Difference between revisions of "Linux Container"

From Proxmox VE
Jump to: navigation, search
 
(12 intermediate revisions by 3 users not shown)
Line 4: Line 4:
 
[[Category:Reference Documentation]]
 
[[Category:Reference Documentation]]
 
<pvehide>
 
<pvehide>
Containers are a lightweight alternative to fully virtualized
+
Containers are a lightweight alternative to fully virtualized machines (VMs).
VMs. Instead of emulating a complete Operating System (OS), containers
+
They use the kernel of the host system that they run on, instead of emulating a
simply use the OS of the host they run on. This implies that all
+
full operating system (OS). This means that containers can access resources on
containers use the same kernel, and that they can access resources
+
the host system directly.
from the host directly.
+
The runtime costs for containers is low, usually negligible. However, there are
This is great because containers do not waste CPU power nor memory due
+
some drawbacks that need be considered:
to kernel emulation. Container run-time costs are close to zero and
+
Only Linux distributions can be run in containers. It is not possible to run
usually negligible. But there are also some drawbacks you need to
+
  other Operating Systems like, for example, FreeBSD or Microsoft Windows
consider:
+
  inside a container.
You can only run Linux based OS inside containers, i.e. it is not
+
For security reasons, access to host resources needs to be restricted.
  possible to run FreeBSD or MS Windows inside.
+
   Containers run in their own separate namespaces. Additionally some syscalls
For security reasons, access to host resources needs to be
+
  are not allowed within containers.
  restricted. This is done with AppArmor, SecComp filters and other
+
Proxmox VE uses Linux Containers (LXC) as underlying
   kernel features. Be prepared that some syscalls are not allowed
+
container technology. The &#8220;Proxmox Container Toolkit&#8221; (pct) simplifies the
  inside containers.
+
usage and management of LXC containers.
Proxmox VE uses LXC as underlying container
+
Containers are tightly integrated with Proxmox VE. This means that they are aware of
technology. We consider LXC as low-level library, which provides
+
the cluster setup, and they can use the same network and storage resources as
countless options. It would be too difficult to use those tools
+
virtual machines. You can also use the Proxmox VE firewall, or manage containers
directly. Instead, we provide a small wrapper called pct, the
+
using the HA framework.
"Proxmox Container Toolkit".
+
Our primary goal is to offer an environment as one would get from a VM, but
The toolkit is tightly coupled with Proxmox VE. That means that it is aware
+
without the additional overhead. We call this &#8220;System Containers&#8221;.
of the cluster setup, and it can use the same network and storage
+
If you want to run micro-containers, for example, Docker or rkt, it
resources as fully virtualized VMs. You can even use the Proxmox VE
 
firewall, or manage containers using the HA framework.
 
Our primary goal is to offer an environment as one would get from a
 
VM, but without the additional overhead. We call this "System
 
Containers".
 
If you want to run micro-containers (with docker, rkt, &#8230;), it
 
 
is best to run them inside a VM.
 
is best to run them inside a VM.
 
Technology Overview
 
Technology Overview
 
LXC (https://linuxcontainers.org/)
 
LXC (https://linuxcontainers.org/)
Integrated into Proxmox VE graphical user interface (GUI)
+
Integrated into Proxmox VE graphical web user interface (GUI)
 
Easy to use command line tool pct
 
Easy to use command line tool pct
 
Access via Proxmox VE REST API
 
Access via Proxmox VE REST API
 
lxcfs to provide containerized /proc file system
 
lxcfs to provide containerized /proc file system
AppArmor/Seccomp to improve security
+
Control groups (cgroups) for resource isolation and limitation
CRIU: for live migration (planned)
+
AppArmor and seccomp to improve security
Use latest available kernels (4.4.X)
+
Modern Linux kernels
 
Image based deployment (templates)
 
Image based deployment (templates)
Use Proxmox VE storage library
+
Uses Proxmox VE storage library
Container setup from host (network, DNS, storage, &#8230;)
+
Container setup from host (network, DNS, storage, etc.)
Security Considerations
 
Containers use the same kernel as the host, so there is a big attack
 
surface for malicious users. You should consider this fact if you
 
provide containers to totally untrusted people. In general, fully
 
virtualized VMs provide better isolation.
 
The good news is that LXC uses many kernel security features like
 
AppArmor, CGroups and PID and user namespaces, which makes containers
 
usage quite secure. We distinguish two types of containers:
 
Privileged Containers
 
Security is done by dropping capabilities, using mandatory access
 
control (AppArmor), SecComp filters and namespaces. The LXC team
 
considers this kind of container as unsafe, and they will not consider
 
new container escape exploits to be security issues worthy of a CVE
 
and quick fix. So you should use this kind of containers only inside a
 
trusted environment, or when no untrusted task is running as root in
 
the container.
 
Unprivileged Containers
 
This kind of containers use a new kernel feature called user
 
namespaces. The root UID 0 inside the container is mapped to an
 
unprivileged user outside the container. This means that most security
 
issues (container escape, resource abuse, &#8230;) in those containers
 
will affect a random unprivileged user, and so would be a generic
 
kernel security bug rather than an LXC issue. The LXC team thinks
 
unprivileged containers are safe by design.
 
Guest Operating System Configuration
 
We normally try to detect the operating system type inside the
 
container, and then modify some files inside the container to make
 
them work as expected. Here is a short list of things we do at
 
container startup:
 
set /etc/hostname
 
to set the container name
 
modify /etc/hosts
 
to allow lookup of the local hostname
 
network setup
 
pass the complete network setup to the container
 
configure DNS
 
pass information about DNS servers
 
adapt the init system
 
for example, fix the number of spawned getty processes
 
set the root password
 
when creating a new container
 
rewrite ssh_host_keys
 
so that each container has unique keys
 
randomize crontab
 
so that cron does not start at the same time on all containers
 
Changes made by Proxmox VE are enclosed by comment markers:
 
# --- BEGIN PVE ---
 
&lt;data&gt;
 
# --- END PVE ---
 
Those markers will be inserted at a reasonable location in the
 
file. If such a section already exists, it will be updated in place
 
and will not be moved.
 
Modification of a file can be prevented by adding a .pve-ignore.
 
file for it.  For instance, if the file /etc/.pve-ignore.hosts
 
exists then the /etc/hosts file will not be touched. This can be a
 
simple empty file creatd via:
 
# touch /etc/.pve-ignore.hosts
 
Most modifications are OS dependent, so they differ between different
 
distributions and versions. You can completely disable modifications
 
by manually setting the ostype to unmanaged.
 
OS type detection is done by testing for certain files inside the
 
container:
 
Ubuntu
 
inspect /etc/lsb-release (DISTRIB_ID=Ubuntu)
 
Debian
 
test /etc/debian_version
 
Fedora
 
test /etc/fedora-release
 
RedHat or CentOS
 
test /etc/redhat-release
 
ArchLinux
 
test /etc/arch-release
 
Alpine
 
test /etc/alpine-release
 
Gentoo
 
test /etc/gentoo-release
 
Container start fails if the configured ostype differs from the auto
 
detected type.
 
 
Container Images
 
Container Images
 
Container images, sometimes also referred to as &#8220;templates&#8221; or
 
Container images, sometimes also referred to as &#8220;templates&#8221; or
&#8220;appliances&#8221;, are tar archives which contain everything to run a
+
&#8220;appliances&#8221;, are tar archives which contain everything to run a container.
container. You can think of it as a tidy container backup. Like most
+
Proxmox VE itself provides a variety of basic templates for the most common Linux
modern container toolkits, pct uses those images when you create a
+
distributions. They can be downloaded using the GUI or the pveam (short for
new container, for example:
+
Proxmox VE Appliance Manager) command line utility.
pct create 999 local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz
+
Additionally, TurnKey Linux container templates
Proxmox VE itself ships a set of basic templates for most common
+
are also available to download.
operating systems, and you can download them using the pveam (short
+
The list of available templates is updated daily through the pve-daily-update
for Proxmox VE Appliance Manager) command line utility. You can also
+
timer. You can also trigger an update manually by executing:
download TurnKey Linux containers using
+
# pveam update
that tool (or the graphical user interface).
+
To view the list of available images run:
Our image repositories contain a list of available images, and there
+
# pveam available
is a cron job run each day to download that list. You can trigger that
 
update manually with:
 
pveam update
 
After that you can view the list of available images using:
 
pveam available
 
 
You can restrict this large list by specifying the section you are
 
You can restrict this large list by specifying the section you are
 
interested in, for example basic system images:
 
interested in, for example basic system images:
 
List available system images
 
List available system images
 
# pveam available --section system
 
# pveam available --section system
system          archlinux-base_2015-24-29-1_x86_64.tar.gz
+
system          alpine-3.10-default_20190626_amd64.tar.xz
system          centos-7-default_20160205_amd64.tar.xz
+
system          alpine-3.9-default_20190224_amd64.tar.xz
system          debian-6.0-standard_6.0-7_amd64.tar.gz
+
system          archlinux-base_20190924-1_amd64.tar.gz
system          debian-7.0-standard_7.0-3_amd64.tar.gz
+
system          centos-6-default_20191016_amd64.tar.xz
system          debian-8.0-standard_8.0-1_amd64.tar.gz
+
system          centos-7-default_20190926_amd64.tar.xz
system          ubuntu-12.04-standard_12.04-1_amd64.tar.gz
+
system          centos-8-default_20191016_amd64.tar.xz
system          ubuntu-14.04-standard_14.04-1_amd64.tar.gz
+
system          debian-10.0-standard_10.0-1_amd64.tar.gz
system          ubuntu-15.04-standard_15.04-1_amd64.tar.gz
+
system          debian-8.0-standard_8.11-1_amd64.tar.gz
system          ubuntu-15.10-standard_15.10-1_amd64.tar.gz
+
system          debian-9.0-standard_9.7-1_amd64.tar.gz
Before you can use such a template, you need to download them into one
+
system          fedora-30-default_20190718_amd64.tar.xz
of your storages. You can simply use storage local for that
+
system          fedora-31-default_20191029_amd64.tar.xz
purpose. For clustered installations, it is preferred to use a shared
+
system          gentoo-current-default_20190718_amd64.tar.xz
storage so that all nodes can access those images.
+
system          opensuse-15.0-default_20180907_amd64.tar.xz
pveam download local debian-8.0-standard_8.0-1_amd64.tar.gz
+
system          opensuse-15.1-default_20190719_amd64.tar.xz
You are now ready to create containers using that image, and you can
+
system          ubuntu-16.04-standard_16.04.5-1_amd64.tar.gz
list all downloaded images on storage local with:
+
system          ubuntu-18.04-standard_18.04.1-1_amd64.tar.gz
 +
system          ubuntu-19.04-standard_19.04-1_amd64.tar.gz
 +
system          ubuntu-19.10-standard_19.10-1_amd64.tar.gz
 +
Before you can use such a template, you need to download them into one of your
 +
storages. If you&#8217;re unsure to which one, you can simply use the local named
 +
storage for that purpose. For clustered installations, it is preferred to use a
 +
shared storage so that all nodes can access those images.
 +
# pveam download local debian-10.0-standard_10.0-1_amd64.tar.gz
 +
You are now ready to create containers using that image, and you can list all
 +
downloaded images on storage local with:
 
# pveam list local
 
# pveam list local
local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz  190.20MB
+
local:vztmpl/debian-10.0-standard_10.0-1_amd64.tar.gz  219.95MB
The above command shows you the full Proxmox VE volume identifiers. They include
+
You can also use the Proxmox VE web interface GUI to download, list and delete
the storage name, and most other Proxmox VE commands can use them. For
+
container templates.
example you can delete that image later with:
+
pct uses them to create a new container, for example:
pveam remove local:vztmpl/debian-8.0-standard_8.0-1_amd64.tar.gz
+
# pct create 999 local:vztmpl/debian-10.0-standard_10.0-1_amd64.tar.gz
Container Storage
+
The above command shows you the full Proxmox VE volume identifiers. They include the
Traditional containers use a very simple storage model, only allowing
+
storage name, and most other Proxmox VE commands can use them. For example you can
a single mount point, the root file system. This was further
+
delete that image later with:
restricted to specific file system types like ext4 and nfs.
+
# pveam remove local:vztmpl/debian-10.0-standard_10.0-1_amd64.tar.gz
Additional mounts are often done by user provided scripts. This turned
 
out to be complex and error prone, so we try to avoid that now.
 
