White Papers

● Continues saving routes received from the peer if the peer advertised it had graceful restart capability. Continues forwarding
traffic to the peer.
● Flags routes from the peer as Stale and sets a timer to delete them if the peer does not perform a graceful restart.
● Deletes all routes from the peer if forwarding state information is not saved.
● Speeds convergence by advertising a special update packet known as an end-of-RIB marker. This marker indicates the peer
has been updated with all routes in the local RIB.
If you configure your system to do so, the system can perform the following actions during a hot failover:
● Save all forwarding information base (FIB) and content addressable memory (CAM) entries on the line card and continue
forwarding traffic while the secondary route processor module (RPM) is coming online.
● Advertise to all BGP neighbors and peer-groups that the forwarding state of all routes has been saved. This prompts all
peers to continue saving the routes they receive and to continue forwarding traffic.
● Bring the secondary RPM online as the primary and re-open sessions with all peers operating in No Shutdown mode.
● Defer best path selection for a certain amount of time. This helps optimize path selection and results in fewer updates being
sent out.
To enable graceful restart, use the configure router bgp graceful-restart command.
● Enable graceful restart for the BGP node.
CONFIG-ROUTER-BGP mode
bgp graceful-restart
● Set maximum restart time for all peers.
CONFIG-ROUTER-BGP mode
bgp graceful-restart [restart-time time-in-seconds]
The default is 120 seconds.
● Set maximum time to retain the restarting peer’s stale paths.
CONFIG-ROUTER-BGP mode
bgp graceful-restart [stale-path-time time-in-seconds]
The default is 360 seconds.
● Local router supports graceful restart as a receiver only.
CONFIG-ROUTER-BGP mode
bgp graceful-restart [role receiver-only]
Filtering on an AS-Path Attribute
You can use the BGP attribute, AS_PATH, to manipulate routing policies.
The AS_PATH attribute contains a sequence of AS numbers representing the route’s path. As the route traverses an AS, the
ASN is prepended to the route. You can manipulate routes based on their AS_PATH to affect interdomain routing. By identifying
certain ASN in the AS_PATH, you can permit or deny routes based on the number in its AS_PATH.
AS-PATH ACLs use regular expressions to search AS_PATH values. AS-PATH ACLs have an “implicit deny.” This means that
routes that do not meet any Match filter are dropped.
To configure an AS-PATH ACL to filter a specific AS_PATH value, use these commands in the following sequence.
1. Assign a name to a AS-PATH ACL and enter AS-PATH ACL mode.
CONFIGURATION mode
ip as-path access-list as-path-name
2. Enter the parameter to match BGP AS-PATH for filtering.
CONFIG-AS-PATH mode
{deny | permit} filter parameter
This is the filter that is used to match the AS-path. The entries can be any format, letters, numbers, or regular expressions.
You can enter this command multiple times if multiple filters are desired.
For accepted expressions, refer to Regular Expressions as Filters.
3. Return to CONFIGURATION mode.
AS-PATH ACL mode
exit
182
Border Gateway Protocol IPv4 (BGPv4)