10 pct - Tool to manage Linux Containers (LXC) on Proxmox VE
16 include::pct.1-synopsis.adoc[]
23 Proxmox Container Toolkit
24 =========================
28 :title: Linux Container
31 Containers are a lightweight alternative to fully virtualized
32 VMs. Instead of emulating a complete Operating System (OS), containers
33 simply use the OS of the host they run on. This implies that all
34 containers use the same kernel, and that they can access resources
35 from the host directly.
37 This is great because containers do not waste CPU power nor memory due
38 to kernel emulation. Container run-time costs are close to zero and
39 usually negligible. But there are also some drawbacks you need to
42 * You can only run Linux based OS inside containers, i.e. it is not
43 possible to run FreeBSD or MS Windows inside.
45 * For security reasons, access to host resources needs to be
46 restricted. This is done with AppArmor, SecComp filters and other
47 kernel features. Be prepared that some syscalls are not allowed
50 {pve} uses https://linuxcontainers.org/[LXC] as underlying container
51 technology. We consider LXC as low-level library, which provides
52 countless options. It would be too difficult to use those tools
53 directly. Instead, we provide a small wrapper called `pct`, the
54 "Proxmox Container Toolkit".
56 The toolkit is tightly coupled with {pve}. That means that it is aware
57 of the cluster setup, and it can use the same network and storage
58 resources as fully virtualized VMs. You can even use the {pve}
59 firewall, or manage containers using the HA framework.
61 Our primary goal is to offer an environment as one would get from a
62 VM, but without the additional overhead. We call this "System
65 NOTE: If you want to run micro-containers (with docker, rkt, ...), it
66 is best to run them inside a VM.
72 * LXC (https://linuxcontainers.org/)
74 * Integrated into {pve} graphical user interface (GUI)
76 * Easy to use command line tool `pct`
78 * Access via {pve} REST API
80 * lxcfs to provide containerized /proc file system
82 * AppArmor/Seccomp to improve security
84 * CRIU: for live migration (planned)
86 * Use latest available kernels (4.4.X)
88 * Image based deployment (templates)
90 * Use {pve} storage library
92 * Container setup from host (network, DNS, storage, ...)
95 Security Considerations
96 -----------------------
98 Containers use the same kernel as the host, so there is a big attack
99 surface for malicious users. You should consider this fact if you
100 provide containers to totally untrusted people. In general, fully
101 virtualized VMs provide better isolation.
103 The good news is that LXC uses many kernel security features like
104 AppArmor, CGroups and PID and user namespaces, which makes containers
107 Guest Operating System Configuration
108 ------------------------------------
110 We normally try to detect the operating system type inside the
111 container, and then modify some files inside the container to make
112 them work as expected. Here is a short list of things we do at
115 set /etc/hostname:: to set the container name
117 modify /etc/hosts:: to allow lookup of the local hostname
119 network setup:: pass the complete network setup to the container
121 configure DNS:: pass information about DNS servers
123 adapt the init system:: for example, fix the number of spawned getty processes
125 set the root password:: when creating a new container
127 rewrite ssh_host_keys:: so that each container has unique keys
129 randomize crontab:: so that cron does not start at the same time on all containers
131 Changes made by {PVE} are enclosed by comment markers:
139 Those markers will be inserted at a reasonable location in the
140 file. If such a section already exists, it will be updated in place
141 and will not be moved.
143 Modification of a file can be prevented by adding a `.pve-ignore.`
144 file for it. For instance, if the file `/etc/.pve-ignore.hosts`
145 exists then the `/etc/hosts` file will not be touched. This can be a
146 simple empty file created via:
148 # touch /etc/.pve-ignore.hosts
150 Most modifications are OS dependent, so they differ between different
151 distributions and versions. You can completely disable modifications
152 by manually setting the `ostype` to `unmanaged`.
154 OS type detection is done by testing for certain files inside the
157 Ubuntu:: inspect /etc/lsb-release (`DISTRIB_ID=Ubuntu`)
159 Debian:: test /etc/debian_version
161 Fedora:: test /etc/fedora-release
163 RedHat or CentOS:: test /etc/redhat-release
165 ArchLinux:: test /etc/arch-release
167 Alpine:: test /etc/alpine-release
169 Gentoo:: test /etc/gentoo-release
171 NOTE: Container start fails if the configured `ostype` differs from the auto
175 [[pct_container_images]]
179 Container images, sometimes also referred to as ``templates'' or
180 ``appliances'', are `tar` archives which contain everything to run a
181 container. You can think of it as a tidy container backup. Like most
182 modern container toolkits, `pct` uses those images when you create a
183 new container, for example:
185 pct create 999 local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz
187 {pve} itself ships a set of basic templates for most common
188 operating systems, and you can download them using the `pveam` (short
189 for {pve} Appliance Manager) command line utility. You can also
190 download https://www.turnkeylinux.org/[TurnKey Linux] containers using
191 that tool (or the graphical user interface).
193 Our image repositories contain a list of available images, and there
194 is a cron job run each day to download that list. You can trigger that
195 update manually with:
199 After that you can view the list of available images using:
203 You can restrict this large list by specifying the `section` you are
204 interested in, for example basic `system` images:
206 .List available system images
208 # pveam available --section system
209 system archlinux-base_2015-24-29-1_x86_64.tar.gz
210 system centos-7-default_20160205_amd64.tar.xz
211 system debian-6.0-standard_6.0-7_amd64.tar.gz
212 system debian-7.0-standard_7.0-3_amd64.tar.gz
213 system debian-8.0-standard_8.0-1_amd64.tar.gz
214 system ubuntu-12.04-standard_12.04-1_amd64.tar.gz
215 system ubuntu-14.04-standard_14.04-1_amd64.tar.gz
216 system ubuntu-15.04-standard_15.04-1_amd64.tar.gz
217 system ubuntu-15.10-standard_15.10-1_amd64.tar.gz
220 Before you can use such a template, you need to download them into one
221 of your storages. You can simply use storage `local` for that
222 purpose. For clustered installations, it is preferred to use a shared
223 storage so that all nodes can access those images.
225 pveam download local debian-8.0-standard_8.0-1_amd64.tar.gz
227 You are now ready to create containers using that image, and you can
228 list all downloaded images on storage `local` with:
232 local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz 190.20MB
235 The above command shows you the full {pve} volume identifiers. They include
236 the storage name, and most other {pve} commands can use them. For
237 example you can delete that image later with:
239 pveam remove local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz
242 [[pct_container_storage]]
246 Traditional containers use a very simple storage model, only allowing
247 a single mount point, the root file system. This was further
248 restricted to specific file system types like `ext4` and `nfs`.
249 Additional mounts are often done by user provided scripts. This turned
250 out to be complex and error prone, so we try to avoid that now.
252 Our new LXC based container model is more flexible regarding
253 storage. First, you can have more than a single mount point. This
254 allows you to choose a suitable storage for each application. For
255 example, you can use a relatively slow (and thus cheap) storage for
256 the container root file system. Then you can use a second mount point
257 to mount a very fast, distributed storage for your database
258 application. See section <<pct_mount_points,Mount Points>> for further
261 The second big improvement is that you can use any storage type
262 supported by the {pve} storage library. That means that you can store
263 your containers on local `lvmthin` or `zfs`, shared `iSCSI` storage,
264 or even on distributed storage systems like `ceph`. It also enables us
265 to use advanced storage features like snapshots and clones. `vzdump`
266 can also use the snapshot feature to provide consistent container
269 Last but not least, you can also mount local devices directly, or
270 mount local directories using bind mounts. That way you can access
271 local storage inside containers with zero overhead. Such bind mounts
272 also provide an easy way to share data between different containers.
278 WARNING: Because of existing issues in the Linux kernel's freezer
279 subsystem the usage of FUSE mounts inside a container is strongly
280 advised against, as containers need to be frozen for suspend or
281 snapshot mode backups.
283 If FUSE mounts cannot be replaced by other mounting mechanisms or storage
284 technologies, it is possible to establish the FUSE mount on the Proxmox host
285 and use a bind mount point to make it accessible inside the container.
