[SOLVED] Online migrate failure - unable to detect remote migration address

[dcsapak]

and the pvesh get

this looks weird because you should definitely see the other nodes as well

It looks like the routing on the private bond0 interface (meshed networking between the nodes) is not working properly for the clustering.

you should fix your cluster before further trying to migrate

what does pvecm status say?

also your corosync conf looks wrong for multiple rings, see https://pve.proxmox.com/wiki/Separate_Cluster_Network#Redundant_Ring_Protocol

 

[Le PAH]

Thanks for the help

Here is the pvecm

root@srv-pve1:~# pvecm status
Cannot initialize CMAP service

I'll have a look at the documentation to try and fix the current situation.

 

[Le PAH]

The problem is solved!

Here what I understand:

1. The meshed interfaces were not correctly identified in the /etc/hosts file. Routing through these interfaces was impossible.
2. I had not configured the corosync.conf properly. The file did not reflected the network topology.
3. I've mistakenly created a /etc/pve/corosync.conf file manually. This file conflicted with modifications made to the /etc/corosync/corosync.conf file that were not taken into account.
4. I've explicitly declared the meshed interface in the datacenter.cfg file in order to use it when doing migration jobs.

Thanks a lot @dcsapak for the assistance.