Our new LXC based container model is more flexible regarding
 
storage. First, you can have more than a single mount point. This
 
allows you to choose a suitable storage for each application. For
 
example, you can use a relatively slow (and thus cheap) storage for
 
the container root file system. Then you can use a second mount point
 
to mount a very fast, distributed storage for your database
 
application. See section Mount Points for further
 
details.
 
The second big improvement is that you can use any storage type
 
supported by the Proxmox VE storage library. That means that you can store
 
your containers on local lvmthin or zfs, shared iSCSI storage,
 
or even on distributed storage systems like ceph. It also enables us
 
to use advanced storage features like snapshots and clones. vzdump
 
can also use the snapshot feature to provide consistent container
 
backups.
 
Last but not least, you can also mount local devices directly, or
 
mount local directories using bind mounts. That way you can access
 
local storage inside containers with zero overhead. Such bind mounts
 
also provide an easy way to share data between different containers.
 
FUSE Mounts
 
Because of existing issues in the Linux kernel&#8217;s freezer
 
subsystem the usage of FUSE mounts inside a container is strongly
 
advised against, as containers need to be frozen for suspend or
 
snapshot mode backups.
 
If FUSE mounts cannot be replaced by other mounting mechanisms or storage
 
technologies, it is possible to establish the FUSE mount on the Proxmox host
 
and use a bind mount point to make it accessible inside the container.
 
Using Quotas Inside Containers
 
Quotas allow to set limits inside a container for the amount of disk
 
space that each user can use.  This only works on ext4 image based
 
storage types and currently does not work with unprivileged
 
containers.
 
Activating the quota option causes the following mount options to be
 
used for a mount point:
 
usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0
 
This allows quotas to be used like you would on any other system. You
 
can initialize the /aquota.user and /aquota.group files by running
 
quotacheck -cmug /
 
quotaon /
 
and edit the quotas via the edquota command. Refer to the documentation
 
of the distribution running inside the container for details.
 
You need to run the above commands for every mount point by passing
 
the mount point&#8217;s path instead of just /.
 
Using ACLs Inside Containers
 
The standard Posix Access Control Lists are also available inside containers.
 
ACLs allow you to set more detailed file ownership than the traditional user/
 
group/others model.
 
 
Container Settings
 
Container Settings
 +
General Settings
 +
General settings of a container include
 +
the Node : the physical server on which the container will run
 +
the CT ID: a unique number in this Proxmox VE installation used to identify your
 +
  container
 +
Hostname: the hostname of the container
 +
Resource Pool: a logical group of containers and VMs
 +
Password: the root password of the container
 +
SSH Public Key: a public key for connecting to the root account over SSH
 +
Unprivileged container: this option allows to choose at creation time
 +
  if you want to create a privileged or unprivileged container.
 +
Unprivileged Containers
 +
Unprivileged containers use a new kernel feature called user namespaces.
 +
The root UID 0 inside the container is mapped to an unprivileged user outside
 +
the container. This means that most security issues (container escape, resource
 +
abuse, etc.) in these containers will affect a random unprivileged user, and
 +
would be a generic kernel security bug rather than an LXC issue. The LXC team
 +
thinks unprivileged containers are safe by design.
 +
This is the default option when creating a new container.
 +
If the container uses systemd as an init system, please be aware the
 +
systemd version running inside the container should be equal to or greater than
 +
220.
 +
Privileged Containers
 +
Security in containers is achieved by using mandatory access control AppArmor
 +
restrictions, seccomp filters and Linux kernel namespaces. The LXC team
 +
considers this kind of container as unsafe, and they will not consider new
 +
container escape exploits to be security issues worthy of a CVE and quick fix.
 +
That&#8217;s why privileged containers should only be used in trusted environments.
 
CPU
 
CPU
You can restrict the number of visible CPUs inside the container using
+
You can restrict the number of visible CPUs inside the container using the
the cores option. This is implemented using the Linux cpuset
+
cores option. This is implemented using the Linux cpuset cgroup
cgroup (control group). A special task inside pvestatd tries
+
(control group).
to distribute running containers among available CPUs. You can view
+
A special task inside pvestatd tries to distribute running containers among
the assigned CPUs using the following command:
+
available CPUs periodically.
 +
To view the assigned CPUs run the following command:
 
# pct cpusets
 
# pct cpusets
 
  ---------------------
 
  ---------------------
Line 233: Line 133:
 
  108:  0 1
 
  108:  0 1
 
  ---------------------
 
  ---------------------
Containers use the host kernel directly, so all task inside a
+
Containers use the host kernel directly. All tasks inside a container are
container are handled by the host CPU scheduler. Proxmox VE uses the Linux
+
handled by the host CPU scheduler. Proxmox VE uses the Linux CFS (Completely
CFS (Completely Fair Scheduler) scheduler by default,
+
Fair Scheduler) scheduler by default, which has additional bandwidth
which has additional bandwidth control options.
+
control options.
 
cpulimit:
 
cpulimit:
You can use this option to further limit assigned CPU
+
You can use this option to further limit assigned CPU time.
time. Please note that this is a floating point number, so it is
+
Please note that this is a floating point number, so it is perfectly valid to
perfectly valid to assign two cores to a container, but restrict
+
assign two cores to a container, but restrict overall CPU consumption to half a
overall CPU consumption to half a core.
+
core.
 
cores: 2
 
cores: 2
 
cpulimit: 0.5
 
cpulimit: 0.5
 
cpuunits:
 
cpuunits:
This is a relative weight passed to the kernel
+
This is a relative weight passed to the kernel scheduler. The
scheduler. The larger the number is, the more CPU time this container
+
larger the number is, the more CPU time this container gets. Number is relative
gets. Number is relative to the weights of all the other running
+
to the weights of all the other running containers. The default is 1024. You
containers. The default is 1024. You can use this setting to
+
can use this setting to prioritize some containers.
prioritize some containers.
 
 
Memory
 
Memory
 
Container memory is controlled using the cgroup memory controller.
 
Container memory is controlled using the cgroup memory controller.
 
memory:
 
memory:
Limit overall memory usage. This corresponds
+
Limit overall memory usage. This corresponds to the
to the memory.limit_in_bytes cgroup setting.
+
memory.limit_in_bytes cgroup setting.
 
swap:
 
swap:
Allows the container to use additional swap memory from the
+
Allows the container to use additional swap memory from the host
host swap space. This corresponds to the memory.memsw.limit_in_bytes
+
swap space. This corresponds to the memory.memsw.limit_in_bytes cgroup
cgroup setting, which is set to the sum of both value (memory +
+
setting, which is set to the sum of both value (memory + swap).
swap).
 
 
Mount Points
 
Mount Points
The root mount point is configured with the rootfs property, and you can
+
The root mount point is configured with the rootfs property. You can
configure up to 10 additional mount points. The corresponding options
+
configure up to 256 additional mount points. The corresponding options are
are called mp0 to mp9, and they can contain the following setting:
+
called mp0 to mp255. They can contain the following settings:
rootfs: [volume=]&lt;volume&gt; [,acl=&lt;1|0&gt;] [,quota=&lt;1|0&gt;] [,ro=&lt;1|0&gt;] [,shared=&lt;1|0&gt;] [,size=&lt;DiskSize&gt;]
+
rootfs: [volume=]&lt;volume&gt; [,acl=&lt;1|0&gt;] [,mountoptions=&lt;opt[;opt...]&gt;] [,quota=&lt;1|0&gt;] [,replicate=&lt;1|0&gt;] [,ro=&lt;1|0&gt;] [,shared=&lt;1|0&gt;] [,size=&lt;DiskSize&gt;]
 
Use volume as container root. See below for a detailed description of all options.
 
Use volume as container root. See below for a detailed description of all options.
mp[n]: [volume=]&lt;volume&gt; ,mp=&lt;Path&gt; [,acl=&lt;1|0&gt;] [,backup=&lt;1|0&gt;] [,quota=&lt;1|0&gt;] [,ro=&lt;1|0&gt;] [,shared=&lt;1|0&gt;] [,size=&lt;DiskSize&gt;]
+
mp[n]: [volume=]&lt;volume&gt; ,mp=&lt;Path&gt; [,acl=&lt;1|0&gt;] [,backup=&lt;1|0&gt;] [,mountoptions=&lt;opt[;opt...]&gt;] [,quota=&lt;1|0&gt;] [,replicate=&lt;1|0&gt;] [,ro=&lt;1|0&gt;] [,shared=&lt;1|0&gt;] [,size=&lt;DiskSize&gt;]
 
Use volume as container mount point.
 
Use volume as container mount point.
 
acl=&lt;boolean&gt;
 
acl=&lt;boolean&gt;
Line 272: Line 170:
 
backup=&lt;boolean&gt;
 
backup=&lt;boolean&gt;
 
Whether to include the mount point in backups (only used for volume mount points).
 
Whether to include the mount point in backups (only used for volume mount points).
 +
mountoptions=&lt;opt[;opt...]&gt;
 +
Extra mount options for rootfs/mps.
 
mp=&lt;Path&gt;
 
mp=&lt;Path&gt;
 
Path to the mount point as seen from inside the container.
 
Path to the mount point as seen from inside the container.
Line 277: Line 177:
 
quota=&lt;boolean&gt;
 
quota=&lt;boolean&gt;
 
Enable user quotas inside the container (not supported with zfs subvolumes)
 
Enable user quotas inside the container (not supported with zfs subvolumes)
 +
replicate=&lt;boolean&gt; (default = 1)
 +
Will include this volume to a storage replica job.
 
ro=&lt;boolean&gt;
 
ro=&lt;boolean&gt;
 
Read-only mount point
 
Read-only mount point
Line 286: Line 188:
 
volume=&lt;volume&gt;
 
volume=&lt;volume&gt;
 
Volume, device or directory to mount into the container.
 
Volume, device or directory to mount into the container.
Currently there are basically three types of mount points: storage backed
+
Currently there are three types of mount points: storage backed mount points,
mount points, bind mounts and device mounts.
+
bind mounts, and device mounts.
 
Typical container rootfs configuration
 
Typical container rootfs configuration
 
rootfs: thin1:base-100-disk-1,size=8G
 
rootfs: thin1:base-100-disk-1,size=8G
Line 299: Line 201:
 
Directories: passing size=0 triggers a special case where instead of a raw
 
Directories: passing size=0 triggers a special case where instead of a raw
 
   image a directory is created.
 
   image a directory is created.
 +
The special option syntax STORAGE_ID:SIZE_IN_GB for storage backed
 +
mount point volumes will automatically allocate a volume of the specified size
 +
on the specified storage. For example, calling
 +
pct set 100 -mp0 thin1:10,mp=/path/in/container
 +
will allocate a 10GB volume on the storage thin1 and replace the volume ID
 +
place holder 10 with the allocated volume ID, and setup the moutpoint in the
 +
container at /path/in/container
 
Bind Mount Points
 
Bind Mount Points
 
Bind mounts allow you to access arbitrary directories from your Proxmox VE host
 
Bind mounts allow you to access arbitrary directories from your Proxmox VE host
Line 310: Line 219:
 
user mapping and cannot use ACLs.
 
user mapping and cannot use ACLs.
 
