Hi,
I loaded the Pre-Release version of Ubuntu 10.10.1 32 bit workstation. It ran mostly fine for about a 2 months. Then I did an online update and it told me that I had to reboot. When I did, I booted to a splash screen and then a dead machine with a black screen where control-alt-delete does nothing, no disk activity, no mouse, no keyboard, totally dead. Had to hit reset button.

I booted to a suse partition and hacked the grub entry to remove the splash and silent to get more spew at boot time. The last thing before the screen goes black is starting NBD client. I have my email and various directories on my central server so I need NFS.

Is there a way I can hack my way to get it to boot even without NFS temporarily to possibly rebuild the nbd kernel module?

I have looked at the /var/log/ syslog, daemon, debug, kern messages and gdm log files and don't see any indication of why it crashed. Where else can I look?

Puzzled,
BrianP

You could try looking at the files in /etc/init.d for whence the NBD client is started and making it unexecutable.

Sonneteer,
I removed the execute permissions. It just crashed a bit farther on.

Thanks for the good idea. I reloaded the whole thing on another partition.

Is there a standard procedure to do a post-mortem analysis on a non-bootable install? I looked at all of the files with the right date stamp in /var/log (boot.log, dmesg, kern.log, syslog, ...). What am I missing?

BrianP

For a "post-mortem analysis" it would be good to look in the log files for exactly what packages were upgraded. And of those, which are the most likely to be related to the bootup sequence. Or is there any indication that the update came from a mirror that was only partially synced, so that not all of the required packages were downloaded?