Generate system core dumps for rp5430

[WL YU]

I remembered that for some old machines (e.g. HP9000 K series), it is necessary to turn a switch at the system unit and type in some key combinations at the console to generate system core dump.

Is this procedure also applicable to rp5430 running on HP-UX 11.0? If not, what is the proper procedure of generating system core dumps after system panic?

[Michael Tully]

The later rp series of systems this is not necessary. If the system has been configured correctly with crashdumps enabled and the disk space is available, the system will create the crashdump wherever it is defined. Have a look at the 'savecrash' man page and the /sbin/init.d/crashconf and /sbin/init.d/savecrash scripts for a more detail explanation.

[Kent Ostby]

The series of boxes you are referring to were the D and K class boxes.

There was a lock on getting to CM so that users didn't accidentally reboot boxes if they were working at the console.

In those series you had a service switch.

On the K, it was a physical key.

On the D, it was a switch on the back.

Best regards,

Kent M. Ostby

[Johan Lorimier]

Hi,

you can also generate a crashdump by sending a TOC via the GSP using the TC command.

Johan

[Olivier Decorse]

Hello,
on D systems with 10.20 OS, you have to modify the /etc/rc.config.d/savecore, that contains :
# SAVECORE: Set to 0 to disable saving system core files.
SAVECORE=0

# SAVECORE_DIR: Directory name for system core files. Note: the filesystem
# in which this directory is located should have as much free
# space as your system has RAM.
SAVECORE_DIR=/var/adm/crash
[...]

This is not anymore necessary on 11.00.
But if you want to automaticly generate a reset/TOC when a system hang (for example) is detected :
- just log on your GSP (CTRL-B),
- modify write access (CTRL-E, c, f),
- configure Automatique Reset (AR),
- Enable Automatic System Restart (E),
- add your triggering alert levels (13),
- quit your GSP (CO).

Olivier.