Is PVE using incorrect interfaces for disk and VM migration?

[nologin]

Hello dear All.

I have a small cluster with 2 nodes and a storage.
Each node has 3 Ethernet interfaces - 1 1Gbe in vmbr0 for internet and services access (in a vlan) on 1Gbe switch, 1 10Gbe in vlan on 10Gbe switch for storage traffic and 1 10Gbe in vlan on 10Gbe switch for cluster traffic, storage is also 10Gbe.
The traffic between nodes and storage is fine - it goes on 10G (it has no other option) but i noticed that traffic between nodes is not using 10G interfaces and always goes over 1Gbe interface.
Here are also some screenshots.

Also i noticed that it tries to use internet interface for migration but not cluster interface (i just stopped internet interface on one of the nodes)

2024-06-11 02:48:50 # /usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=vm01' -o 'UserKnownHostsFile=/etc/pve/nodes/vm01/ssh_known_hosts' -o 'GlobalKnownHostsFile=none' root@172.20.250.168 /bin/true
2024-06-11 02:48:50 ssh: connect to host 172.20.250.168 port 22: No route to host
2024-06-11 02:48:50 ERROR: migration aborted (duration 00:00:03): Can't connect to destination address using public key
TASK ERROR: migration aborted

Can you please advice!
Thank You!

Cheers

 

[nologin]

Well
I've just fixed it.
editing /etc/hosts on all nodes with desired IPs fixed the issue.

172.25.240.20 vm02.nrg-home.bg vm02
172.25.240.10 vm01.nrg-home.bg vm01