I have another question that may or may not be related. What prompted the customer to try to reboot were several instances of this error: OLE error code 0x80131014: Unknown COM status code. Similar to what I've reported for other clients, these are showing up in the sendmail routine under wwdotnetbridge - trying to create an instance of "Westwind.wwSmtp". (We worked around the problem by switching them to classic mail mode...)
Is the fact that they are having issues with wwdotnetbridge something we should worry about when considering moving their whole site to the .NET Module?
Hi Stein,
Yes unfortunately this happens on some machines where the Windows startup causes delays in loading of some of the Windows features that COM needs to run properly when thread marshalling. It's rare (i've never been able to duplicate it, but there are a few customers who have run into this).
A couple of things to try. Can you run the Web Connection .NET Module? It doesn't use thread marshalling but has dedicated threads so the COM calls are much simpler and don't require all of the DCOM thread management infrastructure so that might actually work better.
If you're running on IIS 7 or later I'd recommend doing this no matter what assuming you don't have apps that span multiple virtual directories for a single backend server.
+++ Rick ---
Just ran into a new issue. The customer moved our WC app to a new 2008 Server. It's configured in wc.ini to load 2 instances of our COM server. Things seemed to be OK until they rebooted the server and started getting 505 errors. Checking the Task Manager showed numerous instances of the COM process, with more being added each time the site was hit.
We eventually got it stop by loading the COM server into the DCOM editor - the settings looked OK but we reapplied them - and then resetting IIS. Not sure the DCOM step was necessary, but after the IIS reset, it stopped adding process, began responding to web hits, and in fact started unloading them until we got back down to the 2 that were specified. But then we tried rebooting and it started all over again.
The logs show numerous instances of Events 5009 (process terminated unexpectedly) and 5011 (fatal communications error with the Windows Process Activation Service) - all associated with the WW WebConnection application pool.
Any idea what could be causing this on every reboot?
--stein