Frequest BGP notification logs.

<timestamp, hostname> rpd[10705]: %DAEMON-4-BGP_NLRI_MISMATCH: bgp_process_caps: mismatch NLRI with 10.x.x.9 (Internal AS 65500): peer: <inet-unicast>(1) us: <evpn>(8388608) (instance master)

<timestamp, hostname> rpd[10705]: %DAEMON-4: bgp_process_caps:4181: NOTIFICATION sent to 10.x.x.9 (Internal AS 65500): code 2 (Open Mess age Error) subcode 7 (unsupported capability) MP capability afi 1, safi 1 <inet-unicast>

The IBGP configurations on the route reflectors are using the loopback IP addresses.

 group BGP-ROUTE-REFLECTOR {

type internal;

local-address 10.x.x.1;

cluster 10.x.x.1;

neighbor 10.x.x.9;

}

group BGP-ROUTE-REFLECTOR {

type internal;

local-address 10.x.x.2;

cluster 10.x.x.2;

neighbor 10.x.x.9;

}

But the leaf switch are using the internal loopback addresses.

group BGP-ROUTE-REFLECTOR {

local-address 127.x.x.9;

neighbor 127.x.x.1;

neighbor 127.x.x.2;

}

The following log messages are triggered frequently.

<timestamp, hostname> rpd[10705]: %DAEMON-4-BGP_NLRI_MISMATCH: bgp_process_caps: mismatch NLRI with 10.x.x.9 (Internal AS 65500): peer: <inet-unicast>(1) us: <evpn>(8388608) (instance master)

<timestamp, hostname> rpd[10705]: %DAEMON-4: bgp_process_caps:4181: NOTIFICATION sent to 10.x.x.9 (Internal AS 65500): code 2 (Open Mess age Error) subcode 7 (unsupported capability) MP capability afi 1, safi 1 <inet-unicast>

The log messages stopped after configuring the right IP loopback addresses on the leaf switch.

  group BGP-ROUTE-REFLECTOR {

local-address 10.x.x.9;

neighbor 10.x.x.1;

neighbor 10.x.x.2;

}