The contents of bind mount points are not backed up when using vzdump.
 
The contents of bind mount points are not backed up when using vzdump.
For security reasons, bind mounts should only be established
+
For security reasons, bind mounts should only be established using
using source directories especially reserved for this purpose, e.g., a
+
source directories especially reserved for this purpose, e.g., a directory
directory hierarchy under /mnt/bindmounts. Never bind mount system
+
hierarchy under /mnt/bindmounts. Never bind mount system directories like
directories like /, /var or /etc into a container - this poses a
+
/, /var or /etc into a container - this poses a great security risk.
great security risk.
 
 
The bind mount source path must not contain any symlinks.
 
The bind mount source path must not contain any symlinks.
 
For example, to make the directory /mnt/bindmounts/shared accessible in the
 
For example, to make the directory /mnt/bindmounts/shared accessible in the
Line 328: Line 236:
 
most cases a storage backed mount point offers the same performance and a lot
 
most cases a storage backed mount point offers the same performance and a lot
 
more features.
 
more features.
The contents of device mount points are not backed up when using vzdump.
+
The contents of device mount points are not backed up when using
 +
vzdump.
 
Network
 
Network
You can configure up to 10 network interfaces for a single
+
You can configure up to 10 network interfaces for a single container.
container. The corresponding options are called net0 to net9, and
+
The corresponding options are called net0 to net9, and they can contain the
they can contain the following setting:
+
following setting:
net[n]: name=&lt;string&gt; [,bridge=&lt;bridge&gt;] [,firewall=&lt;1|0&gt;] [,gw=&lt;GatewayIPv4&gt;] [,gw6=&lt;GatewayIPv6&gt;] [,hwaddr=&lt;XX:XX:XX:XX:XX:XX&gt;] [,ip=&lt;IPv4Format/CIDR&gt;] [,ip6=&lt;IPv6Format/CIDR&gt;] [,mtu=&lt;integer&gt;] [,rate=&lt;mbps&gt;] [,tag=&lt;integer&gt;] [,trunks=&lt;vlanid[;vlanid...]&gt;] [,type=&lt;veth&gt;]
+
net[n]: name=&lt;string&gt; [,bridge=&lt;bridge&gt;] [,firewall=&lt;1|0&gt;] [,gw=&lt;GatewayIPv4&gt;] [,gw6=&lt;GatewayIPv6&gt;] [,hwaddr=&lt;XX:XX:XX:XX:XX:XX&gt;] [,ip=&lt;(IPv4/CIDR|dhcp|manual)&gt;] [,ip6=&lt;(IPv6/CIDR|auto|dhcp|manual)&gt;] [,mtu=&lt;integer&gt;] [,rate=&lt;mbps&gt;] [,tag=&lt;integer&gt;] [,trunks=&lt;vlanid[;vlanid...]&gt;] [,type=&lt;veth&gt;]
 
Specifies network interfaces for the container.
 
Specifies network interfaces for the container.
 
bridge=&lt;bridge&gt;
 
bridge=&lt;bridge&gt;
Line 344: Line 253:
 
Default gateway for IPv6 traffic.
 
Default gateway for IPv6 traffic.
 
hwaddr=&lt;XX:XX:XX:XX:XX:XX&gt;
 
hwaddr=&lt;XX:XX:XX:XX:XX:XX&gt;
The interface MAC address. This is dynamically allocated by default, but you can set that statically if needed, for example to always have the same link-local IPv6 address. (lxc.network.hwaddr)
+
A common MAC address with the I/G (Individual/Group) bit not set.
ip=&lt;IPv4Format/CIDR&gt;
+
ip=&lt;(IPv4/CIDR|dhcp|manual)&gt;
 
IPv4 address in CIDR format.
 
IPv4 address in CIDR format.
ip6=&lt;IPv6Format/CIDR&gt;
+
ip6=&lt;(IPv6/CIDR|auto|dhcp|manual)&gt;
 
IPv6 address in CIDR format.
 
IPv6 address in CIDR format.
 
mtu=&lt;integer&gt; (64 - N)
 
mtu=&lt;integer&gt; (64 - N)
Line 361: Line 270:
 
type=&lt;veth&gt;
 
type=&lt;veth&gt;
 
Network interface type.
 
Network interface type.
 +
Automatic Start and Shutdown of Containers
 +
To automatically start a container when the host system boots, select the
 +
option Start at boot in the Options panel of the container in the web
 +
interface or run the following command:
 +
# pct set CTID -onboot 1
 +
Start and Shutdown Order
 +
If you want to fine tune the boot order of your containers, you can use the
 +
following parameters:
 +
Start/Shutdown order: Defines the start order priority. For example, set it
 +
  to 1 if you want the CT to be the first to be started. (We use the reverse
 +
  startup order for shutdown, so a container with a start order of 1 would be
 +
  the last to be shut down)
 +
Startup delay: Defines the interval between this container start and
 +
  subsequent containers starts. For example, set it to 240 if you want to wait
 +
  240 seconds before starting other containers.
 +
Shutdown timeout: Defines the duration in seconds Proxmox VE should wait
 +
  for the container to be offline after issuing a shutdown command.
 +
  By default this value is set to 60, which means that Proxmox VE will issue a
 +
  shutdown request, wait 60s for the machine to be offline, and if after 60s
 +
  the machine is still online will notify that the shutdown action failed.
 +
Please note that containers without a Start/Shutdown order parameter will
 +
always start after those where the parameter is set, and this parameter only
 +
makes sense between the machines running locally on a host, and not
 +
cluster-wide.
 +
Hookscripts
 +
You can add a hook script to CTs with the config property hookscript.
 +
# pct set 100 -hookscript local:snippets/hookscript.pl
 +
It will be called during various phases of the guests lifetime.  For an example
 +
and documentation see the example script under
 +
/usr/share/pve-docs/examples/guest-example-hookscript.pl.
 +
Security Considerations
 +
Containers use the kernel of the host system. This exposes an attack surface
 +
for malicious users. In general, full virtual machines provide better
 +
isolation. This should be considered if containers are provided to unknown or
 +
untrusted people.
 +
To reduce the attack surface, LXC uses many security features like AppArmor,
 +
CGroups and kernel namespaces.
 +
AppArmor
 +
AppArmor profiles are used to restrict access to possibly dangerous actions.
 +
Some system calls, i.e. mount, are prohibited from execution.
 +
To trace AppArmor activity, use:
 +
# dmesg | grep apparmor
 +
Although it is not recommended, AppArmor can be disabled for a container. This
 +
brings security risks with it. Some syscalls can lead to privilege escalation
 +
when executed within a container if the system is misconfigured or if a LXC or
 +
Linux Kernel vulnerability exists.
 +
To disable AppArmor for a container, add the following line to the container
 +
configuration file located at /etc/pve/lxc/CTID.conf:
 +
lxc.apparmor.profile = unconfined
 +
Please note that this is not recommended for production use.
 +
Guest Operating System Configuration
 +
Proxmox VE tries to detect the Linux distribution in the container, and modifies
 +
some files. Here is a short list of things done at container startup:
 +
set /etc/hostname
 +
to set the container name
 +
modify /etc/hosts
 +
to allow lookup of the local hostname
 +
network setup
 +
pass the complete network setup to the container
 +
configure DNS
 +
pass information about DNS servers
 +
adapt the init system
 +
for example, fix the number of spawned getty processes
 +
set the root password
 +
when creating a new container
 +
rewrite ssh_host_keys
 +
so that each container has unique keys
 +
randomize crontab
 +
so that cron does not start at the same time on all containers
 +
Changes made by Proxmox VE are enclosed by comment markers:
 +
# --- BEGIN PVE ---
 +
&lt;data&gt;
 +
# --- END PVE ---
 +
Those markers will be inserted at a reasonable location in the file. If such a
 +
section already exists, it will be updated in place and will not be moved.
 +
Modification of a file can be prevented by adding a .pve-ignore. file for it.
 +
For instance, if the file /etc/.pve-ignore.hosts exists then the /etc/hosts
 +
file will not be touched. This can be a simple empty file created via:
 +
# touch /etc/.pve-ignore.hosts
 +
Most modifications are OS dependent, so they differ between different
 +
distributions and versions. You can completely disable modifications by
 +
manually setting the ostype to unmanaged.
 +
OS type detection is done by testing for certain files inside the
 +
container. Proxmox VE first checks the /etc/os-release file
 +
[/etc/os-release replaces the multitude of per-distribution
 +
release files https://manpages.debian.org/stable/systemd/os-release.5.en.html].
 +
If that file is not present, or it does not contain a clearly recognizable
 +
distribution identifier the following distribution specific release files are
 +
checked.
 +
Ubuntu
 +
inspect /etc/lsb-release (DISTRIB_ID=Ubuntu)
 +
Debian
 +
test /etc/debian_version
 +
Fedora
 +
test /etc/fedora-release
 +
RedHat or CentOS
 +
test /etc/redhat-release
 +
ArchLinux
 +
test /etc/arch-release
 +
Alpine
 +
test /etc/alpine-release
 +
Gentoo
 +
test /etc/gentoo-release
 +
Container start fails if the configured ostype differs from the auto
 +
detected type.
 +
Container Storage
 +
The Proxmox VE LXC container storage model is more flexible than traditional
 +
container storage models. A container can have multiple mount points. This
 +
makes it possible to use the best suited storage for each application.
 +
For example the root file system of the container can be on slow and cheap
 +
storage while the database can be on fast and distributed storage via a second
 +
mount point. See section  Mount Points for further
 +
details.
 +
Any storage type supported by the Proxmox VE storage library can be used. This means
 +
that containers can be stored on local (for example lvm, zfs or directory),
 +
shared external (like iSCSI, NFS) or even distributed storage systems like
 +
Ceph. Advanced storage features like snapshots or clones can be used if the
 +
underlying storage supports them. The vzdump backup tool can use snapshots to
 +
provide consistent container backups.
 +
Furthermore, local devices or local directories can be mounted directly using
 +
bind mounts. This gives access to local resources inside a container with
 +
practically zero overhead. Bind mounts can be used as an easy way to share data
 +
between containers.
 +
FUSE Mounts
 +
Because of existing issues in the Linux kernel&#8217;s freezer subsystem the
 +
usage of FUSE mounts inside a container is strongly advised against, as
 +
containers need to be frozen for suspend or snapshot mode backups.
 +
If FUSE mounts cannot be replaced by other mounting mechanisms or storage
 +
technologies, it is possible to establish the FUSE mount on the Proxmox host
 +
and use a bind mount point to make it accessible inside the container.
 +
Using Quotas Inside Containers
 +
Quotas allow to set limits inside a container for the amount of disk space that
 +
each user can use.
 +
This only works on ext4 image based storage types and currently only
 +
works with privileged containers.
 +
Activating the quota option causes the following mount options to be used for
 +
a mount point:
 +
usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0
 +
This allows quotas to be used like on any other system. You can initialize the
 +
/aquota.user and /aquota.group files by running:
 +
# quotacheck -cmug /
 +
# quotaon /
 +
Then edit the quotas using the edquota command. Refer to the documentation of
 +
the distribution running inside the container for details.
 +
You need to run the above commands for every mount point by passing the
 +
mount point&#8217;s path instead of just /.
 +
Using ACLs Inside Containers
 +
The standard Posix Access Control Lists are also available inside
 +
containers. ACLs allow you to set more detailed file ownership than the
 +
traditional user/group/others model.
 +
Backup of Container mount points
 +
To include a mount point in backups, enable the backup option for it in the
 +
container configuration. For an existing mount point mp0
 +
mp0: guests:subvol-100-disk-1,mp=/root/files,size=8G
 +
add backup=1 to enable it.
 +
mp0: guests:subvol-100-disk-1,mp=/root/files,size=8G,backup=1
 +
When creating a new mount point in the GUI, this option is enabled by
 +
default.
 +
To disable backups for a mount point, add backup=0 in the way described
 +
above, or uncheck the Backup checkbox on the GUI.
 +
Replication of Containers mount points
 +
By default, additional mount points are replicated when the Root Disk is
 +
replicated. If you want the Proxmox VE storage replication mechanism to skip a mount
 +
point, you can set the Skip replication option for that mount point.
 +
As of Proxmox VE 5.0, replication requires a storage of type zfspool. Adding a
 +
mount point to a different type of storage when the container has replication
 +
configured requires to have Skip replication enabled for that mount point.
 
