1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng{nbsp}{revision}?
77 LTTng{nbsp}{revision} bears the name _Lafontaine_. This modern
78 https://en.wikipedia.org/wiki/saison[saison] from the
79 https://oshlag.com/[Oshlag] microbrewery is a refreshing--zesty--rice
80 beer with hints of fruit and spices. Some even say it makes for a great
81 https://en.wikipedia.org/wiki/Somaek[Somaek] when mixed with
82 Chamisul Soju, not that we've tried!
84 New features and changes in LTTng{nbsp}{revision}:
86 * Just like you can typically perform
87 https://en.wikipedia.org/wiki/Log_rotation[log rotation], you can
88 now <<session-rotation,_rotate_ a tracing session>>, that
89 is, according to man:lttng-rotate(1), archive the current trace
90 chunk (all the tracing session's trace data since the last rotation
91 or since its inception) so that LTTng does not manage it anymore.
93 Once LTTng archives a trace chunk, you are free to read it, modify it,
94 move it, or remove it.
96 You can rotate a tracing session immediately or set a rotation schedule
97 to automate rotations.
99 * When you <<enabling-disabling-events,create an event rule>>, the
100 filter expression syntax now supports the following new operators:
104 ** `<<` (bitwise left shift)
105 ** `>>` (bitwise right shift)
111 The syntax also supports array indexing with the usual square brackets:
114 regs[3][1] & 0xff7 == 0x240
117 There are peculiarities for both the new operators and the array
118 indexing brackets, like a custom precedence table and implicit casting.
119 See man:lttng-enable-event(1) to get all the details about the filter
122 * You can now dynamically instrument any application's or library's
123 function entry by symbol name thanks to the new
124 opt:lttng-enable-event(1):--userspace-probe option of
125 the `lttng enable-event` command:
129 $ lttng enable-event --kernel \
130 --userspace-probe=/usr/lib/libc.so.6:malloc libc_malloc
133 The option also supports tracing existing
134 https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
135 Statically Defined Tracing] (USDT) probe (DTrace-style marker). For
136 example, given the following probe:
140 DTRACE_PROBE2("server", "accept-request", request_id, ip_addr);
143 You can trace this probe with:
146 $ lttng enable-event --kernel \
147 --userspace-probe=sdt:/path/to/server:server:accept-request \
148 server_accept_request
151 This feature makes use of Linux's
152 https://www.kernel.org/doc/Documentation/trace/uprobetracer.txt[uprobe]
153 mechanism, therefore you must use the `--userspace-probe`
154 instrumentation option with the opt:lttng-enable-event(1):--kernel
157 NOTE: As of LTTng{nbsp}{revision}, LTTng does not record function
158 parameters with the opt:lttng-enable-event(1):--userspace-probe option.
160 * Two new <<adding-context,context>> fields are available for Linux
161 kernel <<channel,channels>>:
164 ** `callstack-kernel`
168 Thanks to those, you can record the Linux kernel and user call stacks
169 when a kernel event occurs. For example:
173 $ lttng enable-event --kernel --syscall open
174 $ lttng add-context --kernel --type=callstack-kernel --type=callstack-user
177 When an man:open(2) system call occurs, LTTng attaches the kernel and
178 user call stacks to the recorded event.
180 NOTE: LTTng cannot always sample the user space call stack reliably.
181 For instance, LTTng cannot sample the call stack of user applications
182 and libraries compiled with the
183 https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html[`-fomit-frame-pointer`]
184 option. In such a case, the tracing is not affected, but the sampled
185 user space call stack may only contain the user call stack's topmost
188 * User applications and libraries instrumented with
189 <<lttng-ust,LTTng-UST>> can now safely unload (man:dlclose(3)) a
191 <<building-tracepoint-providers-and-user-application,tracepoint
194 * The <<lttng-relayd,relay daemon>> is more efficient and presents fewer
195 connectivity issues, especially when a large number of targets send
196 trace data to a given relay daemon.
198 * LTTng-UST uses https://github.com/numactl/numactl[libnuma]
199 when available to allocate <<def-sub-buffer,sub-buffers>>, making them
201 https://en.wikipedia.org/wiki/Non-uniform_memory_access[NUMA] node.
203 This change makes the tracer more efficient on NUMA systems.
209 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
210 generation_ is a modern toolkit for tracing Linux systems and
211 applications. So your first question might be:
218 As the history of software engineering progressed and led to what
219 we now take for granted--complex, numerous and
220 interdependent software applications running in parallel on
221 sophisticated operating systems like Linux--the authors of such
222 components, software developers, began feeling a natural
223 urge to have tools that would ensure the robustness and good performance
224 of their masterpieces.
226 One major achievement in this field is, inarguably, the
227 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
228 an essential tool for developers to find and fix bugs. But even the best
229 debugger won't help make your software run faster, and nowadays, faster
230 software means either more work done by the same hardware, or cheaper
231 hardware for the same work.
233 A _profiler_ is often the tool of choice to identify performance
234 bottlenecks. Profiling is suitable to identify _where_ performance is
235 lost in a given software. The profiler outputs a profile, a statistical
236 summary of observed events, which you may use to discover which
237 functions took the most time to execute. However, a profiler won't
238 report _why_ some identified functions are the bottleneck. Bottlenecks
239 might only occur when specific conditions are met, conditions that are
240 sometimes impossible to capture by a statistical profiler, or impossible
241 to reproduce with an application altered by the overhead of an
242 event-based profiler. For a thorough investigation of software
243 performance issues, a history of execution is essential, with the
244 recorded values of variables and context fields you choose, and
245 with as little influence as possible on the instrumented software. This
246 is where tracing comes in handy.
248 _Tracing_ is a technique used to understand what goes on in a running
249 software system. The software used for tracing is called a _tracer_,
250 which is conceptually similar to a tape recorder. When recording,
251 specific instrumentation points placed in the software source code
252 generate events that are saved on a giant tape: a _trace_ file. You
253 can trace user applications and the operating system at the same time,
254 opening the possibility of resolving a wide range of problems that would
255 otherwise be extremely challenging.
257 Tracing is often compared to _logging_. However, tracers and loggers are
258 two different tools, serving two different purposes. Tracers are
259 designed to record much lower-level events that occur much more
260 frequently than log messages, often in the range of thousands per
261 second, with very little execution overhead. Logging is more appropriate
262 for a very high-level analysis of less frequent events: user accesses,
263 exceptional conditions (errors and warnings, for example), database
264 transactions, instant messaging communications, and such. Simply put,
265 logging is one of the many use cases that can be satisfied with tracing.
267 The list of recorded events inside a trace file can be read manually
268 like a log file for the maximum level of detail, but it is generally
269 much more interesting to perform application-specific analyses to
270 produce reduced statistics and graphs that are useful to resolve a
271 given problem. Trace viewers and analyzers are specialized tools
274 In the end, this is what LTTng is: a powerful, open source set of
275 tools to trace the Linux kernel and user applications at the same time.
276 LTTng is composed of several components actively maintained and
277 developed by its link:/community/#where[community].
280 [[lttng-alternatives]]
281 === Alternatives to noch:{LTTng}
283 Excluding proprietary solutions, a few competing software tracers
286 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
287 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
288 user scripts and is responsible for loading code into the
289 Linux kernel for further execution and collecting the outputted data.
290 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
291 subsystem in the Linux kernel in which a virtual machine can execute
292 programs passed from the user space to the kernel. You can attach
293 such programs to tracepoints and kprobes thanks to a system call, and
294 they can output data to the user space when executed thanks to
295 different mechanisms (pipe, VM register values, and eBPF maps, to name
297 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
298 is the de facto function tracer of the Linux kernel. Its user
299 interface is a set of special files in sysfs.
300 * https://perf.wiki.kernel.org/[perf] is
301 a performance analysis tool for Linux which supports hardware
302 performance counters, tracepoints, as well as other counters and
303 types of probes. perf's controlling utility is the cmd:perf command
305 * http://linux.die.net/man/1/strace[strace]
306 is a command-line utility which records system calls made by a
307 user process, as well as signal deliveries and changes of process
308 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
309 to fulfill its function.
310 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
311 analyze Linux kernel events. You write scripts, or _chisels_ in
312 sysdig's jargon, in Lua and sysdig executes them while it traces the
313 system or afterwards. sysdig's interface is the cmd:sysdig
314 command-line tool as well as the curses-based cmd:csysdig tool.
315 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
316 user space tracer which uses custom user scripts to produce plain text
317 traces. SystemTap converts the scripts to the C language, and then
318 compiles them as Linux kernel modules which are loaded to produce
319 trace data. SystemTap's primary user interface is the cmd:stap
322 The main distinctive features of LTTng is that it produces correlated
323 kernel and user space traces, as well as doing so with the lowest
324 overhead amongst other solutions. It produces trace files in the
325 http://diamon.org/ctf[CTF] format, a file format optimized
326 for the production and analyses of multi-gigabyte data.
328 LTTng is the result of more than 10{nbsp}years of active open source
329 development by a community of passionate developers.
330 LTTng{nbsp}{revision} is currently available on major desktop and server
333 The main interface for tracing control is a single command-line tool
334 named cmd:lttng. The latter can create several tracing sessions, enable
335 and disable events on the fly, filter events efficiently with custom
336 user expressions, start and stop tracing, and much more. LTTng can
337 record the traces on the file system or send them over the network, and
338 keep them totally or partially. You can view the traces once tracing
339 becomes inactive or in real-time.
341 <<installing-lttng,Install LTTng now>> and
342 <<getting-started,start tracing>>!
348 **LTTng** is a set of software <<plumbing,components>> which interact to
349 <<instrumenting,instrument>> the Linux kernel and user applications, and
350 to <<controlling-tracing,control tracing>> (start and stop
351 tracing, enable and disable event rules, and the rest). Those
352 components are bundled into the following packages:
355 Libraries and command-line interface to control tracing.
358 Linux kernel modules to instrument and trace the kernel.
361 Libraries and Java/Python packages to instrument and trace user
364 Most distributions mark the LTTng-modules and LTTng-UST packages as
365 optional when installing LTTng-tools (which is always required). In the
366 following sections, we always provide the steps to install all three,
369 * You only need to install LTTng-modules if you intend to trace the
371 * You only need to install LTTng-UST if you intend to trace user
375 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 30 March 2020.
377 |Distribution |Available in releases
379 |https://www.ubuntu.com/[Ubuntu]
380 |Ubuntu{nbsp}16.04 _Xenial Xerus_ and Ubuntu{nbsp}18.04 _Bionic Beaver_:
381 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
383 |https://www.debian.org/[Debian]
384 |<<debian,Debian "bullseye" (testing)>>.
386 |https://www.archlinux.org/[Arch Linux]
387 |<<arch-linux,_Community_ repository and AUR>>.
389 |https://getfedora.org/[Fedora]
390 |xref:fedora[Fedora{nbsp}32 and Fedora{nbsp}33].
392 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
393 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
395 |https://buildroot.org/[Buildroot]
396 |xref:buildroot[Buildroot{nbsp}2019.11 and Buildroot{nbsp}2020.02].
401 === Ubuntu: noch:{LTTng} Stable {revision} PPA
403 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
404 Stable{nbsp}{revision} PPA] offers the latest stable
405 LTTng{nbsp}{revision} packages for Ubuntu{nbsp}16.04 _Xenial Xerus_ and
406 Ubuntu{nbsp}18.04 _Bionic Beaver_.
408 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision}
411 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
417 # apt-add-repository ppa:lttng/stable-2.10
422 . Install the main LTTng{nbsp}{revision} packages:
427 # apt-get install lttng-tools
428 # apt-get install lttng-modules-dkms
429 # apt-get install liblttng-ust-dev
433 . **If you need to instrument and trace
434 <<java-application,Java applications>>**, install the LTTng-UST
440 # apt-get install liblttng-ust-agent-java
444 . **If you need to instrument and trace
445 <<python-application,Python{nbsp}3 applications>>**, install the
446 LTTng-UST Python agent:
451 # apt-get install python3-lttngust
459 To install LTTng{nbsp}{revision} on Debian "bullseye" (testing):
461 . Install the main LTTng{nbsp}{revision} packages:
466 # apt-get install lttng-modules-dkms
467 # apt-get install liblttng-ust-dev
468 # apt-get install lttng-tools
472 . **If you need to instrument and trace <<java-application,Java
473 applications>>**, install the LTTng-UST Java agent:
478 # apt-get install liblttng-ust-agent-java
482 . **If you need to instrument and trace <<python-application,Python
483 applications>>**, install the LTTng-UST Python agent:
488 # apt-get install python3-lttngust
496 LTTng-UST{nbsp}{revision} is available in Arch Linux's _community_
497 repository, while LTTng-tools{nbsp}{revision} and
498 LTTng-modules{nbsp}{revision} are available in the
499 https://aur.archlinux.org/[AUR].
501 To install LTTng{nbsp}{revision} on Arch Linux, using
502 https://github.com/Jguer/yay[yay] for the AUR packages:
504 . Install the main LTTng{nbsp}{revision} packages:
509 # pacman -Sy lttng-ust
510 $ yay -Sy lttng-tools
511 $ yay -Sy lttng-modules
515 . **If you need to instrument and trace <<python-application,Python
516 applications>>**, install the LTTng-UST Python agent:
521 # pacman -Sy python-lttngust
522 # pacman -Sy python2-lttngust
530 To install LTTng{nbsp}{revision} on Fedora{nbsp}32 and Fedora{nbsp}33:
532 . Install the LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision}
538 # yum install lttng-tools
539 # yum install lttng-ust
543 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
549 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
550 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
551 cd lttng-modules-2.11.* &&
553 sudo make modules_install &&
559 .Java and Python application instrumentation and tracing
561 If you need to instrument and trace <<java-application,Java
562 applications>> on Fedora, you need to build and install
563 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
564 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
565 `--enable-java-agent-all` options to the `configure` script, depending
566 on which Java logging framework you use.
568 If you need to instrument and trace <<python-application,Python
569 applications>> on Fedora, you need to build and install
570 LTTng-UST{nbsp}{revision} from source and pass the
571 `--enable-python-agent` option to the `configure` script.
578 To install LTTng{nbsp}{revision} on Buildroot{nbsp}2019.11 and
579 Buildroot{nbsp}2020.02:
581 . Launch the Buildroot configuration tool:
590 . In **Kernel**, check **Linux kernel**.
591 . In **Toolchain**, check **Enable WCHAR support**.
592 . In **Target packages**{nbsp}→ **Debugging, profiling and benchmark**,
593 check **lttng-modules** and **lttng-tools**.
594 . In **Target packages**{nbsp}→ **Libraries**{nbsp}→
595 **Other**, check **lttng-libust**.
598 [[building-from-source]]
599 === Build from source
601 To build and install LTTng{nbsp}{revision} from source:
603 . Using your distribution's package manager, or from source, install
604 the following dependencies of LTTng-tools and LTTng-UST:
607 * https://sourceforge.net/projects/libuuid/[libuuid]
608 * http://directory.fsf.org/wiki/Popt[popt]
609 * http://liburcu.org/[Userspace RCU]
610 * http://www.xmlsoft.org/[libxml2]
611 * **Optional**: https://github.com/numactl/numactl[numactl]
614 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
620 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
621 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
622 cd lttng-modules-2.11.* &&
624 sudo make modules_install &&
629 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
635 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.11.tar.bz2 &&
636 tar -xf lttng-ust-latest-2.11.tar.bz2 &&
637 cd lttng-ust-2.11.* &&
645 Add `--disable-numa` to `./configure` if you don't have
646 https://github.com/numactl/numactl[numactl].
650 .Java and Python application tracing
652 If you need to instrument and trace <<java-application,Java
653 applications>>, pass the `--enable-java-agent-jul`,
654 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
655 `configure` script, depending on which Java logging framework you use.
657 If you need to instrument and trace <<python-application,Python
658 applications>>, pass the `--enable-python-agent` option to the
659 `configure` script. You can set the `PYTHON` environment variable to the
660 path to the Python interpreter for which to install the LTTng-UST Python
668 By default, LTTng-UST libraries are installed to
669 dir:{/usr/local/lib}, which is the de facto directory in which to
670 keep self-compiled and third-party libraries.
672 When <<building-tracepoint-providers-and-user-application,linking an
673 instrumented user application with `liblttng-ust`>>:
675 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
677 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
678 man:gcc(1), man:g++(1), or man:clang(1).
682 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
688 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
689 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
690 cd lttng-tools-2.11.* &&
698 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
699 previous steps automatically for a given version of LTTng and confine
700 the installed files in a specific directory. This can be useful to test
701 LTTng without installing it on your system.
707 This is a short guide to get started quickly with LTTng kernel and user
710 Before you follow this guide, make sure to <<installing-lttng,install>>
713 This tutorial walks you through the steps to:
715 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
716 . <<tracing-your-own-user-application,Trace a user application>> written
718 . <<viewing-and-analyzing-your-traces,View and analyze the
722 [[tracing-the-linux-kernel]]
723 === Trace the Linux kernel
725 The following command lines start with the `#` prompt because you need
726 root privileges to trace the Linux kernel. You can also trace the kernel
727 as a regular user if your Unix user is a member of the
728 <<tracing-group,tracing group>>.
730 . Create a <<tracing-session,tracing session>> which writes its traces
731 to dir:{/tmp/my-kernel-trace}:
736 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
740 . List the available kernel tracepoints and system calls:
745 # lttng list --kernel
746 # lttng list --kernel --syscall
750 . Create <<event,event rules>> which match the desired instrumentation
751 point names, for example the `sched_switch` and `sched_process_fork`
752 tracepoints, and the man:open(2) and man:close(2) system calls:
757 # lttng enable-event --kernel sched_switch,sched_process_fork
758 # lttng enable-event --kernel --syscall open,close
762 You can also create an event rule which matches _all_ the Linux kernel
763 tracepoints (this will generate a lot of data when tracing):
768 # lttng enable-event --kernel --all
772 . <<basic-tracing-session-control,Start tracing>>:
781 . Do some operation on your system for a few seconds. For example,
782 load a website, or list the files of a directory.
783 . <<creating-destroying-tracing-sessions,Destroy>> the current
793 The man:lttng-destroy(1) command does not destroy the trace data; it
794 only destroys the state of the tracing session.
796 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
797 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
798 session>>). You need to stop tracing to make LTTng flush the remaining
799 trace data and make the trace readable.
801 . For the sake of this example, make the recorded trace accessible to
807 # chown -R $(whoami) /tmp/my-kernel-trace
811 See <<viewing-and-analyzing-your-traces,View and analyze the
812 recorded events>> to view the recorded events.
815 [[tracing-your-own-user-application]]
816 === Trace a user application
818 This section steps you through a simple example to trace a
819 _Hello world_ program written in C.
821 To create the traceable user application:
823 . Create the tracepoint provider header file, which defines the
824 tracepoints and the events they can generate:
830 #undef TRACEPOINT_PROVIDER
831 #define TRACEPOINT_PROVIDER hello_world
833 #undef TRACEPOINT_INCLUDE
834 #define TRACEPOINT_INCLUDE "./hello-tp.h"
836 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
839 #include <lttng/tracepoint.h>
849 ctf_string(my_string_field, my_string_arg)
850 ctf_integer(int, my_integer_field, my_integer_arg)
854 #endif /* _HELLO_TP_H */
856 #include <lttng/tracepoint-event.h>
860 . Create the tracepoint provider package source file:
866 #define TRACEPOINT_CREATE_PROBES
867 #define TRACEPOINT_DEFINE
869 #include "hello-tp.h"
873 . Build the tracepoint provider package:
878 $ gcc -c -I. hello-tp.c
882 . Create the _Hello World_ application source file:
889 #include "hello-tp.h"
891 int main(int argc, char *argv[])
895 puts("Hello, World!\nPress Enter to continue...");
898 * The following getchar() call is only placed here for the purpose
899 * of this demonstration, to pause the application in order for
900 * you to have time to list its tracepoints. It is not
906 * A tracepoint() call.
908 * Arguments, as defined in hello-tp.h:
910 * 1. Tracepoint provider name (required)
911 * 2. Tracepoint name (required)
912 * 3. my_integer_arg (first user-defined argument)
913 * 4. my_string_arg (second user-defined argument)
915 * Notice the tracepoint provider and tracepoint names are
916 * NOT strings: they are in fact parts of variables that the
917 * macros in hello-tp.h create.
919 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
921 for (x = 0; x < argc; ++x) {
922 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
925 puts("Quitting now!");
926 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
933 . Build the application:
942 . Link the application with the tracepoint provider package,
943 `liblttng-ust`, and `libdl`:
948 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
952 Here's the whole build process:
955 .User space tracing tutorial's build steps.
956 image::ust-flow.png[]
958 To trace the user application:
960 . Run the application with a few arguments:
965 $ ./hello world and beyond
974 Press Enter to continue...
978 . Start an LTTng <<lttng-sessiond,session daemon>>:
983 $ lttng-sessiond --daemonize
987 Note that a session daemon might already be running, for example as
988 a service that the distribution's service manager started.
990 . List the available user space tracepoints:
995 $ lttng list --userspace
999 You see the `hello_world:my_first_tracepoint` tracepoint listed
1000 under the `./hello` process.
1002 . Create a <<tracing-session,tracing session>>:
1007 $ lttng create my-user-space-session
1011 . Create an <<event,event rule>> which matches the
1012 `hello_world:my_first_tracepoint` event name:
1017 $ lttng enable-event --userspace hello_world:my_first_tracepoint
1021 . <<basic-tracing-session-control,Start tracing>>:
1030 . Go back to the running `hello` application and press Enter. The
1031 program executes all `tracepoint()` instrumentation points and exits.
1032 . <<creating-destroying-tracing-sessions,Destroy>> the current
1042 The man:lttng-destroy(1) command does not destroy the trace data; it
1043 only destroys the state of the tracing session.
