November 13, 2012 at 1:16 pm
Hey folks,
I am going to be setting up a new sql instance in the next while at work and I want to make it have high availability/fail-over. I want to put the database files up on our SAN and have 2 physical servers clustered together.
Is there a better way to do this, perhaps utilizing mirroring with SQL (though I believe that would eliminate the ability to use the SAN) or some other technology that would produce good performance?
I know this is a pretty broad question but I want to see what others think before I pull the trigger on this.
Thanks
November 13, 2012 at 2:16 pm
Not really enough info there.
To decide on a HA/DR setup you need to start with the data loss and downtime SLAs and go from there, also budget and some idea of what kinds of disasters the solution has to withstand. Also, clustering has a single-point of failure in the storage and no redundant copy of the data. Log shipping and mirroring have two (or more) copies of the data.
Ha/DR setups have nothing to do with performance. You cluster or mirror or logship to survive the failure of some component or components
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
November 14, 2012 at 8:37 am
The SAN has redundancy built in so that is why I am leaning towards it.
November 14, 2012 at 8:45 am
chrisph (11/14/2012)
The SAN has redundancy built in so that is why I am leaning towards it.
Err......
Check with your SAN vendor that whatever form of 'redundancy' they have is certified to work with SQL Server databases, otherwise your 'redundant' copy may be a waste of bits.
Everything that I said about HA/DR stands, redundancy in the SAN or not. HA/DR design starts with your SLAs, not your hardware.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
November 14, 2012 at 9:05 am
Gail wasn't advocating not using a SAN. She was merely saying not to count on the SAN's redundancy to save you in a disaster. If the whole SAN goes down, built-in redundancy is gone too. Mirroring and log shipping gives you multiple copies of the data, but if you put all partners on the same SAN, you lost that redundancy as well if you lose the whole SAN.
As Gail stated, it depends on what you are trying to protect against and what your SLAs are (how much data can you lose and how long can you be down for in the case of a disaster).
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply