]> git.proxmox.com Git - pve-docs.git/blame_incremental - pct.adoc
add correct wiki titles
[pve-docs.git] / pct.adoc
... / ...
CommitLineData
1ifdef::manvolnum[]
2PVE({manvolnum})
3================
4include::attributes.txt[]
5
6:pve-toplevel:
7
8NAME
9----
10
11pct - Tool to manage Linux Containers (LXC) on Proxmox VE
12
13
14SYNOPSIS
15--------
16
17include::pct.1-synopsis.adoc[]
18
19DESCRIPTION
20-----------
21endif::manvolnum[]
22
23ifndef::manvolnum[]
24Proxmox Container Toolkit
25=========================
26include::attributes.txt[]
27endif::manvolnum[]
28
29ifdef::wiki[]
30:pve-toplevel:
31:title: Linux Container
32endif::wiki[]
33
34Containers are a lightweight alternative to fully virtualized
35VMs. Instead of emulating a complete Operating System (OS), containers
36simply use the OS of the host they run on. This implies that all
37containers use the same kernel, and that they can access resources
38from the host directly.
39
40This is great because containers do not waste CPU power nor memory due
41to kernel emulation. Container run-time costs are close to zero and
42usually negligible. But there are also some drawbacks you need to
43consider:
44
45* You can only run Linux based OS inside containers, i.e. it is not
46 possible to run FreeBSD or MS Windows inside.
47
48* For security reasons, access to host resources needs to be
49 restricted. This is done with AppArmor, SecComp filters and other
50 kernel features. Be prepared that some syscalls are not allowed
51 inside containers.
52
53{pve} uses https://linuxcontainers.org/[LXC] as underlying container
54technology. We consider LXC as low-level library, which provides
55countless options. It would be too difficult to use those tools
56directly. Instead, we provide a small wrapper called `pct`, the
57"Proxmox Container Toolkit".
58
59The toolkit is tightly coupled with {pve}. That means that it is aware
60of the cluster setup, and it can use the same network and storage
61resources as fully virtualized VMs. You can even use the {pve}
62firewall, or manage containers using the HA framework.
63
64Our primary goal is to offer an environment as one would get from a
65VM, but without the additional overhead. We call this "System
66Containers".
67
68NOTE: If you want to run micro-containers (with docker, rkt, ...), it
69is best to run them inside a VM.
70
71
72Security Considerations
73-----------------------
74
75Containers use the same kernel as the host, so there is a big attack
76surface for malicious users. You should consider this fact if you
77provide containers to totally untrusted people. In general, fully
78virtualized VMs provide better isolation.
79
80The good news is that LXC uses many kernel security features like
81AppArmor, CGroups and PID and user namespaces, which makes containers
82usage quite secure. We distinguish two types of containers:
83
84
85Privileged Containers
86~~~~~~~~~~~~~~~~~~~~~
87
88Security is done by dropping capabilities, using mandatory access
89control (AppArmor), SecComp filters and namespaces. The LXC team
90considers this kind of container as unsafe, and they will not consider
91new container escape exploits to be security issues worthy of a CVE
92and quick fix. So you should use this kind of containers only inside a
93trusted environment, or when no untrusted task is running as root in
94the container.
95
96
97Unprivileged Containers
98~~~~~~~~~~~~~~~~~~~~~~~
99
100This kind of containers use a new kernel feature called user
101namespaces. The root UID 0 inside the container is mapped to an
102unprivileged user outside the container. This means that most security
103issues (container escape, resource abuse, ...) in those containers
104will affect a random unprivileged user, and so would be a generic
105kernel security bug rather than an LXC issue. The LXC team thinks
106unprivileged containers are safe by design.
107
108
109Configuration
110-------------
111
112The `/etc/pve/lxc/<CTID>.conf` file stores container configuration,
113where `<CTID>` is the numeric ID of the given container. Like all
114other files stored inside `/etc/pve/`, they get automatically
115replicated to all other cluster nodes.
116
117NOTE: CTIDs < 100 are reserved for internal purposes, and CTIDs need to be
118unique cluster wide.
