A very good idea. I think the first thing to is to search for existing tools. This is surely the first thing you thought of ;-). We don't have to reinvent the wheel but to improve or add new features or even writing new applications from scratch. I like the idea of something like a Swiss Army Knife. We have very handy Swiss Army knives here and some of them come with a lot of tools which can be useful in whatever situation. Anyway whatever the name you give it should serve its purpose. Well for bad HW, I am thinking of a Linux embedded diagnostic board. What about if the first sector of a HDD is damaged. Is it possible to make the read/write heads start reading from another position on the disk? This might be possible with Assembly language programming. There are a lot of nifty tool like that which can be part of this recovery set. I think we have a lot of data recovery tools out there.
http://freshmeat.net/projects/crk/ (refer to crash recovery)
(first thing I'll do is backup my user and configuration data)
What about putting autonomic computing features in Linux e.g. would be Nitix
Ok the list could grow longer and longer but I like the idea and am willing to help.
Good Luck Tod
David
On 8/23/06, Tod Merley <todbot88@xxxxxxxxx> wrote:
Hi All!
I would like to start a project (OK, perhaps there is already one
ongoing, if so, I would like to know about it) dedicated to recovery
from a system crash (broken X, bad HW, anything making the machine
non-functional). I am thinking of first a set of applications which
gather information about the running system (or what one should look
like when it is running)(HW list, SW list, Boot sector copy, config
files, critical files md5s, results of lspci, lsmod, "normal" log
files (/var/log) etc.... Then a set of applications designed to go in
(not booting from the crippled system) and gather the same sorts of
information for comparison and understanding the problem now crippling
the computer in the shop. And then a set of applications which fix,
flag, or tell the owner the bad news determined about the problem
computer.
Perhaps the process could make for the developers a standard "Crash
Cart Packet" consisting of parts of the logs and command results
(maybe an X-packet, Kernel-packet, Audio-packet, etc).
Well, that is it. Just an idea.
I really do not know how to start or be part of such a thing. I guess
I will follow those before me and send an e-mail.
If interested in the motivation - read on:
Last night I wrote an e-mail to a gentleman who lost some computers.
He thought it might be FC5 (He was doing a fresh install or upgrade).
I have seen several on the lists needing crash help.
I also responded to another gentleman who was experiencing an
"updated" Xorg which broke (no GUI!!). I encouraged him to copy some
of the basic log files and help us all heal from our very human
tendency to break things.
Today I woke up, pressed the on button on my computer, got a cup of
coffee, and came back to an ncurses error screen partially obscured by
an exit to a prompt. My silly old video card already has some
irritations with X/Dapper (hash on the screen after monitor has put
itself to sleep) so my prompt was without cursor and the screen was
changing colors as I went (I think part of the ncurses applet was
still running somehow along with the shell).
I thought I had swallowed a bug!
However when Puppy and then Dapper Live booted - xorg.conf the same
but the Xorg binary of very recent date - I remembered the problems I
commented on and replaced my Xorg from the live CD and was happy.
Please let me know if there is interest!
Thanks!
Tod
--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
