Commit | Line | Data |
---|---|---|
b4867b3b PP |
1 | lttng-enable-channel(1) |
2 | ======================= | |
3 | ||
4 | ||
5 | NAME | |
6 | ---- | |
7 | lttng-enable-channel - Create or enable LTTng channels | |
8 | ||
9 | ||
10 | SYNOPSIS | |
11 | -------- | |
12 | Create a Linux kernel channel: | |
13 | ||
14 | [verse] | |
ce19b9ed | 15 | *lttng* ['linkgenoptions:(GENERAL OPTIONS)'] *enable-channel* option:--kernel |
240311ba | 16 | [option:--overwrite] [option:--output=(`mmap` | `splice`)] |
b4867b3b PP |
17 | [option:--subbuf-size='SIZE'] [option:--num-subbuf='COUNT'] |
18 | [option:--switch-timer='PERIODUS'] [option:--read-timer='PERIODUS'] | |
19 | [option:--tracefile-size='SIZE'] [option:--tracefile-count='COUNT'] | |
20 | [option:--session='SESSION'] 'CHANNEL' | |
21 | ||
22 | Create a user space channel: | |
23 | ||
24 | [verse] | |
ce19b9ed | 25 | *lttng* ['linkgenoptions:(GENERAL OPTIONS)'] *enable-channel* option:--userspace |
240311ba | 26 | [option:--overwrite] [option:--buffers-pid] |
b4867b3b PP |
27 | [option:--subbuf-size='SIZE'] [option:--num-subbuf='COUNT'] |
28 | [option:--switch-timer='PERIODUS'] [option:--read-timer='PERIODUS'] | |
29 | [option:--tracefile-size='SIZE'] [option:--tracefile-count='COUNT'] | |
30 | [option:--session='SESSION'] 'CHANNEL' | |
31 | ||
32 | Enable existing channel(s): | |
33 | ||
34 | [verse] | |
ce19b9ed | 35 | *lttng* ['linkgenoptions:(GENERAL OPTIONS)'] *enable-channel* (option:--userspace | option:--kernel) |
b4867b3b PP |
36 | [option:--session='SESSION'] 'CHANNEL'[,'CHANNEL']... |
37 | ||
38 | ||
39 | DESCRIPTION | |
40 | ----------- | |
41 | The `lttng enable-channel` command can create a new channel, or enable | |
42 | one or more existing and disabled ones. | |
43 | ||
44 | A channel is the owner of sub-buffers holding recorded events. Event, | |
7c1a4458 | 45 | rules, when created using man:lttng-enable-event(1), are always |
b4867b3b PP |
46 | assigned to a channel. When creating a new channel, many parameters |
47 | related to those sub-buffers can be fine-tuned. They are described in | |
48 | the subsections below. | |
49 | ||
50 | When 'CHANNEL' does not name an existing channel, a channel named | |
51 | 'CHANNEL' is created. Otherwise, the disabled channel named 'CHANNEL' | |
52 | is enabled. | |
53 | ||
7c1a4458 | 54 | Note that the man:lttng-enable-event(1) command can automatically |
b4867b3b PP |
55 | create default channels when no channel exist. |
56 | ||
57 | A channel is always contained in a tracing session | |
7c1a4458 | 58 | (see man:lttng-create(1) for creating a tracing session). The |
b4867b3b PP |
59 | session in which a channel is created using `lttng enable-channel` can |
60 | be specified using the option:--session option. If the option:--session | |
61 | option is omitted, the current tracing session is targeted. | |
62 | ||
63 | Existing enabled channels can be disabled using | |
7c1a4458 PP |
64 | man:lttng-disable-channel(1). Channels of a given session can be |
65 | listed using man:lttng-list(1). | |
b4867b3b | 66 | |
1076f2b7 PP |
67 | See the <<limitations,LIMITATIONS>> section below for a list of |
68 | limitations of this command to consider. | |
b4867b3b PP |
69 | |
70 | ||
71 | Event loss modes | |
72 | ~~~~~~~~~~~~~~~~ | |
73 | LTTng tracers are non-blocking: when no empty sub-buffer exists, | |
74 | losing events is acceptable when the alternative would be to cause | |
75 | substantial delays in the instrumented application's execution. | |
76 | ||
77 | LTTng privileges performance over integrity, aiming at perturbing the | |
78 | traced system as little as possible in order to make tracing of subtle | |
79 | race conditions and rare interrupt cascades possible. | |
80 | ||
81 | When it comes to losing events because no empty sub-buffer is available, | |
82 | the channel's event loss mode, specified by one of the option:--discard | |
83 | and option:--overwrite options, determines what to do amongst: | |
84 | ||
85 | Discard:: | |
86 | Drop the newest events until a sub-buffer is released. | |
87 | ||
88 | Overwrite:: | |
89 | Clear the sub-buffer containing the oldest recorded events and start | |
90 | recording the newest events there. This mode is sometimes called | |
91 | _flight recorder mode_ because it behaves like a flight recorder: | |
92 | always keep a fixed amount of the latest data. | |
93 | ||
