XCP-NG VM migration between servers question

Software & Web Applications
1

So I just got a new switch with 10gb connections and all my XCP-NG servers and storage are now on 10gb, my XO is still on an external computer with a 1gb connection. I decided to play and see how fast I could migrate one VM from server 1 to server 3 to play around and see what changed.

Logged into my switch and looked at the live throughput page, and then started a migration. What I saw was that I was only getting 1gb of throughput during this process. This makes me think that the migration isn’t happening from server to server directly, but happens server to XO to server. Does this sound correct? I’ll have to put an XO server up as a VM and see if things change, but wanted to ask if other people are seeing the same thing.

I’m using XO from sources for this on an old HP T630 thin client.

[edit] just tried another and maybe this vm is too small as it migrates much faster than in the past. I’ll have to bloat it up to using more RAM and test again. Maybe everything is working and the live bandwidth just cant grab the peaks fast enough to show me.

2

You can select a migration network when you migrate a VM, even within a pool since version 5.45 so as far as I know it does not loop through XO.

3

Yeah I did that, but I’m still seeing gigabit data moving into and out of the XO computer which is the confusing part. I think I just need to inflate the RAM usage so I can get a decent amount of data moving during the migrate process. The process is happening too quickly as this machine is bare bones Server 2022 doing nothing yet so only around 2GB of RAM in use and that’s only a couple of seconds to transfer over either speed.