On Fri, Jun 18, 1999 at 08:39:11PM +0200, Werner Reisberger wrote: > I am running a debian machine (PC PII 350 MHZ, 128 MB RAM) as an Internet > server. Today the server crashed the second times after running one week > without any problems. The crash was caused by a kernel panic. Nothing has > been written to a log file. Below is an excerpt which I found on the > console: > > unable to handle kernel paging request at virtual address e7fde95c > > current -> tss.cr3=00101000, %cr3=00101000 > *pde=00000000 > Oops: 0002 > CPU: 0 > > The panic happened when a cronjob was running (glimpse ...) which does a > lot of IO and consumes a good amount of memory. Maybe connected with this > problem is a kernel message which occasional appears at this times but > doesn't cause a crash: > > Jun 5 05:47:25 memo kernel: attempt to access beyond end of device > Jun 5 05:47:25 memo kernel: 08:0c: rw=0, want=67136178, limit=610438 > Jun 5 05:47:26 memo kernel: EXT2-fs error (device 08:0c): > ext2_free_blocks: Freeing blocks not in datazone - block = 67136177, count > = 1
Some type of filesystem corruption, try booting in single-user mode and running e2fsck -c /dev/(whatever) to detect physical hard drive problems. > > I am running a kernel 2.2.3 with a hamm system. Could there be any > incompatibility between this kernel version and parts of the hamm > distribution? Should I deinstall glimpse and compile a binary myself on the > system? I don't think glimpse itself is the problem. > > I switched to debian because I thought it would be at least as stable as > Suse or RH, but unfortunately my experiences until now are bad. I had > to reboot the system too many times. You shouldn't be running 2.2.3, there are some filesystem corruption problems in the early 2.2.x releases. Upgrade to 2.2.10, that should be better. > > Any hints appreciated. > > --Werner > > > -- > Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null > > -- Stephen Pitts [EMAIL PROTECTED] webmaster - http://www.mschess.org