]>
Commit | Line | Data |
---|---|---|
55fc19a1 SG |
1 | <!-- |
2 | ||
3 | lxc: linux Container library | |
4 | ||
5 | (C) Copyright IBM Corp. 2007, 2008 | |
6 | ||
7 | Authors: | |
8 | Daniel Lezcano <daniel.lezcano at free.fr> | |
9 | ||
10 | This library is free software; you can redistribute it and/or | |
11 | modify it under the terms of the GNU Lesser General Public | |
12 | License as published by the Free Software Foundation; either | |
13 | version 2.1 of the License, or (at your option) any later version. | |
14 | ||
15 | This library is distributed in the hope that it will be useful, | |
16 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
17 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
18 | Lesser General Public License for more details. | |
19 | ||
20 | You should have received a copy of the GNU Lesser General Public | |
21 | License along with this library; if not, write to the Free Software | |
22 | Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | |
23 | ||
24 | --> | |
25 | ||
26 | <!DOCTYPE refentry PUBLIC @docdtd@ [ | |
27 | ||
28 | <!ENTITY seealso SYSTEM "@builddir@/see_also.sgml"> | |
29 | ]> | |
30 | ||
31 | <refentry> | |
32 | ||
33 | <docinfo><date>@LXC_GENERATE_DATE@</date></docinfo> | |
34 | ||
35 | <refmeta> | |
36 | <refentrytitle>lxc.container.conf</refentrytitle> | |
37 | <manvolnum>5</manvolnum> | |
38 | </refmeta> | |
39 | ||
40 | <refnamediv> | |
41 | <refname>lxc.container.conf</refname> | |
42 | ||
43 | <refpurpose> | |
44 | LXC container configuration file | |
45 | </refpurpose> | |
46 | </refnamediv> | |
47 | ||
48 | <refsect1> | |
49 | <title>Description</title> | |
50 | ||
51 | <para> | |
52 | The linux containers (<command>lxc</command>) are always created | |
53 | before being used. This creation defines a set of system | |
54 | resources to be virtualized / isolated when a process is using | |
55 | the container. By default, the pids, sysv ipc and mount points | |
56 | are virtualized and isolated. The other system resources are | |
57 | shared across containers, until they are explicitly defined in | |
58 | the configuration file. For example, if there is no network | |
59 | configuration, the network will be shared between the creator of | |
60 | the container and the container itself, but if the network is | |
61 | specified, a new network stack is created for the container and | |
62 | the container can no longer use the network of its ancestor. | |
63 | </para> | |
64 | ||
65 | <para> | |
66 | The configuration file defines the different system resources to | |
67 | be assigned for the container. At present, the utsname, the | |
68 | network, the mount points, the root file system, the user namespace, | |
69 | and the control groups are supported. | |
70 | </para> | |
71 | ||
72 | <para> | |
73 | Each option in the configuration file has the form <command>key | |
74 | = value</command> fitting in one line. The '#' character means | |
ff689149 MH |
75 | the line is a comment. List options, like capabilities and cgroups |
76 | options, can be used with no value to clear any previously | |
77 | defined values of that option. | |
55fc19a1 SG |
78 | </para> |
79 | ||
80 | <refsect2> | |
81 | <title>Configuration</title> | |
82 | <para> | |
c464fd7e SG |
83 | In order to ease administration of multiple related containers, it |
84 | is possible to have a container configuration file cause another | |
85 | file to be loaded. For instance, network configuration | |
86 | can be defined in one common file which is included by multiple | |
87 | containers. Then, if the containers are moved to another host, | |
88 | only one file may need to be updated. | |
55fc19a1 SG |
89 | </para> |
90 | ||
91 | <variablelist> | |
c464fd7e SG |
92 | <varlistentry> |
93 | <term> | |
94 | <option>lxc.include</option> | |
95 | </term> | |
96 | <listitem> | |
97 | <para> | |
98 | Specify the file to be included. The included file must be | |
99 | in the same valid lxc configuration file format. | |
100 | </para> | |
101 | </listitem> | |
102 | </varlistentry> | |
55fc19a1 SG |
103 | </variablelist> |
104 | </refsect2> | |
105 | ||
106 | <refsect2> | |
107 | <title>Architecture</title> | |
108 | <para> | |
c464fd7e SG |
109 | Allows one to set the architecture for the container. For example, |
110 | set a 32bits architecture for a container running 32bits | |
111 | binaries on a 64bits host. This fixes the container scripts | |
112 | which rely on the architecture to do some work like | |
113 | downloading the packages. | |
55fc19a1 SG |
114 | </para> |
115 | ||
116 | <variablelist> | |
c464fd7e SG |
117 | <varlistentry> |
118 | <term> | |
119 | <option>lxc.arch</option> | |
120 | </term> | |
121 | <listitem> | |
122 | <para> | |
123 | Specify the architecture for the container. | |
124 | </para> | |
125 | <para> | |
126 | Valid options are | |
127 | <option>x86</option>, | |
128 | <option>i686</option>, | |
129 | <option>x86_64</option>, | |
130 | <option>amd64</option> | |
131 | </para> | |
132 | </listitem> | |
133 | </varlistentry> | |
55fc19a1 SG |
134 | </variablelist> |
135 | ||
136 | </refsect2> | |
137 | ||
138 | <refsect2> | |
139 | <title>Hostname</title> | |
140 | <para> | |
c464fd7e SG |
141 | The utsname section defines the hostname to be set for the |
142 | container. That means the container can set its own hostname | |
143 | without changing the one from the system. That makes the | |
144 | hostname private for the container. | |
55fc19a1 SG |
145 | </para> |
146 | <variablelist> | |
c464fd7e SG |
147 | <varlistentry> |
148 | <term> | |
149 | <option>lxc.utsname</option> | |
150 | </term> | |
151 | <listitem> | |
152 | <para> | |
153 | specify the hostname for the container | |
154 | </para> | |
155 | </listitem> | |
156 | </varlistentry> | |
55fc19a1 SG |
157 | </variablelist> |
158 | </refsect2> | |
159 | ||
160 | <refsect2> | |
161 | <title>Halt signal</title> | |
162 | <para> | |
936762f3 BP |
163 | Allows one to specify signal name or number, sent by lxc-stop to the |
164 | container's init process to cleanly shutdown the container. Different | |
165 | init systems could use different signals to perform clean shutdown | |
166 | sequence. This option allows the signal to be specified in kill(1) | |
167 | fashion, e.g. SIGPWR, SIGRTMIN+14, SIGRTMAX-10 or plain number. The | |
168 | default signal is SIGPWR. | |
55fc19a1 SG |
169 | </para> |
170 | <variablelist> | |
936762f3 BP |
171 | <varlistentry> |
172 | <term> | |
173 | <option>lxc.haltsignal</option> | |
174 | </term> | |
175 | <listitem> | |
176 | <para> | |
177 | specify the signal used to halt the container | |
178 | </para> | |
179 | </listitem> | |
180 | </varlistentry> | |
181 | </variablelist> | |
182 | </refsect2> | |
183 | ||
184 | <refsect2> | |
185 | <title>Reboot signal</title> | |
186 | <para> | |
187 | Allows one to specify signal name or number, sent by lxc-stop to | |
188 | reboot the container. This option allows signal to be specified in | |
189 | kill(1) fashion, e.g. SIGTERM, SIGRTMIN+14, SIGRTMAX-10 or plain number. | |
190 | The default signal is SIGINT. | |
191 | </para> | |
192 | <variablelist> | |
193 | <varlistentry> | |
194 | <term> | |
195 | <option>lxc.rebootsignal</option> | |
196 | </term> | |
197 | <listitem> | |
198 | <para> | |
199 | specify the signal used to reboot the container | |
200 | </para> | |
201 | </listitem> | |
202 | </varlistentry> | |
55fc19a1 SG |
203 | </variablelist> |
204 | </refsect2> | |
205 | ||
206 | <refsect2> | |
207 | <title>Stop signal</title> | |
208 | <para> | |
936762f3 BP |
209 | Allows one to specify signal name or number, sent by lxc-stop to forcibly |
210 | shutdown the container. This option allows signal to be specified in | |
211 | kill(1) fashion, e.g. SIGKILL, SIGRTMIN+14, SIGRTMAX-10 or plain number. | |
212 | The default signal is SIGKILL. | |
213 | </para> | |
214 | <variablelist> | |
215 | <varlistentry> | |
216 | <term> | |
217 | <option>lxc.stopsignal</option> | |
218 | </term> | |
219 | <listitem> | |
220 | <para> | |
221 | specify the signal used to stop the container | |
222 | </para> | |
223 | </listitem> | |
224 | </varlistentry> | |
55fc19a1 SG |
225 | </variablelist> |
226 | </refsect2> | |
227 | ||
67c660d0 SG |
228 | <refsect2> |
229 | <title>Init command</title> | |
230 | <para> | |
231 | Sets the command to use as the init system for the containers. | |
232 | ||
233 | This option is ignored when using lxc-execute. | |
234 | ||
235 | Defaults to: /sbin/init | |
236 | </para> | |
237 | <variablelist> | |
936762f3 BP |
238 | <varlistentry> |
239 | <term> | |
240 | <option>lxc.init_cmd</option> | |
241 | </term> | |
242 | <listitem> | |
243 | <para> | |
244 | Absolute path from container rootfs to the binary to use as init. | |
245 | </para> | |
246 | </listitem> | |
247 | </varlistentry> | |
67c660d0 SG |
248 | </variablelist> |
249 | </refsect2> | |
250 | ||
dbca9237 PT |
251 | <refsect2> |
252 | <title>Init ID</title> | |
253 | <para> | |
254 | Sets the UID/GID to use for the init system, and subsequent command, executed by lxc-execute. | |
255 | ||
256 | These options are only used when lxc-execute is started in a private user namespace. | |
257 | ||
258 | Defaults to: UID(0), GID(0) | |
259 | </para> | |
260 | <variablelist> | |
261 | <varlistentry> | |
262 | <term> | |
263 | <option>lxc.init_uid</option> | |
264 | </term> | |
265 | <listitem> | |
266 | <para> | |
267 | UID to use within a private user namesapce for init. | |
268 | </para> | |
269 | </listitem> | |
270 | </varlistentry> | |
271 | <varlistentry> | |
272 | <term> | |
273 | <option>lxc.init_gid</option> | |
274 | </term> | |
275 | <listitem> | |
276 | <para> | |
277 | GID to use within a private user namesapce for init. | |
278 | </para> | |
279 | </listitem> | |
280 | </varlistentry> | |
281 | </variablelist> | |
282 | </refsect2> | |
283 | ||
4e6eb26b CB |
284 | <refsect2> |
285 | <title>Ephemeral</title> | |
286 | <para> | |
287 | Allows one to specify whether a container will be destroyed on shutdown. | |
288 | </para> | |
289 | <variablelist> | |
