SUMMARY: Sparc 10 Ethernet Maddness

From: jeff goldfarb (
Date: Fri Mar 19 1993 - 07:16:43 CST

I received several suggestions and some feedback from others
experiencing unusual ethernet collision rates with Sparc 10s...
Unfortunatly, none on point, and no sites reporting exactly what
we have seen!

My original post (below) could have been a bit clearer. The application
was a direct upgrade from an SS2, and is not realy *doing* any unusual
processing or network activity... but it *is* inducing trouble under load.

Sun has not explained what we have seen, but has provided a fix which
has, for the moment, removed (or masked) the problem. The key here is
that we experienced the difficulty *with* the 4m suppliment, but
when it was *removed*, the network problem disappeared! [Though no
network problems were reported as the subject of the supplement].

We have been given a 'rev A' version (seems to be 704-3416-11),
Patch-ID# 100743-02 SS10 Jumbo Fix. This seems to NOT break
the network... and we hope, does what the original patch set
was meant to do.

Although the patch date is Nov 92 (for the new rev) this does
not seem to be shipping with, at least *our* machines, and
is not the first thing on Sun supports mind...

If our experience is at all typical, I would be most wary of
the earlier version of the 4m supplement.

Any follow-up from the net will be appreciated -- I will re-summarize
if anything of substance appears.

Thanks, as always,


Jeff Goldfarb |
OTRA, Inc. | +1 908 389 4613 (direct)
6 Industrial Way West| +1 908 542 6908 (fax)
Eatontown, NJ 07724 | att mail: !jgoldfarb
+1 908 542 6900 | applelink: OTRA


Sun Managers:

We are experiencing network problems with several of our new SS 10s
(Model 20/30). Running 4.1.3 *with* 4m suppliment.

Under network load (moderate to heavy) the systems 'go mad' and
apparently flood the network... the collision rate goes to
'infinity' and 'bad' things happen to other machines on the net...
Simply removing the '10' brings things back to normal. Same
effect whether AUI port or twisted pair. Same effect on different

We have had several recommendations from Sun regarding patches... no
success yet in getting either an explaination or a stable configuration.

Have any of you experienced similar problems... or a solution?

We are attempting to isolate the problem further... I will report
on advice/solutions.




This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:07:37 CDT