Hi All,
It may sound like a clustering question but it's also service restart one.
We have a Windows 2008 cluster with one instance of SQL Server 2008R2.
There are 10,000+ databases on it. While user_db access is fast, cluster failover or server/service restart takes almost an hour, even
on high performing hardware. We would like to drop the failover time.
1/3 of the databases are not active.
Almost all of them are on SIMPLE recovery mode.
I am considering;
- Setting AUTO_CLOSE ON for inactive databases, or OFFLINE
- Playing with recovery interval
- Moving to SQL 2012 Availability Groups instead of failover clustering.
Anything else you might suggest?
Cheers,
Kuzey