290 | <varlistentry> | |
291 | <term> | |
292 | <option>lxc.ephemeral</option> | |
293 | </term> | |
294 | <listitem> | |
295 | <para> | |
296 | The only allowed values are 0 and 1. Set this to 1 to destroy a | |
297 | container on shutdown. | |
298 | </para> | |
299 | </listitem> | |
300 | </varlistentry> | |
301 | </variablelist> | |
302 | </refsect2> | |
303 | ||
55fc19a1 SG |
304 | <refsect2> |
305 | <title>Network</title> | |
306 | <para> | |
c464fd7e SG |
307 | The network section defines how the network is virtualized in |
308 | the container. The network virtualization acts at layer | |
309 | two. In order to use the network virtualization, parameters | |
310 | must be specified to define the network interfaces of the | |
311 | container. Several virtual interfaces can be assigned and used | |
312 | in a container even if the system has only one physical | |
313 | network interface. | |
55fc19a1 SG |
314 | </para> |
315 | <variablelist> | |
020104c3 MH |
316 | <varlistentry> |
317 | <term> | |
318 | <option>lxc.network</option> | |
319 | </term> | |
320 | <listitem> | |
321 | <para> | |
322 | may be used without a value to clear all previous network options. | |
323 | </para> | |
324 | </listitem> | |
325 | </varlistentry> | |
c464fd7e SG |
326 | <varlistentry> |
327 | <term> | |
328 | <option>lxc.network.type</option> | |
329 | </term> | |
330 | <listitem> | |
331 | <para> | |
332 | specify what kind of network virtualization to be used | |
333 | for the container. Each time | |
334 | a <option>lxc.network.type</option> field is found a new | |
335 | round of network configuration begins. In this way, | |
336 | several network virtualization types can be specified | |
337 | for the same container, as well as assigning several | |
338 | network interfaces for one container. The different | |
339 | virtualization types can be: | |
340 | </para> | |
341 | ||
342 | <para> | |
343 | <option>none:</option> will cause the container to share | |
344 | the host's network namespace. This means the host | |
345 | network devices are usable in the container. It also | |
346 | means that if both the container and host have upstart as | |
347 | init, 'halt' in a container (for instance) will shut down the | |
348 | host. | |
349 | </para> | |
350 | ||
351 | <para> | |
352 | <option>empty:</option> will create only the loopback | |
353 | interface. | |
354 | </para> | |
355 | ||
356 | <para> | |
38005c54 MA |
357 | <option>veth:</option> a virtual ethernet pair |
358 | device is created with one side assigned to the container | |
359 | and the other side attached to a bridge specified by | |
360 | the <option>lxc.network.link</option> option. | |
361 | If the bridge is not specified, then the veth pair device | |
362 | will be created but not attached to any bridge. | |
363 | Otherwise, the bridge has to be created on the system | |
364 | before starting the container. | |
365 | <command>lxc</command> won't handle any | |
366 | configuration outside of the container. | |
367 | By default, <command>lxc</command> chooses a name for the | |
c464fd7e | 368 | network device belonging to the outside of the |
38005c54 MA |
369 | container, but if you wish to handle |
370 | this name yourselves, you can tell <command>lxc</command> | |
c464fd7e SG |
371 | to set a specific name with |
372 | the <option>lxc.network.veth.pair</option> option (except for | |
373 | unprivileged containers where this option is ignored for security | |
374 | reasons). | |
375 | </para> | |
376 | ||
377 | <para> | |
378 | <option>vlan:</option> a vlan interface is linked with | |
379 | the interface specified by | |
380 | the <option>lxc.network.link</option> and assigned to | |
381 | the container. The vlan identifier is specified with the | |
382 | option <option>lxc.network.vlan.id</option>. | |
383 | </para> | |
384 | ||
385 | <para> | |
386 | <option>macvlan:</option> a macvlan interface is linked | |
387 | with the interface specified by | |
388 | the <option>lxc.network.link</option> and assigned to | |
389 | the container. | |
390 | <option>lxc.network.macvlan.mode</option> specifies the | |
391 | mode the macvlan will use to communicate between | |
392 | different macvlan on the same upper device. The accepted | |
c15ea607 EL |
393 | modes are <option>private</option>, <option>vepa</option>, |
394 | <option>bridge</option> and <option>passthru</option>. | |
395 | In <option>private</option> mode, the device never | |
396 | communicates with any other device on the same upper_dev (default). | |
397 | In <option>vepa</option> mode, the new Virtual Ethernet Port | |
c464fd7e SG |
398 | Aggregator (VEPA) mode, it assumes that the adjacent |
399 | bridge returns all frames where both source and | |
400 | destination are local to the macvlan port, i.e. the | |
401 | bridge is set up as a reflective relay. Broadcast | |
402 | frames coming in from the upper_dev get flooded to all | |
403 | macvlan interfaces in VEPA mode, local frames are not | |
c15ea607 | 404 | delivered locally. In <option>bridge</option> mode, it |
c464fd7e SG |
405 | provides the behavior of a simple bridge between |
406 | different macvlan interfaces on the same port. Frames | |
407 | from one interface to another one get delivered directly | |
408 | and are not sent out externally. Broadcast frames get | |
409 | flooded to all other bridge ports and to the external | |
410 | interface, but when they come back from a reflective | |
411 | relay, we don't deliver them again. Since we know all | |
412 | the MAC addresses, the macvlan bridge mode does not | |
c15ea607 EL |
413 | require learning or STP like the bridge module does. In |
414 | <option>passthru</option> mode, all frames received by | |
415 | the physical interface are forwarded to the macvlan | |
416 | interface. Only one macvlan interface in <option>passthru</option> | |
417 | mode is possible for one physical interface. | |
c464fd7e SG |
418 | </para> |
419 | ||
420 | <para> | |
421 | <option>phys:</option> an already existing interface | |
422 | specified by the <option>lxc.network.link</option> is | |
423 | assigned to the container. | |
424 | </para> | |
425 | </listitem> | |
426 | </varlistentry> | |
427 | ||
428 | <varlistentry> | |
429 | <term> | |
430 | <option>lxc.network.flags</option> | |
431 | </term> | |
432 | <listitem> | |
433 | <para> | |
434 | specify an action to do for the | |
435 | network. | |
436 | </para> | |
437 | ||
438 | <para><option>up:</option> activates the interface. | |
439 | </para> | |
440 | </listitem> | |
441 | </varlistentry> | |
442 | ||
443 | <varlistentry> | |
444 | <term> | |
445 | <option>lxc.network.link</option> | |
446 | </term> | |
447 | <listitem> | |
448 | <para> | |
449 | specify the interface to be used for real network | |
450 | traffic. | |
451 | </para> | |
452 | </listitem> | |
453 | </varlistentry> | |
454 | ||
455 | <varlistentry> | |
456 | <term> | |
457 | <option>lxc.network.mtu</option> | |
458 | </term> | |
459 | <listitem> | |
460 | <para> | |
461 | specify the maximum transfer unit for this interface. | |
462 | </para> | |
463 | </listitem> | |
464 | </varlistentry> | |
465 | ||
466 | <varlistentry> | |
467 | <term> | |
468 | <option>lxc.network.name</option> | |
469 | </term> | |
470 | <listitem> | |
471 | <para> | |
472 | the interface name is dynamically allocated, but if | |
473 | another name is needed because the configuration files | |
474 | being used by the container use a generic name, | |
475 | eg. eth0, this option will rename the interface in the | |
476 | container. | |
477 | </para> | |
478 | </listitem> | |
479 | </varlistentry> | |
480 | ||
481 | <varlistentry> | |
482 | <term> | |
483 | <option>lxc.network.hwaddr</option> | |
484 | </term> | |
485 | <listitem> | |
486 | <para> | |
487 | the interface mac address is dynamically allocated by | |
488 | default to the virtual interface, but in some cases, | |
489 | this is needed to resolve a mac address conflict or to | |
490 | always have the same link-local ipv6 address. | |
491 | Any "x" in address will be replaced by random value, | |
492 | this allows setting hwaddr templates. | |
493 | </para> | |
494 | </listitem> | |
495 | </varlistentry> | |
496 | ||
497 | <varlistentry> | |
498 | <term> | |
499 | <option>lxc.network.ipv4</option> | |
500 | </term> | |
501 | <listitem> | |
502 | <para> | |
503 | specify the ipv4 address to assign to the virtualized | |
504 | interface. Several lines specify several ipv4 addresses. | |
505 | The address is in format x.y.z.t/m, | |
506 | eg. 192.168.1.123/24. The broadcast address should be | |
507 | specified on the same line, right after the ipv4 | |
508 | address. | |
509 | </para> | |
510 | </listitem> | |
511 | </varlistentry> | |
512 | ||
513 | <varlistentry> | |
514 | <term> | |
515 | <option>lxc.network.ipv4.gateway</option> | |
516 | </term> | |
517 | <listitem> | |
518 | <para> | |
519 | specify the ipv4 address to use as the gateway inside the | |
520 | container. The address is in format x.y.z.t, eg. | |
521 | 192.168.1.123. | |
522 | ||
523 | Can also have the special value <option>auto</option>, | |
524 | which means to take the primary address from the bridge | |
525 | interface (as specified by the | |
526 | <option>lxc.network.link</option> option) and use that as | |
527 | the gateway. <option>auto</option> is only available when | |
528 | using the <option>veth</option> and | |
529 | <option>macvlan</option> network types. | |
530 | </para> | |
531 | </listitem> | |
532 | </varlistentry> | |
533 | ||
534 | ||
535 | <varlistentry> | |
536 | <term> | |
537 | <option>lxc.network.ipv6</option> | |
538 | </term> | |
539 | <listitem> | |
540 | <para> | |
541 | specify the ipv6 address to assign to the virtualized | |
542 | interface. Several lines specify several ipv6 addresses. | |
543 | The address is in format x::y/m, | |
544 | eg. 2003:db8:1:0:214:1234:fe0b:3596/64 | |
545 | </para> | |
546 | </listitem> | |
547 | </varlistentry> | |
548 | ||
549 | <varlistentry> | |
550 | <term> | |
551 | <option>lxc.network.ipv6.gateway</option> | |
552 | </term> | |
