Support Center > Search Results > SecureKnowledge Details
Security Gateway randomly stops forwarding the IGMP / PIM Sparse Mode multicast traffic
Symptoms
  • Security Gateway / Cluster randomly stops forwarding the IGMP traffic.

  • IGMP reports are received from the clients.
    PIM joins are not sent to PIM Rendezvous Point.
    Upon sending new IGMP reports, PIM prune messages are sent to both the PIM Rendezvous Point and to the Senders of the stream.

  • Disabling and re-enabling the PIM Rendezvous Point resolves the issue.

  • PIM and Multicast debug (per sk92598) shows messages like these:

    MMM DD HH:MM:SS.SSSSSS mfc_data_driven_prune: deleting kernel forwarding entry for (230.0.0.1, X.X.X.X)/64 to trigger data driven prune
    MMM DD HH:MM:SS.SSSSSS mfc_get_krt_pktcnt: retrieve usage counts for kernel multicast forwarding entry for (230.0.0.1, X.X.X.X)/64
    MMM DD HH:MM:SS.SSSSSS task_set_option: task IGMP socket 22 option MulticastDelMfc(46) group 230.0.0.1 source X.X.X.X
    MMM DD HH:MM:SS.SSSSSS PIM MRT: join is already a part of j/p listof nbr Y.Y.Y.Y
    MMM DD HH:MM:SS.SSSSSS PIM MRT: Added prune to triggered list of nbr Z.Z.Z.Z
    MMM DD HH:MM:SS.SSSSSS PIM MRT: sending join/prune(54) to Z.Z.Z.Z on eth2-01.2417 (Z.Z.Z.Z)
    MMM DD HH:MM:SS.SSSSSS task_send_packet_iovec: task PIM_IO socket 23 length 54 if_index 24 to 224.0.0.13
    PIM SENT Z.Z.Z.Z -> 224.0.0.13 len 34 join/prune:
    upstream nbr Z.Z.Z.Z num_groups 1 holdtime 210
    group 230.0.0.1/32 num_joins 0 num_prunes 1
    prune source X.X.X.W/32 SPARSE
    MMM DD HH:MM:SS.SSSSSS
    MMM DD HH:MM:SS.SSSSSS mfc_del_oif: deleting oif eth2-01.2417 from (230.0.0.1, X.X.X.W)/64
    MMM DD HH:MM:SS.SSSSSS mfc_krt: parent IF eth1-02.1016
    MMM DD HH:MM:SS.SSSSSS group 230.0.0.1 source X.X.X.W
Cause
  1. (Issue 01714924) Multicast traffic times out and does not resume because no PIM null register packets are being sent when Security Gateway is DR for local sender of PIM traffic (PIM null register packets are sent every 60 seconds while PIM traffic is still arriving from local senders).
  2. (Issue 01952634) In PIM Sparse Mode, low amount of arriving multicast traffic causes RouteD daemon to crash.
  3. (Issue 01946518) In PIM Sparse Mode, low amount of arriving multicast traffic causes the local multicast members to be pruned early.

Solution
Note: To view this solution you need to Sign In .