topic Re: Windows server 2003 SP1 breaks clustering... in Windows Server 2003

Windows server 2003 SP1 breaks clustering...

mujzeptu — Thu, 20 Oct 2005 12:53:17 GMT

I have a windows server 2003, 2-node cluster setup for failover filesharing. They are MLM370 nodes, in a fabric, connected to a VA7400 array.

All was fine and well in the world until I try to upgrade to SP1. Once I do this the cluster service fails to start. I have to wait for the service to fail and windows to finally let me login. Once I try to manually start hte cluster service it says I dont have valid permissions. So I switch it to the local account, and then it just says it cannot see the hardware device, cant see the cluster on the network (I am assuming cant find the quorum partition)

I try fixing quorum, I make sure the nodes can see the array still through command view, etc. Still no luck. Once I do an un-install of SP1, the nodes come up fine, clustering starts no issue, and all is well again.

I cant find anything online about SP1 except people talking about it breaking exchange/application server.

Any thoughts?

Re: Windows server 2003 SP1 breaks clustering...

Roger Faucher — Thu, 20 Oct 2005 13:09:31 GMT

Adam:

Do you have ICF enabled? Look here:

http://support.microsoft.com/default.aspx?scid=kb;en-us;883398

Make a great day!

Roger

Re: Windows server 2003 SP1 breaks clustering...

mujzeptu — Thu, 20 Oct 2005 13:22:35 GMT

ICF is turned off on both nodes. Does SP1 turn it on once installed?

Re: Windows server 2003 SP1 breaks clustering...

Roger Faucher — Thu, 20 Oct 2005 17:30:53 GMT

Adam:

Not to my knowledge. Did you read the entire article and perform the SCW functions suggested in the article?

Roger

Re: Windows server 2003 SP1 breaks clustering...

mujzeptu — Tue, 01 Nov 2005 14:41:06 GMT

Sorry for the delayed response.

I tried SP1, I installed SCM and set it up appropriately based on the given document. It still does not work.

Once sp1 is installed I get errors about the scsi mappings to the array and losing persistent reservations. However, the switch is not setup with persistent reservations. It also gives me an error saying the given user doesnt have rights to start the clustering service. It is a domain service account with admin rights to the local machine...

Also, when the service is switched to a local system account, it takes 20-30 seconds for the service to attempt to start, then says there is an error connecting to attached storage.

Any thoughts?

The firmware on the va7400 is older, could this be related?

Re: Windows server 2003 SP1 breaks clustering...

David Filas — Wed, 02 Nov 2005 21:42:16 GMT

Check this: http://support.microsoft.com/default.aspx?scid=kb;en-us;269229

You will probably have to assign the cluster service account some of these permissions manually. Most likely it's mission the permission to act as part of the operating system.

The lack of that exact permission has caused me the exact same problem you are having when upgrading some of our clusters to SP1.

Microsoft greatly increased how sensitive cluster services is to permissions in SP1, although the events/errors produced by cluster services remain useless and cryptic at best.

Re: Windows server 2003 SP1 breaks clustering...

Bcnbit — Thu, 24 Nov 2005 12:20:51 GMT

I've got the same problem. The cluster worked perfectly but when I'd installed the SP 1 it brokes down.

I'd talked to Microsoft and tell me that the RPC and MSDTC services were not working correctly with the SP1. Then, he send me the fix for the SP1 (KB896729).

I'd put this fix and now is working perfectly. To download this fix I talked to Microsoft and they send it by mail.

See this link http://support.microsoft.com/kb/896729/en-us

I've got a SQL and IIS cluster. I hope this solve your problem.

regards