Based on my analysis, this may caused by the FHR's "Null Register Suppress" implementation which may not comply with PIM-SM standard (RFC 7761). This bug may exist for Nexus 7000 platform as well.
The possible solutions described in the bug may not comply with the RFC 7761. The possible solution potentially could be: After the FHR suppress the "Null Register" due to "no data from source" and the expiration of the "Register Stop Timer", the FHR router should remove the "Received Register stop" state for the (S,G). The FHR should punt future data packet to CPU upon the expiration of the "register stop timer" and restart the "source register" process. This will trigger the RP to build/refresh the (S,G) tree to the FHR. Changing the leak frequency cannot fix the issue since there still will be a delay. Punt data packet to CPU after " SG expiry timer expired" solution may not work since the (S,G) entry actually didn't expire. So I feel it is more reasonable to make the "register stop timer" expiration as the trigger to punt the multicast data packet to CPU.
1. After the (S,G) entry is built on the FHR, the first hop router marks it as "Received Register stop"
2. According to RFC 7761, if there is multicast data to the group G, the FHR supposes to send "Null Register Msg" every "Register Stop Timer". 3. Nexus implementation of the FHR will suppress this "Null Register Msg" if there is no multicast data to the group with reason "due to no data from source". This makes senses, but what FHR does next may not comply with the RFC.
4. FHR SHOULD also remove the ""Received Register stop" state for the (S,G), for future received data packet to this group, FHR SHOULD encapsulate it and start a "Source-Register Process". But Nexus FHR doesn't implemented this way.
Here is the RFC except:
The Designated Router (DR) on a LAN or point-to-point link encapsulates multicast packets from local sources to the RP for the relevant group unless it recently received a Register-Stop message for that (S,G) or (*,G) from the RP. When the DR receives a Register-Stop message from the RP, it starts a Register-Stop Timer to maintain this state. Just before the Register-Stop Timer expires, the DR sends a Null-Register message to the RP to allow the RP to refresh the Register-Stop information at the DR. If the Register-Stop Timer actually expires, the DR will resume encapsulating packets from the source to the RP.
... View more
.jpg "VIP Advisor"){kind=icon}
