Below topology was used for this post, and all the configuration happened on two Cisco CSR1000v
BGP Nonstop Forwarding
- During normal NSF operation, CEF on the active RP synchronizes its current FIB and adjacency databases with the FIB and adjacency databases on the standby RP
- While switching over, the traffic is depended on CEF, once the routing protocol is converged, FIB will be updated
- RIB repopulating happens prefix-by-prefix, thus the same for FIB and adjacency table
- For BGP NSF, graceful-restart needs to be configured on both ends of a BGP session. Although one end could be only NSF-aware (not SSO capable)
- BFD can’t be enabled simultaneously with NSF for BGP
- SSO is not integrated into EIGRP, hence only NSF awareness is supported
By entering bgp graceful-restart
some default parameters will be also configured:
CSR1(config)#do s run | s bgp router bgp 1 .... bgp graceful-restart restart-time 120 bgp graceful-restart stalepath-time 360 bgp graceful-restart ... CSR1#sh ip bgp neighbors | i Grace Graceful Restart Capability: advertised Graceful-Restart is enabled, restart-time 120 seconds, stalepath-time 360 seconds
The optional restart-time
determine how long peer routers will wait to delete stale routes before a BGP open message is received.
The optional stalepath-time
determine how long a router will wait before deleting stale routes after an end of record (EOR) message is received from the restarting router.
When graceful-restart is not enabled:
CSR2#sh ip bgp neighbors | in Grace Graceful Restart Capability: received Graceful-Restart is disabled
In case this feature is configured after session establishments, BGP sessions should be cleared!
CSR2(config-router)#bgp graceful-restart All BGP sessions must be reset to take the new GR config