Making SQL server service continuously available in Windows Server 2012
-
Saturday, October 13, 2012 10:25 PM
I am planning on making SQL Server continuously available. I read about Windows 2012 Scale out File Server. Okay, I have planned to host my database on SMB 3.0 share, make it continuous available and all but this only makes data access continuously available.
What is the machine where the SQL server service is running crashes ? How can I make the SQL server service continuously available such that if the machine where the service is hosted crashes, it fails over to another machine.
Same with IIS. Can I create virtual directories on SMB continuous available share ?
- Moved by Yan Li_Microsoft Contingent Staff Monday, October 15, 2012 6:46 AM (From:Windows Server 2012 General)
All Replies
-
Sunday, October 14, 2012 1:52 PM
Look into Microsoft Failover Clusters. That is required to create the Scale Out File Server, so you will need to understand clustering to create that. The same software, on another pair of machines, can be used to make SQL highly available. You should read up on SQL clustering, too.
Go to the TechNet library and start reading. It's almost impossible to answer your question in forum because there is so much information that you need.
tim
-
Sunday, October 14, 2012 7:31 PM
I am planning on making SQL Server continuously available. I read about Windows 2012 Scale out File Server. Okay, I have planned to host my database on SMB 3.0 share, make it continuous available and all but this only makes data access continuously available.
What is the machine where the SQL server service is running crashes ? How can I make the SQL server service continuously available such that if the machine where the service is hosted crashes, it fails over to another machine.
Same with IIS. Can I create virtual directories on SMB continuous available share ?
The way you want to go is neither fastest nor the most cost-effective... For SQL Server you really should start looking at native, built-in options like database mirroring and log shipping:
http://msdn.microsoft.com/en-us/library/ms187016.aspx
Making long story short with database mirroring you don't run reads over network (at THAT would be the natural way for SMB 3.0) so storage runs at DAS performance and with local disk latency - that's why it's faster then SoFS. And with both DB and LS you need only a pair of boxes and no shared storage - for SoFS you need an extra pair of Windows Server 2012 licenses and physical servers to run them of course and also either hardware or software providing shared storage for SoFS as they cannot run on their own - CSV should be configured to put shared content on. That's why SoFS approach is at least twice more expensive to implement.
For IIS putting content on failover share is only part of the game - you also need to configure HA / load-balancing for IIS. Take a look at these links:
http://support.microsoft.com/kb/970759
http://blogs.msdn.com/b/clustering/archive/2009/06/01/9674799.aspx
And this is another great and still active discussion started by person willing to do exactly what you want to do - HA IIS and SQL Server:
http://community.spiceworks.com/topic/264123-failover-for-iis-and-sql-do-i-need-to-cluster
A LOT of a valuable feedback.
Hope this helped :)
- Proposed As Answer by VR38DETTMicrosoft Community Contributor Sunday, October 14, 2012 7:32 PM
- Marked As Answer by Shawn Dan Friday, October 19, 2012 6:01 AM

