From Funtoo
Revision as of 21:32, October 21, 2019 by Drobbins (talk | contribs)
Jump to: navigation, search
Other languages:
English • ‎português do Brasil


LXD is a container "hypervisor" designed to provide an easy set of tools to manage Linux containers, and its development is currently being led by employees at Canonical. You can learn more about the project in general at https://linuxcontainers.org/lxd/ .

LXD is currently used for container infrastructure for Funtoo Containers and is also very well-supported under Funtoo Linux. For this reason, it's recommended that you check out LXD and see what it can do for you.

Basic Setup on Funtoo

The following steps will show you how to set up a basic LXD environment under Funtoo Linux. This environment will essentially use the default LXD setup -- a will be created called lxdbr0 which will use NAT to provide Internet access to your containers. In addition, a default storage pool will be created that will simply use your existing filesystem's storage, creating a directory at /var/lib/lxd/storage-pools/default to store any containers you create. More sophisticated configurations are possible that use dedicated network bridges connected to physical interfaces without NAT, as well as dedicated storage pools that use ZFS and btrfs -- however, these types of configurations are generally overkill for a developer workstation and should only be attempted by advanced users. So we won't cover them here.


This section will guide you through setting up the basic requirements for creating an LXD environment.

The first step is to emerge LXD and its dependencies. Perform the following:

root # emerge -a lxd

Once LXD is done emerging, we will want to enable it to start by default:

root # rc-update add lxd default

In addition, we will want to set up the following files. /etc/security/limits.conf should be modified to have the following lines in it:

*       soft    nofile  1048576
*       hard    nofile  1048576
root    soft    nofile  1048576
root    hard    nofile  1048576
*       soft    memlock unlimited
*       hard    memlock unlimited
# End of file

In addition, we will want to map a set of user ids and group ids to the root user so they are available for its use. Do this by creating the /etc/subuid and /etc/subgid files with the following identical contents:


At this point we are ready to initialize and start LXD.


To configure LXD, first we will need to start LXD. This can be done as follows:

root # /etc/init.d/lxd start

At this point, we can run lxd init to run a configuration wizard to set up LXD:

root # lxd init
Would you like to use LXD clustering? (yes/no) [default=no]: 
Do you want to configure a new storage pool? (yes/no) [default=yes]: 
Name of the new storage pool [default=default]: 
Name of the storage backend to use (btrfs, dir, lvm) [default=btrfs]: dir↵
Would you like to connect to a MAAS server? (yes/no) [default=no]: 
Would you like to create a new local network bridge? (yes/no) [default=yes]: 
What should the new bridge be called? [default=lxdbr0]: 
What IPv4 address should be used? (CIDR subnet notation, “auto” or “none”) [default=auto]: 
What IPv6 address should be used? (CIDR subnet notation, “auto” or “none”) [default=auto]: 
Would you like LXD to be available over the network? (yes/no) [default=no]: 
Would you like stale cached images to be updated automatically? (yes/no) [default=yes] 
Would you like a YAML "lxd init" preseed to be printed? (yes/no) [default=no]: 
root #

As you can see, we chose all the default 'except' for storage pool, where we opted for using a directory-based container storage rather than btrfs. Now, we should be able to run lxc image list and get a response from the LXD daemon:

root # lxc image list
root #

If you are able to do this, you have successfully set up the core parts of LXD! Note that we used the command lxc and not lxd like we did for lxd init -- from this point forward, you will use the lxc command. Don't let this confuse you -- the lxc command is the primary command-line tool for working with LXD containers.

Above, you can see that no images are installed. Images are installable snapshots of containers that we can use to create new containers ourselves. So, as a first step, let's go ahead and grab an image we can use. You will want to browse https://build.funtoo.org for an LXD image that will work on your computer hardware. For example, I was able to download the following file using wget:

root # wget https://build.funtoo.org/1.3-release-std/x86-64bit/intel64-skylake/lxd-intel64-skylake-1.3-release-std-2019-06-11.tar.xz

Once downloaded, this image can be installed using the following command:

root # lxc image import lxd-intel64-skylake-1.3-release-std-2019-06-11.tar.xz --alias funtoo
Image imported with fingerprint: fe4d27fb31bfaf3bd4f470e0ea43d26a6c05991de2a504b9e0a3b1a266dddc69

Now you will see the image available in our image list:

root # lxc image list
 ALIAS  | FINGERPRINT  | PUBLIC |                DESCRIPTION                 |  ARCH  |   SIZE   |         UPLOAD DATE          |
 funtoo | fe4d27fb31bf | no     | 1.3 Release Skylake 64bit [std] 2019-06-14 | x86_64 | 279.35MB | Jun 15, 2019 at 3:09am (UTC) |
root #

