--Fingers, what you really mean to say is use Ctrl-Alt-Backspace to try and kill the X-server. Alt-Bkspc alone won't do it.

Quote Originally Posted by fingers99
1.If at all possible, use alt backspace (it may be that only X is locking up) if that fails, use the magic keys rather than a hard reset. Press:

Alt + SysRq

(with these two held down) next press in order:

r s e i u b

Often (but not always!) one of these two will do the job.
--To me, it sounds maybe like a DMA error on the HD.

--Cuddles, pls post results of (root):
' hdparm /dev/hda ' as well as hdc... DMA usually gets enabled in /etc/init.d/bootmisc.sh.

--As an aside, Cuddles - the Alt-Sysrq sequence is still a 3-finger combo. What the keys do, in order:

Alt-Sysrq-R: unRaw (not needed usually)
Alt-Sysrq-S: SYNC all pending writes to HD's
Alt-Sysrq-E: Send sigTerm to all processes
Alt-Sysrq-I: Send sigKill to all
Alt-Sysrq-U: Umount (remount all partitions R/O)
Alt-Sysrq-B: ReBoot

--Note: You should wait a few (2-5) seconds between each sysrq request, so all the writes can commit and things have a chance to settle down. To get help on Sysrq, get to a text console and hit Alt-Sysrq-H. (Make sure to use the LEFT Alt key.)

--AFA testing the HD: Boot knoppix cd into runlevel 2:
boot: ' knoppix 2 vga=normal ' (and whatever other codes you normally use)

' badblocks -c 128 -s -v /dev/hda ' == Equivalent to "surface scan"; tests R/O for bad sectors on entire disk.

--Also run badblocks for /dev/hdc. (There is additionally a non-destructive R/W mode; see ' man badblocks '.)

--1st thing I did after ordering a new 80GB Hitachi drive about a month ago, even before partitioning, was run badblocks R/W on the entire drive - and sure enough it went bad on me during the test. I sent it back and exchanged it for a Seagate 80GB. The Seagate tested OK and is doing fine. Remember to put a HD fan on the drive if it's 7200rpm or greater - they will help the drive last longer.

--Finally, if you have to replace the HD I would recommend the following:

o Test the entire new HD with badblocks in "destructive R/W mode" before partitioning it, to see if it fails.

o If it passes, consider formatting it with Reiserfs instead of ext3 (yes, I'm just a little biased) with the "noatime" mount option:

Code:
/dev/hdb11 /mnt/bkps2  reiserfs  defaults,noatime,notail,rw 0 0
--If you will likely have more Large files than small on the partition, (or if you value speed over space-savings) "notail" makes sense as well. But I basically put that "noatime" option on *all* my Linux mounts, and 99% of my partitions are Reiserfs; and I haven't had a significant problem with that setup for YEARS.

Quote Originally Posted by Cuddles
I can not just re-install - this is NOT an option. I did the "just re-install" thing three times now. Once to get my old system over to my new system, Second to get my modem to work, and Third to get ALSA working - now, as I am getting something else on my system going - the solution AGAIN is to re-install - NO - NO WAY...

As for the magic keys - shesh, can you have "invented" anything less complicated ? In windows you have the "three finger salute", whereas in Linux, you have to do the "four finger salute" - now we have this "contorsionistic" set of key presses to do what??? Regain system control ???? The keyboard is locked, the system is locked, nothing is running, its dead when this happens...

How about this as an idea -=- The sound file (speach 102) is on my new 80 gig hard drive, the ReZound "main" 45 to 50 minute file used to generate these "smaller" speach files IS ON THE SAME DRIVE. K3b was trying to burn from files on the SAME DRIVE - when the system locks up, it is when I am doing "something" on the 80gig hard drive - this drive is hdc2 - whereas my system is located on hda1...

Could this be a failing hdc2 device???

How about some routines, or utilities, that can test the device? Or the file system on the device? Or the files on that device?

fsck doesn't seem to think that the "file system" is a problem, does it also test the files? I also noted during one of these "fsck" checks on the hdc2 device (80gig) that 24.5% was non-continueous - the device is not even 20% full - I thought that fsck checked and corrected these "defrag" issues?

Anyone have a disk test utility, something that I can ensure that my new hard drive, which apparently seems to be at the "root" of a common problem (system lock ups) may be causing these problems???

Thanks...
Cuddles