Can anyone help with my understanding of the virtual fastethernet interface in the Cisco 3020 switches. I have installed two Cisco 3020's which are working independently in a HP blade server enclosure. The network has multiple VLAN's and one (VLAN 20) is used for managment and one (VLAN 10) is used for servers. Each 3020 has a 802.1q trunk back to the core with all the VLANs on it. VLAN 20 has a subnet of 172.20.0.0/16 and the onboard administrator on the HP blade enclosure and all the switches in the enclosure including the 3020's on their fastethernet F0 interface have IP addresses in the 172.20.0.0/16 range. If i give my laptop a 172.20.0.0/16 address and plug directly into the onboard administrator port of the enclosure, i can ping/telnet all the switches including the 3020's. So all in the enclosure is working ok.
I have configured an external interface on the 3020's as a VLAN 20 access interface and plugged the HP onboard administrator port into it to give access to the enclosure from the rest of the network. Everything is working fine except for one strange issue. I can now ping the onboard administrator and every switch in the enclosure except for the 3020 switch the onboard administrator is plugged into this 3020 can't be accessed even via the onboard admin GUI. If i swop the onboard administrator onto the other 3020 then i can ping the first 3020 again but not the one that now has the onboard administrator in it this now becomes unaccessible via the 172.16.0.0/16 address.
The 3020 Fastethernet F0 is not able to have VLAN's assigned to it and is not accessible externally, it just connects internally to the onboard administrator.
This means with the present setup at least one 3020 is unaccessible.
I have attached the 3020 configs below... port g0/17 is the uplink to the cores, port g0/22 is the link to the onboard admin and g0/23-24 will link to other switches in the enclosure.
Eventually there will be a redundent onboard admin so both g0/22 ports on the 3020's will be used.