3 lxc: linux Container library
5 (C) Copyright IBM Corp. 2007, 2008
8 Daniel Lezcano <daniel.lezcano at free.fr>
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.
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.
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
26 <!DOCTYPE refentry PUBLIC @docdtd@ [
28 <!ENTITY seealso SYSTEM "@builddir@/see_also.sgml">
33 <docinfo><date>@LXC_GENERATE_DATE@</date></docinfo>
36 <refentrytitle>lxc.container.conf</refentrytitle>
37 <manvolnum>5</manvolnum>
41 <refname>lxc.container.conf</refname>
44 LXC container configuration file
49 <title>Description</title>
52 LXC is the well-known and heavily tested low-level Linux container
53 runtime. It is in active development since 2008 and has proven itself in
54 critical production environments world-wide. Some of its core contributors
55 are the same people that helped to implement various well-known
56 containerization features inside the Linux kernel.
60 LXC's main focus is system containers. That is, containers which offer an
61 environment as close as possible as the one you'd get from a VM but
62 without the overhead that comes with running a separate kernel and
63 simulating all the hardware.
67 This is achieved through a combination of kernel security features such as
68 namespaces, mandatory access control and control groups.
72 LXC has supports unprivileged containers. Unprivileged containers are
73 containers that are run without any privilege. This requires support for
74 user namespaces in the kernel that the container is run on. LXC was the
75 first runtime to support unprivileged containers after user namespaces
76 were merged into the mainline kernel.
80 In essence, user namespaces isolate given sets of UIDs and GIDs. This is
81 achieved by establishing a mapping between a range of UIDs and GIDs on the
82 host to a different (unprivileged) range of UIDs and GIDs in the
83 container. The kernel will translate this mapping in such a way that
84 inside the container all UIDs and GIDs appear as you would expect from the
85 host whereas on the host these UIDs and GIDs are in fact unprivileged. For
86 example, a process running as UID and GID 0 inside the container might
87 appear as UID and GID 100000 on the host. The implementation and working
88 details can be gathered from the corresponding user namespace man page.
89 UID and GID mappings can be defined with the <option>lxc.idmap</option>
94 Linux containers are defined with a simple configuration file. Each
95 option in the configuration file has the form <command>key =
96 value</command> fitting in one line. The "#" character means the line is a
97 comment. List options, like capabilities and cgroups options, can be used
98 with no value to clear any previously defined values of that option.
102 LXC namespaces configuration keys by using single dots. This means complex
103 configuration keys such as <option>lxc.net.0</option> expose various
104 subkeys such as <option>lxc.net.0.type</option>,
105 <option>lxc.net.0.link</option>, <option>lxc.net.0.ipv6.address</option>, and
106 others for even more fine-grained configuration.
110 <title>Configuration</title>
112 In order to ease administration of multiple related containers, it is
113 possible to have a container configuration file cause another file to be
114 loaded. For instance, network configuration can be defined in one common
115 file which is included by multiple containers. Then, if the containers
116 are moved to another host, only one file may need to be updated.
122 <option>lxc.include</option>
126 Specify the file to be included. The included file must be
127 in the same valid lxc configuration file format.
135 <title>Architecture</title>
137 Allows one to set the architecture for the container. For example, set a
138 32bits architecture for a container running 32bits binaries on a 64bits
139 host. This fixes the container scripts which rely on the architecture to
140 do some work like downloading the packages.
146 <option>lxc.arch</option>
150 Specify the architecture for the container.
153 Some valid options are
154 <option>x86</option>,
155 <option>i686</option>,
156 <option>x86_64</option>,
157 <option>amd64</option>
166 <title>Hostname</title>
168 The utsname section defines the hostname to be set for the container.
169 That means the container can set its own hostname without changing the
170 one from the system. That makes the hostname private for the container.
175 <option>lxc.uts.name</option>
179 specify the hostname for the container
187 <title>Halt signal</title>
189 Allows one to specify signal name or number sent to the container's
190 init process to cleanly shutdown the container. Different init systems
191 could use different signals to perform clean shutdown sequence. This
192 option allows the signal to be specified in kill(1) fashion, e.g.
193 SIGPWR, SIGRTMIN+14, SIGRTMAX-10 or plain number. The default signal is
199 <option>lxc.signal.halt</option>
203 specify the signal used to halt the container
211 <title>Reboot signal</title>
213 Allows one to specify signal name or number to reboot the container.
214 This option allows signal to be specified in kill(1) fashion, e.g.
215 SIGTERM, SIGRTMIN+14, SIGRTMAX-10 or plain number. The default signal
221 <option>lxc.signal.reboot</option>
225 specify the signal used to reboot the container
233 <title>Stop signal</title>
235 Allows one to specify signal name or number to forcibly shutdown the
236 container. This option allows signal to be specified in kill(1) fashion,
237 e.g. SIGKILL, SIGRTMIN+14, SIGRTMAX-10 or plain number. The default
243 <option>lxc.signal.stop</option>
247 specify the signal used to stop the container
255 <title>Init command</title>
257 Sets the command to use as the init system for the containers.
262 <option>lxc.execute.cmd</option>
266 Absolute path from container rootfs to the binary to run by default. This
267 mostly makes sense for <command>lxc-execute</command>.
275 <option>lxc.init.cmd</option>
279 Absolute path from container rootfs to the binary to use as init. This
280 mostly makes sense for <command>lxc-start</command>. Default is <command>/sbin/init</command>.
288 <title>Init working directory</title>
290 Sets the absolute path inside the container as the working directory for the containers.
291 LXC will switch to this directory before executing init.
296 <option>lxc.init.cwd</option>
300 Absolute path inside the container to use as the working directory.
308 <title>Init ID</title>
310 Sets the UID/GID to use for the init system, and subsequent commands.
311 Note that using a non-root uid when booting a system container will
312 likely not work due to missing privileges. Setting the UID/GID is mostly
313 useful when running application container.
315 Defaults to: UID(0), GID(0)
320 <option>lxc.init.uid</option>
330 <option>lxc.init.gid</option>
344 Configure proc filesystem for the container.
349 <option>lxc.proc.[proc file name]</option>
353 Specify the proc file name to be set. The file name available
354 are those listed under /proc/PID/.
358 lxc.proc.oom_score_adj = 10
366 <title>Ephemeral</title>
368 Allows one to specify whether a container will be destroyed on shutdown.
373 <option>lxc.ephemeral</option>
377 The only allowed values are 0 and 1. Set this to 1 to destroy a
378 container on shutdown.
