Affected: Juniper Networks Junos OS, Juniper Networks Junos OS Evolved

On Juniper Networks Junos OS and Junos OS Evolved devices, BGP session flapping can lead to a routing process daemon (RPD) crash and restart, limiting the attack surface to configured BGP peers. This issue only affects devices with BGP damping in combination with accepted-prefix-limit configuration. When the issue occurs the following messages will appear in the /var/log/messages: rpd[6046]: %DAEMON-4-BGP_PREFIX_THRESH_EXCEEDED: XXXX (External AS x): Configured maximum accepted prefix-limit threshold(1800) exceeded for inet6-unicast nlri: 1984 (instance master) rpd[6046]: %DAEMON-3-BGP_CEASE_PREFIX_LIMIT_EXCEEDED: 2001:x:x:x::2 (External AS x): Shutting down peer due to exceeding configured maximum accepted prefix-limit(2000) for inet6-unicast nlri: 2001 (instance master) rpd[6046]: %DAEMON-4: bgp_rt_maxprefixes_check_common:9284: NOTIFICATION sent to 2001:x:x:x::2 (External AS x): code 6 (Cease) subcode 1 (Maximum Number of Prefixes Reached) AFI: 2 SAFI: 1 prefix limit 2000 kernel: %KERN-5: mastership_relinquish_on_process_exit: RPD crashed on master RE. Sending SIGUSR2 to chassisd (5612:chassisd) to trigger RE switchover This issue affects: Juniper Networks Junos OS: 17.2R3-S3; 17.3 version 17.3R3-S3 and later versions, prior to 17.3R3-S8; 17.4 version 17.4R2-S4, 17.4R3 and later versions, prior to 17.4R2-S10, 17.4R3-S2; 18.1 version 18.1R3-S6 and later versions, prior to 18.1R3-S10; 18.2 version 18.2R3 and later versions, prior to 18.2R3-S4; 18.2X75 version 18.2X75-D50, 1…