[2388] in Consulting_FYI
crash dump instructions were wrong
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Thu Oct 5 13:04:29 1989
Date: Thu, 5 Oct 89 13:03:55 -0400
From: Jonathan I. Kamens <jik@ATHENA.MIT.EDU>
To: cfyi@ATHENA.MIT.EDU
The instructions I sent in earlier for getting a crash dump weren't
quite right. Below is a corrected version.
jik
*************************
1) Bring up the machine in single user mode. To do this, halt it with the
halt button on back of the machine (if it's a VS2000) or on the front of
the machine (if it's a VSII), and type "b/2" at the ">>>" prompt. If you
don't get a prompt when you halt the machine (this applies only to VSII's),
that means that the halt switch on the back is set wrong.... it's not worth
dealing with it.
2) When you get a shell # prompt, type "cat /etc/fstab" to find out what the
device name for root is. It should be something like "/dev/rd0a". Type
"fsck -p /dev/rd0a" (or whatever).
3) If it reports any errors (and says "FILE SYSTEM WAS MODIFIED" or
something similar), then type "/etc/reboot -n" *immediately*. Forget
about the crash dump, it's not worth it.
4) Type "ed /etc/rc.conf". Type "/SAVECORE/s/false/true/p" and hit
return, and should print out one line with SAVECORE=true. Type "1"
and hit return. Type "/PUBLIC/s/true/false/p" and hit return, and
it should print out one line with PUBLIC=false. Type "wq" to save
and exit.
5) Type ctrl-d to exit the shell. The machine will continue to reboot and
save the core image at the proper time.
6) When it's rebooted, login as root and edit /etc/rc.conf again to turn off
SAFECORE. Type "/SAVECORE/s/true/false/p" inside ed, then "1", then
"/PUBLIC/s/false/true/p", and do wq again.
7) Send mail to hotline and tell them the workstation number that rebooted,
and mention that you saved a core image using savecore.
