SQL Server - CLuster Servicee error - Asked By Kanagasabapathy T on 17-May-10 07:48 AM

Hi ,

We are using SQL server 2005,SP2 two node active active cluster ruuning on Windows server 2003.

On node 1 , the Cluster Service service is down. It's throwing the below error msg when starting up the cluster service.


Could not start the Cluster Service service on Local Computer.Error 1067: The process terminated unexpectedly.

On Cluster Administrator,  I could not see the node 1 at the below (only node 2 (server name)  and the move group option on SQL instance was grayed out ( not useable ).

Cluster Error: 

[CS] Cluster Service started - Cluster Node Version 4.3790

                               OS Version 5.2.2026 - Service Pack 2 (ADS 03000112L)

                               Local Time is  2010/05/17-20:19:09.897

[INIT] ClusterInitialize called to start cluster.

[EP] Initialization...

[DM] Initialization

[DM] DmInitialize: The hive was loaded- rollback, unload and reload again

[DM] DmpRestartFlusher: Entry

[DM] DmpUnloadHive: unloading the hive

[Qfs] QfsSetFileAttributes C:\WINDOWS\Cluster\CLUSDB.BKP$ 80, status 2

[Qfs] QfsDeleteFile C:\WINDOWS\Cluster\CLUSDB.BKP$, status 2

[DM] Loading cluster database from C:\WINDOWS\Cluster\CLUSDB

[DM] DmpStartFlusher: Entry

[DM] DmpStartFlusher: thread created

[NM] Initializing...

[NM] Local node name = NFTSQL01.

[NM] Local node ID = 2.

[NM] Failed to open node key, status 2

[NM] Initialization failed 5052

[NM] Shutdown starting...

[NM] Interface cleanup starting...

[NM] Network interface cleanup complete

[NM] Network cleanup starting...

[NM] Network cleanup complete

[NM] Node cleanup starting...

[NM] Node cleanup complete

[NM] Shutdown of the Cluster Network driver failed, status 21

[NM] Shutdown complete.

[CS] ClusterInitialize failed 5052

[INIT] The cluster service is shutting down.

[FM] Shutdown: Failover Manager requested to shutdown groups.

[Dm] DmShutdown

[DM] DmpShutdownFlusher: Entry

[DM] DmpShutdownFlusher: Setting event

[DM] DmpRegistryFlusher: got 0

[DM] DmpRegistryFlusher: exiting

[MM] MMLeave is called when rgp=NULL.

[CS] Service Stopped. exit code = 5052

Please let me know on how to proceed further regarding this issue.

Thanks,
Kanagasabapathy. 


 

Web Star replied to Kanagasabapathy T on 17-May-10 01:21 PM
The error message can occur if the quorum log file (Quolog.log) cannot be found or read at Cluster service startup. This file must be readable to confirm that the cluster configuration on the local node is current. If the Cluster service cannot read the log, the service is not prevented from loading stale configuration data.

The log file may be corrupted if the time and date stamp of the Chkxxx.tmp file does not match the stamp of the Quolog.log file in the \MSCS folder on the quorum drive. The corruption can occur if you recently performed a restore operation, or if you experienced a power outage.

http://support.microsoft.com/kb/257905#top

RESOLUTION
To resolve this issue, start the Cluster service with the -ResetQuorumLog switch...

To resolve this issue, start the Cluster service with the -ResetQuorumLog switch:
  1. Start the Computer Management snap-in.
  2. Open the properties of the Cluster service.
  3. In the Start Parameters box, type -ResetQuorumLog, and then start the Cluster service.
  4. To make sure that everything is running correctly, stop the Cluster service, and then start it again without any parameters. If this procedure does not start the Cluster service, you may have to completely re-create the checkpoint and the Quolog.log file from the local cluster registry. To do so:
    1. Start Computer Management.
    2. Open the properties of the Cluster service.
    3. In the Start Parameters box, type -NoQuorumLogging, and then start the Cluster service.
    4. Start Windows Explorer, and then move the Quolog.log file and all the Chkxxx.tmp files that are in the MSCS folder that is on the quorum drive to the root of the quorum drive so that they do not exist in the MSCS folder.
    5. Switch to Computer Management, and then stop the Cluster service.
    6. In the Start Parameters box, type -ResetQuorumLog, and then start the Cluster service again.

      Note If the cluster is running Windows 2000 Service Pack 4 (SP4), and if the KB 872970 hotfix is installed, do not use the -ResetQuorumLog switch. You may start the Cluster service normally.
    7. Stop the Cluster service, and then start it again without the startup parameters.

details is here http://support.microsoft.com/kb/257905