1045 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
1046 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
1047 session>>). You need to stop tracing to make LTTng flush the remaining
1048 trace data and make the trace readable.
1050 By default, LTTng saves the traces in
1051 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1052 where +__name__+ is the tracing session name. The
1053 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1055 See <<viewing-and-analyzing-your-traces,View and analyze the
1056 recorded events>> to view the recorded events.
1059 [[viewing-and-analyzing-your-traces]]
1060 === View and analyze the recorded events
1062 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1063 kernel>> and <<tracing-your-own-user-application,Trace a user
1064 application>> tutorials, you can inspect the recorded events.
1066 There are many tools you can use to read LTTng traces:
1068 * **cmd:babeltrace** is a command-line utility which converts trace
1069 formats; it supports the format that LTTng produces, CTF, as well as a
1070 basic text output which can be ++grep++ed. The cmd:babeltrace command
1071 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1072 * Babeltrace also includes
1073 **https://www.python.org/[Python{nbsp}3] bindings** so
1074 that you can easily open and read an LTTng trace with your own script,
1075 benefiting from the power of Python.
1076 * http://tracecompass.org/[**Trace Compass**]
1077 is a graphical user interface for viewing and analyzing any type of
1078 logs or traces, including LTTng's.
1079 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1080 project which includes many high-level analyses of LTTng kernel
1081 traces, like scheduling statistics, interrupt frequency distribution,
1082 top CPU usage, and more.
1084 NOTE: This section assumes that LTTng saved the traces it recorded
1085 during the previous tutorials to their default location, in the
1086 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
1087 environment variable defaults to `$HOME` if not set.
1090 [[viewing-and-analyzing-your-traces-bt]]
1091 ==== Use the cmd:babeltrace command-line tool
1093 The simplest way to list all the recorded events of a trace is to pass
1094 its path to cmd:babeltrace with no options:
1098 $ babeltrace ~/lttng-traces/my-user-space-session*
1101 cmd:babeltrace finds all traces recursively within the given path and
1102 prints all their events, merging them in chronological order.
1104 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1109 $ babeltrace /tmp/my-kernel-trace | grep _switch
1112 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1113 count the recorded events:
1117 $ babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
1121 [[viewing-and-analyzing-your-traces-bt-python]]
1122 ==== Use the Babeltrace{nbsp}1 Python bindings
1124 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1125 is useful to isolate events by simple matching using man:grep(1) and
1126 similar utilities. However, more elaborate filters, such as keeping only
1127 event records with a field value falling within a specific range, are
1128 not trivial to write using a shell. Moreover, reductions and even the
1129 most basic computations involving multiple event records are virtually
1130 impossible to implement.
1132 Fortunately, Babeltrace{nbsp}1 ships with Python{nbsp}3 bindings which
1133 makes it easy to read the event records of an LTTng trace sequentially
1134 and compute the desired information.
1136 The following script accepts an LTTng Linux kernel trace path as its
1137 first argument and prints the short names of the top five running
1138 processes on CPU{nbsp}0 during the whole trace:
1143 from collections import Counter
1149 if len(sys.argv) != 2:
1150 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1151 print(msg, file=sys.stderr)
1154 # A trace collection contains one or more traces
1155 col = babeltrace.TraceCollection()
1157 # Add the trace provided by the user (LTTng traces always have
1159 if col.add_trace(sys.argv[1], 'ctf') is None:
1160 raise RuntimeError('Cannot add trace')
1162 # This counter dict contains execution times:
1164 # task command name -> total execution time (ns)
1165 exec_times = Counter()
1167 # This contains the last `sched_switch` timestamp
1171 for event in col.events:
1172 # Keep only `sched_switch` events
1173 if event.name != 'sched_switch':
1176 # Keep only events which happened on CPU 0
1177 if event['cpu_id'] != 0:
1181 cur_ts = event.timestamp
1187 # Previous task command (short) name
1188 prev_comm = event['prev_comm']
1190 # Initialize entry in our dict if not yet done
1191 if prev_comm not in exec_times:
1192 exec_times[prev_comm] = 0
1194 # Compute previous command execution time
1195 diff = cur_ts - last_ts
1197 # Update execution time of this command
1198 exec_times[prev_comm] += diff
1200 # Update last timestamp
1204 for name, ns in exec_times.most_common(5):
1206 print('{:20}{} s'.format(name, s))
1211 if __name__ == '__main__':
1212 sys.exit(0 if top5proc() else 1)
1219 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1225 swapper/0 48.607245889 s
1226 chromium 7.192738188 s
1227 pavucontrol 0.709894415 s
1228 Compositor 0.660867933 s
1229 Xorg.bin 0.616753786 s
1232 Note that `swapper/0` is the "idle" process of CPU{nbsp}0 on Linux;
1233 since we weren't using the CPU that much when tracing, its first
1234 position in the list makes sense.
1238 == [[understanding-lttng]]Core concepts
1240 From a user's perspective, the LTTng system is built on a few concepts,
1241 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1242 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1243 Understanding how those objects relate to eachother is key in mastering
1246 The core concepts are:
1248 * <<tracing-session,Tracing session>>
1249 * <<domain,Tracing domain>>
1250 * <<channel,Channel and ring buffer>>
1251 * <<"event","Instrumentation point, event rule, event, and event record">>
1257 A _tracing session_ is a stateful dialogue between you and
1258 a <<lttng-sessiond,session daemon>>. You can
1259 <<creating-destroying-tracing-sessions,create a new tracing
1260 session>> with the `lttng create` command.
1262 Anything that you do when you control LTTng tracers happens within a
1263 tracing session. In particular, a tracing session:
1266 * Has its own set of trace files.
1267 * Has its own state of activity (started or stopped).
1268 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1270 * Has its own <<channel,channels>> to which are associated their own
1271 <<event,event rules>>.
1274 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1275 image::concepts.png[]
1277 Those attributes and objects are completely isolated between different
1280 A tracing session is analogous to a cash machine session:
1281 the operations you do on the banking system through the cash machine do
1282 not alter the data of other users of the same system. In the case of
1283 the cash machine, a session lasts as long as your bank card is inside.
1284 In the case of LTTng, a tracing session lasts from the `lttng create`
1285 command to the `lttng destroy` command.
1288 .Each Unix user has its own set of tracing sessions.
1289 image::many-sessions.png[]
1292 [[tracing-session-mode]]
1293 ==== Tracing session mode
1295 LTTng can send the generated trace data to different locations. The
1296 _tracing session mode_ dictates where to send it. The following modes
1297 are available in LTTng{nbsp}{revision}:
1300 LTTng writes the traces to the file system of the machine it traces
1303 Network streaming mode::
1304 LTTng sends the traces over the network to a
1305 <<lttng-relayd,relay daemon>> running on a remote system.
1308 LTTng does not write the traces by default. Instead, you can request
1309 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1310 tracing session's current sub-buffers, and to write it to the
1311 target's file system or to send it over the network to a
1312 <<lttng-relayd,relay daemon>> running on a remote system.
1315 This mode is similar to the network streaming mode, but a live
1316 trace viewer can connect to the distant relay daemon to
1317 <<lttng-live,view event records as LTTng generates them>>.
1323 A _tracing domain_ is a namespace for event sources. A tracing domain
1324 has its own properties and features.
1326 There are currently five available tracing domains:
1330 * `java.util.logging` (JUL)
1334 You must specify a tracing domain when using some commands to avoid
1335 ambiguity. For example, since all the domains support named tracepoints
1336 as event sources (instrumentation points that you manually insert in the
1337 source code), you need to specify a tracing domain when
1338 <<enabling-disabling-events,creating an event rule>> because all the
1339 tracing domains could have tracepoints with the same names.
1341 You can create <<channel,channels>> in the Linux kernel and user space
1342 tracing domains. The other tracing domains have a single default
1347 === Channel and ring buffer
1349 A _channel_ is an object which is responsible for a set of ring buffers.
1350 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1351 tracer emits an event, it can record it to one or more
1352 sub-buffers. The attributes of a channel determine what to do when
1353 there's no space left for a new event record because all sub-buffers
1354 are full, where to send a full sub-buffer, and other behaviours.
1356 A channel is always associated to a <<domain,tracing domain>>. The
1357 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1358 a default channel which you cannot configure.
1360 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1361 an event, it records it to the sub-buffers of all
1362 the enabled channels with a satisfied event rule, as long as those
1363 channels are part of active <<tracing-session,tracing sessions>>.
1366 [[channel-buffering-schemes]]
1367 ==== Per-user vs. per-process buffering schemes
1369 A channel has at least one ring buffer _per CPU_. LTTng always
1370 records an event to the ring buffer associated to the CPU on which it
1373 Two _buffering schemes_ are available when you
1374 <<enabling-disabling-channels,create a channel>> in the
1375 user space <<domain,tracing domain>>:
1377 Per-user buffering::
1378 Allocate one set of ring buffers--one per CPU--shared by all the
1379 instrumented processes of each Unix user.
1383 .Per-user buffering scheme.
1384 image::per-user-buffering.png[]
1387 Per-process buffering::
1388 Allocate one set of ring buffers--one per CPU--for each
1389 instrumented process.
1393 .Per-process buffering scheme.
1394 image::per-process-buffering.png[]
1397 The per-process buffering scheme tends to consume more memory than the
1398 per-user option because systems generally have more instrumented
1399 processes than Unix users running instrumented processes. However, the
1400 per-process buffering scheme ensures that one process having a high
1401 event throughput won't fill all the shared sub-buffers of the same
1404 The Linux kernel tracing domain has only one available buffering scheme
1405 which is to allocate a single set of ring buffers for the whole system.
1406 This scheme is similar to the per-user option, but with a single, global
1407 user "running" the kernel.
1410 [[channel-overwrite-mode-vs-discard-mode]]
1411 ==== Overwrite vs. discard event record loss modes
1413 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1414 arc in the following animations) of a specific channel's ring buffer.
1415 When there's no space left in a sub-buffer, the tracer marks it as
1416 consumable (red) and another, empty sub-buffer starts receiving the
1417 following event records. A <<lttng-consumerd,consumer daemon>>
1418 eventually consumes the marked sub-buffer (returns to white).
1421 [role="docsvg-channel-subbuf-anim"]
1426 In an ideal world, sub-buffers are consumed faster than they are filled,
1427 as it is the case in the previous animation. In the real world,
1428 however, all sub-buffers can be full at some point, leaving no space to
1429 record the following events.
1431 By default, LTTng-modules and LTTng-UST are _non-blocking_ tracers: when
1432 no empty sub-buffer is available, it is acceptable to lose event records
1433 when the alternative would be to cause substantial delays in the
1434 instrumented application's execution. LTTng privileges performance over
1435 integrity; it aims at perturbing the target system as little as possible
1436 in order to make tracing of subtle race conditions and rare interrupt
1439 Since LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST, supports
1440 a _blocking mode_. See the <<blocking-timeout-example,blocking timeout
1441 example>> to learn how to use the blocking mode.
1443 When it comes to losing event records because no empty sub-buffer is
1444 available, or because the <<opt-blocking-timeout,blocking timeout>> is
1445 reached, the channel's _event record loss mode_ determines what to do.
1446 The available event record loss modes are:
1449 Drop the newest event records until the tracer releases a sub-buffer.
1451 This is the only available mode when you specify a
1452 <<opt-blocking-timeout,blocking timeout>>.
1455 Clear the sub-buffer containing the oldest event records and start
1456 writing the newest event records there.
1458 This mode is sometimes called _flight recorder mode_ because it's
1460 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1461 always keep a fixed amount of the latest data.
1463 Which mechanism you should choose depends on your context: prioritize
1464 the newest or the oldest event records in the ring buffer?
1466 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1467 as soon as a there's no space left for a new event record, whereas in
1468 discard mode, the tracer only discards the event record that doesn't
1471 In discard mode, LTTng increments a count of lost event records when an
1472 event record is lost and saves this count to the trace. In overwrite
1473 mode, since LTTng{nbsp}2.8, LTTng increments a count of lost sub-buffers
1474 when a sub-buffer is lost and saves this count to the trace. In this
1475 mode, LTTng does not write to the trace the exact number of lost event
1476 records in those lost sub-buffers. Trace analyses can use the trace's
1477 saved discarded event record and sub-buffer counts to decide whether or
1478 not to perform the analyses even if trace data is known to be missing.
1480 There are a few ways to decrease your probability of losing event
1482 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1483 how you can fine-tune the sub-buffer count and size of a channel to
1484 virtually stop losing event records, though at the cost of greater
1488 [[channel-subbuf-size-vs-subbuf-count]]
1489 ==== Sub-buffer count and size
1491 When you <<enabling-disabling-channels,create a channel>>, you can
1492 set its number of sub-buffers and their size.
1494 Note that there is noticeable CPU overhead introduced when
1495 switching sub-buffers (marking a full one as consumable and switching
1496 to an empty one for the following events to be recorded). Knowing this,
1497 the following list presents a few practical situations along with how
1498 to configure the sub-buffer count and size for them:
1500 * **High event throughput**: In general, prefer bigger sub-buffers to
1501 lower the risk of losing event records.
1503 Having bigger sub-buffers also ensures a lower
1504 <<channel-switch-timer,sub-buffer switching frequency>>.
1506 The number of sub-buffers is only meaningful if you create the channel
1507 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1508 other sub-buffers are left unaltered.
1510 * **Low event throughput**: In general, prefer smaller sub-buffers
1511 since the risk of losing event records is low.
1513 Because events occur less frequently, the sub-buffer switching frequency
1514 should remain low and thus the tracer's overhead should not be a
1517 * **Low memory system**: If your target system has a low memory
1518 limit, prefer fewer first, then smaller sub-buffers.
1520 Even if the system is limited in memory, you want to keep the
1521 sub-buffers as big as possible to avoid a high sub-buffer switching
1524 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1525 which means event data is very compact. For example, the average
1526 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1527 sub-buffer size of 1{nbsp}MiB is considered big.
1529 The previous situations highlight the major trade-off between a few big
1530 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1531 frequency vs. how much data is lost in overwrite mode. Assuming a
1532 constant event throughput and using the overwrite mode, the two
1533 following configurations have the same ring buffer total size:
1536 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1541 * **Two sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1542 switching frequency, but if a sub-buffer overwrite happens, half of
1543 the event records so far (4{nbsp}MiB) are definitely lost.
1544 * **Eight sub-buffers of 1{nbsp}MiB each**: Expect four times the tracer's
1545 overhead as the previous configuration, but if a sub-buffer
1546 overwrite happens, only the eighth of event records so far are
1549 In discard mode, the sub-buffers count parameter is pointless: use two
1550 sub-buffers and set their size according to the requirements of your
1554 [[channel-switch-timer]]
1555 ==== Switch timer period
1557 The _switch timer period_ is an important configurable attribute of
1558 a channel to ensure periodic sub-buffer flushing.
1560 When the _switch timer_ expires, a sub-buffer switch happens. You can
1561 set the switch timer period attribute when you
1562 <<enabling-disabling-channels,create a channel>> to ensure that LTTng
1563 consumes and commits trace data to trace files or to a distant relay
1564 daemon periodically in case of a low event throughput.
1567 [role="docsvg-channel-switch-timer"]
1572 This attribute is also convenient when you use big sub-buffers to cope
1573 with a sporadic high event throughput, even if the throughput is
1577 [[channel-read-timer]]
1578 ==== Read timer period
1580 By default, the LTTng tracers use a notification mechanism to signal a
1581 full sub-buffer so that a consumer daemon can consume it. When such
1582 notifications must be avoided, for example in real-time applications,
1583 you can use the channel's _read timer_ instead. When the read timer
1584 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1585 consumable sub-buffers.
1588 [[tracefile-rotation]]
1589 ==== Trace file count and size
1591 By default, trace files can grow as large as needed. You can set the
1592 maximum size of each trace file that a channel writes when you
1593 <<enabling-disabling-channels,create a channel>>. When the size of
1594 a trace file reaches the channel's fixed maximum size, LTTng creates
1595 another file to contain the next event records. LTTng appends a file
1596 count to each trace file name in this case.
1598 If you set the trace file size attribute when you create a channel, the
1599 maximum number of trace files that LTTng creates is _unlimited_ by
1600 default. To limit them, you can also set a maximum number of trace
1601 files. When the number of trace files reaches the channel's fixed
1602 maximum count, the oldest trace file is overwritten. This mechanism is
1603 called _trace file rotation_.
1607 Even if you don't limit the trace file count, you cannot assume that
1608 LTTng doesn't manage any trace file.
1610 In other words, there is no safe way to know if LTTng still holds a
1611 given trace file open with the trace file rotation feature.
1613 The only way to obtain an unmanaged, self-contained LTTng trace before
1614 you <<creating-destroying-tracing-sessions,destroy>> the tracing session
1615 is with the <<session-rotation,tracing session rotation>> feature
1616 (available since LTTng{nbsp}2.11).
1621 === Instrumentation point, event rule, event, and event record
1623 An _event rule_ is a set of conditions which must be **all** satisfied
1624 for LTTng to record an occuring event.
1626 You set the conditions when you <<enabling-disabling-events,create
1629 You always attach an event rule to a <<channel,channel>> when you create
1632 When an event passes the conditions of an event rule, LTTng records it
1633 in one of the attached channel's sub-buffers.
1635 The available conditions, as of LTTng{nbsp}{revision}, are:
1637 * The event rule _is enabled_.
1638 * The instrumentation point's type _is{nbsp}T_.
1639 * The instrumentation point's name (sometimes called _event name_)
1640 _matches{nbsp}N_, but _is not{nbsp}E_.
1641 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1642 _is exactly{nbsp}L_.
1643 * The fields of the event's payload _satisfy_ a filter
1644 expression{nbsp}__F__.
1646 As you can see, all the conditions but the dynamic filter are related to
1647 the event rule's status or to the instrumentation point, not to the
1648 occurring events. This is why, without a filter, checking if an event
1649 passes an event rule is not a dynamic task: when you create or modify an
1650 event rule, all the tracers of its tracing domain enable or disable the
1651 instrumentation points themselves once. This is possible because the
1652 attributes of an instrumentation point (type, name, and log level) are
1653 defined statically. In other words, without a dynamic filter, the tracer
1654 _does not evaluate_ the arguments of an instrumentation point unless it
1655 matches an enabled event rule.
1657 Note that, for LTTng to record an event, the <<channel,channel>> to
1658 which a matching event rule is attached must also be enabled, and the
1659 <<tracing-session,tracing session>> owning this channel must be active
1663 .Logical path from an instrumentation point to an event record.
1664 image::event-rule.png[]
1666 .Event, event record, or event rule?
1668 With so many similar terms, it's easy to get confused.
1670 An **event** is the consequence of the execution of an _instrumentation
1671 point_, like a tracepoint that you manually place in some source code,
1672 or a Linux kernel kprobe. An event is said to _occur_ at a specific
1673 time. Different actions can be taken upon the occurrence of an event,
1674 like record the event's payload to a buffer.
1676 An **event record** is the representation of an event in a sub-buffer. A
1677 tracer is responsible for capturing the payload of an event, current
1678 context variables, the event's ID, and the event's timestamp. LTTng
1679 can append this sub-buffer to a trace file.
1681 An **event rule** is a set of conditions which must _all_ be satisfied
1682 for LTTng to record an occuring event. Events still occur without
1683 satisfying event rules, but LTTng does not record them.
1688 == Components of noch:{LTTng}
1690 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1691 to call LTTng a simple _tool_ since it is composed of multiple
1692 interacting components. This section describes those components,
1693 explains their respective roles, and shows how they connect together to
1694 form the LTTng ecosystem.
1696 The following diagram shows how the most important components of LTTng
1697 interact with user applications, the Linux kernel, and you:
1700 .Control and trace data paths between LTTng components.
1701 image::plumbing.png[]
1703 The LTTng project incorporates:
1705 * **LTTng-tools**: Libraries and command-line interface to
1706 control tracing sessions.
1707 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1708 ** <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
1709 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1710 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1711 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1712 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1714 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1715 headers to instrument and trace any native user application.
1716 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1717 *** `liblttng-ust-libc-wrapper`
1718 *** `liblttng-ust-pthread-wrapper`
1719 *** `liblttng-ust-cyg-profile`
1720 *** `liblttng-ust-cyg-profile-fast`
1721 *** `liblttng-ust-dl`
1722 ** User space tracepoint provider source files generator command-line
1723 tool (man:lttng-gen-tp(1)).
1724 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1725 Java applications using `java.util.logging` or
1726 Apache log4j{nbsp}1.2 logging.
1727 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1728 Python applications using the standard `logging` package.
1729 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1731 ** LTTng kernel tracer module.
1732 ** Tracing ring buffer kernel modules.
1733 ** Probe kernel modules.
1734 ** LTTng logger kernel module.
1738 === Tracing control command-line interface
1741 .The tracing control command-line interface.
1742 image::plumbing-lttng-cli.png[]
1744 The _man:lttng(1) command-line tool_ is the standard user interface to
1745 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1746 is part of LTTng-tools.
1748 The cmd:lttng tool is linked with
1749 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1750 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1752 The cmd:lttng tool has a Git-like interface:
1756 $ lttng <GENERAL OPTIONS> <COMMAND> <COMMAND OPTIONS>
1759 The <<controlling-tracing,Tracing control>> section explores the
1760 available features of LTTng using the cmd:lttng tool.
1763 [[liblttng-ctl-lttng]]
1764 === Tracing control library
1767 .The tracing control library.
1768 image::plumbing-liblttng-ctl.png[]
1770 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1771 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1772 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1774 The <<lttng-cli,cmd:lttng command-line tool>>
1775 is linked with `liblttng-ctl`.
1777 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1782 #include <lttng/lttng.h>
1785 Some objects are referenced by name (C string), such as tracing
1786 sessions, but most of them require to create a handle first using
1787 `lttng_create_handle()`.
1789 As of LTTng{nbsp}{revision}, the best available developer documentation for
1790 `liblttng-ctl` is its installed header files. Every function and structure is
1791 thoroughly documented.
1795 === User space tracing library
1798 .The user space tracing library.
1799 image::plumbing-liblttng-ust.png[]
1801 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1802 is the LTTng user space tracer. It receives commands from a
1803 <<lttng-sessiond,session daemon>>, for example to
1804 enable and disable specific instrumentation points, and writes event
1805 records to ring buffers shared with a
1806 <<lttng-consumerd,consumer daemon>>.
1807 `liblttng-ust` is part of LTTng-UST.
1809 Public C header files are installed beside `liblttng-ust` to
1810 instrument any <<c-application,C or $$C++$$ application>>.
1812 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1813 packages, use their own library providing tracepoints which is
1814 linked with `liblttng-ust`.
1816 An application or library does not have to initialize `liblttng-ust`
1817 manually: its constructor does the necessary tasks to properly register
1818 to a session daemon. The initialization phase also enables the
1819 instrumentation points matching the <<event,event rules>> that you
1823 [[lttng-ust-agents]]
1824 === User space tracing agents
1827 .The user space tracing agents.
1828 image::plumbing-lttng-ust-agents.png[]
1830 The _LTTng-UST Java and Python agents_ are regular Java and Python
1831 packages which add LTTng tracing capabilities to the
1832 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1834 In the case of Java, the
1835 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1836 core logging facilities] and
1837 https://logging.apache.org/log4j/1.2/[Apache log4j{nbsp}1.2] are supported.
1838 Note that Apache Log4{nbsp}2 is not supported.
1840 In the case of Python, the standard
1841 https://docs.python.org/3/library/logging.html[`logging`] package
1842 is supported. Both Python{nbsp}2 and Python{nbsp}3 modules can import the
1843 LTTng-UST Python agent package.
1845 The applications using the LTTng-UST agents are in the
1846 `java.util.logging` (JUL),
1847 log4j, and Python <<domain,tracing domains>>.
1849 Both agents use the same mechanism to trace the log statements. When an
1850 agent initializes, it creates a log handler that attaches to the root
1851 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1852 When the application executes a log statement, the root logger passes it
1853 to the agent's log handler. The agent's log handler calls a native
1854 function in a tracepoint provider package shared library linked with
1855 <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1856 other fields, like its logger name and its log level. This native
1857 function contains a user space instrumentation point, hence tracing the
1860 The log level condition of an
1861 <<event,event rule>> is considered when tracing
1862 a Java or a Python application, and it's compatible with the standard
1863 JUL, log4j, and Python log levels.
1867 === LTTng kernel modules
1870 .The LTTng kernel modules.
1871 image::plumbing-lttng-modules.png[]
1873 The _LTTng kernel modules_ are a set of Linux kernel modules
1874 which implement the kernel tracer of the LTTng project. The LTTng
1875 kernel modules are part of LTTng-modules.
1877 The LTTng kernel modules include:
1879 * A set of _probe_ modules.
1881 Each module attaches to a specific subsystem
1882 of the Linux kernel using its tracepoint instrument points. There are
1883 also modules to attach to the entry and return points of the Linux
1884 system call functions.
1886 * _Ring buffer_ modules.
1888 A ring buffer implementation is provided as kernel modules. The LTTng
1889 kernel tracer writes to the ring buffer; a
1890 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1892 * The _LTTng kernel tracer_ module.
1893 * The _LTTng logger_ module.
1895 The LTTng logger module implements the special path:{/proc/lttng-logger}
1896 file so that any executable can generate LTTng events by opening and
1897 writing to this file.
1899 See <<proc-lttng-logger-abi,LTTng logger>>.
1901 Generally, you do not have to load the LTTng kernel modules manually
1902 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1903 daemon>> loads the necessary modules when starting. If you have extra
1904 probe modules, you can specify to load them to the session daemon on
1907 The LTTng kernel modules are installed in
1908 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1909 the kernel release (see `uname --kernel-release`).
1916 .The session daemon.
1917 image::plumbing-sessiond.png[]
1919 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1920 managing tracing sessions and for controlling the various components of
1921 LTTng. The session daemon is part of LTTng-tools.
1923 The session daemon sends control requests to and receives control
1926 * The <<lttng-ust,user space tracing library>>.
1928 Any instance of the user space tracing library first registers to
1929 a session daemon. Then, the session daemon can send requests to
1930 this instance, such as:
1933 ** Get the list of tracepoints.
1934 ** Share an <<event,event rule>> so that the user space tracing library
1935 can enable or disable tracepoints. Amongst the possible conditions
1936 of an event rule is a filter expression which `liblttng-ust` evalutes
1937 when an event occurs.
1938 ** Share <<channel,channel>> attributes and ring buffer locations.
1941 The session daemon and the user space tracing library use a Unix
1942 domain socket for their communication.
1944 * The <<lttng-ust-agents,user space tracing agents>>.
1946 Any instance of a user space tracing agent first registers to
1947 a session daemon. Then, the session daemon can send requests to
1948 this instance, such as:
1951 ** Get the list of loggers.
1952 ** Enable or disable a specific logger.
1955 The session daemon and the user space tracing agent use a TCP connection
1956 for their communication.
1958 * The <<lttng-modules,LTTng kernel tracer>>.
1959 * The <<lttng-consumerd,consumer daemon>>.
1961 The session daemon sends requests to the consumer daemon to instruct
1962 it where to send the trace data streams, amongst other information.
1964 * The <<lttng-relayd,relay daemon>>.
1966 The session daemon receives commands from the
1967 <<liblttng-ctl-lttng,tracing control library>>.
1969 The root session daemon loads the appropriate
1970 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1971 a <<lttng-consumerd,consumer daemon>> as soon as you create
1972 an <<event,event rule>>.
1974 The session daemon does not send and receive trace data: this is the
1975 role of the <<lttng-consumerd,consumer daemon>> and
1976 <<lttng-relayd,relay daemon>>. It does, however, generate the
1977 http://diamon.org/ctf/[CTF] metadata stream.
1979 Each Unix user can have its own session daemon instance. The
1980 tracing sessions which different session daemons manage are completely
1983 The root user's session daemon is the only one which is
1984 allowed to control the LTTng kernel tracer, and its spawned consumer
1985 daemon is the only one which is allowed to consume trace data from the
1986 LTTng kernel tracer. Note, however, that any Unix user which is a member
1987 of the <<tracing-group,tracing group>> is allowed
1988 to create <<channel,channels>> in the
1989 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1992 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1993 session daemon when using its `create` command if none is currently
1994 running. You can also start the session daemon manually.
2001 .The consumer daemon.
2002 image::plumbing-consumerd.png[]
2004 The _consumer daemon_, cmd:lttng-consumerd, is a daemon which shares
2005 ring buffers with user applications or with the LTTng kernel modules to
2006 collect trace data and send it to some location (on disk or to a
2007 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2008 is part of LTTng-tools.
2010 You do not start a consumer daemon manually: a consumer daemon is always
2011 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2012 <<event,event rule>>, that is, before you start tracing. When you kill
2013 its owner session daemon, the consumer daemon also exits because it is
2014 the session daemon's child process. Command-line options of
2015 man:lttng-sessiond(8) target the consumer daemon process.
2017 There are up to two running consumer daemons per Unix user, whereas only
2018 one session daemon can run per user. This is because each process can be
2019 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2020 and 64-bit processes, it is more efficient to have separate
2021 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2022 exception: it can have up to _three_ running consumer daemons: 32-bit
2023 and 64-bit instances for its user applications, and one more
2024 reserved for collecting kernel trace data.
2032 image::plumbing-relayd.png[]
2034 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2035 between remote session and consumer daemons, local trace files, and a
2036 remote live trace viewer. The relay daemon is part of LTTng-tools.
2038 The main purpose of the relay daemon is to implement a receiver of
2039 <<sending-trace-data-over-the-network,trace data over the network>>.
2040 This is useful when the target system does not have much file system
2041 space to record trace files locally.
2043 The relay daemon is also a server to which a
2044 <<lttng-live,live trace viewer>> can
2045 connect. The live trace viewer sends requests to the relay daemon to
2046 receive trace data as the target system emits events. The
2047 communication protocol is named _LTTng live_; it is used over TCP
2050 Note that you can start the relay daemon on the target system directly.
2051 This is the setup of choice when the use case is to view events as
2052 the target system emits them without the need of a remote system.
2056 == [[using-lttng]]Instrumentation
2058 There are many examples of tracing and monitoring in our everyday life:
2060 * You have access to real-time and historical weather reports and
2061 forecasts thanks to weather stations installed around the country.
2062 * You know your heart is safe thanks to an electrocardiogram.
2063 * You make sure not to drive your car too fast and to have enough fuel
2064 to reach your destination thanks to gauges visible on your dashboard.
2066 All the previous examples have something in common: they rely on
2067 **instruments**. Without the electrodes attached to the surface of your
2068 body's skin, cardiac monitoring is futile.
2070 LTTng, as a tracer, is no different from those real life examples. If
2071 you're about to trace a software system or, in other words, record its
2072 history of execution, you better have **instrumentation points** in the
2073 subject you're tracing, that is, the actual software.
2075 Various ways were developed to instrument a piece of software for LTTng
2076 tracing. The most straightforward one is to manually place
2077 instrumentation points, called _tracepoints_, in the software's source
2078 code. It is also possible to add instrumentation points dynamically in
2079 the Linux kernel <<domain,tracing domain>>.
2081 If you're only interested in tracing the Linux kernel, your
2082 instrumentation needs are probably already covered by LTTng's built-in
2083 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2084 user application which is already instrumented for LTTng tracing.
2085 In such cases, you can skip this whole section and read the topics of
2086 the <<controlling-tracing,Tracing control>> section.
2088 Many methods are available to instrument a piece of software for LTTng
2091 * <<c-application,User space instrumentation for C and $$C++$$
2093 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2094 * <<java-application,User space Java agent>>.
2095 * <<python-application,User space Python agent>>.
2096 * <<proc-lttng-logger-abi,LTTng logger>>.
2097 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2101 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2103 The procedure to instrument a C or $$C++$$ user application with
2104 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2106 . <<tracepoint-provider,Create the source files of a tracepoint provider
2108 . <<probing-the-application-source-code,Add tracepoints to
2109 the application's source code>>.
2110 . <<building-tracepoint-providers-and-user-application,Build and link
2111 a tracepoint provider package and the user application>>.
2113 If you need quick, man:printf(3)-like instrumentation, you can skip
2114 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2117 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2118 instrument a user application with `liblttng-ust`.
2121 [[tracepoint-provider]]
2122 ==== Create the source files of a tracepoint provider package
2124 A _tracepoint provider_ is a set of compiled functions which provide
2125 **tracepoints** to an application, the type of instrumentation point
2126 supported by LTTng-UST. Those functions can emit events with
2127 user-defined fields and serialize those events as event records to one
2128 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2129 macro, which you <<probing-the-application-source-code,insert in a user
2130 application's source code>>, calls those functions.
2132 A _tracepoint provider package_ is an object file (`.o`) or a shared
2133 library (`.so`) which contains one or more tracepoint providers.
2134 Its source files are:
2136 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2137 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2139 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2140 the LTTng user space tracer, at run time.
2143 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2144 image::ust-app.png[]
2146 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2147 skip creating and using a tracepoint provider and use
2148 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2152 ===== Create a tracepoint provider header file template
2154 A _tracepoint provider header file_ contains the tracepoint
2155 definitions of a tracepoint provider.
2157 To create a tracepoint provider header file:
2159 . Start from this template:
2163 .Tracepoint provider header file template (`.h` file extension).
2165 #undef TRACEPOINT_PROVIDER
2166 #define TRACEPOINT_PROVIDER provider_name
2168 #undef TRACEPOINT_INCLUDE
2169 #define TRACEPOINT_INCLUDE "./tp.h"
2171 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2174 #include <lttng/tracepoint.h>
2177 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2178 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2183 #include <lttng/tracepoint-event.h>
2189 * `provider_name` with the name of your tracepoint provider.
2190 * `"tp.h"` with the name of your tracepoint provider header file.
2192 . Below the `#include <lttng/tracepoint.h>` line, put your
2193 <<defining-tracepoints,tracepoint definitions>>.
2195 Your tracepoint provider name must be unique amongst all the possible
2196 tracepoint provider names used on the same target system. We
2197 suggest to include the name of your project or company in the name,
2198 for example, `org_lttng_my_project_tpp`.
2200 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2201 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2202 write are the <<defining-tracepoints,tracepoint definitions>>.
2205 [[defining-tracepoints]]
2206 ===== Create a tracepoint definition
2208 A _tracepoint definition_ defines, for a given tracepoint:
2210 * Its **input arguments**. They are the macro parameters that the
2211 `tracepoint()` macro accepts for this particular tracepoint
2212 in the user application's source code.
2213 * Its **output event fields**. They are the sources of event fields
2214 that form the payload of any event that the execution of the
2215 `tracepoint()` macro emits for this particular tracepoint.
2217 You can create a tracepoint definition by using the
2218 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2220 <<tpp-header,tracepoint provider header file template>>.
2222 The syntax of the `TRACEPOINT_EVENT()` macro is:
2225 .`TRACEPOINT_EVENT()` macro syntax.
2228 /* Tracepoint provider name */
2231 /* Tracepoint name */
2234 /* Input arguments */
2239 /* Output event fields */
2248 * `provider_name` with your tracepoint provider name.
2249 * `tracepoint_name` with your tracepoint name.
2250 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2251 * `fields` with the <<tpp-def-output-fields,output event field>>
2254 This tracepoint emits events named `provider_name:tracepoint_name`.
2257 .Event name's length limitation
2259 The concatenation of the tracepoint provider name and the
2260 tracepoint name must not exceed **254{nbsp}characters**. If it does, the
2261 instrumented application compiles and runs, but LTTng throws multiple
2262 warnings and you could experience serious issues.
2265 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2268 .`TP_ARGS()` macro syntax.
2277 * `type` with the C type of the argument.
2278 * `arg_name` with the argument name.
2280 You can repeat `type` and `arg_name` up to 10{nbsp}times to have
2281 more than one argument.
2283 .`TP_ARGS()` usage with three arguments.
2295 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2296 tracepoint definition with no input arguments.
2298 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2299 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2300 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2301 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2304 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2305 C expression that the tracer evalutes at the `tracepoint()` macro site
2306 in the application's source code. This expression provides a field's
2307 source of data. The argument expression can include input argument names
2308 listed in the `TP_ARGS()` macro.
2310 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2311 must be unique within a given tracepoint definition.
2313 Here's a complete tracepoint definition example:
2315 .Tracepoint definition.
2317 The following tracepoint definition defines a tracepoint which takes
2318 three input arguments and has four output event fields.
2322 #include "my-custom-structure.h"
2328 const struct my_custom_structure*, my_custom_structure,
2333 ctf_string(query_field, query)
2334 ctf_float(double, ratio_field, ratio)
2335 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2336 ctf_integer(int, send_size, my_custom_structure->send_size)
2341 You can refer to this tracepoint definition with the `tracepoint()`
2342 macro in your application's source code like this:
2346 tracepoint(my_provider, my_tracepoint,
2347 my_structure, some_ratio, the_query);
2351 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2352 if they satisfy an enabled <<event,event rule>>.
2355 [[using-tracepoint-classes]]
2356 ===== Use a tracepoint class
2358 A _tracepoint class_ is a class of tracepoints which share the same
2359 output event field definitions. A _tracepoint instance_ is one
2360 instance of such a defined tracepoint class, with its own tracepoint
2363 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2364 shorthand which defines both a tracepoint class and a tracepoint
2365 instance at the same time.
2367 When you build a tracepoint provider package, the C or $$C++$$ compiler
2368 creates one serialization function for each **tracepoint class**. A
2369 serialization function is responsible for serializing the event fields
2370 of a tracepoint to a sub-buffer when tracing.
2372 For various performance reasons, when your situation requires multiple
2373 tracepoint definitions with different names, but with the same event
2374 fields, we recommend that you manually create a tracepoint class
2375 and instantiate as many tracepoint instances as needed. One positive
2376 effect of such a design, amongst other advantages, is that all
2377 tracepoint instances of the same tracepoint class reuse the same
2378 serialization function, thus reducing
2379 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2381 .Use a tracepoint class and tracepoint instances.
2383 Consider the following three tracepoint definitions:
2395 ctf_integer(int, userid, userid)
2396 ctf_integer(size_t, len, len)
2408 ctf_integer(int, userid, userid)
2409 ctf_integer(size_t, len, len)
2421 ctf_integer(int, userid, userid)
2422 ctf_integer(size_t, len, len)
2427 In this case, we create three tracepoint classes, with one implicit
2428 tracepoint instance for each of them: `get_account`, `get_settings`, and
2429 `get_transaction`. However, they all share the same event field names
2430 and types. Hence three identical, yet independent serialization
2431 functions are created when you build the tracepoint provider package.
2433 A better design choice is to define a single tracepoint class and three
2434 tracepoint instances:
2438 /* The tracepoint class */
2439 TRACEPOINT_EVENT_CLASS(
2440 /* Tracepoint provider name */
2443 /* Tracepoint class name */
2446 /* Input arguments */
2452 /* Output event fields */
2454 ctf_integer(int, userid, userid)
2455 ctf_integer(size_t, len, len)
2459 /* The tracepoint instances */
2460 TRACEPOINT_EVENT_INSTANCE(
2461 /* Tracepoint provider name */
2464 /* Tracepoint class name */
2467 /* Tracepoint name */
2470 /* Input arguments */
2476 TRACEPOINT_EVENT_INSTANCE(
2485 TRACEPOINT_EVENT_INSTANCE(
2498 [[assigning-log-levels]]
2499 ===== Assign a log level to a tracepoint definition
2501 You can assign an optional _log level_ to a
2502 <<defining-tracepoints,tracepoint definition>>.
2504 Assigning different levels of severity to tracepoint definitions can
2505 be useful: when you <<enabling-disabling-events,create an event rule>>,
2506 you can target tracepoints having a log level as severe as a specific
2509 The concept of LTTng-UST log levels is similar to the levels found
2510 in typical logging frameworks:
2512 * In a logging framework, the log level is given by the function
2513 or method name you use at the log statement site: `debug()`,
2514 `info()`, `warn()`, `error()`, and so on.
2515 * In LTTng-UST, you statically assign the log level to a tracepoint
2516 definition; any `tracepoint()` macro invocation which refers to
2517 this definition has this log level.
2519 You can assign a log level to a tracepoint definition with the
2520 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2521 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2522 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2525 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2528 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2530 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2535 * `provider_name` with the tracepoint provider name.
2536 * `tracepoint_name` with the tracepoint name.
2537 * `log_level` with the log level to assign to the tracepoint
2538 definition named `tracepoint_name` in the `provider_name`
2539 tracepoint provider.
2541 See man:lttng-ust(3) for a list of available log level names.
2543 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2547 /* Tracepoint definition */
2556 ctf_integer(int, userid, userid)
2557 ctf_integer(size_t, len, len)
2561 /* Log level assignment */
2562 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2568 ===== Create a tracepoint provider package source file
2570 A _tracepoint provider package source file_ is a C source file which
2571 includes a <<tpp-header,tracepoint provider header file>> to expand its
2572 macros into event serialization and other functions.
2574 You can always use the following tracepoint provider package source
2578 .Tracepoint provider package source file template.
2580 #define TRACEPOINT_CREATE_PROBES
2585 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2586 header file>> name. You may also include more than one tracepoint
2587 provider header file here to create a tracepoint provider package
2588 holding more than one tracepoint providers.
2591 [[probing-the-application-source-code]]
2592 ==== Add tracepoints to an application's source code
2594 Once you <<tpp-header,create a tracepoint provider header file>>, you
2595 can use the `tracepoint()` macro in your application's
2596 source code to insert the tracepoints that this header
2597 <<defining-tracepoints,defines>>.
2599 The `tracepoint()` macro takes at least two parameters: the tracepoint
2600 provider name and the tracepoint name. The corresponding tracepoint
2601 definition defines the other parameters.
2603 .`tracepoint()` usage.
2605 The following <<defining-tracepoints,tracepoint definition>> defines a
2606 tracepoint which takes two input arguments and has two output event
2610 .Tracepoint provider header file.
2612 #include "my-custom-structure.h"
2619 const char*, cmd_name
2622 ctf_string(cmd_name, cmd_name)
2623 ctf_integer(int, number_of_args, argc)
2628 You can refer to this tracepoint definition with the `tracepoint()`
2629 macro in your application's source code like this:
2632 .Application's source file.
2636 int main(int argc, char* argv[])
2638 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2644 Note how the application's source code includes
2645 the tracepoint provider header file containing the tracepoint
2646 definitions to use, path:{tp.h}.
2649 .`tracepoint()` usage with a complex tracepoint definition.
2651 Consider this complex tracepoint definition, where multiple event
2652 fields refer to the same input arguments in their argument expression
2656 .Tracepoint provider header file.
2658 /* For `struct stat` */
2659 #include <sys/types.h>
2660 #include <sys/stat.h>
2672 ctf_integer(int, my_constant_field, 23 + 17)
2673 ctf_integer(int, my_int_arg_field, my_int_arg)
2674 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2675 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2676 my_str_arg[2] + my_str_arg[3])
2677 ctf_string(my_str_arg_field, my_str_arg)
2678 ctf_integer_hex(off_t, size_field, st->st_size)
2679 ctf_float(double, size_dbl_field, (double) st->st_size)
2680 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2681 size_t, strlen(my_str_arg) / 2)
2686 You can refer to this tracepoint definition with the `tracepoint()`
2687 macro in your application's source code like this:
2690 .Application's source file.