Backup and Restore
 
Backup and Restore
 
Container Backup
 
Container Backup
It is possible to use the vzdump tool for container backup. Please
+
It is possible to use the vzdump tool for container backup. Please refer to
refer to the vzdump manual page for details.
+
the vzdump manual page for details.
 
Restoring Container Backups
 
Restoring Container Backups
Restoring container backups made with vzdump is possible using the
+
Restoring container backups made with vzdump is possible using the pct
pct restore command. By default, pct restore will attempt to restore as much
+
restore command. By default, pct restore will attempt to restore as much of
of the backed up container configuration as possible. It is possible to override
+
the backed up container configuration as possible. It is possible to override
the backed up configuration by manually setting container options on the command
+
the backed up configuration by manually setting container options on the
line (see the pct manual page for details).
+
command line (see the pct manual page for details).
 
pvesm extractconfig can be used to view the backed up configuration
 
pvesm extractconfig can be used to view the backed up configuration
 
contained in a vzdump archive.
 
contained in a vzdump archive.
Line 376: Line 452:
 
points:
 
points:
 
&#8220;Simple&#8221; Restore Mode
 
&#8220;Simple&#8221; Restore Mode
If neither the rootfs parameter nor any of the optional mpX parameters
+
If neither the rootfs parameter nor any of the optional mpX parameters are
are explicitly set, the mount point configuration from the backed up
+
explicitly set, the mount point configuration from the backed up configuration
configuration file is restored using the following steps:
+
file is restored using the following steps:
 
Extract mount points and their options from backup
 
Extract mount points and their options from backup
 
Create volumes for storage backed mount points (on storage provided with the
 
Create volumes for storage backed mount points (on storage provided with the
storage parameter, or default local storage if unset)
+
  storage parameter, or default local storage if unset)
 
Extract files from backup archive
 
Extract files from backup archive
Add bind and device mount points to restored configuration (limited to root user)
+
Add bind and device mount points to restored configuration (limited to root
 +
  user)
 
Since bind and device mount points are never backed up, no files are
 
Since bind and device mount points are never backed up, no files are
 
restored in the last step, but only the configuration options. The assumption
 
restored in the last step, but only the configuration options. The assumption
Line 395: Line 472:
 
parameters), the pct restore command is automatically switched into an
 
parameters), the pct restore command is automatically switched into an
 
advanced mode. This advanced mode completely ignores the rootfs and mpX
 
advanced mode. This advanced mode completely ignores the rootfs and mpX
configuration options contained in the backup archive, and instead only
+
configuration options contained in the backup archive, and instead only uses
uses the options explicitly provided as parameters.
+
the options explicitly provided as parameters.
This mode allows flexible configuration of mount point settings at restore time,
+
This mode allows flexible configuration of mount point settings at restore
for example:
+
time, for example:
 
Set target storages, volume sizes and other options for each mount point
 
Set target storages, volume sizes and other options for each mount point
individually
+
  individually
 
Redistribute backed up files according to new mount point scheme
 
Redistribute backed up files according to new mount point scheme
 
Restore to device and/or bind mount points (limited to root user)
 
Restore to device and/or bind mount points (limited to root user)
 
Managing Containers with pct
 
Managing Containers with pct
pct is the tool to manage Linux Containers on Proxmox VE. You can create
+
The &#8220;Proxmox Container Toolkit&#8221; (pct) is the command line tool to manage
and destroy containers, and control execution (start, stop, migrate,
+
Proxmox VE containers. It enables you to create or destroy containers, as well as
&#8230;). You can use pct to set parameters in the associated config file,
+
control the container execution (start, stop, reboot, migrate, etc.). It can be
like network configuration or memory limits.
+
used to set parameters in the config file of a container, for example the
 +
network configuration or memory limits.
 
CLI Usage Examples
 
CLI Usage Examples
Create a container based on a Debian template (provided you have
+
Create a container based on a Debian template (provided you have already
already downloaded the template via the web interface)
+
downloaded the template via the web interface)
pct create 100 /var/lib/vz/template/cache/debian-8.0-standard_8.0-1_amd64.tar.gz
+
# pct create 100 /var/lib/vz/template/cache/debian-10.0-standard_10.0-1_amd64.tar.gz
 
Start container 100
 
Start container 100
pct start 100
+
# pct start 100
 
Start a login session via getty
 
Start a login session via getty
pct console 100
+
# pct console 100
 
Enter the LXC namespace and run a shell as root user
 
Enter the LXC namespace and run a shell as root user
pct enter 100
+
# pct enter 100
 
Display the configuration
 
Display the configuration
pct config 100
+
# pct config 100
Add a network interface called eth0, bridged to the host bridge vmbr0,
+
Add a network interface called eth0, bridged to the host bridge vmbr0, set
set the address and gateway, while it&#8217;s running
+
the address and gateway, while it&#8217;s running
pct set 100 -net0 name=eth0,bridge=vmbr0,ip=192.168.15.147/24,gw=192.168.15.1
+
# pct set 100 -net0 name=eth0,bridge=vmbr0,ip=192.168.15.147/24,gw=192.168.15.1
 
Reduce the memory of the container to 512MB
 
Reduce the memory of the container to 512MB
pct set 100 -memory 512
+
# pct set 100 -memory 512
 
Obtaining Debugging Logs
 
Obtaining Debugging Logs
 
In case pct start is unable to start a specific container, it might be
 
In case pct start is unable to start a specific container, it might be
 
helpful to collect debugging output by running lxc-start (replace ID with
 
helpful to collect debugging output by running lxc-start (replace ID with
 
the container&#8217;s ID):
 
the container&#8217;s ID):
lxc-start -n ID -F -l DEBUG -o /tmp/lxc-ID.log
+
# lxc-start -n ID -F -l DEBUG -o /tmp/lxc-ID.log
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.
+
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.
 
The collected debug log is written to /tmp/lxc-ID.log.
 
The collected debug log is written to /tmp/lxc-ID.log.
 
If you have changed the container&#8217;s configuration since the last start
 
If you have changed the container&#8217;s configuration since the last start
 
attempt with pct start, you need to run pct start at least once to also
 
attempt with pct start, you need to run pct start at least once to also
 
update the configuration used by lxc-start.
 
update the configuration used by lxc-start.
 +
Migration
 +
If you have a cluster, you can migrate your Containers with
 +
# pct migrate &lt;ctid&gt; &lt;target&gt;
 +
This works as long as your Container is offline. If it has local volumes or
 +
mount points defined, the migration will copy the content over the network to
 +
the target host if the same storage is defined there.
 +
Running containers cannot live-migrated due to technical limitations. You can
 +
do a restart migration, which shuts down, moves and then starts a container
 +
again on the target node. As containers are very lightweight, this results
 +
normally only in a downtime of some hundreds of milliseconds.
 +
A restart migration can be done through the web interface or by using the
 +
--restart flag with the pct migrate command.
 +
A restart migration will shut down the Container and kill it after the
 +
specified timeout (the default is 180 seconds). Then it will migrate the
 +
Container like an offline migration and when finished, it starts the Container
 +
on the target node.
 
Configuration
 
Configuration
The /etc/pve/lxc/&lt;CTID&gt;.conf file stores container configuration,
+
The /etc/pve/lxc/&lt;CTID&gt;.conf file stores container configuration, where
where &lt;CTID&gt; is the numeric ID of the given container. Like all
+
&lt;CTID&gt; is the numeric ID of the given container. Like all other files stored
other files stored inside /etc/pve/, they get automatically
+
inside /etc/pve/, they get automatically replicated to all other cluster
replicated to all other cluster nodes.
+
nodes.
 
CTIDs &lt; 100 are reserved for internal purposes, and CTIDs need to be
 
CTIDs &lt; 100 are reserved for internal purposes, and CTIDs need to be
 
unique cluster wide.
 
unique cluster wide.
Line 450: Line 545:
 
net0: bridge=vmbr0,hwaddr=66:64:66:64:64:36,ip=dhcp,name=eth0,type=veth
 
net0: bridge=vmbr0,hwaddr=66:64:66:64:64:36,ip=dhcp,name=eth0,type=veth
 
rootfs: local:107/vm-107-disk-1.raw,size=7G
 
rootfs: local:107/vm-107-disk-1.raw,size=7G
Those configuration files are simple text files, and you can edit them
+
The configuration files are simple text files. You can edit them using a normal
using a normal text editor (vi, nano, &#8230;). This is sometimes
+
text editor, for example, vi or nano.
useful to do small corrections, but keep in mind that you need to
+
This is sometimes useful to do small corrections, but keep in mind that you
restart the container to apply such changes.
+
need to restart the container to apply such changes.
For that reason, it is usually better to use the pct command to
+
For that reason, it is usually better to use the pct command to generate and
generate and modify those files, or do the whole thing using the GUI.
+
modify those files, or do the whole thing using the GUI.
Our toolkit is smart enough to instantaneously apply most changes to
+
Our toolkit is smart enough to instantaneously apply most changes to running
running containers. This feature is called "hot plug", and there is no
+
containers. This feature is called &#8220;hot plug&#8221;, and there is no need to restart
need to restart the container in that case.
+
the container in that case.
 +
In cases where a change cannot be hot-plugged, it will be registered as a
 +
pending change (shown in red color in the GUI).
 +
They will only be applied after rebooting the container.
 
File Format
 
File Format
Container configuration files use a simple colon separated key/value
+
The container configuration file uses a simple colon separated key/value
 
format. Each line has the following format:
 
format. Each line has the following format:
 
# this is a comment
 
# this is a comment
 
OPTION: value
 
OPTION: value
Blank lines in those files are ignored, and lines starting with a #
+
Blank lines in those files are ignored, and lines starting with a # character
character are treated as comments and are also ignored.
+
are treated as comments and are also ignored.
It is possible to add low-level, LXC style configuration directly, for
+
It is possible to add low-level, LXC style configuration directly, for example:
example:
 
 
lxc.init_cmd: /sbin/my_own_init
 
lxc.init_cmd: /sbin/my_own_init
 
or
 
or
 
lxc.init_cmd = /sbin/my_own_init
 
lxc.init_cmd = /sbin/my_own_init
Those settings are directly passed to the LXC low-level tools.
+
The settings are passed directly to the LXC low-level tools.
 
Snapshots
 
Snapshots
When you create a snapshot, pct stores the configuration at snapshot
+
When you create a snapshot, pct stores the configuration at snapshot time
time into a separate snapshot section within the same configuration
+
into a separate snapshot section within the same configuration file. For
file. For example, after creating a snapshot called &#8220;testsnapshot&#8221;,
+
example, after creating a snapshot called &#8220;testsnapshot&#8221;, your configuration
your configuration file will look like this:
+
file will look like this:
 
Container configuration with snapshot
 
Container configuration with snapshot
 
memory: 512
 
memory: 512
Line 487: Line 584:
 
snaptime: 1457170803
 
snaptime: 1457170803
 
...
 
