crash dumps

[navin]

Hello ,
i had a a problem shuting down my server and i had to recycle the power .i had to manyally pull and replug the power cable. But now when system boots up ..it continues says system crash occured and dumping is done ,and then system reboots again with the same messages.
How do i clear these message and boot my system back.
Thanks for the help.Most appreciated.

[Steven E. Protter · Accepted solution]

Shalom,

Your system has been damaged by what happened to it.

You need to boot the system off a Core OS CD/DVD and see if you can use the automated recovery tools.

If not, have hardware service come in and look at your machine,its broken. The problem that forced you to boot may have been the entire cause.

SEP

[A. Clay Stephenson]

You may have (and probably have) more serious problems than you know but to answer your question:

1) Interupt the boot process and bring the box up in single user mode
BOOT PRI
Interact with ISL/IPL? y
hpux -is

2) You will be logged in as root.
mount /usr
mount /var
(You may have to first fsck these filesystems before they will mount)
3) vi /etc/rc.config.d/savecrash and set SAVECRASH=0
save the file.
4)
umount /var
umount /usr
sync; sync; sync
reboot

This will stop the saving/compressing of the crash dumps.

[navin]

Thanks. Is there any thing new in power cycling PA-RISC servers. I did press off the front power button .I saw it went down ,but not the LAN LED. But i did not see the power went down for a server in the console. SO i had to manually un plug the power cable. What was wrong.How the power cycling works.please help.

[A. Clay Stephenson]

Nothing has really changed in the power-cyling, it's just as safe (and as dangerous) as it has always been. I suspect (although I have no way of knowing) that the damage was already done (e.g. a critical file was altered or removed) so that your system was an accident waiting to happen and would have failed no matter what on your next boot. Of course, yanking the power out from under a UNIX is never a good idea but almost every time, the box will come right back up --- note that little word, almost.

I would bring the box up in single user mode and then step it through each init level. e.g. init 2, init 3. That won't even be necessary if you can find where the system failed in /etc/rc.log.

[A. Clay Stephenson]

By the way, this is the reason that your have Ignite images and/or lifeboats --- and I'm sure that you have recent versions; in fact, I'm sure that you do a make_tape_recovery each and every week. If not then you deserve whatever happens to you. However, even a backup may not fix you because you may have a basic hardware problem; there are just not enough data (at least you haven't provided enough data) to tell.

[navin]

Thank you all for your reply. It is a test box so i do not take os backups (tape recovery). But i was able to boot from cd and recover the box using the recovery tool.Thanks again