Last weekend I completed my first clustered SQL Server 2008 instance install, as I have blogged previously, this was into an existing Windows 2003 Server – SQL Server 2005 cluster. I got stung by several gotchas.
First is: don’t forget the pesky filestream hotfix for Windows 2003 server. You need this for installing SQL Server 2008 even if you aren’t using filestream – and it doesn’t error on this until right at the end of the setup choices, just before installation. So annoying, so frustrating, and so easy to forget., Add it to your list of prerequisites/preparation steps – and it needs a reboot, just like .Net Framework 3.5 SP1, and the Windows Installer hotfix.. grrr …. I have blogged about this before, and the relevant MSKB Hotfix is 937444.
I encountered one major problem – I got the following error (setup failed on the first node)
The computer account for Cluster resource ‘SQL Network Name (SQLNAME)’ in domain domain.com.net could not be created for the following reason: Unable to create computer account.
The text for the associated error code is: Your computer could not be joined to the domain. You have exceeded the maximum number of computer accounts you are allowed to create in this domain. Contact your system administrator to have this limit reset or increased.
The Cluster Service Account may lack the proper access rights to Active Directory. The domain administrator should be contacted to assist with resolving this issue.
Follow the instructions from this article, under the first section marked ‘To verify that the Cluster service account has the Add Workstations to a Domain user right’. Tedious, annoying, and easy to fix, and it took me a whole week to fix it (I had to play cricket, and couldn’t spend the whole weekend researching the issue!)
Second weekend, and second attempt, on Node 2, (Node 1 now fixed as above), got nearly to the end of setup, it failed with the error ‘Not enough resources’. BAH! I had left the 2 existing SQL Server 2005 instances running on this Node with fixed Max Memories and I think the setup ran out of memory. Had to fail over these 2 instances to the passive node, then uninstall the instance totally – it left it installed, but kind of not clustered – more to come in a following post. After getting myself out of this jam, I made sure I got the 2005 instances off Node 3 and Node 4 off before running Add Node.
Finally, all ok – spangly new 2008 (with SP1 sexily slipstreamed already!) instance ready to play with/configure/use/hand over to users.
I love clustering .. when it works…