[openib-general] Question

[Hal Rosenstock]

On Mon, 2005-02-28 at 18:16, Ronald G. Minnich wrote:
> It does seem there's a gremlin in the network. We need simple ways for 
> people to find such gremlins. It's also worrisome that one bad "thing" 
> (whatever it is) can get opensm confused to the point that you can't bring 
> up the network at all.

Some bad things are supposedly rules which are _not_ to be violated
(perhaps like LinearFDBTop one you cited). 

Others are failure modes that clearly need to be dealt with better both
on the front end (by OpenSM) and the diagnostics to find them.

-- Hal