119
120.Example Container Configuration
121----
122ostype: debian
123arch: amd64
124hostname: www
125memory: 512
126swap: 512
127net0: bridge=vmbr0,hwaddr=66:64:66:64:64:36,ip=dhcp,name=eth0,type=veth
128rootfs: local:107/vm-107-disk-1.raw,size=7G
129----
130
131Those configuration files are simple text files, and you can edit them
132using a normal text editor (`vi`, `nano`, ...). This is sometimes
133useful to do small corrections, but keep in mind that you need to
134restart the container to apply such changes.
135
136For that reason, it is usually better to use the `pct` command to
137generate and modify those files, or do the whole thing using the GUI.
138Our toolkit is smart enough to instantaneously apply most changes to
139running containers. This feature is called "hot plug", and there is no
140need to restart the container in that case.
141
142
143File Format
144~~~~~~~~~~~
145
146Container configuration files use a simple colon separated key/value
147format. Each line has the following format:
148
149-----
150# this is a comment
151OPTION: value
152-----
153
154Blank lines in those files are ignored, and lines starting with a `#`
155character are treated as comments and are also ignored.
156
157It is possible to add low-level, LXC style configuration directly, for
158example:
159
160 lxc.init_cmd: /sbin/my_own_init
161
162or
163
164 lxc.init_cmd = /sbin/my_own_init
165
166Those settings are directly passed to the LXC low-level tools.
167
168
169Snapshots
170~~~~~~~~~
171
172When you create a snapshot, `pct` stores the configuration at snapshot
173time into a separate snapshot section within the same configuration
174file. For example, after creating a snapshot called ``testsnapshot'',
175your configuration file will look like this:
176
177.Container configuration with snapshot
178----
179memory: 512
180swap: 512
181parent: testsnaphot
182...
183
184[testsnaphot]
185memory: 512
186swap: 512
187snaptime: 1457170803
188...
189----
190
191There are a few snapshot related properties like `parent` and
192`snaptime`. The `parent` property is used to store the parent/child
193relationship between snapshots. `snaptime` is the snapshot creation
194time stamp (Unix epoch).
195
196
197Guest Operating System Configuration
198~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
199
200We normally try to detect the operating system type inside the
201container, and then modify some files inside the container to make
202them work as expected. Here is a short list of things we do at
203container startup:
204
205set /etc/hostname:: to set the container name
206
207modify /etc/hosts:: to allow lookup of the local hostname
208
209network setup:: pass the complete network setup to the container
210
211configure DNS:: pass information about DNS servers
212
213adapt the init system:: for example, fix the number of spawned getty processes
214
215set the root password:: when creating a new container
216
217rewrite ssh_host_keys:: so that each container has unique keys
218
219randomize crontab:: so that cron does not start at the same time on all containers
220
221Changes made by {PVE} are enclosed by comment markers:
222
223----
224# --- BEGIN PVE ---
225<data>
226# --- END PVE ---
227----
228
229Those markers will be inserted at a reasonable location in the
230file. If such a section already exists, it will be updated in place
231and will not be moved.
232
233Modification of a file can be prevented by adding a `.pve-ignore.`
234file for it. For instance, if the file `/etc/.pve-ignore.hosts`
235exists then the `/etc/hosts` file will not be touched. This can be a
236simple empty file creatd via:
237
238 # touch /etc/.pve-ignore.hosts
239
240Most modifications are OS dependent, so they differ between different
241distributions and versions. You can completely disable modifications
242by manually setting the `ostype` to `unmanaged`.
243
244OS type detection is done by testing for certain files inside the
245container:
246
247Ubuntu:: inspect /etc/lsb-release (`DISTRIB_ID=Ubuntu`)
248
249Debian:: test /etc/debian_version
250
251Fedora:: test /etc/fedora-release
252
253RedHat or CentOS:: test /etc/redhat-release
254
255ArchLinux:: test /etc/arch-release
256
257Alpine:: test /etc/alpine-release
258
259Gentoo:: test /etc/gentoo-release
260
261NOTE: Container start fails if the configured `ostype` differs from the auto
262detected type.
263
264
265Options
266~~~~~~~
267
268include::pct.conf.5-opts.adoc[]
269
270
271Container Images
272----------------
273
274Container images, sometimes also referred to as ``templates'' or
275``appliances'', are `tar` archives which contain everything to run a
276container. You can think of it as a tidy container backup. Like most
277modern container toolkits, `pct` uses those images when you create a
278new container, for example:
279
280 pct create 999 local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz
281
282{pve} itself ships a set of basic templates for most common
283operating systems, and you can download them using the `pveam` (short
284for {pve} Appliance Manager) command line utility. You can also
285download https://www.turnkeylinux.org/[TurnKey Linux] containers using
286that tool (or the graphical user interface).
