How does a migration affect a VM's network connection?

[churnd]

In my homelab where I have a nested PVE cluster, migration of a container from one node to the other interrupts the network briefly for the container. I would guess maybe 15 pings worth. I wonder how much of the delay is due to the slowness of my setup (which is to be expected) & how much the delay would be with a real PVE cluster setup (fast hypervisor servers & fast NFS SAN)? Ideally I'd like no noticeable interruption, but not sure if that's possible.

 

[MeyrNL]

I have the same issue after upgrading all hypervisors to version 3.4. The “Here I am” stage in the QEMU migration progress looks like not being triggered. Before the upgrade the functionality was working fine without any problems. Is this may be related to the - E1000/disconnected - functionality changes in this release?

I am currently not able to do any migration, because the virtual server will be unavailable and in some cases completely freeze!

Any help appreciated.