[Beowulf] Keeping the Athlon MP cluster limping along

[David Mathog]

> Both AMD and Tyan have been pretty decent about fullfilling the terms of
> their 3 year mfrs warranty on both the CPUs and motherboards. 

Unfortunately these nodes fail at relatively long intervals
(days to weeks).  I did speak to Tyan support and they offered to
RMA the motherboards, but they didn't want the CPUs to go
with them.  AMD is presumably the other way
around.  One can easily imagine the resulting "it's their problem"
finger pointing following a 24 diagnostic hour test by both
manufacturers during which neither's component failed.
I wouldn't hesitate to RMA if a node failed dramatically,
preferably with a nicely scorched component to indicate the
exact point of failure.

Regards

David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech