I'm using udpcast for 1.5 year now and since I upgraded to a 3.2.x
kernel, I'm experiencing some weird issue on udpcast.
The setup didn't changed at all since the time it worked except the
kernel (same hw devices / network /..).
The behavior we saw is that after a while (~ 30Gb of data) some clients
looks like frozen.
We started to investigate on it and the can see the following:
- the multicast stream is still reaching the host (proc/net/igmp is ok,
tcpdump -pni show the traffic too)
- the udp-receiver seems block on a read() call
- udp-cast doesn't exist due to wrong data even if the timeout is
specified (the timeout worked by the past)
We are using 2011.07.10 version of updcast for a while now. It didn't
got changed since the time it worked.
Does anyone experienced that or does some have some ideas to understand
what's happening ?
Show replies by date