepoint provider.
2634 See man:lttng-ust(3) for a list of available log level names.
2636 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2640 /* Tracepoint definition */
2649 ctf_integer(int, userid, userid)
2650 ctf_integer(size_t, len, len)
2654 /* Log level assignment */
2655 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2661 ===== Create a tracepoint provider package source file
2663 A _tracepoint provider package source file_ is a C source file which
2664 includes a <<tpp-header,tracepoint provider header file>> to expand its
2665 macros into event serialization and other functions.
2667 You can always use the following tracepoint provider package source
2671 .Tracepoint provider package source file template.
2673 #define TRACEPOINT_CREATE_PROBES
2678 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2679 header file>> name. You may also include more than one tracepoint
2680 provider header file here to create a tracepoint provider package
2681 holding more than one tracepoint providers.
2684 [[probing-the-application-source-code]]
2685 ==== Add tracepoints to an application's source code
2687 Once you <<tpp-header,create a tracepoint provider header file>>, you
2688 can use the `tracepoint()` macro in your application's
2689 source code to insert the tracepoints that this header
2690 <<defining-tracepoints,defines>>.
2692 The `tracepoint()` macro takes at least two parameters: the tracepoint
2693 provider name and the tracepoint name. The corresponding tracepoint
2694 definition defines the other parameters.
2696 .`tracepoint()` usage.
2698 The following <<defining-tracepoints,tracepoint definition>> defines a
2699 tracepoint which takes two input arguments and has two output event
2703 .Tracepoint provider header file.
2705 #include "my-custom-structure.h"
2712 const char*, cmd_name
2715 ctf_string(cmd_name, cmd_name)
2716 ctf_integer(int, number_of_args, argc)
2721 You can refer to this tracepoint definition with the `tracepoint()`
2722 macro in your application's source code like this:
2725 .Application's source file.
2729 int main(int argc, char* argv[])
2731 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2737 Note how the application's source code includes
2738 the tracepoint provider header file containing the tracepoint
2739 definitions to use, path:{tp.h}.
2742 .`tracepoint()` usage with a complex tracepoint definition.
2744 Consider this complex tracepoint definition, where multiple event
2745 fields refer to the same input arguments in their argument expression
2749 .Tracepoint provider header file.
2751 /* For `struct stat` */
2752 #include <sys/types.h>
2753 #include <sys/stat.h>
2765 ctf_integer(int, my_constant_field, 23 + 17)
2766 ctf_integer(int, my_int_arg_field, my_int_arg)
2767 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2768 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2769 my_str_arg[2] + my_str_arg[3])
2770 ctf_string(my_str_arg_field, my_str_arg)
2771 ctf_integer_hex(off_t, size_field, st->st_size)
2772 ctf_float(double, size_dbl_field, (double) st->st_size)
2773 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2774 size_t, strlen(my_str_arg) / 2)
2779 You can refer to this tracepoint definition with the `tracepoint()`
2780 macro in your application's source code like this:
2783 .Application's source file.
2785 #define TRACEPOINT_DEFINE
2792 stat("/etc/fstab", &s);
2793 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2799 If you look at the event record that LTTng writes when tracing this
2800 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2801 it should look like this:
2803 .Event record fields
2805 |Field's name |Field's value
2806 |`my_constant_field` |40
2807 |`my_int_arg_field` |23
2808 |`my_int_arg_field2` |529
2810 |`my_str_arg_field` |`Hello, World!`
2811 |`size_field` |0x12d
2812 |`size_dbl_field` |301.0
2813 |`half_my_str_arg_field` |`Hello,`
2817 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2818 compute--they use the call stack, for example. To avoid this
2819 computation when the tracepoint is disabled, you can use the
2820 `tracepoint_enabled()` and `do_tracepoint()` macros.
2822 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2826 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2828 tracepoint_enabled(provider_name, tracepoint_name)
2829 do_tracepoint(provider_name, tracepoint_name, ...)
2834 * `provider_name` with the tracepoint provider name.
2835 * `tracepoint_name` with the tracepoint name.
2837 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2838 `tracepoint_name` from the provider named `provider_name` is enabled
2841 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2842 if the tracepoint is enabled. Using `tracepoint()` with
2843 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2844 the `tracepoint_enabled()` check, thus a race condition is
2845 possible in this situation:
2848 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2850 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2851 stuff = prepare_stuff();
2854 tracepoint(my_provider, my_tracepoint, stuff);
2857 If the tracepoint is enabled after the condition, then `stuff` is not
2858 prepared: the emitted event will either contain wrong data, or the whole
2859 application could crash (segmentation fault, for example).
2861 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2862 `STAP_PROBEV()` call. If you need it, you must emit
2866 [[building-tracepoint-providers-and-user-application]]
2867 ==== Build and link a tracepoint provider package and an application
2869 Once you have one or more <<tpp-header,tracepoint provider header
2870 files>> and a <<tpp-source,tracepoint provider package source file>>,
2871 you can create the tracepoint provider package by compiling its source
2872 file. From here, multiple build and run scenarios are possible. The
2873 following table shows common application and library configurations
2874 along with the required command lines to achieve them.
2876 In the following diagrams, we use the following file names:
2879 Executable application.
2882 Application's object file.
2885 Tracepoint provider package object file.
2888 Tracepoint provider package archive file.
2891 Tracepoint provider package shared object file.
2894 User library object file.
2897 User library shared object file.
2899 We use the following symbols in the diagrams of table below:
2902 .Symbols used in the build scenario diagrams.
2903 image::ust-sit-symbols.png[]
2905 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2906 variable in the following instructions.
2908 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2909 .Common tracepoint provider package scenarios.
2911 |Scenario |Instructions
2914 The instrumented application is statically linked with
2915 the tracepoint provider package object.
2917 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2920 include::../common/ust-sit-step-tp-o.txt[]
2922 To build the instrumented application:
2924 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2929 #define TRACEPOINT_DEFINE
2933 . Compile the application source file:
2942 . Build the application:
2947 $ gcc -o app app.o tpp.o -llttng-ust -ldl
2951 To run the instrumented application:
2953 * Start the application:
2963 The instrumented application is statically linked with the
2964 tracepoint provider package archive file.
2966 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2969 To create the tracepoint provider package archive file:
2971 . Compile the <<tpp-source,tracepoint provider package source file>>:
2980 . Create the tracepoint provider package archive file:
2985 $ ar rcs tpp.a tpp.o
2989 To build the instrumented application:
2991 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2996 #define TRACEPOINT_DEFINE
3000 . Compile the application source file:
3009 . Build the application:
3014 $ gcc -o app app.o tpp.a -llttng-ust -ldl
3018 To run the instrumented application:
3020 * Start the application:
3030 The instrumented application is linked with the tracepoint provider
3031 package shared object.
3033 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
3036 include::../common/ust-sit-step-tp-so.txt[]
3038 To build the instrumented application:
3040 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3045 #define TRACEPOINT_DEFINE
3049 . Compile the application source file:
3058 . Build the application:
3063 $ gcc -o app app.o -ldl -L. -ltpp
3067 To run the instrumented application:
3069 * Start the application:
3079 The tracepoint provider package shared object is preloaded before the
3080 instrumented application starts.
3082 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
3085 include::../common/ust-sit-step-tp-so.txt[]
3087 To build the instrumented application:
3089 . In path:{app.c}, before including path:{tpp.h}, add the
3095 #define TRACEPOINT_DEFINE
3096 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3100 . Compile the application source file:
3109 . Build the application:
3114 $ gcc -o app app.o -ldl
3118 To run the instrumented application with tracing support:
3120 * Preload the tracepoint provider package shared object and
3121 start the application:
3126 $ LD_PRELOAD=./libtpp.so ./app
3130 To run the instrumented application without tracing support:
3132 * Start the application:
3142 The instrumented application dynamically loads the tracepoint provider
3143 package shared object.
3145 See the <<dlclose-warning,warning about `dlclose()`>>.
3147 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3150 include::../common/ust-sit-step-tp-so.txt[]
3152 To build the instrumented application:
3154 . In path:{app.c}, before including path:{tpp.h}, add the
3160 #define TRACEPOINT_DEFINE
3161 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3165 . Compile the application source file:
3174 . Build the application:
3179 $ gcc -o app app.o -ldl
3183 To run the instrumented application:
3185 * Start the application:
3195 The application is linked with the instrumented user library.
3197 The instrumented user library is statically linked with the tracepoint
3198 provider package object file.
3200 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3203 include::../common/ust-sit-step-tp-o-fpic.txt[]
3205 To build the instrumented user library:
3207 . In path:{emon.c}, before including path:{tpp.h}, add the
3213 #define TRACEPOINT_DEFINE
3217 . Compile the user library source file:
3222 $ gcc -I. -fpic -c emon.c
3226 . Build the user library shared object:
3231 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3235 To build the application:
3237 . Compile the application source file:
3246 . Build the application:
3251 $ gcc -o app app.o -L. -lemon
3255 To run the application:
3257 * Start the application:
3267 The application is linked with the instrumented user library.
3269 The instrumented user library is linked with the tracepoint provider
3270 package shared object.
3272 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3275 include::../common/ust-sit-step-tp-so.txt[]
3277 To build the instrumented user library:
3279 . In path:{emon.c}, before including path:{tpp.h}, add the
3285 #define TRACEPOINT_DEFINE
3289 . Compile the user library source file:
3294 $ gcc -I. -fpic -c emon.c
3298 . Build the user library shared object:
3303 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3307 To build the application:
3309 . Compile the application source file:
3318 . Build the application:
3323 $ gcc -o app app.o -L. -lemon
3327 To run the application:
3329 * Start the application:
3339 The tracepoint provider package shared object is preloaded before the
3342 The application is linked with the instrumented user library.
3344 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3347 include::../common/ust-sit-step-tp-so.txt[]
3349 To build the instrumented user library:
3351 . In path:{emon.c}, before including path:{tpp.h}, add the
3357 #define TRACEPOINT_DEFINE
3358 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3362 . Compile the user library source file:
3367 $ gcc -I. -fpic -c emon.c
3371 . Build the user library shared object:
3376 $ gcc -shared -o libemon.so emon.o -ldl
3380 To build the application:
3382 . Compile the application source file:
3391 . Build the application:
3396 $ gcc -o app app.o -L. -lemon
3400 To run the application with tracing support:
3402 * Preload the tracepoint provider package shared object and
3403 start the application:
3408 $ LD_PRELOAD=./libtpp.so ./app
3412 To run the application without tracing support:
3414 * Start the application:
3424 The application is linked with the instrumented user library.
3426 The instrumented user library dynamically loads the tracepoint provider
3427 package shared object.
3429 See the <<dlclose-warning,warning about `dlclose()`>>.
3431 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3434 include::../common/ust-sit-step-tp-so.txt[]
3436 To build the instrumented user library:
3438 . In path:{emon.c}, before including path:{tpp.h}, add the
3444 #define TRACEPOINT_DEFINE
3445 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3449 . Compile the user library source file:
3454 $ gcc -I. -fpic -c emon.c
3458 . Build the user library shared object:
3463 $ gcc -shared -o libemon.so emon.o -ldl
3467 To build the application:
3469 . Compile the application source file:
3478 . Build the application:
3483 $ gcc -o app app.o -L. -lemon
3487 To run the application:
3489 * Start the application:
3499 The application dynamically loads the instrumented user library.
3501 The instrumented user library is linked with the tracepoint provider
3502 package shared object.
3504 See the <<dlclose-warning,warning about `dlclose()`>>.
3506 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3509 include::../common/ust-sit-step-tp-so.txt[]
3511 To build the instrumented user library:
3513 . In path:{emon.c}, before including path:{tpp.h}, add the
3519 #define TRACEPOINT_DEFINE
3523 . Compile the user library source file:
3528 $ gcc -I. -fpic -c emon.c
3532 . Build the user library shared object:
3537 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3541 To build the application:
3543 . Compile the application source file:
3552 . Build the application:
3557 $ gcc -o app app.o -ldl -L. -lemon
3561 To run the application:
3563 * Start the application:
3573 The application dynamically loads the instrumented user library.
3575 The instrumented user library dynamically loads the tracepoint provider
3576 package shared object.