386 <title>Network</title>
388 The network section defines how the network is virtualized in
389 the container. The network virtualization acts at layer
390 two. In order to use the network virtualization, parameters
391 must be specified to define the network interfaces of the
392 container. Several virtual interfaces can be assigned and used
393 in a container even if the system has only one physical
399 <option>lxc.net</option>
403 may be used without a value to clear all previous network options.
409 <option>lxc.net.[i].type</option>
413 specify what kind of network virtualization to be used
415 Multiple networks can be specified by using an additional index
417 after all <option>lxc.net.*</option> keys. For example,
418 <option>lxc.net.0.type = veth</option> and
419 <option>lxc.net.1.type = veth</option> specify two different
420 networks of the same type. All keys sharing the same index
421 <option>i</option> will be treated as belonging to the same
422 network. For example, <option>lxc.net.0.link = br0</option>
423 will belong to <option>lxc.net.0.type</option>.
424 Currently, the different virtualization types can be:
428 <option>none:</option> will cause the container to share
429 the host's network namespace. This means the host
430 network devices are usable in the container. It also
431 means that if both the container and host have upstart as
432 init, 'halt' in a container (for instance) will shut down the
437 <option>empty:</option> will create only the loopback
442 <option>veth:</option> a virtual ethernet pair
443 device is created with one side assigned to the container
444 and the other side attached to a bridge specified by
445 the <option>lxc.net.[i].link</option> option.
446 If the bridge is not specified, then the veth pair device
447 will be created but not attached to any bridge.
448 Otherwise, the bridge has to be created on the system
449 before starting the container.
450 <command>lxc</command> won't handle any
451 configuration outside of the container.
452 By default, <command>lxc</command> chooses a name for the
453 network device belonging to the outside of the
454 container, but if you wish to handle
455 this name yourselves, you can tell <command>lxc</command>
456 to set a specific name with
457 the <option>lxc.net.[i].veth.pair</option> option (except for
458 unprivileged containers where this option is ignored for security
463 <option>vlan:</option> a vlan interface is linked with
464 the interface specified by
465 the <option>lxc.net.[i].link</option> and assigned to
466 the container. The vlan identifier is specified with the
467 option <option>lxc.net.[i].vlan.id</option>.
471 <option>macvlan:</option> a macvlan interface is linked
472 with the interface specified by
473 the <option>lxc.net.[i].link</option> and assigned to
475 <option>lxc.net.[i].macvlan.mode</option> specifies the
476 mode the macvlan will use to communicate between
477 different macvlan on the same upper device. The accepted
478 modes are <option>private</option>, <option>vepa</option>,
479 <option>bridge</option> and <option>passthru</option>.
480 In <option>private</option> mode, the device never
481 communicates with any other device on the same upper_dev (default).
482 In <option>vepa</option> mode, the new Virtual Ethernet Port
483 Aggregator (VEPA) mode, it assumes that the adjacent
484 bridge returns all frames where both source and
485 destination are local to the macvlan port, i.e. the
486 bridge is set up as a reflective relay. Broadcast
487 frames coming in from the upper_dev get flooded to all
488 macvlan interfaces in VEPA mode, local frames are not
489 delivered locally. In <option>bridge</option> mode, it
490 provides the behavior of a simple bridge between
491 different macvlan interfaces on the same port. Frames
492 from one interface to another one get delivered directly
493 and are not sent out externally. Broadcast frames get
494 flooded to all other bridge ports and to the external
495 interface, but when they come back from a reflective
496 relay, we don't deliver them again. Since we know all
497 the MAC addresses, the macvlan bridge mode does not
498 require learning or STP like the bridge module does. In
499 <option>passthru</option> mode, all frames received by
500 the physical interface are forwarded to the macvlan
501 interface. Only one macvlan interface in <option>passthru</option>
502 mode is possible for one physical interface.
506 <option>phys:</option> an already existing interface
507 specified by the <option>lxc.net.[i].link</option> is
508 assigned to the container.
515 <option>lxc.net.[i].flags</option>
519 Specify an action to do for the network.
522 <para><option>up:</option> activates the interface.
529 <option>lxc.net.[i].link</option>
533 Specify the interface to be used for real network traffic.
540 <option>lxc.net.[i].mtu</option>
544 Specify the maximum transfer unit for this interface.
551 <option>lxc.net.[i].name</option>
555 The interface name is dynamically allocated, but if another name
556 is needed because the configuration files being used by the
557 container use a generic name, eg. eth0, this option will rename
558 the interface in the container.
565 <option>lxc.net.[i].hwaddr</option>
569 The interface mac address is dynamically allocated by default to
570 the virtual interface, but in some cases, this is needed to
571 resolve a mac address conflict or to always have the same
572 link-local ipv6 address. Any "x" in address will be replaced by
573 random value, this allows setting hwaddr templates.
580 <option>lxc.net.[i].ipv4.address</option>
584 Specify the ipv4 address to assign to the virtualized interface.
585 Several lines specify several ipv4 addresses. The address is in
586 format x.y.z.t/m, eg. 192.168.1.123/24.
593 <option>lxc.net.[i].ipv4.gateway</option>
597 Specify the ipv4 address to use as the gateway inside the
598 container. The address is in format x.y.z.t, eg. 192.168.1.123.
600 Can also have the special value <option>auto</option>,
601 which means to take the primary address from the bridge
602 interface (as specified by the
603 <option>lxc.net.[i].link</option> option) and use that as
604 the gateway. <option>auto</option> is only available when
605 using the <option>veth</option> and
606 <option>macvlan</option> network types.
614 <option>lxc.net.[i].ipv6.address</option>
618 Specify the ipv6 address to assign to the virtualized
619 interface. Several lines specify several ipv6 addresses. The
620 address is in format x::y/m, eg.
621 2003:db8:1:0:214:1234:fe0b:3596/64
628 <option>lxc.net.[i].ipv6.gateway</option>
632 Specify the ipv6 address to use as the gateway inside the
633 container. The address is in format x::y, eg. 2003:db8:1:0::1
635 Can also have the special value <option>auto</option>,
636 which means to take the primary address from the bridge
637 interface (as specified by the
638 <option>lxc.net.[i].link</option> option) and use that as
639 the gateway. <option>auto</option> is only available when
640 using the <option>veth</option> and
641 <option>macvlan</option> network types.