553 | <listitem> | |
554 | <para> | |
555 | specify the ipv6 address to use as the gateway inside the | |
556 | container. The address is in format x::y, | |
557 | eg. 2003:db8:1:0::1 | |
558 | ||
559 | Can also have the special value <option>auto</option>, | |
560 | which means to take the primary address from the bridge | |
561 | interface (as specified by the | |
562 | <option>lxc.network.link</option> option) and use that as | |
563 | the gateway. <option>auto</option> is only available when | |
564 | using the <option>veth</option> and | |
565 | <option>macvlan</option> network types. | |
566 | </para> | |
567 | </listitem> | |
568 | </varlistentry> | |
569 | ||
570 | <varlistentry> | |
571 | <term> | |
572 | <option>lxc.network.script.up</option> | |
573 | </term> | |
574 | <listitem> | |
575 | <para> | |
576 | add a configuration option to specify a script to be | |
577 | executed after creating and configuring the network used | |
578 | from the host side. The following arguments are passed | |
579 | to the script: container name and config section name | |
580 | (net) Additional arguments depend on the config section | |
581 | employing a script hook; the following are used by the | |
582 | network system: execution context (up), network type | |
583 | (empty/veth/macvlan/phys), Depending on the network | |
584 | type, other arguments may be passed: | |
585 | veth/macvlan/phys. And finally (host-sided) device name. | |
586 | </para> | |
587 | <para> | |
588 | Standard output from the script is logged at debug level. | |
589 | Standard error is not logged, but can be captured by the | |
590 | hook redirecting its standard error to standard output. | |
591 | </para> | |
592 | </listitem> | |
593 | </varlistentry> | |
594 | ||
595 | <varlistentry> | |
596 | <term> | |
597 | <option>lxc.network.script.down</option> | |
598 | </term> | |
599 | <listitem> | |
600 | <para> | |
601 | add a configuration option to specify a script to be | |
602 | executed before destroying the network used from the | |
603 | host side. The following arguments are passed to the | |
604 | script: container name and config section name (net) | |
605 | Additional arguments depend on the config section | |
606 | employing a script hook; the following are used by the | |
607 | network system: execution context (down), network type | |
608 | (empty/veth/macvlan/phys), Depending on the network | |
609 | type, other arguments may be passed: | |
610 | veth/macvlan/phys. And finally (host-sided) device name. | |
611 | </para> | |
612 | <para> | |
613 | Standard output from the script is logged at debug level. | |
614 | Standard error is not logged, but can be captured by the | |
615 | hook redirecting its standard error to standard output. | |
616 | </para> | |
617 | </listitem> | |
618 | </varlistentry> | |
55fc19a1 SG |
619 | </variablelist> |
620 | </refsect2> | |
621 | ||
622 | <refsect2> | |
623 | <title>New pseudo tty instance (devpts)</title> | |
624 | <para> | |
c464fd7e SG |
625 | For stricter isolation the container can have its own private |
626 | instance of the pseudo tty. | |
55fc19a1 SG |
627 | </para> |
628 | <variablelist> | |
c464fd7e SG |
629 | <varlistentry> |
630 | <term> | |
631 | <option>lxc.pts</option> | |
632 | </term> | |
633 | <listitem> | |
634 | <para> | |
635 | If set, the container will have a new pseudo tty | |
636 | instance, making this private to it. The value specifies | |
55fc19a1 SG |
637 | the maximum number of pseudo ttys allowed for a pts |
638 | instance (this limitation is not implemented yet). | |
c464fd7e SG |
639 | </para> |
640 | </listitem> | |
641 | </varlistentry> | |
55fc19a1 SG |
642 | </variablelist> |
643 | </refsect2> | |
644 | ||
645 | <refsect2> | |
646 | <title>Container system console</title> | |
647 | <para> | |
c464fd7e SG |
648 | If the container is configured with a root filesystem and the |
649 | inittab file is setup to use the console, you may want to specify | |
650 | where the output of this console goes. | |
55fc19a1 SG |
651 | </para> |
652 | <variablelist> | |
c464fd7e SG |
653 | <varlistentry> |
654 | <term> | |
655 | <option>lxc.console.logfile</option> | |
656 | </term> | |
657 | <listitem> | |
658 | <para> | |
659 | Specify a path to a file where the console output will | |
660 | be written. | |
661 | </para> | |
662 | </listitem> | |
663 | </varlistentry> | |
664 | <varlistentry> | |
665 | <term> | |
666 | <option>lxc.console</option> | |
667 | </term> | |
668 | <listitem> | |
669 | <para> | |
670 | Specify a path to a device to which the console will be | |
6e3bb289 CB |
671 | attached. The keyword 'none' will simply disable the |
672 | console. Note, when specifying 'none' and creating a device node | |
673 | for the console in the container at /dev/console or bind-mounting | |
674 | the hosts's /dev/console into the container at /dev/console the | |
675 | container will have direct access to the hosts's /dev/console. | |
676 | This is dangerous when the container has write access to the | |
677 | device and should thus be used with caution. | |
c464fd7e SG |
678 | </para> |
679 | </listitem> | |
680 | </varlistentry> | |
55fc19a1 SG |
681 | </variablelist> |
682 | </refsect2> | |
683 | ||
684 | <refsect2> | |
685 | <title>Console through the ttys</title> | |
686 | <para> | |
c464fd7e SG |
687 | This option is useful if the container is configured with a root |
688 | filesystem and the inittab file is setup to launch a getty on the | |
689 | ttys. The option specifies the number of ttys to be available for | |
690 | the container. The number of gettys in the inittab file of the | |
691 | container should not be greater than the number of ttys specified | |
692 | in this option, otherwise the excess getty sessions will die and | |
693 | respawn indefinitely giving annoying messages on the console or in | |
694 | <filename>/var/log/messages</filename>. | |
55fc19a1 SG |
695 | </para> |
696 | <variablelist> | |
c464fd7e SG |
697 | <varlistentry> |
698 | <term> | |
699 | <option>lxc.tty</option> | |
700 | </term> | |
701 | <listitem> | |
702 | <para> | |
703 | Specify the number of tty to make available to the | |
704 | container. | |
705 | </para> | |
706 | </listitem> | |
707 | </varlistentry> | |
55fc19a1 SG |
708 | </variablelist> |
709 | </refsect2> | |
710 | ||
711 | <refsect2> | |
712 | <title>Console devices location</title> | |
713 | <para> | |
714 | LXC consoles are provided through Unix98 PTYs created on the | |
c464fd7e SG |
715 | host and bind-mounted over the expected devices in the container. |
716 | By default, they are bind-mounted over <filename>/dev/console</filename> | |
717 | and <filename>/dev/ttyN</filename>. This can prevent package upgrades | |
718 | in the guest. Therefore you can specify a directory location (under | |
719 | <filename>/dev</filename> under which LXC will create the files and | |
720 | bind-mount over them. These will then be symbolically linked to | |
721 | <filename>/dev/console</filename> and <filename>/dev/ttyN</filename>. | |
722 | A package upgrade can then succeed as it is able to remove and replace | |
723 | the symbolic links. | |
55fc19a1 SG |
724 | </para> |
725 | <variablelist> | |
c464fd7e SG |
726 | <varlistentry> |
727 | <term> | |
728 | <option>lxc.devttydir</option> | |
729 | </term> | |
730 | <listitem> | |
731 | <para> | |
732 | Specify a directory under <filename>/dev</filename> | |
6e3bb289 CB |
733 | under which to create the container console devices. Note that LXC |
734 | will move any bind-mounts or device nodes for /dev/console into | |
735 | this directory. | |
c464fd7e SG |
736 | </para> |
737 | </listitem> | |
738 | </varlistentry> | |
55fc19a1 SG |
739 | </variablelist> |
740 | </refsect2> | |
741 | ||
742 | <refsect2> | |
743 | <title>/dev directory</title> | |
744 | <para> | |
c464fd7e SG |
745 | By default, lxc creates a few symbolic links (fd,stdin,stdout,stderr) |
746 | in the container's <filename>/dev</filename> directory but does not | |
747 | automatically create device node entries. This allows the container's | |
748 | <filename>/dev</filename> to be set up as needed in the container | |
749 | rootfs. If lxc.autodev is set to 1, then after mounting the container's | |
750 | rootfs LXC will mount a fresh tmpfs under <filename>/dev</filename> | |
c35d2909 | 751 | (limited to 500k) and fill in a minimal set of initial devices. |
55fc19a1 SG |
752 | This is generally required when starting a container containing |
753 | a "systemd" based "init" but may be optional at other times. Additional | |
754 | devices in the containers /dev directory may be created through the | |
755 | use of the <option>lxc.hook.autodev</option> hook. | |
756 | </para> | |
757 | <variablelist> | |
c464fd7e SG |
758 | <varlistentry> |
759 | <term> | |
760 | <option>lxc.autodev</option> | |
761 | </term> | |
762 | <listitem> | |
763 | <para> | |
124fa0a8 | 764 | Set this to 0 to stop LXC from mounting and populating a minimal |
c464fd7e SG |
765 | <filename>/dev</filename> when starting the container. |
766 | </para> | |
767 | </listitem> | |
768 | </varlistentry> | |
55fc19a1 SG |
769 | </variablelist> |
770 | </refsect2> | |
771 | ||
772 | <refsect2> | |
773 | <title>Enable kmsg symlink</title> | |
774 | <para> | |
d89de239 | 775 | Enable creating /dev/kmsg as symlink to /dev/console. This defaults to 0. |
55fc19a1 SG |
776 | </para> |
777 | <variablelist> | |
778 | <varlistentry> | |
779 | <term> | |
780 | <option>lxc.kmsg</option> | |
781 | </term> | |
782 | <listitem> | |
783 | <para> | |
d89de239 | 784 | Set this to 1 to enable /dev/kmsg symlinking. |
55fc19a1 SG |
785 | </para> |
786 | </listitem> | |
787 | </varlistentry> | |
788 | </variablelist> | |
789 | </refsect2> | |
790 | ||
791 | <refsect2> | |
792 | <title>Mount points</title> | |
793 | <para> | |
c464fd7e SG |
794 | The mount points section specifies the different places to be |
795 | mounted. These mount points will be private to the container | |
796 | and won't be visible by the processes running outside of the | |
797 | container. This is useful to mount /etc, /var or /home for | |
