When I disconnect node from cluster, HA kicks in, then the disconnected node self reboots

[logui]

I have a 3 node cluster, I have zfs + replication of 2 critical VMs (running on node 3) to all the other nodes (node 1 and 2), also both VMs are included in HA.

They are running on node 3, I intentionally disconnect node 3 from the network (pull the ethernet cable) to simulate a node failure, after 1-2 minutes, node 3 will self reboot or crash, I am not sure. Then, 3 minutes from the initial disconnect the 2 VMs will be up and running on a different node (N1 in this case as configured in HA).

From HA perspective, everything worked as planned and expected, my question is: is the node 3 reboot (or crash) part of the HA process or is it a bug in Proxmox and it's crashing? Thanks

 

[waltar]

It's expected behavior by ha as node cannot reach the other members anymore and try to reconnect by reboot, so all is well.

 

[logui]

It's expected behavior by ha as node cannot reach the other members anymore and try to reconnect by reboot, so all is well.

Great, thanks for confirming my theory, because it was happening consistently all the time, if it's a planned reboot, means it is a safe shutdown and no corruption. I didn't know that an isolated node will proceed that way

 

[UdoB]

I didn't know that an isolated node will proceed that way

For reference: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#ha_manager_fencing