648 <option>lxc.net.[i].script.up</option>
652 Add a configuration option to specify a script to be
653 executed after creating and configuring the network used
654 from the host side. The following arguments are passed
655 to the script: container name and config section name
656 (net) Additional arguments depend on the config section
657 employing a script hook; the following are used by the
658 network system: execution context (up), network type
659 (empty/veth/macvlan/phys), Depending on the network
660 type, other arguments may be passed:
661 veth/macvlan/phys. And finally (host-sided) device name.
664 Standard output from the script is logged at debug level.
665 Standard error is not logged, but can be captured by the
666 hook redirecting its standard error to standard output.
673 <option>lxc.net.[i].script.down</option>
677 Add a configuration option to specify a script to be
678 executed before destroying the network used from the
679 host side. The following arguments are passed to the
680 script: container name and config section name (net)
681 Additional arguments depend on the config section
682 employing a script hook; the following are used by the
683 network system: execution context (down), network type
684 (empty/veth/macvlan/phys), Depending on the network
685 type, other arguments may be passed:
686 veth/macvlan/phys. And finally (host-sided) device name.
689 Standard output from the script is logged at debug level.
690 Standard error is not logged, but can be captured by the
691 hook redirecting its standard error to standard output.
699 <title>New pseudo tty instance (devpts)</title>
701 For stricter isolation the container can have its own private
702 instance of the pseudo tty.
707 <option>lxc.pty.max</option>
711 If set, the container will have a new pseudo tty
712 instance, making this private to it. The value specifies
713 the maximum number of pseudo ttys allowed for a pts
714 instance (this limitation is not implemented yet).
722 <title>Container system console</title>
724 If the container is configured with a root filesystem and the
725 inittab file is setup to use the console, you may want to specify
726 where the output of this console goes.
732 <option>lxc.console.buffer.size</option>
736 Setting this option instructs liblxc to allocate an in-memory
737 ringbuffer. The container's console output will be written to the
738 ringbuffer. Note that ringbuffer must be at least as big as a
739 standard page size. When passed a value smaller than a single page
740 size liblxc will allocate a ringbuffer of a single page size. A page
743 The keyword 'auto' will cause liblxc to allocate a ringbuffer of
746 When manually specifying a size for the ringbuffer the value should
747 be a power of 2 when converted to bytes. Valid size prefixes are
748 'kB', 'MB', 'GB'. (Note that all conversions are based on multiples
749 of 1024. That means 'kb' == 'KiB', 'MB' == 'MiB', 'GB' == 'GiB'.)
756 <option>lxc.console.buffer.logfile</option>
760 Setting this option instructs liblxc to write the in-memory
761 ringbuffer to disk. For performance reasons liblxc will only write
762 the in-memory ringbuffer to disk when requested. Note that the this
763 option is only used by liblxc when
764 <option>lxc.console.buffer.size</option> is set.
766 By default liblxc will dump the contents of the in-memory ringbuffer
767 to disk when the container terminates. This allows users to diagnose
768 boot failures when the container crashed before an API request to
769 retrieve the in-memory ringbuffer could be sent or handled.
776 <option>lxc.console.logfile</option>
780 Specify a path to a file where the console output will be written.
781 Note that in contrast to the on-disk ringbuffer logfile this file
782 will keep growing potentially filling up the users disks if not
783 rotated and deleted. This problem can also be avoided by using the
784 in-memory ringbuffer options
785 <option>lxc.console.buffer.size</option> and
786 <option>lxc.console.buffer.logfile</option>.
793 <option>lxc.console.rotate</option>
797 Whether to rotate the console logfile specified in
798 <option>lxc.console.logfile</option>. Users can send an API
799 request to rotate the logfile. Note that the old logfile will have
800 the same name as the original with the suffix ".1" appended.
802 Users wishing to prevent the console log file from filling the
803 disk should rotate the logfile and delete it if unneeded. This
804 problem can also be avoided by using the in-memory ringbuffer
805 options <option>lxc.console.buffer.size</option> and
806 <option>lxc.console.buffer.logfile</option>.
813 <option>lxc.console.path</option>
817 Specify a path to a device to which the console will be
818 attached. The keyword 'none' will simply disable the
819 console. Note, when specifying 'none' and creating a device node
820 for the console in the container at /dev/console or bind-mounting
821 the hosts's /dev/console into the container at /dev/console the
822 container will have direct access to the hosts's /dev/console.
823 This is dangerous when the container has write access to the
824 device and should thus be used with caution.
832 <title>Console through the ttys</title>
834 This option is useful if the container is configured with a root
835 filesystem and the inittab file is setup to launch a getty on the
836 ttys. The option specifies the number of ttys to be available for
837 the container. The number of gettys in the inittab file of the
838 container should not be greater than the number of ttys specified
839 in this option, otherwise the excess getty sessions will die and
840 respawn indefinitely giving annoying messages on the console or in
841 <filename>/var/log/messages</filename>.
846 <option>lxc.tty.max</option>
850 Specify the number of tty to make available to the
859 <title>Console devices location</title>
861 LXC consoles are provided through Unix98 PTYs created on the
862 host and bind-mounted over the expected devices in the container.
863 By default, they are bind-mounted over <filename>/dev/console</filename>
864 and <filename>/dev/ttyN</filename>. This can prevent package upgrades
865 in the guest. Therefore you can specify a directory location (under
866 <filename>/dev</filename> under which LXC will create the files and
867 bind-mount over them. These will then be symbolically linked to
868 <filename>/dev/console</filename> and <filename>/dev/ttyN</filename>.
869 A package upgrade can then succeed as it is able to remove and replace
875 <option>lxc.tty.dir</option>
879 Specify a directory under <filename>/dev</filename>
880 under which to create the container console devices. Note that LXC
881 will move any bind-mounts or device nodes for /dev/console into
890 <title>/dev directory</title>
892 By default, lxc creates a few symbolic links (fd,stdin,stdout,stderr)
893 in the container's <filename>/dev</filename> directory but does not
894 automatically create device node entries. This allows the container's
895 <filename>/dev</filename> to be set up as needed in the container
896 rootfs. If lxc.autodev is set to 1, then after mounting the container's
897 rootfs LXC will mount a fresh tmpfs under <filename>/dev</filename>
898 (limited to 500k) and fill in a minimal set of initial devices.
899 This is generally required when starting a container containing
900 a "systemd" based "init" but may be optional at other times. Additional
901 devices in the containers /dev directory may be created through the
902 use of the <option>lxc.hook.autodev</option> hook.
907 <option>lxc.autodev</option>
911 Set this to 0 to stop LXC from mounting and populating a minimal
912 <filename>/dev</filename> when starting the container.