288 Using Quotas Inside Containers
289 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
291 Quotas allow to set limits inside a container for the amount of disk
292 space that each user can use. This only works on ext4 image based
293 storage types and currently does not work with unprivileged
296 Activating the `quota` option causes the following mount options to be
297 used for a mount point:
298 `usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0`
300 This allows quotas to be used like you would on any other system. You
301 can initialize the `/aquota.user` and `/aquota.group` files by running
308 and edit the quotas via the `edquota` command. Refer to the documentation
309 of the distribution running inside the container for details.
311 NOTE: You need to run the above commands for every mount point by passing
312 the mount point's path instead of just `/`.
315 Using ACLs Inside Containers
316 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
318 The standard Posix **A**ccess **C**ontrol **L**ists are also available inside containers.
319 ACLs allow you to set more detailed file ownership than the traditional user/
323 Backup of Containers mount points
324 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
326 By default additional mount points besides the Root Disk mount point are not
327 included in backups. You can reverse this default behavior by setting the
328 *Backup* option on a mount point.
329 // see PVE::VZDump::LXC::prepare()
331 Replication of Containers mount points
332 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
334 By default additional mount points are replicated when the Root Disk
335 is replicated. If you want the {pve} storage replication mechanism to skip a
336 mount point when starting a replication job, you can set the
337 *Skip replication* option on that mount point. +
338 As of {pve} 5.0, replication requires a storage of type `zfspool`, so adding a
339 mount point to a different type of storage when the container has replication
340 configured requires to *Skip replication* for that mount point.
351 [thumbnail="screenshot/gui-create-ct-general.png"]
353 General settings of a container include
355 * the *Node* : the physical server on which the container will run
356 * the *CT ID*: a unique number in this {pve} installation used to identify your container
357 * *Hostname*: the hostname of the container
358 * *Resource Pool*: a logical group of containers and VMs
359 * *Password*: the root password of the container
360 * *SSH Public Key*: a public key for connecting to the root account over SSH
361 * *Unprivileged container*: this option allows to choose at creation time
362 if you want to create a privileged or unprivileged container.
365 Privileged Containers
366 ^^^^^^^^^^^^^^^^^^^^^
368 Security is done by dropping capabilities, using mandatory access
369 control (AppArmor), SecComp filters and namespaces. The LXC team
370 considers this kind of container as unsafe, and they will not consider
371 new container escape exploits to be security issues worthy of a CVE
372 and quick fix. So you should use this kind of containers only inside a
373 trusted environment, or when no untrusted task is running as root in
377 Unprivileged Containers
378 ^^^^^^^^^^^^^^^^^^^^^^^
380 This kind of containers use a new kernel feature called user
381 namespaces. The root UID 0 inside the container is mapped to an
382 unprivileged user outside the container. This means that most security
383 issues (container escape, resource abuse, ...) in those containers
384 will affect a random unprivileged user, and so would be a generic
385 kernel security bug rather than an LXC issue. The LXC team thinks
386 unprivileged containers are safe by design.
388 NOTE: If the container uses systemd as an init system, please be
389 aware the systemd version running inside the container should be equal
396 [thumbnail="screenshot/gui-create-ct-cpu.png"]
398 You can restrict the number of visible CPUs inside the container using
399 the `cores` option. This is implemented using the Linux 'cpuset'
400 cgroup (**c**ontrol *group*). A special task inside `pvestatd` tries
401 to distribute running containers among available CPUs. You can view
402 the assigned CPUs using the following command:
406 ---------------------
410 ---------------------
413 Containers use the host kernel directly, so all task inside a
414 container are handled by the host CPU scheduler. {pve} uses the Linux
415 'CFS' (**C**ompletely **F**air **S**cheduler) scheduler by default,
416 which has additional bandwidth control options.
420 `cpulimit`: :: You can use this option to further limit assigned CPU
421 time. Please note that this is a floating point number, so it is
422 perfectly valid to assign two cores to a container, but restrict
423 overall CPU consumption to half a core.
430 `cpuunits`: :: This is a relative weight passed to the kernel
431 scheduler. The larger the number is, the more CPU time this container
432 gets. Number is relative to the weights of all the other running
433 containers. The default is 1024. You can use this setting to
434 prioritize some containers.
441 [thumbnail="screenshot/gui-create-ct-memory.png"]
443 Container memory is controlled using the cgroup memory controller.
447 `memory`: :: Limit overall memory usage. This corresponds
448 to the `memory.limit_in_bytes` cgroup setting.
450 `swap`: :: Allows the container to use additional swap memory from the
451 host swap space. This corresponds to the `memory.memsw.limit_in_bytes`
452 cgroup setting, which is set to the sum of both value (`memory +
460 [thumbnail="screenshot/gui-create-ct-root-disk.png"]
462 The root mount point is configured with the `rootfs` property, and you can
463 configure up to 10 additional mount points. The corresponding options
464 are called `mp0` to `mp9`, and they can contain the following setting:
466 include::pct-mountpoint-opts.adoc[]
468 Currently there are basically three types of mount points: storage backed
469 mount points, bind mounts and device mounts.
471 .Typical container `rootfs` configuration
473 rootfs: thin1:base-100-disk-1,size=8G
477 Storage Backed Mount Points
478 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
480 Storage backed mount points are managed by the {pve} storage subsystem and come
481 in three different flavors:
483 - Image based: these are raw images containing a single ext4 formatted file
485 - ZFS subvolumes: these are technically bind mounts, but with managed storage,
486 and thus allow resizing and snapshotting.
487 - Directories: passing `size=0` triggers a special case where instead of a raw
488 image a directory is created.
490 NOTE: The special option syntax `STORAGE_ID:SIZE_IN_GB` for storage backed
491 mount point volumes will automatically allocate a volume of the specified size
492 on the specified storage. E.g., calling
493 `pct set 100 -mp0 thin1:10,mp=/path/in/container` will allocate a 10GB volume
494 on the storage `thin1` and replace the volume ID place holder `10` with the
501 Bind mounts allow you to access arbitrary directories from your Proxmox VE host
502 inside a container. Some potential use cases are:
504 - Accessing your home directory in the guest
505 - Accessing an USB device directory in the guest
506 - Accessing an NFS mount from the host in the guest
508 Bind mounts are considered to not be managed by the storage subsystem, so you
509 cannot make snapshots or deal with quotas from inside the container. With
510 unprivileged containers you might run into permission problems caused by the
511 user mapping and cannot use ACLs.
513 NOTE: The contents of bind mount points are not backed up when using `vzdump`.
515 WARNING: For security reasons, bind mounts should only be established
516 using source directories especially reserved for this purpose, e.g., a
517 directory hierarchy under `/mnt/bindmounts`. Never bind mount system
518 directories like `/`, `/var` or `/etc` into a container - this poses a
521 NOTE: The bind mount source path must not contain any symlinks.
523 For example, to make the directory `/mnt/bindmounts/shared` accessible in the
524 container with ID `100` under the path `/shared`, use a configuration line like
525 `mp0: /mnt/bindmounts/shared,mp=/shared` in `/etc/pve/lxc/100.conf`.
526 Alternatively, use `pct set 100 -mp0 /mnt/bindmounts/shared,mp=/shared` to
527 achieve the same result.
533 Device mount points allow to mount block devices of the host directly into the
534 container. Similar to bind mounts, device mounts are not managed by {PVE}'s
535 storage subsystem, but the `quota` and `acl` options will be honored.
