On Thu, Mar 18, 2004 at 09:19:26AM +0100, Alexander Troppmann wrote: > I'm running a webserver with Fedora Core 1 and have troubles with an (yet > unkown) daily running cronjob. It seems that a process started from > /etc/cron.daily at 4:02h in the morning freezes the linux kernel so the > server returns back to work by a hardware reset only... :-( Yes I had a similar thing a few months ago. My hang was at 4:09AM. It was due to a bug in the SMP fedora kernel. That bug is still in the latest FC1 kernel. Are you running the SMP kernel? Here are some interesting bugzillas to read if you are: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=109962 https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=109497 > Unfortunatly I cannot find out which one is the process responsible for the > server crash. I tried to run each job from console myself but everything was > fine, no errors or kernel freezes. You can get more information by enabling the sysreq key. That is, in /etc/sysctl.conf, set this: # Controls the System Request debugging functionality of the kernel kernel.sysrq = 1 then run "sysctl -p" Then, when the system hangs, on the console hit 'alt-prntscrn-P' to see where the processors are locked. alt-prntscrn-H will show you other commands you might want to try. A serial console is needed to collect or view all the output. -- Norman Gaywood, Systems Administrator School of Mathematics, Statistics and Computer Science University of New England, Armidale, NSW 2351, Australia norm@xxxxxxxxxxxxxxxxx Phone: +61 (0)2 6773 2412 http://turing.une.edu.au/~norm Fax: +61 (0)2 6773 3312 Please avoid sending me Word or PowerPoint attachments. See http://www.fsf.org/philosophy/no-word-attachments.html
