November 20, 2019 at 7:03 pm
Hi,Currently we have a sql failover cluster with an attached iscsi target. This consists of 3 windows server 2012 R2 machines w/SQL server 2012. We are looking to move to an always on group instead. Can this be accomplished with using the same sql servers in the cluster? What is the best path to accomplish our goal?
November 21, 2019 at 8:49 am
You can't use shared storage with Availability Groups, so that might answer your question, not sure.
November 25, 2019 at 6:52 pm
Here's one way that requires a temporary 3rd node: add a standalone SQL server; configure new AG between your existing FCI and the standalone; evict one of the FCI nodes and rebuild it (with new non-shared storage) as a 2nd standalone and add it to the AG; finally retire/decommission last remaining FCI node and the shared storage. Note there is no automatic failover between the FCI and AG, so once you evict the first FCI node your HA is manual until you finish with the 2nd AG node.
PS - You may be tempted to be even more efficient, but be aware that even though it's possible to add a new standalone SQL instance to one of your servers hosting the FCI, I've learned the hard way SQL Server will not allow you to have the same AG database in two different SQL instances running on the same physical node.
PPS - After using FCIs for HA along with log-shipping or block-based DR solutions for many years, back in 2012 we switched to sync AGs for HA and async AGs for DR. Worked pretty well. But by 2014-2015 we realized we preferred the FCI HA experience, and converted back. Now our preferred approach is FCIs for HA, async AGs for DR--this has worked very well. I expect this is an unpopular opinion.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply