Isolated nodes

[ioanv]

Hi

I'm running a v4 cluster with 7 nodes.
Once every 3-4 days, each node seems isolated from the rest. Howevver communication seems fine (i can login from one another, ping, ssh everything else seems ok. However I cannot manage one node from the interface of another.
To solve this, I shut down all VM and perform a restart of the nodes. Sometimes the situation is remediated only after restarting the first node.
VMs have local storage.



The same problem occurs on a V3 cluster with 5 machines that I run in the same network but the frequency is much lower (1 per month maybe).

Any ideas?

Thanks
Vlad

 

[tom]

Any ideas?

First, update your hosts to latest stable. 4.2-2 is not the current one.

Follow this update howto:

http://pve.proxmox.com/wiki/Downloads#Update_a_running_Proxmox_Virtual_Environment_4.x_to_latest_4.2

 

[ioanv]

OK, I'll do that. One more thing, not sure if it is related.

I perform a backup of all VMs on a NFS storage (FreeNAS). Some VMs seem to fail as I show in this thread:
https://forum.proxmox.com/threads/b...te-write-error-broken-pipe.28480/#post-145182

Every time I find the nodes isolated, I also find the backup process hanged (although it seems to go to the end):