UNIX for Advanced & Expert Users

Hi,

We have one application (c language) installed on our AIX 5.2. When we enter some large amount of data in one screen and do update, it is dumping core some times without any error (like bus error, segmentation fault). Next we open the application again and enter the same data and do update. This time updates go through fine without any problem.

This problem (core dump) is happening only with that particular screen and occasionally, not always for the same data update.

We tried the same on another machine. It is not dumping core on that. Updates are going through fine always.

What could be the cause for this problem? Is this problem due to some memory settings? How to troubleshoot this problem?

Please help ...

Thanks,
Venkat.

I would open a ticket with IBM. Sounds a little more complex than that average questions here.

Check out:
1. are the runtime (shared libraries) exactly the same release levels for all the software?
2. what patches have been applied to one machine but not to another?

If 1 checks out and there are no differences in patches, then I'd call IBM as well.
It is probably a hardware problem.

Adding to the thread,

I had a similiar kind of problem when I was working on a [robust] e-sql c program.
The Program was working fine on our development server without any issues and any hiccups.
When the same was tried in one of our test boxes, it crashed at the point of fetch from a non-dynamic cursor.
To add to the fascination, I tried the same program on 2 more test servers and it worked perfect, and that was the only server where it crashed.
That was 6months old problem nobody was able to found out still.
Unfortunately that request was dropped.

Only difference was in the Informix version we had, upgraded version in our test boxes and that too upgraded in all the test boxes.

I am just bringing it to notice that such rare problems do occur where the execution is guaranteed on one server and crash is guaranteed on other server.
Which entity to be blamed?
whether it is the program ?
or the hardware in which it is run?
else the last one software which supports it ?

This sounds still as a mystery to me.

Experts please throw on more light in to this issue.

Thank you friends,

Mr Jim, please see the answers for your questions.

1. are the runtime (shared libraries) exactly the same release levels for all the software?
2. what patches have been applied to one machine but not to another?

Sorry I did not tell that this problem has been happening on that particular machine for the last 3 years. Within the last 3 years many upgrades happened (machine upgrades, software upgrades ..). But the problem is there from the begining. So I don't think this problem would be due to these upgrades.

Moreover, the problem occured when the shared libraries and pathes were same on both the machines also.

I have the core files generated. But I could not find any useful information from those files. These core files are from a stripped a program or executable.

Can we get any useful information from the core files?


Thank you very much for your help till now.
Please continue helping me to find the solution or cause for the problem ...