high availability watch dog bug

[mora]

I've got 10 nodes in cluster, 3 of them connected in HA group (for example name them 1, 2, 3). If I connect container to this group on server 2 and shut down cluster network on server 3 then:
3 server lose quorum and should reboot, (and it works as suggested)
2 server should work as nothing changes (it's got 9 of 10 nodes in quorum) but it reboots.

How can I fix that?

 

[dietmar]

and shut down cluster network on server 3 then:

What you do mean by "shut down cluster network" exactly?

 

[mora]

What you do mean by "shut down cluster network" exactly?

Cluster nodes contact each other by local network, so I can turn off local interface