Thursday, February 3, 2011

Windows SharePoint Services Search Failure and Errors 10036 and 17836

Introduction

This procedure provides one resolution to a failure in Windows SharePoint Server 2007 Enterprise Search along with the appearance of Errors 10036 and 17836 in the Windows Server 2003 Enterprise Edition Application Event log.  It provides some general background for the error and the steps that were taken to successfully resolve it.

Background

An unexpected power outage led to the abrupt shutdown of a blade server hosting a customer SharePoint 2007 instance.  Both the SharePoint 2007 instance and its SQL Server instance were co-located on this blade.  The outage occurred over a weekend.  The blade server was restarted the following business day, and the SharePoint instance restarted automatically, presenting all pages normally.  In all ways, the SharePoint instance appeared to be fully operational.  A few business days later, a user attempted to perform a search and immediately experienced an error page.  Further attention by technical support verified the user experience and the matter was placed under investigation.

Review

SharePoint Shared Services was found to be running.  The crawl record showed that crawls executed previous to the outage had completed successfully without unexpected errors.  Crawls were configured to execute daily.  No crawls were logged after the power outage.  A manual crawl was attempted but was unsuccessful.

A review of the Windows System Event log found no unexpected errors.  A review of the Windows Application Event log identified the appearance of several unexpected warnings and errors:
  • Error 17836: MSSQLSERVER
  • Error 10036: Office Server Search
  • Error 10036 Windows SharePoint Services 3 Search
  • Warning 2423: Windows SharePoint Services 3 Search
  • Warning 2423: Office Server Search
These errors and warnings appeared at the time of the outage but did not recur.

A review of Windows Component Services found the following:
  • Office SharePoint Server Search | Started | Automatic...
  • Windows SharePoint Services Search | [Stopped] | Manual...
These findings suggested that there was database corruption, specifically the search index database.  Given these findings, the following resolution steps were performed.

Method

The Office SharePoint Server Search service was restarted. The Windows SharePoint Services Search service was started and set to Automatic.  A search was then performed, in the SharePoint 2007 website, but was found to still be unsuccessful, returning an error page.

In SharePoint Shared Services, a Reset all crawled content was performed.  After the reset completed, a new crawl was initated manually. The crawl completed successfully.  A new search was then performed, in the SharePoint 2007 website, and this time it was completed successfully, returning expected results.  The issue was found to be resolved.

Summary

From review of the findings and resolution steps, the following is our best understanding of the sequence of events involved.  A power outage apparently led to a corruption in the SQL Server database servicing a SharePoint 2007 instance.  The specific database corrupted involved the SharePoint search index database.  As a result of the search index corruption, a routine scheduled crawl failed, leading to a shutdown of the Windows SharePoint Services Search service, which stopped further crawls being performed.  The stoppage of the SharePoint Services Search service caused users to experience an error whenever searches were performed.

References

No comments: