[5047] in testers
"krtld: load_exec: fail to expand cpu/$CPU" on reboot
daemon@ATHENA.MIT.EDU (Craig A. Counterman)
Sun Jul 22 18:54:40 2001
Message-Id: <200107222254.SAA00699@birdland.mit.edu>
To: testers@MIT.EDU
cc: ccount@MIT.EDU
Reply-To: ccount@MIT.EDU
Date: Sun, 22 Jul 2001 18:54:37 -0400
From: "Craig A. Counterman" <ccount@MIT.EDU>
OK, here's what happened:
Machine 'birdland' to be given to a developer.
I had Abby add it to the list so it'd take 9.0 early so I could set it
up for the developer.
It updated to 9.0.10.
I did I thought the same the next day to another machine, which took
9.0.11.
So, I wanted to give this machine 9.0.11 as well, so did
detach -a
attach -O dev-sun4sys-90
update_ws
and it said that it was already at that release. Ok, I thought, maybe
it's in the wrong cluster, I'll stick with 9.0.10 and let nature take
its course later this week.
I rebooted.
It failed very early in the boot, with the message
krtld: load_exec: fail to expand cpu/$CPU
"Oy", I thought, "I've screwed it up. I'll reinstall."
So i did boot net:18.179.0.20,cnboot,18.152.0.122 and reinstalled, it
got 8.4. Then I did the attach -O bit, and it took 9.0.10 again.
I started to configure for oracle installation: adding an oracle user
to passwd.local, shadow.local, group and changing kernel params in
/etc/system.
When I rebooted after that, the
krtld: load_exec: fail to expand cpu/$CPU
error was back.
On the advise zacheiss, I tried "boot kernel/unix" It booted OK, but
again failed on the next reboot without that argument.
Now I'm booted with kernel/unix again, and the changes I made in /etc
are all gone (!).
I'm going to give up on installing oracle now and leave it as is till
I hear from you what to do.
Thanks,
Craig