2692 #define TRACEPOINT_DEFINE
2699 stat("/etc/fstab", &s);
2700 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2706 If you look at the event record that LTTng writes when tracing this
2707 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2708 it should look like this:
2710 .Event record fields
2712 |Field's name |Field's value
2713 |`my_constant_field` |40
2714 |`my_int_arg_field` |23
2715 |`my_int_arg_field2` |529
2717 |`my_str_arg_field` |`Hello, World!`
2718 |`size_field` |0x12d
2719 |`size_dbl_field` |301.0
2720 |`half_my_str_arg_field` |`Hello,`
2724 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2725 compute--they use the call stack, for example. To avoid this
2726 computation when the tracepoint is disabled, you can use the
2727 `tracepoint_enabled()` and `do_tracepoint()` macros.
2729 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2733 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2735 tracepoint_enabled(provider_name, tracepoint_name)
2736 do_tracepoint(provider_name, tracepoint_name, ...)
2741 * `provider_name` with the tracepoint provider name.
2742 * `tracepoint_name` with the tracepoint name.
2744 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2745 `tracepoint_name` from the provider named `provider_name` is enabled
2748 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2749 if the tracepoint is enabled. Using `tracepoint()` with
2750 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2751 the `tracepoint_enabled()` check, thus a race condition is
2752 possible in this situation:
2755 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2757 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2758 stuff = prepare_stuff();
2761 tracepoint(my_provider, my_tracepoint, stuff);
2764 If the tracepoint is enabled after the condition, then `stuff` is not
2765 prepared: the emitted event will either contain wrong data, or the whole
2766 application could crash (segmentation fault, for example).
2768 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2769 `STAP_PROBEV()` call. If you need it, you must emit
2773 [[building-tracepoint-providers-and-user-application]]
2774 ==== Build and link a tracepoint provider package and an application
2776 Once you have one or more <<tpp-header,tracepoint provider header
2777 files>> and a <<tpp-source,tracepoint provider package source file>>,
2778 you can create the tracepoint provider package by compiling its source
2779 file. From here, multiple build and run scenarios are possible. The
2780 following table shows common application and library configurations
2781 along with the required command lines to achieve them.
2783 In the following diagrams, we use the following file names:
2786 Executable application.
2789 Application's object file.
2792 Tracepoint provider package object file.
2795 Tracepoint provider package archive file.
2798 Tracepoint provider package shared object file.
2801 User library object file.
2804 User library shared object file.
2806 We use the following symbols in the diagrams of table below:
2809 .Symbols used in the build scenario diagrams.
2810 image::ust-sit-symbols.png[]
2812 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2813 variable in the following instructions.
2815 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2816 .Common tracepoint provider package scenarios.
2818 |Scenario |Instructions
2821 The instrumented application is statically linked with
2822 the tracepoint provider package object.
2824 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2827 include::../common/ust-sit-step-tp-o.txt[]
2829 To build the instrumented application:
2831 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2836 #define TRACEPOINT_DEFINE
2840 . Compile the application source file:
2849 . Build the application:
2854 $ gcc -o app app.o tpp.o -llttng-ust -ldl
2858 To run the instrumented application:
2860 * Start the application:
2870 The instrumented application is statically linked with the
2871 tracepoint provider package archive file.
2873 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2876 To create the tracepoint provider package archive file:
2878 . Compile the <<tpp-source,tracepoint provider package source file>>:
2887 . Create the tracepoint provider package archive file:
2892 $ ar rcs tpp.a tpp.o
2896 To build the instrumented application:
2898 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2903 #define TRACEPOINT_DEFINE
2907 . Compile the application source file:
2916 . Build the application:
2921 $ gcc -o app app.o tpp.a -llttng-ust -ldl
2925 To run the instrumented application:
2927 * Start the application:
2937 The instrumented application is linked with the tracepoint provider
2938 package shared object.
2940 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2943 include::../common/ust-sit-step-tp-so.txt[]
2945 To build the instrumented application:
2947 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2952 #define TRACEPOINT_DEFINE
2956 . Compile the application source file:
2965 . Build the application:
2970 $ gcc -o app app.o -ldl -L. -ltpp
2974 To run the instrumented application:
2976 * Start the application:
2986 The tracepoint provider package shared object is preloaded before the
2987 instrumented application starts.
2989 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2992 include::../common/ust-sit-step-tp-so.txt[]
2994 To build the instrumented application:
2996 . In path:{app.c}, before including path:{tpp.h}, add the
3002 #define TRACEPOINT_DEFINE
3003 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3007 . Compile the application source file:
3016 . Build the application:
3021 $ gcc -o app app.o -ldl
3025 To run the instrumented application with tracing support:
3027 * Preload the tracepoint provider package shared object and
3028 start the application:
3033 $ LD_PRELOAD=./libtpp.so ./app
3037 To run the instrumented application without tracing support:
3039 * Start the application:
3049 The instrumented application dynamically loads the tracepoint provider
3050 package shared object.
3052 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3055 include::../common/ust-sit-step-tp-so.txt[]
3057 To build the instrumented application:
3059 . In path:{app.c}, before including path:{tpp.h}, add the
3065 #define TRACEPOINT_DEFINE
3066 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3070 . Compile the application source file:
3079 . Build the application:
3084 $ gcc -o app app.o -ldl
3088 To run the instrumented application:
3090 * Start the application:
3100 The application is linked with the instrumented user library.
3102 The instrumented user library is statically linked with the tracepoint
3103 provider package object file.
3105 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3108 include::../common/ust-sit-step-tp-o-fpic.txt[]
3110 To build the instrumented user library:
3112 . In path:{emon.c}, before including path:{tpp.h}, add the
3118 #define TRACEPOINT_DEFINE
3122 . Compile the user library source file:
3127 $ gcc -I. -fpic -c emon.c
3131 . Build the user library shared object:
3136 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3140 To build the application:
3142 . Compile the application source file:
3151 . Build the application:
3156 $ gcc -o app app.o -L. -lemon
3160 To run the application:
3162 * Start the application:
3172 The application is linked with the instrumented user library.
3174 The instrumented user library is linked with the tracepoint provider
3175 package shared object.
3177 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3180 include::../common/ust-sit-step-tp-so.txt[]
3182 To build the instrumented user library:
3184 . In path:{emon.c}, before including path:{tpp.h}, add the
3190 #define TRACEPOINT_DEFINE
3194 . Compile the user library source file:
3199 $ gcc -I. -fpic -c emon.c
3203 . Build the user library shared object:
3208 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3212 To build the application:
3214 . Compile the application source file:
3223 . Build the application:
3228 $ gcc -o app app.o -L. -lemon
3232 To run the application:
3234 * Start the application:
3244 The tracepoint provider package shared object is preloaded before the
3247 The application is linked with the instrumented user library.
3249 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3252 include::../common/ust-sit-step-tp-so.txt[]
3254 To build the instrumented user library:
3256 . In path:{emon.c}, before including path:{tpp.h}, add the
3262 #define TRACEPOINT_DEFINE
3263 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3267 . Compile the user library source file:
3272 $ gcc -I. -fpic -c emon.c
3276 . Build the user library shared object:
3281 $ gcc -shared -o libemon.so emon.o -ldl
3285 To build the application:
3287 . Compile the application source file:
3296 . Build the application:
3301 $ gcc -o app app.o -L. -lemon
3305 To run the application with tracing support:
3307 * Preload the tracepoint provider package shared object and
3308 start the application:
3313 $ LD_PRELOAD=./libtpp.so ./app
3317 To run the application without tracing support:
3319 * Start the application:
3329 The application is linked with the instrumented user library.
3331 The instrumented user library dynamically loads the tracepoint provider
3332 package shared object.
3334 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3337 include::../common/ust-sit-step-tp-so.txt[]
3339 To build the instrumented user library:
3341 . In path:{emon.c}, before including path:{tpp.h}, add the
3347 #define TRACEPOINT_DEFINE
3348 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3352 . Compile the user library source file:
3357 $ gcc -I. -fpic -c emon.c
3361 . Build the user library shared object:
3366 $ gcc -shared -o libemon.so emon.o -ldl
3370 To build the application:
3372 . Compile the application source file:
3381 . Build the application:
3386 $ gcc -o app app.o -L. -lemon
3390 To run the application:
3392 * Start the application:
3402 The application dynamically loads the instrumented user library.
3404 The instrumented user library is linked with the tracepoint provider
3405 package shared object.
3407 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3410 include::../common/ust-sit-step-tp-so.txt[]
3412 To build the instrumented user library:
3414 . In path:{emon.c}, before including path:{tpp.h}, add the
3420 #define TRACEPOINT_DEFINE
3424 . Compile the user library source file:
3429 $ gcc -I. -fpic -c emon.c
3433 . Build the user library shared object:
3438 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3442 To build the application:
3444 . Compile the application source file:
3453 . Build the application:
3458 $ gcc -o app app.o -ldl -L. -lemon
3462 To run the application:
3464 * Start the application:
3474 The application dynamically loads the instrumented user library.
3476 The instrumented user library dynamically loads the tracepoint provider
3477 package shared object.
3479 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3482 include::../common/ust-sit-step-tp-so.txt[]
3484 To build the instrumented user library:
3486 . In path:{emon.c}, before including path:{tpp.h}, add the
3492 #define TRACEPOINT_DEFINE
3493 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3497 . Compile the user library source file:
3502 $ gcc -I. -fpic -c emon.c
3506 . Build the user library shared object:
3511 $ gcc -shared -o libemon.so emon.o -ldl
3515 To build the application:
3517 . Compile the application source file:
3526 . Build the application:
3531 $ gcc -o app app.o -ldl -L. -lemon
3535 To run the application:
3537 * Start the application:
3547 The tracepoint provider package shared object is preloaded before the
3550 The application dynamically loads the instrumented user library.
3552 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3555 include::../common/ust-sit-step-tp-so.txt[]
3557 To build the instrumented user library:
3559 . In path:{emon.c}, before including path:{tpp.h}, add the
3565 #define TRACEPOINT_DEFINE
3566 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3570 . Compile the user library source file:
3575 $ gcc -I. -fpic -c emon.c
3579 . Build the user library shared object:
3584 $ gcc -shared -o libemon.so emon.o -ldl
3588 To build the application:
3590 . Compile the application source file:
3599 . Build the application:
3604 $ gcc -o app app.o -L. -lemon
3608 To run the application with tracing support:
3610 * Preload the tracepoint provider package shared object and
3611 start the application:
3616 $ LD_PRELOAD=./libtpp.so ./app
3620 To run the application without tracing support:
3622 * Start the application:
3632 The application is statically linked with the tracepoint provider
3633 package object file.
3635 The application is linked with the instrumented user library.
3637 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3640 include::../common/ust-sit-step-tp-o.txt[]
3642 To build the instrumented user library:
3644 . In path:{emon.c}, before including path:{tpp.h}, add the
3650 #define TRACEPOINT_DEFINE
3654 . Compile the user library source file:
3659 $ gcc -I. -fpic -c emon.c
3663 . Build the user library shared object:
3668 $ gcc -shared -o libemon.so emon.o
3672 To build the application:
3674 . Compile the application source file:
3683 . Build the application:
3688 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3692 To run the instrumented application:
3694 * Start the application:
3704 The application is statically linked with the tracepoint provider
3705 package object file.
3707 The application dynamically loads the instrumented user library.
3709 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3712 include::../common/ust-sit-step-tp-o.txt[]
3714 To build the application:
3716 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3721 #define TRACEPOINT_DEFINE
3725 . Compile the application source file:
3734 . Build the application:
3739 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
3744 The `--export-dynamic` option passed to the linker is necessary for the
3745 dynamically loaded library to ``see'' the tracepoint symbols defined in
3748 To build the instrumented user library:
3750 . Compile the user library source file:
3755 $ gcc -I. -fpic -c emon.c
3759 . Build the user library shared object:
3764 $ gcc -shared -o libemon.so emon.o
3768 To run the application:
3770 * Start the application:
3781 [[using-lttng-ust-with-daemons]]
3782 ===== Use noch:{LTTng-UST} with daemons
3784 If your instrumented application calls man:fork(2), man:clone(2),
3785 or BSD's man:rfork(2), without a following man:exec(3)-family
3786 system call, you must preload the path:{liblttng-ust-fork.so} shared
3787 object when you start the application.
3791 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
3794 If your tracepoint provider package is
3795 a shared library which you also preload, you must put both
3796 shared objects in env:LD_PRELOAD:
3800 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3806 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3808 If your instrumented application closes one or more file descriptors
3809 which it did not open itself, you must preload the
3810 path:{liblttng-ust-fd.so} shared object when you start the application:
3814 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
3817 Typical use cases include closing all the file descriptors after
3818 man:fork(2) or man:rfork(2) and buggy applications doing
3822 [[lttng-ust-pkg-config]]
3823 ===== Use noch:{pkg-config}
3825 On some distributions, LTTng-UST ships with a
3826 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3827 metadata file. If this is your case, then you can use cmd:pkg-config to
3828 build an application on the command line:
3832 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3836 [[instrumenting-32-bit-app-on-64-bit-system]]
3837 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3839 In order to trace a 32-bit application running on a 64-bit system,
3840 LTTng must use a dedicated 32-bit
3841 <<lttng-consumerd,consumer daemon>>.
3843 The following steps show how to build and install a 32-bit consumer
3844 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3845 build and install the 32-bit LTTng-UST libraries, and how to build and
3846 link an instrumented 32-bit application in that context.
3848 To build a 32-bit instrumented application for a 64-bit target system,
3849 assuming you have a fresh target system with no installed Userspace RCU
3852 . Download, build, and install a 32-bit version of Userspace RCU:
3857 $ cd $(mktemp -d) &&
3858 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3859 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3860 cd userspace-rcu-0.9.* &&
3861 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3863 sudo make install &&
3868 . Using your distribution's package manager, or from source, install
3869 the following 32-bit versions of the following dependencies of
3870 LTTng-tools and LTTng-UST:
3873 * https://sourceforge.net/projects/libuuid/[libuuid]
3874 * http://directory.fsf.org/wiki/Popt[popt]
3875 * http://www.xmlsoft.org/[libxml2]
3878 . Download, build, and install a 32-bit version of the latest
3879 LTTng-UST{nbsp}{revision}:
3884 $ cd $(mktemp -d) &&
3885 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.11.tar.bz2 &&
3886 tar -xf lttng-ust-latest-2.11.tar.bz2 &&
3887 cd lttng-ust-2.11.* &&
3888 ./configure --libdir=/usr/local/lib32 \
3889 CFLAGS=-m32 CXXFLAGS=-m32 \
3890 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3892 sudo make install &&
3899 Depending on your distribution,
3900 32-bit libraries could be installed at a different location than
3901 `/usr/lib32`. For example, Debian is known to install
3902 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3904 In this case, make sure to set `LDFLAGS` to all the
3905 relevant 32-bit library paths, for example:
3909 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3913 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3914 the 32-bit consumer daemon:
3919 $ cd $(mktemp -d) &&
3920 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
3921 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
3922 cd lttng-tools-2.11.* &&
3923 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3924 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
3925 --disable-bin-lttng --disable-bin-lttng-crash \
3926 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
3928 cd src/bin/lttng-consumerd &&
3929 sudo make install &&
3934 . From your distribution or from source,
3935 <<installing-lttng,install>> the 64-bit versions of
3936 LTTng-UST and Userspace RCU.
3937 . Download, build, and install the 64-bit version of the
3938 latest LTTng-tools{nbsp}{revision}:
3943 $ cd $(mktemp -d) &&
3944 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
3945 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
3946 cd lttng-tools-2.11.* &&
3947 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3948 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3950 sudo make install &&
3955 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3956 when linking your 32-bit application:
3959 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3960 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3963 For example, let's rebuild the quick start example in
3964 <<tracing-your-own-user-application,Trace a user application>> as an
3965 instrumented 32-bit application:
3970 $ gcc -m32 -c -I. hello-tp.c
3971 $ gcc -m32 -c hello.c
3972 $ gcc -m32 -o hello hello.o hello-tp.o \
3973 -L/usr/lib32 -L/usr/local/lib32 \
3974 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3979 No special action is required to execute the 32-bit application and
3980 to trace it: use the command-line man:lttng(1) tool as usual.
3987 man:tracef(3) is a small LTTng-UST API designed for quick,
3988 man:printf(3)-like instrumentation without the burden of
3989 <<tracepoint-provider,creating>> and
3990 <<building-tracepoint-providers-and-user-application,building>>
3991 a tracepoint provider package.
3993 To use `tracef()` in your application:
3995 . In the C or C++ source files where you need to use `tracef()`,
3996 include `<lttng/tracef.h>`:
4001 #include <lttng/tracef.h>
4005 . In the application's source code, use `tracef()` like you would use
4013 tracef("my message: %d (%s)", my_integer, my_string);
4019 . Link your application with `liblttng-ust`:
4024 $ gcc -o app app.c -llttng-ust
4028 To trace the events that `tracef()` calls emit:
4030 * <<enabling-disabling-events,Create an event rule>> which matches the
4031 `lttng_ust_tracef:*` event name:
4036 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
4041 .Limitations of `tracef()`
4043 The `tracef()` utility function was developed to make user space tracing
4044 super simple, albeit with notable disadvantages compared to
4045 <<defining-tracepoints,user-defined tracepoints>>:
4047 * All the emitted events have the same tracepoint provider and
4048 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4049 * There is no static type checking.
4050 * The only event record field you actually get, named `msg`, is a string
4051 potentially containing the values you passed to `tracef()`
4052 using your own format string. This also means that you cannot filter
4053 events with a custom expression at run time because there are no
4055 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4056 function behind the scenes to format the strings at run time, its
4057 expected performance is lower than with user-defined tracepoints,
4058 which do not require a conversion to a string.
4060 Taking this into consideration, `tracef()` is useful for some quick
4061 prototyping and debugging, but you should not consider it for any
4062 permanent and serious applicative instrumentation.
4068 ==== Use `tracelog()`
4070 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4071 the difference that it accepts an additional log level parameter.
4073 The goal of `tracelog()` is to ease the migration from logging to
4076 To use `tracelog()` in your application:
4078 . In the C or C++ source files where you need to use `tracelog()`,
4079 include `<lttng/tracelog.h>`:
4084 #include <lttng/tracelog.h>
4088 . In the application's source code, use `tracelog()` like you would use
4089 man:printf(3), except for the first parameter which is the log
4097 tracelog(TRACE_WARNING, "my message: %d (%s)",
4098 my_integer, my_string);
4104 See man:lttng-ust(3) for a list of available log level names.
4106 . Link your application with `liblttng-ust`:
4111 $ gcc -o app app.c -llttng-ust
4115 To trace the events that `tracelog()` calls emit with a log level
4116 _as severe as_ a specific log level:
4118 * <<enabling-disabling-events,Create an event rule>> which matches the
4119 `lttng_ust_tracelog:*` event name and a minimum level
4125 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4126 --loglevel=TRACE_WARNING
4130 To trace the events that `tracelog()` calls emit with a
4131 _specific log level_:
4133 * Create an event rule which matches the `lttng_ust_tracelog:*`
4134 event name and a specific log level:
4139 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4140 --loglevel-only=TRACE_INFO
4145 [[prebuilt-ust-helpers]]
4146 === Prebuilt user space tracing helpers
4148 The LTTng-UST package provides a few helpers in the form or preloadable
4149 shared objects which automatically instrument system functions and
4152 The helper shared objects are normally found in dir:{/usr/lib}. If you
4153 built LTTng-UST <<building-from-source,from source>>, they are probably
4154 located in dir:{/usr/local/lib}.
4156 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4159 path:{liblttng-ust-libc-wrapper.so}::
4160 path:{liblttng-ust-pthread-wrapper.so}::
4161 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4162 memory and POSIX threads function tracing>>.
4164 path:{liblttng-ust-cyg-profile.so}::
4165 path:{liblttng-ust-cyg-profile-fast.so}::
4166 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4168 path:{liblttng-ust-dl.so}::
4169 <<liblttng-ust-dl,Dynamic linker tracing>>.
4171 To use a user space tracing helper with any user application:
4173 * Preload the helper shared object when you start the application:
4178 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4182 You can preload more than one helper:
4187 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4193 [[liblttng-ust-libc-pthread-wrapper]]
4194 ==== Instrument C standard library memory and POSIX threads functions
4196 The path:{liblttng-ust-libc-wrapper.so} and
4197 path:{liblttng-ust-pthread-wrapper.so} helpers
4198 add instrumentation to some C standard library and POSIX
4202 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4204 |TP provider name |TP name |Instrumented function
4206 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4207 |`calloc` |man:calloc(3)
4208 |`realloc` |man:realloc(3)
4209 |`free` |man:free(3)
4210 |`memalign` |man:memalign(3)
4211 |`posix_memalign` |man:posix_memalign(3)
4215 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4217 |TP provider name |TP name |Instrumented function
4219 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4220 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4221 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4222 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4225 When you preload the shared object, it replaces the functions listed
4226 in the previous tables by wrappers which contain tracepoints and call
4227 the replaced functions.
4230 [[liblttng-ust-cyg-profile]]
4231 ==== Instrument function entry and exit
4233 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4234 to the entry and exit points of functions.
4236 man:gcc(1) and man:clang(1) have an option named
4237 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4238 which generates instrumentation calls for entry and exit to functions.
4239 The LTTng-UST function tracing helpers,
4240 path:{liblttng-ust-cyg-profile.so} and
4241 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4242 to add tracepoints to the two generated functions (which contain
4243 `cyg_profile` in their names, hence the helper's name).
4245 To use the LTTng-UST function tracing helper, the source files to
4246 instrument must be built using the `-finstrument-functions` compiler
4249 There are two versions of the LTTng-UST function tracing helper:
4251 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4252 that you should only use when it can be _guaranteed_ that the
4253 complete event stream is recorded without any lost event record.