3578 See the <<dlclose-warning,warning about `dlclose()`>>.
3580 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3583 include::../common/ust-sit-step-tp-so.txt[]
3585 To build the instrumented user library:
3587 . In path:{emon.c}, before including path:{tpp.h}, add the
3593 #define TRACEPOINT_DEFINE
3594 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3598 . Compile the user library source file:
3603 $ gcc -I. -fpic -c emon.c
3607 . Build the user library shared object:
3612 $ gcc -shared -o libemon.so emon.o -ldl
3616 To build the application:
3618 . Compile the application source file:
3627 . Build the application:
3632 $ gcc -o app app.o -ldl -L. -lemon
3636 To run the application:
3638 * Start the application:
3648 The tracepoint provider package shared object is preloaded before the
3651 The application dynamically loads the instrumented user library.
3653 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3656 include::../common/ust-sit-step-tp-so.txt[]
3658 To build the instrumented user library:
3660 . In path:{emon.c}, before including path:{tpp.h}, add the
3666 #define TRACEPOINT_DEFINE
3667 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3671 . Compile the user library source file:
3676 $ gcc -I. -fpic -c emon.c
3680 . Build the user library shared object:
3685 $ gcc -shared -o libemon.so emon.o -ldl
3689 To build the application:
3691 . Compile the application source file:
3700 . Build the application:
3705 $ gcc -o app app.o -L. -lemon
3709 To run the application with tracing support:
3711 * Preload the tracepoint provider package shared object and
3712 start the application:
3717 $ LD_PRELOAD=./libtpp.so ./app
3721 To run the application without tracing support:
3723 * Start the application:
3733 The application is statically linked with the tracepoint provider
3734 package object file.
3736 The application is linked with the instrumented user library.
3738 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3741 include::../common/ust-sit-step-tp-o.txt[]
3743 To build the instrumented user library:
3745 . In path:{emon.c}, before including path:{tpp.h}, add the
3751 #define TRACEPOINT_DEFINE
3755 . Compile the user library source file:
3760 $ gcc -I. -fpic -c emon.c
3764 . Build the user library shared object:
3769 $ gcc -shared -o libemon.so emon.o
3773 To build the application:
3775 . Compile the application source file:
3784 . Build the application:
3789 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3793 To run the instrumented application:
3795 * Start the application:
3805 The application is statically linked with the tracepoint provider
3806 package object file.
3808 The application dynamically loads the instrumented user library.
3810 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3813 include::../common/ust-sit-step-tp-o.txt[]
3815 To build the application:
3817 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3822 #define TRACEPOINT_DEFINE
3826 . Compile the application source file:
3835 . Build the application:
3840 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
3845 The `--export-dynamic` option passed to the linker is necessary for the
3846 dynamically loaded library to ``see'' the tracepoint symbols defined in
3849 To build the instrumented user library:
3851 . Compile the user library source file:
3856 $ gcc -I. -fpic -c emon.c
3860 . Build the user library shared object:
3865 $ gcc -shared -o libemon.so emon.o
3869 To run the application:
3871 * Start the application:
3883 .Do not use man:dlclose(3) on a tracepoint provider package
3885 Never use man:dlclose(3) on any shared object which:
3887 * Is linked with, statically or dynamically, a tracepoint provider
3889 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3890 package shared object.
3892 This is currently considered **unsafe** due to a lack of reference
3893 counting from LTTng-UST to the shared object.
3895 A known workaround (available since glibc 2.2) is to use the
3896 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3897 effect of not unloading the loaded shared object, even if man:dlclose(3)
3900 You can also preload the tracepoint provider package shared object with
3901 the env:LD_PRELOAD environment variable to overcome this limitation.
3905 [[using-lttng-ust-with-daemons]]
3906 ===== Use noch:{LTTng-UST} with daemons
3908 If your instrumented application calls man:fork(2), man:clone(2),
3909 or BSD's man:rfork(2), without a following man:exec(3)-family
3910 system call, you must preload the path:{liblttng-ust-fork.so} shared
3911 object when you start the application.
3915 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
3918 If your tracepoint provider package is
3919 a shared library which you also preload, you must put both
3920 shared objects in env:LD_PRELOAD:
3924 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3930 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3932 If your instrumented application closes one or more file descriptors
3933 which it did not open itself, you must preload the
3934 path:{liblttng-ust-fd.so} shared object when you start the application:
3938 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
3941 Typical use cases include closing all the file descriptors after
3942 man:fork(2) or man:rfork(2) and buggy applications doing
3946 [[lttng-ust-pkg-config]]
3947 ===== Use noch:{pkg-config}
3949 On some distributions, LTTng-UST ships with a
3950 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3951 metadata file. If this is your case, then you can use cmd:pkg-config to
3952 build an application on the command line:
3956 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3960 [[instrumenting-32-bit-app-on-64-bit-system]]
3961 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3963 In order to trace a 32-bit application running on a 64-bit system,
3964 LTTng must use a dedicated 32-bit
3965 <<lttng-consumerd,consumer daemon>>.
3967 The following steps show how to build and install a 32-bit consumer
3968 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3969 build and install the 32-bit LTTng-UST libraries, and how to build and
3970 link an instrumented 32-bit application in that context.
3972 To build a 32-bit instrumented application for a 64-bit target system,
3973 assuming you have a fresh target system with no installed Userspace RCU
3976 . Download, build, and install a 32-bit version of Userspace RCU:
3981 $ cd $(mktemp -d) &&
3982 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3983 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3984 cd userspace-rcu-0.9.* &&
3985 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3987 sudo make install &&
3992 . Using your distribution's package manager, or from source, install
3993 the following 32-bit versions of the following dependencies of
3994 LTTng-tools and LTTng-UST:
3997 * https://sourceforge.net/projects/libuuid/[libuuid]
3998 * http://directory.fsf.org/wiki/Popt[popt]
3999 * http://www.xmlsoft.org/[libxml2]
4002 . Download, build, and install a 32-bit version of the latest
4003 LTTng-UST{nbsp}{revision}:
4008 $ cd $(mktemp -d) &&
4009 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.10.tar.bz2 &&
4010 tar -xf lttng-ust-latest-2.10.tar.bz2 &&
4011 cd lttng-ust-2.10.* &&
4012 ./configure --libdir=/usr/local/lib32 \
4013 CFLAGS=-m32 CXXFLAGS=-m32 \
4014 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
4016 sudo make install &&
4023 Depending on your distribution,
4024 32-bit libraries could be installed at a different location than
4025 `/usr/lib32`. For example, Debian is known to install
4026 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
4028 In this case, make sure to set `LDFLAGS` to all the
4029 relevant 32-bit library paths, for example:
4033 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
4037 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
4038 the 32-bit consumer daemon:
4043 $ cd $(mktemp -d) &&
4044 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.10.tar.bz2 &&
4045 tar -xf lttng-tools-latest-2.10.tar.bz2 &&
4046 cd lttng-tools-2.10.* &&
4047 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
4048 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
4049 --disable-bin-lttng --disable-bin-lttng-crash \
4050 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
4052 cd src/bin/lttng-consumerd &&
4053 sudo make install &&
4058 . From your distribution or from source,
4059 <<installing-lttng,install>> the 64-bit versions of
4060 LTTng-UST and Userspace RCU.
4061 . Download, build, and install the 64-bit version of the
4062 latest LTTng-tools{nbsp}{revision}:
4067 $ cd $(mktemp -d) &&
4068 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.10.tar.bz2 &&
4069 tar -xf lttng-tools-latest-2.10.tar.bz2 &&
4070 cd lttng-tools-2.10.* &&
4071 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
4072 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
4074 sudo make install &&
4079 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
4080 when linking your 32-bit application:
4083 -m32 -L/usr/lib32 -L/usr/local/lib32 \
4084 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
4087 For example, let's rebuild the quick start example in
4088 <<tracing-your-own-user-application,Trace a user application>> as an
4089 instrumented 32-bit application:
4094 $ gcc -m32 -c -I. hello-tp.c
4095 $ gcc -m32 -c hello.c
4096 $ gcc -m32 -o hello hello.o hello-tp.o \
4097 -L/usr/lib32 -L/usr/local/lib32 \
4098 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
4103 No special action is required to execute the 32-bit application and
4104 to trace it: use the command-line man:lttng(1) tool as usual.
4111 man:tracef(3) is a small LTTng-UST API designed for quick,
4112 man:printf(3)-like instrumentation without the burden of
4113 <<tracepoint-provider,creating>> and
4114 <<building-tracepoint-providers-and-user-application,building>>
4115 a tracepoint provider package.
4117 To use `tracef()` in your application:
4119 . In the C or C++ source files where you need to use `tracef()`,
4120 include `<lttng/tracef.h>`:
4125 #include <lttng/tracef.h>
4129 . In the application's source code, use `tracef()` like you would use
4137 tracef("my message: %d (%s)", my_integer, my_string);
4143 . Link your application with `liblttng-ust`:
4148 $ gcc -o app app.c -llttng-ust
4152 To trace the events that `tracef()` calls emit:
4154 * <<enabling-disabling-events,Create an event rule>> which matches the
4155 `lttng_ust_tracef:*` event name:
4160 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
4165 .Limitations of `tracef()`
4167 The `tracef()` utility function was developed to make user space tracing
4168 super simple, albeit with notable disadvantages compared to
4169 <<defining-tracepoints,user-defined tracepoints>>:
4171 * All the emitted events have the same tracepoint provider and
4172 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4173 * There is no static type checking.
4174 * The only event record field you actually get, named `msg`, is a string
4175 potentially containing the values you passed to `tracef()`
4176 using your own format string. This also means that you cannot filter
4177 events with a custom expression at run time because there are no
4179 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4180 function behind the scenes to format the strings at run time, its
4181 expected performance is lower than with user-defined tracepoints,
4182 which do not require a conversion to a string.
4184 Taking this into consideration, `tracef()` is useful for some quick
4185 prototyping and debugging, but you should not consider it for any
4186 permanent and serious applicative instrumentation.
4192 ==== Use `tracelog()`
4194 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4195 the difference that it accepts an additional log level parameter.
4197 The goal of `tracelog()` is to ease the migration from logging to
4200 To use `tracelog()` in your application:
4202 . In the C or C++ source files where you need to use `tracelog()`,
4203 include `<lttng/tracelog.h>`:
4208 #include <lttng/tracelog.h>
4212 . In the application's source code, use `tracelog()` like you would use
4213 man:printf(3), except for the first parameter which is the log
4221 tracelog(TRACE_WARNING, "my message: %d (%s)",
4222 my_integer, my_string);
4228 See man:lttng-ust(3) for a list of available log level names.
4230 . Link your application with `liblttng-ust`:
4235 $ gcc -o app app.c -llttng-ust
4239 To trace the events that `tracelog()` calls emit with a log level
4240 _as severe as_ a specific log level:
4242 * <<enabling-disabling-events,Create an event rule>> which matches the
4243 `lttng_ust_tracelog:*` event name and a minimum level
4249 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4250 --loglevel=TRACE_WARNING
4254 To trace the events that `tracelog()` calls emit with a
4255 _specific log level_:
4257 * Create an event rule which matches the `lttng_ust_tracelog:*`
4258 event name and a specific log level:
4263 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4264 --loglevel-only=TRACE_INFO
4269 [[prebuilt-ust-helpers]]
4270 === Prebuilt user space tracing helpers
4272 The LTTng-UST package provides a few helpers in the form or preloadable
4273 shared objects which automatically instrument system functions and
4276 The helper shared objects are normally found in dir:{/usr/lib}. If you
4277 built LTTng-UST <<building-from-source,from source>>, they are probably
4278 located in dir:{/usr/local/lib}.
4280 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4283 path:{liblttng-ust-libc-wrapper.so}::
4284 path:{liblttng-ust-pthread-wrapper.so}::
4285 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4286 memory and POSIX threads function tracing>>.
4288 path:{liblttng-ust-cyg-profile.so}::
4289 path:{liblttng-ust-cyg-profile-fast.so}::
4290 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4292 path:{liblttng-ust-dl.so}::
4293 <<liblttng-ust-dl,Dynamic linker tracing>>.
4295 To use a user space tracing helper with any user application:
4297 * Preload the helper shared object when you start the application:
4302 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4306 You can preload more than one helper:
4311 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4317 [[liblttng-ust-libc-pthread-wrapper]]
4318 ==== Instrument C standard library memory and POSIX threads functions
4320 The path:{liblttng-ust-libc-wrapper.so} and
4321 path:{liblttng-ust-pthread-wrapper.so} helpers
4322 add instrumentation to some C standard library and POSIX
4326 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4328 |TP provider name |TP name |Instrumented function
4330 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4331 |`calloc` |man:calloc(3)
4332 |`realloc` |man:realloc(3)
4333 |`free` |man:free(3)
4334 |`memalign` |man:memalign(3)
4335 |`posix_memalign` |man:posix_memalign(3)
4339 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4341 |TP provider name |TP name |Instrumented function
4343 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4344 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4345 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4346 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4349 When you preload the shared object, it replaces the functions listed
4350 in the previous tables by wrappers which contain tracepoints and call
4351 the replaced functions.
4354 [[liblttng-ust-cyg-profile]]
4355 ==== Instrument function entry and exit
4357 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4358 to the entry and exit points of functions.