287
288Our image repositories contain a list of available images, and there
289is a cron job run each day to download that list. You can trigger that
290update manually with:
291
292 pveam update
293
294After that you can view the list of available images using:
295
296 pveam available
297
298You can restrict this large list by specifying the `section` you are
299interested in, for example basic `system` images:
300
301.List available system images
302----
303# pveam available --section system
304system archlinux-base_2015-24-29-1_x86_64.tar.gz
305system centos-7-default_20160205_amd64.tar.xz
306system debian-6.0-standard_6.0-7_amd64.tar.gz
307system debian-7.0-standard_7.0-3_amd64.tar.gz
308system debian-8.0-standard_8.0-1_amd64.tar.gz
309system ubuntu-12.04-standard_12.04-1_amd64.tar.gz
310system ubuntu-14.04-standard_14.04-1_amd64.tar.gz
311system ubuntu-15.04-standard_15.04-1_amd64.tar.gz
312system ubuntu-15.10-standard_15.10-1_amd64.tar.gz
313----
314
315Before you can use such a template, you need to download them into one
316of your storages. You can simply use storage `local` for that
317purpose. For clustered installations, it is preferred to use a shared
318storage so that all nodes can access those images.
319
320 pveam download local debian-8.0-standard_8.0-1_amd64.tar.gz
321
322You are now ready to create containers using that image, and you can
323list all downloaded images on storage `local` with:
324
325----
326# pveam list local
327local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz 190.20MB
328----
329
330The above command shows you the full {pve} volume identifiers. They include
331the storage name, and most other {pve} commands can use them. For
332example you can delete that image later with:
333
334 pveam remove local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz
335
336
337Container Storage
338-----------------
339
340Traditional containers use a very simple storage model, only allowing
341a single mount point, the root file system. This was further
342restricted to specific file system types like `ext4` and `nfs`.
343Additional mounts are often done by user provided scripts. This turned
344out to be complex and error prone, so we try to avoid that now.
345
346Our new LXC based container model is more flexible regarding
347storage. First, you can have more than a single mount point. This
348allows you to choose a suitable storage for each application. For
349example, you can use a relatively slow (and thus cheap) storage for
350the container root file system. Then you can use a second mount point
351to mount a very fast, distributed storage for your database
352application.
353
354The second big improvement is that you can use any storage type
355supported by the {pve} storage library. That means that you can store
356your containers on local `lvmthin` or `zfs`, shared `iSCSI` storage,
357or even on distributed storage systems like `ceph`. It also enables us
358to use advanced storage features like snapshots and clones. `vzdump`
359can also use the snapshot feature to provide consistent container
360backups.
361
362Last but not least, you can also mount local devices directly, or
363mount local directories using bind mounts. That way you can access
364local storage inside containers with zero overhead. Such bind mounts
365also provide an easy way to share data between different containers.
366
367
368Mount Points
369~~~~~~~~~~~~
370
371The root mount point is configured with the `rootfs` property, and you can
372configure up to 10 additional mount points. The corresponding options
373are called `mp0` to `mp9`, and they can contain the following setting:
374
375include::pct-mountpoint-opts.adoc[]
376
377Currently there are basically three types of mount points: storage backed
378mount points, bind mounts and device mounts.
379
380.Typical container `rootfs` configuration
381----
382rootfs: thin1:base-100-disk-1,size=8G
383----
384
385
386Storage Backed Mount Points
387^^^^^^^^^^^^^^^^^^^^^^^^^^^
388
389Storage backed mount points are managed by the {pve} storage subsystem and come
390in three different flavors:
391
392- Image based: these are raw images containing a single ext4 formatted file
393 system.
394- ZFS subvolumes: these are technically bind mounts, but with managed storage,
395 and thus allow resizing and snapshotting.
396- Directories: passing `size=0` triggers a special case where instead of a raw
397 image a directory is created.
