Thankyou Dr. Allan Rollow, Paul Thompson, Thomas Leitner and Craig
Farrington for their responses.
Normally if you lost your utmp file the resolution would be
touch /var/adm/utmp
chown root:system /var/adm/utmp
chmod 755 /var/adm/utmp then a reboot.
The reason our system wouldn't come up properly, however, was because a
var backup had been restored onto our system while /var was not mounted.
When the system was coming up, it relied on information from these old
files rather than from the proper /var filesystem causing the system to
not know what run-level it was in. It seems the boot process will use
information from a non file-system /var directory prior to mounting the
/var filesystem if it exists.
>
> Dear Group
>
> I am in a bit of bother. We had a root disk crash and although I can now
> bring the system up, when it executes the /sbin/rc2 and /sbin/rc3 scripts,
> the set `who -r` is stopping the execution of the start-up scripts due to
> the lack of a /var/adm/utmp file (I presume).
>
> Any ideas - I need to get this system functioning before tomorrow morning.
> (We're running an 8400 in a 2 node cluster using 4.0D).
>
> Thanking you
>
> Duncan Webbe
>
> Unix Systems Administrator
> David Jones Ltd. (Australia)
>
> email: dwebbe_at_davidjones.com.au
>
>
>
Received on Tue May 12 1998 - 03:58:00 NZST