4360 man:gcc(1) and man:clang(1) have an option named
4361 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4362 which generates instrumentation calls for entry and exit to functions.
4363 The LTTng-UST function tracing helpers,
4364 path:{liblttng-ust-cyg-profile.so} and
4365 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4366 to add tracepoints to the two generated functions (which contain
4367 `cyg_profile` in their names, hence the helper's name).
4369 To use the LTTng-UST function tracing helper, the source files to
4370 instrument must be built using the `-finstrument-functions` compiler
4373 There are two versions of the LTTng-UST function tracing helper:
4375 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4376 that you should only use when it can be _guaranteed_ that the
4377 complete event stream is recorded without any lost event record.
4378 Any kind of duplicate information is left out.
4380 Assuming no event record is lost, having only the function addresses on
4381 entry is enough to create a call graph, since an event record always
4382 contains the ID of the CPU that generated it.
4384 You can use a tool like man:addr2line(1) to convert function addresses
4385 back to source file names and line numbers.
4387 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4388 which also works in use cases where event records might get discarded or
4389 not recorded from application startup.
4390 In these cases, the trace analyzer needs more information to be
4391 able to reconstruct the program flow.
4393 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4394 points of this helper.
4396 All the tracepoints that this helper provides have the
4397 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4399 TIP: It's sometimes a good idea to limit the number of source files that
4400 you compile with the `-finstrument-functions` option to prevent LTTng
4401 from writing an excessive amount of trace data at run time. When using
4402 man:gcc(1), you can use the
4403 `-finstrument-functions-exclude-function-list` option to avoid
4404 instrument entries and exits of specific function names.
4409 ==== Instrument the dynamic linker
4411 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4412 man:dlopen(3) and man:dlclose(3) function calls.
4414 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4419 [[java-application]]
4420 === User space Java agent
4422 You can instrument any Java application which uses one of the following
4425 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4426 (JUL) core logging facilities.
4427 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4428 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4431 .LTTng-UST Java agent imported by a Java application.
4432 image::java-app.png[]
4434 Note that the methods described below are new in LTTng{nbsp}2.8.
4435 Previous LTTng versions use another technique.
4437 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4438 and https://ci.lttng.org/[continuous integration], thus this version is
4439 directly supported. However, the LTTng-UST Java agent is also tested
4440 with OpenJDK{nbsp}7.
4445 ==== Use the LTTng-UST Java agent for `java.util.logging`
4447 To use the LTTng-UST Java agent in a Java application which uses
4448 `java.util.logging` (JUL):
4450 . In the Java application's source code, import the LTTng-UST
4451 log handler package for `java.util.logging`:
4456 import org.lttng.ust.agent.jul.LttngLogHandler;
4460 . Create an LTTng-UST JUL log handler:
4465 Handler lttngUstLogHandler = new LttngLogHandler();
4469 . Add this handler to the JUL loggers which should emit LTTng events:
4474 Logger myLogger = Logger.getLogger("some-logger");
4476 myLogger.addHandler(lttngUstLogHandler);
4480 . Use `java.util.logging` log statements and configuration as usual.
4481 The loggers with an attached LTTng-UST log handler can emit
4484 . Before exiting the application, remove the LTTng-UST log handler from
4485 the loggers attached to it and call its `close()` method:
4490 myLogger.removeHandler(lttngUstLogHandler);
4491 lttngUstLogHandler.close();
4495 This is not strictly necessary, but it is recommended for a clean
4496 disposal of the handler's resources.
4498 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4499 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4501 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4502 path] when you build the Java application.
4504 The JAR files are typically located in dir:{/usr/share/java}.
4506 IMPORTANT: The LTTng-UST Java agent must be
4507 <<installing-lttng,installed>> for the logging framework your
4510 .Use the LTTng-UST Java agent for `java.util.logging`.
4515 import java.io.IOException;
4516 import java.util.logging.Handler;
4517 import java.util.logging.Logger;
4518 import org.lttng.ust.agent.jul.LttngLogHandler;
4522 private static final int answer = 42;
4524 public static void main(String[] argv) throws Exception
4527 Logger logger = Logger.getLogger("jello");
4529 // Create an LTTng-UST log handler
4530 Handler lttngUstLogHandler = new LttngLogHandler();
4532 // Add the LTTng-UST log handler to our logger
4533 logger.addHandler(lttngUstLogHandler);
4536 logger.info("some info");
4537 logger.warning("some warning");
4539 logger.finer("finer information; the answer is " + answer);
4541 logger.severe("error!");
4543 // Not mandatory, but cleaner
4544 logger.removeHandler(lttngUstLogHandler);
4545 lttngUstLogHandler.close();
4554 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4557 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4558 <<enabling-disabling-events,create an event rule>> matching the
4559 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4564 $ lttng enable-event --jul jello
4568 Run the compiled class:
4572 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4575 <<basic-tracing-session-control,Stop tracing>> and inspect the
4585 In the resulting trace, an <<event,event record>> generated by a Java
4586 application using `java.util.logging` is named `lttng_jul:event` and
4587 has the following fields:
4590 Log record's message.
4596 Name of the class in which the log statement was executed.
4599 Name of the method in which the log statement was executed.
4602 Logging time (timestamp in milliseconds).
4605 Log level integer value.
4608 ID of the thread in which the log statement was executed.
4610 You can use the opt:lttng-enable-event(1):--loglevel or
4611 opt:lttng-enable-event(1):--loglevel-only option of the
4612 man:lttng-enable-event(1) command to target a range of JUL log levels
4613 or a specific JUL log level.
4618 ==== Use the LTTng-UST Java agent for Apache log4j
4620 To use the LTTng-UST Java agent in a Java application which uses
4623 . In the Java application's source code, import the LTTng-UST
4624 log appender package for Apache log4j:
4629 import org.lttng.ust.agent.log4j.LttngLogAppender;
4633 . Create an LTTng-UST log4j log appender:
4638 Appender lttngUstLogAppender = new LttngLogAppender();
4642 . Add this appender to the log4j loggers which should emit LTTng events:
4647 Logger myLogger = Logger.getLogger("some-logger");
4649 myLogger.addAppender(lttngUstLogAppender);
4653 . Use Apache log4j log statements and configuration as usual. The
4654 loggers with an attached LTTng-UST log appender can emit LTTng events.
4656 . Before exiting the application, remove the LTTng-UST log appender from
4657 the loggers attached to it and call its `close()` method:
4662 myLogger.removeAppender(lttngUstLogAppender);
4663 lttngUstLogAppender.close();
4667 This is not strictly necessary, but it is recommended for a clean
4668 disposal of the appender's resources.
4670 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4671 files, path:{lttng-ust-agent-common.jar} and
4672 path:{lttng-ust-agent-log4j.jar}, in the
4673 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4674 path] when you build the Java application.
4676 The JAR files are typically located in dir:{/usr/share/java}.
4678 IMPORTANT: The LTTng-UST Java agent must be
4679 <<installing-lttng,installed>> for the logging framework your
4682 .Use the LTTng-UST Java agent for Apache log4j.
4687 import org.apache.log4j.Appender;
4688 import org.apache.log4j.Logger;
4689 import org.lttng.ust.agent.log4j.LttngLogAppender;
4693 private static final int answer = 42;
4695 public static void main(String[] argv) throws Exception
4698 Logger logger = Logger.getLogger("jello");
4700 // Create an LTTng-UST log appender
4701 Appender lttngUstLogAppender = new LttngLogAppender();
4703 // Add the LTTng-UST log appender to our logger
4704 logger.addAppender(lttngUstLogAppender);
4707 logger.info("some info");
4708 logger.warn("some warning");
4710 logger.debug("debug information; the answer is " + answer);
4712 logger.fatal("error!");
4714 // Not mandatory, but cleaner
4715 logger.removeAppender(lttngUstLogAppender);
4716 lttngUstLogAppender.close();
4722 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4727 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4730 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4731 <<enabling-disabling-events,create an event rule>> matching the
4732 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4737 $ lttng enable-event --log4j jello
4741 Run the compiled class:
4745 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4748 <<basic-tracing-session-control,Stop tracing>> and inspect the
4758 In the resulting trace, an <<event,event record>> generated by a Java
4759 application using log4j is named `lttng_log4j:event` and
4760 has the following fields:
4763 Log record's message.
4769 Name of the class in which the log statement was executed.
4772 Name of the method in which the log statement was executed.
4775 Name of the file in which the executed log statement is located.
4778 Line number at which the log statement was executed.
4784 Log level integer value.
4787 Name of the Java thread in which the log statement was executed.
4789 You can use the opt:lttng-enable-event(1):--loglevel or
4790 opt:lttng-enable-event(1):--loglevel-only option of the
4791 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4792 or a specific log4j log level.
4796 [[java-application-context]]
4797 ==== Provide application-specific context fields in a Java application
4799 A Java application-specific context field is a piece of state provided
4800 by the application which <<adding-context,you can add>>, using the
4801 man:lttng-add-context(1) command, to each <<event,event record>>
4802 produced by the log statements of this application.
4804 For example, a given object might have a current request ID variable.
4805 You can create a context information retriever for this object and
4806 assign a name to this current request ID. You can then, using the
4807 man:lttng-add-context(1) command, add this context field by name to
4808 the JUL or log4j <<channel,channel>>.
4810 To provide application-specific context fields in a Java application:
4812 . In the Java application's source code, import the LTTng-UST
4813 Java agent context classes and interfaces:
4818 import org.lttng.ust.agent.context.ContextInfoManager;
4819 import org.lttng.ust.agent.context.IContextInfoRetriever;
4823 . Create a context information retriever class, that is, a class which
4824 implements the `IContextInfoRetriever` interface:
4829 class MyContextInfoRetriever implements IContextInfoRetriever
4832 public Object retrieveContextInfo(String key)
4834 if (key.equals("intCtx")) {
4836 } else if (key.equals("strContext")) {
4837 return "context value!";
4846 This `retrieveContextInfo()` method is the only member of the
4847 `IContextInfoRetriever` interface. Its role is to return the current
4848 value of a state by name to create a context field. The names of the
4849 context fields and which state variables they return depends on your
4852 All primitive types and objects are supported as context fields.
4853 When `retrieveContextInfo()` returns an object, the context field
4854 serializer calls its `toString()` method to add a string field to
4855 event records. The method can also return `null`, which means that
4856 no context field is available for the required name.
4858 . Register an instance of your context information retriever class to
4859 the context information manager singleton:
4864 IContextInfoRetriever cir = new MyContextInfoRetriever();
4865 ContextInfoManager cim = ContextInfoManager.getInstance();
4866 cim.registerContextInfoRetriever("retrieverName", cir);
4870 . Before exiting the application, remove your context information
4871 retriever from the context information manager singleton:
4876 ContextInfoManager cim = ContextInfoManager.getInstance();
4877 cim.unregisterContextInfoRetriever("retrieverName");
4881 This is not strictly necessary, but it is recommended for a clean
4882 disposal of some manager's resources.
4884 . Build your Java application with LTTng-UST Java agent support as
4885 usual, following the procedure for either the <<jul,JUL>> or
4886 <<log4j,Apache log4j>> framework.
4889 .Provide application-specific context fields in a Java application.
4894 import java.util.logging.Handler;
4895 import java.util.logging.Logger;
4896 import org.lttng.ust.agent.jul.LttngLogHandler;
4897 import org.lttng.ust.agent.context.ContextInfoManager;
4898 import org.lttng.ust.agent.context.IContextInfoRetriever;
4902 // Our context information retriever class
4903 private static class MyContextInfoRetriever
4904 implements IContextInfoRetriever
4907 public Object retrieveContextInfo(String key) {
4908 if (key.equals("intCtx")) {
4910 } else if (key.equals("strContext")) {
4911 return "context value!";
4918 private static final int answer = 42;
4920 public static void main(String args[]) throws Exception
4922 // Get the context information manager instance
4923 ContextInfoManager cim = ContextInfoManager.getInstance();
4925 // Create and register our context information retriever
4926 IContextInfoRetriever cir = new MyContextInfoRetriever();
4927 cim.registerContextInfoRetriever("myRetriever", cir);
4930 Logger logger = Logger.getLogger("jello");
4932 // Create an LTTng-UST log handler
4933 Handler lttngUstLogHandler = new LttngLogHandler();
4935 // Add the LTTng-UST log handler to our logger
4936 logger.addHandler(lttngUstLogHandler);
4939 logger.info("some info");
4940 logger.warning("some warning");
4942 logger.finer("finer information; the answer is " + answer);
4944 logger.severe("error!");
4946 // Not mandatory, but cleaner
4947 logger.removeHandler(lttngUstLogHandler);
4948 lttngUstLogHandler.close();
4949 cim.unregisterContextInfoRetriever("myRetriever");
4958 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4961 <<creating-destroying-tracing-sessions,Create a tracing session>>
4962 and <<enabling-disabling-events,create an event rule>> matching the
4968 $ lttng enable-event --jul jello
4971 <<adding-context,Add the application-specific context fields>> to the
4976 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
4977 $ lttng add-context --jul --type='$app.myRetriever:strContext'
4980 <<basic-tracing-session-control,Start tracing>>:
4987 Run the compiled class:
4991 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4994 <<basic-tracing-session-control,Stop tracing>> and inspect the
5006 [[python-application]]
5007 === User space Python agent
5009 You can instrument a Python 2 or Python 3 application which uses the
5010 standard https://docs.python.org/3/library/logging.html[`logging`]
5013 Each log statement emits an LTTng event once the
5014 application module imports the
5015 <<lttng-ust-agents,LTTng-UST Python agent>> package.
5018 .A Python application importing the LTTng-UST Python agent.
5019 image::python-app.png[]
5021 To use the LTTng-UST Python agent:
5023 . In the Python application's source code, import the LTTng-UST Python
5033 The LTTng-UST Python agent automatically adds its logging handler to the
5034 root logger at import time.
5036 Any log statement that the application executes before this import does
5037 not emit an LTTng event.
5039 IMPORTANT: The LTTng-UST Python agent must be
5040 <<installing-lttng,installed>>.
5042 . Use log statements and logging configuration as usual.
5043 Since the LTTng-UST Python agent adds a handler to the _root_
5044 logger, you can trace any log statement from any logger.
5046 .Use the LTTng-UST Python agent.
5057 logging.basicConfig()
5058 logger = logging.getLogger('my-logger')
5061 logger.debug('debug message')
5062 logger.info('info message')
5063 logger.warn('warn message')
5064 logger.error('error message')
5065 logger.critical('critical message')
5069 if __name__ == '__main__':
5073 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
5074 logging handler which prints to the standard error stream, is not
5075 strictly required for LTTng-UST tracing to work, but in versions of
5076 Python preceding 3.2, you could see a warning message which indicates
5077 that no handler exists for the logger `my-logger`.
5079 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5080 <<enabling-disabling-events,create an event rule>> matching the
5081 `my-logger` Python logger, and <<basic-tracing-session-control,start
5087 $ lttng enable-event --python my-logger
5091 Run the Python script:
5098 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5108 In the resulting trace, an <<event,event record>> generated by a Python
5109 application is named `lttng_python:event` and has the following fields:
5112 Logging time (string).
5115 Log record's message.
5121 Name of the function in which the log statement was executed.
5124 Line number at which the log statement was executed.
5127 Log level integer value.
5130 ID of the Python thread in which the log statement was executed.
5133 Name of the Python thread in which the log statement was executed.
5135 You can use the opt:lttng-enable-event(1):--loglevel or
5136 opt:lttng-enable-event(1):--loglevel-only option of the
5137 man:lttng-enable-event(1) command to target a range of Python log levels
5138 or a specific Python log level.
5140 When an application imports the LTTng-UST Python agent, the agent tries
5141 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5142 <<start-sessiond,start the session daemon>> _before_ you run the Python
5143 application. If a session daemon is found, the agent tries to register
5144 to it during 5{nbsp}seconds, after which the application continues
5145 without LTTng tracing support. You can override this timeout value with
5146 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5149 If the session daemon stops while a Python application with an imported
5150 LTTng-UST Python agent runs, the agent retries to connect and to
5151 register to a session daemon every 3{nbsp}seconds. You can override this
5152 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5157 [[proc-lttng-logger-abi]]
5160 The `lttng-tracer` Linux kernel module, part of
5161 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
5162 path:{/proc/lttng-logger} when it's loaded. Any application can write
5163 text data to this file to emit an LTTng event.
5166 .An application writes to the LTTng logger file to emit an LTTng event.
5167 image::lttng-logger.png[]
5169 The LTTng logger is the quickest method--not the most efficient,
5170 however--to add instrumentation to an application. It is designed
5171 mostly to instrument shell scripts:
5175 $ echo "Some message, some $variable" > /proc/lttng-logger
5178 Any event that the LTTng logger emits is named `lttng_logger` and
5179 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5180 other instrumentation points in the kernel tracing domain, **any Unix
5181 user** can <<enabling-disabling-events,create an event rule>> which
5182 matches its event name, not only the root user or users in the
5183 <<tracing-group,tracing group>>.
5185 To use the LTTng logger:
5187 * From any application, write text data to the path:{/proc/lttng-logger}
5190 The `msg` field of `lttng_logger` event records contains the
5193 NOTE: The maximum message length of an LTTng logger event is
5194 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5195 than one event to contain the remaining data.
5197 You should not use the LTTng logger to trace a user application which
5198 can be instrumented in a more efficient way, namely:
5200 * <<c-application,C and $$C++$$ applications>>.
5201 * <<java-application,Java applications>>.
5202 * <<python-application,Python applications>>.
5204 .Use the LTTng logger.
5209 echo 'Hello, World!' > /proc/lttng-logger
5211 df --human-readable --print-type / > /proc/lttng-logger
5214 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5215 <<enabling-disabling-events,create an event rule>> matching the
5216 `lttng_logger` Linux kernel tracepoint, and
5217 <<basic-tracing-session-control,start tracing>>:
5222 $ lttng enable-event --kernel lttng_logger
5226 Run the Bash script:
5233 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5244 [[instrumenting-linux-kernel]]
5245 === LTTng kernel tracepoints
5247 NOTE: This section shows how to _add_ instrumentation points to the
5248 Linux kernel. The kernel's subsystems are already thoroughly
5249 instrumented at strategic places for LTTng when you
5250 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5254 There are two methods to instrument the Linux kernel:
5256 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5257 tracepoint which uses the `TRACE_EVENT()` API.
5259 Choose this if you want to instrumentation a Linux kernel tree with an
5260 instrumentation point compatible with ftrace, perf, and SystemTap.
5262 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5263 instrument an out-of-tree kernel module.
5265 Choose this if you don't need ftrace, perf, or SystemTap support.
5269 [[linux-add-lttng-layer]]
5270 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5272 This section shows how to add an LTTng layer to existing ftrace
5273 instrumentation using the `TRACE_EVENT()` API.
5275 This section does not document the `TRACE_EVENT()` macro. You can
5276 read the following articles to learn more about this API:
5278 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
5279 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
5280 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
5282 The following procedure assumes that your ftrace tracepoints are
5283 correctly defined in their own header and that they are created in
5284 one source file using the `CREATE_TRACE_POINTS` definition.
5286 To add an LTTng layer over an existing ftrace tracepoint:
5288 . Make sure the following kernel configuration options are
5294 * `CONFIG_HIGH_RES_TIMERS`
5295 * `CONFIG_TRACEPOINTS`
5298 . Build the Linux source tree with your custom ftrace tracepoints.
5299 . Boot the resulting Linux image on your target system.
5301 Confirm that the tracepoints exist by looking for their names in the
5302 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5303 is your subsystem's name.
5305 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5310 $ cd $(mktemp -d) &&
5311 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.10.tar.bz2 &&
5312 tar -xf lttng-modules-latest-2.10.tar.bz2 &&
5313 cd lttng-modules-2.10.*
5317 . In dir:{instrumentation/events/lttng-module}, relative to the root
5318 of the LTTng-modules source tree, create a header file named
5319 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5320 LTTng-modules tracepoint definitions using the LTTng-modules
5323 Start with this template:
5327 .path:{instrumentation/events/lttng-module/my_subsys.h}
5330 #define TRACE_SYSTEM my_subsys
5332 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5333 #define _LTTNG_MY_SUBSYS_H
5335 #include "../../../probes/lttng-tracepoint-event.h"
5336 #include <linux/tracepoint.h>
5338 LTTNG_TRACEPOINT_EVENT(
5340 * Format is identical to TRACE_EVENT()'s version for the three
5341 * following macro parameters:
5344 TP_PROTO(int my_int, const char *my_string),
5345 TP_ARGS(my_int, my_string),
5347 /* LTTng-modules specific macros */
5349 ctf_integer(int, my_int_field, my_int)
5350 ctf_string(my_bar_field, my_bar)
5354 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5356 #include "../../../probes/define_trace.h"
5360 The entries in the `TP_FIELDS()` section are the list of fields for the
5361 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5362 ftrace's `TRACE_EVENT()` macro.
5364 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5365 complete description of the available `ctf_*()` macros.
5367 . Create the LTTng-modules probe's kernel module C source file,
5368 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5373 .path:{probes/lttng-probe-my-subsys.c}
5375 #include <linux/module.h>
5376 #include "../lttng-tracer.h"
5379 * Build-time verification of mismatch between mainline
5380 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5381 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5383 #include <trace/events/my_subsys.h>
5385 /* Create LTTng tracepoint probes */
5386 #define LTTNG_PACKAGE_BUILD
5387 #define CREATE_TRACE_POINTS
5388 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5390 #include "../instrumentation/events/lttng-module/my_subsys.h"
5392 MODULE_LICENSE("GPL and additional rights");
5393 MODULE_AUTHOR("Your name <your-email>");
5394 MODULE_DESCRIPTION("LTTng my_subsys probes");
5395 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5396 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5397 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5398 LTTNG_MODULES_EXTRAVERSION);
5402 . Edit path:{probes/KBuild} and add your new kernel module object
5403 next to the existing ones:
5407 .path:{probes/KBuild}
5411 obj-m += lttng-probe-module.o
5412 obj-m += lttng-probe-power.o
5414 obj-m += lttng-probe-my-subsys.o
5420 . Build and install the LTTng kernel modules:
5425 $ make KERNELDIR=/path/to/linux
5426 # make modules_install && depmod -a
5430 Replace `/path/to/linux` with the path to the Linux source tree where
5431 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5433 Note that you can also use the
5434 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5435 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5436 C code that need to be executed before the event fields are recorded.
5438 The best way to learn how to use the previous LTTng-modules macros is to
5439 inspect the existing LTTng-modules tracepoint definitions in the
5440 dir:{instrumentation/events/lttng-module} header files. Compare them
5441 with the Linux kernel mainline versions in the
5442 dir:{include/trace/events} directory of the Linux source tree.
5446 [[lttng-tracepoint-event-code]]
5447 ===== Use custom C code to access the data for tracepoint fields
5449 Although we recommended to always use the
5450 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5451 the arguments and fields of an LTTng-modules tracepoint when possible,
5452 sometimes you need a more complex process to access the data that the
5453 tracer records as event record fields. In other words, you need local
5454 variables and multiple C{nbsp}statements instead of simple
5455 argument-based expressions that you pass to the
5456 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5458 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5459 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5460 a block of C{nbsp}code to be executed before LTTng records the fields.
5461 The structure of this macro is:
5464 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5466 LTTNG_TRACEPOINT_EVENT_CODE(
5468 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5469 * version for the following three macro parameters:
5472 TP_PROTO(int my_int, const char *my_string),
5473 TP_ARGS(my_int, my_string),
5475 /* Declarations of custom local variables */
5478 unsigned long b = 0;
5479 const char *name = "(undefined)";
5480 struct my_struct *my_struct;
5484 * Custom code which uses both tracepoint arguments
5485 * (in TP_ARGS()) and local variables (in TP_locvar()).
5487 * Local variables are actually members of a structure pointed
5488 * to by the special variable tp_locvar.
5492 tp_locvar->a = my_int + 17;
5493 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5494 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5495 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5496 put_my_struct(tp_locvar->my_struct);
5505 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5506 * version for this, except that tp_locvar members can be
5507 * used in the argument expression parameters of
5508 * the ctf_*() macros.
5511 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5512 ctf_integer(int, my_struct_a, tp_locvar->a)
5513 ctf_string(my_string_field, my_string)
5514 ctf_string(my_struct_name, tp_locvar->name)
5519 IMPORTANT: The C code defined in `TP_code()` must not have any side
5520 effects when executed. In particular, the code must not allocate
5521 memory or get resources without deallocating this memory or putting
5522 those resources afterwards.
5525 [[instrumenting-linux-kernel-tracing]]
5526 ==== Load and unload a custom probe kernel module
5528 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5529 kernel module>> in the kernel before it can emit LTTng events.
5531 To load the default probe kernel modules and a custom probe kernel
5534 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5535 probe modules to load when starting a root <<lttng-sessiond,session
5539 .Load the `my_subsys`, `usb`, and the default probe modules.
5543 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5548 You only need to pass the subsystem name, not the whole kernel module
5551 To load _only_ a given custom probe kernel module:
5553 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5554 modules to load when starting a root session daemon:
5557 .Load only the `my_subsys` and `usb` probe modules.
