July 10, 2009 at 1:58 pm
Hey everyone,
We are working on installing SQL Server 2008 SP1 on a development SQL Server Cluster running Windows 2003 R2 SP2 (32bit). We also have a SQL Server 2005 installation on this server that is running SP2 with Hotfix 3159 with no issues.
We have tried numerous times to install 2008 SP1 and each time it corrupts our master database. When we use the RTM install of 2008, everything is up and running successfully. We have tried to the following procedures to install SP1.
1. Open Cluster Administrator and remove the passive node as a possible owner from all the SQL Server 2008 resources.
2. Upgrade the passive node and ensure there are no errors in the Event Log (which it has been clean)
3. Add the passive node as a possible owner to the SQL 2008 resources in Cluster Adminstrator.
4. Move the SQL 2008 Resource Group over to the passive node
5. Remove the now passive node as a possible owner of the SQL 2008 resources in Cluster Administrator
6. Upgrade the passive node and ensure there are no errors in the Event Log (which it has been clean)
7. Add the passive node as a possible owner to the SQL 2008 resources in Cluster Adminstrator.
Each time we try this, at some point following the initial failover, we start receiving the following error message in the Application Event Logs:
Event Type:Error
Event Source:MSSQL$DEVSQL2008
Event Category:Server
Event ID:3417
Date:7/10/2009
Time:3:48:39 PM
User:N/A
Computer:DEVSQL2008
Description:
Cannot recover the master database. SQL Server is unable to run. Restore master from a full backup, repair it, or rebuild it. For more information about how to rebuild the master database, see SQL Server Books Online.
For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.
Data:
0000: 59 0d 00 00 15 00 00 00 Y.......
0008: 16 00 00 00 44 00 45 00 ....D.E.
0010: 56 00 53 00 51 00 4c 00 V.S.Q.L.
0018: 32 00 30 00 30 00 38 00 2.0.0.8.
0020: 5c 00 44 00 45 00 56 00 \.D.E.V.
0028: 53 00 51 00 4c 00 32 00 S.Q.L.2.
0030: 30 00 30 00 38 00 00 00 0.0.8...
0038: 00 00 00 00 ....
I've googled this and tried doing the installs many different ways (including stopping the SQL Services before we do the upgrade) and still have the same result. Any help would be greatly appreciated.
Thanks!
Brian
July 10, 2009 at 2:18 pm
I'd suggest calling customer support at this point.
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
July 10, 2009 at 2:22 pm
That is about where I am at right now. Luckily this is the Dev Cluster, so I'm going to just clean up both boxes, get a clean SQL Server 2008 RTM instance up and running, then probably call MS support and see what they say.
If anyone has any ideas though, let me know.
Thanks!
Brian
July 10, 2009 at 2:32 pm
Just one question. Why are you patching the two nodes separately? SQL is cluster aware and should be able to patch both nodes with you run the patch on the active node.
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
July 10, 2009 at 2:43 pm
I have actually found that the SQL 2008 installs vary tremendously from the clustered installs of SQL Server 2005.
In 2005, it would determine what nodes were part of the cluster and install SQL on all the nodes you selected. In SQL Server 2008, you have to install SQL 2008 on a per node basis (which seems like a step backwards to me).
The SP1 upgrade procedures were directly from the SP1 download details as I originally tried running the upgrade on the active node and that failed. So I followed the procedures step by step from MS, which outlined the steps I had listed above to apply SP1 to a clustered SQL Server 2008 install.
Not entirely sure what MS's reasoning was for changing this, but it has certainly added a few steps to the setup process.
July 10, 2009 at 3:07 pm
Interesting. Last cluster patching I did was 2005. I knew the install for the cluster had changed in 2008, didn't realise the patch had changed too.
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
July 10, 2009 at 3:57 pm
Just to make matters worse. After uninstalling SQL Server 2008 all together from our cluster, now the SQL Server 2005 install is experiencing the same behavior. Prior to the SQL Server 2008 SP1 install, all was working just fine, but now we have a corrupt master database in the SQL 2005 instance as well.
What a day....
July 11, 2009 at 4:53 am
I think definitely time to give Product support a call
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
July 14, 2009 at 12:42 pm
I know this is a bit late, have you tried to patch the sql server 2008 RTM with sp1 w/update 2... I have a simliar issue with applying sp1 on a cluster and sadly enough i had to run the sp1 w/update 2 to get the error i needed to trouble shoot my problem.
July 14, 2009 at 1:11 pm
Thanks for the post. We have completely removed SQL from this cluster and are rebuilding from scratch again (good for documentation purposes).
I will try and download that update and give that a shot. Thanks for the help, I'll let you know how it goes.
July 14, 2009 at 1:25 pm
I feel your pain =/, I've already gone through different scenario's with clustering 08... Best option so far is a fresh OS install, everything else has always caused different errors and long phone calls with MS support.
PM me anytime if you come across any errors I have a 14 page document of errors and fixes related to sql server 2008 cluster upgrade paths so far =(
July 14, 2009 at 1:41 pm
Since you are rebuilding - I would avoid the SP1 upgrade issue altogether by building a slipstreamed version of the installation media. Here are the instructions I used to update the installation media:
Jeffrey Williams
“We are all faced with a series of great opportunities brilliantly disguised as impossible situations.”
― Charles R. Swindoll
How to post questions to get better answers faster
Managing Transaction Logs
July 16, 2009 at 12:07 pm
Jeffrey Williams (7/14/2009)
Since you are rebuilding - I would avoid the SP1 upgrade issue altogether by building a slipstreamed version of the installation media. Here are the instructions I used to update the installation media:
Just wanted to say thanks, I tested this on a win 08 VM 😎 +3 kool pts
July 21, 2009 at 10:51 am
Just in case someone happens to stumble across this issue, here is some clarification on this issue after working with Microsoft support.
1. SQL 2008 has been changed in the way that you install SQL 2008 in a clustered environment. It is still cluster aware, but MS has changed their stance on "pushing" the installs \ updates to all the nodes in a cluster and rather, will have you do the installs \ upgrades on a node by node basis. Essentially this was done to help simplify troubleshooting as in the event an install fails, you will still have another node in tact that resources could be failed over to and brought back online. In addition, you do not need to modify the clustered resource name that I mentioned in a previous post. This is only necessary if you are going to be doing a rolling update.
2. As for the errors I had, it was related to a temp database not being properly deleted. After reviewing many of the bootstrap setup logs with support, we found the following errors:
Error: 5170, Severity: 16, State: 1.
2009-07-20 18:24:26.96 spid7s Cannot create file 'H:\MSSQL10.DEVSQL2008\MSSQL\DATA\temp_MS_AgentSigningCertificate_database.mdf' because it already exists. Change the file path or the file name, and retry the operation.
2009-07-20 18:24:26.96 spid7s Error: 5170, Severity: 16, State: 1.
2009-07-20 18:24:26.96 spid7s Cannot create file 'H:\MSSQL10.DEVSQL2008\MSSQL\DATA\temp_MS_AgentSigningCertificate_database.mdf' because it already exists. Change the file path or the file name, and retry the operation.
We went out to that shared path, renamed the file and then started the upgrade again and then it completed successfully. Once that was finished, the SQL 2008 clustered resources all came online just fine and we are now running SQL 2008 SP1 in a clustered environment.
July 21, 2009 at 3:34 pm
Glad to hear that you've resolved the problem.
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
Viewing 15 posts - 1 through 14 (of 14 total)
You must be logged in to reply to this topic. Login to reply