The set up I have now is the followng:
I have a router running eigrp whose FastE interface is conencted to an ethernet port on a catos switch. That catos switch's MSFC is also running an instance of eigrp. Both devices are exchanging eigrp updates.
Here are the specifics:
Th FastE interface of the router has an IP address within the range of vlan 38. Vlan 38 is configured, along with its SVI, on the L3 switch, too. The eigrp process on the MSFC has, of course, the "no passive interface vlan38" command configured to allow routing updates to pass through to the router and back...
Now that L3 switch is going to have its MSFC disabled and perform only L2 switching. Furthermore, that switch will be trunked up to an IOS distro switch with the SVI for vlan 38 configured on it. In other words, the L3 duties originally being performed by the catos switch have been migrated UP to a new distro switch, which is running IOS, not catos.
So, to for the router to continue to be able to exchange eigrp updates within the new environment, that router connection must be migrated up to the new distro switch (lets say to interface gigabitethernet 3/1) and then that ge3/1 interface must be allowed to pass eigrp updates with the "no passive interface" command on the IOS switch's eigrp process.
Am I right? Or is it sufficient to keep the router connected to what will become the L2-only switch, as long as vlan38 is allowed on th etrunk? I dont think the latter is correct.
There is no need to change the physical connectivity, if Router can be reached from int vlan 38 on MSFC.
For peers to form eigrp neighbors they should be able to reach each other directly so that eigrp multicast/unicast packets can be exchanged. In ur case vlan38 will be considered as one broadcast network spanning accross multiple switches. So Eigrp peers would be able to talk to each other.
Please rate this if the info was of any use.
Deepinder Singh Babbar