I did a fresh install on my oldworld tonight, because I had been having problems since I upgraded it to woody a week or two ago. It had been locking up on me when I started dselect, it had done this 4 or 5 times. It locks up completely, no disk activity, no keyboard response, etc. Usually this was when I was scrolling through dselect entries, or making a selection with - or +.
Since the machine had crashed this way several times, I began to get nervous about my filesystems. They had been fsck'd each time, but I was suspecting maybe some hard errors were creeping in. I started to get segfaults from jed when I invoked it, and tried to use nano to edit a config file when it failed. When I use ^O in nano to write the file, it said it wrote 15 lines and then froze the computer solid. When I got back in, I checked that file and it had not actually been modified. So I started witha brand new set of partitions, freshly initialized, and did a new install with b-f 3.0.16. It went very smoothly, although still waiting 1/2 hour with the extracting whiptail dialog on the screen. I used tasksel to ask for the desktop package, and then asked for dselect so I could get jed also. I started to scroll through dselect, and no more than 30 seconds later it had frozen solid again. I went back in, and avoided dselect. I asked for the desktop package again, and it downloaded and installed everything, though I had to repeat a couple of times to get it all set up. Finally, I did apt-get install jed, and it installed fine. I verified I could invoke it without getting a segfault. Then I opened nano to edit a config file again. It gave me the exact same behavior, freezing at the exact same point. (This is the completely fresh install.) I have the same versions of dselect and nano, and jed for that matter, installed on my newworld machine, and they give me no problems whatsoever. I think the jed problem was due to all the crashing. But the nano and dselect problems seem to be hard failures on the oldworld. They are both dependent on libncurses5. When I went back in to the machine, it failed to start. init.d/rcS tried to clean /var/lock, /var/run, and /var/run/utmp, and couldn't find any of them. The system halted after trying to start syslogd. These are obviously very damaging type crashes. Let me know what steps I can take to narrow down this problem further. -- *----------------------------------------------------------------* | .''`. | Debian GNU/Linux: <http://www.debian.org> | | : :' : | debian-imac: <http://debian-imac.sourceforge.net> | | `. `'` | Chris Tillman [EMAIL PROTECTED] | | `- | May the Source be with you | *----------------------------------------------------------------*
