🚀 go-pugleaf

Hello to you all,

I finally fetched the guy from the local support of Cisco and he brought the
documentation of 9.0 to help us.  It was essentially as follows:

ip forward-protocol spanning-tree
ip forward-protocol udp 111

interface ethernet 0
 .
 .
bridge-group 1

interface ethernet 1
 .
 .
bridge-group 1

 .
 .

bridge 1 protocol dec

This allows the broadcasts to UDP port 111 (portmapper) go to all the subnets,
the portmappers will take care of everything else.  We allready killed a server
on a couple of subnets and after killing/restartting ypbind on the clients
on these subnets things seem to work just as we planned (i.e. the broadcast
reached the servers on the other subnets and get processed correctly).

That's all.  It turns out that it is very important to enter the commands in
the right order, while doing trial-and-error we had to enter some commands
more than once in several combinations before we got the effect we wanted.

One last problem is that one of the segments (132.65.16.0) doesn't have
broadcasts routed properly (i.e. it behaves just as before).  The subnet mask
as defined on the Cisco is 255.255.224.0 (top 3 bits are subnet), which means
that subnet 132.65.16.0 is subnet 0, could be that this is the reason things
work differently there?  Another thing which might matter is that the "helper-
address" used to point to a host under this subnet (but not any longer). The
configuration right now is just the same for all the interfaces:

interface Ethernet 1
ip address 132.65.x.1 255.255.224.0
! seems like the following doesn't matter
ip broadcast-address 132.65.0.0
bridge-group 1

Thank you all for your help.

Regards,


--

--Amos Shapira (Jumper Extraordiner)
C.S. System Group, Hebrew University, Jerusalem, Israel
amoss@cs.huji.ac.il