4254 Any kind of duplicate information is left out.
4256 Assuming no event record is lost, having only the function addresses on
4257 entry is enough to create a call graph, since an event record always
4258 contains the ID of the CPU that generated it.
4260 You can use a tool like man:addr2line(1) to convert function addresses
4261 back to source file names and line numbers.
4263 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4264 which also works in use cases where event records might get discarded or
4265 not recorded from application startup.
4266 In these cases, the trace analyzer needs more information to be
4267 able to reconstruct the program flow.
4269 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4270 points of this helper.
4272 All the tracepoints that this helper provides have the
4273 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4275 TIP: It's sometimes a good idea to limit the number of source files that
4276 you compile with the `-finstrument-functions` option to prevent LTTng
4277 from writing an excessive amount of trace data at run time. When using
4278 man:gcc(1), you can use the
4279 `-finstrument-functions-exclude-function-list` option to avoid
4280 instrument entries and exits of specific function names.
4285 ==== Instrument the dynamic linker
4287 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4288 man:dlopen(3) and man:dlclose(3) function calls.
4290 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4295 [[java-application]]
4296 === User space Java agent
4298 You can instrument any Java application which uses one of the following
4301 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4302 (JUL) core logging facilities.
4303 * http://logging.apache.org/log4j/1.2/[**Apache log4j{nbsp}1.2**], since
4304 LTTng{nbsp}2.6. Note that Apache Log4j{nbsp}2 is not supported.
4307 .LTTng-UST Java agent imported by a Java application.
4308 image::java-app.png[]
4310 Note that the methods described below are new in LTTng{nbsp}{revision}.
4311 Previous LTTng versions use another technique.
4313 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4314 and https://ci.lttng.org/[continuous integration], thus this version is
4315 directly supported. However, the LTTng-UST Java agent is also tested
4316 with OpenJDK{nbsp}7.
4321 ==== Use the LTTng-UST Java agent for `java.util.logging`
4323 To use the LTTng-UST Java agent in a Java application which uses
4324 `java.util.logging` (JUL):
4326 . In the Java application's source code, import the LTTng-UST
4327 log handler package for `java.util.logging`:
4332 import org.lttng.ust.agent.jul.LttngLogHandler;
4336 . Create an LTTng-UST JUL log handler:
4341 Handler lttngUstLogHandler = new LttngLogHandler();
4345 . Add this handler to the JUL loggers which should emit LTTng events:
4350 Logger myLogger = Logger.getLogger("some-logger");
4352 myLogger.addHandler(lttngUstLogHandler);
4356 . Use `java.util.logging` log statements and configuration as usual.
4357 The loggers with an attached LTTng-UST log handler can emit
4360 . Before exiting the application, remove the LTTng-UST log handler from
4361 the loggers attached to it and call its `close()` method:
4366 myLogger.removeHandler(lttngUstLogHandler);
4367 lttngUstLogHandler.close();
4371 This is not strictly necessary, but it is recommended for a clean
4372 disposal of the handler's resources.
4374 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4375 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4377 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4378 path] when you build the Java application.
4380 The JAR files are typically located in dir:{/usr/share/java}.
4382 IMPORTANT: The LTTng-UST Java agent must be
4383 <<installing-lttng,installed>> for the logging framework your
4386 .Use the LTTng-UST Java agent for `java.util.logging`.
4391 import java.io.IOException;
4392 import java.util.logging.Handler;
4393 import java.util.logging.Logger;
4394 import org.lttng.ust.agent.jul.LttngLogHandler;
4398 private static final int answer = 42;
4400 public static void main(String[] argv) throws Exception
4403 Logger logger = Logger.getLogger("jello");
4405 // Create an LTTng-UST log handler
4406 Handler lttngUstLogHandler = new LttngLogHandler();
4408 // Add the LTTng-UST log handler to our logger
4409 logger.addHandler(lttngUstLogHandler);
4412 logger.info("some info");
4413 logger.warning("some warning");
4415 logger.finer("finer information; the answer is " + answer);
4417 logger.severe("error!");
4419 // Not mandatory, but cleaner
4420 logger.removeHandler(lttngUstLogHandler);
4421 lttngUstLogHandler.close();
4430 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4433 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4434 <<enabling-disabling-events,create an event rule>> matching the
4435 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4440 $ lttng enable-event --jul jello
4444 Run the compiled class:
4448 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4451 <<basic-tracing-session-control,Stop tracing>> and inspect the
4461 In the resulting trace, an <<event,event record>> generated by a Java
4462 application using `java.util.logging` is named `lttng_jul:event` and
4463 has the following fields:
4466 Log record's message.
4472 Name of the class in which the log statement was executed.
4475 Name of the method in which the log statement was executed.
4478 Logging time (timestamp in milliseconds).
4481 Log level integer value.
4484 ID of the thread in which the log statement was executed.
4486 You can use the opt:lttng-enable-event(1):--loglevel or
4487 opt:lttng-enable-event(1):--loglevel-only option of the
4488 man:lttng-enable-event(1) command to target a range of JUL log levels
4489 or a specific JUL log level.
4494 ==== Use the LTTng-UST Java agent for Apache log4j
4496 To use the LTTng-UST Java agent in a Java application which uses
4497 Apache log4j{nbsp}1.2:
4499 . In the Java application's source code, import the LTTng-UST
4500 log appender package for Apache log4j:
4505 import org.lttng.ust.agent.log4j.LttngLogAppender;
4509 . Create an LTTng-UST log4j log appender:
4514 Appender lttngUstLogAppender = new LttngLogAppender();
4518 . Add this appender to the log4j loggers which should emit LTTng events:
4523 Logger myLogger = Logger.getLogger("some-logger");
4525 myLogger.addAppender(lttngUstLogAppender);
4529 . Use Apache log4j log statements and configuration as usual. The
4530 loggers with an attached LTTng-UST log appender can emit LTTng events.
4532 . Before exiting the application, remove the LTTng-UST log appender from
4533 the loggers attached to it and call its `close()` method:
4538 myLogger.removeAppender(lttngUstLogAppender);
4539 lttngUstLogAppender.close();
4543 This is not strictly necessary, but it is recommended for a clean
4544 disposal of the appender's resources.
4546 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4547 files, path:{lttng-ust-agent-common.jar} and
4548 path:{lttng-ust-agent-log4j.jar}, in the
4549 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4550 path] when you build the Java application.
4552 The JAR files are typically located in dir:{/usr/share/java}.
4554 IMPORTANT: The LTTng-UST Java agent must be
4555 <<installing-lttng,installed>> for the logging framework your
4558 .Use the LTTng-UST Java agent for Apache log4j.
4563 import org.apache.log4j.Appender;
4564 import org.apache.log4j.Logger;
4565 import org.lttng.ust.agent.log4j.LttngLogAppender;
4569 private static final int answer = 42;
4571 public static void main(String[] argv) throws Exception
4574 Logger logger = Logger.getLogger("jello");
4576 // Create an LTTng-UST log appender
4577 Appender lttngUstLogAppender = new LttngLogAppender();
4579 // Add the LTTng-UST log appender to our logger
4580 logger.addAppender(lttngUstLogAppender);
4583 logger.info("some info");
4584 logger.warn("some warning");
4586 logger.debug("debug information; the answer is " + answer);
4588 logger.fatal("error!");
4590 // Not mandatory, but cleaner
4591 logger.removeAppender(lttngUstLogAppender);
4592 lttngUstLogAppender.close();
4598 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4603 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4606 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4607 <<enabling-disabling-events,create an event rule>> matching the
4608 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4613 $ lttng enable-event --log4j jello
4617 Run the compiled class:
4621 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4624 <<basic-tracing-session-control,Stop tracing>> and inspect the
4634 In the resulting trace, an <<event,event record>> generated by a Java
4635 application using log4j is named `lttng_log4j:event` and
4636 has the following fields:
4639 Log record's message.
4645 Name of the class in which the log statement was executed.
4648 Name of the method in which the log statement was executed.
4651 Name of the file in which the executed log statement is located.
4654 Line number at which the log statement was executed.
4660 Log level integer value.
4663 Name of the Java thread in which the log statement was executed.
4665 You can use the opt:lttng-enable-event(1):--loglevel or
4666 opt:lttng-enable-event(1):--loglevel-only option of the
4667 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4668 or a specific log4j log level.
4672 [[java-application-context]]
4673 ==== Provide application-specific context fields in a Java application
4675 A Java application-specific context field is a piece of state provided
4676 by the application which <<adding-context,you can add>>, using the
4677 man:lttng-add-context(1) command, to each <<event,event record>>
4678 produced by the log statements of this application.
4680 For example, a given object might have a current request ID variable.
4681 You can create a context information retriever for this object and
4682 assign a name to this current request ID. You can then, using the
4683 man:lttng-add-context(1) command, add this context field by name to
4684 the JUL or log4j <<channel,channel>>.
4686 To provide application-specific context fields in a Java application:
4688 . In the Java application's source code, import the LTTng-UST
4689 Java agent context classes and interfaces:
4694 import org.lttng.ust.agent.context.ContextInfoManager;
4695 import org.lttng.ust.agent.context.IContextInfoRetriever;
4699 . Create a context information retriever class, that is, a class which
4700 implements the `IContextInfoRetriever` interface:
4705 class MyContextInfoRetriever implements IContextInfoRetriever
4708 public Object retrieveContextInfo(String key)
4710 if (key.equals("intCtx")) {
4712 } else if (key.equals("strContext")) {
4713 return "context value!";
4722 This `retrieveContextInfo()` method is the only member of the
4723 `IContextInfoRetriever` interface. Its role is to return the current
4724 value of a state by name to create a context field. The names of the
4725 context fields and which state variables they return depends on your
4728 All primitive types and objects are supported as context fields.
4729 When `retrieveContextInfo()` returns an object, the context field
4730 serializer calls its `toString()` method to add a string field to
4731 event records. The method can also return `null`, which means that
4732 no context field is available for the required name.
4734 . Register an instance of your context information retriever class to
4735 the context information manager singleton:
4740 IContextInfoRetriever cir = new MyContextInfoRetriever();
4741 ContextInfoManager cim = ContextInfoManager.getInstance();
4742 cim.registerContextInfoRetriever("retrieverName", cir);
4746 . Before exiting the application, remove your context information
4747 retriever from the context information manager singleton:
4752 ContextInfoManager cim = ContextInfoManager.getInstance();
4753 cim.unregisterContextInfoRetriever("retrieverName");
4757 This is not strictly necessary, but it is recommended for a clean
4758 disposal of some manager's resources.
4760 . Build your Java application with LTTng-UST Java agent support as
4761 usual, following the procedure for either the <<jul,JUL>> or
4762 <<log4j,Apache log4j>> framework.
4765 .Provide application-specific context fields in a Java application.
4770 import java.util.logging.Handler;
4771 import java.util.logging.Logger;
4772 import org.lttng.ust.agent.jul.LttngLogHandler;
4773 import org.lttng.ust.agent.context.ContextInfoManager;
4774 import org.lttng.ust.agent.context.IContextInfoRetriever;
4778 // Our context information retriever class
4779 private static class MyContextInfoRetriever
4780 implements IContextInfoRetriever
4783 public Object retrieveContextInfo(String key) {
4784 if (key.equals("intCtx")) {
4786 } else if (key.equals("strContext")) {
4787 return "context value!";
4794 private static final int answer = 42;
4796 public static void main(String args[]) throws Exception
4798 // Get the context information manager instance
4799 ContextInfoManager cim = ContextInfoManager.getInstance();
4801 // Create and register our context information retriever
4802 IContextInfoRetriever cir = new MyContextInfoRetriever();
4803 cim.registerContextInfoRetriever("myRetriever", cir);
4806 Logger logger = Logger.getLogger("jello");
4808 // Create an LTTng-UST log handler
4809 Handler lttngUstLogHandler = new LttngLogHandler();
4811 // Add the LTTng-UST log handler to our logger
4812 logger.addHandler(lttngUstLogHandler);
4815 logger.info("some info");
4816 logger.warning("some warning");
4818 logger.finer("finer information; the answer is " + answer);
4820 logger.severe("error!");
4822 // Not mandatory, but cleaner
4823 logger.removeHandler(lttngUstLogHandler);
4824 lttngUstLogHandler.close();
4825 cim.unregisterContextInfoRetriever("myRetriever");
4834 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4837 <<creating-destroying-tracing-sessions,Create a tracing session>>
4838 and <<enabling-disabling-events,create an event rule>> matching the
4844 $ lttng enable-event --jul jello
4847 <<adding-context,Add the application-specific context fields>> to the
4852 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
4853 $ lttng add-context --jul --type='$app.myRetriever:strContext'
4856 <<basic-tracing-session-control,Start tracing>>:
4863 Run the compiled class:
4867 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4870 <<basic-tracing-session-control,Stop tracing>> and inspect the
4882 [[python-application]]
4883 === User space Python agent
4885 You can instrument a Python{nbsp}2 or Python{nbsp}3 application which
4887 https://docs.python.org/3/library/logging.html[`logging`] package.
4889 Each log statement emits an LTTng event once the
4890 application module imports the
4891 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4894 .A Python application importing the LTTng-UST Python agent.
4895 image::python-app.png[]
4897 To use the LTTng-UST Python agent:
4899 . In the Python application's source code, import the LTTng-UST Python
4909 The LTTng-UST Python agent automatically adds its logging handler to the
4910 root logger at import time.
4912 Any log statement that the application executes before this import does
4913 not emit an LTTng event.
4915 IMPORTANT: The LTTng-UST Python agent must be
4916 <<installing-lttng,installed>>.
4918 . Use log statements and logging configuration as usual.
4919 Since the LTTng-UST Python agent adds a handler to the _root_
4920 logger, you can trace any log statement from any logger.
4922 .Use the LTTng-UST Python agent.
4933 logging.basicConfig()
4934 logger = logging.getLogger('my-logger')
4937 logger.debug('debug message')
4938 logger.info('info message')
4939 logger.warn('warn message')
4940 logger.error('error message')
4941 logger.critical('critical message')
4945 if __name__ == '__main__':
4949 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4950 logging handler which prints to the standard error stream, is not
4951 strictly required for LTTng-UST tracing to work, but in versions of
4952 Python preceding{nbsp}3.2, you could see a warning message which indicates
4953 that no handler exists for the logger `my-logger`.
4955 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4956 <<enabling-disabling-events,create an event rule>> matching the
4957 `my-logger` Python logger, and <<basic-tracing-session-control,start
4963 $ lttng enable-event --python my-logger
4967 Run the Python script:
4974 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4984 In the resulting trace, an <<event,event record>> generated by a Python
4985 application is named `lttng_python:event` and has the following fields:
4988 Logging time (string).
4991 Log record's message.
4997 Name of the function in which the log statement was executed.
5000 Line number at which the log statement was executed.
5003 Log level integer value.
5006 ID of the Python thread in which the log statement was executed.
5009 Name of the Python thread in which the log statement was executed.
5011 You can use the opt:lttng-enable-event(1):--loglevel or
5012 opt:lttng-enable-event(1):--loglevel-only option of the
5013 man:lttng-enable-event(1) command to target a range of Python log levels
5014 or a specific Python log level.
5016 When an application imports the LTTng-UST Python agent, the agent tries
5017 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5018 <<start-sessiond,start the session daemon>> _before_ you run the Python
5019 application. If a session daemon is found, the agent tries to register
5020 to it during five seconds, after which the application continues
5021 without LTTng tracing support. You can override this timeout value with
5022 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5025 If the session daemon stops while a Python application with an imported
5026 LTTng-UST Python agent runs, the agent retries to connect and to
5027 register to a session daemon every three seconds. You can override this
5028 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5033 [[proc-lttng-logger-abi]]
5036 The `lttng-tracer` Linux kernel module, part of
5037 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
5038 path:{/proc/lttng-logger} when it's loaded. Any application can write
5039 text data to this file to emit an LTTng event.
5042 .An application writes to the LTTng logger file to emit an LTTng event.
5043 image::lttng-logger.png[]
5045 The LTTng logger is the quickest method--not the most efficient,
5046 however--to add instrumentation to an application. It is designed
5047 mostly to instrument shell scripts:
5051 $ echo "Some message, some $variable" > /proc/lttng-logger
5054 Any event that the LTTng logger emits is named `lttng_logger` and
5055 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5056 other instrumentation points in the kernel tracing domain, **any Unix
5057 user** can <<enabling-disabling-events,create an event rule>> which
5058 matches its event name, not only the root user or users in the
5059 <<tracing-group,tracing group>>.
5061 To use the LTTng logger:
5063 * From any application, write text data to the path:{/proc/lttng-logger}
5066 The `msg` field of `lttng_logger` event records contains the
5069 NOTE: The maximum message length of an LTTng logger event is
5070 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5071 than one event to contain the remaining data.
5073 You should not use the LTTng logger to trace a user application which
5074 can be instrumented in a more efficient way, namely:
5076 * <<c-application,C and $$C++$$ applications>>.
5077 * <<java-application,Java applications>>.
5078 * <<python-application,Python applications>>.
5080 .Use the LTTng logger.
5085 echo 'Hello, World!' > /proc/lttng-logger
5087 df --human-readable --print-type / > /proc/lttng-logger
5090 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5091 <<enabling-disabling-events,create an event rule>> matching the
5092 `lttng_logger` Linux kernel tracepoint, and
5093 <<basic-tracing-session-control,start tracing>>:
5098 $ lttng enable-event --kernel lttng_logger
5102 Run the Bash script:
5109 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5120 [[instrumenting-linux-kernel]]
5121 === LTTng kernel tracepoints
5123 NOTE: This section shows how to _add_ instrumentation points to the
5124 Linux kernel. The kernel's subsystems are already thoroughly
5125 instrumented at strategic places for LTTng when you
5126 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5130 There are two methods to instrument the Linux kernel:
5132 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5133 tracepoint which uses the `TRACE_EVENT()` API.
5135 Choose this if you want to instrumentation a Linux kernel tree with an
5136 instrumentation point compatible with ftrace, perf, and SystemTap.
5138 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5139 instrument an out-of-tree kernel module.
5141 Choose this if you don't need ftrace, perf, or SystemTap support.
5145 [[linux-add-lttng-layer]]
5146 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5148 This section shows how to add an LTTng layer to existing ftrace
5149 instrumentation using the `TRACE_EVENT()` API.
5151 This section does not document the `TRACE_EVENT()` macro. You can
5152 read the following articles to learn more about this API:
5154 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part{nbsp}1)]
5155 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part{nbsp}2)]
5156 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part{nbsp}3)]
5158 The following procedure assumes that your ftrace tracepoints are
5159 correctly defined in their own header and that they are created in
5160 one source file using the `CREATE_TRACE_POINTS` definition.
5162 To add an LTTng layer over an existing ftrace tracepoint:
5164 . Make sure the following kernel configuration options are
5170 * `CONFIG_HIGH_RES_TIMERS`
5171 * `CONFIG_TRACEPOINTS`
5174 . Build the Linux source tree with your custom ftrace tracepoints.
5175 . Boot the resulting Linux image on your target system.
5177 Confirm that the tracepoints exist by looking for their names in the
5178 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5179 is your subsystem's name.
5181 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5186 $ cd $(mktemp -d) &&
5187 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
5188 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
5189 cd lttng-modules-2.11.*
5193 . In dir:{instrumentation/events/lttng-module}, relative to the root
5194 of the LTTng-modules source tree, create a header file named
5195 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5196 LTTng-modules tracepoint definitions using the LTTng-modules
5199 Start with this template:
5203 .path:{instrumentation/events/lttng-module/my_subsys.h}
5206 #define TRACE_SYSTEM my_subsys
5208 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5209 #define _LTTNG_MY_SUBSYS_H
5211 #include "../../../probes/lttng-tracepoint-event.h"
5212 #include <linux/tracepoint.h>
5214 LTTNG_TRACEPOINT_EVENT(
5216 * Format is identical to TRACE_EVENT()'s version for the three
5217 * following macro parameters:
5220 TP_PROTO(int my_int, const char *my_string),
5221 TP_ARGS(my_int, my_string),
5223 /* LTTng-modules specific macros */
5225 ctf_integer(int, my_int_field, my_int)
5226 ctf_string(my_bar_field, my_bar)
5230 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5232 #include "../../../probes/define_trace.h"
5236 The entries in the `TP_FIELDS()` section are the list of fields for the
5237 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5238 ftrace's `TRACE_EVENT()` macro.
5240 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5241 complete description of the available `ctf_*()` macros.
5243 . Create the LTTng-modules probe's kernel module C source file,
5244 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5249 .path:{probes/lttng-probe-my-subsys.c}
5251 #include <linux/module.h>
5252 #include "../lttng-tracer.h"
5255 * Build-time verification of mismatch between mainline
5256 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5257 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5259 #include <trace/events/my_subsys.h>
5261 /* Create LTTng tracepoint probes */
5262 #define LTTNG_PACKAGE_BUILD
5263 #define CREATE_TRACE_POINTS
5264 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5266 #include "../instrumentation/events/lttng-module/my_subsys.h"
5268 MODULE_LICENSE("GPL and additional rights");
5269 MODULE_AUTHOR("Your name <your-email>");
5270 MODULE_DESCRIPTION("LTTng my_subsys probes");
5271 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5272 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5273 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5274 LTTNG_MODULES_EXTRAVERSION);
5278 . Edit path:{probes/KBuild} and add your new kernel module object
5279 next to the existing ones:
5283 .path:{probes/KBuild}
5287 obj-m += lttng-probe-module.o
5288 obj-m += lttng-probe-power.o
5290 obj-m += lttng-probe-my-subsys.o
5296 . Build and install the LTTng kernel modules:
5301 $ make KERNELDIR=/path/to/linux
5302 # make modules_install && depmod -a
5306 Replace `/path/to/linux` with the path to the Linux source tree where
5307 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5309 Note that you can also use the
5310 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5311 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5312 C code that need to be executed before the event fields are recorded.
5314 The best way to learn how to use the previous LTTng-modules macros is to
5315 inspect the existing LTTng-modules tracepoint definitions in the
5316 dir:{instrumentation/events/lttng-module} header files. Compare them
5317 with the Linux kernel mainline versions in the
5318 dir:{include/trace/events} directory of the Linux source tree.
5322 [[lttng-tracepoint-event-code]]
5323 ===== Use custom C code to access the data for tracepoint fields
5325 Although we recommended to always use the
5326 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5327 the arguments and fields of an LTTng-modules tracepoint when possible,
5328 sometimes you need a more complex process to access the data that the
5329 tracer records as event record fields. In other words, you need local
5330 variables and multiple C{nbsp}statements instead of simple
5331 argument-based expressions that you pass to the
5332 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5334 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5335 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5336 a block of C{nbsp}code to be executed before LTTng records the fields.
5337 The structure of this macro is:
5340 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5342 LTTNG_TRACEPOINT_EVENT_CODE(
5344 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5345 * version for the following three macro parameters:
5348 TP_PROTO(int my_int, const char *my_string),
5349 TP_ARGS(my_int, my_string),
5351 /* Declarations of custom local variables */
5354 unsigned long b = 0;
5355 const char *name = "(undefined)";
5356 struct my_struct *my_struct;
5360 * Custom code which uses both tracepoint arguments
5361 * (in TP_ARGS()) and local variables (in TP_locvar()).
5363 * Local variables are actually members of a structure pointed
5364 * to by the special variable tp_locvar.
5368 tp_locvar->a = my_int + 17;
5369 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5370 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5371 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5372 put_my_struct(tp_locvar->my_struct);
5381 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5382 * version for this, except that tp_locvar members can be
5383 * used in the argument expression parameters of
5384 * the ctf_*() macros.
5387 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5388 ctf_integer(int, my_struct_a, tp_locvar->a)
5389 ctf_string(my_string_field, my_string)
5390 ctf_string(my_struct_name, tp_locvar->name)
5395 IMPORTANT: The C code defined in `TP_code()` must not have any side
5396 effects when executed. In particular, the code must not allocate
5397 memory or get resources without deallocating this memory or putting
5398 those resources afterwards.
5401 [[instrumenting-linux-kernel-tracing]]
5402 ==== Load and unload a custom probe kernel module
5404 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5405 kernel module>> in the kernel before it can emit LTTng events.
5407 To load the default probe kernel modules and a custom probe kernel
5410 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5411 probe modules to load when starting a root <<lttng-sessiond,session
5415 .Load the `my_subsys`, `usb`, and the default probe modules.
5419 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5424 You only need to pass the subsystem name, not the whole kernel module
5427 To load _only_ a given custom probe kernel module:
5429 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5430 modules to load when starting a root session daemon:
5433 .Load only the `my_subsys` and `usb` probe modules.
5437 # lttng-sessiond --kmod-probes=my_subsys,usb
5442 To confirm that a probe module is loaded:
5449 $ lsmod | grep lttng_probe_usb
5453 To unload the loaded probe modules:
5455 * Kill the session daemon with `SIGTERM`:
5460 # pkill lttng-sessiond
5464 You can also use man:modprobe(8)'s `--remove` option if the session
5465 daemon terminates abnormally.
5468 [[controlling-tracing]]
5471 Once an application or a Linux kernel is
5472 <<instrumenting,instrumented>> for LTTng tracing,
5475 This section is divided in topics on how to use the various
5476 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5477 command-line tool>>, to _control_ the LTTng daemons and tracers.
5479 NOTE: In the following subsections, we refer to an man:lttng(1) command
5480 using its man page name. For example, instead of _Run the `create`
5481 command to..._, we use _Run the man:lttng-create(1) command to..._.
5485 === Start a session daemon
5487 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5488 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5491 You will see the following error when you run a command while no session
5495 Error: No session daemon is available
5498 The only command that automatically runs a session daemon is
5499 man:lttng-create(1), which you use to
5500 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5501 this is most of the time the first operation that you do, sometimes it's
5502 not. Some examples are:
5504 * <<list-instrumentation-points,List the available instrumentation points>>.
5505 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5507 [[tracing-group]] Each Unix user must have its own running session
5508 daemon to trace user applications. The session daemon that the root user
5509 starts is the only one allowed to control the LTTng kernel tracer. Users
5510 that are part of the _tracing group_ can control the root session
5511 daemon. The default tracing group name is `tracing`; you can set it to
5512 something else with the opt:lttng-sessiond(8):--group option when you
5513 start the root session daemon.
5515 To start a user session daemon:
5517 * Run man:lttng-sessiond(8):
5522 $ lttng-sessiond --daemonize
5526 To start the root session daemon:
5528 * Run man:lttng-sessiond(8) as the root user:
5533 # lttng-sessiond --daemonize
5537 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5538 start the session daemon in foreground.
5540 To stop a session daemon, use man:kill(1) on its process ID (standard
5543 Note that some Linux distributions could manage the LTTng session daemon
5544 as a service. In this case, you should use the service manager to
5545 start, restart, and stop session daemons.
5548 [[creating-destroying-tracing-sessions]]
5549 === Create and destroy a tracing session
5551 Almost all the LTTng control operations happen in the scope of
5552 a <<tracing-session,tracing session>>, which is the dialogue between the
5553 <<lttng-sessiond,session daemon>> and you.
5555 To create a tracing session with a generated name:
5557 * Use the man:lttng-create(1) command:
5566 The created tracing session's name is `auto` followed by the
5569 To create a tracing session with a specific name:
5571 * Use the optional argument of the man:lttng-create(1) command:
5576 $ lttng create my-session
5580 Replace `my-session` with the specific tracing session name.
5582 LTTng appends the creation date to the created tracing session's name.
5584 LTTng writes the traces of a tracing session in
5585 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5586 name of the tracing session. Note that the env:LTTNG_HOME environment
5587 variable defaults to `$HOME` if not set.
5589 To output LTTng traces to a non-default location:
5591 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5596 $ lttng create my-session --output=/tmp/some-directory
5600 You may create as many tracing sessions as you wish.
5602 To list all the existing tracing sessions for your Unix user:
5604 * Use the man:lttng-list(1) command:
5613 When you create a tracing session, it is set as the _current tracing
5614 session_. The following man:lttng(1) commands operate on the current
5615 tracing session when you don't specify one:
5617 [role="list-3-cols"]
5618 * man:lttng-add-context(1)
5619 * man:lttng-destroy(1)
5620 * man:lttng-disable-channel(1)
5621 * man:lttng-disable-event(1)
5622 * man:lttng-disable-rotation(1)
5623 * man:lttng-enable-channel(1)
5624 * man:lttng-enable-event(1)
5625 * man:lttng-enable-rotation(1)
5627 * man:lttng-regenerate(1)
5628 * man:lttng-rotate(1)
5630 * man:lttng-snapshot(1)
5631 * man:lttng-start(1)
5632 * man:lttng-status(1)
5634 * man:lttng-track(1)
5635 * man:lttng-untrack(1)
5638 To change the current tracing session:
5640 * Use the man:lttng-set-session(1) command:
5645 $ lttng set-session new-session
5649 Replace `new-session` by the name of the new current tracing session.
5651 When you are done tracing in a given tracing session, you can destroy
5652 it. This operation frees the resources taken by the tracing session
5653 to destroy; it does not destroy the trace data that LTTng wrote for
5654 this tracing session.
5656 To destroy the current tracing session:
5658 * Use the man:lttng-destroy(1) command:
5667 The man:lttng-destroy(1) command also runs the man:lttng-stop(1)
5668 command implicitly (see <<basic-tracing-session-control,Start and stop a
5669 tracing session>>). You need to stop tracing to make LTTng flush the
5670 remaining trace data and make the trace readable.
5673 [[list-instrumentation-points]]
5674 === List the available instrumentation points
5676 The <<lttng-sessiond,session daemon>> can query the running instrumented
5677 user applications and the Linux kernel to get a list of available
5678 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5679 they are tracepoints and system calls. For the user space tracing
5680 domain, they are tracepoints. For the other tracing domains, they are
5683 To list the available instrumentation points:
5685 * Use the man:lttng-list(1) command with the requested tracing domain's
5689 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5690 must be a root user, or it must be a member of the
5691 <<tracing-group,tracing group>>).
5692 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5693 kernel system calls (your Unix user must be a root user, or it must be
5694 a member of the tracing group).
5695 * opt:lttng-list(1):--userspace: user space tracepoints.
5696 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5697 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5698 * opt:lttng-list(1):--python: Python loggers.
5701 .List the available user space tracepoints.
5705 $ lttng list --userspace
5709 .List the available Linux kernel system call tracepoints.
5713 $ lttng list --kernel --syscall
5718 [[enabling-disabling-events]]
5719 === Create and enable an event rule
5721 Once you <<creating-destroying-tracing-sessions,create a tracing
5722 session>>, you can create <<event,event rules>> with the
5723 man:lttng-enable-event(1) command.
5725 You specify each condition with a command-line option. The available
5726 condition arguments are shown in the following table.
5728 [role="growable",cols="asciidoc,asciidoc,default"]
5729 .Condition command-line arguments for the man:lttng-enable-event(1) command.
5731 |Argument |Description |Applicable tracing domains
5737 . +--probe=__ADDR__+
5738 . +--function=__ADDR__+
5739 . +--userspace-probe=__PATH__:__SYMBOL__+
5740 . +--userspace-probe=sdt:__PATH__:__PROVIDER__:__NAME__+
5743 Instead of using the default _tracepoint_ instrumentation type, use:
5745 . A Linux system call (entry and exit).
5746 . A Linux https://lwn.net/Articles/132196/[kprobe] (symbol or address).
5747 . The entry and return points of a Linux function (symbol or address).
5748 . The entry point of a user application or library function (path to
5749 application/library and symbol).
5750 . A https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
5751 Statically Defined Tracing] (USDT) probe (path to application/library,
5752 provider and probe names).
5756 |First positional argument.
5759 Tracepoint or system call name.
5761 With the opt:lttng-enable-event(1):--probe,
5762 opt:lttng-enable-event(1):--function, and
5763 opt:lttng-enable-event(1):--userspace-probe options, this is a custom
5764 name given to the event rule. With the JUL, log4j, and Python domains,
5765 this is a logger name.
5767 With a tracepoint, logger, or system call name, you can use the special
5768 `*` globbing character to match anything (for example, `sched_*`,
5776 . +--loglevel=__LEVEL__+
5777 . +--loglevel-only=__LEVEL__+
5780 . Match only tracepoints or log statements with a logging level at
5781 least as severe as +__LEVEL__+.
5782 . Match only tracepoints or log statements with a logging level
5783 equal to +__LEVEL__+.
5785 See man:lttng-enable-event(1) for the list of available logging level
5788 |User space, JUL, log4j, and Python.
5790 |+--exclude=__EXCLUSIONS__+
5793 When you use a `*` character at the end of the tracepoint or logger
5794 name (first positional argument), exclude the specific names in the
5795 comma-delimited list +__EXCLUSIONS__+.
5798 User space, JUL, log4j, and Python.
5800 |+--filter=__EXPR__+
5803 Match only events which satisfy the expression +__EXPR__+.
5805 See man:lttng-enable-event(1) to learn more about the syntax of a
5812 You attach an event rule to a <<channel,channel>> on creation. If you do
5813 not specify the channel with the opt:lttng-enable-event(1):--channel
5814 option, and if the event rule to create is the first in its
5815 <<domain,tracing domain>> for a given tracing session, then LTTng
5816 creates a _default channel_ for you. This default channel is reused in
5817 subsequent invocations of the man:lttng-enable-event(1) command for the
5818 same tracing domain.
5820 An event rule is always enabled at creation time.
5822 The following examples show how you can combine the previous
5823 command-line options to create simple to more complex event rules.
5825 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5829 $ lttng enable-event --kernel sched_switch
5833 .Create an event rule matching four Linux kernel system calls (default channel).
5837 $ lttng enable-event --kernel --syscall open,write,read,close
5841 .Create event rules matching tracepoints with filter expressions (default channel).
5845 $ lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5850 $ lttng enable-event --kernel --all \
5851 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5856 $ lttng enable-event --jul my_logger \
5857 --filter='$app.retriever:cur_msg_id > 3'
5860 IMPORTANT: Make sure to always quote the filter string when you
5861 use man:lttng(1) from a shell.
5864 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5868 $ lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5871 IMPORTANT: Make sure to always quote the wildcard character when you
5872 use man:lttng(1) from a shell.
5875 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5879 $ lttng enable-event --python my-app.'*' \
5880 --exclude='my-app.module,my-app.hello'
5884 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5888 $ lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5892 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5896 $ lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5900 .Create an event rule matching the `malloc` function entry in path:{/usr/lib/libc.so.6}:
5904 $ lttng enable-event --kernel --userspace-probe=/usr/lib/libc.so.6:malloc \
5909 .Create an event rule matching the `server`/`accept_request` https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[USDT probe] in path:{/usr/bin/serv}:
5913 $ lttng enable-event --kernel --userspace-probe=sdt:serv:server:accept_request \
5914 server_accept_request
5918 The event rules of a given channel form a whitelist: as soon as an
5919 emitted event passes one of them, LTTng can record the event. For
5920 example, an event named `my_app:my_tracepoint` emitted from a user space
5921 tracepoint with a `TRACE_ERROR` log level passes both of the following
5926 $ lttng enable-event --userspace my_app:my_tracepoint
5927 $ lttng enable-event --userspace my_app:my_tracepoint \
5928 --loglevel=TRACE_INFO
5931 The second event rule is redundant: the first one includes
5935 [[disable-event-rule]]
5936 === Disable an event rule
5938 To disable an event rule that you <<enabling-disabling-events,created>>
5939 previously, use the man:lttng-disable-event(1) command. This command
5940 disables _all_ the event rules (of a given tracing domain and channel)
5941 which match an instrumentation point. The other conditions are not
5942 supported as of LTTng{nbsp}{revision}.
5944 The LTTng tracer does not record an emitted event which passes
5945 a _disabled_ event rule.
5947 .Disable an event rule matching a Python logger (default channel).
5951 $ lttng disable-event --python my-logger
5955 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5959 $ lttng disable-event --jul '*'
5963 .Disable _all_ the event rules of the default channel.
5965 The opt:lttng-disable-event(1):--all-events option is not, like the
5966 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5967 equivalent of the event name `*` (wildcard): it disables _all_ the event
5968 rules of a given channel.
5972 $ lttng disable-event --jul --all-events
5976 NOTE: You cannot delete an event rule once you create it.
5980 === Get the status of a tracing session
5982 To get the status of the current tracing session, that is, its
5983 parameters, its channels, event rules, and their attributes:
5985 * Use the man:lttng-status(1) command:
5995 To get the status of any tracing session:
5997 * Use the man:lttng-list(1) command with the tracing session's name:
6002 $ lttng list my-session
6006 Replace `my-session` with the desired tracing session's name.
6009 [[basic-tracing-session-control]]
6010 === Start and stop a tracing session
6012 Once you <<creating-destroying-tracing-sessions,create a tracing
6014 <<enabling-disabling-events,create one or more event rules>>,
6015 you can start and stop the tracers for this tracing session.
6017 To start tracing in the current tracing session:
6019 * Use the man:lttng-start(1) command:
6028 LTTng is very flexible: you can launch user applications before
6029 or after the you start the tracers. The tracers only record the events
6030 if they pass enabled event rules and if they occur while the tracers are
6033 To stop tracing in the current tracing session:
6035 * Use the man:lttng-stop(1) command:
6044 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
6045 records>> or lost sub-buffers since the last time you ran
6046 man:lttng-start(1), warnings are printed when you run the
6047 man:lttng-stop(1) command.
6049 IMPORTANT: You need to stop tracing to make LTTng flush the remaining
6050 trace data and make the trace readable. Note that the
6051 man:lttng-destroy(1) command (see
6052 <<creating-destroying-tracing-sessions,Create and destroy a tracing
6053 session>>) also runs the man:lttng-stop(1) command implicitly.
6056 [[enabling-disabling-channels]]
6057 === Create a channel
6059 Once you create a tracing session, you can create a <<channel,channel>>
6060 with the man:lttng-enable-channel(1) command.
6062 Note that LTTng automatically creates a default channel when, for a
6063 given <<domain,tracing domain>>, no channels exist and you
6064 <<enabling-disabling-events,create>> the first event rule. This default
6065 channel is named `channel0` and its attributes are set to reasonable
6066 values. Therefore, you only need to create a channel when you need
6067 non-default attributes.
6069 You specify each non-default channel attribute with a command-line
6070 option when you use the man:lttng-enable-channel(1) command. The
6071 available command-line options are:
6073 [role="growable",cols="asciidoc,asciidoc"]
6074 .Command-line options for the man:lttng-enable-channel(1) command.
6076 |Option |Description
6082 <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> instead
6083 of the default _discard_ mode.
6085 |`--buffers-pid` (user space tracing domain only)
6088 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6089 instead of the default per-user buffering scheme.
6091 |+--subbuf-size=__SIZE__+
6094 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6095 either for each Unix user (default), or for each instrumented process.
6097 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6099 |+--num-subbuf=__COUNT__+
6102 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6103 for each Unix user (default), or for each instrumented process.
6105 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6107 |+--tracefile-size=__SIZE__+
6110 Set the maximum size of each trace file that this channel writes within
6111 a stream to +__SIZE__+ bytes instead of no maximum.
6113 See <<tracefile-rotation,Trace file count and size>>.
6115 |+--tracefile-count=__COUNT__+
6118 Limit the number of trace files that this channel creates to
6119 +__COUNT__+ channels instead of no limit.
6121 See <<tracefile-rotation,Trace file count and size>>.
6123 |+--switch-timer=__PERIODUS__+
6126 Set the <<channel-switch-timer,switch timer period>>
6127 to +__PERIODUS__+{nbsp}µs.
6129 |+--read-timer=__PERIODUS__+
6132 Set the <<channel-read-timer,read timer period>>
6133 to +__PERIODUS__+{nbsp}µs.
6135 |[[opt-blocking-timeout]]+--blocking-timeout=__TIMEOUTUS__+
6138 Set the timeout of user space applications which load LTTng-UST
6139 in blocking mode to +__TIMEOUTUS__+:
6142 Never block (non-blocking mode).
6145 Block forever until space is available in a sub-buffer to record
6148 __n__, a positive value::
6149 Wait for at most __n__ µs when trying to write into a sub-buffer.
6151 Note that, for this option to have any effect on an instrumented
6152 user space application, you need to run the application with a set
6153 env:LTTNG_UST_ALLOW_BLOCKING environment variable.
6155 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6158 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
6162 You can only create a channel in the Linux kernel and user space
6163 <<domain,tracing domains>>: other tracing domains have their own channel
6164 created on the fly when <<enabling-disabling-events,creating event
6169 Because of a current LTTng limitation, you must create all channels
6170 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6171 tracing session, that is, before the first time you run
6174 Since LTTng automatically creates a default channel when you use the
6175 man:lttng-enable-event(1) command with a specific tracing domain, you
6176 cannot, for example, create a Linux kernel event rule, start tracing,
6177 and then create a user space event rule, because no user space channel
6178 exists yet and it's too late to create one.
6180 For this reason, make sure to configure your channels properly
6181 before starting the tracers for the first time!
6184 The following examples show how you can combine the previous
6185 command-line options to create simple to more complex channels.
6187 .Create a Linux kernel channel with default attributes.
6191 $ lttng enable-channel --kernel my-channel
6195 .Create a user space channel with four sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6199 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6200 --buffers-pid my-channel
6204 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6206 <<creating-destroying-tracing-sessions,Create a tracing-session>>,
6207 create the channel, <<enabling-disabling-events,create an event rule>>,
6208 and <<basic-tracing-session-control,start tracing>>:
6213 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
6214 $ lttng enable-event --userspace --channel=blocking-channel --all
6218 Run an application instrumented with LTTng-UST and allow it to block:
6222 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6226 .Create a Linux kernel channel which rotates eight trace files of 4{nbsp}MiB each for each stream
6230 $ lttng enable-channel --kernel --tracefile-count=8 \
6231 --tracefile-size=4194304 my-channel
6235 .Create a user space channel in overwrite (or _flight recorder_) mode.
6239 $ lttng enable-channel --userspace --overwrite my-channel
6243 You can <<enabling-disabling-events,create>> the same event rule in
6244 two different channels:
6248 $ lttng enable-event --userspace --channel=my-channel app:tp
6249 $ lttng enable-event --userspace --channel=other-channel app:tp
6252 If both channels are enabled, when a tracepoint named `app:tp` is
6253 reached, LTTng records two events, one for each channel.
6257 === Disable a channel
6259 To disable a specific channel that you <<enabling-disabling-channels,created>>
6260 previously, use the man:lttng-disable-channel(1) command.
6262 .Disable a specific Linux kernel channel.
6266 $ lttng disable-channel --kernel my-channel
6270 The state of a channel precedes the individual states of event rules
6271 attached to it: event rules which belong to a disabled channel, even if
6272 they are enabled, are also considered disabled.
6276 === Add context fields to a channel
6278 Event record fields in trace files provide important information about
6279 events that occured previously, but sometimes some external context may
6280 help you solve a problem faster. Examples of context fields are:
6282 * The **process ID**, **thread ID**, **process name**, and
6283 **process priority** of the thread in which the event occurs.
6284 * The **hostname** of the system on which the event occurs.
