I've been using Web Connection V6.10 on a production site for over a year now. The web connection software holds up very well and once the COM instances have a problem they reset themselves; this normally happens about every 48 hours or so and has been tolerable to us. Lately, since my last upgrade, the site runs the usual course of clean operation for about 48 hours, but now instead of resetting itself it locks up until I reset IIS. There are no errors in the Request Log, but in the Module Error Log I see the following: "Unable to retrieve COM Instance". In looking through the OS (Windows Server 2012 R2) Event logs, I noticed this Warning occurred just before the Module Error: "A desktop heap allocation failed", then this Error: "The server {A4527DDC-4CD4-4C57-A47F-CC557E895BA7} did not register with DCOM within the required timeout". When I looked in the registry, I see that the UUID matches my COM server (wctrac.wcTRACServer). I can't locate any other clues to the problem. If someone could provide some insight or other things to check I would much appreciate it.

Thanks, Rob

Does anyone have a suggestion? ...Rick? Please...I need help with this.

Thanks, Rob

Unable to retrieve COM Instance usually means that all COM servers are busy and requests are basically not able to get a server instance and timing out. This can happen when servers are either too busy (unlikely) or if you have a problem with servers hanging and timing out.

The Heap Error indicates a memory issue, but based on your description I can't tell whether that's coming from the Application Pool (ie. the .NET module) or the FoxPro servers that are running out of memory. Based on the error message it sounds like it's the FoxPro servers that are running out of memory (desktop app). You should check and see if the FoxPro server memory usage keeps creeping up over time after they've been running for a while. If you have a memory leak and FoxPro servers die due to memory issues eventually they'll hang in such a way they can't be released and that's what the issue may be.

Also check memory usage on the Application Pool (w3p.exe) - if that's causing the problem there health checks you can use to recycle the app pool automatically if memory usage hits a certain threshold.

The Windows COM system inside of an application like an IIS AppPool can get corrupted if COM servers die frequently as the COM system reserves a certain amount of memory for managing COM STA instances. If COM servers crash a lot and are effectively killed rather than released this memory can be exhausted and there's nothing that can be done about it other than recycle the application pool.

It might be a good idea to set your health checks up to recycle the application pool at least once a day or maybe more often to avoid cumulative memory/COM heap exhaustion creep.

None of this is definite but some ideas.

Ok, thanks for the suggestions Rick. I'll post here whatever I find.

-Rob