Re: [cobalt-users] Raq Mysteriously reboots

[flash22@xxxxxxx]

On Fri, 17 Nov 2000, Brian Kosick wrote:

> Hello all,
> 	our Raq3 seems to be possessed, it reboots on its own.  I have not been 
> able to put a finger on why it does it.  There seems to be nothing unusual 
> happening in the logs, "except" for what appears to me to be a memory 
> error.  Even the memory error doesn't show up all the time.  It happens 

Well, from the logs it crashed while reading from the disk while running
log_rotate...given that the log rotate scripts use quite a bit of memory
it's possible you have some bad ram that doesn't get used untill this
point, thus the sporadic crashes...but you really can't nail this down
without testing it...it's equally possible you have bad power, head (bad
cpu fan?) a bad power supply, a bad block on the swap partition...

I agree that the rest of the logs don't look particularly suspiscous, but
remember the logs don't show everything, for example they won't show
'weird' packets sent to the machine if they don't access any service that
logs and don't trip inetd...older linux kernels are known to be crashable
in some cases this way...(tho it's my imprssion that the kernel shipping
witht the raq's doesn't have these kinds of bugs)

If you had a floppy drive i'd recommend memtest86...hmm

Sometimes just taking the ram out and putting it back in is the trick, the
sockets get flaky with age in damp environments...

> very sporadically, for example, it rebooted itself four times in one day 
> last week, and we go along fine for a month maybe more, then all of a 
> sudden it'll do it all over again.

Which would make me think power quality....bad ram rarely 'gets better'

G