I am getting odd cluster behaviour on my mail cluster.
We have 2 mail servers in a cluster, with no load balancing configured. All clients use NRPC and connect to server 1 as their primary server. They should only failover to server 2 if server 1 is unavailable.
The notes log on both servers is filling up with messages as follows:
“Failing over from Server1!!mail\512622.nsf, directing open to Server2”
This is happeneing for many mailfiles and it is failing over to both server 2 and server 1(itself!)
We cannot find any explanation for this as server 1 is available all the time and we are not using any load balancing.
The only client connectivity is offline inotes users, synchronizing with server1 over NRPC.
We cannot understand what fails over and why. Can anyone help?
Subject: odd cluster failover issue
First of all I would check netstat on port 1352 to see what state its in to determine why the clients are being automatically failed over back and forthe. Check server Stats to see availabity index, it its too low then that would be the reason. Also check server config to see if server is set to restricted mode
Subject: RE: odd cluster failover issue
thanks for the response. Surely the availability index is irrelevent though. I have an availability threshold of zero. Surely there should therefore be no failover, even if the availability index is low because I am not trying to use any load balancing, but only want failover to kick in when the server is completely unavailable. If my threshold is zero, the index should never trigger failover. Am I right in this?
Subject: RE: odd cluster failover issue
also - how do I check whether a server is in restricted mode? I am not familiar with this term.
Subject: RE: odd cluster failover issue
You can see it with console-command "show config server_restricted"0 = no restriction
1 = no user access for the current server-session
2 = no user access permanently
But I dont’t think that this is the reason for your failover-problem.