6285 * The Linux kernel and user call stacks (since
6286 LTTng{nbsp}{revision}).
6287 * The current values of many possible **performance counters** using
6289 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6291 ** Branch instructions, misses, and loads.
6293 * Any context defined at the application level (supported for the
6294 JUL and log4j <<domain,tracing domains>>).
6296 To get the full list of available context fields, see
6297 `lttng add-context --list`. Some context fields are reserved for a
6298 specific <<domain,tracing domain>> (Linux kernel or user space).
6300 You add context fields to <<channel,channels>>. All the events
6301 that a channel with added context fields records contain those fields.
6303 To add context fields to one or all the channels of a given tracing
6306 * Use the man:lttng-add-context(1) command.
6308 .Add context fields to all the channels of the current tracing session.
6310 The following command line adds the virtual process identifier and
6311 the per-thread CPU cycles count fields to all the user space channels
6312 of the current tracing session.
6316 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6320 .Add performance counter context fields by raw ID
6322 See man:lttng-add-context(1) for the exact format of the context field
6323 type, which is partly compatible with the format used in
6328 $ lttng add-context --userspace --type=perf:thread:raw:r0110:test
6329 $ lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6333 .Add context fields to a specific channel.
6335 The following command line adds the thread identifier and user call
6336 stack context fields to the Linux kernel channel named `my-channel` in
6337 the current tracing session.
6341 $ lttng add-context --kernel --channel=my-channel \
6342 --type=tid --type=callstack-user
6346 .Add an application-specific context field to a specific channel.
6348 The following command line adds the `cur_msg_id` context field of the
6349 `retriever` context retriever for all the instrumented
6350 <<java-application,Java applications>> recording <<event,event records>>
6351 in the channel named `my-channel`:
6355 $ lttng add-context --kernel --channel=my-channel \
6356 --type='$app:retriever:cur_msg_id'
6359 IMPORTANT: Make sure to always quote the `$` character when you
6360 use man:lttng-add-context(1) from a shell.
6363 NOTE: You cannot remove context fields from a channel once you add it.
6368 === Track process IDs
6370 It's often useful to allow only specific process IDs (PIDs) to emit
6371 events. For example, you may wish to record all the system calls made by
6372 a given process (Ă la http://linux.die.net/man/1/strace[strace]).
6374 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6375 purpose. Both commands operate on a whitelist of process IDs. You _add_
6376 entries to this whitelist with the man:lttng-track(1) command and remove
6377 entries with the man:lttng-untrack(1) command. Any process which has one
6378 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6379 an enabled <<event,event rule>>.
6381 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6382 process with a given tracked ID exit and another process be given this
6383 ID, then the latter would also be allowed to emit events.
6385 .Track and untrack process IDs.
6387 For the sake of the following example, assume the target system has
6388 16{nbsp}possible PIDs.
6391 <<creating-destroying-tracing-sessions,create a tracing session>>,
6392 the whitelist contains all the possible PIDs:
6395 .All PIDs are tracked.
6396 image::track-all.png[]
6398 When the whitelist is full and you use the man:lttng-track(1) command to
6399 specify some PIDs to track, LTTng first clears the whitelist, then it
6400 tracks the specific PIDs. After:
6404 $ lttng track --pid=3,4,7,10,13
6410 .PIDs 3, 4, 7, 10, and 13 are tracked.
6411 image::track-3-4-7-10-13.png[]
6413 You can add more PIDs to the whitelist afterwards:
6417 $ lttng track --pid=1,15,16
6423 .PIDs 1, 15, and 16 are added to the whitelist.
6424 image::track-1-3-4-7-10-13-15-16.png[]
6426 The man:lttng-untrack(1) command removes entries from the PID tracker's
6427 whitelist. Given the previous example, the following command:
6431 $ lttng untrack --pid=3,7,10,13
6434 leads to this whitelist:
6437 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6438 image::track-1-4-15-16.png[]
6440 LTTng can track all possible PIDs again using the
6441 opt:lttng-track(1):--all option:
6445 $ lttng track --pid --all
6448 The result is, again:
6451 .All PIDs are tracked.
6452 image::track-all.png[]
6455 .Track only specific PIDs
6457 A very typical use case with PID tracking is to start with an empty
6458 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6459 then add PIDs manually while tracers are active. You can accomplish this
6460 by using the opt:lttng-untrack(1):--all option of the
6461 man:lttng-untrack(1) command to clear the whitelist after you
6462 <<creating-destroying-tracing-sessions,create a tracing session>>:
6466 $ lttng untrack --pid --all
6472 .No PIDs are tracked.
6473 image::untrack-all.png[]
6475 If you trace with this whitelist configuration, the tracer records no
6476 events for this <<domain,tracing domain>> because no processes are
6477 tracked. You can use the man:lttng-track(1) command as usual to track
6478 specific PIDs, for example:
6482 $ lttng track --pid=6,11
6488 .PIDs 6 and 11 are tracked.
6489 image::track-6-11.png[]
6494 [[saving-loading-tracing-session]]
6495 === Save and load tracing session configurations
6497 Configuring a <<tracing-session,tracing session>> can be long. Some of
6498 the tasks involved are:
6500 * <<enabling-disabling-channels,Create channels>> with
6501 specific attributes.
6502 * <<adding-context,Add context fields>> to specific channels.
6503 * <<enabling-disabling-events,Create event rules>> with specific log
6504 level and filter conditions.
6506 If you use LTTng to solve real world problems, chances are you have to
6507 record events using the same tracing session setup over and over,
6508 modifying a few variables each time in your instrumented program
6509 or environment. To avoid constant tracing session reconfiguration,
6510 the man:lttng(1) command-line tool can save and load tracing session
6511 configurations to/from XML files.
6513 To save a given tracing session configuration:
6515 * Use the man:lttng-save(1) command:
6520 $ lttng save my-session
6524 Replace `my-session` with the name of the tracing session to save.
6526 LTTng saves tracing session configurations to
6527 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6528 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6529 the opt:lttng-save(1):--output-path option to change this destination
6532 LTTng saves all configuration parameters, for example:
6534 * The tracing session name.
6535 * The trace data output path.
6536 * The channels with their state and all their attributes.
6537 * The context fields you added to channels.
6538 * The event rules with their state, log level and filter conditions.
6540 To load a tracing session:
6542 * Use the man:lttng-load(1) command:
6547 $ lttng load my-session
6551 Replace `my-session` with the name of the tracing session to load.
6553 When LTTng loads a configuration, it restores your saved tracing session
6554 as if you just configured it manually.
6556 See man:lttng-load(1) for the complete list of command-line options. You
6557 can also save and load many sessions at a time, and decide in which
6558 directory to output the XML files.
6561 [[sending-trace-data-over-the-network]]
6562 === Send trace data over the network
6564 LTTng can send the recorded trace data to a remote system over the
6565 network instead of writing it to the local file system.
6567 To send the trace data over the network:
6569 . On the _remote_ system (which can also be the target system),
6570 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6579 . On the _target_ system, create a tracing session configured to
6580 send trace data over the network:
6585 $ lttng create my-session --set-url=net://remote-system
6589 Replace `remote-system` by the host name or IP address of the
6590 remote system. See man:lttng-create(1) for the exact URL format.
6592 . On the target system, use the man:lttng(1) command-line tool as usual.
6593 When tracing is active, the target's consumer daemon sends sub-buffers
6594 to the relay daemon running on the remote system instead of flushing
6595 them to the local file system. The relay daemon writes the received
6596 packets to the local file system.
6598 The relay daemon writes trace files to
6599 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6600 +__hostname__+ is the host name of the target system and +__session__+
6601 is the tracing session name. Note that the env:LTTNG_HOME environment
6602 variable defaults to `$HOME` if not set. Use the
6603 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6604 trace files to another base directory.
6609 === View events as LTTng emits them (noch:{LTTng} live)
6611 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6612 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6613 display events as LTTng emits them on the target system while tracing is
6616 The relay daemon creates a _tee_: it forwards the trace data to both
6617 the local file system and to connected live viewers:
6620 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6625 . On the _target system_, create a <<tracing-session,tracing session>>
6631 $ lttng create my-session --live
6635 This spawns a local relay daemon.
6637 . Start the live viewer and configure it to connect to the relay
6638 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6643 $ babeltrace --input-format=lttng-live \
6644 net://localhost/host/hostname/my-session
6651 * `hostname` with the host name of the target system.
6652 * `my-session` with the name of the tracing session to view.
6655 . Configure the tracing session as usual with the man:lttng(1)
6656 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6658 You can list the available live tracing sessions with Babeltrace:
6662 $ babeltrace --input-format=lttng-live net://localhost
6665 You can start the relay daemon on another system. In this case, you need
6666 to specify the relay daemon's URL when you create the tracing session
6667 with the opt:lttng-create(1):--set-url option. You also need to replace
6668 `localhost` in the procedure above with the host name of the system on
6669 which the relay daemon is running.
6671 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6672 command-line options.
6676 [[taking-a-snapshot]]
6677 === Take a snapshot of the current sub-buffers of a tracing session
6679 The normal behavior of LTTng is to append full sub-buffers to growing
6680 trace data files. This is ideal to keep a full history of the events
6681 that occurred on the target system, but it can
6682 represent too much data in some situations. For example, you may wish
6683 to trace your application continuously until some critical situation
6684 happens, in which case you only need the latest few recorded
6685 events to perform the desired analysis, not multi-gigabyte trace files.
6687 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6688 current sub-buffers of a given <<tracing-session,tracing session>>.
6689 LTTng can write the snapshot to the local file system or send it over
6693 .A snapshot is a copy of the current sub-buffers, which are not cleared after the operation.
6694 image::snapshot.png[]
6696 If you wish to create unmanaged, self-contained, non-overlapping
6697 trace chunk archives instead of a simple copy of the current
6698 sub-buffers, see the <<session-rotation,tracing session rotation>>
6699 feature (available since LTTng{nbsp}2.11).
6703 . Create a tracing session in _snapshot mode_:
6708 $ lttng create my-session --snapshot
6712 The <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> of
6713 <<channel,channels>> created in this mode is automatically set to
6714 _overwrite_ (flight recorder mode).
6716 . Configure the tracing session as usual with the man:lttng(1)
6717 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6719 . **Optional**: When you need to take a snapshot,
6720 <<basic-tracing-session-control,stop tracing>>.
6722 You can take a snapshot when the tracers are active, but if you stop
6723 them first, you are sure that the data in the sub-buffers does not
6724 change before you actually take the snapshot.
6731 $ lttng snapshot record --name=my-first-snapshot
6735 LTTng writes the current sub-buffers of all the current tracing
6736 session's channels to trace files on the local file system. Those trace
6737 files have `my-first-snapshot` in their name.
6739 There is no difference between the format of a normal trace file and the
6740 format of a snapshot: viewers of LTTng traces also support LTTng
6743 By default, LTTng writes snapshot files to the path shown by
6744 `lttng snapshot list-output`. You can change this path or decide to send
6745 snapshots over the network using either:
6747 . An output path or URL that you specify when you
6748 <<creating-destroying-tracing-sessions,create the tracing session>>.
6749 . A snapshot output path or URL that you add using
6750 `lttng snapshot add-output`.
6751 . An output path or URL that you provide directly to the
6752 `lttng snapshot record` command.
6754 Method{nbsp}3 overrides method{nbsp}2, which overrides method 1. When
6755 you specify a URL, a relay daemon must listen on a remote system (see
6756 <<sending-trace-data-over-the-network,Send trace data over the
6761 [[session-rotation]]
6762 === Archive the current trace chunk (rotate a tracing session)
6764 The <<taking-a-snapshot,snapshot user guide>> shows how you can dump
6765 a tracing session's current sub-buffers to the file system or send them
6766 over the network. When you take a snapshot, LTTng does not clear the
6767 tracing session's ring buffers: if you take another snapshot immediately
6768 after, both snapshots could contain overlapping trace data.
6770 Inspired by https://en.wikipedia.org/wiki/Log_rotation[log rotation],
6771 _tracing session rotation_ is a feature which appends the content of the
6772 ring buffers to what's already on the file system or sent over the
6773 network since the tracing session's creation or since the last
6774 rotation, and then clears those ring buffers to avoid trace data
6777 What LTTng is about to write when performing a tracing session rotation
6778 is called the _current trace chunk_. When this current trace chunk is
6779 written to the file system or sent over the network, it becomes a _trace
6780 chunk archive_. Therefore, a tracing session rotation _archives_ the
6781 current trace chunk.
6784 .A tracing session rotation operation _archives_ the current trace chunk.
6785 image::rotation.png[]
6787 A trace chunk archive is a self-contained LTTng trace which LTTng
6788 doesn't manage anymore: you can read it, modify it, move it, or remove
6791 There are two methods to perform a tracing session rotation: immediately
6792 or with a rotation schedule.
6794 To perform an immediate tracing session rotation:
6796 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6797 in _normal mode_ or _network streaming mode_
6798 (only those two creation modes support tracing session rotation):
6803 $ lttng create my-session
6807 . <<enabling-disabling-events,Create one or more event rules>>
6808 and <<basic-tracing-session-control,start tracing>>:
6813 $ lttng enable-event --kernel sched_'*'
6818 . When needed, immediately rotate the current tracing session:
6827 The cmd:lttng-rotate command prints the path to the created trace
6828 chunk archive. See man:lttng-rotate(1) to learn about the format
6829 of trace chunk archive directory names.
6831 You can perform other immediate rotations while the tracing session is
6832 active. It is guaranteed that all the trace chunk archives do not
6833 contain overlapping trace data. You can also perform an immediate
6834 rotation once you have <<basic-tracing-session-control,stopped>> the
6837 . When you are done tracing,
6838 <<creating-destroying-tracing-sessions,destroy the current tracing
6848 The tracing session destruction operation creates one last trace
6849 chunk archive from the current trace chunk.
6851 A tracing session rotation schedule is a planned rotation which LTTng
6852 performs automatically based on one of the following conditions:
6854 * A timer with a configured period times out.
6856 * The total size of the flushed part of the current trace chunk
6857 becomes greater than or equal to a configured value.
6859 To schedule a tracing session rotation, set a _rotation schedule_:
6861 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6862 in _normal mode_ or _network streaming mode_
6863 (only those two creation modes support tracing session rotation):
6868 $ lttng create my-session
6872 . <<enabling-disabling-events,Create one or more event rules>>:
6877 $ lttng enable-event --kernel sched_'*'
6881 . Set a tracing session rotation schedule:
6886 $ lttng enable-rotation --timer=10s
6890 In this example, we set a rotation schedule so that LTTng performs a
6891 tracing session rotation every ten seconds.
6893 See man:lttng-enable-rotation(1) to learn more about other ways to set a
6896 . <<basic-tracing-session-control,Start tracing>>:
6905 LTTng performs tracing session rotations automatically while the tracing
6906 session is active thanks to the rotation schedule.
6908 . When you are done tracing,
6909 <<creating-destroying-tracing-sessions,destroy the current tracing
6919 The tracing session destruction operation creates one last trace chunk
6920 archive from the current trace chunk.
6922 You can use man:lttng-disable-rotation(1) to unset a tracing session
6925 NOTE: man:lttng-rotate(1) and man:lttng-enable-rotation(1) list
6926 limitations regarding those two commands.
6931 === Use the machine interface
6933 With any command of the man:lttng(1) command-line tool, you can set the
6934 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6935 XML machine interface output, for example:
6939 $ lttng --mi=xml enable-event --kernel --syscall open
6942 A schema definition (XSD) is
6943 https://github.com/lttng/lttng-tools/blob/stable-2.11/src/common/mi-lttng-3.0.xsd[available]
6944 to ease the integration with external tools as much as possible.
6948 [[metadata-regenerate]]
6949 === Regenerate the metadata of an LTTng trace
6951 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6952 data stream files and a metadata file. This metadata file contains,
6953 amongst other things, information about the offset of the clock sources
6954 used to timestamp <<event,event records>> when tracing.
6956 If, once a <<tracing-session,tracing session>> is
6957 <<basic-tracing-session-control,started>>, a major
6958 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6959 happens, the trace's clock offset also needs to be updated. You
6960 can use the `metadata` item of the man:lttng-regenerate(1) command
6963 The main use case of this command is to allow a system to boot with
6964 an incorrect wall time and trace it with LTTng before its wall time
6965 is corrected. Once the system is known to be in a state where its
6966 wall time is correct, it can run `lttng regenerate metadata`.
6968 To regenerate the metadata of an LTTng trace:
6970 * Use the `metadata` item of the man:lttng-regenerate(1) command:
6975 $ lttng regenerate metadata
6981 `lttng regenerate metadata` has the following limitations:
6983 * Tracing session <<creating-destroying-tracing-sessions,created>>
6985 * User space <<channel,channels>>, if any, are using
6986 <<channel-buffering-schemes,per-user buffering>>.
6991 [[regenerate-statedump]]
6992 === Regenerate the state dump of a tracing session
6994 The LTTng kernel and user space tracers generate state dump
6995 <<event,event records>> when the application starts or when you
6996 <<basic-tracing-session-control,start a tracing session>>. An analysis
6997 can use the state dump event records to set an initial state before it
6998 builds the rest of the state from the following event records.
6999 http://tracecompass.org/[Trace Compass] is a notable example of an
7000 application which uses the state dump of an LTTng trace.
7002 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
7003 state dump event records are not included in the snapshot because they
7004 were recorded to a sub-buffer that has been consumed or overwritten
7007 You can use the `lttng regenerate statedump` command to emit the state
7008 dump event records again.
7010 To regenerate the state dump of the current tracing session, provided
7011 create it in snapshot mode, before you take a snapshot:
7013 . Use the `statedump` item of the man:lttng-regenerate(1) command:
7018 $ lttng regenerate statedump
7022 . <<basic-tracing-session-control,Stop the tracing session>>:
7031 . <<taking-a-snapshot,Take a snapshot>>:
7036 $ lttng snapshot record --name=my-snapshot
7040 Depending on the event throughput, you should run steps 1 and 2
7041 as closely as possible.
7043 NOTE: To record the state dump events, you need to
7044 <<enabling-disabling-events,create event rules>> which enable them.
7045 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
7046 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
7050 [[persistent-memory-file-systems]]
7051 === Record trace data on persistent memory file systems
7053 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
7054 (NVRAM) is random-access memory that retains its information when power
7055 is turned off (non-volatile). Systems with such memory can store data
7056 structures in RAM and retrieve them after a reboot, without flushing
7057 to typical _storage_.
7059 Linux supports NVRAM file systems thanks to either
7060 http://pramfs.sourceforge.net/[PRAMFS] or
7061 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
7062 (requires Linux{nbsp}4.1+).
7064 This section does not describe how to operate such file systems;
7065 we assume that you have a working persistent memory file system.
7067 When you create a <<tracing-session,tracing session>>, you can specify
7068 the path of the shared memory holding the sub-buffers. If you specify a
7069 location on an NVRAM file system, then you can retrieve the latest
7070 recorded trace data when the system reboots after a crash.
7072 To record trace data on a persistent memory file system and retrieve the
7073 trace data after a system crash:
7075 . Create a tracing session with a sub-buffer shared memory path located
7076 on an NVRAM file system:
7081 $ lttng create my-session --shm-path=/path/to/shm
7085 . Configure the tracing session as usual with the man:lttng(1)
7086 command-line tool, and <<basic-tracing-session-control,start tracing>>.
7088 . After a system crash, use the man:lttng-crash(1) command-line tool to
7089 view the trace data recorded on the NVRAM file system:
7094 $ lttng-crash /path/to/shm
7098 The binary layout of the ring buffer files is not exactly the same as
7099 the trace files layout. This is why you need to use man:lttng-crash(1)
7100 instead of your preferred trace viewer directly.
7102 To convert the ring buffer files to LTTng trace files:
7104 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
7109 $ lttng-crash --extract=/path/to/trace /path/to/shm
7115 [[notif-trigger-api]]
7116 === Get notified when a channel's buffer usage is too high or too low
7118 With LTTng's $$C/C++$$ notification and trigger API, your user
7119 application can get notified when the buffer usage of one or more
7120 <<channel,channels>> becomes too low or too high. You can use this API
7121 and enable or disable <<event,event rules>> during tracing to avoid
7122 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>.
7124 .Have a user application get notified when an LTTng channel's buffer usage is too high.
7126 In this example, we create and build an application which gets notified
7127 when the buffer usage of a specific LTTng channel is higher than
7128 75{nbsp}%. We only print that it is the case in the example, but we
7129 could as well use the API of <<liblttng-ctl-lttng,`liblttng-ctl`>> to
7130 disable event rules when this happens.
7132 . Create the application's C source file:
7140 #include <lttng/domain.h>
7141 #include <lttng/action/action.h>
7142 #include <lttng/action/notify.h>
7143 #include <lttng/condition/condition.h>
7144 #include <lttng/condition/buffer-usage.h>
7145 #include <lttng/condition/evaluation.h>
7146 #include <lttng/notification/channel.h>
7147 #include <lttng/notification/notification.h>
7148 #include <lttng/trigger/trigger.h>
7149 #include <lttng/endpoint.h>
7151 int main(int argc, char *argv[])
7153 int exit_status = 0;
7154 struct lttng_notification_channel *notification_channel;
7155 struct lttng_condition *condition;
7156 struct lttng_action *action;
7157 struct lttng_trigger *trigger;
7158 const char *tracing_session_name;
7159 const char *channel_name;
7162 tracing_session_name = argv[1];
7163 channel_name = argv[2];
7166 * Create a notification channel. A notification channel
7167 * connects the user application to the LTTng session daemon.
7168 * This notification channel can be used to listen to various
7169 * types of notifications.
7171 notification_channel = lttng_notification_channel_create(
7172 lttng_session_daemon_notification_endpoint);
7175 * Create a "high buffer usage" condition. In this case, the
7176 * condition is reached when the buffer usage is greater than or
7177 * equal to 75 %. We create the condition for a specific tracing
7178 * session name, channel name, and for the user space tracing
7181 * The "low buffer usage" condition type also exists.
