Ask a question

Search Forums:

• Search SharePoint - Search Forum
• Search All SharePoint Products and Technologies Forums
• Search All Microsoft TechNet Forums

 

Weirdest MOSS Crawl problem ever!

• Monday, November 05, 2007 4:02 AMRanbou

   
  

  0

  Sign In to Vote
  I am encountering an incredibly weird problem while trying to build my index for a Microsoft Office SharePoint Server (MOSS) 2007 Web Application. I set my Content Source to the top site of the application and kicked off a full crawl. For a period of time the IIS CPU is pegged at 100% and everything is crawling nicely. The Search Settings page in Central Administration shows it is crawling my Content Source and refreshing the page shows the Items in index count steadily rising. This is where things get weird and the crawl hangs.
  
  I will do my best to describe the cycle.
  

  • The CPU utilization on the IIS box falls to nothing and the W3WP.EXE process is no longer getting CPU time. If left alone it would stay in this hung state.
  • At the same time the SQL Server 2005 database starts throwing errors. The exception log shows "spid [number] Exception 0xc0000005 EXCEPTION_ACCESS_VIOLATION reading address 00000000 at 0x0110577" and the SQL error log shows "A user request from the session with SPID [number] generated a fatal exception. SQL Server is terminating this session. Contact Product Support Services with the dump produced in the log directory."
  • The error on the SQL Server I described above happens approximately every .04 seconds and it creates a SQLDump[number].log, SQLDump[number].mdmp and SQLDump[number].txt file. I can post these files if needed. Since the files are created so quickly it eats up all of the available space on the network share and effectively brings SQL to grinding hault.
  • Here is the weird part. While the crawl is hung and the SQL box is spitting out errors, I can go to the Search Settings page, click on the Authoritative pages link, make sure the Refresh now check box is checked, click the OK button and it will temporarily unhang the crawl. Once you click the OK button it returns you to the Search Settings page. You can see in the Indexing Status indicator is has appended | Computing ranking on the end.
  • Once the ranking is finished computing (usually 5 - 10 seconds) everything returns to normal. The crawl starts working, the IIS CPU utilization pegs at 100%, the Items Indexed count starts rising and the SQL errors stop.
  • The nirvana lasts for a few minutes (anywhere from 1 to about 15) and then the cycle starts over.

  Any help solving this issue would be greatly appreciated. It is definitely a head scratcher. I can provide more information, if needed, but I wasn't sure what would be relevant. The SQLDump[number].txt files has a lot of information in it, but it doesn't have any meaning to me.
  
  I thank you in advance for any help you can provide.
  
  • Reply
  • Quote