Brocade FastIron Ethernet Switch Layer 3 Routing Configuration Guide - Manual (Page 466)

IPv4 BGP4 commands for different configuration levels (Continued) TABLE 93
Command Global (iPv4 and IPv6) IPv4 address family unicast
update-time x
Entering and exiting the address family configuration level
The BGP4 address family contains a unicast sub-level.
To go to the IPv4 BGP4 unicast address family configuration level, enter the following command.
device(config-bgp)# address-family ipv4 unicast
device(config-bgp)#
NOTE
The CLI prompt for the global BGP4 level and the BGP4 address-family IPv4 unicast level is the same.
Syntax: [no] address-family ipv4 unicast [ vrf vrf-name ]
The default is the IPv4 unicast address family level.
The vrf option allows you to configure a unicast instance for the VRF specified by the vrf-name
variable.
To exit an address family configuration level, enter the following command.
device(config-bgp)# exit-address-family
device(config-bgp)#
Syntax: [no] exit-address-family
BGP route reflector
A BGP device selects a preferred BGP4 route for a specific prefix learned from multiple peers by using
the BGP best path selection algorithm, and installs the BGP4 route in the Routing Table Manager
(RTM). The BGP device marks the preferred BGP4 route as the best route, and advertises the route to
other BGP4 neighbors. Generally, the RTM route table size is larger than the number of unique BGP4
routes in the BGP4 route table. All preferred BGP4 routes are installed in RTM and are marked as the
best BGP4 routes.
However, in certain configurations it is possible that the total number of preferred BGP4 routes may
exceed the RTM route table size limit. Therefore, some preferred BGP4 routes may not be installed in
the RTM, and the BGP device is not able to forward traffic correctly for those BGP4 routes. Those
BGP4 routes are not considered as the best BGP4 routes, and are not advertised to other BGP4
neighbors because traffic miss-forwarding or packet drop can occur.
When a BGP device is configured as only a route reflector server, and is not placed directly in the
forwarding path, it is possible to mark all preferred BGP4 routes as the best routes to be advertised to
other BGP4 neighbors even if the routes are not installed in the RTM. To support the behavior of a
BGP device as a route reflector server in such a scenario, use the always-propagate command and
the rib-route-limit command.
Entering and exiting the address family configuration level
466 FastIron Ethernet Switch Layer 3 Routing Configuration Guide
53-1003087-04