7183 condition = lttng_condition_buffer_usage_high_create();
7184 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
7185 lttng_condition_buffer_usage_set_session_name(
7186 condition, tracing_session_name);
7187 lttng_condition_buffer_usage_set_channel_name(condition,
7189 lttng_condition_buffer_usage_set_domain_type(condition,
7193 * Create an action (get a notification) to take when the
7194 * condition created above is reached.
7196 action = lttng_action_notify_create();
7199 * Create a trigger. A trigger associates a condition to an
7200 * action: the action is executed when the condition is reached.
7202 trigger = lttng_trigger_create(condition, action);
7204 /* Register the trigger to LTTng. */
7205 lttng_register_trigger(trigger);
7208 * Now that we have registered a trigger, a notification will be
7209 * emitted everytime its condition is met. To receive this
7210 * notification, we must subscribe to notifications that match
7211 * the same condition.
7213 lttng_notification_channel_subscribe(notification_channel,
7217 * Notification loop. You can put this in a dedicated thread to
7218 * avoid blocking the main thread.
7221 struct lttng_notification *notification;
7222 enum lttng_notification_channel_status status;
7223 const struct lttng_evaluation *notification_evaluation;
7224 const struct lttng_condition *notification_condition;
7225 double buffer_usage;
7227 /* Receive the next notification. */
7228 status = lttng_notification_channel_get_next_notification(
7229 notification_channel, ¬ification);
7232 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
7234 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
7236 * The session daemon can drop notifications if
7237 * a monitoring application is not consuming the
7238 * notifications fast enough.
7241 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
7243 * The notification channel has been closed by the
7244 * session daemon. This is typically caused by a session
7245 * daemon shutting down.
7249 /* Unhandled conditions or errors. */
7255 * A notification provides, amongst other things:
7257 * * The condition that caused this notification to be
7259 * * The condition evaluation, which provides more
7260 * specific information on the evaluation of the
7263 * The condition evaluation provides the buffer usage
7264 * value at the moment the condition was reached.
7266 notification_condition = lttng_notification_get_condition(
7268 notification_evaluation = lttng_notification_get_evaluation(
7271 /* We're subscribed to only one condition. */
7272 assert(lttng_condition_get_type(notification_condition) ==
7273 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
7276 * Get the exact sampled buffer usage from the
7277 * condition evaluation.
7279 lttng_evaluation_buffer_usage_get_usage_ratio(
7280 notification_evaluation, &buffer_usage);
7283 * At this point, instead of printing a message, we
7284 * could do something to reduce the channel's buffer
7285 * usage, like disable specific events.
7287 printf("Buffer usage is %f %% in tracing session \"%s\", "
7288 "user space channel \"%s\".\n", buffer_usage * 100,
7289 tracing_session_name, channel_name);
7290 lttng_notification_destroy(notification);
7294 lttng_action_destroy(action);
7295 lttng_condition_destroy(condition);
7296 lttng_trigger_destroy(trigger);
7297 lttng_notification_channel_destroy(notification_channel);
7303 . Build the `notif-app` application, linking it to `liblttng-ctl`:
7308 $ gcc -o notif-app notif-app.c -llttng-ctl
7312 . <<creating-destroying-tracing-sessions,Create a tracing session>>,
7313 <<enabling-disabling-events,create an event rule>> matching all the
7314 user space tracepoints, and
7315 <<basic-tracing-session-control,start tracing>>:
7320 $ lttng create my-session
7321 $ lttng enable-event --userspace --all
7326 If you create the channel manually with the man:lttng-enable-channel(1)
7327 command, you can control how frequently are the current values of the
7328 channel's properties sampled to evaluate user conditions with the
7329 opt:lttng-enable-channel(1):--monitor-timer option.
7331 . Run the `notif-app` application. This program accepts the
7332 <<tracing-session,tracing session>> name and the user space channel
7333 name as its two first arguments. The channel which LTTng automatically
7334 creates with the man:lttng-enable-event(1) command above is named
7340 $ ./notif-app my-session channel0
7344 . In another terminal, run an application with a very high event
7345 throughput so that the 75{nbsp}% buffer usage condition is reached.
7347 In the first terminal, the application should print lines like this:
7350 Buffer usage is 81.45197 % in tracing session "my-session", user space
7354 If you don't see anything, try modifying the condition in
7355 path:{notif-app.c} to a lower value (0.1, for example), rebuilding it
7356 (step{nbsp}2) and running it again (step{nbsp}4).
7363 [[lttng-modules-ref]]
7364 === noch:{LTTng-modules}
7368 [[lttng-tracepoint-enum]]
7369 ==== `LTTNG_TRACEPOINT_ENUM()` usage
7371 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
7375 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
7380 * `name` with the name of the enumeration (C identifier, unique
7381 amongst all the defined enumerations).
7382 * `entries` with a list of enumeration entries.
7384 The available enumeration entry macros are:
7386 +ctf_enum_value(__name__, __value__)+::
7387 Entry named +__name__+ mapped to the integral value +__value__+.
7389 +ctf_enum_range(__name__, __begin__, __end__)+::
7390 Entry named +__name__+ mapped to the range of integral values between
7391 +__begin__+ (included) and +__end__+ (included).
7393 +ctf_enum_auto(__name__)+::
7394 Entry named +__name__+ mapped to the integral value following the
7395 last mapping's value.
7397 The last value of a `ctf_enum_value()` entry is its +__value__+
7400 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
7402 If `ctf_enum_auto()` is the first entry in the list, its integral
7405 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
7406 to use a defined enumeration as a tracepoint field.
7408 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
7412 LTTNG_TRACEPOINT_ENUM(
7415 ctf_enum_auto("AUTO: EXPECT 0")
7416 ctf_enum_value("VALUE: 23", 23)
7417 ctf_enum_value("VALUE: 27", 27)
7418 ctf_enum_auto("AUTO: EXPECT 28")
7419 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
7420 ctf_enum_auto("AUTO: EXPECT 304")
7428 [[lttng-modules-tp-fields]]
7429 ==== Tracepoint fields macros (for `TP_FIELDS()`)
7431 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
7432 tracepoint fields, which must be listed within `TP_FIELDS()` in
7433 `LTTNG_TRACEPOINT_EVENT()`, are:
7435 [role="func-desc growable",cols="asciidoc,asciidoc"]
7436 .Available macros to define LTTng-modules tracepoint fields
7438 |Macro |Description and parameters
7441 +ctf_integer(__t__, __n__, __e__)+
7443 +ctf_integer_nowrite(__t__, __n__, __e__)+
7445 +ctf_user_integer(__t__, __n__, __e__)+
7447 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
7449 Standard integer, displayed in base{nbsp}10.
7452 Integer C type (`int`, `long`, `size_t`, ...).
7458 Argument expression.
7461 +ctf_integer_hex(__t__, __n__, __e__)+
7463 +ctf_user_integer_hex(__t__, __n__, __e__)+
7465 Standard integer, displayed in base{nbsp}16.
7474 Argument expression.
7476 |+ctf_integer_oct(__t__, __n__, __e__)+
7478 Standard integer, displayed in base{nbsp}8.
7487 Argument expression.
7490 +ctf_integer_network(__t__, __n__, __e__)+
7492 +ctf_user_integer_network(__t__, __n__, __e__)+
7494 Integer in network byte order (big-endian), displayed in base{nbsp}10.
7503 Argument expression.
7506 +ctf_integer_network_hex(__t__, __n__, __e__)+
7508 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
7510 Integer in network byte order, displayed in base{nbsp}16.
7519 Argument expression.
7522 +ctf_enum(__N__, __t__, __n__, __e__)+
7524 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
7526 +ctf_user_enum(__N__, __t__, __n__, __e__)+
7528 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
7533 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
7536 Integer C type (`int`, `long`, `size_t`, ...).
7542 Argument expression.
7545 +ctf_string(__n__, __e__)+
7547 +ctf_string_nowrite(__n__, __e__)+
7549 +ctf_user_string(__n__, __e__)+
7551 +ctf_user_string_nowrite(__n__, __e__)+
7553 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7559 Argument expression.
7562 +ctf_array(__t__, __n__, __e__, __s__)+
7564 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7566 +ctf_user_array(__t__, __n__, __e__, __s__)+
7568 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7570 Statically-sized array of integers.
7573 Array element C type.
7579 Argument expression.
7585 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7587 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7589 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7591 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7593 Statically-sized array of bits.
7595 The type of +__e__+ must be an integer type. +__s__+ is the number
7596 of elements of such type in +__e__+, not the number of bits.
7599 Array element C type.
7605 Argument expression.
7611 +ctf_array_text(__t__, __n__, __e__, __s__)+
7613 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7615 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7617 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7619 Statically-sized array, printed as text.
7621 The string does not need to be null-terminated.
7624 Array element C type (always `char`).
7630 Argument expression.
7636 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7638 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7640 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7642 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7644 Dynamically-sized array of integers.
7646 The type of +__E__+ must be unsigned.
7649 Array element C type.
7655 Argument expression.
7658 Length expression C type.
7664 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7666 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7668 Dynamically-sized array of integers, displayed in base{nbsp}16.
7670 The type of +__E__+ must be unsigned.
7673 Array element C type.
7679 Argument expression.
7682 Length expression C type.
7687 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7689 Dynamically-sized array of integers in network byte order (big-endian),
7690 displayed in base{nbsp}10.
7692 The type of +__E__+ must be unsigned.
7695 Array element C type.
7701 Argument expression.
7704 Length expression C type.
7710 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7712 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7714 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7716 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7718 Dynamically-sized array of bits.
7720 The type of +__e__+ must be an integer type. +__s__+ is the number
7721 of elements of such type in +__e__+, not the number of bits.
7723 The type of +__E__+ must be unsigned.
7726 Array element C type.
7732 Argument expression.
7735 Length expression C type.
7741 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7743 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7745 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7747 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7749 Dynamically-sized array, displayed as text.
7751 The string does not need to be null-terminated.
7753 The type of +__E__+ must be unsigned.
7755 The behaviour is undefined if +__e__+ is `NULL`.
7758 Sequence element C type (always `char`).
7764 Argument expression.
7767 Length expression C type.
7773 Use the `_user` versions when the argument expression, `e`, is
7774 a user space address. In the cases of `ctf_user_integer*()` and
7775 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7778 The `_nowrite` versions omit themselves from the session trace, but are
7779 otherwise identical. This means the `_nowrite` fields won't be written
7780 in the recorded trace. Their primary purpose is to make some
7781 of the event context available to the
7782 <<enabling-disabling-events,event filters>> without having to
7783 commit the data to sub-buffers.
7789 Terms related to LTTng and to tracing in general:
7792 The http://diamon.org/babeltrace[Babeltrace] project, which includes:
7794 * The cmd:babeltrace (Babeltrace{nbsp}1) or cmd:babeltrace2
7795 (Babeltrace{nbsp}2) command.
7796 * Libraries with a C{nbsp}API.
7797 * Python{nbsp}3 bindings.
7798 * Plugins (Babeltrace{nbsp}2).
7800 [[def-buffering-scheme]]<<channel-buffering-schemes,buffering scheme>>::
7801 A layout of <<def-sub-buffer,sub-buffers>> applied to a given channel.
7803 [[def-channel]]<<channel,channel>>::
7804 An entity which is responsible for a set of
7805 <<def-ring-buffer,ring buffers>>.
7807 <<def-event-rule,Event rules>> are always attached to a specific
7811 A source of time for a <<def-tracer,tracer>>.
7813 [[def-consumer-daemon]]<<lttng-consumerd,consumer daemon>>::
7814 A process which is responsible for consuming the full
7815 <<def-sub-buffer,sub-buffers>> and write them to a file system or
7816 send them over the network.
7818 [[def-current-trace-chunk]]current trace chunk::
7819 A <<def-trace-chunk,trace chunk>> which includes the current content
7820 of all the <<def-tracing-session-rotation,tracing session>>'s
7821 <<def-sub-buffer,sub-buffers>> and the stream files produced since the
7822 latest event amongst:
7824 * The creation of the <<def-tracing-session,tracing session>>.
7825 * The last tracing session rotation, if any.
7827 <<channel-overwrite-mode-vs-discard-mode,discard mode>>::
7828 The <<def-event-record-loss-mode,event record loss mode>> in which
7829 the <<def-tracer,tracer>> _discards_ new event records when there's no
7830 <<def-sub-buffer,sub-buffer>> space left to store them.
7832 [[def-event]]event::
7833 The consequence of the execution of an
7834 <<def-instrumentation-point,instrumentation point>>, like a
7835 <<def-tracepoint,tracepoint>> that you manually place in some source
7836 code, or a Linux kernel kprobe.
7838 An event is said to _occur_ at a specific time. <<def-lttng,LTTng>> can
7839 take various actions upon the occurrence of an event, like record the
7840 event's payload to a <<def-sub-buffer,sub-buffer>>.
7842 [[def-event-name]]event name::
7843 The name of an <<def-event,event>>, which is also the name of the
7844 <<def-event-record,event record>>.
7846 This is also called the _instrumentation point name_.
7848 [[def-event-record]]event record::
7849 A record, in a <<def-trace,trace>>, of the payload of an
7850 <<def-event,event>> which occured.
7852 [[def-event-record-loss-mode]]<<channel-overwrite-mode-vs-discard-mode,event record loss mode>>::
7853 The mechanism by which event records of a given
7854 <<def-channel,channel>> are lost (not recorded) when there is no
7855 <<def-sub-buffer,sub-buffer>> space left to store them.
7857 [[def-event-rule]]<<event,event rule>>::
7858 Set of conditions which must be satisfied for one or more occuring
7859 <<def-event,events>> to be recorded.
7861 `java.util.logging`::
7863 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7865 <<instrumenting,instrumentation>>::
7866 The use of <<def-lttng,LTTng>> probes to make a piece of software
7869 [[def-instrumentation-point]]instrumentation point::
7870 A point in the execution path of a piece of software that, when
7871 reached by this execution, can emit an <<def-event,event>>.
7873 instrumentation point name::
7874 See _<<def-event-name,event name>>_.
7877 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7878 developed by the Apache Software Foundation.
7881 Level of severity of a log statement or user space
7882 <<def-instrumentation-point,instrumentation point>>.
7884 [[def-lttng]]LTTng::
7885 The _Linux Trace Toolkit: next generation_ project.
7887 <<lttng-cli,cmd:lttng>>::
7888 A command-line tool provided by the <<def-lttng-tools,LTTng-tools>>
7889 project which you can use to send and receive control messages to and
7890 from a <<def-session-daemon,session daemon>>.
7893 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7894 which is a set of analyzing programs that you can use to obtain a
7895 higher level view of an <<def-lttng,LTTng>> <<def-trace,trace>>.
7897 cmd:lttng-consumerd::
7898 The name of the <<def-consumer-daemon,consumer daemon>> program.
7901 A utility provided by the <<def-lttng-tools,LTTng-tools>> project
7902 which can convert <<def-ring-buffer,ring buffer>> files (usually
7903 <<persistent-memory-file-systems,saved on a persistent memory file
7904 system>>) to <<def-trace,trace>> files.
7906 See man:lttng-crash(1).
7908 LTTng Documentation::
7911 <<lttng-live,LTTng live>>::
7912 A communication protocol between the <<lttng-relayd,relay daemon>> and
7913 live viewers which makes it possible to see <<def-event-record,event
7914 records>> "live", as they are received by the
7915 <<def-relay-daemon,relay daemon>>.
7917 <<lttng-modules,LTTng-modules>>::
7918 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7919 which contains the Linux kernel modules to make the Linux kernel
7920 <<def-instrumentation-point,instrumentation points>> available for
7921 <<def-lttng,LTTng>> tracing.
7924 The name of the <<def-relay-daemon,relay daemon>> program.
7926 cmd:lttng-sessiond::
7927 The name of the <<def-session-daemon,session daemon>> program.
7929 [[def-lttng-tools]]LTTng-tools::
7930 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7931 contains the various programs and libraries used to
7932 <<controlling-tracing,control tracing>>.
7934 [[def-lttng-ust]]<<lttng-ust,LTTng-UST>>::
7935 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7936 contains libraries to instrument
7937 <<def-user-application,user applications>>.
7939 <<lttng-ust-agents,LTTng-UST Java agent>>::
7940 A Java package provided by the <<def-lttng-ust,LTTng-UST>> project to
7941 allow the LTTng instrumentation of `java.util.logging` and Apache
7942 log4j{nbsp}1.2 logging statements.
7944 <<lttng-ust-agents,LTTng-UST Python agent>>::
7945 A Python package provided by the <<def-lttng-ust,LTTng-UST>> project
7946 to allow the <<def-lttng,LTTng>> instrumentation of Python logging
7949 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7950 The <<def-event-record-loss-mode,event record loss mode>> in which new
7951 <<def-event-record,event records>> _overwrite_ older event records
7952 when there's no <<def-sub-buffer,sub-buffer>> space left to store
7955 <<channel-buffering-schemes,per-process buffering>>::
7956 A <<def-buffering-scheme,buffering scheme>> in which each instrumented
7957 process has its own <<def-sub-buffer,sub-buffers>> for a given user
7958 space <<def-channel,channel>>.
7960 <<channel-buffering-schemes,per-user buffering>>::
7961 A <<def-buffering-scheme,buffering scheme>> in which all the processes
7962 of a Unix user share the same <<def-sub-buffer,sub-buffers>> for a
7963 given user space <<def-channel,channel>>.
7965 [[def-relay-daemon]]<<lttng-relayd,relay daemon>>::
7966 A process which is responsible for receiving the <<def-trace,trace>>
7967 data which a distant <<def-consumer-daemon,consumer daemon>> sends.
7969 [[def-ring-buffer]]ring buffer::
7970 A set of <<def-sub-buffer,sub-buffers>>.
7973 See _<<def-tracing-session-rotation,tracing session rotation>>_.
7975 [[def-session-daemon]]<<lttng-sessiond,session daemon>>::
7976 A process which receives control commands from you and orchestrates
7977 the <<def-tracer,tracers>> and various <<def-lttng,LTTng>> daemons.
7979 <<taking-a-snapshot,snapshot>>::
7980 A copy of the current data of all the <<def-sub-buffer,sub-buffers>>
7981 of a given <<def-tracing-session,tracing session>>, saved as
7982 <<def-trace,trace>> files.
7984 [[def-sub-buffer]]sub-buffer::
7985 One part of an <<def-lttng,LTTng>> <<def-ring-buffer,ring buffer>>
7986 which contains <<def-event-record,event records>>.
7989 The time information attached to an <<def-event,event>> when it is
7992 [[def-trace]]trace (_noun_)::
7995 * One http://diamon.org/ctf/[CTF] metadata stream file.
7996 * One or more CTF data stream files which are the concatenations of one
7997 or more flushed <<def-sub-buffer,sub-buffers>>.
7999 [[def-trace-verb]]trace (_verb_)::
8000 The action of recording the <<def-event,events>> emitted by an
8001 application or by a system, or to initiate such recording by
8002 controlling a <<def-tracer,tracer>>.
8004 [[def-trace-chunk]]trace chunk::
8005 A self-contained <<def-trace,trace>> which is part of a
8006 <<def-tracing-session,tracing session>>. Each
8007 <<def-tracing-session-rotation, tracing session rotation>> produces a
8008 <<def-trace-chunk-archive,trace chunk archive>>.
8010 [[def-trace-chunk-archive]]trace chunk archive::
8011 The result of a <<def-tracing-session-rotation, tracing session rotation>>.
8013 <<def-lttng,LTTng>> does not manage any trace chunk archive, even if its
8014 containing <<def-tracing-session,tracing session>> is still active: you
8015 are free to read it, modify it, move it, or remove it.
8018 The http://tracecompass.org[Trace Compass] project and application.
8020 [[def-tracepoint]]tracepoint::
8021 An instrumentation point using the tracepoint mechanism of the Linux
8022 kernel or of <<def-lttng-ust,LTTng-UST>>.
8024 tracepoint definition::
8025 The definition of a single <<def-tracepoint,tracepoint>>.
8028 The name of a <<def-tracepoint,tracepoint>>.
8030 [[def-tracepoint-provider]]tracepoint provider::
8031 A set of functions providing <<def-tracepoint,tracepoints>> to an
8032 instrumented <<def-user-application,user application>>.
8034 Not to be confused with a <<def-tracepoint-provider-package,tracepoint
8035 provider package>>: many tracepoint providers can exist within a
8036 tracepoint provider package.
8038 [[def-tracepoint-provider-package]]tracepoint provider package::
8039 One or more <<def-tracepoint-provider,tracepoint providers>> compiled
8040 as an https://en.wikipedia.org/wiki/Object_file[object file] or as a
8041 link:https://en.wikipedia.org/wiki/Library_(computing)#Shared_libraries[shared
8044 [[def-tracer]]tracer::
8045 A software which records emitted <<def-event,events>>.
8047 <<domain,tracing domain>>::
8048 A namespace for <<def-event,event>> sources.
8050 <<tracing-group,tracing group>>::
8051 The Unix group in which a Unix user can be to be allowed to
8052 <<def-trace-verb,trace>> the Linux kernel.
8054 [[def-tracing-session]]<<tracing-session,tracing session>>::
8055 A stateful dialogue between you and a <<lttng-sessiond,session daemon>>.
8057 [[def-tracing-session-rotation]]<<session-rotation,tracing session rotation>>::
8058 The action of archiving the
8059 <<def-current-trace-chunk,current trace chunk>> of a
8060 <<def-tracing-session,tracing session>>.
8062 [[def-user-application]]user application::
8063 An application running in user space, as opposed to a Linux kernel
8064 module, for example.