Reset Search
 

 

Article

Delayed BGP next-hop router update when BFD for BGP session goes down.

« Go Back

Information

 
TitleDelayed BGP next-hop router update when BFD for BGP session goes down.
Objective
  • Next-hop route failover takes around 120 seconds when BFD is enabled on a BGP VLAN. This can be observed in the logs below. 
11/27/2017 04:03:56.48 <Verb:BGP.RtMgr.RtImprt> [VR 2] BGP,ipv4-unicast: 10.1.1.0/24 -> 100.127.160.2 [nhVR 2] rt origin IBGP flags 0x90000C13 ttl 44 rank 1 opqlen 0

11/27/2017 04:03:56.48 <Verb:BGP.RtMgr.RtImprt> [VR 2] BGP,ipv4-unicast: 100.127.0.48/32 -> 0.0.0.0 [nhVR 0] rt origin BGP flags 0x80000C03 ttl 0 rank 0 opqlen 0

11/27/2017 04:03:56.48 <Data:BGP.NeighborMgr.TxMsgDetail> [VR 2] Sending UPDATE to Peer 100.127.160.2 (0), Id 1686077442

11/27/2017 04:01:56.48 <Noti:BGP.NeighborMgr.PeerEstTrans> [VR 0x00000002] Peer 100.127.160.33 (0) has entered or left Established state, EST? 0, holdtime: 45, passive: 0

11/27/2017 04:01:56.48 <Info:BGP.NeighborMgr.PeerFSMDegrade> [VR 2] Peer FSM State Deteriorated: Peer 100.127.160.33 (0), Established -> Idle, Input: 20

  • Without BFD, the delay is only a second. 
2017-10-16 15:27:36.08 <Verb:BGP.RtMgr.RtImprt> [VR 2] BGP,ipv4-unicast: 0.0.0.0/0 -> 0.0.0.0 [nhVR 0] rt origin BGP flags 0x80000C03 ttl 0 rank 0 opqlen 0 
2017-10-16 15:27:36.09 <Verb:BGP.RtMgr.RtImprt> [VR 2] BGP,ipv4-unicast: 0.0.0.0/0 -> 100.127.160.2 [nhVR 2] rt origin IBGP flags 0x90000C13 ttl 10252 rank 1 opqlen 0 
2017-10-16 15:27:36.08 <Info:BGP.NeighborMgr.PeerFSMDegrade> [VR 2] Peer FSM State Deteriorated: Peer 100.127.160.33 (0), Established -> Idle, Input: 3 

 
Environment
  • EXOS
  • Summit
  • BlackDiamond
Procedure
  • This behavior is expected when BGP graceful restart is configured along with BFD on that BGP enabled interface. 
  • Once BFD session goes down, BFD notifies this to the BGP process and BGP process treats it as graceful restart and waits for restart time (default is 120 seconds) before marks its peer as down. Till this period (120 seconds) the e-BGP routes are marked as stale and kept active in BGP routing table.
  • Run the command : configure bgp restart none  to get around this problem. 
  • BGP Graceful restart cannot co-exist with BFP on a BGP interface.
Additional notes
Refer to xos0070492 for more information. 

Feedback

 

Was this article helpful?


   

Feedback

Please tell us how we can make this article more useful.

Characters Remaining: 255