0 Replies Latest reply on Oct 8, 2014 5:52 AM by pvandermn

    CF8 stops returning pages to browser and service must be restarted

    pvandermn

      We have been experiencing problems with Coldfusion and we cannot figure out what is happening.  We don’t know if it is a server issue, or a code issue.  It occurs primarily on one of our two production servers in one instance (but has rarely occurred on other servers).  What happens is that the transactions slow dramatically or stop being returned at all.  We use the Seefusion monitor and can see CF accept and process the transactions, which start to get older and older and never finish.  It will occur in one instance, and we have to restart that instance to fix it.  Then often, about an hour or two later, the same will occur in the other instance on the server.  When we look at the database server, there are no transactions running in SQL for any of the aging jobs.  Usually, when traffic slows or backs up due to volume, we can see a SQL job running for all of the jobs displayed in SeeFusion.


      A review of the application and exception logs does not reveal any specific error that would have a common denominator.  What we see is actually a suspension of log entries entirely until the service is restarted.  Something is happening to CF but we cannot determine what it is or how to rectify it.  This is CF 8.01 with all patches up to date.  We are in the process of building a new web server to deploy CF 11 as the only alternative we can think of with the hopes that it will solve the problem.


      It could possibly be code related, as we did release a new version of our application at the approximate time this started, though I don’t know for sure if there was a direct cause and effect.  But if it is code related, how can we possibly determine the cause if there is nothing in the logs to indicate a problem?  The Windows logs do not show any problem either. 

       

      Has anyone ever encountered something like this, or do you have any thoughts about how we might fix it?  We are current in our patches.