Administrator Guide

Route-refresh
This section explains how the soft-reconfiguration and route-refresh works. Soft-reconfiguration has to be configured explicitly for a
neighbor unlike route refresh, which is automatically negotiated between BGP peers when establishing a peer session.
The route-refresh updates will be sent, only if the neighbor soft-reconfiguration inbound command is not configured in a
BGP neighbor and when you do a soft reset using clear ip bgp {neighbor-address | peer-group-name} soft in
command.
If the neighbor soft-reconfiguration inbound command is already configured in the BGP neighbor, the route refresh updates
will not be sent. If you want to use route-refresh updates instead, remove the neighbor soft-reconfiguration inbound
configuration and do a hard reset once, using clear ip bgp {* | as-number | ipv4| ipv6 | neighbor-address |
peer-group-name} command. If you remove neighbor soft-reconfiguration inbound configuration for an individual
neighbor (not part of peer-group), it is sufficient to do a hard reset only for the individual neighbor. If the neighbor is part of a peer-group
and when neighbor soft-reconfiguration inbound is removed from the peer group, you need to do a hard reset for the peer-
group.
If neighbor soft-reconfiguration inbound command is not configured ever in the router, then doing a soft reset is enough
for the route-refresh updates to be sent.
Route-refresh updates for IPv4 and IPv6 prefixes
This section explains the route-refresh functionality in different combinations for IPv4 or IPv6 prefix configured with IPv4 or IPv6
neighbors.
By default, the IPv4 prefixes is sent for all the neighbors irrespective of IPv4 address family is enabled or disabled. Following are the
different scenarios during which the route-refresh message is sent for the corresponding neighbors:
By default, the route-refresh message for IPv4 prefixes is sent for both the IPv4 and IPv6 neighbors. Following is an example configuration
in which IPv4 prefixes are enabled by default and the corresponding route-refresh message:
DellEMC(conf-router_bgp)# show config
!
router bgp 100
redistribute connected
neighbor 20.1.1.2 remote-as 200
neighbor 20.1.1.2 no shutdown
neighbor 20::2 remote-as 200
neighbor 20::2 no shutdown
!
address-family ipv6 unicast
redistribute connected
exit-address-family
!
DellEMC(conf-router_bgp)#do clear ip bgp 20.1.1.2 soft in
May 8 15:28:11 : BGP: 20.1.1.2 sending ROUTE_REFRESH AFI/SAFI (1/1)
May 8 15:28:12 : BGP: 20.1.1.2 UPDATE rcvd packet len 56
May 8 15:28:12 : BGP: 20.1.1.2 rcvd UPDATE w/ attr: origin ?, path 200, nexthop 20.1.1.2,
metric 0,
<output truncated>
DellEMC(conf-router_bgp)#do clear ip bgp 20::2 soft in
May 8 15:28:04 : BGP: 20::2 sending ROUTE_REFRESH AFI/SAFI (1/1)
May 8 15:28:04 : BGP: 20::2 UPDATE rcvd packet len 56
May 8 15:28:04 : BGP: 20::2 rcvd UPDATE w/ attr: origin ?, path 200, nexthop 20.1.1.2,
metric 0,
<output truncated>
In the below example, under the IPv6 address family configuration, only the IPv6 neighbor is enabled using neighbor ipv6–address
activate command. If you execute, clear ip bgp neighbor-ipv4–address soft in command, only the IPv4 route-refresh
update is sent. If you execute clear ip bgp neighbor-ipv6–address soft incommand, both the IPv4 and IPv6 route-
refresh updates are sent. Following is an example configuration in which IPv6 prefixes is enabled for a IPv6 neighbor and the
corresponding route-refresh message:
DellEMC(conf-router_bgp)# show config
!
router bgp 100
redistribute connected
neighbor 20.1.1.2 remote-as 200
182
Border Gateway Protocol (BGP)