Cluster Drives not failing over.

[Gavin Yeoh]

Greetings All,

I'm sure this is just a config option I've missed, but any help would be great.

Basically we have an EVA4000 that I have attached 2 BL460C blades too. I've built up a working cluster with Windows 2003 Ent(RC2-SP1) + SQL 2005 Ent(SP2).

When performing a move of the SQL resource group from one node to another, it works perfectly, only taking about 20 seconds to become fully online on the second node.

But when I just pull the power on the active blade, it fails to move to the remaining node. IP address and Network name appear on the second node, but the SQLLog & SQLBackup drives fail to mount on the second node.

If I turn the original active node back on, it picks up the drives straight away and SQL is up an running again.

It looks like the original Active node is locking the SAN drives so that they can't be mounted by the pasive node.

Thanks in Advance.

[Mike Bollman_1]

Gavin, what SP level are you running on the OS, SP1 or SP2?

If you are on SP1 there are quite a few post SP1 hotfixes needed and some are related to this type of issue.

KB908510
KB911030
KB915691
KB916048

all should be looked at closely. As a matter of fact, check this website for details on other hotfixes.

SP1:
http://support.microsoft.com/kb/923830/en-us

SP2:
http://support.microsoft.com/kb/935640

[Mike Bollman_1]

Gavin,
It very well could be a known bug as there are quite a few issues surrounding clusters and failover events.

What SP level is your OS at?

Go here for SP1:
http://support.microsoft.com/kb/923830/en-us

and here for SP2:
http://support.microsoft.com/kb/935640

These websites list the post SP hotfixes that you need to look into.

[Gavin Yeoh]

I'm using Windows 2003 Server Enterprise RC2. It has SP1 built into it and I have not applied SP2 yet.

Thanks for the info Michael, I'll check the links back and let you know how it goes.