798 | examples. | |
55fc19a1 | 799 | </para> |
592fd47a SH |
800 | <para> |
801 | NOTE - LXC will generally ensure that mount targets and relative | |
802 | bind-mount sources are properly confined under the container | |
803 | root, to avoid attacks involving over-mounting host directories | |
804 | and files. (Symbolic links in absolute mount sources are ignored) | |
805 | However, if the container configuration first mounts a directory which | |
806 | is under the control of the container user, such as /home/joe, into | |
807 | the container at some <filename>path</filename>, and then mounts | |
808 | under <filename>path</filename>, then a TOCTTOU attack would be | |
809 | possible where the container user modifies a symbolic link under | |
810 | his home directory at just the right time. | |
811 | </para> | |
55fc19a1 | 812 | <variablelist> |
c464fd7e SG |
813 | <varlistentry> |
814 | <term> | |
815 | <option>lxc.mount</option> | |
816 | </term> | |
817 | <listitem> | |
818 | <para> | |
819 | specify a file location in | |
820 | the <filename>fstab</filename> format, containing the | |
821 | mount information. The mount target location can and in | |
822 | most cases should be a relative path, which will become | |
823 | relative to the mounted container root. For instance, | |
824 | </para> | |
6191f4f4 SH |
825 | <screen> |
826 | proc proc proc nodev,noexec,nosuid 0 0 | |
827 | </screen> | |
c464fd7e SG |
828 | <para> |
829 | Will mount a proc filesystem under the container's /proc, | |
830 | regardless of where the root filesystem comes from. This | |
831 | is resilient to block device backed filesystems as well as | |
832 | container cloning. | |
833 | </para> | |
834 | <para> | |
835 | Note that when mounting a filesystem from an | |
836 | image file or block device the third field (fs_vfstype) | |
837 | cannot be auto as with | |
55fc19a1 | 838 | <citerefentry> |
c464fd7e | 839 | <refentrytitle>mount</refentrytitle> |
55fc19a1 SG |
840 | <manvolnum>8</manvolnum> |
841 | </citerefentry> | |
842 | but must be explicitly specified. | |
c464fd7e SG |
843 | </para> |
844 | </listitem> | |
845 | </varlistentry> | |
846 | ||
847 | <varlistentry> | |
848 | <term> | |
849 | <option>lxc.mount.entry</option> | |
850 | </term> | |
851 | <listitem> | |
852 | <para> | |
853 | specify a mount point corresponding to a line in the | |
854 | fstab format. | |
f5b67b36 NC |
855 | |
856 | Moreover lxc add two options to mount. | |
857 | <option>optional</option> don't fail if mount does not work. | |
858 | <option>create=dir</option> or <option>create=file</option> | |
859 | to create dir (or file) when the point will be mounted. | |
c464fd7e SG |
860 | </para> |
861 | </listitem> | |
862 | </varlistentry> | |
863 | ||
864 | <varlistentry> | |
865 | <term> | |
866 | <option>lxc.mount.auto</option> | |
867 | </term> | |
868 | <listitem> | |
869 | <para> | |
870 | specify which standard kernel file systems should be | |
871 | automatically mounted. This may dramatically simplify | |
872 | the configuration. The file systems are: | |
873 | </para> | |
874 | <itemizedlist> | |
875 | <listitem> | |
876 | <para> | |
877 | <option>proc:mixed</option> (or <option>proc</option>): | |
878 | mount <filename>/proc</filename> as read-write, but | |
879 | remount <filename>/proc/sys</filename> and | |
880 | <filename>/proc/sysrq-trigger</filename> read-only | |
881 | for security / container isolation purposes. | |
882 | </para> | |
883 | </listitem> | |
884 | <listitem> | |
885 | <para> | |
886 | <option>proc:rw</option>: mount | |
887 | <filename>/proc</filename> as read-write | |
888 | </para> | |
889 | </listitem> | |
890 | <listitem> | |
891 | <para> | |
f24a52d5 SG |
892 | <option>sys:mixed</option> (or <option>sys</option>): |
893 | mount <filename>/sys</filename> as read-only but with | |
894 | /sys/devices/virtual/net writable. | |
895 | </para> | |
896 | </listitem> | |
897 | <listitem> | |
898 | <para> | |
899 | <option>sys:ro</option>: | |
c464fd7e SG |
900 | mount <filename>/sys</filename> as read-only |
901 | for security / container isolation purposes. | |
902 | </para> | |
903 | </listitem> | |
904 | <listitem> | |
905 | <para> | |
906 | <option>sys:rw</option>: mount | |
907 | <filename>/sys</filename> as read-write | |
908 | </para> | |
909 | </listitem> | |
910 | <listitem> | |
911 | <para> | |
912 | <option>cgroup:mixed</option>: | |
913 | mount a tmpfs to <filename>/sys/fs/cgroup</filename>, | |
914 | create directories for all hierarchies to which | |
915 | the container is added, create subdirectories | |
916 | there with the name of the cgroup, and bind-mount | |
917 | the container's own cgroup into that directory. | |
918 | The container will be able to write to its own | |
919 | cgroup directory, but not the parents, since they | |
4608594e | 920 | will be remounted read-only. |
c464fd7e SG |
921 | </para> |
922 | </listitem> | |
923 | <listitem> | |
924 | <para> | |
925 | <option>cgroup:ro</option>: similar to | |
926 | <option>cgroup:mixed</option>, but everything will | |
927 | be mounted read-only. | |
928 | </para> | |
929 | </listitem> | |
930 | <listitem> | |
931 | <para> | |
932 | <option>cgroup:rw</option>: similar to | |
933 | <option>cgroup:mixed</option>, but everything will | |
934 | be mounted read-write. Note that the paths leading | |
935 | up to the container's own cgroup will be writable, | |
936 | but will not be a cgroup filesystem but just part | |
937 | of the tmpfs of <filename>/sys/fs/cgroup</filename> | |
938 | </para> | |
939 | </listitem> | |
940 | <listitem> | |
941 | <para> | |
942 | <option>cgroup</option> (without specifier): | |
943 | defaults to <option>cgroup:rw</option> if the | |
944 | container retains the CAP_SYS_ADMIN capability, | |
945 | <option>cgroup:mixed</option> otherwise. | |
946 | </para> | |
947 | </listitem> | |
948 | <listitem> | |
949 | <para> | |
950 | <option>cgroup-full:mixed</option>: | |
951 | mount a tmpfs to <filename>/sys/fs/cgroup</filename>, | |
952 | create directories for all hierarchies to which | |
953 | the container is added, bind-mount the hierarchies | |
954 | from the host to the container and make everything | |
955 | read-only except the container's own cgroup. Note | |
956 | that compared to <option>cgroup</option>, where | |
957 | all paths leading up to the container's own cgroup | |
958 | are just simple directories in the underlying | |
959 | tmpfs, here | |
960 | <filename>/sys/fs/cgroup/$hierarchy</filename> | |
961 | will contain the host's full cgroup hierarchy, | |
962 | albeit read-only outside the container's own cgroup. | |
963 | This may leak quite a bit of information into the | |
964 | container. | |
965 | </para> | |
966 | </listitem> | |
967 | <listitem> | |
968 | <para> | |
969 | <option>cgroup-full:ro</option>: similar to | |
970 | <option>cgroup-full:mixed</option>, but everything | |
971 | will be mounted read-only. | |
972 | </para> | |
973 | </listitem> | |
974 | <listitem> | |
975 | <para> | |
976 | <option>cgroup-full:rw</option>: similar to | |
977 | <option>cgroup-full:mixed</option>, but everything | |
978 | will be mounted read-write. Note that in this case, | |
979 | the container may escape its own cgroup. (Note also | |
980 | that if the container has CAP_SYS_ADMIN support | |
981 | and can mount the cgroup filesystem itself, it may | |
982 | do so anyway.) | |
983 | </para> | |
984 | </listitem> | |
985 | <listitem> | |
986 | <para> | |
987 | <option>cgroup-full</option> (without specifier): | |
988 | defaults to <option>cgroup-full:rw</option> if the | |
989 | container retains the CAP_SYS_ADMIN capability, | |
990 | <option>cgroup-full:mixed</option> otherwise. | |
991 | </para> | |
992 | </listitem> | |
993 | </itemizedlist> | |
4608594e SH |
994 | <para> |
995 | If cgroup namespaces are enabled, then any <option>cgroup</option> | |
996 | auto-mounting request will be ignored, since the container can | |
997 | mount the filesystems itself, and automounting can confuse the | |
998 | container init. | |
999 | </para> | |
c464fd7e SG |
1000 | <para> |
1001 | Note that if automatic mounting of the cgroup filesystem | |
1002 | is enabled, the tmpfs under | |
1003 | <filename>/sys/fs/cgroup</filename> will always be | |
1004 | mounted read-write (but for the <option>:mixed</option> | |
1005 | and <option>:ro</option> cases, the individual | |
1006 | hierarchies, | |
1007 | <filename>/sys/fs/cgroup/$hierarchy</filename>, will be | |
1008 | read-only). This is in order to work around a quirk in | |
1009 | Ubuntu's | |
b46f0553 | 1010 | <citerefentry> |
c464fd7e | 1011 | <refentrytitle>mountall</refentrytitle> |
b46f0553 CS |
1012 | <manvolnum>8</manvolnum> |
1013 | </citerefentry> | |
c464fd7e SG |
1014 | command that will cause containers to wait for user |
1015 | input at boot if | |
1016 | <filename>/sys/fs/cgroup</filename> is mounted read-only | |
1017 | and the container can't remount it read-write due to a | |
1018 | lack of CAP_SYS_ADMIN. | |
1019 | </para> | |
1020 | <para> | |
1021 | Examples: | |
1022 | </para> | |
1023 | <programlisting> | |
1024 | lxc.mount.auto = proc sys cgroup | |
1025 | lxc.mount.auto = proc:rw sys:rw cgroup-full:rw | |
1026 | </programlisting> | |
1027 | </listitem> | |
1028 | </varlistentry> | |
55fc19a1 SG |
1029 | |
1030 | </variablelist> | |
1031 | </refsect2> | |
1032 | ||
1033 | <refsect2> | |
1034 | <title>Root file system</title> | |
1035 | <para> | |
c464fd7e SG |
1036 | The root file system of the container can be different than that |
1037 | of the host system. | |
55fc19a1 SG |
1038 | </para> |
1039 | <variablelist> | |
c464fd7e SG |
1040 | <varlistentry> |
1041 | <term> | |
1042 | <option>lxc.rootfs</option> | |
1043 | </term> | |
1044 | <listitem> | |
1045 | <para> | |
1046 | specify the root file system for the container. It can | |
1047 | be an image file, a directory or a block device. If not | |
1048 | specified, the container shares its root file system | |
1049 | with the host. | |
1050 | </para> | |
1051 | <para> | |
f1c26f2c SH |
