Saturday, January 28, 2012

Disaster Recovery - Server down

Scenario: For some reason, if the SharePoint installation gets corrupt on one of the front end's or application server in a SharePoint farm. If this is a front end server, users will get service errors when a request is sent to the faulting front end. If there is a load balancer that would greatly help the situation.


*If you have a load balancing solution, remove the faulting server from the load balancer first.

1. Uninstall SharePoint binaries from the server. If Project Server binaries are installed, uninstall that also from the Control Panel. This will automatically remove the server from the farm.

2. Reboot server
-After reboot open IIS and check if there are any orphan sites. If you see something like IIS_Site0, click on the Sites tab above it and remove the site.

3. Reinstall all binaries on the server, e.g.
   a. SharePoint binaries
   b. Project Server binaries
   c. SharePoint Foundation - SP1 
   d. SharePoint Server - SP1
   e. Foundation - June CU
   f. Server - June CU

4. Run psconfig using Configuration Wizard or powershell and choose to add the server to an existing farm. Provide configuration database information. The server will automatically be configured and added to the farm. If there is no other server in the farm hosting central admin, then this server will automatically be enforced to host central admin.

5. If custom solutions were installed on the server, during psconfig all the 14\hive folders will be automatically be provisioned. If any custom changes were made not using wsp's in the 14 hive then they will have to be manually re-done as the re-install process wipes out the 14 hive.

6. If the faulting server was the server where the search admin component was hosted and also if it served the query role, search will be broken. 
To restore the admin component use the below script (assuming there is only 1 search service application)

$varInstance = Get-SPEnterpriseSearchServiceInstance -local  
$varSearchApp = get-spenterprisesearchserviceapplication
set-spenterprisesearchadministrationcomponent –searchapplication $varSearchApp –searchserviceinstance $varInstance

7. Sometimes the crawl component may get stuck in a Recovering state. To resolve this, create a new crawl component with the same parameters on the same server. Once this is complete, you will notice that the old and new crawl component both come online. Delete the old crawl component.


1 comment:

  1. Thanks for best information about disaster recovery server. Really nice and helpful post. thanks again!! :)