views:

109

answers:

1

Hello everyone,

I'm facing an extremely weird bug here and I'm not really sure If I'm following the right path to solving it or even how to solve it.

Here is the problem I'm facing: I start debugging a WPF application which uses log4net, NHibernate and LINQ to NHibernate, and when I try to get an Entity from the database my application and sometimes VS hang for a lot of time, and after a while an exception dialog opens showing a message containing the following information on a ContextSwitchDeadlock MDA:

The CLR has been unable to transition from COM context 0x34fc1a0 to COM context 0x34fc258 for 60 seconds. The thread that owns the destination context/apartment is most likely either doing a non pumping wait or processing a very long running operation without pumping Windows messages. This situation generally has a negative performance impact and may even lead to the application becoming non responsive or memory usage accumulating continually over time. To avoid this

I copied the code files to a new project and deleted the old project to see If I could make this message disappear, thinking it had something to do with my configuration. I started adding few things at a time to see what was causing it, and when I included log4net configuration code the bug appeared again. First I included it through AssemblyInfo and later trough code configuration on application startup, and absolutely nothing changed at all :(

So, here are my findings:

  • It only happens when I'm using log4net.
  • It happens when NHibernate loads an Entity from database (lazy loading).

I don't know what might be the source of this bug. It only happens when debugging in Visual Studio. I've tried following the steps on the "Enabling and Disabling MDAs" section of the following page: http://msdn.microsoft.com/en-us/library/d21c150d.aspx, but that doesn't work either, VS still hangs and it's memory usage increases.

When I run the program normally none of this happens, so I'm pretty sure this is not a deadlock situation, as this question suggests: http://stackoverflow.com/questions/334360/c-contextswitchdeadlock (I've also tried the solutions posted there).

Because of that, I've decided to disable log4net and enable it again when deploying my app.

I'm posting this question to find out if somebody else has faced this bug or if somebody has some suggestions on how to solve it. Finally, it might help somebody else facing this very same problem.

Thanks in advance,

Jorge Vargas.

A: 

When using the DebugAppender all the entities in the databases are loaded and all its data written to the debug output. That was causing the ContextSwitchDeadlock MDA since it tookmore than 60 seconds to run.

Disabling the DebugAppender solved my problem.

Thanks to Mauricio Scheffer for the tip.

Jorge Vargas