Fixing network after Ubuntu 19.04 to 19.10 upgrade
I updated my server from Ubuntu 19.04 to 19.10 recently, and ran into a rather curious issue: My virtual machines connected to the network bridge would continue to work fine, and I could connect to the host via SSH as well, but any kind of DNS lookup on the host would fail. I.e. you’d run
nslookup anteru.net and you’d get a
SERVFAIL response. After you manually restarted the resolver using
systemctl restart systemd-resolved, things would immediately work again on the host. I’ve been poking around without much luck and eventually turned to askubuntu.com. That paid off big time – I did get the right hint there to solve my issue, which is: Migrate to netplan!
The original guide linked is a bit shy on how to do it properly and how we ended up here in the first place, so here’s the long version: Before Ubuntu 18.04, you’d manage your network using
ifupdown. Additionally, the DNS resolver was configured using
resolveconf. Since 18.04, Ubuntu server uses
systemd-networkd for network setup, and to ease the configuration,
netplan is used to create configuration files (and
systemd-resolved is taking care of resolving.) You’ll be wondering what this
netplan thing is good for: It’s a unified way to create configurations for both
NetworkManager, the latter being most likely what you’re using on your desktop.
With that information at hand, our goal is clear – move to
netplan manually. This is fairly simple now that we know what we’re doing. I’ll recommend using some kind of console access to do this, as I had the network disconnect me over SSH.
Let’s start by writing the
netplan configuration. This is pretty much a 1:1 port of your
/etc/network/interfaces file. What I needed was a bridge (there’s a simple example of how to use it) and I wanted to set the forward delay parameter on it. I stuffed this into
network: version: 2 renderer: networkd ethernets: eno1: dhcp4: no dhcp6: no bridges: br0: interfaces: [eno1] dhcp4: yes dhcp6: yes parameters: forward-delay: 5
We’re nearly done here. After a quick
netplan try to make sure the config is valid, let’s run
netplan apply to generate the configuration files. From here on, the network should be configured correctly. The only remaining bit is to remove things we don’t need:
apt remove --purge ifupdown resolveconf– no longer needed, and
--purgemakes sure we remove all related configuration files.
systemctl stop NetworkManager && systemctl disable NetworkManager– we’re using
systemd-networkd, so no need for
NetworkManagerto stick around.
rm /etc/network/interfacesjust in case they’re still floating around
And that’s it – your server should be working again, without having to reinstall from scratch!