ARP learning and p2p ping stopped working after LMIC restart/swap on MX304 platforms

This issue is observed on MX304 platforms during an LMIC restart/swap when firewall filters with interface or ARP policers are applied across interfaces belonging to two different LMICs. In such cases, policers on different LMICs can share the same memory address on separate PFE partitions, which is valid during normal operation. However, when one of the LMICs is restarted, the policer configured on the other LMIC is also incorrectly triggered for an update. As a result, when the update for the actual restarting LMIC is processed, a memory address collision occurs during policer reprogramming. This collision in memory usage leads to discard filter installation, resulting in traffic drops until the FPC is restarted to restore normal forwarding.

This issue can be identified when ARP learning and P2P (Point-to-Point) ping fail immediately after an LMIC restart, while traffic is restored upon removing firewall filters with policers and PFE trace shows packets being dropped.

Restart the FPC to resolve the issue temporary.

Issue fixed in below releases :

evo:25.4R2-EVO 

evo:26.2R2-EVO 

evo:26.3R1-EVO 

junos:25.2R2-S2 

junos:25.4R2 

junos:25.4R2-S1 

junos:26.2R1 

junos:26.2R2 

junos:26.3R1