I'm trying to monitor BGP peers under VRF IPv4 unicast, using SNMP context. But seems as BGP peers have the same IPv4 addresses, I'm unable to retrieve the correct BGP Status for the peer on SNMP, despite context/VRF definition.
Here is my configuration for test :
vrf definition 100 address-family ipv4 snmp context CTX_100 vrf definition 101 address-family ipv4 snmp context CTX_101 ! snmp-server context CTX_100 snmp-server context CTX_101 ! snmp-server user SNMP_100 SNMP_100 v3 auth sha PrivAuth priv aes 128 PrivPass snmp-server user SNMP_101 SNMP_101 v3 auth sha PrivAuth priv aes 128 PrivPass ! snmp-server group SNMP_100 v3 priv context CTX_100 read VIEW_100 snmp-server view VIEW_100 iso included snmp-server group SNMP_101 v3 priv context CTX_101 read VIEW_101 snmp-server view VIEW_101 iso included
Here are my 2 BGP neighbors configured in 2 differents VRFs
The problem happens when I shutdown one peer, the MIB does not reflect it, I can still see it up and running. If I shutdown the other BGP peer, then the MIB reflects the shutdown. It means that only one peer is shown under the MIB, despite VRF/SNMP context polling?
Is there a way on Catalyst 9300 / IOS-XE to reflect BGP peer status using SNMP polling under VRF correctly ? Seems it works fine under IOS-XR, but can't find a way on IOS-XE 16.12 or 17.04 to run it correctly. And can't find a documentation that could explain it's not possible or possible.
Can it be linked to the fact that I've 2 BGP peers under 2 differents VRF running with the same IP addresses ?