This backport to 18.04 LTS broke all our OpenStack compute nodes. We use
systemd-network to configure interfaces etc. and neutron-linuxbridge-
agent with some flat networks. In this scenario, the neutron agent
creates bridges for the flat networks and assigns them to physical
interfaces based on a mapping.
The backport now sets nomaster on these interfaces, effectively removing
the bridges and cutting of connectivity of all VMs using these flat
networks. This happened already when the systemd package was upgraded
(and systemd-networkd restarted).
Example:
```
10: flat-net1@data: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue
master brq88363244-5f state UP group default qlen 1000
link/ether b0:22:28:1c:f7:12 brd ff:ff:ff:ff:ff:ff
```
The interface (VLAN here) is created and managed by systemd-networkd,
the bridge `brq88363244-5f` is managed by neutron. Of course, systemd-
networkd doesn't know anything about the bridge, this wasn't a problem
until now, but the latest patch update always sets nomaster to `flat-
net1`. We were really surprise by this change in 18.04 LTS.
Is there any recommend workaround to have systemd-networkd not set
nomaster or do we have to not use systemd-network at all?
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1929560
Title:
link mac isn't set if already up
To manage notifications about this bug go to:
https://bugs.launchpad.net/systemd/+bug/1929560/+subscriptions
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
