RE: question about NETBIOS, RIF fields, and DLSw+

From: Jim Mathers (jmathers@xxxxxxxxxx)
Date: Sun Jun 17 2001 - 18:50:51 GMT-3


   
Hello and thanks for all the awesome responses!

First of all, "source-bridge spanning" is on. Roman--It looks like I was
seeing the behavior that you are describing with some exceptions. The hosts
were sending NETBIOS-NAME-QUERIES and since I did not have transparent
bridging on, they did not work. However, according to your email, the WIN98
host should have then sent a frame with a RIF and this is not occuring. When
I get back to my lab, I will do that registry hack that looks like it will
hard code the 98 box to always include a RIF? Could someone confirm this?
Also, I am using the drivers that came with the card. They are Madge 4/16
cards that I bought at Goodwill. I believe they are Microsoft drivers. How
can I find out for sure? What difference would putting IBM drivers on the
box make? And where can I download those? Please comment. Thanks again Roman
and everyone else for your input on this issue.

Jim Mathers

-----Original Message-----
From: Roman Rodichev
To: psimmons@netcomuk.co.uk; Jim Mathers; ccielab@groupstudy.com
Sent: 6/17/01 2:59 PM
Subject: Re: question about NETBIOS, RIF fields, and DLSw+

Jim,

Peter is right, you need "source-route spanning" turned on.

I never tried it, but I know that Windows host will first send out
NetBIOS
name query without source information field. So if you have transparent
bridging turned off they won't work. Then if the host doesn't see
responses
it will send out name query with RIF set.

try this also

HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\VXD\Netbeui

and set QueryAlwaysUsesSourceRouting to 1

And also, are you using IBM or Microsoft drivers for your TR NICs?

>From: "Peter Simmons" <psimmons@netcomuk.co.uk>
>Reply-To: "Peter Simmons" <psimmons@netcomuk.co.uk>
>To: "Jim Mathers" <jmathers@inflow.com>, <ccielab@groupstudy.com>
>Subject: Re: question about NETBIOS, RIF fields, and DLSw+
>Date: Sun, 10 Jun 2001 19:26:57 +0100
>
>Jim,
>
>Have you got the "source-bridge spanning" command on the TR interfaces?
>
>By default, NetBIOS uses spanning explorers, and SNA uses All-routes
>explorers. If you don't use the IBM spanning tree protocol to build a
>spanning tree path for STE explorers, then the interface drops them, as
it
>regards itself to be in "blocking" mode for these frames. Using the
>"source-bridge spanning" command puts the interface automatically into
>"forwarding" mode for the NetBIOS STE frames, and allows them to be
>forwarded onto the DLSw process (or wherever else).
>
>This is an SRB thing, as you've seen TB just forwards the frames as you
>expect.
>
>Try a "debug DLSw reach" before and after applying the command and see
the
>difference.....
>
>Lots more info on this on CCO, search on "source-bridge spanning" as a
>starter.
>
>Hope this helps,
>
>Regards
>
>Pete S.
>
>(Four days 'til Brussels.......)
>
>
>-----Original Message-----
>From: Jim Mathers <jmathers@inflow.com>
>To: Jim Mathers <jmathers@inflow.com>; 'ccielab@groupstudy.com'
><ccielab@groupstudy.com>
>Date: 17 June 2001 17:21
>Subject: RE: question about NETBIOS, RIF fields, and DLSw+
>
>
> >Further, I just changed the Token Ring interfaces from SRB to
Transparent
> >bridging by removing the statement "source-bridge 20 1 100" and
adding
>the
> >statement "bridge-group 20" on both of the routers and it works fine.
I
>also
> >had to remove the "source-bridge ring-group 100" from the DLSw+
configs
>and
> >add a "dlsw bridge-group 20". Now everything is working fine. I am
forced
>to
> >assume that WIN98 is NOT generating a RIF, but instead is sending a
>NETBIOS
> >name-query. I am still looking for feedback on this. Any help would
be
> >appreciated. Also, all Cisco documentation that I have read says that
you
> >have to do SRB for Token Ring. I guess that I am misunderstanding
>something
> >because I am successfully doing transparent bridging between two
remote
> >Token-Rings. Any feedback? Please comment. Jim
> >
> >-----Original Message-----
> >From: Jim Mathers
> >Sent: Sunday, June 17, 2001 10:35 AM
> >To: 'ccielab@groupstudy.com'
> >Subject: question about NETBIOS, RIF fields, and DLSw+
> >
> >
> >Hello,
> >
> >I have two hosts connected to two different MAUs that in turn connect
to
>two
> >different routers via Token-Ring interfaces in my lab. The two
routers
>are
> >separated via an IP backbone. I have configured DLSw+ to bridge
between
>the
> >two Token Rings. The hosts are running WIN98 with only NETBEUI
>configured.
> >When I click on Network Neighborhood, the hosts do not find each
other.
>Nor
> >does DLSw build a circuit from end to end. I get my DLSw+ Peers to
>connect
> >fine. I have been troubleshooting this all night and after getting
very
> >frustrated, I pulled up the DLSw+ troubleshooting document on Cisco's
web
> >site. It says that it is possible that the Operating System (WIN 98)
on
>my
> >hosts will not generate a RIF, and, if this is the case, DLSw+ will
fail
> >because the router will look for a RIF on the incoming frames. If the
>frames
> >do not contain RIFs, then the router will not be able to source-route
>bridge
> >the packet onto the virtual ring. Has anyone seen this? I want to put

>this
> >behind me. I just finally joined the list and thought it would be a
good
> >idea to bounce this of off the crowd. My lab is scheduled for
September
>23rd
> >in San Jose. Thanks in advance.
> >
> >Jim Mathers
> >**Please read:http://www.groupstudy.com/list/posting.html
> >**Please read:http://www.groupstudy.com/list/posting.html
>**Please read:http://www.groupstudy.com/list/posting.html



This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:31:25 GMT-3