David Gotz wrote:
> I have a problem with an IP Multicast application running on Red Hat
> Linux with mrouted version 3.9 beta 3.
>
> The problem is that when my application drops a group (first by using
> IP_DROP_MEMBERSHIP, then closing the socket), the nearest mrouted
> router continues to forward packets from the group to the client.
>
> There occurs even when I have just 1 client running, so it isn't that
> there are other clients on the same LAN requesting the same multicast
> group.
>
> Is there any way to make mrouted act upon the IP_DROP_MEMBERSHIP
> command? How can I get the smallest latency between the IP_DROP_MEMBERSHIP
> command and the actual end of packets being forwarded along my
> local link?
>
> Thanks,
> Dave
>
>
it's very strange because when a client LEAVE the group multicast send a
"igmp v2/v3 leave" then the mrouted daemon have few seconds to leave the
client from group.
Which type of leave send your client to mrouted v2 or v3?
Platform of client?


pppee