I'm working on a CentOS system, as requested by one of my customers.
Recently a bug has been reported, mentioning that in a particular case, a core-file was generated.
After a huge amount of work, I realised that the problem was not due to my program behaving badly, but to internal configuration: my program (being a Java program) was started up without -Xss option. The machine automatically set a maximum stack size which was too small and as a result the program crashed.
Now I've solved that bug simply by adding -Xss2m to the Java startup parameters of my program.
Now I'm wondering: imagine that a program goes bad on such a CentOS system. Is it possible to configure the system in such a way that I can easily know whether the problem is caused by a malfunctioning program or by a configuration/startup parameters issue?
Thanks in advance
You are assuming a distinction where non-exists.
When you supply a program, you supply it with all relevant arguments and environment. If your program crashes, then you have a bug. It might be a packaging bug (though it is strange to me that it resulted in an actual core generating crash, rather than a more diagnosable error message), but a bug it still is, and in your supplied delivery.
So, there is no such system, but none would make much sense.