core dump

[Shivkumar]

Dear Sirs,

How to force a core dump on hpux 11i ? What are the percautions ?

Thanks,
Shiv

[Rick Garland · Accepted solution]

Depending on hardware and/or OS version, you will interact with the GSP or the EFI.

Go through the menus. There is a TC option, also known as a TOC (Transfer of Control). Invoke this option and follow the prompts. It will verify if you truely want to do this as it will reboot the system.

Do not do the RS option (reset). This will reboot system but will not save crash/core.

Make sure you have room in your crash directory. Make sure you are SAVECRASH=1 (in /etc/rc.config.d/savecrash)

NOTE: This will reboot the system. But if you want to force a system crash...

[A. Clay Stephenson]

Whether or not a core file will be generated depends upon a number of factors: 1) permission to write in the CWD, 2) UID's match; 3) core file does not already exist or if it exists is writable

However, this still does not guarantee the creation of a core file because it also depends upon what signal handlers are actually in play in the process in question.

In general, you can send a SIGABRT to a process to trigger a signal handler which calls the abort() functions which in turn triggers the creation of a core file.

e.g. kill -6 PID will generally result in a core file.

Man 5 signal, 3 abort for details.

[A. Clay Stephenson]

When you say "core dump" I assume you mean a process as opposed to a "crash" which means that the kernel itself crashes and then attempts to save the core image.

By definition, however a given process responds to a kill depends upon the process itself except for all signals which cannot be ignored or caught (e.g. kill -9).

[Mahesh Kumar Malik]

Hi Shiv

Following pages of hp-ux document may help you

http://docs.hp.com/en/J2237-90005/ch06s05.html

Regards
Mahesh

[generic_1]

TOC forces it from GSP.
I would use that as a last resort IE your system locked up, but youd like to know why.

Then do a Q4 Analysis, and send the data into HP with your case. They are used to walking users through a Q4 if you need some help they can email you some very helpful docs.
Look at at your /var/adm/syslog/syslog.logc, GSP logs
and tombstone logs too. Hopefully between these you can find out what went wrong and if you need to replace any hardware.

Good Luck.