920 <title>Mount points</title>
922 The mount points section specifies the different places to be
923 mounted. These mount points will be private to the container
924 and won't be visible by the processes running outside of the
925 container. This is useful to mount /etc, /var or /home for
929 NOTE - LXC will generally ensure that mount targets and relative
930 bind-mount sources are properly confined under the container
931 root, to avoid attacks involving over-mounting host directories
932 and files. (Symbolic links in absolute mount sources are ignored)
933 However, if the container configuration first mounts a directory which
934 is under the control of the container user, such as /home/joe, into
935 the container at some <filename>path</filename>, and then mounts
936 under <filename>path</filename>, then a TOCTTOU attack would be
937 possible where the container user modifies a symbolic link under
938 his home directory at just the right time.
943 <option>lxc.mount.fstab</option>
947 specify a file location in
948 the <filename>fstab</filename> format, containing the
949 mount information. The mount target location can and in
950 most cases should be a relative path, which will become
951 relative to the mounted container root. For instance,
954 proc proc proc nodev,noexec,nosuid 0 0
957 Will mount a proc filesystem under the container's /proc,
958 regardless of where the root filesystem comes from. This
959 is resilient to block device backed filesystems as well as
963 Note that when mounting a filesystem from an
964 image file or block device the third field (fs_vfstype)
965 cannot be auto as with
967 <refentrytitle>mount</refentrytitle>
968 <manvolnum>8</manvolnum>
970 but must be explicitly specified.
977 <option>lxc.mount.entry</option>
981 specify a mount point corresponding to a line in the
984 Moreover lxc add two options to mount.
985 <option>optional</option> don't fail if mount does not work.
986 <option>create=dir</option> or <option>create=file</option>
987 to create dir (or file) when the point will be mounted.
994 <option>lxc.mount.auto</option>
998 specify which standard kernel file systems should be
999 automatically mounted. This may dramatically simplify
1000 the configuration. The file systems are:
1005 <option>proc:mixed</option> (or <option>proc</option>):
1006 mount <filename>/proc</filename> as read-write, but
1007 remount <filename>/proc/sys</filename> and
1008 <filename>/proc/sysrq-trigger</filename> read-only
1009 for security / container isolation purposes.
1014 <option>proc:rw</option>: mount
1015 <filename>/proc</filename> as read-write
1020 <option>sys:mixed</option> (or <option>sys</option>):
1021 mount <filename>/sys</filename> as read-only but with
1022 /sys/devices/virtual/net writable.
1027 <option>sys:ro</option>:
1028 mount <filename>/sys</filename> as read-only
1029 for security / container isolation purposes.
1034 <option>sys:rw</option>: mount
1035 <filename>/sys</filename> as read-write
1040 <option>cgroup:mixed</option>:
1041 mount a tmpfs to <filename>/sys/fs/cgroup</filename>,
1042 create directories for all hierarchies to which
1043 the container is added, create subdirectories
1044 there with the name of the cgroup, and bind-mount
1045 the container's own cgroup into that directory.
1046 The container will be able to write to its own
1047 cgroup directory, but not the parents, since they
1048 will be remounted read-only.
1053 <option>cgroup:ro</option>: similar to
1054 <option>cgroup:mixed</option>, but everything will
1055 be mounted read-only.
1060 <option>cgroup:rw</option>: similar to
1061 <option>cgroup:mixed</option>, but everything will
1062 be mounted read-write. Note that the paths leading
1063 up to the container's own cgroup will be writable,
1064 but will not be a cgroup filesystem but just part
1065 of the tmpfs of <filename>/sys/fs/cgroup</filename>
1070 <option>cgroup</option> (without specifier):
1071 defaults to <option>cgroup:rw</option> if the
1072 container retains the CAP_SYS_ADMIN capability,
1073 <option>cgroup:mixed</option> otherwise.
1078 <option>cgroup-full:mixed</option>:
1079 mount a tmpfs to <filename>/sys/fs/cgroup</filename>,
1080 create directories for all hierarchies to which
1081 the container is added, bind-mount the hierarchies
1082 from the host to the container and make everything
1083 read-only except the container's own cgroup. Note
1084 that compared to <option>cgroup</option>, where
1085 all paths leading up to the container's own cgroup
1086 are just simple directories in the underlying
1088 <filename>/sys/fs/cgroup/$hierarchy</filename>
1089 will contain the host's full cgroup hierarchy,
1090 albeit read-only outside the container's own cgroup.
1091 This may leak quite a bit of information into the
1097 <option>cgroup-full:ro</option>: similar to
1098 <option>cgroup-full:mixed</option>, but everything
1099 will be mounted read-only.
1104 <option>cgroup-full:rw</option>: similar to
1105 <option>cgroup-full:mixed</option>, but everything
1106 will be mounted read-write. Note that in this case,
1107 the container may escape its own cgroup. (Note also
1108 that if the container has CAP_SYS_ADMIN support
1109 and can mount the cgroup filesystem itself, it may
1115 <option>cgroup-full</option> (without specifier):
1116 defaults to <option>cgroup-full:rw</option> if the
1117 container retains the CAP_SYS_ADMIN capability,
1118 <option>cgroup-full:mixed</option> otherwise.
1123 If cgroup namespaces are enabled, then any <option>cgroup</option>
1124 auto-mounting request will be ignored, since the container can
1125 mount the filesystems itself, and automounting can confuse the
1129 Note that if automatic mounting of the cgroup filesystem
1130 is enabled, the tmpfs under
1131 <filename>/sys/fs/cgroup</filename> will always be
1132 mounted read-write (but for the <option>:mixed</option>
1133 and <option>:ro</option> cases, the individual
1135 <filename>/sys/fs/cgroup/$hierarchy</filename>, will be
1136 read-only). This is in order to work around a quirk in
1139 <refentrytitle>mountall</refentrytitle>
1140 <manvolnum>8</manvolnum>
1142 command that will cause containers to wait for user
1144 <filename>/sys/fs/cgroup</filename> is mounted read-only
1145 and the container can't remount it read-write due to a
1146 lack of CAP_SYS_ADMIN.
1152 lxc.mount.auto = proc sys cgroup
1153 lxc.mount.auto = proc:rw sys:rw cgroup-full:rw
1162 <title>Root file system</title>
1164 The root file system of the container can be different than that
1170 <option>lxc.rootfs.path</option>
1174 specify the root file system for the container. It can
1175 be an image file, a directory or a block device. If not
1176 specified, the container shares its root file system
1180 For directory or simple block-device backed containers,
1181 a pathname can be used. If the rootfs is backed by a nbd
1182 device, then <filename>nbd:file:1</filename> specifies that
1183 <filename>file</filename> should be attached to a nbd device,
1184 and partition 1 should be mounted as the rootfs.
1185 <filename>nbd:file</filename> specifies that the nbd device
1186 itself should be mounted. <filename>overlayfs:/lower:/upper</filename>
1187 specifies that the rootfs should be an overlay with <filename>/upper</filename>
1188 being mounted read-write over a read-only mount of <filename>/lower</filename>.
1189 <filename>aufs:/lower:/upper</filename> does the same using aufs in place
1190 of overlayfs. For both <filename>overlayfs</filename> and
1191 <filename>aufs</filename> multiple <filename>/lower</filename>
1192 directories can be specified. <filename>loop:/file</filename> tells lxc to attach
1193 <filename>/file</filename> to a loop device and mount the loop device.
1200 <option>lxc.rootfs.mount</option>
1204 where to recursively bind <option>lxc.rootfs.path</option>
1205 before pivoting. This is to ensure success of the
1207 <refentrytitle><command>pivot_root</command></refentrytitle>
1208 <manvolnum>8</manvolnum>
1210 syscall. Any directory suffices, the default should
1218 <option>lxc.rootfs.options</option>
1222 extra mount options to use when mounting the rootfs.
1231 <title>Control group</title>
1233 The control group section contains the configuration for the
1234 different subsystem. <command>lxc</command> does not check the
1235 correctness of the subsystem name. This has the disadvantage
1236 of not detecting configuration errors until the container is
1237 started, but has the advantage of permitting any future
1243 <option>lxc.cgroup.[subsystem name]</option>
1247 specify the control group value to be set. The
1248 subsystem name is the literal name of the control group
1249 subsystem. The permitted names and the syntax of their
1250 values is not dictated by LXC, instead it depends on the
1251 features of the Linux kernel running at the time the
1252 container is started,
1253 eg. <option>lxc.cgroup.cpuset.cpus</option>
1259 <option>lxc.cgroup.dir</option>
1263 specify a directory or path in which the container's cgroup will
1264 be created. For example, setting
1265 <option>lxc.cgroup.dir = my-cgroup/first</option> for a container
1266 named "c1" will create the container's cgroup as a sub-cgroup of
1267 "my-cgroup". For example, if the user's current cgroup "my-user"
1268 is located in the root cgroup of the cpuset controller in a
1269 cgroup v1 hierarchy this would create the cgroup
1270 "/sys/fs/cgroup/cpuset/my-user/my-cgroup/first/c1" for the
1271 container. Any missing cgroups will be created by LXC. This
1272 presupposes that the user has write access to its current cgroup.
1280 <title>Capabilities</title>
1282 The capabilities can be dropped in the container if this one
1288 <option>lxc.cap.drop</option>
1292 Specify the capability to be dropped in the container. A
1293 single line defining several capabilities with a space
1294 separation is allowed. The format is the lower case of
1295 the capability definition without the "CAP_" prefix,
1296 eg. CAP_SYS_MODULE should be specified as
1299 <refentrytitle><command>capabilities</command></refentrytitle>
1300 <manvolnum>7</manvolnum>
1302 If used with no value, lxc will clear any drop capabilities
1303 specified up to this point.
1309 <option>lxc.cap.keep</option>
1313 Specify the capability to be kept in the container. All other
1314 capabilities will be dropped. When a special value of "none" is
1315 encountered, lxc will clear any keep capabilities specified up
1316 to this point. A value of "none" alone can be used to drop all
1325 <title>Namespace Inheritance</title>
1327 A namespace can be inherited from another container or process.
1332 <option>lxc.namespace.[namespace identifier]</option>
1336 Specify a namespace to inherit from another container or process.
1337 The <option>[namespace identifier]</option> suffix needs to be
1338 replaced with one of the namespaces that appear in the
1339 <filename>/proc/PID/ns</filename> directory.
1343 To inherit the namespace from another process set the
1344 <option>lxc.namespace.[namespace identifier]</option> to the PID of
1345 the process, e.g. <option>lxc.namespace.net=42</option>.
1349 To inherit the namespace from another container set the
1350 <option>lxc.namespace.[namespace identifier]</option> to the name of
1351 the container, e.g. <option>lxc.namespace.pid=c3</option>.
1355 To inherit the namespace from another container located in a
1356 different path than the standard liblxc path set the
1357 <option>lxc.namespace.[namespace identifier]</option> to the full
1358 path to the container, e.g.
1359 <option>lxc.namespace.user=/opt/c3</option>.
1363 In order to inherit namespaces the caller needs to have sufficient
1364 privilege over the process or container.
1368 Note that sharing pid namespaces between system containers will
1369 likely not work with most init systems.
1373 Note that if two processes are in different user namespaces and one
1374 process wants to inherit the other's network namespace it usually
1375 needs to inherit the user namespace as well.
1383 <title>Resource limits</title>
1385 The soft and hard resource limits for the container can be changed.
1386 Unprivileged containers can only lower them. Resources which are not
1387 explicitly specified will be inherited.
1392 <option>lxc.prlimit.[limit name]</option>
1396 Specify the resource limit to be set. A limit is specified as two
1397 colon separated values which are either numeric or the word
1398 'unlimited'. A single value can be used as a shortcut to set both
1399 soft and hard limit to the same value. The permitted names the
1400 "RLIMIT_" resource names in lowercase without the "RLIMIT_"
1401 prefix, eg. RLIMIT_NOFILE should be specified as "nofile". See
1403 <refentrytitle><command>setrlimit</command></refentrytitle>
1404 <manvolnum>2</manvolnum>
1406 If used with no value, lxc will clear the resource limit
1407 specified up to this point. A resource with no explicitly
1408 configured limitation will be inherited from the process starting
1417 <title>Sysctl</title>
1419 Configure kernel parameters for the container.
1424 <option>lxc.sysctl.[kernel parameters name]</option>
1428 Specify the kernel parameters to be set. The parameters available
1429 are those listed under /proc/sys/.
1430 Note that not all sysctls are namespaced. Changing Non-namespaced
1431 sysctls will cause the system-wide setting to be modified.
1433 <refentrytitle><command>sysctl</command></refentrytitle>
1434 <manvolnum>8</manvolnum>
1436 If used with no value, lxc will clear the parameters specified up
1445 <title>Apparmor profile</title>
1447 If lxc was compiled and installed with apparmor support, and the host
1448 system has apparmor enabled, then the apparmor profile under which the
1449 container should be run can be specified in the container
1450 configuration. The default is <command>lxc-container-default-cgns</command>
1451 if the host kernel is cgroup namespace aware, or
1452 <command>lxc-container-default</command> othewise.
