Users are facing lot of TC crashes.
Issue is inconsistent and unable to reproduce the issue at will.
Is there a way to debug the Teamcenter crash in the production environment.
Is there any best practices to root cause Teamcenter crash
The first step would be to look for errors in Teamcenter session syslogs, journal logs, user RAC logs, custom dll logs, log4j logs (if you are using log4j)
The next step would be to enable SQL debugging but won't recommend at the onset as you mentioned that crash is observed in production environment.
Can you share the Teamcenter syslogs to better understand the reason for the crash
Thanks for your response.
How to get the separate log for custom dlls, please let me know.
Also please provide details of log4j log, how to enable and what is the use of the this log, please let me know.
To share the log, I need to get a consent from the client, else will share the error from the syslog
You will have to analyse log files as HarisVenkat suggested.
When Teamcenter crash, do you get error like - tcserver failed to respond? If that's the case then error is on the server side. You will need to look in syslog/jnl file on the server. This is most likely scenario.
If crash is occuring from Teamcenter client, then you will have to look in recent RAC log file to see what JAVA exception is thrown. This is unlike to be the case.
When Teamcenter crash, you are also likely to see crash dump file (.dmp) file. If your custom DLL is causing the crash, you should be able to debug the dmp file in Visual Studio to see what code/data condition is causing the crash. If crash is occuring due to error/bug in one of the core Teamcenter library, then the dmp file may not be of much help. You might get some information out of it, but not enough to pinpoint the cause of the crash.
If you have access to past PLMWorld presentations, then look for my presentation from 2013. It is titled - "Invalid Memory Access, Segmentation Violation, and Signal 11! Oh My". It outlines in detail how to debug a Teamcenter session in Visual Studio. This is helpful if bug in your custom code.
If crash is due to bug in core DLL, then you may have to enable detail journaling and then analyse jnl/syslog file.
If all else fails, contact GTAC :-).
Thanks for the valuable inputs, and informative PLM World presentation.
I will try to implement in my TC crash analysis.
Yes, I have raised an IR with GTAc, but not recieved a good reply, suggesting to enable the debugging variables.