Here's my 2-cents:
We have two SGI workstations: an O2 connected to the spectrometer and
a separate Indigo2 in the same room used as a separate data processing
station. The O2 on the instrument has (probably) never crashed for
anything short of a power outage. (I second the motion that putting the
spect on the ec0 interface gets rid of a lot of problems with instrument
communications.) The Indigo2 has ALWAYS had memory errors which usually
PANIC it. The only memory in it has been SGI supplied memory. This has
happened as the os has been upgraded from the original 5.3 to the
current 6.5.11, and our extra copy of xwinnmr has progressed from 1.? to
2.6v4. All the hardware (memory, motherboard, powersupply, etc.) except
the disk, tape drive, CD and external SCSI peripherals has been swapped,
but the problem still persists (it happened two weeks ago). The
frequency of this happening has been reduced, and there were some times
when the frequency of errors was higher. It seems manageable now that it
happens only about every few months. The crash files indicate many
different processes were running when the crashes occur, so no specific
pattern is apparent. This is perplexing to me, but not enough to make me
want to use NT!
Thanks,
Dan
Guillermo Moyna wrote:
>
> Oh boy, I created a monster! Anyway, glad (in a wired way) that I'm
> not the only one seeing this. By the way, after one month of aparent
> calm, kaput, another crash. This time I'm looking into the
> 'analysis.n' files, generated after each crash, and I see that they
> all point to problems with the ec1 (the second ethernet card)
> statistics, and also memory-like problems. I will check the make of
> the memory as suggested by Sebastien last week, and I'll let you all
> know...
>
> Cheers,
>
> Guillermo
>
--
***************************************************************
Daniel W. Bearden (Dan) Phone: (843) 762-8527
National Ocean Service FAX: (843) 762-8700
219 Ft. Johnson Rd. e-mail: ccehbr.nmrlab_at_noaa.gov
Charleston, SC 29412
***************************************************************
Received on Mon Nov 26 2001 - 13:59:21 MST