Reboot of all cluster nodes after lost network/quorum

[hape]

Hello all,

does a VE-cluster, or all cluster-nodes, reboot automatically after they lost the network-connection on which the cluster-quorum is running?

I've updated a router and just in that moment the complete cluster, with a configured ceph-filesystem on a seperate net, have rebooted immediately.

All i want to know is: Is that a normal thing, and how can i prevent such a reaction?

Regards

Hans-Peter Straub

 

[dietmar]

does a VE-cluster, or all cluster-nodes, reboot automatically after they lost the network-connection on which the cluster-quorum is running?

Only if you run HA - this is called sef-fencing. Please use redundant network to avoid such situations.