94 | Which mechanism to choose depends on the context: prioritize the newest | |
95 | or the oldest events in the ring buffer? | |
96 | ||
97 | Beware that, in overwrite mode (option:--overwrite option), a whole | |
98 | sub-buffer is abandoned as soon as a new event doesn't find an empty | |
99 | sub-buffer, whereas in discard mode (option:--discard option), only the | |
100 | event that doesn't fit is discarded. | |
101 | ||
102 | Also note that a count of lost events is incremented and saved in the | |
103 | trace itself when an event is lost in discard mode, whereas no | |
104 | information is kept when a sub-buffer gets overwritten before being | |
105 | committed. | |
106 | ||
107 | The probability of losing events, if it is experience in a given | |
108 | context, can be reduced by fine-tuning the sub-buffers count and size | |
109 | (see next subsection). | |
110 | ||
111 | ||
112 | Sub-buffers count and size | |
113 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
114 | The option:--num-subbuf and option:--subbuf-size options respectively | |
115 | set the number of sub-buffers and their individual size when creating | |
116 | a new channel. | |
117 | ||
118 | Note that there is a noticeable tracer's CPU overhead introduced when | |
119 | switching sub-buffers (marking a full one as consumable and switching | |
120 | to an empty one for the following events to be recorded). Knowing this, | |
121 | the following list presents a few practical situations along with how | |
122 | to configure sub-buffers for them when creating a channel in overwrite | |
123 | mode (option:--overwrite option): | |
124 | ||
125 | High event throughput:: | |
126 | In general, prefer bigger sub-buffers to lower the risk of losing | |
127 | events. Having bigger sub-buffers also ensures a lower sub-buffer | |
128 | switching frequency. The number of sub-buffers is only meaningful | |
129 | if the channel is enabled in overwrite mode: in this case, if a | |
130 | sub-buffer overwrite happens, the other sub-buffers | |
131 | are left unaltered. | |
132 | ||
133 | Low event throughput:: | |
134 | In general, prefer smaller sub-buffers since the risk of losing | |
135 | events is already low. Since events happen less frequently, the | |
136 | sub-buffer switching frequency should remain low and thus the | |
137 | tracer's overhead should not be a problem. | |
138 | ||
139 | Low memory system:: | |
140 | If the target system has a low memory limit, prefer fewer first, | |
141 | then smaller sub-buffers. Even if the system is limited in memory, | |
142 | it is recommended to keep the sub-buffers as big as possible to | |
143 | avoid a high sub-buffer switching frequency. | |
144 | ||
145 | In discard mode (option:--discard option), the sub-buffers count | |
146 | parameter is pointless: using two sub-buffers and setting their size | |
147 | according to the requirements of the context is fine. | |
148 | ||
149 | ||
150 | Switch and read timers | |
151 | ~~~~~~~~~~~~~~~~~~~~~~ | |
152 | When a channel's switch timer fires, a sub-buffer switch happens. This | |
153 | timer may be used to ensure that event data is consumed and committed | |
154 | to trace files periodically in case of a low event throughput. | |
155 | ||
156 | It's also convenient when big sub-buffers are used to cope with sporadic | |
157 | high event throughput, even if the throughput is normally lower. | |
158 | ||
159 | By default, a notification mechanism is used to signal a full sub-buffer | |
160 | so that it can be consumed. When such notifications must be avoided, | |
161 | for example in real-time applications, the channel's read timer can be | |
162 | used instead. When the read timer fires, sub-buffers are checked for | |
163 | consumption when they are full. | |
164 | ||
165 | ||
166 | Buffering scheme | |
167 | ~~~~~~~~~~~~~~~~ | |
168 | In the user space tracing domain, two buffering schemes are available | |
169 | when creating a channel: | |
170 | ||
171 | Per-process buffering (option:--buffers-pid option):: | |
172 | Keep one ring buffer per process. | |
173 | ||
174 | Per-user buffering (option:--buffers-uid option):: | |
175 | Keep one ring buffer for all the processes of a single user. | |
176 | ||
177 | The per-process buffering scheme consumes more memory than the per-user | |
178 | option if more than one process is instrumented for LTTng-UST. | |