...
There are a few snapshot related properties like parent and
+
There are a few snapshot related properties like parent and snaptime. The
snaptime. The parent property is used to store the parent/child
+
parent property is used to store the parent/child relationship between
relationship between snapshots. snaptime is the snapshot creation
+
snapshots. snaptime is the snapshot creation time stamp (Unix epoch).
time stamp (Unix epoch).
 
 
Options
 
Options
arch: &lt;amd64 | i386&gt; (default = amd64)
+
arch: &lt;amd64 | arm64 | armhf | i386&gt; (default = amd64)
 
OS architecture type.
 
OS architecture type.
 
cmode: &lt;console | shell | tty&gt; (default = tty)
 
cmode: &lt;console | shell | tty&gt; (default = tty)
Line 508: Line 604:
 
description: &lt;string&gt;
 
description: &lt;string&gt;
 
Container description. Only used on the configuration web interface.
 
Container description. Only used on the configuration web interface.
 +
features: [force_rw_sys=&lt;1|0&gt;] [,fuse=&lt;1|0&gt;] [,keyctl=&lt;1|0&gt;] [,mknod=&lt;1|0&gt;] [,mount=&lt;fstype;fstype;...&gt;] [,nesting=&lt;1|0&gt;]
 +
Allow containers access to advanced features.
 +
force_rw_sys=&lt;boolean&gt; (default = 0)
 +
Mount /sys in unprivileged containers as rw instead of mixed. This can break networking under newer (&gt;= v245) systemd-network use.
 +
fuse=&lt;boolean&gt; (default = 0)
 +
Allow using fuse file systems in a container. Note that interactions between fuse and the freezer cgroup can potentially cause I/O deadlocks.
 +
keyctl=&lt;boolean&gt; (default = 0)
 +
For unprivileged containers only: Allow the use of the keyctl() system call. This is required to use docker inside a container. By default unprivileged containers will see this system call as non-existent. This is mostly a workaround for systemd-networkd, as it will treat it as a fatal error when some keyctl() operations are denied by the kernel due to lacking permissions. Essentially, you can choose between running systemd-networkd or docker.
 +
mknod=&lt;boolean&gt; (default = 0)
 +
Allow unprivileged containers to use mknod() to add certain device nodes. This requires a kernel with seccomp trap to user space support (5.3 or newer). This is experimental.
 +
mount=&lt;fstype;fstype;...&gt;
 +
Allow mounting file systems of specific types. This should be a list of file system types as used with the mount command. Note that this can have negative effects on the container&#8217;s security. With access to a loop device, mounting a file can circumvent the mknod permission of the devices cgroup, mounting an NFS file system can block the host&#8217;s I/O completely and prevent it from rebooting, etc.
 +
nesting=&lt;boolean&gt; (default = 0)
 +
Allow nesting. Best used with unprivileged containers with additional id mapping. Note that this will expose procfs and sysfs contents of the host to the guest.
 +
hookscript: &lt;string&gt;
 +
Script that will be exectued during various steps in the containers lifetime.
 
hostname: &lt;string&gt;
 
hostname: &lt;string&gt;
 
Set a host name for the container.
 
Set a host name for the container.
lock: &lt;backup | migrate | rollback | snapshot&gt;
+
lock: &lt;backup | create | destroyed | disk | fstrim | migrate | mounted | rollback | snapshot | snapshot-delete&gt;
 
Lock/unlock the VM.
 
Lock/unlock the VM.
 
memory: &lt;integer&gt; (16 - N) (default = 512)
 
memory: &lt;integer&gt; (16 - N) (default = 512)
 
Amount of RAM for the VM in MB.
 
Amount of RAM for the VM in MB.
mp[n]: [volume=]&lt;volume&gt; ,mp=&lt;Path&gt; [,acl=&lt;1|0&gt;] [,backup=&lt;1|0&gt;] [,quota=&lt;1|0&gt;] [,ro=&lt;1|0&gt;] [,shared=&lt;1|0&gt;] [,size=&lt;DiskSize&gt;]
+
mp[n]: [volume=]&lt;volume&gt; ,mp=&lt;Path&gt; [,acl=&lt;1|0&gt;] [,backup=&lt;1|0&gt;] [,mountoptions=&lt;opt[;opt...]&gt;] [,quota=&lt;1|0&gt;] [,replicate=&lt;1|0&gt;] [,ro=&lt;1|0&gt;] [,shared=&lt;1|0&gt;] [,size=&lt;DiskSize&gt;]
 
Use volume as container mount point.
 
Use volume as container mount point.
 
acl=&lt;boolean&gt;
 
acl=&lt;boolean&gt;
Line 520: Line 632:
 
backup=&lt;boolean&gt;
 
backup=&lt;boolean&gt;
 
Whether to include the mount point in backups (only used for volume mount points).
 
Whether to include the mount point in backups (only used for volume mount points).
 +
mountoptions=&lt;opt[;opt...]&gt;
 +
Extra mount options for rootfs/mps.
 
mp=&lt;Path&gt;
 
mp=&lt;Path&gt;
 
Path to the mount point as seen from inside the container.
 
Path to the mount point as seen from inside the container.
Line 525: Line 639:
 
quota=&lt;boolean&gt;
 
quota=&lt;boolean&gt;
 
Enable user quotas inside the container (not supported with zfs subvolumes)
 
Enable user quotas inside the container (not supported with zfs subvolumes)
 +
replicate=&lt;boolean&gt; (default = 1)
 +
Will include this volume to a storage replica job.
 
ro=&lt;boolean&gt;
 
ro=&lt;boolean&gt;
 
Read-only mount point
 
Read-only mount point
Line 536: Line 652:
 
nameserver: &lt;string&gt;
 
nameserver: &lt;string&gt;
 
Sets DNS server IP address for a container. Create will automatically use the setting from the host if you neither set searchdomain nor nameserver.
 
Sets DNS server IP address for a container. Create will automatically use the setting from the host if you neither set searchdomain nor nameserver.
net[n]: name=&lt;string&gt; [,bridge=&lt;bridge&gt;] [,firewall=&lt;1|0&gt;] [,gw=&lt;GatewayIPv4&gt;] [,gw6=&lt;GatewayIPv6&gt;] [,hwaddr=&lt;XX:XX:XX:XX:XX:XX&gt;] [,ip=&lt;IPv4Format/CIDR&gt;] [,ip6=&lt;IPv6Format/CIDR&gt;] [,mtu=&lt;integer&gt;] [,rate=&lt;mbps&gt;] [,tag=&lt;integer&gt;] [,trunks=&lt;vlanid[;vlanid...]&gt;] [,type=&lt;veth&gt;]
+
net[n]: name=&lt;string&gt; [,bridge=&lt;bridge&gt;] [,firewall=&lt;1|0&gt;] [,gw=&lt;GatewayIPv4&gt;] [,gw6=&lt;GatewayIPv6&gt;] [,hwaddr=&lt;XX:XX:XX:XX:XX:XX&gt;] [,ip=&lt;(IPv4/CIDR|dhcp|manual)&gt;] [,ip6=&lt;(IPv6/CIDR|auto|dhcp|manual)&gt;] [,mtu=&lt;integer&gt;] [,rate=&lt;mbps&gt;] [,tag=&lt;integer&gt;] [,trunks=&lt;vlanid[;vlanid...]&gt;] [,type=&lt;veth&gt;]
 
Specifies network interfaces for the container.
 
Specifies network interfaces for the container.
 
bridge=&lt;bridge&gt;
 
bridge=&lt;bridge&gt;
Line 547: Line 663:
 
Default gateway for IPv6 traffic.
 
Default gateway for IPv6 traffic.
 
hwaddr=&lt;XX:XX:XX:XX:XX:XX&gt;
 
hwaddr=&lt;XX:XX:XX:XX:XX:XX&gt;
The interface MAC address. This is dynamically allocated by default, but you can set that statically if needed, for example to always have the same link-local IPv6 address. (lxc.network.hwaddr)
+
A common MAC address with the I/G (Individual/Group) bit not set.
ip=&lt;IPv4Format/CIDR&gt;
+
ip=&lt;(IPv4/CIDR|dhcp|manual)&gt;
 
IPv4 address in CIDR format.
 
IPv4 address in CIDR format.
ip6=&lt;IPv6Format/CIDR&gt;
+
ip6=&lt;(IPv6/CIDR|auto|dhcp|manual)&gt;
 
IPv6 address in CIDR format.
 
IPv6 address in CIDR format.
 
mtu=&lt;integer&gt; (64 - N)
 
mtu=&lt;integer&gt; (64 - N)
Line 570: Line 686:
 
protection: &lt;boolean&gt; (default = 0)
 
protection: &lt;boolean&gt; (default = 0)
 
Sets the protection flag of the container. This will prevent the CT or CT&#8217;s disk remove/update operation.
 
Sets the protection flag of the container. This will prevent the CT or CT&#8217;s disk remove/update operation.
rootfs: [volume=]&lt;volume&gt; [,acl=&lt;1|0&gt;] [,quota=&lt;1|0&gt;] [,ro=&lt;1|0&gt;] [,shared=&lt;1|0&gt;] [,size=&lt;DiskSize&gt;]
+
rootfs: [volume=]&lt;volume&gt; [,acl=&lt;1|0&gt;] [,mountoptions=&lt;opt[;opt...]&gt;] [,quota=&lt;1|0&gt;] [,replicate=&lt;1|0&gt;] [,ro=&lt;1|0&gt;] [,shared=&lt;1|0&gt;] [,size=&lt;DiskSize&gt;]
 
Use volume as container root.
 
Use volume as container root.
 
acl=&lt;boolean&gt;
 
acl=&lt;boolean&gt;
 
Explicitly enable or disable ACL support.
 
Explicitly enable or disable ACL support.
 +
mountoptions=&lt;opt[;opt...]&gt;
 +
Extra mount options for rootfs/mps.
 
quota=&lt;boolean&gt;
 
quota=&lt;boolean&gt;
 
Enable user quotas inside the container (not supported with zfs subvolumes)
 
Enable user quotas inside the container (not supported with zfs subvolumes)
 +
replicate=&lt;boolean&gt; (default = 1)
 +
Will include this volume to a storage replica job.
 
ro=&lt;boolean&gt;
 
ro=&lt;boolean&gt;
 
Read-only mount point
 
Read-only mount point
Line 591: Line 711:
 
swap: &lt;integer&gt; (0 - N) (default = 512)
 
swap: &lt;integer&gt; (0 - N) (default = 512)
 
Amount of SWAP for the VM in MB.
 
Amount of SWAP for the VM in MB.
 +
tags: &lt;string&gt;
 +
Tags of the Container. This is only meta information.
 
template: &lt;boolean&gt; (default = 0)
 
template: &lt;boolean&gt; (default = 0)
 
Enable/disable Template.
 
Enable/disable Template.
Line 597: Line 719:
 
unprivileged: &lt;boolean&gt; (default = 0)
 
unprivileged: &lt;boolean&gt; (default = 0)
 
Makes the container run as unprivileged user. (Should not be modified manually.)
 
Makes the container run as unprivileged user. (Should not be modified manually.)
unused[n]: &lt;string&gt;
+
unused[n]: [volume=]&lt;volume&gt;
 
Reference to unused volumes. This is used internally, and should not be modified manually.
 
Reference to unused volumes. This is used internally, and should not be modified manually.
 +
volume=&lt;volume&gt;
 +
The volume that is not used currently.
 
Locks
 
Locks
Container migrations, snapshots and backups (vzdump) set a lock to
+
Container migrations, snapshots and backups (vzdump) set a lock to prevent
prevent incompatible concurrent actions on the affected container. Sometimes
+
incompatible concurrent actions on the affected container. Sometimes you need
you need to remove such a lock manually (e.g., after a power failure).
+
to remove such a lock manually (e.g., after a power failure).
pct unlock &lt;CTID&gt;
+
# pct unlock &lt;CTID&gt;
Only do that if you are sure the action which set the lock is
+
Only do this if you are sure the action which set the lock is no
no longer running.
+
longer running.
 
