[Info-vax] hub better than switch?!

[John Wallace]

On Mar 7, 7:36 am, hel... at astro.multiCLOTHESvax.de (Phillip Helbig---
undress to reply) wrote:
> In article
> <d8fc7c70-0d91-4abe-a7a7-0326ec1ed... at y10g2000vbn.googlegroups.com>,
>
> John Wallace <johnwalla... at yahoo.co.uk> writes:
> > Carrier check failures aren't a good sign, but there aren't many of
> > them, and it's been a few days, and presumably you've been tinkering
> > with cables or interfaces or both?
>
> No, no tinkering!
>
> > I've lost track of what box/interface is configured with what
> > settings.
>
> It's an XP1000 but I don't know the settings.
>
> > I presume the above display isn't from an interface you are expecting
> > to be full duplex, right?
>
> Not sure.
>
> If there are dumber cards on the network, presumably this one can't be
> full duplex, or can it talk full duplex to one and not to another?
>
> > My understanding was that with full duplex properly configured, you'd
> > not see things like "initially deferred" or indeed anything collision
> > related because the properly configured switch with full duplex at
> > both ends of a cable handles all that by magick.
>
> Some of what we see might be from the hub.
>
> > Zero the counters and come back in a day or so?
>
> OK, if there's time.  I'm really busy the next couple of weeks.

"If there are dumber cards on the network, presumably this one can't
be full duplex, or can it talk full duplex to one and not to another?"

A proper switch hides this kind of thing from each of its connected
devices. Your XP1000 or any other device could be configured as 10 or
100, full or half duplex, and as long as the switch was doing the
right thing, it wouldn't matter (other than there'd be a potential
performance impact). The silicon inside the switch sorts it all out.

See you in a few days?