IPs address changed after restore from backup - how can I fix everything to point to new IPs?

LXD Support
1

The nodes are running in Amazon EC2 and the primary interface gets assigned an internal IP that AWS assigns at the instance start time and can’t be changed.

How can I reconfigure LXD so everything points to the new IPs and not the old ones? Currently, everything is in a broken state.

2

Do you mean the LXD instance IPs or the cluster IPs?

If the LXD cluster IPs, take a look at

https://documentation.ubuntu.com/lxd/en/stable-5.21/howto/cluster_recover/#recover-cluster-members-with-changed-addresses

3

Thanks for the reply @tomp .

I looked at that link and tried the command sudo lxd cluster edit but it didn’t respond. What I found was that when you have a 3 node LXD cluster and they can’t connect to each other then you can’t do anything at all. Seemed to me like raft wants a minimum of 2 nodes for consensus and because it can’t then all the lxc and lxd commands I tried to run wouldn’t work at all.

In the end, I did fix the problem which took a long time and I can’t remember all the steps but it was something like this:

  • Went back to the original servers rather than the restored backup with the different IPs.
  • Reinitialized the primary node to make a 1 node cluster which got the main containers running again.
  • Reinitialized the second node and joined the cluster losing all the container data I had on node 2
  • Through many zpool and zfs commands and the help of ChatGPT, I somehow managed to restore all the containers back into node 2.

So all good now but not a fun experience.

1 Like