First Containeer

It is now time to launch our first container. This can be done as follows:

root # lxc launch funtoo testcontainer
Creating testcontainer
Starting testcontainer

We can now see the container running via lxc list:

root # lxc list
     NAME      |  STATE  | IPV4 |                     IPV6                      |    TYPE    | SNAPSHOTS |
 testcontainer | RUNNING |      | fd42:8063:81cb:988c:216:3eff:fe2a:f901 (eth0) | PERSISTENT |           |
root #

By default, our new container testcontainer will use the default profile, which will connect an eth0 interface in the container to NAT, and will also use our directory-based LXD storage pool. We can now enter the container as follows:

root # lxc exec testcontainer -- su --login
testcontainer #

As you might have noticed, we do not yet have any IPv4 networking configured. While LXD has set up a bridge and NAT for us, along with a DHCP server to query, we actually need to use dhcpcd to query for an IP address, so let's get that set up:

testcontainer # echo "template=dhcpcd" > /etc/conf.d/netif.eth0
testcontainer # cd /etc/init.d
testcontainer # ln -s netif.tmpl netif.eth0
testcontainer # rc-update add netif.eth0 default
 * service netif.eth0 added to runlevel default
testcontainer # rc
 * rc is deprecated, please use openrc instead.
 * Caching service dependencies ...                             [ ok ]
 * Starting DHCP Client Daemon ...                              [ ok ]
 * Network dhcpcd eth0 up ...                                   [ ok ]
testcontainer # 

You can now see that eth0 has a valid IPv4 address:

testcontainer # ifconfig
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet  netmask  broadcast
        inet6 fd42:8063:81cb:988c:25ea:b5bd:603d:8b0d  prefixlen 64  scopeid 0x0<global>
        inet6 fe80::216:3eff:fe2a:f901  prefixlen 64  scopeid 0x20<link>
        ether 00:16:3e:2a:f9:01  txqueuelen 1000  (Ethernet)
        RX packets 45  bytes 5385 (5.2 KiB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 20  bytes 2232 (2.1 KiB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

Time to have some fun!

testcontainer # ego sync

Running systemd container on a non-systemd host

To use systemd in the container, a recent enough (>=4.6) kernel version with support for cgroup namespaces is needed. Funtoo's openrc has the fix to mount systemd cgroups, which is sufficient to run systemd based distributions lxd containers.

If you want to get systemd hierarchy mounted automatically on system startup, using /etc/fstab will not work, but the dev-libs/libcgroup

can be used for this. First you needed to edit the /etc/cgroup/cgconfig.conf and add:
mount {
    "name=systemd" = /sys/fs/cgroup/systemd;

Then you need to start the cgconfig daemon:

root # rc-service cgconfig start

The daemon can be started as needed, or automatically at system start by simply adding it to default group:

root # rc-update add cgconfig default


PART Y - Docker in LXD


List of tested and working images

These are images from the https://images.linuxcontainers.org repository available by default in lxd. You can list all available images by typing following command (beware the list is very long):

root # lxc image list images:
|              ALIAS              | FINGERPRINT  | PUBLIC |               DESCRIPTION                |  ARCH   |   SIZE   |          UPLOAD DATE          |
| alpine/3.3 (3 more)             | ef69c8dc37f6 | yes    | Alpine 3.3 amd64 (20171018_17:50)        | x86_64  | 2.00MB   | Oct 18, 2017 at 12:00am (UTC) |
| alpine/3.3/armhf (1 more)       | 5ce4c80edcf3 | yes    | Alpine 3.3 armhf (20170103_17:50)        | armv7l  | 1.53MB   | Jan 3, 2017 at 12:00am (UTC)  |
| alpine/3.3/i386 (1 more)        | cd1700cb7c97 | yes    | Alpine 3.3 i386 (20171018_17:50)         | i686    | 1.84MB   | Oct 18, 2017 at 12:00am (UTC) |
| alpine/3.4 (3 more)             | bd4f1ccfabb5 | yes    | Alpine 3.4 amd64 (20171018_17:50)        | x86_64  | 2.04MB   | Oct 18, 2017 at 12:00am (UTC) |
| alpine/3.4/armhf (1 more)       | 9fe7c201924c | yes    | Alpine 3.4 armhf (20170111_20:27)        | armv7l  | 1.58MB   | Jan 11, 2017 at 12:00am (UTC) |
| alpine/3.4/i386 (1 more)        | 188a31315773 | yes    | Alpine 3.4 i386 (20171018_17:50)         | i686    | 1.88MB   | Oct 18, 2017 at 12:00am (UTC) |
| alpine/3.5 (3 more)             | 63bebc672163 | yes    | Alpine 3.5 amd64 (20171018_17:50)        | x86_64  | 1.70MB   | Oct 18, 2017 at 12:00am (UTC) |
| alpine/3.5/i386 (1 more)        | 48045e297515 | yes    | Alpine 3.5 i386 (20171018_17:50)         | i686    | 1.73MB   | Oct 18, 2017 at 12:00am (UTC) |
|                                 | fd95a7a754a0 | yes    | Alpine 3.5 amd64 (20171016_17:50)        | x86_64  | 1.70MB   | Oct 16, 2017 at 12:00am (UTC) |
|                                 | fef66668f5a2 | yes    | Debian stretch arm64 (20171016_22:42)    | aarch64 | 96.56MB  | Oct 16, 2017 at 12:00am (UTC) |
|                                 | ff18aa2c11d7 | yes    | Opensuse 42.3 amd64 (20171017_00:53)     | x86_64  | 58.92MB  | Oct 17, 2017 at 12:00am (UTC) |
|                                 | ff4ef0d824b6 | yes    | Ubuntu zesty s390x (20171017_03:49)      | s390x   | 86.88MB  | Oct 17, 2017 at 12:00am (UTC) |

