I currently have mirroring setup between three computers, principle, mirror, and witness. During the day, all these systems are fine; they failover with the failover command, and the failover when unplugging any of the stations. My problem is at night I lose Quorum between the three stations, and the only way to get it back up is to restart the witness SQL Server. The worst thing about losing Quorum is that all stations revert to recovery mode which renders the database unusable. It’s almost as if the router connecting all the stations went down.
I have tried computer names and IP addresses without any luck. I have similar setups running at a bunch of other customers, so I don’t think it’s the way we have setup the SQL servers or mirroring. I have the router on a UPS with serge protection and the router has been replaced multiple times, so I know it’s not the router. The mirror station has been replaced, hardware and software, because of some other issues that are still occurring. My only other solution is to replace the witness, but if this doesn’t work, I’m all out of ideas.
Below is a log file of the error that is being produced when quorum fails. This error message is shown on all of the SQL Servers.
09/09/2009 10:07:48,spid20s,Unknown,Database mirroring connection error 4 ‘An error occurred while receiving data: ’64(The specified network name is no longer available.)’.’ for ‘TCP://Silver1:15111’.
09/09/2009 10:07:48,spid20s,Unknown,Error: 1474 Severity: 16 State: 1.
09/09/2009 10:06:16,spid19s,Unknown,Database mirroring connection error 2 ‘DNS lookup failed with error: ‘11001(No such host is known.)’.’ for ‘TCP://Silver2:15111’.
09/09/2009 10:06:16,spid19s,Unknown,Error: 1474 Severity: 16 State: 1.
Also note that this setup is not using certificates or anything. All the users have the same username and passwords, so I know this is a working solution without having to be on the domain.
This pretty much sums up everything. Your networking guys better have an explanation why the DNS name ‘Silver2’ is lost.