Is there a way to have a smooth cluster fail over from the users perspective?
If server A gets down and server B is still up they will get a server not responding message.
http://www-01.ibm.com/support/docview.wss?uid=swg21086484
Is there any settings I can tweak?
Subject: No real changes in 8.5 in this area
So called - smoother Cluster Failover is something that is being investigated for a future release…
For now, the general rule of thumb is when user get Server Not Responding instruct them to exit from the database they are currently working in and re-open and it will fail-over at that point.
Subject: replica
If they are in the mailbox they got a respons like; “Do you want to switch replica”.
Ok, thanks for answer.
Subject: Silent Cluster Failover…
I thought I read somewhere that either 8.5 or 8.5.1 had an INI setting that allowed for a cluster failover without the user being prompted.
Am I mistaken?
Subject: What about Smooth Fail back?
One of our servers was down last week and like I would expect users failed to a cluster server in Tulsa. Looks like they were using mail on that server all week. Now Tulsa has a power outage and when they logged in today they got a “Server not responding message”. Why didn’t the client just know that Tulsa server was unavailable and fail over back to their home mail server which is a cluster member? I had to get them to manually open replica of their mail server to Conway server… Why? I looked at my own Cluster.ncf and it showed a date a week old. Any ideas?