SQL2014 Failover Cluster config sanity check

• jasona.work

  SSC Guru

  Points: 50083

  More actions

  December 30, 2015 at 7:20 am

  #307410

  I'm planning out what I want to try to get when we migrate our SQL instances to SQL2014 on Server 2012 next year, and for various reasons I'm looking at going with a Failover cluster (not Availability Groups.) What I'm looking for at the moment is a "sanity check" of what I think I need to implement what I've got.

  First, all our servers are virtualized, but we've had one instance of an OS issue on a production SQL, so virtual machine migration wouldn't (and didn't) help. We do have a SAN that I can request the cluster shared volumes be placed on, rather than as virtual disks.

  So, here's what I'd like to do, what I've got already, and what I think I need:

  Like to do, and what I think I need:

  Two server cluster, 5 cluster shared volumes (two for each SQL instance {Data / Log}, one for Filestream), each server hosting one named instance of SQL. One instance will require Filestream enabled. Each server will have sufficient resources to host both instances.

  Have now:

  Two servers, separate virtual disks for data and log files, one server has a Filestream-enabled database.

  What I'm hoping to accomplish with this:

  First and foremost, better protection against OS issues.

  Second, quicker recovery in the event of a physical VM host failure, as I would request the VMs be configured such that they NEVER are on the same physical host at the same time (the VMware Admin has told me this is possible.)

  Third, less weekend time for applying OS updates to the servers, by performing a manual failover of an instance, then patching the now-standby server. I know there will still be an outage while the instance is brought up on the other server, but this can be coordinated with the customer(s) to happen when it would have low impact.

  I'm planning to set up on my home VMs a rough-and-ready test to see if I'm both on the right track for the Filestream, and to get familiar with setting up something like this. Some of it would be handled by our network / storage admins, but the OS side of things is my bailiwick.

  Thanks all,

  Jason A.