Cluster service configuration wizard does not start Cluster service (296594)


The information in this article applies to:

  • Microsoft Windows 2000 Advanced Server
  • Microsoft Windows 2000 Datacenter Server
This article was previously published under Q296594

SYMPTOMS

When you configure the Cluster service on a Windows 2000-based computer, the setup wizard reaches the final completion stage and then you receive the following error message:
Cluster Service Configuration Wizard

Cluster service could not be started. Cluster service Configuration has completed successfully, but the cluster service failed on startup. Check the event log for the reason. Before attempting to run Cluster service Configuration again, you must uninstall the cluster software from the Add/Remove Programs applet in the control panel.
The following events may be logged in Event Viewer:
Event ID 7023:

The Cluster Service terminated with the following error. The quorum disk could not be located by the cluster service.

Event ID 7031:

The Cluster Service service terminated unexpectedly. It has done this X time(s). The following corrective action will be taken in 60000 milliseconds: Restart the service.

Event ID 1009:

The Clustering Service could not join an existing cluster and could not form a new cluster. The Clustering Service has terminated.
If you set the Cluster disk and Cluster service to Disabled, and you restart the node while the other node is powered off, you can access the hard disks by using Windows Explorer or by accessing the My Computer folder. The MSCS folder may be empty, but the Clusdb file exists; it is complete and it is loaded in the registry.

CAUSE

This error occurs if the external disk arrays are connected to a Dell PERC RAID controller with Write-Back caching enabled and Cluster Option disabled.

RESOLUTION

To resolve this issue, disable Write-Back caching and enable Write-Thru caching on the PERC cards of both nodes. Also, enable the Cluster Mode option.

To properly configure your Dell PERC RAID controllers, follow these steps for each logical drive in the array:
  1. During POST, press CTRL+M when you are prompted to enter the PERC BIOS Configuration Utility.
  2. In the Management menu, click Objects\Logical Drive/View\Update Parameters.
  3. Note If you have a PERC4 controller, you may skip this step and execute step 4.
    Click Write policy, and then press ENTER. Configure the policy for Write-Thru. This option should now appear as:

    Write Policy = WRTHRU

  4. On the Management menu, click Objects\Adapter and then press ENTER. Click Cluster Mode, and then press ENTER. Change the option to Enabled. This option should now appear as:

    Cluster Mode = Enabled


    Note If you enable Cluster Mode when you configure your Perc Controllers, you are automatically prompted to enable the Write-Thru cache. When you enable the Write-Thru cache, the PERC BIOS Configuration Utility resets the Initiator ID (SCSI ID) to the default of 7. The Initator ID must be unique on each node. Otherwise, disk reservation will be lost when you power up any node that is connected to the shared resources.
  5. Save the configuration, exit the utility, and then restart the computer.

MORE INFORMATION

If you attempt to start the Cluster service by using the command

c:\winnt\cluster>clussvc -resetquorumlog -debug

at a command prompt, you may see output similar to the following text:

[ClRtl] Load WMI for ClusSvc. Status 126.
[DM] DmInitialize: The hive was loaded- rollback, unload and reload again
[JOIN] Spawning thread to connect to sponsor 192.1.1.1
[JOIN] Asking 201.99.99.42 to sponsor us.
[JOIN] JoinVersion data for sponsor 192.1.1.1 is invalid, status 1722.
[JOIN] Unable to connect to any sponsor node.
[INIT] Failed to join cluster, status 53
[FM] arbitrate for quorum resource id 3b1a7008-a269-4c93-a89d-f5b21a0722ca.
[FM] FmGetQuorumResource failed, error 170.
[INIT] ClusterForm: Could not get quorum resource. No fixup attempted. Status = 5086
[INIT] Failed to form cluster, status 5086.
[CS] ClusterInitialize failed 5086
[INIT] The cluster service is shutting down.
[FM] Shutdown: Failover Manager requested to shutdown groups.
[MM] MMLeave is called when rgp=NULL.

Modification Type:MinorLast Reviewed:1/27/2006
Keywords:kberrmsg kbprb kbsetup KB296594
©2004 Microsoft Corporation. All rights reserved.