1457 <option>lxc.apparmor.profile</option>
1461 Specify the apparmor profile under which the container should
1462 be run. To specify that the container should be unconfined,
1465 <programlisting>lxc.apparmor.profile = unconfined</programlisting>
1467 If the apparmor profile should remain unchanged (i.e. if you
1468 are nesting containers and are already confined), then use
1470 <programlisting>lxc.apparmor.profile = unchanged</programlisting>
1475 <option>lxc.apparmor.allow_incomplete</option>
1479 Apparmor profiles are pathname based. Therefore many file
1480 restrictions require mount restrictions to be effective against
1481 a determined attacker. However, these mount restrictions are not
1482 yet implemented in the upstream kernel. Without the mount
1483 restrictions, the apparmor profiles still protect against accidental
1487 If this flag is 0 (default), then the container will not be
1488 started if the kernel lacks the apparmor mount features, so that a
1489 regression after a kernel upgrade will be detected. To start the
1490 container under partial apparmor protection, set this flag to 1.
1498 <title>SELinux context</title>
1500 If lxc was compiled and installed with SELinux support, and the host
1501 system has SELinux enabled, then the SELinux context under which the
1502 container should be run can be specified in the container
1503 configuration. The default is <command>unconfined_t</command>,
1504 which means that lxc will not attempt to change contexts.
1505 See @DATADIR@/lxc/selinux/lxc.te for an example policy and more
1511 <option>lxc.selinux.context</option>
1515 Specify the SELinux context under which the container should
1516 be run or <command>unconfined_t</command>. For example
1518 <programlisting>lxc.selinux.context = system_u:system_r:lxc_t:s0:c22</programlisting>
1525 <title>Seccomp configuration</title>
1527 A container can be started with a reduced set of available
1528 system calls by loading a seccomp profile at startup. The
1529 seccomp configuration file must begin with a version number
1530 on the first line, a policy type on the second line, followed
1531 by the configuration.
1534 Versions 1 and 2 are currently supported. In version 1, the
1535 policy is a simple whitelist. The second line therefore must
1536 read "whitelist", with the rest of the file containing one (numeric)
1537 sycall number per line. Each syscall number is whitelisted,
1538 while every unlisted number is blacklisted for use in the container
1542 In version 2, the policy may be blacklist or whitelist,
1543 supports per-rule and per-policy default actions, and supports
1544 per-architecture system call resolution from textual names.
1547 An example blacklist policy, in which all system calls are
1548 allowed except for mknod, which will simply do nothing and
1549 return 0 (success), looks like:
1561 <option>lxc.seccomp.profile</option>
1565 Specify a file containing the seccomp configuration to
1566 load before the container starts.
1574 <title>PR_SET_NO_NEW_PRIVS</title>
1576 With PR_SET_NO_NEW_PRIVS active execve() promises not to grant
1577 privileges to do anything that could not have been done without
1578 the execve() call (for example, rendering the set-user-ID and
1579 set-group-ID mode bits, and file capabilities non-functional).
1580 Once set, this bit cannot be unset. The setting of this bit is
1581 inherited by children created by fork() and clone(), and preserved
1583 Note that PR_SET_NO_NEW_PRIVS is applied after the container has
1584 changed into its intended AppArmor profile or SElinux context.
1589 <option>lxc.no_new_privs</option>
1593 Specify whether the PR_SET_NO_NEW_PRIVS flag should be set for the
1594 container. Set to 1 to activate.
1602 <title>UID mappings</title>
1604 A container can be started in a private user namespace with
1605 user and group id mappings. For instance, you can map userid
1606 0 in the container to userid 200000 on the host. The root
1607 user in the container will be privileged in the container,
1608 but unprivileged on the host. Normally a system container
1609 will want a range of ids, so you would map, for instance,
1610 user and group ids 0 through 20,000 in the container to the
1611 ids 200,000 through 220,000.
1616 <option>lxc.idmap</option>
1620 Four values must be provided. First a character, either
1621 'u', or 'g', to specify whether user or group ids are
1622 being mapped. Next is the first userid as seen in the
1623 user namespace of the container. Next is the userid as
1624 seen on the host. Finally, a range indicating the number
1625 of consecutive ids to map.
1633 <title>Container hooks</title>
1635 Container hooks are programs or scripts which can be executed
1636 at various times in a container's lifetime.
1639 When a container hook is executed, information is passed both
1640 as command line arguments and through environment variables.
1643 <listitem><para> Container name. </para></listitem>
1644 <listitem><para> Section (always 'lxc'). </para></listitem>
1645 <listitem><para> The hook type (i.e. 'clone' or 'pre-mount'). </para></listitem>
1646 <listitem><para> Additional arguments. In the
1647 case of the clone hook, any extra arguments passed to
1648 lxc-clone will appear as further arguments to the hook.
1649 In the case of the stop hook, paths to filedescriptors
1650 for each of the container's namespaces along with their types
1651 are passed. </para></listitem>
1653 The following environment variables are set:
1655 <listitem><para> LXC_NAME: is the container's name. </para></listitem>
1656 <listitem><para> LXC_ROOTFS_MOUNT: the path to the mounted root filesystem. </para></listitem>
1657 <listitem><para> LXC_CONFIG_FILE: the path to the container configuration file. </para></listitem>
1658 <listitem><para> LXC_SRC_NAME: in the case of the clone hook, this is the original container's name. </para></listitem>
1659 <listitem><para> LXC_ROOTFS_PATH: this is the lxc.rootfs.path 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>
1660 <listitem><para> LXC_CGNS_AWARE: indicated whether the container is cgroup namespace aware. </para></listitem>
1661 <listitem><para> LXC_LOG_LEVEL: the container's log level. </para></listitem>
1665 Standard output from the hooks is logged at debug level.
1666 Standard error is not logged, but can be captured by the
1667 hook redirecting its standard error to standard output.
1672 <option>lxc.hook.pre-start</option>
1676 A hook to be run in the host's namespace before the
1677 container ttys, consoles, or mounts are up.
1685 <option>lxc.hook.pre-mount</option>
1689 A hook to be run in the container's fs namespace but before
1690 the rootfs has been set up. This allows for manipulation
1691 of the rootfs, i.e. to mount an encrypted filesystem. Mounts
1692 done in this hook will not be reflected on the host (apart from
1693 mounts propagation), so they will be automatically cleaned up
1694 when the container shuts down.
