Sunday, 28 September 2008

Can not start windows 2003 cluster after restore from backup

I have recently been working with backup software and even using system state in windows i have come across problems with restoring windows clusters.
This is a detail of problems with open files not being fully backed up and when restored preventing the cluster from coming online the cluster registry keys are not restored

If the HKEY_LOCAL_MACHINE\Cluster hive is not available or has missing entry's the cluster service will fail, to restore these settings you will need to restore the hive by restoring the clusdb.

The clusdb holds the HKEY_LOCAL_MACHINE\Cluster registry keys the file its self is located in the %windir%\cluster\clustdb this file can be restored from the quorum drive.

To do this you will need to follow these steps


1. Boot node A to form the cluster (with node B powered-off).
2. Right Click on My Computer, select Manage, Click on Device Manager.
3. click on view and then click on Show hidden devices.
4. Click on Non-Plug and Play Drivers, double click on Cluster Disk Driver, click Driver and change the start-up option to Manual, click OK, and then click Close.
5. Click the Services, and then double-click Cluster Service and change the start-up to Manual click OK, and then click Close.
6. Reboot node A (this starts Windows on node A, without loading the Clusdisk.sys file)
7. Once node A is booted you'll be able to access the quorum drive. Copy the latest quorum_device:\Mscs\Chksequential_number.tmp file to %SystemRoot%\Cluster\Clusdb.
8. Open regedit by clicking Start, click Run type regedit and click ok.
9. Expand the registry and select the HKEY_LOCAL_MACHINE\Cluster key then select File and Unload Hive then select the HKEY_LOCAL_MACHINE and click file and load hive now browse to the clusdb file located under %windir%\cluster\clustdb.
10. At a command prompt, type net start clusdisk, and then press ENTER.
11. At a command prompt, type net start clussvc, and then press ENTER.
12. Restart node B to join the cluster, and replicate the Clusdb file from the sponsor node A.
13. On node A, after both nodes are running, Right Click on My Computer, select Manage, click Device Manager.
14. Click on Non-Plug and Play Drivers, Double click Cluster Disk Driver, click Driver and change the start-up option to Automatic, click OK, and then click Close
15. Click the Services, and then double-click Cluster Service and change the start-up to Automatic, click OK, and then click Close.


You are now finished and should have a working cluster once more.






Wednesday, 3 September 2008

SharePoint "Thread was being aborted"

"Thread was being aborted" while enabling Full-text indexing.

This is one of those little known areas of SharePoint, I can think of only to reason that you will come across this message, first is where you have a large SharePoint farm and you are enabling fulltext search for the first time/re-enabling it and the second is if you are enabling it on very slow hardware.

Are to actions you can follow to try to resolve this first is to do the following steps.

On the SQL server stop and start the:
"Microsoft Search service" and "MSSQLServer service"
then open SQL Query Analyzer and for each content database run:

exec proc_DisableFullTextSearch
exec proc_EnableFullTextSearch


This enables the FTS outside of Sharepoint. After these steps are performed, Sharepoint must still be configured to show the search box.
To do this, check the enable full-text indexing component in the central admin.

If the error still occures, increase the time-out value for the Central Admin site.
This can be done by increasing the "executionTimeout" value (default 600 seconds) in the web.config in
C:\Program Files\Common Files\Microsoft Shared\web server extensions\60\TEMPLATE\ADMIN\1033

The line you are looking for is (httpRuntime executionTimeout="600") as you can see the default is 10minutes I had to reset this to 4 hours on one SharePoint farm where the content was over 1TB