topic Re: CORE in Operating System - HP-UX

CORE

Reuben M. Ordonio — Mon, 05 May 2003 04:11:53 GMT

What does it mean when your server is creating a file called "core"? Is this an indication of a hardware related problem? I'm also experiencing an error that says "There is not enough memory" from time to time. Is this related to this "core" thing?

Re: CORE

Eugeny Brychkov — Mon, 05 May 2003 04:21:52 GMT

Usually coredumping is caused by incorrect software behavior - in your case maybe software is trying to allocate some memory, but do not track that allocation function returns 'no memory available' and tries to write to 'null' area - causing coredumping ('memory page protection fault' in windows...).
If there's a hardware issue, server crashdumps and reboots
Eugeny

Re: CORE

Michael Tully — Mon, 05 May 2003 04:26:56 GMT

Basically a core dump from an application is what was in memory as a page fault for a particular function or program.
You can find where it comes from by:

$ what core

Hardware problems do not dump core, they may issue a panic and reboot your system.

Re: CORE

Suhas_2 — Mon, 05 May 2003 04:30:17 GMT

Hi,
This problem is related to resources on the server. When a particular application is trying to allocate some memory using systems calls like "malloc" or "calloc" it is facing a problem. When this system call fails the application will dump core.
Secondly, such problem may occur because of wrong way of allocating memory in the programs. Some programs go on allocating memory but never release it, this is called as "Memory leakage", a typical solution would be to reboot the server to release the memory and tuning the application code to avoid any future incidents.

Rgds..
Suhas

Re: CORE

Steve Steel — Mon, 05 May 2003 09:06:30 GMT

Hi

file core

Will tell you what happened

You may need to debug it or tune your system

Please post the output.

Look in /var/adm/syslog/syslog.log at the same time as the core.

Steve Steel

Re: CORE

Mark Greene_1 — Mon, 05 May 2003 11:33:20 GMT

You can use the file command to determine what process aborted (and thus created the core file).

You can also use the strings command to see what's in the core file:

strings core |pg

Common causes include hardware problems, telnet or socket sessions dropping without properly shutting down the associated processes on the HP side, and there's a couple of SAM bugs that will dump core if you aren't up-to-date with your patches.

HTH
mark

Re: CORE

Michael Murphy_2 — Wed, 07 May 2003 13:01:10 GMT

I agree with running a the "file" command on the core to see which binary is creating it. Also, examine the size of the core. Compare it to the value of maxdsize. If they are around the same size, the exec is reaching the memory limit in your kernel. Bump up the kernel param and re-run - but be aware that if the process has a memory leak it will get the error again - watch the memory usage with ps.

Re: CORE

Donny Jekels — Wed, 07 May 2003 15:11:15 GMT

you can also analyse the core dump

#strings core | more

this would give you a closer idea og what cuased the core dump. don't try to go way deep into the file. usually the answer is within the 1st couple of lines.

peace
Donny

Re: CORE

Dario_1 — Wed, 07 May 2003 16:11:24 GMT

Hi!

Check this post. Paula gives a very nice explanation about how to check a core file.

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x4f1f9607df6ed711abdc0090277a778c,00.html

Regards,

Dario