Adding a bridge to a guest breaks the host

[greenseeker]

Hi all,

I was trying to spin up a matterv single-node nested deployment on my PVE cluster, just so I could poke around and check it out. The install goes fine, but as soon as I get to the step where I configure the VM bridge on the matterv host, the PVE host where it's running loses all network communication.

The matterv bridge doesn't have any IP range assigned to it yet. It's also a different naming scheme than the bridge on the PVE host, though I can't imagine that makes a difference.

Why would this happen? How can I troubleshoot?

 

[bkry]

Hi,
Have you been able to find a solution to this problem?
If not, have you checked the journalctl log on the PVE host during the issue time? This might show what’s happening.
Which NIC model type are you using for the matterv VM? Have you tried using a different model, and did you experience the same issue? [0]

[0] https://localhost:8006/pve-docs/chapter-qm.html#qm_network_device

 

[greenseeker]

Hi,
Have you been able to find a solution to this problem?
If not, have you checked the journalctl log on the PVE host during the issue time? This might show what’s happening.
Which NIC model type are you using for the matterv VM? Have you tried using a different model, and did you experience the same issue? [0]

[0] https://localhost:8006/pve-docs/chapter-qm.html#qm_network_device

I concluded having a bridge on the vnic which is then plugged into a bridge on the host was creating arp chaos and/or routing loops.