Server disconnected (code: 1006)

[masakrator]

I now have this problem after:
upgrading first node to pve 9
adding new node
deleting certs (because was unable to browse second node from gui)
I was then able to manage new node from gui
then I discovered I have error 1006 on shell of node0 (first one) and all VMs console connection fail
I am albe to log in via ssh with no issues so I guess this is pve cert issue
however I am not able to edit known hosts - file is read only
is it safe /ok to force the change ? can I change mod on this file to be read/write ?

 

[Bent]

I now have this problem after:
upgrading first node to pve 9
adding new node
deleting certs (because was unable to browse second node from gui)
I was then able to manage new node from gui
then I discovered I have error 1006 on shell of node0 (first one) and all VMs console connection fail
I am albe to log in via ssh with no issues so I guess this is pve cert issue
however I am not able to edit known hosts - file is read only
is it safe /ok to force the change ? can I change mod on this file to be read/write ?

From what you’ve described, there could be several potential causes for this issue. Let’s start with a few basic checks to narrow it down:

1. Network connectivity:
   Verify that your first node can reach all other nodes via ICMP (ping) without any issues.
2. Hosts file configuration:
   Make sure that the /etc/hosts file on all nodes contains the correct hostname mappings, for example:
   10.127.12.10 node0.example.local
   Each node’s hostname should correctly resolve to its respective IP. Using hostname --ip-address should return the correct IP.
3. SSH access between nodes:
   Test if you can SSH from node0 to your other nodes (node1, node2, etc.) without warnings or errors.
4. Check for related errors:
   Run the following command and look for anything relevant:
   journalctl --boot 0 --since today --no-pager --priority=err
   You can post the output here if you’re unsure how to interpret it.
5. Regarding the known_hosts file being read-only -> Likely due to your cluster not having quorum and pmxcfs being blocked to avoid a slit brain situation.

Let me know if this helps you!