July 19, 2012 at 1:17 pm
Does 2012 Always On clustering require the same database restart as Sql 2008's use of clustering? If so, a TB sized database might "fail over" in hours rather than minutes
*****
Give a man a fish, and you've given him a meal. Teach a man to fish and he'll ask to borrow your boat.:-P
July 20, 2012 at 1:07 am
No, databases that are part of an AlwaysOn availability groups are already synchronised so recovery should be fairly quick. AlwaysOn is basically database mirroring on steroids 😉
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
July 20, 2012 at 5:10 am
It depends on two things, neither one database size, transactions in transit and your connection speed. So, no, you should see reasonably fast transitions depending on what's going on with the transactions. My local testing (which is largely meaningless, I'll grant you) is near instantaneous. Much better than anything I've ever seen with failover clustering.
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt
Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning
July 21, 2012 at 11:56 pm
AlwaysOn Failover Clustering Instance (FCI) (which is not the same thing as AlwaysOn Availability Groups) does require the same database start up as earlier versions of failover clustering. So, yes, the time it takes your database to start up is still a factor in failover time.
I used to do operations for a 2 TB database on a cluster, and failover took about 3 to 5 minutes for us.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply