(Apologies for the OT nature, but for completness' sake, it should stay on-list)
On Fri, Nov 25, 2005 at 07:29:05AM +1100, Felix Karpfen wrote: > Right at the start of the install Linux (Debian) asked if I wanted to > check the HD for faulty Sectors. To which I foolishly said "No". That's not necessarily a mistake per se, Felix. It's also conceivable these errors were realised sometime after install. But, in general, running badblocks (which is what that question was essentially asking) is the more prudent course of action. > It appears that I have maligned both fvwm2 and Gnome. And, if the > problem is greater than faulty Sectors and cannot be fixed by fsck, > then I had better invest in a second HD. Indeed. Given the HW symptoms, I'm more likely to make the assumption that these played a part in the errors you were receiving with FVWM and GNOME in this instance. If it does crash again, however, I would be very interested in seeing a backtrace from a corefile. > The advice in the replies has helped to avoid total disaster. I am > fairly conscientious about making backups. But I usually leave them > on a separate partition of the HD; once every 3 months or so, I burn This is perhaps OK in the short-term, but no so good in the long-term. What happens if one critical partition goes -- or a specific sector is beyond repair? That could have a knock-on effect in accessing the other partitions on that drive. > the backups onto a CD. It looks as though I need to do that much more > frequently - while awaiting the arrival of the next crash. *nods*. I also tend to (if you have the options) scp some of my important files as well, as a backup. As well as a CD backup, both locally stored here, and one remotely --- just in case. :) -- Thomas Adam -- I'm brutal, honest, and afraid of you.