1052 | For directory or simple block-device backed containers, |
1053 | a pathname can be used. If the rootfs is backed by a nbd | |
1054 | device, then <filename>nbd:file:1</filename> specifies that | |
1055 | <filename>file</filename> should be attached to a nbd device, | |
1056 | and partition 1 should be mounted as the rootfs. | |
1057 | <filename>nbd:file</filename> specifies that the nbd device | |
1058 | itself should be mounted. <filename>overlayfs:/lower:/upper</filename> | |
1059 | specifies that the rootfs should be an overlay with <filename>/upper</filename> | |
1060 | being mounted read-write over a read-only mount of <filename>/lower</filename>. | |
1061 | <filename>aufs:/lower:/upper</filename> does the same using aufs in place | |
280d2379 CB |
1062 | of overlayfs. For both <filename>overlayfs</filename> and |
1063 | <filename>aufs</filename> multiple <filename>/lower</filename> | |
1064 | directories can be specified. <filename>loop:/file</filename> tells lxc to attach | |
f1c26f2c | 1065 | <filename>/file</filename> to a loop device and mount the loop device. |
c464fd7e SG |
1066 | </para> |
1067 | </listitem> | |
1068 | </varlistentry> | |
1069 | ||
1070 | <varlistentry> | |
1071 | <term> | |
1072 | <option>lxc.rootfs.mount</option> | |
1073 | </term> | |
1074 | <listitem> | |
1075 | <para> | |
1076 | where to recursively bind <option>lxc.rootfs</option> | |
1077 | before pivoting. This is to ensure success of the | |
1078 | <citerefentry> | |
1079 | <refentrytitle><command>pivot_root</command></refentrytitle> | |
1080 | <manvolnum>8</manvolnum> | |
1081 | </citerefentry> | |
1082 | syscall. Any directory suffices, the default should | |
1083 | generally work. | |
1084 | </para> | |
1085 | </listitem> | |
1086 | </varlistentry> | |
1087 | ||
1088 | <varlistentry> | |
1089 | <term> | |
1090 | <option>lxc.rootfs.options</option> | |
1091 | </term> | |
1092 | <listitem> | |
1093 | <para> | |
1094 | extra mount options to use when mounting the rootfs. | |
1095 | </para> | |
1096 | </listitem> | |
1097 | </varlistentry> | |
a17b1e65 | 1098 | |
f9039861 SH |
1099 | <varlistentry> |
1100 | <term> | |
1101 | <option>lxc.rootfs.backend</option> | |
1102 | </term> | |
1103 | <listitem> | |
1104 | <para> | |
1105 | specify the rootfs backend type to use, for instance 'dir' or | |
1106 | 'zfs'. While this can be guessed by lxc at container startup, | |
1107 | doing so takes time. Specifying it here avoids extra | |
1108 | processing. | |
1109 | </para> | |
1110 | </listitem> | |
1111 | </varlistentry> | |
1112 | ||
55fc19a1 SG |
1113 | </variablelist> |
1114 | </refsect2> | |
1115 | ||
1116 | <refsect2> | |
1117 | <title>Control group</title> | |
1118 | <para> | |
c464fd7e SG |
1119 | The control group section contains the configuration for the |
1120 | different subsystem. <command>lxc</command> does not check the | |
1121 | correctness of the subsystem name. This has the disadvantage | |
1122 | of not detecting configuration errors until the container is | |
1123 | started, but has the advantage of permitting any future | |
1124 | subsystem. | |
55fc19a1 SG |
1125 | </para> |
1126 | <variablelist> | |
c464fd7e SG |
1127 | <varlistentry> |
1128 | <term> | |
1129 | <option>lxc.cgroup.[subsystem name]</option> | |
1130 | </term> | |
1131 | <listitem> | |
1132 | <para> | |
1133 | specify the control group value to be set. The | |
1134 | subsystem name is the literal name of the control group | |
1135 | subsystem. The permitted names and the syntax of their | |
1136 | values is not dictated by LXC, instead it depends on the | |
1137 | features of the Linux kernel running at the time the | |
1138 | container is started, | |
1139 | eg. <option>lxc.cgroup.cpuset.cpus</option> | |
1140 | </para> | |
1141 | </listitem> | |
1142 | </varlistentry> | |
55fc19a1 SG |
1143 | </variablelist> |
1144 | </refsect2> | |
1145 | ||
1146 | <refsect2> | |
1147 | <title>Capabilities</title> | |
1148 | <para> | |
c464fd7e SG |
1149 | The capabilities can be dropped in the container if this one |
1150 | is run as root. | |
55fc19a1 SG |
1151 | </para> |
1152 | <variablelist> | |
c464fd7e SG |
1153 | <varlistentry> |
1154 | <term> | |
1155 | <option>lxc.cap.drop</option> | |
1156 | </term> | |
1157 | <listitem> | |
1158 | <para> | |
1159 | Specify the capability to be dropped in the container. A | |
1160 | single line defining several capabilities with a space | |
1161 | separation is allowed. The format is the lower case of | |
1162 | the capability definition without the "CAP_" prefix, | |
1163 | eg. CAP_SYS_MODULE should be specified as | |
1164 | sys_module. See | |
1165 | <citerefentry> | |
1166 | <refentrytitle><command>capabilities</command></refentrytitle> | |
1167 | <manvolnum>7</manvolnum> | |
7eff30fd MH |
1168 | </citerefentry>. |
1169 | If used with no value, lxc will clear any drop capabilities | |
1170 | specified up to this point. | |
c464fd7e SG |
1171 | </para> |
1172 | </listitem> | |
1173 | </varlistentry> | |
1174 | <varlistentry> | |
1175 | <term> | |
1176 | <option>lxc.cap.keep</option> | |
1177 | </term> | |
1178 | <listitem> | |
1179 | <para> | |
1180 | Specify the capability to be kept in the container. All other | |
1181 | capabilities will be dropped. When a special value of "none" is | |
1182 | encountered, lxc will clear any keep capabilities specified up | |
1183 | to this point. A value of "none" alone can be used to drop all | |
1184 | capabilities. | |
1185 | </para> | |
1186 | </listitem> | |
1187 | </varlistentry> | |
55fc19a1 SG |
1188 | </variablelist> |
1189 | </refsect2> | |
1190 | ||
93f9e90d WB |
1191 | <refsect2> |
1192 | <title>Resource limits</title> | |
1193 | <para> | |
1194 | The soft and hard resource limits for the container can be changed. | |
1195 | Unprivileged containers can only lower them. Resources which are not | |
1196 | explicitly specified will be inherited. | |
1197 | </para> | |
1198 | <variablelist> | |
1199 | <varlistentry> | |
1200 | <term> | |
1201 | <option>lxc.limit.[limit name]</option> | |
1202 | </term> | |
1203 | <listitem> | |
1204 | <para> | |
1205 | Specify the resource limit to be set. A limit is specified as two | |
1206 | colon separated values which are either numeric or the word | |
1207 | 'unlimited'. A single value can be used as a shortcut to set both | |
1208 | soft and hard limit to the same value. The permitted names the | |
1209 | "RLIMIT_" resource names in lowercase without the "RLIMIT_" | |
1210 | prefix, eg. RLIMIT_NOFILE should be specified as "nofile". See | |
1211 | <citerefentry> | |
1212 | <refentrytitle><command>setrlimit</command></refentrytitle> | |
1213 | <manvolnum>2</manvolnum> | |
1214 | </citerefentry>. | |
1215 | If used with no value, lxc will clear the resource limit | |
1216 | specified up to this point. A resource with no explicitly | |
1217 | configured limitation will be inherited from the process starting | |
1218 | up the container. | |
1219 | </para> | |
1220 | </listitem> | |
1221 | </varlistentry> | |
1222 | </variablelist> | |
1223 | </refsect2> | |
1224 | ||
55fc19a1 SG |
1225 | <refsect2> |
1226 | <title>Apparmor profile</title> | |
1227 | <para> | |
c464fd7e SG |
1228 | If lxc was compiled and installed with apparmor support, and the host |
1229 | system has apparmor enabled, then the apparmor profile under which the | |
1230 | container should be run can be specified in the container | |
7a126ae1 SH |
1231 | configuration. The default is <command>lxc-container-default-cgns</command> |
1232 | if the host kernel is cgroup namespace aware, or | |
1233 | <command>lxc-container-default</command> othewise. | |
55fc19a1 SG |
1234 | </para> |
1235 | <variablelist> | |
c464fd7e SG |
1236 | <varlistentry> |
1237 | <term> | |
1238 | <option>lxc.aa_profile</option> | |
1239 | </term> | |
1240 | <listitem> | |
1241 | <para> | |
1242 | Specify the apparmor profile under which the container should | |
1243 | be run. To specify that the container should be unconfined, | |
1244 | use | |
1245 | </para> | |
1246 | <programlisting>lxc.aa_profile = unconfined</programlisting> | |
7a126ae1 SH |
1247 | <para> |
1248 | If the apparmor profile should remain unchanged (i.e. if you | |
1249 | are nesting containers and are already confined), then use | |
1250 | </para> | |
1251 | <programlisting>lxc.aa_profile = unchanged</programlisting> | |
c464fd7e SG |
1252 | </listitem> |
1253 | </varlistentry> | |
1254 | <varlistentry> | |
1255 | <term> | |
1256 | <option>lxc.aa_allow_incomplete</option> | |
1257 | </term> | |
1258 | <listitem> | |
1259 | <para> | |
1260 | Apparmor profiles are pathname based. Therefore many file | |
1261 | restrictions require mount restrictions to be effective against | |
1262 | a determined attacker. However, these mount restrictions are not | |
1263 | yet implemented in the upstream kernel. Without the mount | |
1264 | restrictions, the apparmor profiles still protect against accidental | |
1265 | damager. | |
1266 | </para> | |
1267 | <para> | |
1268 | If this flag is 0 (default), then the container will not be | |
1269 | started if the kernel lacks the apparmor mount features, so that a | |
1270 | regression after a kernel upgrade will be detected. To start the | |
1271 | container under partial apparmor protection, set this flag to 1. | |
1272 | </para> | |
1273 | </listitem> | |
1274 | </varlistentry> | |
55fc19a1 SG |
1275 | </variablelist> |
1276 | </refsect2> | |
1277 | ||
1278 | <refsect2> | |
1279 | <title>SELinux context</title> | |
1280 | <para> | |
c464fd7e SG |
1281 | If lxc was compiled and installed with SELinux support, and the host |
1282 | system has SELinux enabled, then the SELinux context under which the | |
1283 | container should be run can be specified in the container | |
1284 | configuration. The default is <command>unconfined_t</command>, | |
1285 | which means that lxc will not attempt to change contexts. | |
1286 | See @DATADIR@/lxc/selinux/lxc.te for an example policy and more | |
1287 | information. | |
55fc19a1 SG |
1288 | </para> |
1289 | <variablelist> | |
c464fd7e SG |
1290 | <varlistentry> |