</pvehide>
 
</pvehide>
 
<!--PVE_IMPORT_END_MARKER-->
 
<!--PVE_IMPORT_END_MARKER-->
  
 
 
 
=== Security ===
 
LXC Containers use an AppArmor profile to provide ressource isolation in the container. This works by blocking system calls
 
like 'mount' who are denied being executed in the container.
 
You can trace the AppArmor activity with:
 
 
dmesg | grep apparmor
 
 
If you want to disable AppArmor for a container, you can add the stanza
 
 
lxc.aa_profile = unconfined
 
 
at the end of the configuration file ( located in /etc/pve/lxc/CTID.conf ) Note that this is not a recommended setup for production.
 
  
  
Line 635: Line 744:
 
* [https://github.com/lxc GIT Linux Container]
 
* [https://github.com/lxc GIT Linux Container]
  
[[Category:Installation]][[Category:Technology]][[Category:HOWTO]]
+
[[Category:Installation]][[Category:HOWTO]]

Latest revision as of 10:51, 12 May 2020

Containers are a lightweight alternative to fully virtualized machines (VMs). They use the kernel of the host system that they run on, instead of emulating a full operating system (OS). This means that containers can access resources on the host system directly.

The runtime costs for containers is low, usually negligible. However, there are some drawbacks that need be considered:

  • Only Linux distributions can be run in containers. It is not possible to run other Operating Systems like, for example, FreeBSD or Microsoft Windows inside a container.

  • For security reasons, access to host resources needs to be restricted. Containers run in their own separate namespaces. Additionally some syscalls are not allowed within containers.

Proxmox VE uses Linux Containers (LXC) as underlying container technology. The “Proxmox Container Toolkit” (pct) simplifies the usage and management of LXC containers.

Containers are tightly integrated with Proxmox VE. This means that they are aware of the cluster setup, and they can use the same network and storage resources as virtual machines. You can also use the Proxmox VE firewall, or manage containers using the HA framework.

Our primary goal is to offer an environment as one would get from a VM, but without the additional overhead. We call this “System Containers”.

Note If you want to run micro-containers, for example, Docker or rkt, it is best to run them inside a VM.

Technology Overview

  • LXC (https://linuxcontainers.org/)

  • Integrated into Proxmox VE graphical web user interface (GUI)

  • Easy to use command line tool pct

  • Access via Proxmox VE REST API

  • lxcfs to provide containerized /proc file system

  • Control groups (cgroups) for resource isolation and limitation

  • AppArmor and seccomp to improve security

  • Modern Linux kernels

  • Image based deployment (templates)

  • Uses Proxmox VE storage library

  • Container setup from host (network, DNS, storage, etc.)

Container Images

Container images, sometimes also referred to as “templates” or “appliances”, are tar archives which contain everything to run a container.

Proxmox VE itself provides a variety of basic templates for the most common Linux distributions. They can be downloaded using the GUI or the pveam (short for Proxmox VE Appliance Manager) command line utility. Additionally, TurnKey Linux container templates are also available to download.

The list of available templates is updated daily through the pve-daily-update timer. You can also trigger an update manually by executing:

# pveam update

To view the list of available images run:

# pveam available

You can restrict this large list by specifying the section you are interested in, for example basic system images:

List available system images
# pveam available --section system
system          alpine-3.10-default_20190626_amd64.tar.xz
system          alpine-3.9-default_20190224_amd64.tar.xz
system          archlinux-base_20190924-1_amd64.tar.gz
system          centos-6-default_20191016_amd64.tar.xz
system          centos-7-default_20190926_amd64.tar.xz
system          centos-8-default_20191016_amd64.tar.xz
system          debian-10.0-standard_10.0-1_amd64.tar.gz
system          debian-8.0-standard_8.11-1_amd64.tar.gz
system          debian-9.0-standard_9.7-1_amd64.tar.gz
system          fedora-30-default_20190718_amd64.tar.xz
system          fedora-31-default_20191029_amd64.tar.xz
system          gentoo-current-default_20190718_amd64.tar.xz
system          opensuse-15.0-default_20180907_amd64.tar.xz
system          opensuse-15.1-default_20190719_amd64.tar.xz
system          ubuntu-16.04-standard_16.04.5-1_amd64.tar.gz
system          ubuntu-18.04-standard_18.04.1-1_amd64.tar.gz
system          ubuntu-19.04-standard_19.04-1_amd64.tar.gz
system          ubuntu-19.10-standard_19.10-1_amd64.tar.gz

Before you can use such a template, you need to download them into one of your storages. If you’re unsure to which one, you can simply use the local named storage for that purpose. For clustered installations, it is preferred to use a shared storage so that all nodes can access those images.

# pveam download local debian-10.0-standard_10.0-1_amd64.tar.gz

You are now ready to create containers using that image, and you can list all downloaded images on storage local with:

# pveam list local
local:vztmpl/debian-10.0-standard_10.0-1_amd64.tar.gz  219.95MB
Tip You can also use the Proxmox VE web interface GUI to download, list and delete container templates.

pct uses them to create a new container, for example:

# pct create 999 local:vztmpl/debian-10.0-standard_10.0-1_amd64.tar.gz

The above command shows you the full Proxmox VE volume identifiers. They include the storage name, and most other Proxmox VE commands can use them. For example you can delete that image later with:

# pveam remove local:vztmpl/debian-10.0-standard_10.0-1_amd64.tar.gz

Container Settings

General Settings

screenshot/gui-create-ct-general.png

General settings of a container include

  • the Node : the physical server on which the container will run

  • the CT ID: a unique number in this Proxmox VE installation used to identify your container

  • Hostname: the hostname of the container

  • Resource Pool: a logical group of containers and VMs

  • Password: the root password of the container

  • SSH Public Key: a public key for connecting to the root account over SSH

  • Unprivileged container: this option allows to choose at creation time if you want to create a privileged or unprivileged container.

Unprivileged Containers

Unprivileged containers use a new kernel feature called user namespaces. The root UID 0 inside the container is mapped to an unprivileged user outside the container. This means that most security issues (container escape, resource abuse, etc.) in these containers will affect a random unprivileged user, and would be a generic kernel security bug rather than an LXC issue. The LXC team thinks unprivileged containers are safe by design.

This is the default option when creating a new container.

Note If the container uses systemd as an init system, please be aware the systemd version running inside the container should be equal to or greater than 220.

Privileged Containers

Security in containers is achieved by using mandatory access control AppArmor restrictions, seccomp filters and Linux kernel namespaces. The LXC team considers this kind of container as unsafe, and they will not consider new container escape exploits to be security issues worthy of a CVE and quick fix. That’s why privileged containers should only be used in trusted environments.

CPU

screenshot/gui-create-ct-cpu.png

You can restrict the number of visible CPUs inside the container using the cores option. This is implemented using the Linux cpuset cgroup (control group). A special task inside pvestatd tries to distribute running containers among available CPUs periodically. To view the assigned CPUs run the following command:

# pct cpusets
 ---------------------
 102:              6 7
 105:      2 3 4 5
 108:  0 1
 ---------------------

Containers use the host kernel directly. All tasks inside a container are handled by the host CPU scheduler. Proxmox VE uses the Linux CFS (Completely Fair Scheduler) scheduler by default, which has additional bandwidth control options.

cpulimit:

You can use this option to further limit assigned CPU time. Please note that this is a floating point number, so it is perfectly valid to assign two cores to a container, but restrict overall CPU consumption to half a core.

cores: 2
cpulimit: 0.5
cpuunits:

This is a relative weight passed to the kernel scheduler. The larger the number is, the more CPU time this container gets. Number is relative to the weights of all the other running containers. The default is 1024. You can use this setting to prioritize some containers.

Memory

screenshot/gui-create-ct-memory.png

Container memory is controlled using the cgroup memory controller.

memory:

Limit overall memory usage. This corresponds to the memory.limit_in_bytes cgroup setting.

swap:

Allows the container to use additional swap memory from the host swap space. This corresponds to the memory.memsw.limit_in_bytes cgroup setting, which is set to the sum of both value (memory + swap).

Mount Points

screenshot/gui-create-ct-root-disk.png

The root mount point is configured with the rootfs property. You can configure up to 256 additional mount points. The corresponding options are called mp0 to mp255. They can contain the following settings:

rootfs: [volume=]<volume> [,acl=<1|0>] [,mountoptions=<opt[;opt...]>] [,quota=<1|0>] [,replicate=<1|0>] [,ro=<1|0>] [,shared=<1|0>] [,size=<DiskSize>]

Use volume as container root. See below for a detailed description of all options.

mp[n]: [volume=]<volume> ,mp=<Path> [,acl=<1|0>] [,backup=<1|0>] [,mountoptions=<opt[;opt...]>] [,quota=<1|0>] [,replicate=<1|0>] [,ro=<1|0>] [,shared=<1|0>] [,size=<DiskSize>]

Use volume as container mount point.

acl=<boolean>

Explicitly enable or disable ACL support.

backup=<boolean>

Whether to include the mount point in backups (only used for volume mount points).

mountoptions=<opt[;opt...]>

Extra mount options for rootfs/mps.

mp=<Path>

Path to the mount point as seen from inside the container.

Note Must not contain any symlinks for security reasons.
quota=<boolean>

Enable user quotas inside the container (not supported with zfs subvolumes)

replicate=<boolean> (default = 1)

Will include this volume to a storage replica job.

ro=<boolean>

Read-only mount point

shared=<boolean> (default = 0)

Mark this non-volume mount point as available on all nodes.

Warning This option does not share the mount point automatically, it assumes it is shared already!
size=<DiskSize>

Volume size (read only value).

volume=<volume>

Volume, device or directory to mount into the container.

Currently there are three types of mount points: storage backed mount points, bind mounts, and device mounts.

Typical container rootfs configuration
rootfs: thin1:base-100-disk-1,size=8G

Storage Backed Mount Points

Storage backed mount points are managed by the Proxmox VE storage subsystem and come in three different flavors:

  • Image based: these are raw images containing a single ext4 formatted file system.

  • ZFS subvolumes: these are technically bind mounts, but with managed storage, and thus allow resizing and snapshotting.

  • Directories: passing size=0 triggers a special case where instead of a raw image a directory is created.

Note The special option syntax STORAGE_ID:SIZE_IN_GB for storage backed mount point volumes will automatically allocate a volume of the specified size on the specified storage. For example, calling
pct set 100 -mp0 thin1:10,mp=/path/in/container

will allocate a 10GB volume on the storage thin1 and replace the volume ID place holder 10 with the allocated volume ID, and setup the moutpoint in the container at /path/in/container

Bind Mount Points

Bind mounts allow you to access arbitrary directories from your Proxmox VE host inside a container. Some potential use cases are:

  • Accessing your home directory in the guest

  • Accessing an USB device directory in the guest

  • Accessing an NFS mount from the host in the guest

Bind mounts are considered to not be managed by the storage subsystem, so you cannot make snapshots or deal with quotas from inside the container. With unprivileged containers you might run into permission problems caused by the user mapping and cannot use ACLs.

Note The contents of bind mount points are not backed up when using vzdump.
Warning For security reasons, bind mounts should only be established using source directories especially reserved for this purpose, e.g., a directory hierarchy under /mnt/bindmounts. Never bind mount system directories like /, /var or /etc into a container - this poses a great security risk.
Note The bind mount source path must not contain any symlinks.

For example, to make the directory /mnt/bindmounts/shared accessible in the container with ID 100 under the path /shared, use a configuration line like mp0: /mnt/bindmounts/shared,mp=/shared in /etc/pve/lxc/100.conf. Alternatively, use pct set 100 -mp0 /mnt/bindmounts/shared,mp=/shared to achieve the same result.

Device Mount Points

Device mount points allow to mount block devices of the host directly into the container. Similar to bind mounts, device mounts are not managed by Proxmox VE’s storage subsystem, but the quota and acl options will be honored.

Note Device mount points should only be used under special circumstances. In most cases a storage backed mount point offers the same performance and a lot more features.
Note The contents of device mount points are not backed up when using vzdump.

Network

screenshot/gui-create-ct-network.png

You can configure up to 10 network interfaces for a single container. The corresponding options are called net0 to net9, and they can contain the following setting:

net[n]: name=<string> [,bridge=<bridge>] [,firewall=<1|0>] [,gw=<GatewayIPv4>] [,gw6=<GatewayIPv6>] [,hwaddr=<XX:XX:XX:XX:XX:XX>] [,ip=<(IPv4/CIDR|dhcp|manual)>] [,ip6=<(IPv6/CIDR|auto|dhcp|manual)>] [,mtu=<integer>] [,rate=<mbps>] [,tag=<integer>] [,trunks=<vlanid[;vlanid...]>] [,type=<veth>]

Specifies network interfaces for the container.

bridge=<bridge>

Bridge to attach the network device to.

firewall=<boolean>

Controls whether this interface’s firewall rules should be used.

gw=<GatewayIPv4>

Default gateway for IPv4 traffic.

gw6=<GatewayIPv6>

Default gateway for IPv6 traffic.

hwaddr=<XX:XX:XX:XX:XX:XX>

A common MAC address with the I/G (Individual/Group) bit not set.

ip=<(IPv4/CIDR|dhcp|manual)>

IPv4 address in CIDR format.

ip6=<(IPv6/CIDR|auto|dhcp|manual)>

IPv6 address in CIDR format.

mtu=<integer> (64 - N)

Maximum transfer unit of the interface. (lxc.network.mtu)

name=<string>

Name of the network device as seen from inside the container. (lxc.network.name)

rate=<mbps>

Apply rate limiting to the interface

tag=<integer> (1 - 4094)

VLAN tag for this interface.

trunks=<vlanid[;vlanid...]>

VLAN ids to pass through the interface

type=<veth>

Network interface type.

Automatic Start and Shutdown of Containers

To automatically start a container when the host system boots, select the option Start at boot in the Options panel of the container in the web interface or run the following command:

# pct set CTID -onboot 1
screenshot/gui-qemu-edit-start-order.png
Start and Shutdown Order

If you want to fine tune the boot order of your containers, you can use the following parameters:

  • Start/Shutdown order: Defines the start order priority. For example, set it to 1 if you want the CT to be the first to be started. (We use the reverse startup order for shutdown, so a container with a start order of 1 would be the last to be shut down)

  • Startup delay: Defines the interval between this container start and subsequent containers starts. For example, set it to 240 if you want to wait 240 seconds before starting other containers.

  • Shutdown timeout: Defines the duration in seconds Proxmox VE should wait for the container to be offline after issuing a shutdown command. By default this value is set to 60, which means that Proxmox VE will issue a shutdown request, wait 60s for the machine to be offline, and if after 60s the machine is still online will notify that the shutdown action failed.

Please note that containers without a Start/Shutdown order parameter will always start after those where the parameter is set, and this parameter only makes sense between the machines running locally on a host, and not cluster-wide.

Hookscripts

You can add a hook script to CTs with the config property hookscript.

# pct set 100 -hookscript local:snippets/hookscript.pl

It will be called during various phases of the guests lifetime. For an example and documentation see the example script under /usr/share/pve-docs/examples/guest-example-hookscript.pl.

Security Considerations

Containers use the kernel of the host system. This exposes an attack surface for malicious users. In general, full virtual machines provide better isolation. This should be considered if containers are provided to unknown or untrusted people.

To reduce the attack surface, LXC uses many security features like AppArmor, CGroups and kernel namespaces.

AppArmor

AppArmor profiles are used to restrict access to possibly dangerous actions. Some system calls, i.e. mount, are prohibited from execution.

To trace AppArmor activity, use:

# dmesg | grep apparmor

Although it is not recommended, AppArmor can be disabled for a container. This brings security risks with it. Some syscalls can lead to privilege escalation when executed within a container if the system is misconfigured or if a LXC or Linux Kernel vulnerability exists.

To disable AppArmor for a container, add the following line to the container configuration file located at /etc/pve/lxc/CTID.conf:

lxc.apparmor.profile = unconfined
Warning Please note that this is not recommended for production use.

Guest Operating System Configuration

Proxmox VE tries to detect the Linux distribution in the container, and modifies some files. Here is a short list of things done at container startup:

set /etc/hostname

to set the container name

modify /etc/hosts

to allow lookup of the local hostname

network setup

pass the complete network setup to the container

configure DNS

pass information about DNS servers

adapt the init system

for example, fix the number of spawned getty processes

set the root password

when creating a new container

rewrite ssh_host_keys

so that each container has unique keys

randomize crontab

so that cron does not start at the same time on all containers

Changes made by Proxmox VE are enclosed by comment markers:

# --- BEGIN PVE ---
<data>
# --- END PVE ---

Those markers will be inserted at a reasonable location in the file. If such a section already exists, it will be updated in place and will not be moved.

Modification of a file can be prevented by adding a .pve-ignore. file for it. For instance, if the file /etc/.pve-ignore.hosts exists then the /etc/hosts file will not be touched. This can be a simple empty file created via:

# touch /etc/.pve-ignore.hosts

Most modifications are OS dependent, so they differ between different distributions and versions. You can completely disable modifications by manually setting the ostype to unmanaged.

OS type detection is done by testing for certain files inside the container. Proxmox VE first checks the /etc/os-release file
[/etc/os-release replaces the multitude of per-distribution release files https://manpages.debian.org/stable/systemd/os-release.5.en.html]
. If that file is not present, or it does not contain a clearly recognizable distribution identifier the following distribution specific release files are checked.

Ubuntu

inspect /etc/lsb-release (DISTRIB_ID=Ubuntu)

Debian

test /etc/debian_version

Fedora

test /etc/fedora-release

RedHat or CentOS

test /etc/redhat-release

ArchLinux

test /etc/arch-release

Alpine

test /etc/alpine-release

Gentoo

test /etc/gentoo-release

Note Container start fails if the configured ostype differs from the auto detected type.

Container Storage

The Proxmox VE LXC container storage model is more flexible than traditional container storage models. A container can have multiple mount points. This makes it possible to use the best suited storage for each application.

For example the root file system of the container can be on slow and cheap storage while the database can be on fast and distributed storage via a second mount point. See section Mount Points for further details.

Any storage type supported by the Proxmox VE storage library can be used. This means that containers can be stored on local (for example lvm, zfs or directory), shared external (like iSCSI, NFS) or even distributed storage systems like Ceph. Advanced storage features like snapshots or clones can be used if the underlying storage supports them. The vzdump backup tool can use snapshots to provide consistent container backups.

Furthermore, local devices or local directories can be mounted directly using bind mounts. This gives access to local resources inside a container with practically zero overhead. Bind mounts can be used as an easy way to share data between containers.

FUSE Mounts

Warning Because of existing issues in the Linux kernel’s freezer subsystem the usage of FUSE mounts inside a container is strongly advised against, as containers need to be frozen for suspend or snapshot mode backups.

If FUSE mounts cannot be replaced by other mounting mechanisms or storage technologies, it is possible to establish the FUSE mount on the Proxmox host and use a bind mount point to make it accessible inside the container.

Using Quotas Inside Containers

Quotas allow to set limits inside a container for the amount of disk space that each user can use.

Note This only works on ext4 image based storage types and currently only works with privileged containers.

Activating the quota option causes the following mount options to be used for a mount point: usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0

This allows quotas to be used like on any other system. You can initialize the /aquota.user and /aquota.group files by running:

# quotacheck -cmug /
# quotaon /

Then edit the quotas using the edquota command. Refer to the documentation of the distribution running inside the container for details.

Note You need to run the above commands for every mount point by passing the mount point’s path instead of just /.

Using ACLs Inside Containers

The standard Posix Access Control Lists are also available inside containers. ACLs allow you to set more detailed file ownership than the traditional user/group/others model.

Backup of Container mount points

To include a mount point in backups, enable the backup option for it in the container configuration. For an existing mount point mp0

mp0: guests:subvol-100-disk-1,mp=/root/files,size=8G

add backup=1 to enable it.

mp0: guests:subvol-100-disk-1,mp=/root/files,size=8G,backup=1
Note When creating a new mount point in the GUI, this option is enabled by default.

To disable backups for a mount point, add backup=0 in the way described above, or uncheck the Backup checkbox on the GUI.

Replication of Containers mount points

By default, additional mount points are replicated when the Root Disk is replicated. If you want the Proxmox VE storage replication mechanism to skip a mount point, you can set the Skip replication option for that mount point. As of Proxmox VE 5.0, replication requires a storage of type zfspool. Adding a mount point to a different type of storage when the container has replication configured requires to have Skip replication enabled for that mount point.

Backup and Restore

Container Backup

It is possible to use the vzdump tool for container backup. Please refer to the vzdump manual page for details.

Restoring Container Backups

Restoring container backups made with vzdump is possible using the pct restore command. By default, pct restore will attempt to restore as much of the backed up container configuration as possible. It is possible to override the backed up configuration by manually setting container options on the command line (see the pct manual page for details).

Note pvesm extractconfig can be used to view the backed up configuration contained in a vzdump archive.

There are two basic restore modes, only differing by their handling of mount points:

“Simple” Restore Mode

If neither the rootfs parameter nor any of the optional mpX parameters are explicitly set, the mount point configuration from the backed up configuration file is restored using the following steps:

  1. Extract mount points and their options from backup

  2. Create volumes for storage backed mount points (on storage provided with the storage parameter, or default local storage if unset)

  3. Extract files from backup archive

  4. Add bind and device mount points to restored configuration (limited to root user)

Note Since bind and device mount points are never backed up, no files are restored in the last step, but only the configuration options. The assumption is that such mount points are either backed up with another mechanism (e.g., NFS space that is bind mounted into many containers), or not intended to be backed up at all.

This simple mode is also used by the container restore operations in the web interface.

“Advanced” Restore Mode

By setting the rootfs parameter (and optionally, any combination of mpX parameters), the pct restore command is automatically switched into an advanced mode. This advanced mode completely ignores the rootfs and mpX configuration options contained in the backup archive, and instead only uses the options explicitly provided as parameters.

This mode allows flexible configuration of mount point settings at restore time, for example:

  • Set target storages, volume sizes and other options for each mount point individually

  • Redistribute backed up files according to new mount point scheme

  • Restore to device and/or bind mount points (limited to root user)

Managing Containers with pct

The “Proxmox Container Toolkit” (pct) is the command line tool to manage Proxmox VE containers. It enables you to create or destroy containers, as well as control the container execution (start, stop, reboot, migrate, etc.). It can be used to set parameters in the config file of a container, for example the network configuration or memory limits.

CLI Usage Examples

Create a container based on a Debian template (provided you have already downloaded the template via the web interface)

# pct create 100 /var/lib/vz/template/cache/debian-10.0-standard_10.0-1_amd64.tar.gz

Start container 100

# pct start 100

Start a login session via getty

# pct console 100

Enter the LXC namespace and run a shell as root user

# pct enter 100

Display the configuration

# pct config 100

Add a network interface called eth0, bridged to the host bridge vmbr0, set the address and gateway, while it’s running

# pct set 100 -net0 name=eth0,bridge=vmbr0,ip=192.168.15.147/24,gw=192.168.15.1

Reduce the memory of the container to 512MB

# pct set 100 -memory 512

Obtaining Debugging Logs

In case pct start is unable to start a specific container, it might be helpful to collect debugging output by running lxc-start (replace ID with the container’s ID):

# lxc-start -n ID -F -l DEBUG -o /tmp/lxc-ID.log

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.

