bridging vlan on ISL trunk

From: John Conzone (jkconzone@xxxxxxxx)
Date: Sat Sep 16 2000 - 19:25:52 GMT-3


   
       First, before anyone says why are you doing this, I'm doing this
   because this is the twisted stuff that I might see on the lab. Now for
   the question.
   
   I have three routers connected to a Cat5 as such.
   
                                R4
                                  |
                                  |
   R3---------------CAT5----------------R5
   
       R3 has ethernet into a port on Cat in VLAN 1. R5 has ethernet into
   a port on Cat in VLAN 2. Both in same ip subnet , 192.168.10.3 and
   192.168.10.5.
       R4 is connected via ISL trunk to Cat. R4 has subint fa0/0.1 encap
   ISL 1 and subint fa0/0.2 encap isl 2.
       Now I have no ip on either subint on R4. I create bridge group 1
   on R4, and place both sub interfaces in the bridge group. I try to
   ping betwen R3 and R5 with no luck. I then create BVI, put it in
   subnet, route IP on bridge group and R3 and R5 can ping.
       I need someone to tell me if what I think is happenig when it
   fails without IRB is what is really happening, or if not what is
   happening. I'm thinking the reason the pings fail without the bvi is
   that R3 sends a packet out to the Cat. Say its an arp (cause it is, I
   debugged it). The packet gets to the Cat, gets encapsulated in ISL1 to
   go up the trunk to R4.
       R4 is bridging bewteen the subs, so he takes the packet bridges it
   over to fa0/0.2, encapsulates it again in ISL 2 and sends it down
   fa0/0.2 towards the cat and vlan 2. So now I'm thinking I have a
   packet with 2 ISL headers. The packet goes back down the trunk to the
   Cat to VLAN 2. The cat strips off the outermost ISL header (isl2) and
   forwards the packet to R5 with the ISL1 header still on it. R5 gets a
   ISL encap packet and doesn't know what to do with this. Or perhaps
   becasuse the packet isn't an etehrnet packet headin out the etehrnet
   port, the Cat drops it.
       When I use a BVI on R4, he strips the fist ISL header, so when the
   packet reaches R5, it is a ethernet packet with no ISL header.
       Of course this could be totally wrong, but it sounds good! Someone
   let me know if I'm wrong. (As if I have to ask that with this
   bunch!<G>)
   
               John



This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:24:57 GMT-3