Failed pre-start check for device "eth0": Network "lxdbr0" unavailable on this server

Incus
,
1

Hi, I’m using lxd in Arch Linux in wsl. I can start the container alpine/3.24 and access Internet in the container shell, until today, lxc start mycontainer failes with Failed pre-start check for device "eth0": Network "lxdbr0" unavailable on this server

Content of /var/log/lxd/lxd.log is

time=“2026-09-04T09:39:33+08:00” level=warning msg=“Dqlite last entry” index=0 term=0
time=“2026-09-04T09:41:13+08:00” level=warning msg=“AppArmor support has been disabled because of lack of kernel support”
time=“2026-09-04T09:41:13+08:00” level=warning msg=" - AppArmor support has been disabled, Disabled because of lack of kernel support"
time=“2026-09-04T09:41:13+08:00” level=warning msg=" - Could not find the CGroup hugetlb controller, hugepage limits will be ignored"
time=“2026-09-04T09:41:15+08:00” level=error msg=“Failed initializing network” err=“Failed starting network: The DNS and DHCP service exited prematurely: exit status 2 (“dnsmasq: failed to bind DHCP server socket: Address already in use”)” network=lxdbr0 project=default
time=“2026-09-04T09:41:20+08:00” level=error msg=“Failed initializing network” err=“Failed starting network: The DNS and DHCP service exited prematurely: exit status 2 (“dnsmasq: failed to bind DHCP server socket: Address already in use”)” network=lxdbr0 project=default

The error shows dnsmasq failed to bind DHCP server socket because Address already in use.

netstat -tulpn | grep -E ':53 |:67 ' shows

tcp        0      0 127.0.0.53:53           0.0.0.0:*               LISTEN      84/systemd-resolved
tcp        0      0 10.255.255.254:53       0.0.0.0:*               LISTEN      -
tcp        0      0 127.0.0.54:53           0.0.0.0:*               LISTEN      84/systemd-resolved
udp        0      0 127.0.0.54:53           0.0.0.0:*                           84/systemd-resolved
udp        0      0 127.0.0.53:53           0.0.0.0:*                           84/systemd-resolved
udp        0      0 10.255.255.254:53       0.0.0.0:*                           -

I assume these DNS servers are launched by the system and would like to keep them.

lxc network list shows

+-----------+----------+---------+----------------+---------------------------+-------------+---------+-------------+
|   NAME    |   TYPE   | MANAGED |      IPV4      |           IPV6            | DESCRIPTION | USED BY |    STATE    |
+-----------+----------+---------+----------------+---------------------------+-------------+---------+-------------+
| eth0      | physical | NO      |                |                           |             | 0       |             |
+-----------+----------+---------+----------------+---------------------------+-------------+---------+-------------+
| eth1      | physical | NO      |                |                           |             | 0       |             |
+-----------+----------+---------+----------------+---------------------------+-------------+---------+-------------+
| eth2      | physical | NO      |                |                           |             | 0       |             |
+-----------+----------+---------+----------------+---------------------------+-------------+---------+-------------+
| eth3      | physical | NO      |                |                           |             | 0       |             |
+-----------+----------+---------+----------------+---------------------------+-------------+---------+-------------+
| loopback0 | physical | NO      |                |                           |             | 0       |             |
+-----------+----------+---------+----------------+---------------------------+-------------+---------+-------------+
| lxdbr0    | bridge   | YES     | 10.85.190.1/24 | fd42:5ee6:eec9:ab79::1/64 |             | 2       | UNAVAILABLE |
+-----------+----------+---------+----------------+---------------------------+-------------+---------+-------------+

If relevant, ifconfig shows 3 interfaces: eth1, lo, loopback0

To fix the problem, I’m thinking about disabling dnsmasq’s DNS server. I tried lxc network set lxdbr0 raw.dnsmasq="port=0" but it failed with Error: The DNS and DHCP service exited prematurely: exit status 2 ("dnsmasq: failed to bind DHCP server socket: Address already in use")

How to fix the problem? Thanks very much!

2

This forum no longer provides support for lxd - that’s now Canonical’s project with their own support forum.

However, it’s straightforward to migrate your system from lxd to incus, and then you can get support here.

3

Thanks, I have moved to incus. Weird thing is when I decided to uninstall lxd and tried starting it the last time, lxc start mycontainer succeeded.

After moving to incus, the same problem happens again. At first incus launch images:alpine/3.24 mycontainer succeds and I can run commands in the container’s shell. The next day incus start mycontainer fails with Error: Failed pre-start check for device "eth0": Network "incusbr0" unavailable on this server.

