Path: utzoo!news-server.csri.toronto.edu!rutgers!att!emory!swrinde!cs.utexas.edu!hellgate.utah.edu!csn!boulder!daemon
From: SNMS4@vaxb.acs.unt.edu (Kevin W. Mullet, Univ. of N. Tx)
Newsgroups: comp.dcom.sys.cisco
Subject: Troubles with routing Novell on Cisco.
Message-ID: <33212@boulder.Colorado.EDU>
Date: 12 Mar 91 21:21:31 GMT
Sender: daemon@boulder.Colorado.EDU
Lines: 62


         We have a Cisco AGS+ running 8.1(21). I've been unsuccessful
         at getting Novell IPX to route between E0 and E6 (or any
         other interfaces, for that matter).  Both have been set to
         novell-encapsulaion ARPA, both have been given the right
         Novell (XNS) network numbers, and I've even tried XNS access
         lists, Novell helper lists and helper addressees to no avail.

         Here's the problem.  Normally the Novell network on E0 get's
         it's RIP and SAP updates from a Netware 286 external
         brid..uh.. router that routes between our campus wide (older)
         IBM-PC Network backbone and their local Ethernet.  This works
         fine but slow.

         When we have both the Cisco and the Novell router routing, we
         get weird problems on both sides of the Cisco with SNA
         Gateway problems, NACS problems, none of which we can
         conclusively pin down to the Cisco.  They happen without the
         Cisco, but not as frequently.

         When the Novell network on E0 is using the Cisco exclusively
         to get to the rest of campus, an slist from a workstation on
         the E0 side shows servers going up and down like a mad dog. 
         If I run a looping batch file to keep doing SLIST, I get 39
         servers, 34, 30, 24, 28, 32, 38, 39, and so forth.

         Both sides are doing ARPA encapsulation a la Ethernet II with
         8137 typefield.  Doing a sho servers at the AGS+ appears
         consistent with no dissappearing servser, but servers on the
         E0 side fade in and out to E6, and servers on the E6 side
         fade in and out to E0.  

         The problem appears also in the output from debug
         novell-sap.  It's littered with the likes of "such and such"
         DOWN when it really isn't.

         (I've got a file with the debugging output and configuration
         stuff in it if anyone's interested.)

         That, in a nutshell (or rather a NETshell?) is the problem. 
         I've had Nick Thille from Cisco tech support on the phone for
         about a week with no visible means of success.  I've recently
         FTP'd a 4-5 meg Sniffer capture of each Ethernet to Cisco in
         hopes that they'll be able to figure something out from that.

         1:  Does anything come to mind?  (feel free to look around on
         the router, but I've got the Novell stuff turned off now).

         2:  I notice both our E0 and E6 interfaces have the same
         MAC-level address.  Is this normal?

         3:  Who else, that y'all know of, is doing Novell routing
         with Ciscos on Ethernets with ARPA encapsulation?

         4:  Any other ideas that I can try?

         r.s.v.p.
         (Please reply to me directly and I'll post to the net.
         thanks.)

         -Kevin Mullet
          University of North Texas