Administrator Guide

Excessive traffic is generated when the join process from the RP back to the source is blocked due to a new source
group being permitted in the join-filter. This results in the new source becoming stuck in registering on the DR and the
continuous generation of UDP-encapsulated registration messages between the DR and RP routers which are being sent
to the CPU.
• Prevent the PIM SM router from creating state based on multicast source and/ or group.
ip pim join-filter
Understanding Multicast Traceroute (mtrace)
Multicast Traceroute (mtrace) is a multicast diagnostic facility used for tracing multicast paths.
Mtrace enables you to trace the path that a multicast packet takes from its source to the destination. When you initiate mtrace from a
source to a destination, an mtrace Query packet with IGMP type 0x1F is sent to the last-hop multicast router for the given destination.
The mtrace query packet is forwarded hop-by-hop untill it reaches the last-hop router.
NOTE: If the system initiating the mtrace is the last-hop router, then the Query message will not be initiated. Instead,
the router sends the request message to it previous router.
The last-hop router converts this query packet to a request packet by adding a response data block. This response data block contains the
last-hop router’s interface address. The response data block inserted by the router also contains the following information:
• Incoming interface details
• Outgoing interface details
• Previous-hop router address
• Forwarding Code
• Query Arrival Time
• Routing Protocol details
The last-hop router calculates the path to reach the source in the reverse direction of the multicast data traffic. Based on this calculation,
the last-hop router estimates the possible next-hop neighbor that is located in the direction towards the source and forwards the request
packet to that neighbor.
Each router along the multicast path fills its response block in a similar manner. When the mtrace request reaches the first-hop router, it
sends the response (with IGMP type 0x1E) to the response destination address specified in the mtrace query.
The response may be returned before reaching the first-hop router if a fatal error condition such as "no route" is encountered along the
path.
If a multicast router along the path does not implement the mtrace feature or if there is an outage, no response is returned.
When the initiator does not get a response for a specified time interval, the system performs a hop-by-hop expanding-length search to pin
point the location in the network where the problem has occurred.
NOTE: You cannot configure the wait time. It is fixed to 3 seconds.
Important Points to Remember
• Destination address of the mtrace query message can be either a unicast or a multicast address.
NOTE:
When you use mtrace to trace a specific multicast group, the query is sent with the group's address as the
destination. Retries of the query use the unicast address of the receiver.
• When you issue an mtrace without specifying a group address (weak mtrace), the destination address is considered as the unicast
address of the receiver.
• If the CLI session is terminated after the mtrace command is issued, then the response is ignored.
• System ignores any stray mtrace responses that it receives.
• Duplicate query messages as identified by the IP source, and Query ID (tuple) are ignored. However, duplicate request messages are
not ignored in a similar manner.
• The system supports up to a maximum of eleven mtrace clients at a time.
NOTE: The maximum number of clients are subject to performance restrictions in the new platform.
• Mtrace supports only IPv4 address family.
Multicast Features
581