5561 # lttng-sessiond --kmod-probes=my_subsys,usb
5566 To confirm that a probe module is loaded:
5573 $ lsmod | grep lttng_probe_usb
5577 To unload the loaded probe modules:
5579 * Kill the session daemon with `SIGTERM`:
5584 # pkill lttng-sessiond
5588 You can also use man:modprobe(8)'s `--remove` option if the session
5589 daemon terminates abnormally.
5592 [[controlling-tracing]]
5595 Once an application or a Linux kernel is
5596 <<instrumenting,instrumented>> for LTTng tracing,
5599 This section is divided in topics on how to use the various
5600 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5601 command-line tool>>, to _control_ the LTTng daemons and tracers.
5603 NOTE: In the following subsections, we refer to an man:lttng(1) command
5604 using its man page name. For example, instead of _Run the `create`
5605 command to..._, we use _Run the man:lttng-create(1) command to..._.
5609 === Start a session daemon
5611 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5612 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5615 You will see the following error when you run a command while no session
5619 Error: No session daemon is available
5622 The only command that automatically runs a session daemon is
5623 man:lttng-create(1), which you use to
5624 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5625 this is most of the time the first operation that you do, sometimes it's
5626 not. Some examples are:
5628 * <<list-instrumentation-points,List the available instrumentation points>>.
5629 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5631 [[tracing-group]] Each Unix user must have its own running session
5632 daemon to trace user applications. The session daemon that the root user
5633 starts is the only one allowed to control the LTTng kernel tracer. Users
5634 that are part of the _tracing group_ can control the root session
5635 daemon. The default tracing group name is `tracing`; you can set it to
5636 something else with the opt:lttng-sessiond(8):--group option when you
5637 start the root session daemon.
5639 To start a user session daemon:
5641 * Run man:lttng-sessiond(8):
5646 $ lttng-sessiond --daemonize
5650 To start the root session daemon:
5652 * Run man:lttng-sessiond(8) as the root user:
5657 # lttng-sessiond --daemonize
5661 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5662 start the session daemon in foreground.
5664 To stop a session daemon, use man:kill(1) on its process ID (standard
5667 Note that some Linux distributions could manage the LTTng session daemon
5668 as a service. In this case, you should use the service manager to
5669 start, restart, and stop session daemons.
5672 [[creating-destroying-tracing-sessions]]
5673 === Create and destroy a tracing session
5675 Almost all the LTTng control operations happen in the scope of
5676 a <<tracing-session,tracing session>>, which is the dialogue between the
5677 <<lttng-sessiond,session daemon>> and you.
5679 To create a tracing session with a generated name:
5681 * Use the man:lttng-create(1) command:
5690 The created tracing session's name is `auto` followed by the
5693 To create a tracing session with a specific name:
5695 * Use the optional argument of the man:lttng-create(1) command:
5700 $ lttng create my-session
5704 Replace `my-session` with the specific tracing session name.
5706 LTTng appends the creation date to the created tracing session's name.
5708 LTTng writes the traces of a tracing session in
5709 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5710 name of the tracing session. Note that the env:LTTNG_HOME environment
5711 variable defaults to `$HOME` if not set.
5713 To output LTTng traces to a non-default location:
5715 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5720 $ lttng create my-session --output=/tmp/some-directory
5724 You may create as many tracing sessions as you wish.
5726 To list all the existing tracing sessions for your Unix user:
5728 * Use the man:lttng-list(1) command:
5737 When you create a tracing session, it is set as the _current tracing
5738 session_. The following man:lttng(1) commands operate on the current
5739 tracing session when you don't specify one:
5741 [role="list-3-cols"]
5758 To change the current tracing session:
5760 * Use the man:lttng-set-session(1) command:
5765 $ lttng set-session new-session
5769 Replace `new-session` by the name of the new current tracing session.
5771 When you are done tracing in a given tracing session, you can destroy
5772 it. This operation frees the resources taken by the tracing session
5773 to destroy; it does not destroy the trace data that LTTng wrote for
5774 this tracing session.
5776 To destroy the current tracing session:
5778 * Use the man:lttng-destroy(1) command:
5787 The man:lttng-destroy(1) command also runs the man:lttng-stop(1)
5788 command implicitly (see <<basic-tracing-session-control,Start and stop a
5789 tracing session>>). You need to stop tracing to make LTTng flush the
5790 remaining trace data and make the trace readable.
5793 [[list-instrumentation-points]]
5794 === List the available instrumentation points
5796 The <<lttng-sessiond,session daemon>> can query the running instrumented
5797 user applications and the Linux kernel to get a list of available
5798 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5799 they are tracepoints and system calls. For the user space tracing
5800 domain, they are tracepoints. For the other tracing domains, they are
5803 To list the available instrumentation points:
5805 * Use the man:lttng-list(1) command with the requested tracing domain's
5809 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5810 must be a root user, or it must be a member of the
5811 <<tracing-group,tracing group>>).
5812 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5813 kernel system calls (your Unix user must be a root user, or it must be
5814 a member of the tracing group).
5815 * opt:lttng-list(1):--userspace: user space tracepoints.
5816 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5817 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5818 * opt:lttng-list(1):--python: Python loggers.
5821 .List the available user space tracepoints.
5825 $ lttng list --userspace
5829 .List the available Linux kernel system call tracepoints.
5833 $ lttng list --kernel --syscall
5838 [[enabling-disabling-events]]
5839 === Create and enable an event rule
5841 Once you <<creating-destroying-tracing-sessions,create a tracing
5842 session>>, you can create <<event,event rules>> with the
5843 man:lttng-enable-event(1) command.
5845 You specify each condition with a command-line option. The available
5846 condition options are shown in the following table.
5848 [role="growable",cols="asciidoc,asciidoc,default"]
5849 .Condition command-line options for the man:lttng-enable-event(1) command.
5851 |Option |Description |Applicable tracing domains
5857 . +--probe=__ADDR__+
5858 . +--function=__ADDR__+
5861 Instead of using the default _tracepoint_ instrumentation type, use:
5863 . A Linux system call.
5864 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5865 . The entry and return points of a Linux function (symbol or address).
5869 |First positional argument.
5872 Tracepoint or system call name. In the case of a Linux KProbe or
5873 function, this is a custom name given to the event rule. With the
5874 JUL, log4j, and Python domains, this is a logger name.
5876 With a tracepoint, logger, or system call name, you can use the special
5877 `*` globbing character to match anything (for example, `sched_*`,
5885 . +--loglevel=__LEVEL__+
5886 . +--loglevel-only=__LEVEL__+
5889 . Match only tracepoints or log statements with a logging level at
5890 least as severe as +__LEVEL__+.
5891 . Match only tracepoints or log statements with a logging level
5892 equal to +__LEVEL__+.
5894 See man:lttng-enable-event(1) for the list of available logging level
5897 |User space, JUL, log4j, and Python.
5899 |+--exclude=__EXCLUSIONS__+
5902 When you use a `*` character at the end of the tracepoint or logger
5903 name (first positional argument), exclude the specific names in the
5904 comma-delimited list +__EXCLUSIONS__+.
5907 User space, JUL, log4j, and Python.
5909 |+--filter=__EXPR__+
5912 Match only events which satisfy the expression +__EXPR__+.
5914 See man:lttng-enable-event(1) to learn more about the syntax of a
5921 You attach an event rule to a <<channel,channel>> on creation. If you do
5922 not specify the channel with the opt:lttng-enable-event(1):--channel
5923 option, and if the event rule to create is the first in its
5924 <<domain,tracing domain>> for a given tracing session, then LTTng
5925 creates a _default channel_ for you. This default channel is reused in
5926 subsequent invocations of the man:lttng-enable-event(1) command for the
5927 same tracing domain.
5929 An event rule is always enabled at creation time.
5931 The following examples show how you can combine the previous
5932 command-line options to create simple to more complex event rules.
5934 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5938 $ lttng enable-event --kernel sched_switch
5942 .Create an event rule matching four Linux kernel system calls (default channel).
5946 $ lttng enable-event --kernel --syscall open,write,read,close
5950 .Create event rules matching tracepoints with filter expressions (default channel).
5954 $ lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5959 $ lttng enable-event --kernel --all \
5960 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5965 $ lttng enable-event --jul my_logger \
5966 --filter='$app.retriever:cur_msg_id > 3'
5969 IMPORTANT: Make sure to always quote the filter string when you
5970 use man:lttng(1) from a shell.
5973 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5977 $ lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5980 IMPORTANT: Make sure to always quote the wildcard character when you
5981 use man:lttng(1) from a shell.
5984 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5988 $ lttng enable-event --python my-app.'*' \
5989 --exclude='my-app.module,my-app.hello'
5993 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5997 $ lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
6001 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
6005 $ lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
6009 The event rules of a given channel form a whitelist: as soon as an
6010 emitted event passes one of them, LTTng can record the event. For
6011 example, an event named `my_app:my_tracepoint` emitted from a user space
6012 tracepoint with a `TRACE_ERROR` log level passes both of the following
6017 $ lttng enable-event --userspace my_app:my_tracepoint
6018 $ lttng enable-event --userspace my_app:my_tracepoint \
6019 --loglevel=TRACE_INFO
6022 The second event rule is redundant: the first one includes
6026 [[disable-event-rule]]
6027 === Disable an event rule
6029 To disable an event rule that you <<enabling-disabling-events,created>>
6030 previously, use the man:lttng-disable-event(1) command. This command
6031 disables _all_ the event rules (of a given tracing domain and channel)
6032 which match an instrumentation point. The other conditions are not
6033 supported as of LTTng{nbsp}{revision}.
6035 The LTTng tracer does not record an emitted event which passes
6036 a _disabled_ event rule.
6038 .Disable an event rule matching a Python logger (default channel).
6042 $ lttng disable-event --python my-logger
6046 .Disable an event rule matching all `java.util.logging` loggers (default channel).
6050 $ lttng disable-event --jul '*'
6054 .Disable _all_ the event rules of the default channel.
6056 The opt:lttng-disable-event(1):--all-events option is not, like the
6057 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
6058 equivalent of the event name `*` (wildcard): it disables _all_ the event
6059 rules of a given channel.
6063 $ lttng disable-event --jul --all-events
6067 NOTE: You cannot delete an event rule once you create it.
6071 === Get the status of a tracing session
6073 To get the status of the current tracing session, that is, its
6074 parameters, its channels, event rules, and their attributes:
6076 * Use the man:lttng-status(1) command:
6086 To get the status of any tracing session:
6088 * Use the man:lttng-list(1) command with the tracing session's name:
6093 $ lttng list my-session
6097 Replace `my-session` with the desired tracing session's name.
6100 [[basic-tracing-session-control]]
6101 === Start and stop a tracing session
6103 Once you <<creating-destroying-tracing-sessions,create a tracing
6105 <<enabling-disabling-events,create one or more event rules>>,
6106 you can start and stop the tracers for this tracing session.
6108 To start tracing in the current tracing session:
6110 * Use the man:lttng-start(1) command:
6119 LTTng is very flexible: you can launch user applications before
6120 or after the you start the tracers. The tracers only record the events
6121 if they pass enabled event rules and if they occur while the tracers are
6124 To stop tracing in the current tracing session:
6126 * Use the man:lttng-stop(1) command:
6135 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
6136 records>> or lost sub-buffers since the last time you ran
6137 man:lttng-start(1), warnings are printed when you run the
6138 man:lttng-stop(1) command.
6140 IMPORTANT: You need to stop tracing to make LTTng flush the remaining
6141 trace data and make the trace readable. Note that the
6142 man:lttng-destroy(1) command (see
6143 <<creating-destroying-tracing-sessions,Create and destroy a tracing
6144 session>>) also runs the man:lttng-stop(1) command implicitly.
6147 [[enabling-disabling-channels]]
6148 === Create a channel
6150 Once you create a tracing session, you can create a <<channel,channel>>
6151 with the man:lttng-enable-channel(1) command.
6153 Note that LTTng automatically creates a default channel when, for a
6154 given <<domain,tracing domain>>, no channels exist and you
6155 <<enabling-disabling-events,create>> the first event rule. This default
6156 channel is named `channel0` and its attributes are set to reasonable
6157 values. Therefore, you only need to create a channel when you need
6158 non-default attributes.
6160 You specify each non-default channel attribute with a command-line
6161 option when you use the man:lttng-enable-channel(1) command. The
6162 available command-line options are:
6164 [role="growable",cols="asciidoc,asciidoc"]
6165 .Command-line options for the man:lttng-enable-channel(1) command.
6167 |Option |Description
6173 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
6174 the default _discard_ mode.
