Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

amd64 server upgrade failed - help

I have an HP proliant server (amd64) running current from about March
2013. It has all its filesystems on raid 1 sets and as well as being a
fileserver also acts as a DNS server/mailserver for a few domains. I 
decided to do an update to fix the existing bind security problem. 
CVS'ed up to current as of 31st July. Built new tools and a new GENERIC
kernel - all fine. Installed kernel and modules - kept old ones just in
case. Rebooted and everything was fine. Did a build release and when
that completed successfully did a build install=/. etcupdate. Rebooted.
Oh dear!

Kernel boots up and then panics when init died:

Panic: init died (signal 0, exit 11)
fatal breakpoint trap in supervisor mode
trap type 1 code 0 rip ffffffff8026a18d cs 8 flags 246 cr 2 7f7ff63b20
ilevel 0 rsp fffffe8002f7aaa0

OK, no problem, boot a recent amd64 image from releng from a usb stick.
Aggh - but I have a raided root and the USB kernel plows through and
uses root on the raid rather than root on the usb disk. Use the usb
keyboard to force a 'boot netbsd -a' from the usb stick. That works and
it prompts for a root device. The bloody usb keyboard doesn't work at
this point so I am a bit stuffed.

I had a slight premonition of this as I had an issue before and Brian
Burrow did suggest boot -a may work but the lack of a real keyboard is a
show stopper. Given that usb keyboards are the norm now I would have
thought the kernel (and debugger) should support them better.

My best guess is I need to build an install image with root hardwired to
sd0a so that it will pick up root off the usb stick. This will be a
tedious process as my only other amd64 machine is slow.

Any other suggestions other than a total reinstall ? (I do have a full
backup from 3 days ago but really don't want to have to waste a couple
of days...)


Phone: 07805784357
Open Source O/S:

Home | Main Index | Thread Index | Old Index