asked
What will happened if quorum would not available in case of Disaster Recovery?
It seems I missed something on DR configuration regarding quorum settings which may impact during Disaster Recovery situation. Please review the given below configuration and advice regarding Quorum configuration which should be best fit for DR conditions and also let me know what steps I need to perform quorum side before or after SQL AG force fail-over to DR site. **SQL 2012 AG WSFC with standalone nodes, 2 node on primary site with synchronized auto fail-over and 1 DR node with A-synchronized manual fail-over (Multi-subnet).** Quorum is configured with file server path. Also I didn’t change any vote and anything and all are with default setting so advice if anything needs to change.
You have some answers below. This web site runs by you voting and marking the best answers. For each helpful answer below, indicate this by clicking on the thumbs up symbol next to those answers. If any one answer lead to a solution, indicate this by clicking on the check mark next to that answer. This helps indicate your solution for anyone else with the same question.
3 node cluster with a file share witness can sustain the loss of 1 node. If the 2 nodes in primary site where to go offline due to a site failure so would your DR node. You will need to look at the possibility of a 4 node cluster with fileshare or a 5 node cluster to always ensure you have a majority vote
You would have to force remove the cluster nodes from the cluster so that you only had one node left in the cluster
This is Availability group configuration and if I will remove the node from cluster then AG will be no longer available. I can do force fail-over to SQL on DR node. I just want to know what are the things I need to do on Quorum file share to point this to new location or force failover to DR site.
Your in a 3 node cluster so why the file share, your generating a 4 way vote quorum which is not generally best practice as you generally need an odd number to get majority vote. What are you going to do in the event your primary site goes offline. You have lost 2 nodes. Even with a file share you still have lost quorum due to not having a majority vote, so the cluster is down. You could look at something like DFS and DFSR to replicate files between sites and point to the DFS share instead of a particular filesystem, but unsure how that would work with quorum. You need to establish a quorum (majority vote) some how after failing over and loosing the primary site. Either force remove the 2 primary nodes and remove the fileshare (perfectly acceptable to have a 1 node cluster) What you might be better of doing is if the DR site is a true DR site and not used in HA, is to remove Node 3's quorum vote that way the 2 primary nodes and fileshare provide quorum. DR node is setup for async commit mode with manual failover not automatic failover.
MS has an interesting solution for your File Share Witness: Cloud Witness, but it requires Windows Server 2016 per latest on MS's site. That has changed recently as it was listed as available for Windows Server 2012 R2 until recently.
http://blogs.msdn.com/b/clustering/archive/2014/11/14/10572766.aspx
19 People are following this question.
