We have a pretty simple 3-node multisite DAG setup with DAC mode enabled; two sites, the primary (SiteA) has two mailbox servers and the secondary (SiteB) has one. The cluster is set to use a Node Majority Quorum config:
SiteA
- MBX1
- MBX2
SiteB
- MBX3
On a recent test of our site link system we switched from our primary PTP connection to our VPN backup connection. This resulted in the cluster service stopping on MBX1 and MBX3 which in turn caused an unmountable condition for the Mailbox Databases on all DAG members.
Event ID 7024 is present on nodes MBX1 and MBX3 during the testing:
"The Cluster Service service terminated with service-specific error A quorum of cluster nodes was not present to form a cluster."
After about 5 minutes of going back to the PTP connection, everything works again.
So ignoring the obvious routing issue over the VPN connection, the cluster service should not have failed on the SiteA nodes since that site would have held the node majority (2/3).
This has happened on two seperate tests. Any ideas?
The LAN in SiteA was up and running the whole time, so there was no local LAN failure.
Even better is that there's a huge gap in the cluster log that exempts the date of the site-link testing dates so no clue from there.
Thanks for your time and attention in advance!