Open jeffrson opened 1 year ago
Oh - it starts working as soon as the sender joins the multicast group as well. I always assumed this is only necessary for listeners, while the sender just uses the group as destination address. Hmm.
Furthermore, I really need sudo iptables --policy FORWARD ACCEPT
. Is this a recommended solution? It's apparently not necessary for smcroute.
Finally, I tried pimd though smcroute works, because smcroute had a certain problem after installation (Debian/Ubuntu systemd): it does not recognize the docker0 interface when it's started. Surely because docker is started after smcroute or because it takes some more time. Unfortunately, pimd seems to have the same problem. It only works if I restart the service from terminal after first login after reboot. Any idea how this could be solved? Should pimd rescan network interfaces?
I read https://github.com/troglobit/pimd/issues/121 and my situation is similar, except that I tried the solution (forward accept) and it still doesn't work :-(
I have a docker container running:
docker run --rm -it -p 2222:2222/udp -v $(pwd):/usr/src/app -w /usr/src/app node:alpine sh
In this container I have a node.js script that receives multicast on the host just fine.I installed pimd 2.3.2 and ran it by
sudo /usr/sbin/pimd -d -f -c /etc/pimd.conf 2>&1 | grep 230.185
, in order to grep for the multicast address I'm interested in. The log starts with these lines as soon as I run my script:tcpdump on the host shows incoming UDP packets, however inside the container it doesn't:
tcpdump -i eth0 -s0 -vv host 230.185.192.108
sudo pimd -r shows this:
Network interfaces have "MULTICAST" - static routes with smcroute work. Did I miss something to set up for pimd?