Cluster reboot on loss of connectivity

maxim.webster

Active Member
Nov 12, 2024
252
115
43
Germany
So,

this morning at 3am my 3-node-homelab cluster running PVE9 rebooted.

At first i considered a power outage, but that would have caused traces on other machines and household devices.

Turned out, the switch all nodes are connected to received a firmware update and rebooted. That obviously triggered the cluster's reboot.

Now, that doesn't happen frequently, but is this a configurable reaction from the cluster, as I would love to increase the „toleration time for network outage?
 
If you have HA managed resources on your cluster, the reboot is intended.
You can prevent this by providing a second Corosync link, which of course is running over another switch.
 
  • Like
Reactions: maxim.webster