/var/log/incus/incusd.log has:

time=“2026-09-17T21:06:08+08:00” level=warning msg=“AppArmor support has been disabled because of lack of kernel support”
time=“2026-09-17T21:06:09+08:00” level=warning msg=" - AppArmor support has been disabled, Disabled because of lack of kernel support"
time=“2026-09-17T21:06:09+08:00” level=error msg=“Failed initializing network” err=“Failed starting: The DNS and DHCP service exited prematurely: exit status 2 (“dnsmasq: failed to bind DHCP server socket: Address already in use”)” network=incusbr0 project=default
time=“2026-09-17T21:07:10+08:00” level=error msg=“Failed initializing network” err=“Failed starting: The DNS and DHCP service exited prematurely: exit status 2 (“dnsmasq: failed to bind DHCP server socket: Address already in use”)” network=incusbr0 project=default

/var/log/incus/dnsmasq.incusbr0.log has:

dnsmasq: failed to bind DHCP server socket: Address already in use

incus network list shows:

┌───────────┬──────────┬─────────┬────────────────┬───────────────────────────┬─────────────┬─────────┬─────────────┐
│   NAME    │   TYPE   │ MANAGED │      IPV4      │           IPV6            │ DESCRIPTION │ USED BY │    STATE    │
├───────────┼──────────┼─────────┼────────────────┼───────────────────────────┼─────────────┼─────────┼─────────────┤
│ eth0      │ physical │ NO      │                │                           │             │ 0       │             │
├───────────┼──────────┼─────────┼────────────────┼───────────────────────────┼─────────────┼─────────┼─────────────┤
│ eth1      │ physical │ NO      │                │                           │             │ 0       │             │
├───────────┼──────────┼─────────┼────────────────┼───────────────────────────┼─────────────┼─────────┼─────────────┤
│ eth2      │ physical │ NO      │                │                           │             │ 0       │             │
├───────────┼──────────┼─────────┼────────────────┼───────────────────────────┼─────────────┼─────────┼─────────────┤
│ eth3      │ physical │ NO      │                │                           │             │ 0       │             │
├───────────┼──────────┼─────────┼────────────────┼───────────────────────────┼─────────────┼─────────┼─────────────┤
│ incusbr0  │ bridge   │ YES     │ 10.105.72.1/24 │ fd42:a7a4:ef70:ad1d::1/64 │             │ 3       │ UNAVAILABLE │
├───────────┼──────────┼─────────┼────────────────┼───────────────────────────┼─────────────┼─────────┼─────────────┤
│ lo        │ loopback │ NO      │                │                           │             │ 0       │             │
├───────────┼──────────┼─────────┼────────────────┼───────────────────────────┼─────────────┼─────────┼─────────────┤
│ loopback0 │ physical │ NO      │                │                           │             │ 0       │             │
└───────────┴──────────┴─────────┴────────────────┴───────────────────────────┴─────────────┴─────────┴─────────────┘
4

Seems like maybe lxd is still there. Did you install it from a snap?

What does “which lxc” show? And “lxc version”?

Is it possible that you aliased lxc to incus?

5

So you have a port conflict on either the DHCP or DNS port.
Can you show a full netstat -lnp on the system?

Chances are you have a DNS server of some kind binding tcp/53 or udp/53 on all addresses, if Incus starts before that server, things work, if it starts after, it fails to bring up dnsmasq which causes the failure you’re seeing.

6

For newer systems (which don’t have the legacy ‘netstat’ installed by default), you can do:

ss -lnp sport 53
7

@candlerb

Thanks for the reply. I installed lxd using pacman so I uninstalled it by pacman -R lxd. There is no lxc and lxd.service left.

Result of netstat -lnp is:

Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name
tcp 0 0 127.0.0.54:53 0.0.0.0:* LISTEN 84/systemd-resolved
tcp 0 0 127.0.0.53:53 0.0.0.0:* LISTEN 84/systemd-resolved
tcp 0 0 10.255.255.254:53 0.0.0.0:* LISTEN -
tcp 0 0 0.0.0.0:5355 0.0.0.0:* LISTEN 84/systemd-resolved
tcp6 0 0 :::5355 :::* LISTEN 84/systemd-resolved
udp 0 0 0.0.0.0:5353 0.0.0.0:* 84/systemd-resolved
udp 0 0 0.0.0.0:5355 0.0.0.0:* 84/systemd-resolved
udp 0 0 127.0.0.54:53 0.0.0.0:* 84/systemd-resolved
udp 0 0 127.0.0.53:53 0.0.0.0:* 84/systemd-resolved
udp 0 0 10.255.255.254:53 0.0.0.0:* -
udp 0 0 127.0.0.1:323 0.0.0.0:* -
udp6 0 0 :::5353 :::* 84/systemd-resolved
udp6 0 0 :::5355 :::* 84/systemd-resolved
udp6 0 0 ::1:323 :::* -

I’ve tried disabling systemd-resolved and reboot. There are no more systemd-resolved lines in netstat -lnp result. But incus start mycontainer still failes with the same error

8

dnsmasq says “failed to bind DHCP server socket: Address already in use” which implies something listening on port 67. Here’s my system, with incus’ dnsmasq successfully running on the incusbr0 bridge:

# ss -nlp sport 67
Netid          State           Recv-Q           Send-Q                        Local Address:Port                     Peer Address:Port          Process
udp            UNCONN          0                0                          0.0.0.0%incusbr0:67                            0.0.0.0:*              users:(("dnsmasq",pid=2843694,fd=4))

I am still concerned about you saying that “lxc start mycontainer” works. As I asked before, what do the following two commands show?

which lxc
lxc version
9

Result of which lxc:which: no lxc in (PATH)

Result of lxc version: -bash: lxc: command not found