398
399
400Bind Mount Points
401^^^^^^^^^^^^^^^^^
402
403Bind mounts allow you to access arbitrary directories from your Proxmox VE host
404inside a container. Some potential use cases are:
405
406- Accessing your home directory in the guest
407- Accessing an USB device directory in the guest
408- Accessing an NFS mount from the host in the guest
409
410Bind mounts are considered to not be managed by the storage subsystem, so you
411cannot make snapshots or deal with quotas from inside the container. With
412unprivileged containers you might run into permission problems caused by the
413user mapping and cannot use ACLs.
414
415NOTE: The contents of bind mount points are not backed up when using `vzdump`.
416
417WARNING: For security reasons, bind mounts should only be established
418using source directories especially reserved for this purpose, e.g., a
419directory hierarchy under `/mnt/bindmounts`. Never bind mount system
420directories like `/`, `/var` or `/etc` into a container - this poses a
421great security risk.
422
423NOTE: The bind mount source path must not contain any symlinks.
424
425For example, to make the directory `/mnt/bindmounts/shared` accessible in the
426container with ID `100` under the path `/shared`, use a configuration line like
427`mp0: /mnt/bindmounts/shared,mp=/shared` in `/etc/pve/lxc/100.conf`.
428Alternatively, use `pct set 100 -mp0 /mnt/bindmounts/shared,mp=/shared` to
429achieve the same result.
430
431
432Device Mount Points
433^^^^^^^^^^^^^^^^^^^
434
435Device mount points allow to mount block devices of the host directly into the
436container. Similar to bind mounts, device mounts are not managed by {PVE}'s
437storage subsystem, but the `quota` and `acl` options will be honored.
438
439NOTE: Device mount points should only be used under special circumstances. In
440most cases a storage backed mount point offers the same performance and a lot
441more features.
442
443NOTE: The contents of device mount points are not backed up when using `vzdump`.
444
445
446FUSE Mounts
447~~~~~~~~~~~
448
449WARNING: Because of existing issues in the Linux kernel's freezer
450subsystem the usage of FUSE mounts inside a container is strongly
451advised against, as containers need to be frozen for suspend or
452snapshot mode backups.
453
454If FUSE mounts cannot be replaced by other mounting mechanisms or storage
455technologies, it is possible to establish the FUSE mount on the Proxmox host
456and use a bind mount point to make it accessible inside the container.
457
458
459Using Quotas Inside Containers
460~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
461
462Quotas allow to set limits inside a container for the amount of disk
463space that each user can use. This only works on ext4 image based
464storage types and currently does not work with unprivileged
465containers.
466
467Activating the `quota` option causes the following mount options to be
468used for a mount point:
469`usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0`
470
471This allows quotas to be used like you would on any other system. You
472can initialize the `/aquota.user` and `/aquota.group` files by running
473
474----
475quotacheck -cmug /
476quotaon /
477----
478
479and edit the quotas via the `edquota` command. Refer to the documentation
480of the distribution running inside the container for details.
481
482NOTE: You need to run the above commands for every mount point by passing
483the mount point's path instead of just `/`.
484
485
486Using ACLs Inside Containers
487~~~~~~~~~~~~~~~~~~~~~~~~~~~~
488
489The standard Posix **A**ccess **C**ontrol **L**ists are also available inside containers.
490ACLs allow you to set more detailed file ownership than the traditional user/
491group/others model.
492
493
494Container Network
495-----------------
496
497You can configure up to 10 network interfaces for a single
498container. The corresponding options are called `net0` to `net9`, and
499they can contain the following setting:
500
501include::pct-network-opts.adoc[]
502
503
504Backup and Restore
505------------------
506
507
508Container Backup
509~~~~~~~~~~~~~~~~
510
511It is possible to use the `vzdump` tool for container backup. Please
512refer to the `vzdump` manual page for details.
513
514
515Restoring Container Backups
516~~~~~~~~~~~~~~~~~~~~~~~~~~~
517
518Restoring container backups made with `vzdump` is possible using the
519`pct restore` command. By default, `pct restore` will attempt to restore as much
520of the backed up container configuration as possible. It is possible to override
521the backed up configuration by manually setting container options on the command
522line (see the `pct` manual page for details).
523
524NOTE: `pvesm extractconfig` can be used to view the backed up configuration
525contained in a vzdump archive.
