Re: [nsp] Catalyst 5505/WS-X5225R and 10/100 autodetection

From: Rafi Sadowsky (rafi@noc.ilan.net.il)
Date: Mon Oct 04 1999 - 09:24:59 EDT


On Mon, 4 Oct 1999, Dmitri Kalintsev wrote:

> On Mon, Oct 04, 1999 at 09:42:26AM +0200, Rafi Sadowsky wrote:
> > (auto-negotiation may work depending on Solaris version & patch level
> > you probably want the latest hme patch from the recommended patchset )
patches 102979 SunOS 5.5: hme driver patch
patches 104212 SunOS 5.5.1: /kernel/drv/hme patch
patches 105795 SunOS 5.6: /kernel/drv/hme patch
>
> [ lots of useful info is skipped ]
>
> Hmm. I feel somewhat misunderstood.
now I feel misunderstood so we're even ;-)
(see the attached sun doc)
> The point is not on how to make Sun
> Solaris's hme to run in FDX, the point is:
>
> "Am I doing something wrong? The very same Sun (as I've pointed out later)
> auto-negotiates very nicely with 3com switch, meaning that there is NOTHING
> wrong with NIC's drivers. So I make an assumption that there is something
> wrong with the catalyst switch, as it's behaving erratically in my opinion"
well I can tell you for that for a sun AXi with the built-in 100BT
interface with Solaris 5.6 connected to bay switches ( different models )
        the 5.6 May 98 Solaris release auto-negotiation works
while the original unpatched 5.6 the sun gets set to 100HDX while the
bay gets set to 100FDX ( which for some reason causes a lot of errors :)

I suggest you try setting both sides to 100 FDX before giving up

AFAIK the reason is that the auto-negotiation standards were not stable
when some of the vendors released the card/drivers
so that if auto-negotiation you may try a driver/IOS/?? upgrade
or just set it up statically
if your Solaris kernel doesn't generate a age about 10/100 FDX/HDX messages
when the link goes up ( not just at boot ) then you defintly need a patch

if you install an HME patch don't get suprised if it start doing the
negotiation fine with cisco & stops working with the 3COM switch
(such are the "standards" that we have to live with )
>
> So, what is wrong with this catalyst? Is it known to behave in this manner?
> Because if it is, than I'm *definitely* unsatisfied with this module and I
> will have to have it changed to say 100M-only module.
>
> Thanks,
> --
> (KD1350) CCNA Dmitri E. Kalintsev
> CDPlayer@irc Senior network engineer at NetActive Internet
> dek @ hades.uz phone: +27 (0)11 719-0333 fax: 444-8900
> http://hades.uz UIN:7150410 cell +27 (0)82-336-1033
Hope this helps

Regards
        Rafi
>





This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:06 EDT