1702 <option>lxc.hook.mount</option>
1706 A hook to be run in the container's namespace after
1707 mounting has been done, but before the pivot_root.
1715 <option>lxc.hook.autodev</option>
1719 A hook to be run in the container's namespace after
1720 mounting has been done and after any mount hooks have
1721 run, but before the pivot_root, if
1722 <option>lxc.autodev</option> == 1.
1723 The purpose of this hook is to assist in populating the
1724 /dev directory of the container when using the autodev
1725 option for systemd based containers. The container's /dev
1726 directory is relative to the
1727 ${<option>LXC_ROOTFS_MOUNT</option>} environment
1728 variable available when the hook is run.
1736 <option>lxc.hook.start-host</option>
1740 A hook to be run in the host's namespace after the
1741 container has been setup, and immediately before starting
1750 <option>lxc.hook.start</option>
1754 A hook to be run in the container's namespace immediately
1755 before executing the container's init. This requires the
1756 program to be available in the container.
1764 <option>lxc.hook.stop</option>
1768 A hook to be run in the host's namespace with references
1769 to the container's namespaces after the container has been shut
1770 down. For each namespace an extra argument is passed to the hook
1771 containing the namespace's type and a filename that can be used to
1772 obtain a file descriptor to the corresponding namespace, separated
1773 by a colon. The type is the name as it would appear in the
1774 <filename>/proc/PID/ns</filename> directory.
1775 For instance for the mount namespace the argument usually looks
1776 like <filename>mnt:/proc/PID/fd/12</filename>.
1784 <option>lxc.hook.post-stop</option>
1788 A hook to be run in the host's namespace after the
1789 container has been shut down.
1797 <option>lxc.hook.clone</option>
1801 A hook to be run when the container is cloned to a new one.
1802 See <citerefentry><refentrytitle><command>lxc-clone</command></refentrytitle>
1803 <manvolnum>1</manvolnum></citerefentry> for more information.
1811 <option>lxc.hook.destroy</option>
1815 A hook to be run when the container is destroyed.
1823 <title>Container hooks Environment Variables</title>
1825 A number of environment variables are made available to the startup
1826 hooks to provide configuration information and assist in the
1827 functioning of the hooks. Not all variables are valid in all
1828 contexts. In particular, all paths are relative to the host system
1829 and, as such, not valid during the <option>lxc.hook.start</option> hook.
1834 <option>LXC_NAME</option>
1838 The LXC name of the container. Useful for logging messages
1839 in common log environments. [<option>-n</option>]
1847 <option>LXC_CONFIG_FILE</option>
1851 Host relative path to the container configuration file. This
1852 gives the container to reference the original, top level,
1853 configuration file for the container in order to locate any
1854 additional configuration information not otherwise made
1855 available. [<option>-f</option>]
1863 <option>LXC_CONSOLE</option>
1867 The path to the console output of the container if not NULL.
1868 [<option>-c</option>] [<option>lxc.console.path</option>]
1876 <option>LXC_CONSOLE_LOGPATH</option>
1880 The path to the console log output of the container if not NULL.
1881 [<option>-L</option>]
1889 <option>LXC_ROOTFS_MOUNT</option>
1893 The mount location to which the container is initially bound.
1894 This will be the host relative path to the container rootfs
1895 for the container instance being started and is where changes
1896 should be made for that instance.
1897 [<option>lxc.rootfs.mount</option>]
1905 <option>LXC_ROOTFS_PATH</option>
1909 The host relative path to the container root which has been
1910 mounted to the rootfs.mount location.
1911 [<option>lxc.rootfs.path</option>]
1919 <option>LXC_SRC_NAME</option>
1923 Only for the clone hook. Is set to the original container name.
1931 <option>LXC_TARGET</option>
1935 Only for the stop hook. Is set to "stop" for a container
1936 shutdown or "reboot" for a container reboot.
1944 <option>LXC_CGNS_AWARE</option>
1948 If unset, then this version of lxc is not aware of cgroup
1949 namespaces. If set, it will be set to 1, and lxc is aware
1950 of cgroup namespaces. Note this does not guarantee that
1951 cgroup namespaces are enabled in the kernel. This is used
1952 by the lxcfs mount hook.
1959 <title>Logging</title>
1961 Logging can be configured on a per-container basis. By default,
1962 depending upon how the lxc package was compiled, container startup
1963 is logged only at the ERROR level, and logged to a file named after
1964 the container (with '.log' appended) either under the container path,
1968 Both the default log level and the log file can be specified in the
1969 container configuration file, overriding the default behavior. Note
1970 that the configuration file entries can in turn be overridden by the
1971 command line options to <command>lxc-start</command>.
1976 <option>lxc.log.level</option>
1980 The level at which to log. The log level is an integer in
1981 the range of 0..8 inclusive, where a lower number means more
1982 verbose debugging. In particular 0 = trace, 1 = debug, 2 =
1983 info, 3 = notice, 4 = warn, 5 = error, 6 = critical, 7 =
1984 alert, and 8 = fatal. If unspecified, the level defaults
1985 to 5 (error), so that only errors and above are logged.
1988 Note that when a script (such as either a hook script or a
1989 network interface up or down script) is called, the script's
1990 standard output is logged at level 1, debug.
1996 <option>lxc.log</option>
2000 The file to which logging info should be written.
2006 <option>lxc.log.syslog</option>
2010 Send logging info to syslog. It respects the log level defined in
2011 <command>lxc.log.level</command>. The argument should be the syslog
2012 facility to use, valid ones are: daemon, local0, local1, local2,
2013 local3, local4, local5, local5, local6, local7.
2021 <title>Autostart</title>
2023 The autostart options support marking which containers should be
2024 auto-started and in what order. These options may be used by LXC tools
2025 directly or by external tooling provided by the distributions.
2031 <option>lxc.start.auto</option>
2035 Whether the container should be auto-started.
2036 Valid values are 0 (off) and 1 (on).
2042 <option>lxc.start.delay</option>
2046 How long to wait (in seconds) after the container is
2047 started before starting the next one.
2053 <option>lxc.start.order</option>
2057 An integer used to sort the containers when auto-starting
2058 a series of containers at once.
2064 <option>lxc.monitor.unshare</option>
2068 If not zero the mount namespace will be unshared from the host
2069 before initializing the container (before running any pre-start
2070 hooks). This requires the CAP_SYS_ADMIN capability at startup.
