locked
Best Practices Lib and possible caching RRS feed

  • Question

  • User37428675 posted

    We have two ways our web services communicates to SQL.  ASP.NET and Microsoft.Practices.EnterpriseLibrary.Data.    The best practices call is only used in one specific area where we read the webservice calls headers and authenticate it against SQL.    Once in a blue moon, we start getting access denied on all clients, but one.   That one is never the same, but who ever the lucky client is continues to process.    We end up restarted the web service and everything starts working again.    So now, I've been tasks to figure out what the problem is.

    Just so you understand my comprehension in this.   I've been a professional software engineer/developer since 1995 years, but currently manage a development team.   I write software across 15 different languages fluently, so I fully understand the coding that was done in this project.  I don't always understand the backside of what Microsoft is doing, which is why I'm here.

    My guess on this problem is there some kind of caching going on and at first I was thinking, connection pooling.   However, once I started looking into connection pooling and the best practices lib, I can't find anything.    The reason I think it's a caching issue is because there are no errors anywhere.   Eventlog has nothing, IIS logs have nothing, there isn't a connection failure error.   Everything works as if it was able to query SQL and come back with a value, but the value is incorrect to the query that was performed.   I then started looking for possible caching issues within the best practices lib and I found best processes has it's own caching methods, so I'm flooded with something I'm not really wanting to see.

    Has anyone ever had this issue?  Can someone with some in dept knowledge on best practices lib, point me to some information on connection pooling within this lib or if it's using connection pooling at all?   I'm assuming yes, but finding something that says that hasn't been too easy to find.

    Since this has been occurring every 8 or 9 months, it's very difficult to track and being in prod, I'm unable to test during it's outage.    Any direction would be greatly appreciated.

    Monday, November 5, 2012 4:43 PM

All replies

  • User37428675 posted

    I've ran across another possible option, Connection Pool Fragmentation<!-- google_ad_section_end -->.    Our web service does use many different connection string because we have large DB's that are being flipped once the DB gets to 1TB.

    Tuesday, November 6, 2012 9:15 AM
  • User-37275327 posted

    I am not top of this. Just have a look at following

    http://www.codeproject.com/Articles/115107/Ten-Caching-Mistakes-that-Break-your-App

    Monday, January 14, 2013 9:16 PM
  • User37428675 posted

    I'm not caching, the question does the best practices library from Microsoft do some kind of caching through connection pooling?

    Saturday, January 19, 2013 5:58 PM
  • User1793623445 posted

    We had some connection pooling issue and connections were getting accumalated during the peak time. Finally we figured it out by changing the IIS setting request limit property set to 0 instead of 1000 somehow. Then all connections opened  went down to 1 all the time..

    Monday, February 18, 2013 10:15 AM