SQL Server Performance Forum – Threads Archive
quorum
if you were making a two node cluster, where does the quorum reside? I can’t seem to find a clear answer to this in documentation.For example, would there be a quorum disk on both servers and then data is replicated to the passive node so it is up to date for the time a failure? Or is it done with a third server that has nothing to do with the cluster really? Can someone explain this to me?
Check: http://windowssdk.msdn.microsoft.com/library/en-us/mscs/mscs/quorum_resource.asp
http://www.microsoft.com/technet/pr…ons/1f6ff06c-88dc-4a51-8d6e-2b5ea4bcb69e.mspx
http://support.microsoft.com/kb/280353
Luis Martin
Moderator
SQL-Server-Performance.com One of the symptoms of an approaching nervous breakdown is the belief that one’s work is terribly important
Bertrand Russell Nunca esperes el reconocimiento de tus hijos, eso ocurrirá luego de tu muerte
All postings are provided “AS IS†with no warranties for accuracy.
The quorum disk is on one of the nodes. It resides in it’s own cluster group with the cluster ip and name. There is no data replication going on in a Microsoft cluster, it’s a failover cluster.
so then what happens if the quorum were to fail? Is that creating a single point of failure within the cluster?
The second link you provided, LuisMartin had a reference to "cluster configuration database" which would be exactly what I would need to remove the danger or the single point of failure with the quorum. Is there a doc on configuring that? I searched in MS but I only found definitions for it; nothing really about the setup, etc.
Your shared storage can already be seen as a single point of failure. To minimize risks you have multiple data paths to it and all disks configured in RAID. The quorum is often a RAID1 for example. In Windows 2003 clustering there is also something that is called "Majority Node Set". Where each node is keeping it’s own copy of the quorum. It’s mostly used when you have nodes that far away from eachother (in another city or country etc). The most common setup is still the single quorum setup. If the node that currently own the quorum disk fail another node takes over and becomes the owner.
]]>