Bidirectional Forwarding Directions (BFD) sessions flap after GRES or NSR switchover on the NG-RE and SCBE2 setup.
With BFD flaps, the BGP session goes down and the traffic related to BGP routes get lost.
When BFD turns UP, BGP and traffic will be recovered. The syslog message below could help identify the issue:
user@router> show log syslog.log Aug 25 00:53:28 olympic bfdd[12098]: BFDD_TRAP_SHOP_STATE_DOWN: local discriminator: 1274, new state: down, interface: xe-5/1/4, peer addr: 10.255.161.225
This issue could be caused by PR1522261 - BFD with authentication for BGP flaps after GRES or NSR switchover on the NG-RE and SCBE2 setup .
On devices with NG-RE (Next Generation Routing Engine) and SCBE2 (Enhanced Switch Control Board), when BFD authentication for BGP is enabled, the BFD may flap after the NG-RE switchover. After the flap, the device can self recover.
The RE type and CB type can be found using the ' show chassis hardware models' CLI command.
show chassis hardware models'
NG-RE models:
RE-S-X6-64G: 6-core Haswell CPU, Wellsburg PCH based routing engine with 64GB DRAM and 2x 64GB SSD storage for MX240/MX480/MX960
REMX2K-X8-64G: 8-core Haswell CPU, Wellsburg PCH based routing engine with 64GB DRAM and 2x 64GB SSD storage for MX2020 and MX2010
RE-PTX-X8-64G/CB2-PTX: 8-core Haswell CPU, Wellsburg PCH based routing engine with 64GB DRAM and 2x 64GB SSD storage for PTX5K
RE-PTX-2X00x4(JNP10K-RE0) : 4-core Intel IVY Bridge CPU, Wellsburg PCH based routing engine with 32GB DRAM and 2x 50GB SSD storage for PTX10K
RE X10 (JNP10K-RE1) : 10-core Intel Broadwell-EP CPU, Wellsburg PCH based routing engine with 64GB DRAM and 2x 200G SSD storage for PTX10K