1291 | <term> | |
1292 | <option>lxc.se_context</option> | |
1293 | </term> | |
1294 | <listitem> | |
1295 | <para> | |
1296 | Specify the SELinux context under which the container should | |
1297 | be run or <command>unconfined_t</command>. For example | |
1298 | </para> | |
1299 | <programlisting>lxc.se_context = system_u:system_r:lxc_t:s0:c22</programlisting> | |
1300 | </listitem> | |
1301 | </varlistentry> | |
55fc19a1 SG |
1302 | </variablelist> |
1303 | </refsect2> | |
1304 | ||
1305 | <refsect2> | |
1306 | <title>Seccomp configuration</title> | |
1307 | <para> | |
1308 | A container can be started with a reduced set of available | |
c464fd7e SG |
1309 | system calls by loading a seccomp profile at startup. The |
1310 | seccomp configuration file must begin with a version number | |
1311 | on the first line, a policy type on the second line, followed | |
1312 | by the configuration. | |
55fc19a1 | 1313 | </para> |
a7c27357 SH |
1314 | <para> |
1315 | Versions 1 and 2 are currently supported. In version 1, the | |
c464fd7e SG |
1316 | policy is a simple whitelist. The second line therefore must |
1317 | read "whitelist", with the rest of the file containing one (numeric) | |
1318 | sycall number per line. Each syscall number is whitelisted, | |
1319 | while every unlisted number is blacklisted for use in the container | |
a7c27357 SH |
1320 | </para> |
1321 | ||
1322 | <para> | |
1323 | In version 2, the policy may be blacklist or whitelist, | |
1324 | supports per-rule and per-policy default actions, and supports | |
1325 | per-architecture system call resolution from textual names. | |
1326 | </para> | |
1327 | <para> | |
1328 | An example blacklist policy, in which all system calls are | |
1329 | allowed except for mknod, which will simply do nothing and | |
1330 | return 0 (success), looks like: | |
1331 | </para> | |
1332 | <screen> | |
1333 | 2 | |
1334 | blacklist | |
1335 | mknod errno 0 | |
1336 | </screen> | |
55fc19a1 | 1337 | <variablelist> |
c464fd7e SG |
1338 | <varlistentry> |
1339 | <term> | |
1340 | <option>lxc.seccomp</option> | |
1341 | </term> | |
1342 | <listitem> | |
1343 | <para> | |
1344 | Specify a file containing the seccomp configuration to | |
1345 | load before the container starts. | |
1346 | </para> | |
1347 | </listitem> | |
1348 | </varlistentry> | |
55fc19a1 SG |
1349 | </variablelist> |
1350 | </refsect2> | |
1351 | ||
222ddc91 CB |
1352 | <refsect2> |
1353 | <title>PR_SET_NO_NEW_PRIVS</title> | |
1354 | <para> | |
1355 | With PR_SET_NO_NEW_PRIVS active execve() promises not to grant | |
1356 | privileges to do anything that could not have been done without | |
1357 | the execve() call (for example, rendering the set-user-ID and | |
1358 | set-group-ID mode bits, and file capabilities non-functional). | |
1359 | Once set, this bit cannot be unset. The setting of this bit is | |
1360 | inherited by children created by fork() and clone(), and preserved | |
1361 | across execve(). | |
1362 | Note that PR_SET_NO_NEW_PRIVS is applied after the container has | |
1363 | changed into its intended AppArmor profile or SElinux context. | |
1364 | </para> | |
1365 | <variablelist> | |
1366 | <varlistentry> | |
1367 | <term> | |
1368 | <option>lxc.no_new_privs</option> | |
1369 | </term> | |
1370 | <listitem> | |
1371 | <para> | |
1372 | Specify whether the PR_SET_NO_NEW_PRIVS flag should be set for the | |
1373 | container. Set to 1 to activate. | |
1374 | </para> | |
1375 | </listitem> | |
1376 | </varlistentry> | |
1377 | </variablelist> | |
1378 | </refsect2> | |
1379 | ||
55fc19a1 SG |
1380 | <refsect2> |
1381 | <title>UID mappings</title> | |
1382 | <para> | |
1383 | A container can be started in a private user namespace with | |
c464fd7e SG |
1384 | user and group id mappings. For instance, you can map userid |
1385 | 0 in the container to userid 200000 on the host. The root | |
1386 | user in the container will be privileged in the container, | |
1387 | but unprivileged on the host. Normally a system container | |
1388 | will want a range of ids, so you would map, for instance, | |
1389 | user and group ids 0 through 20,000 in the container to the | |
1390 | ids 200,000 through 220,000. | |
55fc19a1 SG |
1391 | </para> |
1392 | <variablelist> | |
c464fd7e SG |
1393 | <varlistentry> |
1394 | <term> | |
1395 | <option>lxc.id_map</option> | |
1396 | </term> | |
1397 | <listitem> | |
1398 | <para> | |
1399 | Four values must be provided. First a character, either | |
1400 | 'u', or 'g', to specify whether user or group ids are | |
1401 | being mapped. Next is the first userid as seen in the | |
1402 | user namespace of the container. Next is the userid as | |
1403 | seen on the host. Finally, a range indicating the number | |
1404 | of consecutive ids to map. | |
1405 | </para> | |
1406 | </listitem> | |
1407 | </varlistentry> | |
55fc19a1 SG |
1408 | </variablelist> |
1409 | </refsect2> | |
1410 | ||
1411 | <refsect2> | |
1412 | <title>Container hooks</title> | |
1413 | <para> | |
1414 | Container hooks are programs or scripts which can be executed | |
c464fd7e | 1415 | at various times in a container's lifetime. |
55fc19a1 SG |
1416 | </para> |
1417 | <para> | |
1418 | When a container hook is executed, information is passed both | |
c464fd7e SG |
1419 | as command line arguments and through environment variables. |
1420 | The arguments are: | |
1421 | <itemizedlist> | |
1422 | <listitem><para> Container name. </para></listitem> | |
1423 | <listitem><para> Section (always 'lxc'). </para></listitem> | |
1424 | <listitem><para> The hook type (i.e. 'clone' or 'pre-mount'). </para></listitem> | |
0a2b5ab1 | 1425 | <listitem><para> Additional arguments. In the |
c464fd7e | 1426 | case of the clone hook, any extra arguments passed to |
0a2b5ab1 WB |
1427 | lxc-clone will appear as further arguments to the hook. |
1428 | In the case of the stop hook, paths to filedescriptors | |
1429 | for each of the container's namespaces along with their types | |
1430 | are passed. </para></listitem> | |
c464fd7e SG |
1431 | </itemizedlist> |
1432 | The following environment variables are set: | |
1433 | <itemizedlist> | |
1434 | <listitem><para> LXC_NAME: is the container's name. </para></listitem> | |
1435 | <listitem><para> LXC_ROOTFS_MOUNT: the path to the mounted root filesystem. </para></listitem> | |
1436 | <listitem><para> LXC_CONFIG_FILE: the path to the container configuration file. </para></listitem> | |
1437 | <listitem><para> LXC_SRC_NAME: in the case of the clone hook, this is the original container's name. </para></listitem> | |
1438 | <listitem><para> LXC_ROOTFS_PATH: this is the lxc.rootfs entry for the container. Note this is likely not where the mounted rootfs is to be found, use LXC_ROOTFS_MOUNT for that. </para></listitem> | |
1439 | </itemizedlist> | |
55fc19a1 SG |
1440 | </para> |
1441 | <para> | |
1442 | Standard output from the hooks is logged at debug level. | |
1443 | Standard error is not logged, but can be captured by the | |
1444 | hook redirecting its standard error to standard output. | |
1445 | </para> | |
1446 | <variablelist> | |
c464fd7e SG |
1447 | <varlistentry> |
1448 | <term> | |
1449 | <option>lxc.hook.pre-start</option> | |
1450 | </term> | |
1451 | <listitem> | |
1452 | <para> | |
1453 | A hook to be run in the host's namespace before the | |
1454 | container ttys, consoles, or mounts are up. | |
1455 | </para> | |
1456 | </listitem> | |
1457 | </varlistentry> | |
55fc19a1 SG |
1458 | </variablelist> |
1459 | <variablelist> | |
c464fd7e SG |
1460 | <varlistentry> |
1461 | <term> | |
1462 | <option>lxc.hook.pre-mount</option> | |
1463 | </term> | |
1464 | <listitem> | |
1465 | <para> | |
1466 | A hook to be run in the container's fs namespace but before | |
1467 | the rootfs has been set up. This allows for manipulation | |
1468 | of the rootfs, i.e. to mount an encrypted filesystem. Mounts | |
1469 | done in this hook will not be reflected on the host (apart from | |
1470 | mounts propagation), so they will be automatically cleaned up | |
1471 | when the container shuts down. | |
1472 | </para> | |
1473 | </listitem> | |
1474 | </varlistentry> | |
55fc19a1 SG |
1475 | </variablelist> |
1476 | <variablelist> | |
c464fd7e SG |
1477 | <varlistentry> |
1478 | <term> | |
1479 | <option>lxc.hook.mount</option> | |
1480 | </term> | |
1481 | <listitem> | |
1482 | <para> | |
1483 | A hook to be run in the container's namespace after | |
1484 | mounting has been done, but before the pivot_root. | |
1485 | </para> | |
1486 | </listitem> | |
1487 | </varlistentry> | |
55fc19a1 SG |
1488 | </variablelist> |
1489 | <variablelist> | |
c464fd7e SG |
1490 | <varlistentry> |
1491 | <term> | |
1492 | <option>lxc.hook.autodev</option> | |
1493 | </term> | |
1494 | <listitem> | |
1495 | <para> | |
1496 | A hook to be run in the container's namespace after | |
1497 | mounting has been done and after any mount hooks have | |
1498 | run, but before the pivot_root, if | |
1499 | <option>lxc.autodev</option> == 1. | |
1500 | The purpose of this hook is to assist in populating the | |
1501 | /dev directory of the container when using the autodev | |
1502 | option for systemd based containers. The container's /dev | |
1503 | directory is relative to the | |
1504 | ${<option>LXC_ROOTFS_MOUNT</option>} environment | |
1505 | variable available when the hook is run. | |
1506 | </para> | |
1507 | </listitem> | |
1508 | </varlistentry> | |
55fc19a1 SG |
1509 | </variablelist> |
1510 | <variablelist> | |
c464fd7e SG |
1511 | <varlistentry> |
1512 | <term> | |
1513 | <option>lxc.hook.start</option> | |
1514 | </term> | |
1515 | <listitem> | |
1516 | <para> | |
1517 | A hook to be run in the container's namespace immediately | |
1518 | before executing the container's init. This requires the | |
1519 | program to be available in the container. | |
1520 | </para> | |
1521 | </listitem> | |
1522 | </varlistentry> | |
55fc19a1 | 1523 | </variablelist> |
0a2b5ab1 WB |
1524 | <variablelist> |
1525 | <varlistentry> | |
1526 | <term> | |
1527 | <option>lxc.hook.stop</option> | |
1528 | </term> | |