2077 <option>lxc.group</option>
2081 A multi-value key (can be used multiple times) to put the
2082 container in a container group. Those groups can then be
2083 used (amongst other things) to start a series of related
2092 <title>Autostart and System Boot</title>
2094 Each container can be part of any number of groups or no group at all.
2095 Two groups are special. One is the NULL group, i.e. the container does
2096 not belong to any group. The other group is the "onboot" group.
2100 When the system boots with the LXC service enabled, it will first
2101 attempt to boot any containers with lxc.start.auto == 1 that is a member
2102 of the "onboot" group. The startup will be in order of lxc.start.order.
2103 If an lxc.start.delay has been specified, that delay will be honored
2104 before attempting to start the next container to give the current
2105 container time to begin initialization and reduce overloading the host
2106 system. After starting the members of the "onboot" group, the LXC system
2107 will proceed to boot containers with lxc.start.auto == 1 which are not
2108 members of any group (the NULL group) and proceed as with the onboot
2115 <title>Container Environment</title>
2117 If you want to pass environment variables into the container (that
2118 is, environment variables which will be available to init and all of
2119 its descendents), you can use <command>lxc.environment</command>
2120 parameters to do so. Be careful that you do not pass in anything
2121 sensitive; any process in the container which doesn't have its
2122 environment scrubbed will have these variables available to it, and
2123 environment variables are always available via
2124 <command>/proc/PID/environ</command>.
2128 This configuration parameter can be specified multiple times; once
2129 for each environment variable you wish to configure.
2135 <option>lxc.environment</option>
2139 Specify an environment variable to pass into the container.
2143 lxc.environment = APP_ENV=production
2144 lxc.environment = SYSLOG_SERVER=192.0.2.42
2154 <title>Examples</title>
2156 In addition to the few examples given below, you will find
2157 some other examples of configuration file in @DOCDIR@/examples
2160 <title>Network</title>
2161 <para>This configuration sets up a container to use a veth pair
2162 device with one side plugged to a bridge br0 (which has been
2163 configured before on the system by the administrator). The
2164 virtual network device visible in the container is renamed to
2167 lxc.uts.name = myhostname
2168 lxc.net.0.type = veth
2169 lxc.net.0.flags = up
2170 lxc.net.0.link = br0
2171 lxc.net.0.name = eth0
2172 lxc.net.0.hwaddr = 4a:49:43:49:79:bf
2173 lxc.net.0.ipv4.address = 10.2.3.5/24 10.2.3.255
2174 lxc.net.0.ipv6.address = 2003:db8:1:0:214:1234:fe0b:3597
2179 <title>UID/GID mapping</title>
2180 <para>This configuration will map both user and group ids in the
2181 range 0-9999 in the container to the ids 100000-109999 on the host.
2184 lxc.idmap = u 0 100000 10000
2185 lxc.idmap = g 0 100000 10000
2190 <title>Control group</title>
2191 <para>This configuration will setup several control groups for
2192 the application, cpuset.cpus restricts usage of the defined cpu,
2193 cpus.share prioritize the control group, devices.allow makes
2194 usable the specified devices.</para>
2196 lxc.cgroup.cpuset.cpus = 0,1
2197 lxc.cgroup.cpu.shares = 1234
2198 lxc.cgroup.devices.deny = a
2199 lxc.cgroup.devices.allow = c 1:3 rw
2200 lxc.cgroup.devices.allow = b 8:0 rw
2205 <title>Complex configuration</title>
2206 <para>This example show a complex configuration making a complex
2207 network stack, using the control groups, setting a new hostname,
2208 mounting some locations and a changing root file system.</para>
2210 lxc.uts.name = complex
2211 lxc.net.0.type = veth
2212 lxc.net.0.flags = up
2213 lxc.net.0.link = br0
2214 lxc.net.0.hwaddr = 4a:49:43:49:79:bf
2215 lxc.net.0.ipv4.address = 10.2.3.5/24 10.2.3.255
2216 lxc.net.0.ipv6.address = 2003:db8:1:0:214:1234:fe0b:3597
2217 lxc.net.0.ipv6.address = 2003:db8:1:0:214:5432:feab:3588
2218 lxc.net.1.type = macvlan
2219 lxc.net.1.flags = up
2220 lxc.net.1.link = eth0
2221 lxc.net.1.hwaddr = 4a:49:43:49:79:bd
2222 lxc.net.1.ipv4.address = 10.2.3.4/24
2223 lxc.net.1.ipv4.address = 192.168.10.125/24
2224 lxc.net.1.ipv6.address = 2003:db8:1:0:214:1234:fe0b:3596
2225 lxc.net.2.type = phys
2226 lxc.net.2.flags = up
2227 lxc.net.2.link = dummy0
2228 lxc.net.2.hwaddr = 4a:49:43:49:79:ff
2229 lxc.net.2.ipv4.address = 10.2.3.6/24
2230 lxc.net.2.ipv6.address = 2003:db8:1:0:214:1234:fe0b:3297
2231 lxc.cgroup.cpuset.cpus = 0,1
2232 lxc.cgroup.cpu.shares = 1234
2233 lxc.cgroup.devices.deny = a
2234 lxc.cgroup.devices.allow = c 1:3 rw
2235 lxc.cgroup.devices.allow = b 8:0 rw
2236 lxc.mount.fstab = /etc/fstab.complex
2237 lxc.mount.entry = /lib /root/myrootfs/lib none ro,bind 0 0
2238 lxc.rootfs.path = dir:/mnt/rootfs.complex
2239 lxc.cap.drop = sys_module mknod setuid net_raw
2240 lxc.cap.drop = mac_override
2247 <title>See Also</title>
2250 <refentrytitle><command>chroot</command></refentrytitle>
2251 <manvolnum>1</manvolnum>
2255 <refentrytitle><command>pivot_root</command></refentrytitle>
2256 <manvolnum>8</manvolnum>
2260 <refentrytitle><filename>fstab</filename></refentrytitle>
2261 <manvolnum>5</manvolnum>
2265 <refentrytitle><filename>capabilities</filename></refentrytitle>
2266 <manvolnum>7</manvolnum>
2274 <title>Author</title>
2275 <para>Daniel Lezcano <email>daniel.lezcano@free.fr</email></para>
2280 <!-- Keep this comment at the end of the file
2285 sgml-minimize-attributes:nil
2286 sgml-always-quote-attributes:t
2289 sgml-parent-document:nil
2290 sgml-default-dtd-file:nil
2291 sgml-exposed-tags:nil
2292 sgml-local-catalogs:nil
2293 sgml-local-ecat-files:nil