Xcp-ng pool on different bare metal

Computer Hardware & Server Infrastructure Builds
1

Looking for a little advice:

I just added a new Dell R630 to my homelab pool. In my original XCP-NG host (A supermicro X10SDV based board) is still working fine, so after adding the Dell R630 I created an XCP-NG pool for the hosts.

Everything is working well, except for 1 annoying issue: The difference in hardware is causing pool-wide eths to be different. On my supermicro, I have added a 10G SFP+ card, so it now has 5 ETHs (eth0 - eth3 are all 1gb ports) and eth4 (the 10G port). These differ from the R630 config: eth0 & eth1 are 10G SFP+ ports and eth2 & eth3 are 1G ports.

It’s kind of a pain to have to investigate each VM to see what host it’s on so I can properly assign the VM’s nics. I’ve done a little googling and it seems that I need the XCP to use the “interface-rename” script to move the PIFs around. What is the best approach to do this?

2

When needed I rename the NIC’s before I even join the hosts so they are aligned.

3

What could I expect to see if I rename them while the host is already joined? I would expect to see the VM’s already spun up on that host’s ill-named PIF to start failing because that underlying nic was renamed. Would that be the worst thing that could happen?

4

Probably just breaking the VM’s on the box. You could also move the VM’s off, remove that host, fix the networking, join it back.

5

Perfect! I still have a lot to learn! I hadn’t thought of that.