After a rolling reboot, load-balancer prefixes (/32) are not advertised

[trunet]

Required information

• Distribution: Ubuntu
• Distribution version: 22.04
• The output of "incus info" or if that fails:
  • Kernel version: 6.8.0-40-generic
  • LXC version: 6.0.1
  • Incus version: 6.4
  • Storage backend in use: ceph

Issue description

After a rolling restart of my incus cluster, it fails to advertise its load-balancers /32. I could not see them incus query /internal/testing/bgp. However if I create one LB it triggers a refresh somehow and bring all of them back. This needs to be done for each OVN to refresh its LBs.

incus network load-balancer create ovn-mystack 10.[REDACTED]
incus network load-balancer delete ovn-mystack 10.[REDACTED]

Steps to reproduce

1. Rolling reboot all servers
2. LBs are not reacheable

Information to attach

• [ ] Any relevant kernel output (dmesg)
• [ ] Container log (incus info NAME --show-log)
• [ ] Container configuration (incus config show NAME --expanded)
• [ ] Main daemon log (at /var/log/incus/incusd.log)
• [ ] Output of the client with --debug
• [ ] Output of the daemon with --debug (alternatively output of incus monitor --pretty while reproducing the issue)

[stgraber]

Got the issue reproduced here

[stgraber]

Trivial fix...