topic Re: C program error ... Abort(coredump) in Operating System - Linux

C program error ... Abort(coredump)

Ahmed_58 — Thu, 17 May 2007 00:29:54 GMT

Dears,
Below messages appear to our apll team, which indicate swap need to be increased,...
---------------------------------------------
Pid 21573 received a SIGSEGV for stack growth failure.Possible causes: insufficient memory or swap space,or stack size exceeded maxssiz.
----------------------------------------------

And after increasing swap file, now below message display, while running C program.

----------------------------------------------
rate_vpn@bat-bcdr:/appl/rate_vpn/bin/rating$ _vpn/bin/cdr/mp/cdr_intermediate --postpaid --tapout --verbose Abort(coredump)
----------------------------------------------

Previously, this program was running withoug any error.

Regards,
Ahmed

Re: C program error ... Abort(coredump)

Arunvijai_4 — Thu, 17 May 2007 00:41:25 GMT

Hi Ahmed,

Check you have enough memory available in the system with # swapinfo. After that, analyze your program, core with gdb. It could be due to bad programming.

-Arun

Re: C program error ... Abort(coredump)

Ahmed_58 — Thu, 17 May 2007 01:47:08 GMT

Dear Arun,
Memory seems OK;

# swapinfo -mat
Mb Mb Mb PCT START/ Mb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 4096 0 4096 0% 0 - 1 /dev/vg00/lvol2
dev 3424 0 3424 0% 0 - 1 /dev/vg00/lvol10
reserve - 4651 -4651
memory 6513 111 6402 2%
total 14033 4762 9271 34% - 0 -
Regards,
Ahmed

Re: C program error ... Abort(coredump)

Dennis Handly — Thu, 17 May 2007 02:29:35 GMT

As Arun says, this is typically due to a programming error. You probably have a recursive stack overflow.

I assume you checked maxssiz and aren't using a Fortran like application that requires a gigantic stack?

If you are now getting SIGABRT and there were no error messages explaining why, you need to get a stack trace then debug the issue. Use gdb's "bt" command:
$ gdb executable core
(gdb) bt
(gdb) q

Re: C program error ... Abort(coredump)

Ahmed_58 — Thu, 17 May 2007 03:01:59 GMT

Re: C program error ... Abort(coredump)

Dennis Handly — Thu, 17 May 2007 04:28:04 GMT

>How can I find if maxsiz is the problem? currently it shows: 8388608

8 Mb should be fine. If it was a problem you would still get that SIGSEGV error and not SIGABRT. Unless there is a super special signal handler that prints a message then aborts??

>After I add SWAP space the programmer complining "software that should be working fine is also giving problem"

That programmer will have to debug his application. By using gdb and getting a stack trace may make it obvious.

>And they send me below errors;

Ah, that looked like a jumble of just junk. Until I realized this was wrapped lines that seemed to have some error messages in it.
It would have been better to attach a file with this info, so it doesn't wrap.

Anyway, it is still meaningless to me. Just some error in:
ORA-00904: "START_DATE": invalid identifier

Re: C program error ... Abort(coredump)

Ahmed_58 — Thu, 17 May 2007 05:35:51 GMT

Dear Dennis,
Thanks for replay... now the programmrs looking into it.
Ahmed