1529 | <listitem> | |
1530 | <para> | |
1531 | A hook to be run in the host's namespace with references | |
1532 | to the container's namespaces after the container has been shut | |
1533 | down. For each namespace an extra argument is passed to the hook | |
1534 | containing the namespace's type and a filename that can be used to | |
1535 | obtain a file descriptor to the corresponding namespace, separated | |
1536 | by a colon. The type is the name as it would appear in the | |
1537 | <filename>/proc/PID/ns</filename> directory. | |
1538 | For instance for the mount namespace the argument usually looks | |
1539 | like <filename>mnt:/proc/PID/fd/12</filename>. | |
1540 | </para> | |
1541 | </listitem> | |
1542 | </varlistentry> | |
1543 | </variablelist> | |
55fc19a1 | 1544 | <variablelist> |
c464fd7e SG |
1545 | <varlistentry> |
1546 | <term> | |
1547 | <option>lxc.hook.post-stop</option> | |
1548 | </term> | |
1549 | <listitem> | |
1550 | <para> | |
1551 | A hook to be run in the host's namespace after the | |
1552 | container has been shut down. | |
1553 | </para> | |
1554 | </listitem> | |
1555 | </varlistentry> | |
55fc19a1 SG |
1556 | </variablelist> |
1557 | <variablelist> | |
c464fd7e SG |
1558 | <varlistentry> |
1559 | <term> | |
1560 | <option>lxc.hook.clone</option> | |
1561 | </term> | |
1562 | <listitem> | |
1563 | <para> | |
1564 | A hook to be run when the container is cloned to a new one. | |
1565 | See <citerefentry><refentrytitle><command>lxc-clone</command></refentrytitle> | |
1566 | <manvolnum>1</manvolnum></citerefentry> for more information. | |
1567 | </para> | |
1568 | </listitem> | |
1569 | </varlistentry> | |
55fc19a1 | 1570 | </variablelist> |
37cf711b SY |
1571 | <variablelist> |
1572 | <varlistentry> | |
1573 | <term> | |
1574 | <option>lxc.hook.destroy</option> | |
1575 | </term> | |
1576 | <listitem> | |
1577 | <para> | |
1578 | A hook to be run when the container is destroyed. | |
1579 | </para> | |
1580 | </listitem> | |
1581 | </varlistentry> | |
1582 | </variablelist> | |
55fc19a1 SG |
1583 | </refsect2> |
1584 | ||
1585 | <refsect2> | |
1586 | <title>Container hooks Environment Variables</title> | |
1587 | <para> | |
1588 | A number of environment variables are made available to the startup | |
1589 | hooks to provide configuration information and assist in the | |
1590 | functioning of the hooks. Not all variables are valid in all | |
1591 | contexts. In particular, all paths are relative to the host system | |
1592 | and, as such, not valid during the <option>lxc.hook.start</option> hook. | |
1593 | </para> | |
1594 | <variablelist> | |
c464fd7e SG |
1595 | <varlistentry> |
1596 | <term> | |
1597 | <option>LXC_NAME</option> | |
1598 | </term> | |
1599 | <listitem> | |
1600 | <para> | |
1601 | The LXC name of the container. Useful for logging messages | |
1602 | in common log environments. [<option>-n</option>] | |
1603 | </para> | |
1604 | </listitem> | |
1605 | </varlistentry> | |
55fc19a1 SG |
1606 | </variablelist> |
1607 | <variablelist> | |
c464fd7e SG |
1608 | <varlistentry> |
1609 | <term> | |
1610 | <option>LXC_CONFIG_FILE</option> | |
1611 | </term> | |
1612 | <listitem> | |
1613 | <para> | |
1614 | Host relative path to the container configuration file. This | |
1615 | gives the container to reference the original, top level, | |
1616 | configuration file for the container in order to locate any | |
1617 | additional configuration information not otherwise made | |
1618 | available. [<option>-f</option>] | |
1619 | </para> | |
1620 | </listitem> | |
1621 | </varlistentry> | |
55fc19a1 SG |
1622 | </variablelist> |
1623 | <variablelist> | |
c464fd7e SG |
1624 | <varlistentry> |
1625 | <term> | |
1626 | <option>LXC_CONSOLE</option> | |
1627 | </term> | |
1628 | <listitem> | |
1629 | <para> | |
1630 | The path to the console output of the container if not NULL. | |
1631 | [<option>-c</option>] [<option>lxc.console</option>] | |
1632 | </para> | |
1633 | </listitem> | |
1634 | </varlistentry> | |
55fc19a1 SG |
1635 | </variablelist> |
1636 | <variablelist> | |
c464fd7e SG |
1637 | <varlistentry> |
1638 | <term> | |
1639 | <option>LXC_CONSOLE_LOGPATH</option> | |
1640 | </term> | |
1641 | <listitem> | |
1642 | <para> | |
1643 | The path to the console log output of the container if not NULL. | |
1644 | [<option>-L</option>] | |
1645 | </para> | |
1646 | </listitem> | |
1647 | </varlistentry> | |
55fc19a1 SG |
1648 | </variablelist> |
1649 | <variablelist> | |
c464fd7e SG |
1650 | <varlistentry> |
1651 | <term> | |
1652 | <option>LXC_ROOTFS_MOUNT</option> | |
1653 | </term> | |
1654 | <listitem> | |
1655 | <para> | |
1656 | The mount location to which the container is initially bound. | |
1657 | This will be the host relative path to the container rootfs | |
1658 | for the container instance being started and is where changes | |
1659 | should be made for that instance. | |
1660 | [<option>lxc.rootfs.mount</option>] | |
1661 | </para> | |
1662 | </listitem> | |
1663 | </varlistentry> | |
55fc19a1 SG |
1664 | </variablelist> |
1665 | <variablelist> | |
c464fd7e SG |
1666 | <varlistentry> |
1667 | <term> | |
1668 | <option>LXC_ROOTFS_PATH</option> | |
1669 | </term> | |
1670 | <listitem> | |
1671 | <para> | |
1672 | The host relative path to the container root which has been | |
1673 | mounted to the rootfs.mount location. | |
1674 | [<option>lxc.rootfs</option>] | |
1675 | </para> | |
1676 | </listitem> | |
1677 | </varlistentry> | |
55fc19a1 | 1678 | </variablelist> |
07945418 KY |
1679 | <variablelist> |
1680 | <varlistentry> | |
1681 | <term> | |
1682 | <option>LXC_SRC_NAME</option> | |
1683 | </term> | |
1684 | <listitem> | |
1685 | <para> | |
1686 | Only for the clone hook. Is set to the original container name. | |
1687 | </para> | |
1688 | </listitem> | |
1689 | </varlistentry> | |
1690 | </variablelist> | |
c154af98 SG |
1691 | <variablelist> |
1692 | <varlistentry> | |
1693 | <term> | |
1694 | <option>LXC_TARGET</option> | |
1695 | </term> | |
1696 | <listitem> | |
1697 | <para> | |
1698 | Only for the stop hook. Is set to "stop" for a container | |
1699 | shutdown or "reboot" for a container reboot. | |
1700 | </para> | |
1701 | </listitem> | |
1702 | </varlistentry> | |
c4cafa08 SH |
1703 | </variablelist> |
1704 | <variablelist> | |
1705 | <varlistentry> | |
1706 | <term> | |
1707 | <option>LXC_CGNS_AWARE</option> | |
1708 | </term> | |
1709 | <listitem> | |
1710 | <para> | |
1711 | If unset, then this version of lxc is not aware of cgroup | |
1712 | namespaces. If set, it will be set to 1, and lxc is aware | |
1713 | of cgroup namespaces. Note this does not guarantee that | |
1714 | cgroup namespaces are enabled in the kernel. This is used | |
1715 | by the lxcfs mount hook. | |
1716 | </para> | |
1717 | </listitem> | |
1718 | </varlistentry> | |
c154af98 | 1719 | </variablelist> |
55fc19a1 SG |
1720 | </refsect2> |
1721 | <refsect2> | |
1722 | <title>Logging</title> | |
1723 | <para> | |
1724 | Logging can be configured on a per-container basis. By default, | |
1725 | depending upon how the lxc package was compiled, container startup | |
1726 | is logged only at the ERROR level, and logged to a file named after | |
1727 | the container (with '.log' appended) either under the container path, | |
1728 | or under @LOGPATH@. | |
1729 | </para> | |
1730 | <para> | |
1731 | Both the default log level and the log file can be specified in the | |
1732 | container configuration file, overriding the default behavior. Note | |
1733 | that the configuration file entries can in turn be overridden by the | |
1734 | command line options to <command>lxc-start</command>. | |
1735 | </para> | |
1736 | <variablelist> | |
c464fd7e SG |
1737 | <varlistentry> |
1738 | <term> | |
1739 | <option>lxc.loglevel</option> | |
1740 | </term> | |
1741 | <listitem> | |
1742 | <para> | |
1743 | The level at which to log. The log level is an integer in | |
1744 | the range of 0..8 inclusive, where a lower number means more | |
1745 | verbose debugging. In particular 0 = trace, 1 = debug, 2 = | |
1746 | info, 3 = notice, 4 = warn, 5 = error, 6 = critical, 7 = | |
1747 | alert, and 8 = fatal. If unspecified, the level defaults | |
1748 | to 5 (error), so that only errors and above are logged. | |
1749 | </para> | |
1750 | <para> | |
1751 | Note that when a script (such as either a hook script or a | |
1752 | network interface up or down script) is called, the script's | |
1753 | standard output is logged at level 1, debug. | |
1754 | </para> | |
1755 | </listitem> | |
1756 | </varlistentry> | |
1757 | <varlistentry> | |
1758 | <term> | |
1759 | <option>lxc.logfile</option> | |
1760 | </term> | |
1761 | <listitem> | |
1762 | <para> | |
1763 | The file to which logging info should be written. | |
1764 | </para> | |
1765 | </listitem> | |
1766 | </varlistentry> | |
204dfdf2 BD |
1767 | <varlistentry> |
1768 | <term> | |
1769 | <option>lxc.syslog</option> | |
1770 | </term> | |
1771 | <listitem> | |
1772 | <para> | |
1773 | Send logging info to syslog. It respects the log level defined in | |
1774 | <command>lxc.loglevel</command>. The argument should be the syslog | |
1775 | facility to use, valid ones are: daemon, local0, local1, local2, | |
917420dd | 1776 | local3, local4, local5, local5, local6, local7. |
204dfdf2 BD |
1777 | </para> |
1778 | </listitem> | |
1779 | </varlistentry> | |
55fc19a1 SG |
1780 | </variablelist> |
1781 | </refsect2> | |
1782 | ||
1783 | <refsect2> | |
1784 | <title>Autostart</title> | |
1785 | <para> | |
1786 | The autostart options support marking which containers should be | |
1787 | auto-started and in what order. These options may be used by LXC tools | |
1788 | directly or by external tooling provided by the distributions. | |
1789 | </para> | |
1790 | ||
1791 | <variablelist> | |
1792 | <varlistentry> | |
1793 | <term> | |
1794 | <option>lxc.start.auto</option> | |
1795 | </term> | |
1796 | <listitem> | |
1797 | <para> | |
1798 | Whether the container should be auto-started. | |
1799 | Valid values are 0 (off) and 1 (on). | |
1800 | </para> | |
1801 | </listitem> | |
