[lug] Re: error -2147483630 reading header:

[matthew.w.mcillece at lmco.com]

Thanks for all the great responses!  Specifically,

Grant wrote:
>Maybe it is a bad chip that does not like the pressure.  Try pushing on it
while
>installing.  Maybe one of the little bitty microscopic traces got cracked.
>Maybe the whole thing is smoking crack....

I'll try this (pushing on the CPU ;).  I already tried holding the heatsink
on without the fan and clamp.  Now I'll try pushing on it with something
else.  Not my finger though because the CPU gets way too hot without the
heatsink.

Kevin wrote:
>well, that error is coming from 
>/misc/src/anaconda/isys/uncpio.c
>
>        if ((rc = getNextHeader(&fd, &ch))) {
>            fprintf(stderr, _("error %d reading header: %s\n"), rc,
strerror(errno));
>            return CPIOERR_BAD_HEADER;
>        }
>
>It's apparently happening from the loader. The loaders job is to load
>the installer into memory and run it. 
>
>Could it possibly be that the heatsink is touching or interfereing
>with the memory? You might try running xmemtest86 on it (can't hurt). 
>http://reality.sgi.com/cbrady_denver/memtest86/
>
>Also, does the processor have an external cache? might try disabling
>that in the bios. 
>
>Other than that it sounds like the cpu is somehow really flaky...

Thanks for showing me where this error comes from.  I guess I couldn't find
it myself because I didn't have the source code for anaconda installed.  The
heatsink and RAM are several inches apart at the closest.  I guess I could
install the OS with my other CPU since it seems to work better and then run
xmemtest86.  Also could check for an external cache disable option in the
bios.  But I think your last conjecture, that the CPU is flaky, is most
likely correct.

Jason wrote:
>    My gut reaction is overclocked or mislabeled CPU. I know I got errors

It may have been overclocked at one time.  I did get it used.  The
motherboard was jumpered for a 2.8V core CPU voltage when I first got it
with the CPU, and the CPU is labeled as having a 2.2V core.  The actual CPU
to BUS frequency ratio was set correctly for this 350MHz CPU when I got it
though.

Ralf wrote:
>If memory serves me right i think i read an short note in a
>german computer magazine that described severe problems with
>AMDs caused by certain brands of CPU fans (from what i recall
>these fans _wheren't_ intended to run with AMD). If you want
>i can try and dig for the article.

Thanks for the offer to look up the brands of problematic CPU fans, but I'm
pretty sure it's not the fan.  I can disconnect the fan from power and
remove it from the heatsink and I still get the errors as long as the
heatsink is still on the CPU.  I've also tried a different heatsink and fan
and got the same errors.