Thursday, August 24, 2017

Juniper MX480 BGP taking longer to update the BGP table

This would be one interesting problem.

Nice one to encounter if you aren't facing an urgent issue.

Simple solution guys - adjust the tcp-mss to 1500 or there about.

let me know if this sorts your issues.




ME3600 / ASR 920 BGP flaps


Aug 21 07:42:51.456: %FMFP-3-OBJ_DWNLD_TO_CPP_FAILED: F0: fman_fp_image:  PREFIX 64.210.135.0/24 (Table id 0) download to CPP failed
*Aug 21 07:43:21.466: %FMFP-3-OBJ_DWNLD_TO_CPP_FAILED: F0: fman_fp_image:  PREFIX 192.104.147.0/24 (Table id 0) download to CPP failed
*Aug 21 07:43:51.476: %FMFP-3-OBJ_DWNLD_TO_CPP_FAILED: F0: fman_fp_image:  PREFIX 104.84.96.0/20 (Table id 0) download to CPP failed
*Aug 21 07:44:22.519: %FMFP-3-OBJ_DWNLD_TO_CPP_FAILED: F0: fman_fp_image:  PREFIX 210.102.108.0/24 (Table id 0) download to CPP failed
*Aug 21 07:44:52.748: %FMFP-3-OBJ_DWNLD_TO_CPP_FAILED: F0: fman_fp_image:  PREFIX 206.198.240.0/24 (Table id 0) download to CPP failed


Above mainly because the box can't handle a full BGP table, starts having issues installing the routes on the RIB as soon as memory is done.

Good boxes but not for full scale BGP deployments.