[lug] Reboot Problem

D. Stimits stimits at attbi.com
Thu Sep 5 21:51:31 MDT 2002


Peter Hutnick wrote:
>>I am surprised that the supposition is HW related. That would be a mess.
> 
> 
> Someone else said he thinks it is memory.  I tend to agree.  Do you have
> more than one memory module in the system?  If so that is an obvious place
> to start troubleshooting.
> 
> To go back a little on what I said about swapping hard disks, there are
> several diskette and CD "Linux on a disk" distros out there.  That might
> be easier for you.  Make sure you pound the system pretty hard if you do
> this.
> 
> Geez, what am I thinking  . . why don't you start with memtest86? 
> (http://www.memtest86.com/)

I'd also agree it is probably hardware. FYI, temperature can make an 
enormous difference in life, and marginal chips might look good if 
cooled. Compiling a kernel is a nice test as well, see if it dies with 
signal 11. The kernel itself will spontaneously reboot if a triple 
exception occurs, rather than doing an OOPS. For software to cause 
instant reboot without hardware problems, I can only say that I have 
never seen this happen except using the original Cygwin Insight debugger 
front end on SMP (it does this on multiple machines with different 
chipsets). FYI, power spikes and marginal power supplies can do this as 
well, running undervoltage can make bits less reliable in the memory in 
the same way that actual bad memory does. If the problem was a power 
spike, permanent damage may have occurred.

D. Stimits, stimits AT attbi.com

PS: If it is not *instant* reboot, but a proper shutdown, maybe you have 
a network card running the wake-on-lan software, being told to do the 
reboot. A whole lot changes if it is a spontaneous *normal* shutdown, 
versus acting like the reset button was pushed, and ignoring normal 
shutdown scripts.




More information about the LUG mailing list