change corosync timeout

Hi,
I don't know what do you mean?
Do you mean the time when the node will be fenced
or the time when corosync tell you no quorum?

But both are hard coded and can't be set.
 
Hi Wolfgang,
thanks for this topic, we are running 3node cluster over two arista switches with lacp and two weeks ago we got all three nodes rebooted with "no quorum" in logs. Fine. After some googling we assume that some network "lacp packet mismatch" happened and corosync that is set by default too tight/fast (1000ms ?) has rebooted all three nodes.

using https://www.thegeekdiary.com/how-to-change-pacemaker-cluster-heartbeat-timeout-in-centos-rhel-7/ I've filled
to /etc/pve/corosync.conf line "totem: 10000" and I hope that will do

But is still info above about "hard coded corosync" valid?

It seems now it is adjustable by totem value. Or am I missing some point?

ragars
Petr