in

SharePoint Blogs

The Best Place for SharePoint-related Blogs

Russ Houberg's SharePoint Blog

Unlocking the hidden power of SharePoint for document imaging.

SharePoint Database Indexes and Statistics

In the storage whitepaper that I recently released, I talked a bit about the SPDatabaseStatisticsJobDefinition timer job and the fact that it can, in most cases keep SharePoint running smoothly.  However, I recently had to research a bit further into what exactly this timer job does.  This is what I discovered:

This timer job is associated with each web application in the farm.  It is automatically created and configured to run weekly when the web application is created.  I mentioned that it might be a good idea to run some sort of statistics refresh after a large migration (into SharePoint) or similar operation.  Well, the interesting thing is that I traced this timer job definition all the way down to a stored procedure that exists in every content database.  The stored procedure is called proc_UpdateStatistics. 

So if it were my farm, I might be inclined to try to modify the schedule for this timer job but this may not be possible (haven't had time to check).  The other possibility is to just execute the stored procedure manually but that might not be supported by Microsoft.  Interestingly, according to this KB article, Microsoft specifically says that it's OK to update statistics, so maybe it is supported.

Anyway, I just thought all of that was interesting.  On a related note, I want to clarify, that this timer job runs the statistics update on CONTENT DATABASES.  So it is highly likely that you will need to create an additional maintenance plan specifically for the SSP search database.  If the search database is large and/or supports millions of documents, it's a good idea to keep it cleaned up regularly. 

Finally, a bit of a warning.  No matter how you manage your indexes/statistics refresh, you don't want to be updating statistics any time a crawl is running.  Particularly if the crawl has a lot to do (full crawl), if you're trying to update statistics at the same time, you could end up grinding the SQL Server into oblivion.

 

Comments

No Comments

Leave a Comment

(required )  
(optional )
(required )  
Add

About Russ Houberg

Hmmm. Bio. Ok. Well, I've been a techie geek as far back as I can remember. Which is somewhere in the neighborhood of software program called "Delta Draw" on an original IBM XT personal computer owned by a close friend of my folks. These days I'm into SharePoint mostly. I work for a great company called KnowledgeLake. We specialize in document centric transactional content management. That's fancy talk for high volume scanning using SharePoint as a storage repository and document processing workflow. So I'm particularly interested in scalability and the index/search nuances of SharePoint. I have an amazing wife, two great boys, and a labradoodle. I'm also very active in my local church and little league baseball organizations. I hope you find something useful in my blog.

Need SharePoint Training? Attend a SharePoint Bootcamp!

Posts (c) their respective authors. Everything else (c) 2007 SharePoint Experts