6176 |`--buffers-pid` (user space tracing domain only)
6179 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6180 instead of the default per-user buffering scheme.
6182 |+--subbuf-size=__SIZE__+
6185 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6186 either for each Unix user (default), or for each instrumented process.
6188 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6190 |+--num-subbuf=__COUNT__+
6193 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6194 for each Unix user (default), or for each instrumented process.
6196 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6198 |+--tracefile-size=__SIZE__+
6201 Set the maximum size of each trace file that this channel writes within
6202 a stream to +__SIZE__+ bytes instead of no maximum.
6204 See <<tracefile-rotation,Trace file count and size>>.
6206 |+--tracefile-count=__COUNT__+
6209 Limit the number of trace files that this channel creates to
6210 +__COUNT__+ channels instead of no limit.
6212 See <<tracefile-rotation,Trace file count and size>>.
6214 |+--switch-timer=__PERIODUS__+
6217 Set the <<channel-switch-timer,switch timer period>>
6218 to +__PERIODUS__+{nbsp}µs.
6220 |+--read-timer=__PERIODUS__+
6223 Set the <<channel-read-timer,read timer period>>
6224 to +__PERIODUS__+{nbsp}µs.
6226 |[[opt-blocking-timeout]]+--blocking-timeout=__TIMEOUTUS__+
6229 Set the timeout of user space applications which load LTTng-UST
6230 in blocking mode to +__TIMEOUTUS__+:
6233 Never block (non-blocking mode).
6236 Block forever until space is available in a sub-buffer to record
6239 __n__, a positive value::
6240 Wait for at most __n__ µs when trying to write into a sub-buffer.
6242 Note that, for this option to have any effect on an instrumented
6243 user space application, you need to run the application with a set
6244 env:LTTNG_UST_ALLOW_BLOCKING environment variable.
6246 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6249 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
6253 You can only create a channel in the Linux kernel and user space
6254 <<domain,tracing domains>>: other tracing domains have their own channel
6255 created on the fly when <<enabling-disabling-events,creating event
6260 Because of a current LTTng limitation, you must create all channels
6261 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6262 tracing session, that is, before the first time you run
6265 Since LTTng automatically creates a default channel when you use the
6266 man:lttng-enable-event(1) command with a specific tracing domain, you
6267 cannot, for example, create a Linux kernel event rule, start tracing,
6268 and then create a user space event rule, because no user space channel
6269 exists yet and it's too late to create one.
6271 For this reason, make sure to configure your channels properly
6272 before starting the tracers for the first time!
6275 The following examples show how you can combine the previous
6276 command-line options to create simple to more complex channels.
6278 .Create a Linux kernel channel with default attributes.
6282 $ lttng enable-channel --kernel my-channel
6286 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6290 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6291 --buffers-pid my-channel
6295 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6297 <<creating-destroying-tracing-sessions,Create a tracing-session>>,
6298 create the channel, <<enabling-disabling-events,create an event rule>>,
6299 and <<basic-tracing-session-control,start tracing>>:
6304 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
6305 $ lttng enable-event --userspace --channel=blocking-channel --all
6309 Run an application instrumented with LTTng-UST and allow it to block:
6313 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6317 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
6321 $ lttng enable-channel --kernel --tracefile-count=8 \
6322 --tracefile-size=4194304 my-channel
6326 .Create a user space channel in overwrite (or _flight recorder_) mode.
6330 $ lttng enable-channel --userspace --overwrite my-channel
6334 You can <<enabling-disabling-events,create>> the same event rule in
6335 two different channels:
6339 $ lttng enable-event --userspace --channel=my-channel app:tp
6340 $ lttng enable-event --userspace --channel=other-channel app:tp
6343 If both channels are enabled, when a tracepoint named `app:tp` is
6344 reached, LTTng records two events, one for each channel.
6348 === Disable a channel
6350 To disable a specific channel that you <<enabling-disabling-channels,created>>
6351 previously, use the man:lttng-disable-channel(1) command.
6353 .Disable a specific Linux kernel channel.
6357 $ lttng disable-channel --kernel my-channel
6361 The state of a channel precedes the individual states of event rules
6362 attached to it: event rules which belong to a disabled channel, even if
6363 they are enabled, are also considered disabled.
6367 === Add context fields to a channel
6369 Event record fields in trace files provide important information about
6370 events that occured previously, but sometimes some external context may
6371 help you solve a problem faster. Examples of context fields are:
6373 * The **process ID**, **thread ID**, **process name**, and
6374 **process priority** of the thread in which the event occurs.
6375 * The **hostname** of the system on which the event occurs.
6376 * The current values of many possible **performance counters** using
6378 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6380 ** Branch instructions, misses, and loads.
6382 * Any context defined at the application level (supported for the
6383 JUL and log4j <<domain,tracing domains>>).
6385 To get the full list of available context fields, see
6386 `lttng add-context --list`. Some context fields are reserved for a
6387 specific <<domain,tracing domain>> (Linux kernel or user space).
6389 You add context fields to <<channel,channels>>. All the events
6390 that a channel with added context fields records contain those fields.
6392 To add context fields to one or all the channels of a given tracing
6395 * Use the man:lttng-add-context(1) command.
6397 .Add context fields to all the channels of the current tracing session.
6399 The following command line adds the virtual process identifier and
6400 the per-thread CPU cycles count fields to all the user space channels
6401 of the current tracing session.
6405 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6409 .Add performance counter context fields by raw ID
6411 See man:lttng-add-context(1) for the exact format of the context field
6412 type, which is partly compatible with the format used in
6417 $ lttng add-context --userspace --type=perf:thread:raw:r0110:test
6418 $ lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6422 .Add a context field to a specific channel.
6424 The following command line adds the thread identifier context field
6425 to the Linux kernel channel named `my-channel` in the current
6430 $ lttng add-context --kernel --channel=my-channel --type=tid
6434 .Add an application-specific context field to a specific channel.
6436 The following command line adds the `cur_msg_id` context field of the
6437 `retriever` context retriever for all the instrumented
6438 <<java-application,Java applications>> recording <<event,event records>>
6439 in the channel named `my-channel`:
6443 $ lttng add-context --kernel --channel=my-channel \
6444 --type='$app:retriever:cur_msg_id'
6447 IMPORTANT: Make sure to always quote the `$` character when you
6448 use man:lttng-add-context(1) from a shell.
6451 NOTE: You cannot remove context fields from a channel once you add it.
6456 === Track process IDs
6458 It's often useful to allow only specific process IDs (PIDs) to emit
6459 events. For example, you may wish to record all the system calls made by
6460 a given process (Ă la http://linux.die.net/man/1/strace[strace]).
6462 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6463 purpose. Both commands operate on a whitelist of process IDs. You _add_
6464 entries to this whitelist with the man:lttng-track(1) command and remove
6465 entries with the man:lttng-untrack(1) command. Any process which has one
6466 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6467 an enabled <<event,event rule>>.
6469 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6470 process with a given tracked ID exit and another process be given this
6471 ID, then the latter would also be allowed to emit events.
6473 .Track and untrack process IDs.
6475 For the sake of the following example, assume the target system has 16
6479 <<creating-destroying-tracing-sessions,create a tracing session>>,
6480 the whitelist contains all the possible PIDs:
6483 .All PIDs are tracked.
6484 image::track-all.png[]
6486 When the whitelist is full and you use the man:lttng-track(1) command to
6487 specify some PIDs to track, LTTng first clears the whitelist, then it
6488 tracks the specific PIDs. After:
6492 $ lttng track --pid=3,4,7,10,13
6498 .PIDs 3, 4, 7, 10, and 13 are tracked.
6499 image::track-3-4-7-10-13.png[]
6501 You can add more PIDs to the whitelist afterwards:
6505 $ lttng track --pid=1,15,16
6511 .PIDs 1, 15, and 16 are added to the whitelist.
6512 image::track-1-3-4-7-10-13-15-16.png[]
6514 The man:lttng-untrack(1) command removes entries from the PID tracker's
6515 whitelist. Given the previous example, the following command:
6519 $ lttng untrack --pid=3,7,10,13
6522 leads to this whitelist:
6525 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6526 image::track-1-4-15-16.png[]
6528 LTTng can track all possible PIDs again using the
6529 opt:lttng-track(1):--all option:
6533 $ lttng track --pid --all
6536 The result is, again:
6539 .All PIDs are tracked.
6540 image::track-all.png[]
6543 .Track only specific PIDs
6545 A very typical use case with PID tracking is to start with an empty
6546 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6547 then add PIDs manually while tracers are active. You can accomplish this
6548 by using the opt:lttng-untrack(1):--all option of the
6549 man:lttng-untrack(1) command to clear the whitelist after you
6550 <<creating-destroying-tracing-sessions,create a tracing session>>:
6554 $ lttng untrack --pid --all
6560 .No PIDs are tracked.
6561 image::untrack-all.png[]
6563 If you trace with this whitelist configuration, the tracer records no
6564 events for this <<domain,tracing domain>> because no processes are
6565 tracked. You can use the man:lttng-track(1) command as usual to track
6566 specific PIDs, for example:
6570 $ lttng track --pid=6,11
6576 .PIDs 6 and 11 are tracked.
6577 image::track-6-11.png[]
6582 [[saving-loading-tracing-session]]
6583 === Save and load tracing session configurations
6585 Configuring a <<tracing-session,tracing session>> can be long. Some of
6586 the tasks involved are:
6588 * <<enabling-disabling-channels,Create channels>> with
6589 specific attributes.
6590 * <<adding-context,Add context fields>> to specific channels.
6591 * <<enabling-disabling-events,Create event rules>> with specific log
6592 level and filter conditions.
6594 If you use LTTng to solve real world problems, chances are you have to
6595 record events using the same tracing session setup over and over,
6596 modifying a few variables each time in your instrumented program
6597 or environment. To avoid constant tracing session reconfiguration,
6598 the man:lttng(1) command-line tool can save and load tracing session
6599 configurations to/from XML files.
6601 To save a given tracing session configuration:
6603 * Use the man:lttng-save(1) command:
6608 $ lttng save my-session
6612 Replace `my-session` with the name of the tracing session to save.
6614 LTTng saves tracing session configurations to
6615 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6616 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6617 the opt:lttng-save(1):--output-path option to change this destination
6620 LTTng saves all configuration parameters, for example:
6622 * The tracing session name.
6623 * The trace data output path.
6624 * The channels with their state and all their attributes.
6625 * The context fields you added to channels.
6626 * The event rules with their state, log level and filter conditions.
6628 To load a tracing session:
6630 * Use the man:lttng-load(1) command:
6635 $ lttng load my-session
6639 Replace `my-session` with the name of the tracing session to load.
6641 When LTTng loads a configuration, it restores your saved tracing session
6642 as if you just configured it manually.
6644 See man:lttng(1) for the complete list of command-line options. You
6645 can also save and load all many sessions at a time, and decide in which
6646 directory to output the XML files.
6649 [[sending-trace-data-over-the-network]]
6650 === Send trace data over the network
6652 LTTng can send the recorded trace data to a remote system over the
6653 network instead of writing it to the local file system.
6655 To send the trace data over the network:
6657 . On the _remote_ system (which can also be the target system),
6658 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6667 . On the _target_ system, create a tracing session configured to
6668 send trace data over the network:
6673 $ lttng create my-session --set-url=net://remote-system
6677 Replace `remote-system` by the host name or IP address of the
6678 remote system. See man:lttng-create(1) for the exact URL format.
6680 . On the target system, use the man:lttng(1) command-line tool as usual.
6681 When tracing is active, the target's consumer daemon sends sub-buffers
6682 to the relay daemon running on the remote system instead of flushing
6683 them to the local file system. The relay daemon writes the received
6684 packets to the local file system.
6686 The relay daemon writes trace files to
6687 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6688 +__hostname__+ is the host name of the target system and +__session__+
6689 is the tracing session name. Note that the env:LTTNG_HOME environment
6690 variable defaults to `$HOME` if not set. Use the
6691 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6692 trace files to another base directory.
6697 === View events as LTTng emits them (noch:{LTTng} live)
6699 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6700 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6701 display events as LTTng emits them on the target system while tracing is
6704 The relay daemon creates a _tee_: it forwards the trace data to both
6705 the local file system and to connected live viewers:
6708 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6713 . On the _target system_, create a <<tracing-session,tracing session>>
6719 $ lttng create my-session --live
6723 This spawns a local relay daemon.
6725 . Start the live viewer and configure it to connect to the relay
6726 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6731 $ babeltrace --input-format=lttng-live \
6732 net://localhost/host/hostname/my-session
6739 * `hostname` with the host name of the target system.
6740 * `my-session` with the name of the tracing session to view.
6743 . Configure the tracing session as usual with the man:lttng(1)
6744 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6746 You can list the available live tracing sessions with Babeltrace:
6750 $ babeltrace --input-format=lttng-live net://localhost
6753 You can start the relay daemon on another system. In this case, you need
6754 to specify the relay daemon's URL when you create the tracing session
6755 with the opt:lttng-create(1):--set-url option. You also need to replace
6756 `localhost` in the procedure above with the host name of the system on
6757 which the relay daemon is running.
6759 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6760 command-line options.
6764 [[taking-a-snapshot]]
6765 === Take a snapshot of the current sub-buffers of a tracing session
6767 The normal behavior of LTTng is to append full sub-buffers to growing
6768 trace data files. This is ideal to keep a full history of the events
6769 that occurred on the target system, but it can
6770 represent too much data in some situations. For example, you may wish
6771 to trace your application continuously until some critical situation
6772 happens, in which case you only need the latest few recorded
6773 events to perform the desired analysis, not multi-gigabyte trace files.
6775 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6776 current sub-buffers of a given <<tracing-session,tracing session>>.
6777 LTTng can write the snapshot to the local file system or send it over
6782 . Create a tracing session in _snapshot mode_:
6787 $ lttng create my-session --snapshot
6791 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6792 <<channel,channels>> created in this mode is automatically set to
6793 _overwrite_ (flight recorder mode).
6795 . Configure the tracing session as usual with the man:lttng(1)
6796 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6798 . **Optional**: When you need to take a snapshot,
6799 <<basic-tracing-session-control,stop tracing>>.
6801 You can take a snapshot when the tracers are active, but if you stop
6802 them first, you are sure that the data in the sub-buffers does not
6803 change before you actually take the snapshot.
6810 $ lttng snapshot record --name=my-first-snapshot
6814 LTTng writes the current sub-buffers of all the current tracing
6815 session's channels to trace files on the local file system. Those trace
6816 files have `my-first-snapshot` in their name.
6818 There is no difference between the format of a normal trace file and the
6819 format of a snapshot: viewers of LTTng traces also support LTTng
6822 By default, LTTng writes snapshot files to the path shown by
6823 `lttng snapshot list-output`. You can change this path or decide to send
6824 snapshots over the network using either:
6826 . An output path or URL that you specify when you create the
6828 . An snapshot output path or URL that you add using
6829 `lttng snapshot add-output`
6830 . An output path or URL that you provide directly to the
6831 `lttng snapshot record` command.
6833 Method 3 overrides method 2, which overrides method 1. When you
6834 specify a URL, a relay daemon must listen on a remote system (see
6835 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6840 === Use the machine interface
6842 With any command of the man:lttng(1) command-line tool, you can set the
6843 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6844 XML machine interface output, for example:
6848 $ lttng --mi=xml enable-event --kernel --syscall open
6851 A schema definition (XSD) is
6852 https://github.com/lttng/lttng-tools/blob/stable-2.10/src/common/mi-lttng-3.0.xsd[available]
6853 to ease the integration with external tools as much as possible.
6857 [[metadata-regenerate]]
6858 === Regenerate the metadata of an LTTng trace
6860 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6861 data stream files and a metadata file. This metadata file contains,
6862 amongst other things, information about the offset of the clock sources
6863 used to timestamp <<event,event records>> when tracing.
6865 If, once a <<tracing-session,tracing session>> is
6866 <<basic-tracing-session-control,started>>, a major
6867 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6868 happens, the trace's clock offset also needs to be updated. You
6869 can use the `metadata` item of the man:lttng-regenerate(1) command
6872 The main use case of this command is to allow a system to boot with
6873 an incorrect wall time and trace it with LTTng before its wall time
6874 is corrected. Once the system is known to be in a state where its
6875 wall time is correct, it can run `lttng regenerate metadata`.
6877 To regenerate the metadata of an LTTng trace:
6879 * Use the `metadata` item of the man:lttng-regenerate(1) command:
6884 $ lttng regenerate metadata
6890 `lttng regenerate metadata` has the following limitations:
6892 * Tracing session <<creating-destroying-tracing-sessions,created>>
6894 * User space <<channel,channels>>, if any, are using
6895 <<channel-buffering-schemes,per-user buffering>>.
6900 [[regenerate-statedump]]
6901 === Regenerate the state dump of a tracing session
6903 The LTTng kernel and user space tracers generate state dump
6904 <<event,event records>> when the application starts or when you
6905 <<basic-tracing-session-control,start a tracing session>>. An analysis
6906 can use the state dump event records to set an initial state before it
6907 builds the rest of the state from the following event records.
6908 http://tracecompass.org/[Trace Compass] is a notable example of an
6909 application which uses the state dump of an LTTng trace.
6911 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
6912 state dump event records are not included in the snapshot because they
6913 were recorded to a sub-buffer that has been consumed or overwritten
6916 You can use the `lttng regenerate statedump` command to emit the state
6917 dump event records again.
6919 To regenerate the state dump of the current tracing session, provided
6920 create it in snapshot mode, before you take a snapshot:
6922 . Use the `statedump` item of the man:lttng-regenerate(1) command:
6927 $ lttng regenerate statedump
6931 . <<basic-tracing-session-control,Stop the tracing session>>:
6940 . <<taking-a-snapshot,Take a snapshot>>:
6945 $ lttng snapshot record --name=my-snapshot
6949 Depending on the event throughput, you should run steps 1 and 2
6950 as closely as possible.
6952 NOTE: To record the state dump events, you need to
6953 <<enabling-disabling-events,create event rules>> which enable them.
6954 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
6955 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
6959 [[persistent-memory-file-systems]]
6960 === Record trace data on persistent memory file systems
6962 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6963 (NVRAM) is random-access memory that retains its information when power
6964 is turned off (non-volatile). Systems with such memory can store data
6965 structures in RAM and retrieve them after a reboot, without flushing
6966 to typical _storage_.
6968 Linux supports NVRAM file systems thanks to either
6969 http://pramfs.sourceforge.net/[PRAMFS] or
6970 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6971 (requires Linux 4.1+).
6973 This section does not describe how to operate such file systems;
6974 we assume that you have a working persistent memory file system.
6976 When you create a <<tracing-session,tracing session>>, you can specify
6977 the path of the shared memory holding the sub-buffers. If you specify a
6978 location on an NVRAM file system, then you can retrieve the latest
6979 recorded trace data when the system reboots after a crash.
6981 To record trace data on a persistent memory file system and retrieve the
6982 trace data after a system crash:
6984 . Create a tracing session with a sub-buffer shared memory path located
6985 on an NVRAM file system:
6990 $ lttng create my-session --shm-path=/path/to/shm
6994 . Configure the tracing session as usual with the man:lttng(1)
6995 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6997 . After a system crash, use the man:lttng-crash(1) command-line tool to
6998 view the trace data recorded on the NVRAM file system:
7003 $ lttng-crash /path/to/shm
7007 The binary layout of the ring buffer files is not exactly the same as
7008 the trace files layout. This is why you need to use man:lttng-crash(1)
7009 instead of your preferred trace viewer directly.
7011 To convert the ring buffer files to LTTng trace files:
7013 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
7018 $ lttng-crash --extract=/path/to/trace /path/to/shm
7024 [[notif-trigger-api]]
7025 === Get notified when a channel's buffer usage is too high or too low
7027 With LTTng's $$C/C++$$ notification and trigger API, your user
7028 application can get notified when the buffer usage of one or more
7029 <<channel,channels>> becomes too low or too high. You can use this API
7030 and enable or disable <<event,event rules>> during tracing to avoid
7031 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>.
7033 .Have a user application get notified when an LTTng channel's buffer usage is too high.
7035 In this example, we create and build an application which gets notified
7036 when the buffer usage of a specific LTTng channel is higher than
7037 75{nbsp}%. We only print that it is the case in the example, but we
7038 could as well use the API of <<liblttng-ctl-lttng,`liblttng-ctl`>> to
7039 disable event rules when this happens.
7041 . Create the application's C source file:
7049 #include <lttng/domain.h>
7050 #include <lttng/action/action.h>
7051 #include <lttng/action/notify.h>
7052 #include <lttng/condition/condition.h>
7053 #include <lttng/condition/buffer-usage.h>
7054 #include <lttng/condition/evaluation.h>
7055 #include <lttng/notification/channel.h>
7056 #include <lttng/notification/notification.h>
7057 #include <lttng/trigger/trigger.h>
7058 #include <lttng/endpoint.h>
7060 int main(int argc, char *argv[])
7062 int exit_status = 0;
7063 struct lttng_notification_channel *notification_channel;
7064 struct lttng_condition *condition;
7065 struct lttng_action *action;
7066 struct lttng_trigger *trigger;
7067 const char *tracing_session_name;
7068 const char *channel_name;
7071 tracing_session_name = argv[1];
7072 channel_name = argv[2];
7075 * Create a notification channel. A notification channel
7076 * connects the user application to the LTTng session daemon.
7077 * This notification channel can be used to listen to various
7078 * types of notifications.
7080 notification_channel = lttng_notification_channel_create(
7081 lttng_session_daemon_notification_endpoint);
7084 * Create a "high buffer usage" condition. In this case, the
7085 * condition is reached when the buffer usage is greater than or
7086 * equal to 75 %. We create the condition for a specific tracing
7087 * session name, channel name, and for the user space tracing
7090 * The "low buffer usage" condition type also exists.
7092 condition = lttng_condition_buffer_usage_high_create();
7093 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
7094 lttng_condition_buffer_usage_set_session_name(
7095 condition, tracing_session_name);
7096 lttng_condition_buffer_usage_set_channel_name(condition,
7098 lttng_condition_buffer_usage_set_domain_type(condition,
7102 * Create an action (get a notification) to take when the
7103 * condition created above is reached.
7105 action = lttng_action_notify_create();
7108 * Create a trigger. A trigger associates a condition to an
7109 * action: the action is executed when the condition is reached.
7111 trigger = lttng_trigger_create(condition, action);
7113 /* Register the trigger to LTTng. */
7114 lttng_register_trigger(trigger);
7117 * Now that we have registered a trigger, a notification will be
7118 * emitted everytime its condition is met. To receive this
7119 * notification, we must subscribe to notifications that match
7120 * the same condition.
7122 lttng_notification_channel_subscribe(notification_channel,
7126 * Notification loop. You can put this in a dedicated thread to
7127 * avoid blocking the main thread.
7130 struct lttng_notification *notification;
7131 enum lttng_notification_channel_status status;
7132 const struct lttng_evaluation *notification_evaluation;
7133 const struct lttng_condition *notification_condition;
7134 double buffer_usage;
7136 /* Receive the next notification. */
7137 status = lttng_notification_channel_get_next_notification(
7138 notification_channel, ¬ification);
7141 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
7143 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
7145 * The session daemon can drop notifications if
7146 * a monitoring application is not consuming the
7147 * notifications fast enough.
7150 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
7152 * The notification channel has been closed by the
7153 * session daemon. This is typically caused by a session
7154 * daemon shutting down.
7158 /* Unhandled conditions or errors. */
7164 * A notification provides, amongst other things:
7166 * * The condition that caused this notification to be
7168 * * The condition evaluation, which provides more
7169 * specific information on the evaluation of the
7172 * The condition evaluation provides the buffer usage
7173 * value at the moment the condition was reached.
7175 notification_condition = lttng_notification_get_condition(
7177 notification_evaluation = lttng_notification_get_evaluation(
7180 /* We're subscribed to only one condition. */
7181 assert(lttng_condition_get_type(notification_condition) ==
7182 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
7185 * Get the exact sampled buffer usage from the
7186 * condition evaluation.
7188 lttng_evaluation_buffer_usage_get_usage_ratio(
7189 notification_evaluation, &buffer_usage);
7192 * At this point, instead of printing a message, we
7193 * could do something to reduce the channel's buffer
7194 * usage, like disable specific events.
7196 printf("Buffer usage is %f %% in tracing session \"%s\", "
7197 "user space channel \"%s\".\n", buffer_usage * 100,
7198 tracing_session_name, channel_name);
7199 lttng_notification_destroy(notification);
7203 lttng_action_destroy(action);
7204 lttng_condition_destroy(condition);
7205 lttng_trigger_destroy(trigger);
7206 lttng_notification_channel_destroy(notification_channel);
7212 . Build the `notif-app` application, linking it to `liblttng-ctl`:
7217 $ gcc -o notif-app notif-app.c -llttng-ctl
7221 . <<creating-destroying-tracing-sessions,Create a tracing session>>,
7222 <<enabling-disabling-events,create an event rule>> matching all the
7223 user space tracepoints, and
7224 <<basic-tracing-session-control,start tracing>>:
7229 $ lttng create my-session
7230 $ lttng enable-event --userspace --all
7235 If you create the channel manually with the man:lttng-enable-channel(1)
7236 command, you can control how frequently are the current values of the
7237 channel's properties sampled to evaluate user conditions with the
7238 opt:lttng-enable-channel(1):--monitor-timer option.
