Hi,

Thanks for the quick response. Apologies If I wasn't clear on a few things in my initial message. I will give you a bit more information.

The application in question is written in c# and utilises QuickOPC. I have attached the OPC dll references.

The logs I have provided you with are captured from the System Management Console of our automation software platform software (Wonderware Archestra/inTouch) and for sure are not OPC client logs. The CIOPCServer.cpp error I sent in the first message is actually from our Archestra OI Gateway Server which manages the connection to the OPC Server but is itself not OPC code.

We have a applied the fix suggested by OPC and it has eliminated the handle leaks. This fix will be going through the usual testing processes and hopefully it will solve our problems in production but in the interim we are trying to get a greater understanding of what these problems are.

So my question is:

Can you provide any more information on the symptoms associated with Quick OPC handle leaks? If it is a known bug (which it surely must be), then do you have any records of symptoms and problems that other users have experienced with it? Handle leaks can be very tricky things to link specifically with unwanted application behaviour but any information the OPC team could provide on this would be much appreciated.


Thanks again,

Johnny

Hi,

We are using the following setup.

OPC Quick Client v5.35.1173.1.
Wonderware inTouch for HMI
.NET Windows Forms application which references OPC Quick client.

We have been noting regular performance issues in production. Every two weeks or so, performance degradation is noted which leads to OPC event processing hang-ups in our .NET application we end up having to restart it.

What we have also noted is that there seems to be a handle leak stemming from use of the EasyDAClient object we use. From analysis, these leaks are coming from abandoned Mutex objects somewhere within the OPC code or in a dll which you use. This seems to conform to the below article.

www.opclabs.com/forum/general-issues/2123-handle-leak-memory...ing-opclabs-easyopc-dataaccess

The performance issues and hangups are quite subtle at times and sometimes our logs don't seem to give us too much information even though the performance of the system has significantly slowed and we need to restart it. I have provided some sample logs for you to assess. I have some questions in this regard:

1) Are the OPC client errors we are seeing directly related to the handle leak?
2) if not, can you provide any insight on the exception in question?
3) Are there any established symptoms of the handle leaks associated with the above referenced article on the leaks?


Attachments:
Sample OPC Client error message
Sample handle count snapshot of application. (Without handle leak it should run at about 500-600 handles).
Sample Mutant handle creation from Process Explorer
Sample abandoned Mutex information from WinDbg



Thanks,

Johnny