The collected debug log is written to /tmp/lxc-ID.log.

Note If you have changed the container’s configuration since the last start attempt with pct start, you need to run pct start at least once to also update the configuration used by lxc-start.

Migration

If you have a cluster, you can migrate your Containers with

# pct migrate <ctid> <target>

This works as long as your Container is offline. If it has local volumes or mount points defined, the migration will copy the content over the network to the target host if the same storage is defined there.

Running containers cannot live-migrated due to technical limitations. You can do a restart migration, which shuts down, moves and then starts a container again on the target node. As containers are very lightweight, this results normally only in a downtime of some hundreds of milliseconds.

A restart migration can be done through the web interface or by using the --restart flag with the pct migrate command.

A restart migration will shut down the Container and kill it after the specified timeout (the default is 180 seconds). Then it will migrate the Container like an offline migration and when finished, it starts the Container on the target node.

Configuration

The /etc/pve/lxc/<CTID>.conf file stores container configuration, where <CTID> is the numeric ID of the given container. Like all other files stored inside /etc/pve/, they get automatically replicated to all other cluster nodes.

Note CTIDs < 100 are reserved for internal purposes, and CTIDs need to be unique cluster wide.
Example Container Configuration
ostype: debian
arch: amd64
hostname: www
memory: 512
swap: 512
net0: bridge=vmbr0,hwaddr=66:64:66:64:64:36,ip=dhcp,name=eth0,type=veth
rootfs: local:107/vm-107-disk-1.raw,size=7G

The configuration files are simple text files. You can edit them using a normal text editor, for example, vi or nano. This is sometimes useful to do small corrections, but keep in mind that you need to restart the container to apply such changes.

For that reason, it is usually better to use the pct command to generate and modify those files, or do the whole thing using the GUI. Our toolkit is smart enough to instantaneously apply most changes to running containers. This feature is called “hot plug”, and there is no need to restart the container in that case.

In cases where a change cannot be hot-plugged, it will be registered as a pending change (shown in red color in the GUI). They will only be applied after rebooting the container.

File Format

The container configuration file uses a simple colon separated key/value format. Each line has the following format:

# this is a comment
OPTION: value

Blank lines in those files are ignored, and lines starting with a # character are treated as comments and are also ignored.

It is possible to add low-level, LXC style configuration directly, for example:

lxc.init_cmd: /sbin/my_own_init

or

lxc.init_cmd = /sbin/my_own_init

The settings are passed directly to the LXC low-level tools.

Snapshots

When you create a snapshot, pct stores the configuration at snapshot time into a separate snapshot section within the same configuration file. For example, after creating a snapshot called “testsnapshot”, your configuration file will look like this:

Container configuration with snapshot
memory: 512
swap: 512
parent: testsnaphot
...

[testsnaphot]
memory: 512
swap: 512
snaptime: 1457170803
...

There are a few snapshot related properties like parent and snaptime. The parent property is used to store the parent/child relationship between snapshots. snaptime is the snapshot creation time stamp (Unix epoch).

Options

arch: <amd64 | arm64 | armhf | i386> (default = amd64)

OS architecture type.

cmode: <console | shell | tty> (default = tty)

Console mode. By default, the console command tries to open a connection to one of the available tty devices. By setting cmode to console it tries to attach to /dev/console instead. If you set cmode to shell, it simply invokes a shell inside the container (no login).

console: <boolean> (default = 1)

Attach a console device (/dev/console) to the container.

cores: <integer> (1 - 128)

The number of cores assigned to the container. A container can use all available cores by default.

cpulimit: <number> (0 - 128) (default = 0)

Limit of CPU usage.

Note If the computer has 2 CPUs, it has a total of 2 CPU time. Value 0 indicates no CPU limit.
cpuunits: <integer> (0 - 500000) (default = 1024)

CPU weight for a VM. Argument is used in the kernel fair scheduler. The larger the number is, the more CPU time this VM gets. Number is relative to the weights of all the other running VMs.

Note You can disable fair-scheduler configuration by setting this to 0.
description: <string>

Container description. Only used on the configuration web interface.

features: [force_rw_sys=<1|0>] [,fuse=<1|0>] [,keyctl=<1|0>] [,mknod=<1|0>] [,mount=<fstype;fstype;...>] [,nesting=<1|0>]

Allow containers access to advanced features.

force_rw_sys=<boolean> (default = 0)

Mount /sys in unprivileged containers as rw instead of mixed. This can break networking under newer (>= v245) systemd-network use.

fuse=<boolean> (default = 0)

Allow using fuse file systems in a container. Note that interactions between fuse and the freezer cgroup can potentially cause I/O deadlocks.

keyctl=<boolean> (default = 0)

For unprivileged containers only: Allow the use of the keyctl() system call. This is required to use docker inside a container. By default unprivileged containers will see this system call as non-existent. This is mostly a workaround for systemd-networkd, as it will treat it as a fatal error when some keyctl() operations are denied by the kernel due to lacking permissions. Essentially, you can choose between running systemd-networkd or docker.

mknod=<boolean> (default = 0)

Allow unprivileged containers to use mknod() to add certain device nodes. This requires a kernel with seccomp trap to user space support (5.3 or newer). This is experimental.

mount=<fstype;fstype;...>

Allow mounting file systems of specific types. This should be a list of file system types as used with the mount command. Note that this can have negative effects on the container’s security. With access to a loop device, mounting a file can circumvent the mknod permission of the devices cgroup, mounting an NFS file system can block the host’s I/O completely and prevent it from rebooting, etc.

nesting=<boolean> (default = 0)

Allow nesting. Best used with unprivileged containers with additional id mapping. Note that this will expose procfs and sysfs contents of the host to the guest.

hookscript: <string>

Script that will be exectued during various steps in the containers lifetime.

hostname: <string>

Set a host name for the container.

lock: <backup | create | destroyed | disk | fstrim | migrate | mounted | rollback | snapshot | snapshot-delete>

Lock/unlock the VM.

memory: <integer> (16 - N) (default = 512)

Amount of RAM for the VM in MB.

mp[n]: [volume=]<volume> ,mp=<Path> [,acl=<1|0>] [,backup=<1|0>] [,mountoptions=<opt[;opt...]>] [,quota=<1|0>] [,replicate=<1|0>] [,ro=<1|0>] [,shared=<1|0>] [,size=<DiskSize>]

Use volume as container mount point.

acl=<boolean>

Explicitly enable or disable ACL support.

backup=<boolean>

Whether to include the mount point in backups (only used for volume mount points).

mountoptions=<opt[;opt...]>

Extra mount options for rootfs/mps.

mp=<Path>

Path to the mount point as seen from inside the container.

Note Must not contain any symlinks for security reasons.
quota=<boolean>

Enable user quotas inside the container (not supported with zfs subvolumes)

replicate=<boolean> (default = 1)

Will include this volume to a storage replica job.

ro=<boolean>

Read-only mount point

shared=<boolean> (default = 0)

Mark this non-volume mount point as available on all nodes.

Warning This option does not share the mount point automatically, it assumes it is shared already!
size=<DiskSize>

Volume size (read only value).

volume=<volume>

Volume, device or directory to mount into the container.

nameserver: <string>

Sets DNS server IP address for a container. Create will automatically use the setting from the host if you neither set searchdomain nor nameserver.

net[n]: name=<string> [,bridge=<bridge>] [,firewall=<1|0>] [,gw=<GatewayIPv4>] [,gw6=<GatewayIPv6>] [,hwaddr=<XX:XX:XX:XX:XX:XX>] [,ip=<(IPv4/CIDR|dhcp|manual)>] [,ip6=<(IPv6/CIDR|auto|dhcp|manual)>] [,mtu=<integer>] [,rate=<mbps>] [,tag=<integer>] [,trunks=<vlanid[;vlanid...]>] [,type=<veth>]

Specifies network interfaces for the container.

bridge=<bridge>

Bridge to attach the network device to.

firewall=<boolean>

Controls whether this interface’s firewall rules should be used.

gw=<GatewayIPv4>

Default gateway for IPv4 traffic.

gw6=<GatewayIPv6>

Default gateway for IPv6 traffic.

hwaddr=<XX:XX:XX:XX:XX:XX>

A common MAC address with the I/G (Individual/Group) bit not set.

ip=<(IPv4/CIDR|dhcp|manual)>

IPv4 address in CIDR format.

ip6=<(IPv6/CIDR|auto|dhcp|manual)>

IPv6 address in CIDR format.

mtu=<integer> (64 - N)

Maximum transfer unit of the interface. (lxc.network.mtu)

name=<string>

Name of the network device as seen from inside the container. (lxc.network.name)

rate=<mbps>

Apply rate limiting to the interface

tag=<integer> (1 - 4094)

VLAN tag for this interface.

trunks=<vlanid[;vlanid...]>

VLAN ids to pass through the interface

type=<veth>

Network interface type.

onboot: <boolean> (default = 0)

Specifies whether a VM will be started during system bootup.

ostype: <alpine | archlinux | centos | debian | fedora | gentoo | opensuse | ubuntu | unmanaged>

OS type. This is used to setup configuration inside the container, and corresponds to lxc setup scripts in /usr/share/lxc/config/<ostype>.common.conf. Value unmanaged can be used to skip and OS specific setup.

protection: <boolean> (default = 0)

Sets the protection flag of the container. This will prevent the CT or CT’s disk remove/update operation.

rootfs: [volume=]<volume> [,acl=<1|0>] [,mountoptions=<opt[;opt...]>] [,quota=<1|0>] [,replicate=<1|0>] [,ro=<1|0>] [,shared=<1|0>] [,size=<DiskSize>]

Use volume as container root.

acl=<boolean>

Explicitly enable or disable ACL support.

mountoptions=<opt[;opt...]>

Extra mount options for rootfs/mps.

quota=<boolean>

Enable user quotas inside the container (not supported with zfs subvolumes)

replicate=<boolean> (default = 1)

Will include this volume to a storage replica job.

ro=<boolean>

Read-only mount point

shared=<boolean> (default = 0)

Mark this non-volume mount point as available on all nodes.

Warning This option does not share the mount point automatically, it assumes it is shared already!
size=<DiskSize>

Volume size (read only value).

volume=<volume>

Volume, device or directory to mount into the container.

searchdomain: <string>

Sets DNS search domains for a container. Create will automatically use the setting from the host if you neither set searchdomain nor nameserver.

startup: `[[order=]\d+] [,up=\d+] [,down=\d+] `

Startup and shutdown behavior. Order is a non-negative number defining the general startup order. Shutdown in done with reverse ordering. Additionally you can set the up or down delay in seconds, which specifies a delay to wait before the next VM is started or stopped.

swap: <integer> (0 - N) (default = 512)

Amount of SWAP for the VM in MB.

tags: <string>

Tags of the Container. This is only meta information.

template: <boolean> (default = 0)

Enable/disable Template.

tty: <integer> (0 - 6) (default = 2)

Specify the number of tty available to the container

unprivileged: <boolean> (default = 0)

Makes the container run as unprivileged user. (Should not be modified manually.)

unused[n]: [volume=]<volume>

Reference to unused volumes. This is used internally, and should not be modified manually.

volume=<volume>

The volume that is not used currently.

Locks

Container migrations, snapshots and backups (vzdump) set a lock to prevent incompatible concurrent actions on the affected container. Sometimes you need to remove such a lock manually (e.g., after a power failure).

# pct unlock <CTID>
Caution Only do this if you are sure the action which set the lock is no longer running.



Migrate container from OpenVZ to Linux container

Follow this howto:

References