Share-n-dipity

SharePoint serendipity is the effect by which one accidentally discovers something fortunate, especially while looking for something else entirely. In this case, it is the occassional musings, observations, and Ouija board readings about the phabulously

A Memory Issue To Watch in SharePoint 2013 Beta 2

A Memory Issue To Watch in SharePoint 2013 Beta 2

  • Comments 1
  • Likes

As folks start working with the SharePoint 2013 beta 2 bits, you may randomly notice certain errors that once they start, don't go away.  You may get dialog boxes (coming from javascript) that mention a 500 service activation error, or you may see search queries just stop working, or you may look in the ULS log and see issues with the STS activation.  These are just a few examples, you may see others.  One of main things you need to keep on top of in SharePoint 2013 is your memory usage.  We use more RAM than we did in previous versions of SharePoint, and you will find that anything less than 12GB of RAM can be difficult to use.  One of the outcomes from this additional memory usage is that when you have less than 5% of your RAM available, pretty much all WCF calls start failing.  When that happens, you will see the seemingly random type errors that I'm describing above.  That's because so much of the interaction with service applications in SharePoint happens through the WCF endpoints.

Fixing it is resolved by just closing applications, killing processes, limiting the memory allocated to processes (like SQL Server) or of course you can reboot as a last resort if you really need to.  But I'm generally able to get things working again by just cleaning up some of the applications I have runing.  As a related side note to this that I may try and cover in more detail later, during the beta, you may want to just kill the noderunner.exe process that is consuming the most memory.  I would never do this in production, but then again, you are hopefully not running your entire farm on one 8GB server in production either.  The noderunner.exe processes (you will see five if running your whole farm on one server) are used for the different search components - crawl, index, analytics, query, and content processing.  The search admin service takes care of restarting a process if it goes down, so it will end up starting a new one to replace it, but it generally takes some time for it to chew up enough memory that it's a problem.

Comments
Your comment has been posted.   Close
Thank you, your comment requires moderation so it may take a while to appear.   Close
Leave a Comment