Subject: sig 11s, what's my next step?
To: None <>
From: Jamie Heilman <>
List: port-cobalt
Date: 01/31/2001 01:55:28
Well I've been battling spurious signal 11s for the past few weeks on my
qube2.  Its kinda odd as they come in spades after a certain period of
activity.  I'd like to know if anybody can tell me if using the siop driver
(I know it was sorta iffy at release time) could result in a signal 11 but
no other information in the logs.  I'm asking cause I'm at my wits end as
for what could be the cause of all these signals.  First I'm using 128M (2
64M simms) of 3.3v EDO bought from Cobalt's suggested source, its brand-new
and I'm almost certain it isn't bad.  So when the signal 11s first started
appearing (mostly during software builds ofcourse) I thought, "oh crap, a
bad system board".

So I ran out and bought another Qube2 (too much time invested in this now
to be rational).  Upon swapping the base system I found that after a day or
so of uptime and some light activity, as I started to go through build
things from pkgsrc my signal 11s returned.  Surprised, I decided to try the
RAM I'd got with the new Qube, so after swapping the RAM it ran for another
day with light activity and I actually got 10 or so packages built before
suddenly tar dumped core on me.  After that sig 11s became more frequent,
until most recently I came home to find the machine totally wedged.  I
rebooted and it froze and locked up again while fscking my /home partition
(which perhaps pointedly resides on my scsi disk).  At this point I decided
I'd haul out the other disk that came with the other Qube2, put the 128M of
ram back in my original system, tossed the linux-bearing drive in there,
and brought up cobalt's linux distro.  Then I put a kernel build and a
glibc build into two while loops and just let that run for 2 days,
monitoring stderr for possible signs of segmentation faults, etc.  After 2
days nothing bad had happened except that I eventually ran myself out of
diskspace due to cobalt's inane partitioning scheme and wonky rpm spec file
build instructions.

So now I'm hoping that the siop driver could be the cause of my pain.  I
think what I really need at this point is for somebody to either point me
at a tutorial on how to properly test/debug problems with intermittant
signal 11s, or for somebody to pat me on the head and tell me to lay off
the crack because I'm going in the totally wrong direction.  What it all
boils down to is that I've now sunk a substantial amount of time into my
dream of a home network that doesn't consist of 1500BTU pentiums and
constant fan-induced ear death, and I haven't come this far to see it stop
short due to my own lack of competence.

Jamie Heilman         
"I was in love once -- a Sinclair ZX-81.  People said, "No, Holly, she's 
 not for you." She was cheap, she was stupid and she wouldn't load 
 -- well, not for me, anyway."				-Holly