179 | However, per-process buffering ensures that one process having a high | |
180 | event throughput won't fill all the shared sub-buffers, only its own. | |
181 | ||
182 | The Linux kernel tracing domain only has one available buffering scheme | |
183 | which is to use a single ring buffer for the whole system | |
184 | (option:--buffers-global option). | |
185 | ||
186 | ||
187 | Trace files limit and size | |
188 | ~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
189 | By default, trace files can grow as large as needed. The maximum size | |
190 | of each trace file written by a channel can be set on creation using the | |
191 | option:--tracefile-size option. When such a trace file's size reaches | |
192 | the channel's fixed maximum size, another trace file is created to hold | |
193 | the next recorded events. A file count is appended to each trace file | |
194 | name in this case. | |
195 | ||
196 | If the option:--tracefile-size option is used, the maximum number of | |
197 | created trace files is unlimited. To limit them, the | |
198 | option:--tracefile-count option can be used. This option is always used | |
199 | in conjunction with the option:--tracefile-size option. | |
200 | ||
201 | For example, consider this command: | |
202 | ||
d4f093aa | 203 | [role="term"] |
b4867b3b PP |
204 | ----------------------------------------------------- |
205 | lttng enable-channel --kernel --tracefile-size=4096 \ | |
206 | --tracefile-count=32 my-channel | |
207 | ----------------------------------------------------- | |
208 | ||
209 | Here, for each stream, the maximum size of each trace file is | |
210 | 4 kiB and there can be a maximum of 32 different files. When there is | |
211 | no space left in the last file, _trace file rotation_ happens: the first | |
212 | file is cleared and new sub-buffers containing events are written there. | |
213 | ||
214 | ||
215 | include::common-cmd-options-head.txt[] | |
216 | ||
217 | ||
218 | Domain | |
219 | ~~~~~~ | |
220 | One of: | |
221 | ||
222 | option:-k, option:--kernel:: | |
223 | Enable channel in the Linux kernel domain. | |
224 | ||
225 | option:-u, option:--userspace:: | |
226 | Enable channel in the user space domain. | |
227 | ||
228 | ||
229 | Target | |
230 | ~~~~~~ | |
59b19c3c | 231 | option:-s 'SESSION', option:--session='SESSION':: |
b4867b3b PP |
232 | Create or enable channel in the tracing session named 'SESSION' |
233 | instead of the current tracing session. | |
234 | ||
235 | ||
236 | Event loss mode | |
237 | ~~~~~~~~~~~~~~~ | |
238 | One of: | |
239 | ||
240 | option:--discard:: | |
241 | Discard events when sub-buffers are full (default). | |
242 | ||
243 | option:--overwrite:: | |
244 | Flight recorder mode: always keep a fixed amount of the latest | |
245 | data. | |
246 | ||
247 | ||
248 | Sub-buffers | |
249 | ~~~~~~~~~~~ | |
250 | option:--num-subbuf='COUNT':: | |
251 | Use 'COUNT' sub-buffers. Rounded up to the next power of two. | |
252 | + | |
253 | Default values: | |
254 | + | |
c93eadad PP |
255 | * option:--userspace and option:--buffers-uid options: |
256 | {default_ust_uid_channel_subbuf_num} | |
257 | * option:--userspace and option:--buffers-pid options: | |
258 | {default_ust_pid_channel_subbuf_num} | |
259 | * option:--kernel option: {default_kernel_channel_subbuf_num} | |
260 | * `metadata` channel: {default_metadata_subbuf_num} | |
b4867b3b PP |
261 | |
262 | option:--subbuf-size='SIZE':: | |
263 | Set the individual size of sub-buffers to 'SIZE' bytes. | |
264 | The `k` (kiB), `M` (MiB), and `G` (GiB) suffixes are supported. | |
265 | Rounded up to the next power of two. | |
266 | + | |
267 | The minimum sub-buffer size, for each tracer, is the maximum value | |
268 | between the default below and the system's page size. The following | |
269 | command shows the current system's page size: `getconf PAGE_SIZE`. | |
270 | + | |
271 | Default values: | |
272 | + | |
c93eadad PP |
273 | * option:--userspace and option:--buffers-uid options: |
274 | {default_ust_uid_channel_subbuf_size} | |
275 | * option:--userspace and option:--buffers-pid options: | |
276 | {default_ust_pid_channel_subbuf_size} | |
277 | * option:--kernel option: {default_kernel_channel_subbuf_size} | |
278 | * `metadata` channel: {default_metadata_subbuf_size} | |
b4867b3b PP |
279 | |
280 | option:--output='TYPE':: | |
