Tyan Tiger 2460 (Re)

Robert G. Brown rgb at phy.duke.edu
Thu Apr 25 20:21:21 PDT 2002


On Thu, 25 Apr 2002, Steve Gaudet wrote:

> Has anyone reported these issues to Tyan's technical support?  Tyan knows
> that their products sell very well in the Linux cluster market.  So not
> addressing these issues would cause customers to look elsewhere.

We filed a ticket yesterday, but AFAIK no response yet.  I've been
snooping their website (and 3com's, and google, and anything else I can
think of) trying to find some hint that this problem has been reported
before.  It is very strange that such a complete screw up could still
exist -- how can anybody be using Tigers in 1U or 2U cases if this
problem is universal?  I posted here half hoping to hear somebody say
"Ah, you need to change XXX in the bios, you idiot".  I'd even welcome
the idiot part, if only I could get things to work.

Still, it is completely reproducible on the Tigers we have, it has hit
other Tiger users at Duke, some of whom posted here a month or two ago
and finally gave up and replaced their 2460's with 2466's in sheer
frustration (or so we suspect -- we only figured out the details I
posted yesterday and haven't had time to verify that the problems are
indeed the same) and it is a REAL problem.  The 3c905 unfortunately is
about 1 cm too tall to fit in a 2U case straight up if the metal
backplate is taken off, although the ATI Rage actually will fit.

I just registered (shudder, but what's a bit more SPAM in my already
hi-cal email diet:-) at amdmb.com, which looks like a support forum for
amd motherboard and will try a post there.  It looks like it gets the
attention of both AMD and Tyan engineers and may be faster than Tyan's
support page to return something useful.  People are (purportedly)
SELLING cluster nodes with Tiger 2460's in a 1U or 2U chassis with a
riser; it MUST work somehow, for some bios flash or configuration.

If not and the Tiger 2460 has a dark secret (it doesn't, uhh, actually
work in a 1-2U rackmount configuration, oops) well, I'm preparing a few
flaming torches and oiling my pitchfork...unless of course Tyan makes it
right.  Very quickly -- this has cost us at least a month of
potential productivity at this point, and looks to cost us at least a
few days more.

   rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu






More information about the Beowulf mailing list