537 NOTE: Device mount points should only be used under special circumstances. In
538 most cases a storage backed mount point offers the same performance and a lot
541 NOTE: The contents of device mount points are not backed up when using `vzdump`.
544 [[pct_container_network]]
548 [thumbnail="screenshot/gui-create-ct-network.png"]
550 You can configure up to 10 network interfaces for a single
551 container. The corresponding options are called `net0` to `net9`, and
552 they can contain the following setting:
554 include::pct-network-opts.adoc[]
557 [[pct_startup_and_shutdown]]
558 Automatic Start and Shutdown of Containers
559 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
561 After creating your containers, you probably want them to start automatically
562 when the host system boots. For this you need to select the option 'Start at
563 boot' from the 'Options' Tab of your container in the web interface, or set it with
564 the following command:
566 pct set <ctid> -onboot 1
568 .Start and Shutdown Order
569 // use the screenshot from qemu - its the same
570 [thumbnail="screenshot/gui-qemu-edit-start-order.png"]
572 If you want to fine tune the boot order of your containers, you can use the following
575 * *Start/Shutdown order*: Defines the start order priority. E.g. set it to 1 if
576 you want the CT to be the first to be started. (We use the reverse startup
577 order for shutdown, so a container with a start order of 1 would be the last to
579 * *Startup delay*: Defines the interval between this container start and subsequent
580 containers starts . E.g. set it to 240 if you want to wait 240 seconds before starting
582 * *Shutdown timeout*: Defines the duration in seconds {pve} should wait
583 for the container to be offline after issuing a shutdown command.
584 By default this value is set to 60, which means that {pve} will issue a
585 shutdown request, wait 60s for the machine to be offline, and if after 60s
586 the machine is still online will notify that the shutdown action failed.
588 Please note that containers without a Start/Shutdown order parameter will always
589 start after those where the parameter is set, and this parameter only
590 makes sense between the machines running locally on a host, and not
601 It is possible to use the `vzdump` tool for container backup. Please
602 refer to the `vzdump` manual page for details.
605 Restoring Container Backups
606 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
608 Restoring container backups made with `vzdump` is possible using the
609 `pct restore` command. By default, `pct restore` will attempt to restore as much
610 of the backed up container configuration as possible. It is possible to override
611 the backed up configuration by manually setting container options on the command
612 line (see the `pct` manual page for details).
614 NOTE: `pvesm extractconfig` can be used to view the backed up configuration
615 contained in a vzdump archive.
617 There are two basic restore modes, only differing by their handling of mount
621 ``Simple'' Restore Mode
622 ^^^^^^^^^^^^^^^^^^^^^^^
624 If neither the `rootfs` parameter nor any of the optional `mpX` parameters
625 are explicitly set, the mount point configuration from the backed up
626 configuration file is restored using the following steps:
628 . Extract mount points and their options from backup
629 . Create volumes for storage backed mount points (on storage provided with the
630 `storage` parameter, or default local storage if unset)
631 . Extract files from backup archive
632 . Add bind and device mount points to restored configuration (limited to root user)
634 NOTE: Since bind and device mount points are never backed up, no files are
635 restored in the last step, but only the configuration options. The assumption
636 is that such mount points are either backed up with another mechanism (e.g.,
637 NFS space that is bind mounted into many containers), or not intended to be
640 This simple mode is also used by the container restore operations in the web
644 ``Advanced'' Restore Mode
645 ^^^^^^^^^^^^^^^^^^^^^^^^^
647 By setting the `rootfs` parameter (and optionally, any combination of `mpX`
648 parameters), the `pct restore` command is automatically switched into an
649 advanced mode. This advanced mode completely ignores the `rootfs` and `mpX`
650 configuration options contained in the backup archive, and instead only
651 uses the options explicitly provided as parameters.
653 This mode allows flexible configuration of mount point settings at restore time,
656 * Set target storages, volume sizes and other options for each mount point
658 * Redistribute backed up files according to new mount point scheme
659 * Restore to device and/or bind mount points (limited to root user)
662 Managing Containers with `pct`
663 ------------------------------
665 `pct` is the tool to manage Linux Containers on {pve}. You can create
666 and destroy containers, and control execution (start, stop, migrate,
667 ...). You can use pct to set parameters in the associated config file,
668 like network configuration or memory limits.