281 | Set channel's output type to 'TYPE'. | |
282 | + | |
283 | Available types: `mmap` (always available) and `splice` (only available | |
284 | with the option:--kernel option). | |
285 | + | |
286 | Default values: | |
287 | + | |
288 | * option:--userspace and option:--buffers-uid options: `mmap` | |
289 | * option:--userspace and option:--buffers-pid options: `mmap` | |
290 | * option:--kernel option: `splice` | |
291 | * `metadata` channel: `mmap` | |
292 | ||
293 | Buffering scheme | |
294 | ~~~~~~~~~~~~~~~~ | |
295 | One of: | |
296 | ||
297 | option:--buffers-global:: | |
298 | Use shared sub-buffers for the whole system (only available with the | |
299 | option:--kernel option). | |
300 | ||
301 | option:--buffers-pid:: | |
302 | Use different sub-buffers for each traced process (only available | |
303 | with the the option:--userspace option). This is the default | |
304 | buffering scheme for user space channels. | |
305 | ||
306 | option:--buffers-uid:: | |
307 | Use shared sub-buffers for all the processes of the user running | |
308 | the command (only available with the option:--userspace option). | |
309 | ||
310 | ||
311 | Trace files | |
312 | ~~~~~~~~~~~ | |
313 | option:--tracefile-count='COUNT':: | |
314 | Limit the number of trace files created by this channel to | |
c93eadad PP |
315 | 'COUNT'. 0 means unlimited. Default: |
316 | {default_channel_tracefile_count}. | |
b4867b3b PP |
317 | + |
318 | Use this option in conjunction with the option:--tracefile-size option. | |
319 | + | |
320 | The file count within a stream is appended to each created trace | |
321 | file. If 'COUNT' files are created and more events need to be recorded, | |
322 | the first trace file of the stream is cleared and used again. | |
323 | ||
324 | option:--tracefile-size='SIZE':: | |
325 | Set the maximum size of each trace file written by | |
326 | this channel within a stream to 'SIZE' bytes. 0 means unlimited. | |
c93eadad | 327 | Default: {default_channel_tracefile_size}. |
b4867b3b PP |
328 | + |
329 | Note: traces generated with this option may inaccurately report | |
330 | discarded events as of CTF 1.8. | |
331 | ||
332 | ||
333 | Timers | |
334 | ~~~~~~ | |
335 | option:--read-timer:: | |
336 | Set the channel's read timer's period to 'PERIODUS' µs. 0 means a | |
337 | disabled read timer. | |
338 | + | |
339 | Default values: | |
340 | + | |
c93eadad PP |
341 | * option:--userspace and option:--buffers-uid options: |
342 | {default_ust_uid_channel_read_timer} | |
343 | * option:--userspace and option:--buffers-pid options: | |
344 | {default_ust_pid_channel_read_timer} | |
345 | * option:--kernel option: {default_kernel_channel_read_timer} | |
346 | * `metadata` channel: {default_metadata_read_timer} | |
b4867b3b PP |
347 | |
348 | option:--switch-timer='PERIODUS':: | |
349 | Set the channel's switch timer's period to 'PERIODUS' µs. 0 means | |
c93eadad PP |
350 | a disabled switch timer. |
351 | + | |
352 | Default values: | |
353 | + | |
354 | * option:--userspace and option:--buffers-uid options: | |
355 | {default_ust_uid_channel_switch_timer} | |
356 | * option:--userspace and option:--buffers-pid options: | |
357 | {default_ust_pid_channel_switch_timer} | |
358 | * option:--kernel option: {default_kernel_channel_switch_timer} | |
359 | * `metadata` channel: {default_metadata_switch_timer} | |
b4867b3b PP |
360 | |
361 | ||
362 | include::common-cmd-help-options.txt[] | |
363 | ||
364 | ||
1076f2b7 PP |
365 | [[limitations]] |
366 | LIMITATIONS | |
367 | ----------- | |
368 | As of this version of LTTng, it is not possible to perform the following | |
369 | actions with the `lttng enable-channel` command: | |
370 | ||
371 | * Reconfigure a channel once it is created. | |
372 | * Re-enable a disabled channel once its tracing session has been active | |
373 | at least once. | |
374 | * Create a channel once its tracing session has been active | |
375 | at least once. | |
376 | * Create a user space channel with a given buffering scheme | |
377 | (option:--buffers-uid or option:--buffers-pid options) and create | |
378 | a second user space channel with a different buffering scheme in the | |
379 | same tracing session. | |
380 | ||
381 | ||
b4867b3b PP |
382 | include::common-cmd-footer.txt[] |
383 | ||
384 | ||
385 | SEE ALSO | |
386 | -------- | |
7c1a4458 PP |
387 | man:lttng-disable-channel(1), |
388 | man:lttng(1) |