1802 | </varlistentry> | |
1803 | <varlistentry> | |
1804 | <term> | |
1805 | <option>lxc.start.delay</option> | |
1806 | </term> | |
1807 | <listitem> | |
1808 | <para> | |
1809 | How long to wait (in seconds) after the container is | |
1810 | started before starting the next one. | |
1811 | </para> | |
1812 | </listitem> | |
1813 | </varlistentry> | |
1814 | <varlistentry> | |
1815 | <term> | |
1816 | <option>lxc.start.order</option> | |
1817 | </term> | |
1818 | <listitem> | |
1819 | <para> | |
1820 | An integer used to sort the containers when auto-starting | |
1821 | a series of containers at once. | |
1822 | </para> | |
1823 | </listitem> | |
1824 | </varlistentry> | |
a8dfe4e0 WB |
1825 | <varlistentry> |
1826 | <term> | |
1827 | <option>lxc.monitor.unshare</option> | |
1828 | </term> | |
1829 | <listitem> | |
1830 | <para> | |
1831 | If not zero the mount namespace will be unshared from the host | |
1832 | before initializing the container (before running any pre-start | |
6039eaa2 WB |
1833 | hooks). This requires the CAP_SYS_ADMIN capability at startup. |
1834 | Default is 0. | |
a8dfe4e0 WB |
1835 | </para> |
1836 | </listitem> | |
1837 | </varlistentry> | |
55fc19a1 SG |
1838 | <varlistentry> |
1839 | <term> | |
1840 | <option>lxc.group</option> | |
1841 | </term> | |
1842 | <listitem> | |
1843 | <para> | |
1844 | A multi-value key (can be used multiple times) to put the | |
1845 | container in a container group. Those groups can then be | |
1846 | used (amongst other things) to start a series of related | |
1847 | containers. | |
1848 | </para> | |
1849 | </listitem> | |
1850 | </varlistentry> | |
1851 | </variablelist> | |
1852 | </refsect2> | |
015f0dd7 MW |
1853 | |
1854 | <refsect2> | |
1855 | <title>Autostart and System Boot</title> | |
1856 | <para> | |
1857 | Each container can be part of any number of groups or no group at all. | |
1858 | Two groups are special. One is the NULL group, i.e. the container does | |
1859 | not belong to any group. The other group is the "onboot" group. | |
1860 | </para> | |
1861 | ||
1862 | <para> | |
1863 | When the system boots with the LXC service enabled, it will first | |
1864 | attempt to boot any containers with lxc.start.auto == 1 that is a member | |
1865 | of the "onboot" group. The startup will be in order of lxc.start.order. | |
1866 | If an lxc.start.delay has been specified, that delay will be honored | |
1867 | before attempting to start the next container to give the current | |
1868 | container time to begin initialization and reduce overloading the host | |
1869 | system. After starting the members of the "onboot" group, the LXC system | |
1870 | will proceed to boot containers with lxc.start.auto == 1 which are not | |
1871 | members of any group (the NULL group) and proceed as with the onboot | |
1872 | group. | |
1873 | </para> | |
1874 | ||
1875 | </refsect2> | |
7c661726 MP |
1876 | |
1877 | <refsect2> | |
1878 | <title>Container Environment</title> | |
1879 | <para> | |
c464fd7e SG |
1880 | If you want to pass environment variables into the container (that |
1881 | is, environment variables which will be available to init and all of | |
1882 | its descendents), you can use <command>lxc.environment</command> | |
1883 | parameters to do so. Be careful that you do not pass in anything | |
1884 | sensitive; any process in the container which doesn't have its | |
1885 | environment scrubbed will have these variables available to it, and | |
1886 | environment variables are always available via | |
1887 | <command>/proc/PID/environ</command>. | |
7c661726 MP |
1888 | </para> |
1889 | ||
1890 | <para> | |
1891 | This configuration parameter can be specified multiple times; once | |
1892 | for each environment variable you wish to configure. | |
1893 | </para> | |
1894 | ||
1895 | <variablelist> | |
c464fd7e SG |
1896 | <varlistentry> |
1897 | <term> | |
1898 | <option>lxc.environment</option> | |
1899 | </term> | |
1900 | <listitem> | |
1901 | <para> | |
1902 | Specify an environment variable to pass into the container. | |
1903 | Example: | |
1904 | </para> | |
1905 | <programlisting> | |
1906 | lxc.environment = APP_ENV=production | |
1907 | lxc.environment = SYSLOG_SERVER=192.0.2.42 | |
1908 | </programlisting> | |
1909 | </listitem> | |
1910 | </varlistentry> | |
7c661726 MP |
1911 | </variablelist> |
1912 | </refsect2> | |
1913 | ||
55fc19a1 SG |
1914 | </refsect1> |
1915 | ||
1916 | <refsect1> | |
1917 | <title>Examples</title> | |
1918 | <para> | |
c464fd7e SG |
1919 | In addition to the few examples given below, you will find |
1920 | some other examples of configuration file in @DOCDIR@/examples | |
55fc19a1 SG |
1921 | </para> |
1922 | <refsect2> | |
1923 | <title>Network</title> | |
1924 | <para>This configuration sets up a container to use a veth pair | |
c464fd7e SG |
1925 | device with one side plugged to a bridge br0 (which has been |
1926 | configured before on the system by the administrator). The | |
1927 | virtual network device visible in the container is renamed to | |
1928 | eth0.</para> | |
55fc19a1 | 1929 | <programlisting> |
c464fd7e SG |
1930 | lxc.utsname = myhostname |
1931 | lxc.network.type = veth | |
1932 | lxc.network.flags = up | |
1933 | lxc.network.link = br0 | |
1934 | lxc.network.name = eth0 | |
1935 | lxc.network.hwaddr = 4a:49:43:49:79:bf | |
1936 | lxc.network.ipv4 = 10.2.3.5/24 10.2.3.255 | |
1937 | lxc.network.ipv6 = 2003:db8:1:0:214:1234:fe0b:3597 | |
55fc19a1 SG |
1938 | </programlisting> |
1939 | </refsect2> | |
1940 | ||
1941 | <refsect2> | |
1942 | <title>UID/GID mapping</title> | |
1943 | <para>This configuration will map both user and group ids in the | |
1944 | range 0-9999 in the container to the ids 100000-109999 on the host. | |
1945 | </para> | |
1946 | <programlisting> | |
c464fd7e SG |
1947 | lxc.id_map = u 0 100000 10000 |
1948 | lxc.id_map = g 0 100000 10000 | |
55fc19a1 SG |
1949 | </programlisting> |
1950 | </refsect2> | |
1951 | ||
1952 | <refsect2> | |
1953 | <title>Control group</title> | |
1954 | <para>This configuration will setup several control groups for | |
1955 | the application, cpuset.cpus restricts usage of the defined cpu, | |
1956 | cpus.share prioritize the control group, devices.allow makes | |
1957 | usable the specified devices.</para> | |
1958 | <programlisting> | |
c464fd7e SG |
1959 | lxc.cgroup.cpuset.cpus = 0,1 |
1960 | lxc.cgroup.cpu.shares = 1234 | |
1961 | lxc.cgroup.devices.deny = a | |
1962 | lxc.cgroup.devices.allow = c 1:3 rw | |
1963 | lxc.cgroup.devices.allow = b 8:0 rw | |
55fc19a1 SG |
1964 | </programlisting> |
1965 | </refsect2> | |
1966 | ||
1967 | <refsect2> | |
1968 | <title>Complex configuration</title> | |
1969 | <para>This example show a complex configuration making a complex | |
1970 | network stack, using the control groups, setting a new hostname, | |
1971 | mounting some locations and a changing root file system.</para> | |
1972 | <programlisting> | |
c464fd7e SG |
1973 | lxc.utsname = complex |
1974 | lxc.network.type = veth | |
1975 | lxc.network.flags = up | |
1976 | lxc.network.link = br0 | |
1977 | lxc.network.hwaddr = 4a:49:43:49:79:bf | |
1978 | lxc.network.ipv4 = 10.2.3.5/24 10.2.3.255 | |
1979 | lxc.network.ipv6 = 2003:db8:1:0:214:1234:fe0b:3597 | |
1980 | lxc.network.ipv6 = 2003:db8:1:0:214:5432:feab:3588 | |
1981 | lxc.network.type = macvlan | |
1982 | lxc.network.flags = up | |
1983 | lxc.network.link = eth0 | |
1984 | lxc.network.hwaddr = 4a:49:43:49:79:bd | |
1985 | lxc.network.ipv4 = 10.2.3.4/24 | |
1986 | lxc.network.ipv4 = 192.168.10.125/24 | |
1987 | lxc.network.ipv6 = 2003:db8:1:0:214:1234:fe0b:3596 | |
1988 | lxc.network.type = phys | |
1989 | lxc.network.flags = up | |
1990 | lxc.network.link = dummy0 | |
1991 | lxc.network.hwaddr = 4a:49:43:49:79:ff | |
1992 | lxc.network.ipv4 = 10.2.3.6/24 | |
1993 | lxc.network.ipv6 = 2003:db8:1:0:214:1234:fe0b:3297 | |
1994 | lxc.cgroup.cpuset.cpus = 0,1 | |
1995 | lxc.cgroup.cpu.shares = 1234 | |
1996 | lxc.cgroup.devices.deny = a | |
1997 | lxc.cgroup.devices.allow = c 1:3 rw | |
1998 | lxc.cgroup.devices.allow = b 8:0 rw | |
1999 | lxc.mount = /etc/fstab.complex | |
2000 | lxc.mount.entry = /lib /root/myrootfs/lib none ro,bind 0 0 | |
2001 | lxc.rootfs = /mnt/rootfs.complex | |
2002 | lxc.cap.drop = sys_module mknod setuid net_raw | |
2003 | lxc.cap.drop = mac_override | |
55fc19a1 SG |
2004 | </programlisting> |
2005 | </refsect2> | |
2006 | ||
2007 | </refsect1> | |
2008 | ||
2009 | <refsect1> | |
2010 | <title>See Also</title> | |
2011 | <simpara> | |
2012 | <citerefentry> | |
c464fd7e SG |
2013 | <refentrytitle><command>chroot</command></refentrytitle> |
2014 | <manvolnum>1</manvolnum> | |
55fc19a1 SG |
2015 | </citerefentry>, |
2016 | ||
2017 | <citerefentry> | |
c464fd7e SG |
2018 | <refentrytitle><command>pivot_root</command></refentrytitle> |
2019 | <manvolnum>8</manvolnum> | |
55fc19a1 SG |
2020 | </citerefentry>, |
2021 | ||
2022 | <citerefentry> | |
c464fd7e SG |
2023 | <refentrytitle><filename>fstab</filename></refentrytitle> |
2024 | <manvolnum>5</manvolnum> | |
55fc19a1 SG |
2025 | </citerefentry>, |
2026 | ||
2027 | <citerefentry> | |
c464fd7e SG |
2028 | <refentrytitle><filename>capabilities</filename></refentrytitle> |
2029 | <manvolnum>7</manvolnum> | |
55fc19a1 SG |
2030 | </citerefentry> |
2031 | </simpara> | |
2032 | </refsect1> | |
2033 | ||
2034 | &seealso; | |
2035 | ||
2036 | <refsect1> | |
2037 | <title>Author</title> | |
2038 | <para>Daniel Lezcano <email>daniel.lezcano@free.fr</email></para> | |
2039 | </refsect1> | |
2040 | ||
2041 | </refentry> | |
2042 | ||
2043 | <!-- Keep this comment at the end of the file | |
2044 | Local variables: | |
2045 | mode: sgml | |
2046 | sgml-omittag:t | |
2047 | sgml-shorttag:t | |
2048 | sgml-minimize-attributes:nil | |
2049 | sgml-always-quote-attributes:t | |
2050 | sgml-indent-step:2 | |
2051 | sgml-indent-data:t | |
2052 | sgml-parent-document:nil | |
2053 | sgml-default-dtd-file:nil | |
2054 | sgml-exposed-tags:nil | |
2055 | sgml-local-catalogs:nil | |
2056 | sgml-local-ecat-files:nil | |
2057 | End: | |
2058 | --> |