Tom Parrott
Tom Parrott
@stgraber done
@monstermunchkin are you able to take a look at this, sounds like a potential optimized refresh issue. Thanks
What are your reproducer steps @monstermunchkin as I was doing a cross-host migration yesterday (between LVM in that case) and it worked. Is it just dir and ZFS affected?
Does it work with criu unistalled/disabled in the snap? As neither of those errors look like the one the OP posted.
Thanks, I've missed the "live" part of this issue somehow. CRIU live migration of containers basically doesn't work in the majority of normal situations. It only works to my knowledge...
What we could do with doing though @monstermunchkin is to find a way to better report the error to the user, as `websocket: close 1000 (normal)` is not useful but...
Yeah CRIU isn't working with systemd containers AFAIK.
We're just keeping this case open until we fix the migration to better reveal the error. See https://github.com/lxc/lxd/issues/10672#issuecomment-1190032424
I believe you have to remove the NIC from an alpine container to live migrate it.
Yeah it has a retry mechanism in it already, perhaps I'll add some additional debug logging that can be cherry-picked into the latest release so you can enable debug mode...