674 Create a container based on a Debian template (provided you have
675 already downloaded the template via the web interface)
677 pct create 100 /var/lib/vz/template/cache/debian-8.0-standard_8.0-1_amd64.tar.gz
683 Start a login session via getty
687 Enter the LXC namespace and run a shell as root user
691 Display the configuration
695 Add a network interface called `eth0`, bridged to the host bridge `vmbr0`,
696 set the address and gateway, while it's running
698 pct set 100 -net0 name=eth0,bridge=vmbr0,ip=192.168.15.147/24,gw=192.168.15.1
700 Reduce the memory of the container to 512MB
702 pct set 100 -memory 512
705 Obtaining Debugging Logs
706 ~~~~~~~~~~~~~~~~~~~~~~~~
708 In case `pct start` is unable to start a specific container, it might be
709 helpful to collect debugging output by running `lxc-start` (replace `ID` with
712 lxc-start -n ID -F -l DEBUG -o /tmp/lxc-ID.log
714 This command will attempt to start the container in foreground mode, to stop the container run `pct shutdown ID` or `pct stop ID` in a second terminal.
716 The collected debug log is written to `/tmp/lxc-ID.log`.
718 NOTE: If you have changed the container's configuration since the last start
719 attempt with `pct start`, you need to run `pct start` at least once to also
720 update the configuration used by `lxc-start`.
726 If you have a cluster, you can migrate your Containers with
728 pct migrate <vmid> <target>
730 This works as long as your Container is offline. If it has local volumes or
731 mountpoints defined, the migration will copy the content over the network to
732 the target host if there is the same storage defined.
734 If you want to migrate online Containers, the only way is to use
735 restart migration. This can be initiated with the -restart flag and the optional
738 A restart migration will shut down the Container and kill it after the specified
739 timeout (the default is 180 seconds). Then it will migrate the Container
740 like an offline migration and when finished, it starts the Container on the
743 [[pct_configuration]]
747 The `/etc/pve/lxc/<CTID>.conf` file stores container configuration,
748 where `<CTID>` is the numeric ID of the given container. Like all
749 other files stored inside `/etc/pve/`, they get automatically
750 replicated to all other cluster nodes.
752 NOTE: CTIDs < 100 are reserved for internal purposes, and CTIDs need to be
755 .Example Container Configuration
762 net0: bridge=vmbr0,hwaddr=66:64:66:64:64:36,ip=dhcp,name=eth0,type=veth
763 rootfs: local:107/vm-107-disk-1.raw,size=7G
766 Those configuration files are simple text files, and you can edit them
767 using a normal text editor (`vi`, `nano`, ...). This is sometimes
768 useful to do small corrections, but keep in mind that you need to
769 restart the container to apply such changes.
771 For that reason, it is usually better to use the `pct` command to
772 generate and modify those files, or do the whole thing using the GUI.
773 Our toolkit is smart enough to instantaneously apply most changes to
774 running containers. This feature is called "hot plug", and there is no
775 need to restart the container in that case.
781 Container configuration files use a simple colon separated key/value
782 format. Each line has the following format:
789 Blank lines in those files are ignored, and lines starting with a `#`
790 character are treated as comments and are also ignored.
792 It is possible to add low-level, LXC style configuration directly, for
795 lxc.init_cmd: /sbin/my_own_init
799 lxc.init_cmd = /sbin/my_own_init
801 Those settings are directly passed to the LXC low-level tools.
808 When you create a snapshot, `pct` stores the configuration at snapshot
809 time into a separate snapshot section within the same configuration
810 file. For example, after creating a snapshot called ``testsnapshot'',
811 your configuration file will look like this:
813 .Container configuration with snapshot
827 There are a few snapshot related properties like `parent` and
828 `snaptime`. The `parent` property is used to store the parent/child
829 relationship between snapshots. `snaptime` is the snapshot creation
830 time stamp (Unix epoch).
837 include::pct.conf.5-opts.adoc[]
843 Container migrations, snapshots and backups (`vzdump`) set a lock to
844 prevent incompatible concurrent actions on the affected container. Sometimes
845 you need to remove such a lock manually (e.g., after a power failure).
849 CAUTION: Only do that if you are sure the action which set the lock is
858 `/etc/pve/lxc/<CTID>.conf`::
860 Configuration file for the container '<CTID>'.
863 include::pve-copyright.adoc[]