One of my recent requirement is to provide my client with an 100% high availability solution for SQL Database. Thinking about proposing a Clustering solution, 2 Cluster pairs with 2 independent SQL instances running on them, for eg Cluster Pair 1: Node A : 025 N1 Node B : 025 N2 SQL instance : 025SQL Cluster Pair 2: Node A : 026 N1 Node B : 026 N2 SQL Instance : 026 SQL Principal database name : 'Remote db' present in 025SQL Mirrored database name : 'Remote db' present in 026 SQL How should i setup my witness in this scenario for favoring an automatic failover ? Can my Witness be just a standalone VM or Physical server, is there any additional diskprocess requirement to set up a Witness server. Do any of you have any alternate moves for setting up an 100% availability scenario, if so please do help.;HArdware concepts like EMC's SRDF etc or SAN based mirror will be taken from Hardware engineer perspective, am to suggest from DB end only. I know Log ship is out of the box for 100% availability, if there is any difference to Remote based mirror on Cluster , help me know Thanks Eben
I did not understand what you mean by Remote based mirror, please elaborate. If you are configuring a cluster and then mirrored it to another server I guess it is fullfilling HA + DR requirement, why 2 clusters!! One of the project I worked on was having Clustered which is mirrored and then they have storage box replicated at DR site too (hardware level).