Difference between revisions of "Systemd-nspawn"

From ArchWiki
Jump to navigation Jump to search
m
(→‎Installation: systemd-nspawn does not use user namespaces, whether or not they are enabled in the kernel)
(One intermediate revision by the same user not shown)
Line 15: Line 15:
 
This mechanism differs from [[Lxc-systemd]] or [[Libvirt]]-lxc, as it is a much simple tool to configure.
 
This mechanism differs from [[Lxc-systemd]] or [[Libvirt]]-lxc, as it is a much simple tool to configure.
  
== Installation ==
+
== Installation with pacstrap ==
 
 
Before you start installing the container, please take note the following necessities:
 
* You need to build a custom [[Kernel#Compilation]] as the Archlinux [[kernel]] does not enable by default the user namespace. This setting is under {{ic|General setup ---> Namespaces support --->}}.
 
 
 
Once the kernel is built, you can verify the feature is enabled using this command:
 
 
 
{{hc|$ zgrep USER_NS /proc/config.gz|CONFIG_USER_NS = y}}
 
 
 
Information about the setting is included in the [https://bugs.archlinux.org/task/36969?dev=162 feature request].
 
 
 
* You need to add "audit=0" to the kernel parameters, as compatibility with the kernel auditing subsystem is currently broken.
 
* You need to run {{Pkg|systemd}} >= 209. As it is still under heavy development, best is to run the more recent version.
 
 
 
=== Installation with pacstrap ===
 
  
 
You need to [[pacman|install]] the package {{Pkg|arch-install-scripts}} from the [[official repositories]].
 
You need to [[pacman|install]] the package {{Pkg|arch-install-scripts}} from the [[official repositories]].

Revision as of 19:19, 23 April 2014

systemd-nspawn is like the chroot command, but it is a chroot on steroids.

systemd-nspawn may be used to run a command or OS in a light-weight namespace container. It is more powerful than chroot since it fully virtualizes the file system hierarchy, as well as the process tree, the various IPC subsystems and the host and domain name.

systemd-nspawn limits access to various kernel interfaces in the container to read-only, such as /sys, /proc/sys or /sys/fs/selinux. Network interfaces and the system clock may not be changed from within the container. Device nodes may not be created. The host system cannot be rebooted and kernel modules may not be loaded from within the container.

This mechanism differs from Lxc-systemd or Libvirt-lxc, as it is a much simple tool to configure.

Installation with pacstrap

You need to install the package arch-install-scripts from the official repositories. Then create a directory where you want. For example $ mkdir ~/MyContainer.

The next command will install all packages from the base group. It is strongly recommended to install packages from the base-devel group too.

# pacstrap -i -c -d ~/MyContainer base
Tip: the -i option will avoid auto-confirmation of package selections. As you don't need to install the Linux kernel on the container, you want to remove it from the package list selection.

Once your installation is finished, boot the container:

# systemd-nspawn -bD ~/MyContainer

And that's it! Log in as "root" with no password.

Installation with the Arch Linux ISO

Depending on the host machine filesystem setup, pacstrap can leave you with a broken filesystem with a lot of missing libraries. Thus, a safer way to install the container is to boot from the Arch ISO and follow the Installation guide. Unless you plan to mount at boot any external devices, you do not want to edit Fstab. Do not install a Boot loaders neither the Kernel (see Tip above).

Usage

machinectl

This service is used to introspect and control the state of your systemd VM and container. Managing your container is essentially done with the $ machinectl command. This service is used to introspect and control the state of your virtual machines. Please refer to machinectl(1) for an exhaustive list of options.

Boot your container at your machine startup

If you want to use the container frequently, an easy way is to boot it on init of the machine. Then you will be able to login using the machinectl mechanism.

First, you need to register your container on the host. To do this, you can either # mv /path/to/MyContainer /var/lib/container/MyContainer or just create a directory symlink:

# ln -s /path/to/MyContainer /var/lib/container/MyContainer

Following that enable and start the systemd-nspawn@MyContainer.service. To be sure the container is now registered, run the following command:

$ machinectl list
MACHINE                          CONTAINER SERVICE         
MyContainer                      container nspawn          

1 machines listed.
Tip:
  • the systemd-nspawn service will execute this command: /usr/bin/systemd-nspawn --quiet --keep-unit --boot --link-journal=guest --directory=/var/lib/container/%i . You will need to modify this file and add some more options in case your container directory is not symlinked /var/lib/container/MyCoantainer, you want to use an disk image file or set the SELinux security to container. To isolate network setting for your container, please refer to systemd-networkd. Further boot option information can be found in systemd-nspawn(1).
  • When disabling systemd-nspawn@.service, you can manually boot the container by executing # systemd-nspawn -bD /path/to/container. Once you are logged in the container, run # systemctl poweroff to shut it down

If you want to see the control group contents, run $ systemd-cgls

Login/logout to the container

systemd-nspawn command

If you still don't have enable and start the systemd-nspawn@MyContainer.service, run the following command :

# systemd-nspawn -bD /path_to/MyContainer

machinectl command

Open a new terminal window and run the following command:

# machinectl login MyContainer
Connected to container MyContainer. Press ^] three times within 1s to exit session.
Arch Linux Custom Kernel (pts/1)
MyContainer login:

You can open more than one session by loging in from another terminal.

Note: ^] corresponds to the Ctrl+] key sequence. Hold Ctrl, press tree times ] and the connection will terminate. The container will still be UP, but you will be logged out the container

Shutdown the container

# machinectl terminate MyContainer will kill all container processes and deallocate all resources attached to that instance.

See man machinectl for more options.

X environment

See Xhost if you ever need to run X applications under the new container.

Networking

The above described installation will give the container a right out of the box workable network, with no extra configuration needed. If you want a more complex network setup and isolate the container network from your host network, please visit the systemd-networkd Arch wiki

See also