Errors and suggestions

[urayoan]

Hi:
  Since a long time, i am having an error with NetTalk and honestly, i don't know what else to look at. Usually, it pops in the web server one after another, and sometimes it happens every few days, wich makes it more difficult to catch and reproduce.

  The server when the error happens is under heavy load from clients, but in previous versions of NetTalk this does not happen so often. I think i am missing something, but honestly, don't know where else to look at, so as the title said, any suggestions are welcome. Sorry for the low res screen capture, but this is what i got from the client.

[Bruce]

Hi Ura,

is the GPF always int eh same place?

I guess if it is the next is to inspect the code in that routine, near that line number, to see what code is there.

cheers
Bruce

[urayoan]

No, is not always in the same place. Check the other screen i got. related to the same web server.

Sometimes, the server just hangs with a Not Responding message in the window and the task manager, and sometimes i got the GPF.

That procedure (the one called BrowseResByCal, calls the routine to make reports. Those reports are in a separate DLL that i have no control over it (Clarion DLL) but, the call is not called in the lines the GPF message shows me (if i am looking in the correct place and interpreting the message correctly).

To be clear (and gain knowlege of course), in the GPF, what line should i need to look? When i look at the exact lines the message throws, are mostly auto generated code.

Thank you Bruce

[Bruce]

Hi Ura,

The GPF is a Stack-Trace, meaning that what it shows you is everytime the code made a call which added to the stack.
So you read it by following the calls, (bottom to top), through procedures, methods, and routines (R$). Ideally it gets you close to the problem, by allowing you to see if there is any embed code on that route, or if it's always int he same procedure, or same method, or same routine and so on.

cheers
Bruce

[urayoan]

Thank you Bruce. I was looking at the GPF report from top to bottom and as always, order matters. Interestingly, most of the parts that y receive in the report are generated code and some embeds with p_web.Trace that i removed just in case.

There is a part that I create a queue from a view, and maybe this is the part that is giving me problems.

Question related to that, the Queue i am creating, is declared in the procedure i call to generate the reports. In that case, the Queue is an instance of the same procedure and every user that call the same procedure, should have their own Queue declaration giving that case?

Thanks again in advance!

Hi Ura,

The GPF is a Stack-Trace, meaning that what it shows you is everytime the code made a call which added to the stack.
So you read it by following the calls, (bottom to top), through procedures, methods, and routines (R$). Ideally it gets you close to the problem, by allowing you to see if there is any embed code on that route, or if it's always int he same procedure, or same method, or same routine and so on.

cheers
Bruce

[Bruce]

yes, queues declared locally belong to that instance of the procedure.