7240 . Run the `notif-app` application. This program accepts the
7241 <<tracing-session,tracing session>> name and the user space channel
7242 name as its two first arguments. The channel which LTTng automatically
7243 creates with the man:lttng-enable-event(1) command above is named
7249 $ ./notif-app my-session channel0
7253 . In another terminal, run an application with a very high event
7254 throughput so that the 75{nbsp}% buffer usage condition is reached.
7256 In the first terminal, the application should print lines like this:
7259 Buffer usage is 81.45197 % in tracing session "my-session", user space
7263 If you don't see anything, try modifying the condition in
7264 path:{notif-app.c} to a lower value (0.1, for example), rebuilding it
7265 (step 2) and running it again (step 4).
7272 [[lttng-modules-ref]]
7273 === noch:{LTTng-modules}
7277 [[lttng-tracepoint-enum]]
7278 ==== `LTTNG_TRACEPOINT_ENUM()` usage
7280 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
7284 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
7289 * `name` with the name of the enumeration (C identifier, unique
7290 amongst all the defined enumerations).
7291 * `entries` with a list of enumeration entries.
7293 The available enumeration entry macros are:
7295 +ctf_enum_value(__name__, __value__)+::
7296 Entry named +__name__+ mapped to the integral value +__value__+.
7298 +ctf_enum_range(__name__, __begin__, __end__)+::
7299 Entry named +__name__+ mapped to the range of integral values between
7300 +__begin__+ (included) and +__end__+ (included).
7302 +ctf_enum_auto(__name__)+::
7303 Entry named +__name__+ mapped to the integral value following the
7304 last mapping's value.
7306 The last value of a `ctf_enum_value()` entry is its +__value__+
7309 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
7311 If `ctf_enum_auto()` is the first entry in the list, its integral
7314 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
7315 to use a defined enumeration as a tracepoint field.
7317 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
7321 LTTNG_TRACEPOINT_ENUM(
7324 ctf_enum_auto("AUTO: EXPECT 0")
7325 ctf_enum_value("VALUE: 23", 23)
7326 ctf_enum_value("VALUE: 27", 27)
7327 ctf_enum_auto("AUTO: EXPECT 28")
7328 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
7329 ctf_enum_auto("AUTO: EXPECT 304")
7337 [[lttng-modules-tp-fields]]
7338 ==== Tracepoint fields macros (for `TP_FIELDS()`)
7340 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
7341 tracepoint fields, which must be listed within `TP_FIELDS()` in
7342 `LTTNG_TRACEPOINT_EVENT()`, are:
7344 [role="func-desc growable",cols="asciidoc,asciidoc"]
7345 .Available macros to define LTTng-modules tracepoint fields
7347 |Macro |Description and parameters
7350 +ctf_integer(__t__, __n__, __e__)+
7352 +ctf_integer_nowrite(__t__, __n__, __e__)+
7354 +ctf_user_integer(__t__, __n__, __e__)+
7356 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
7358 Standard integer, displayed in base 10.
7361 Integer C type (`int`, `long`, `size_t`, ...).
7367 Argument expression.
7370 +ctf_integer_hex(__t__, __n__, __e__)+
7372 +ctf_user_integer_hex(__t__, __n__, __e__)+
7374 Standard integer, displayed in base 16.
7383 Argument expression.
7385 |+ctf_integer_oct(__t__, __n__, __e__)+
7387 Standard integer, displayed in base 8.
7396 Argument expression.
7399 +ctf_integer_network(__t__, __n__, __e__)+
7401 +ctf_user_integer_network(__t__, __n__, __e__)+
7403 Integer in network byte order (big-endian), displayed in base 10.
7412 Argument expression.
7415 +ctf_integer_network_hex(__t__, __n__, __e__)+
7417 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
7419 Integer in network byte order, displayed in base 16.
7428 Argument expression.
7431 +ctf_enum(__N__, __t__, __n__, __e__)+
7433 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
7435 +ctf_user_enum(__N__, __t__, __n__, __e__)+
7437 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
7442 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
7445 Integer C type (`int`, `long`, `size_t`, ...).
7451 Argument expression.
7454 +ctf_string(__n__, __e__)+
7456 +ctf_string_nowrite(__n__, __e__)+
7458 +ctf_user_string(__n__, __e__)+
7460 +ctf_user_string_nowrite(__n__, __e__)+
7462 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7468 Argument expression.
7471 +ctf_array(__t__, __n__, __e__, __s__)+
7473 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7475 +ctf_user_array(__t__, __n__, __e__, __s__)+
7477 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7479 Statically-sized array of integers.
7482 Array element C type.
7488 Argument expression.
7494 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7496 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7498 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7500 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7502 Statically-sized array of bits.
7504 The type of +__e__+ must be an integer type. +__s__+ is the number
7505 of elements of such type in +__e__+, not the number of bits.
7508 Array element C type.
7514 Argument expression.
7520 +ctf_array_text(__t__, __n__, __e__, __s__)+
7522 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7524 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7526 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7528 Statically-sized array, printed as text.
7530 The string does not need to be null-terminated.
7533 Array element C type (always `char`).
7539 Argument expression.
7545 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7547 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7549 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7551 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7553 Dynamically-sized array of integers.
7555 The type of +__E__+ must be unsigned.
7558 Array element C type.
7564 Argument expression.
7567 Length expression C type.
7573 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7575 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7577 Dynamically-sized array of integers, displayed in base 16.
7579 The type of +__E__+ must be unsigned.
7582 Array element C type.
7588 Argument expression.
7591 Length expression C type.
7596 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7598 Dynamically-sized array of integers in network byte order (big-endian),
7599 displayed in base 10.
7601 The type of +__E__+ must be unsigned.
7604 Array element C type.
7610 Argument expression.
7613 Length expression C type.
7619 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7621 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7623 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7625 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7627 Dynamically-sized array of bits.
7629 The type of +__e__+ must be an integer type. +__s__+ is the number
7630 of elements of such type in +__e__+, not the number of bits.
7632 The type of +__E__+ must be unsigned.
7635 Array element C type.
7641 Argument expression.
7644 Length expression C type.
7650 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7652 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7654 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7656 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7658 Dynamically-sized array, displayed as text.
7660 The string does not need to be null-terminated.
7662 The type of +__E__+ must be unsigned.
7664 The behaviour is undefined if +__e__+ is `NULL`.
7667 Sequence element C type (always `char`).
7673 Argument expression.
7676 Length expression C type.
7682 Use the `_user` versions when the argument expression, `e`, is
7683 a user space address. In the cases of `ctf_user_integer*()` and
7684 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7687 The `_nowrite` versions omit themselves from the session trace, but are
7688 otherwise identical. This means the `_nowrite` fields won't be written
7689 in the recorded trace. Their primary purpose is to make some
7690 of the event context available to the
7691 <<enabling-disabling-events,event filters>> without having to
7692 commit the data to sub-buffers.
7698 Terms related to LTTng and to tracing in general:
7701 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7702 the cmd:babeltrace command, some libraries, and Python bindings.
7704 <<channel-buffering-schemes,buffering scheme>>::
7705 A layout of sub-buffers applied to a given channel.
7707 <<channel,channel>>::
7708 An entity which is responsible for a set of ring buffers.
7710 <<event,Event rules>> are always attached to a specific channel.
7713 A reference of time for a tracer.
7715 <<lttng-consumerd,consumer daemon>>::
7716 A process which is responsible for consuming the full sub-buffers
7717 and write them to a file system or send them over the network.
7719 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
7720 mode in which the tracer _discards_ new event records when there's no
7721 sub-buffer space left to store them.
7724 The consequence of the execution of an instrumentation
7725 point, like a tracepoint that you manually place in some source code,
7726 or a Linux kernel KProbe.
7728 An event is said to _occur_ at a specific time. Different actions can
7729 be taken upon the occurrence of an event, like record the event's payload
7732 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
7733 The mechanism by which event records of a given channel are lost
7734 (not recorded) when there is no sub-buffer space left to store them.
7736 [[def-event-name]]event name::
7737 The name of an event, which is also the name of the event record.
7738 This is also called the _instrumentation point name_.
7741 A record, in a trace, of the payload of an event which occured.
7743 <<event,event rule>>::
7744 Set of conditions which must be satisfied for one or more occuring
7745 events to be recorded.
7747 `java.util.logging`::
7749 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7751 <<instrumenting,instrumentation>>::
7752 The use of LTTng probes to make a piece of software traceable.
7754 instrumentation point::
7755 A point in the execution path of a piece of software that, when
7756 reached by this execution, can emit an event.
7758 instrumentation point name::
7759 See _<<def-event-name,event name>>_.
7762 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7763 developed by the Apache Software Foundation.
7766 Level of severity of a log statement or user space
7767 instrumentation point.
7770 The _Linux Trace Toolkit: next generation_ project.
7772 <<lttng-cli,cmd:lttng>>::
7773 A command-line tool provided by the LTTng-tools project which you
7774 can use to send and receive control messages to and from a
7778 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7779 which is a set of analyzing programs that are used to obtain a
7780 higher level view of an LTTng trace.
7782 cmd:lttng-consumerd::
7783 The name of the consumer daemon program.
7786 A utility provided by the LTTng-tools project which can convert
7787 ring buffer files (usually
7788 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7791 LTTng Documentation::
7794 <<lttng-live,LTTng live>>::
7795 A communication protocol between the relay daemon and live viewers
7796 which makes it possible to see events "live", as they are received by
7799 <<lttng-modules,LTTng-modules>>::
7800 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7801 which contains the Linux kernel modules to make the Linux kernel
7802 instrumentation points available for LTTng tracing.
7805 The name of the relay daemon program.
7807 cmd:lttng-sessiond::
7808 The name of the session daemon program.
7811 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7812 contains the various programs and libraries used to
7813 <<controlling-tracing,control tracing>>.
7815 <<lttng-ust,LTTng-UST>>::
7816 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7817 contains libraries to instrument user applications.
7819 <<lttng-ust-agents,LTTng-UST Java agent>>::
7820 A Java package provided by the LTTng-UST project to allow the
7821 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
7824 <<lttng-ust-agents,LTTng-UST Python agent>>::
7825 A Python package provided by the LTTng-UST project to allow the
7826 LTTng instrumentation of Python logging statements.
7828 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7829 The event loss mode in which new event records overwrite older
7830 event records when there's no sub-buffer space left to store them.
7832 <<channel-buffering-schemes,per-process buffering>>::
7833 A buffering scheme in which each instrumented process has its own
7834 sub-buffers for a given user space channel.
7836 <<channel-buffering-schemes,per-user buffering>>::
7837 A buffering scheme in which all the processes of a Unix user share the
7838 same sub-buffer for a given user space channel.
7840 <<lttng-relayd,relay daemon>>::
7841 A process which is responsible for receiving the trace data sent by
7842 a distant consumer daemon.
7845 A set of sub-buffers.
7847 <<lttng-sessiond,session daemon>>::
7848 A process which receives control commands from you and orchestrates
7849 the tracers and various LTTng daemons.
7851 <<taking-a-snapshot,snapshot>>::
7852 A copy of the current data of all the sub-buffers of a given tracing
7853 session, saved as trace files.
7856 One part of an LTTng ring buffer which contains event records.
7859 The time information attached to an event when it is emitted.
7862 A set of files which are the concatenations of one or more
7863 flushed sub-buffers.
7866 The action of recording the events emitted by an application
7867 or by a system, or to initiate such recording by controlling
7871 The http://tracecompass.org[Trace Compass] project and application.
7874 An instrumentation point using the tracepoint mechanism of the Linux
7875 kernel or of LTTng-UST.
7877 tracepoint definition::
7878 The definition of a single tracepoint.
7881 The name of a tracepoint.
7883 tracepoint provider::
7884 A set of functions providing tracepoints to an instrumented user
7887 Not to be confused with a _tracepoint provider package_: many tracepoint
7888 providers can exist within a tracepoint provider package.
7890 tracepoint provider package::
7891 One or more tracepoint providers compiled as an object file or as
7895 A software which records emitted events.
7897 <<domain,tracing domain>>::
7898 A namespace for event sources.
7900 <<tracing-group,tracing group>>::
7901 The Unix group in which a Unix user can be to be allowed to trace the
7904 <<tracing-session,tracing session>>::
7905 A stateful dialogue between you and a <<lttng-sessiond,session
7909 An application running in user space, as opposed to a Linux kernel
7910 module, for example.