These are the images that are known to work with current LXD setup on Funtoo Linux:

Image Init Status
CentOS 7 systemd Working
Debian Jessie (8) - EOL April/May 2020 systemd Working (systemd - no failed units)
Debian Stretch (9) - EOL June 2022 systemd Working
Fedora 26 systemd with cgroup v2 Not Working
Fedora 25 systemd Working
Fedora 24 systemd Working
Oracle 7 systemd Working (systemd - no failed units)
OpenSUSE 42.2 systemd Working
OpenSUSE 42.3 systemd Working
Ubuntu Xenial (16.04 LTS) - EOL 2021-04 systemd Working
Ubuntu Zesty (17.04) - EOL 2018-01 systemd Working
Alpine 3.3 OpenRC Working
Alpine 3.4 OpenRC Working
Alpine 3.5 OpenRC Working
Alpine 3.6 OpenRC Working
Alpine Edge OpenRC Working
Archlinux systemd with cgroup v2 Not Working
CentOS 6 upstart Working (systemd - no failed units)
Debian Buster systemd with cgroup v2 Not Working
Debian Sid systemd with cgroup v2 Not working
Debian Wheezy (7) - EOL May 2018 ? ? (more testing needed)
Gentoo OpenRC Working (all services started)
Oracle 6 upstart ? (mount outputs nothing)
Plamo 5 ? ?
Plamo 6 ? ?
Sabayon systemd with cgroup v2 Not Working
Ubuntu Artful (17.10) - EOL 2018-07 systemd with cgroup v2 Not Working
Ubuntu Core 16 ? ?
Ubuntu Trusty (14.04 LTS) - EOL 2019-04 upstart Working


Some of the biggest features of LXD are:

  • Secure by design (unprivileged containers, resource restrictions and much more)
  • Scalable (from containers on your laptop to thousand of compute nodes)
  • Intuitive (simple, clear API and crisp command line experience)
  • Image based (no more distribution templates, only good, trusted images)
  • Live migration

Unprivileged Containers

LXD uses unprivileged containers by default. The difference between an unprivileged container and a privileged one is whether the root user in the container is the “real” root user (uid 0 at the kernel level).

The way unprivileged containers are created is by taking a set of normal UIDs and GIDs from the host, usually at least 65536 of each (to be POSIX compliant) and mapping those into the container.

The most common example and what most LXD users will end up with by default is a map of 65536 UIDs and GIDs, with a host base id of 100000. This means that root in the container (uid 0) will be mapped to the host uid 100000 and uid 65535 in the container will be mapped to uid 165535 on the host. UID/GID 65536 and higher in the container aren’t mapped and will return an error if you attempt to use them.

From a security point of view, that means that anything which is not owned by the users and groups mapped into the container will be inaccessible. Any such resource will show up as being owned by uid/gid “-1” (rendered as 65534 or nobody/nogroup in userspace). It also means that should there be a way to escape the container, even root in the container would find itself with just as much privileges on the host as a nobody user.

LXD does offer a number of options related to unprivileged configuration:

  • Increasing the size of the default uid/gid map
  • Setting up per-container maps
  • Punching holes into the map to expose host users and groups

Relationship with LXC

LXD isn't a rewrite of LXC, in fact it's building on top of LXC to provide a new, better user experience. Under the hood, LXD uses LXC through liblxc and its Go binding to create and manage the containers.

It's basically an alternative to LXC's tools and distribution template system with the added features that come from being controllable over the network.