526
527There are two basic restore modes, only differing by their handling of mount
528points:
529
530
531``Simple'' Restore Mode
532^^^^^^^^^^^^^^^^^^^^^^^
533
534If neither the `rootfs` parameter nor any of the optional `mpX` parameters
535are explicitly set, the mount point configuration from the backed up
536configuration file is restored using the following steps:
537
538. Extract mount points and their options from backup
539. Create volumes for storage backed mount points (on storage provided with the
540`storage` parameter, or default local storage if unset)
541. Extract files from backup archive
542. Add bind and device mount points to restored configuration (limited to root user)
543
544NOTE: Since bind and device mount points are never backed up, no files are
545restored in the last step, but only the configuration options. The assumption
546is that such mount points are either backed up with another mechanism (e.g.,
547NFS space that is bind mounted into many containers), or not intended to be
548backed up at all.
549
550This simple mode is also used by the container restore operations in the web
551interface.
552
553
554``Advanced'' Restore Mode
555^^^^^^^^^^^^^^^^^^^^^^^^^
556
557By setting the `rootfs` parameter (and optionally, any combination of `mpX`
558parameters), the `pct restore` command is automatically switched into an
559advanced mode. This advanced mode completely ignores the `rootfs` and `mpX`
560configuration options contained in the backup archive, and instead only
561uses the options explicitly provided as parameters.
562
563This mode allows flexible configuration of mount point settings at restore time,
564for example:
565
566* Set target storages, volume sizes and other options for each mount point
567individually
568* Redistribute backed up files according to new mount point scheme
569* Restore to device and/or bind mount points (limited to root user)
570
571
572Managing Containers with `pct`
573------------------------------
574
575`pct` is the tool to manage Linux Containers on {pve}. You can create
576and destroy containers, and control execution (start, stop, migrate,
577...). You can use pct to set parameters in the associated config file,
578like network configuration or memory limits.
579
580
581CLI Usage Examples
582~~~~~~~~~~~~~~~~~~
583
584Create a container based on a Debian template (provided you have
585already downloaded the template via the web interface)
586
587 pct create 100 /var/lib/vz/template/cache/debian-8.0-standard_8.0-1_amd64.tar.gz
588
589Start container 100
590
591 pct start 100
592
593Start a login session via getty
594
595 pct console 100
596
597Enter the LXC namespace and run a shell as root user
598
599 pct enter 100
600
601Display the configuration
602
603 pct config 100
604
605Add a network interface called `eth0`, bridged to the host bridge `vmbr0`,
606set the address and gateway, while it's running
607
608 pct set 100 -net0 name=eth0,bridge=vmbr0,ip=192.168.15.147/24,gw=192.168.15.1
609
610Reduce the memory of the container to 512MB
611
612 pct set 100 -memory 512
613
614
615Obtaining Debugging Logs
616~~~~~~~~~~~~~~~~~~~~~~~~
617
618In case `pct start` is unable to start a specific container, it might be
619helpful to collect debugging output by running `lxc-start` (replace `ID` with
620the container's ID):
621
622 lxc-start -n ID -F -l DEBUG -o /tmp/lxc-ID.log
623
624This 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.
625
626The collected debug log is written to `/tmp/lxc-ID.log`.
627
628NOTE: If you have changed the container's configuration since the last start
629attempt with `pct start`, you need to run `pct start` at least once to also
630update the configuration used by `lxc-start`.
631
632
633Files
634------
635
636`/etc/pve/lxc/<CTID>.conf`::
637
638Configuration file for the container '<CTID>'.
639
640
641Container Advantages
642--------------------
643
644* Simple, and fully integrated into {pve}. Setup looks similar to a normal
645 VM setup.
646
647** Storage (ZFS, LVM, NFS, Ceph, ...)
648
649** Network
650
651** Authentication
652
653** Cluster
654
655* Fast: minimal overhead, as fast as bare metal
656
657* High density (perfect for idle workloads)
658
659* REST API
660
661* Direct hardware access
662
663
664Technology Overview
665-------------------
666
667* Integrated into {pve} graphical user interface (GUI)
668
669* LXC (https://linuxcontainers.org/)
670
671* lxcfs to provide containerized /proc file system
672
673* AppArmor
674
675* CRIU: for live migration (planned)
676
677* We use latest available kernels (4.4.X)
678
679* Image based deployment (templates)
680
681* Container setup from host (network, DNS, storage, ...)
682
683
684ifdef::manvolnum[]
685include::pve-copyright.adoc[]
686endif::manvolnum[]
687
688
689
690
691
692
693