Thanks for the responses to my previous summary. Here was what
Dr. Tom Blinn, 603-881-0646" <tpb_at_zk3.dec.com> wrote which solved my
problem:
If the /var/adm directory does not exist, then writes to utmp are queued
until the directory can be accessed.
I suspect that in your root file system you have a /var/adm directory,
and
that you're mounting the new /var filesystem on top of the /var
directory,
which is hiding the existing utmp file.
Get your system up to single user mode (boot -fl s or shutdown), and
look in
/var on the root file system (with nothing mounted on it). It should be
an
empty directory. If it's not, mount -u / to get the root filesystem
mounted
read-write and remove any contents of /var (or move /var aside to
/var.old
and create a new, empty /var as the mount point).
I have a separate /var filesystem on my V3.2C system, and I have never
had a
problem with rc scripts failing. I also have workstations that run with
a
dataless environment that mount all their working filesystems via NFS
that I
have never seen such a problem with.
--
John Schaeffer | Connecticut College
Email: jsch_at_conncoll.edu | 270 Mohegan Ave.
Work: (860)439-2340 | New London, CT 06320-4196
Received on Sat Jun 22 1996 - 13:20:25 NZST