Thanks for everyone jumping in and helping out so fast ! - More more
than I'd ever expect !! This - as they say - rocks !
Okay, (as usual) a part of the blame was on me; stupid, panic,
you-name-it. And one part on the system:
The machine is run fully through serial console, including boot. It also
runs on a UPS, APC RS 1000. The latter needs a few changes to the kernel
to enable the USB-UPS: uhidev, ums, wsmouse, ukbd, wskbd, uhid must be
disabled.
Now, since it didn't reboot properly, I went to the server room and
attached a PS2-keyboard to the running machine. This seems to make the
kernel panic and behave strangely, not even accepting the usual debug
commands.
In the end I dunno what happened; I was too nervous to pin down all
steps. I put the original 3.8-kernel back, removed all serial stuff,
UPS, and it would start; spew a lot of hard disc errors, but go through
after quite some time. Then I put back everything one by one, rebooted
always in between, and now it is up and running again.
Most stupid: I didn't want to believe in the keyboard-induced panic and
plugged it into the same type of machine next to it, same configuration
including UPS, also in production, and that one panicked just the same !
I'll buy a box of chewing gums tonite and plug close the PS2-keyboard
connectors tomorrow morning !
Thanks again for your phantastic reactions,
Uwe
