Link to home
Start Free TrialLog in
Avatar of osiexchange
osiexchange

asked on

How to stop a 2008 Cluster running Exchange 2007 SP1 from failing over

Where is the setting to keep a Cluster from failing over if the Passive node experiences a network issue. Here is what happened. I have a two node CCR Cluster. Windows is 2008 SP2 and Exchange is 2007 SP1, update rollup 9. Node 1 is the Active node and Node 2 is the Passive node. We are running on ESX 4.0 and all the servers are virtual. I decided to update the NIC drivers on the Passive node only. I figured because it was the Passive node, I could without an issue. Well, the drivers installed just fine but caused a brief network connectivity problem on Node 2 which was to be expected. This caused Node 1 to fail over all its resources to Node 2. We are up and running on Node 2 now without a problem but if I try to update the NIC drivers on Node 1, I am sure the same thing will happen. I have been poking around in the settings but wasn't sure what to do. I did not know if there was any special process to follow to keep this from happening or is there is some check box that would prevent this.
Avatar of Jamie McKillop
Jamie McKillop
Flag of Canada image

Hello,

Where is your file share witness located?

JJ
Avatar of osiexchange
osiexchange

ASKER

My file share witness is located on my Hub Transport server which of course is separate from my Cluster server
Open Failover Cluster Management and click on your cluster name. Beside "Quorum Configuration" what does it say?

JJ
Node and File Share Majority: \\servername\share
You have it configured incorrectly. It should say "Other (File Share Witness unc path)". Under your current configuration, a failure of any single node will cause a fialover. Click More Actions - Configure Cluster Quorum Settings and set it up correctly.

JJ
Well, that is exactly what I want in most situations but I was just wondering if there was a setting, besides this particular one, that would prevent that from happening. I don't want to mess with this setting because it supposed to be this way according to this article.

http://www.msexchange.org/articles_tutorials/exchange-server-2007/high-availability-recovery/deploying-exchange-2007-sp1-ccr-cluster-windows-server-2008-failover-cluster-part1.html

Sorry, you are correct. I read your post as just "Node Majority" for some reason. Under your current configuration, your cluster should not have failed over. When you say you updated the NIC drivers, do you mean on the ESX host or did you update the virtual drivers in the guest?

JJ
On the virtual Exchange server (guest), not the ESX host. I though I could do this also so I was surprised to see if failover like it did. What I am guessing happened is the NIC tempoarily lost connection with the network because of the driver upgrade. The host being upgraded took this as the other node going offline because it lost contact with it and requested a failover of resources even though the other node was perfectly fine. Not sure how to appoach this particular problem.
ASKER CERTIFIED SOLUTION
Avatar of Jamie McKillop
Jamie McKillop
Flag of Canada image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial