Multicast over a GRE Tunnel posted 01/10/2007
- Subject: Multicast over a GRE Tunnel
- From: Cisco study <groupstudy@xxxxxxxxxxxxxx>
- Date: Wed, 10 Jan 2007 10:41:37 +0000 (GMT)
- Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=btinternet.com; h=X-YMail-OSG:Received:Date:From:Subject:To:MIME-Version:Content-Type:Content-Transfer-Encoding:Message-ID; b=VMXnWdd4xvUOHXyhysnXyPiXbA/yzmCqwLwuwt/EMuONDTranqjGXVmxpIgIG/m22I02KBAqognwpqp6iUx/oeVgZxYMcM4Rz04eWaG0f49WfpD8sBfYpby+Uixg0pj0Leiq60RV5CDFwo83IOZwqNe2e66EVZyUJDdC/7xr9jc=;
Hope you are well.
I have a multicast question.
We have setup ip pim sparse-dense-mode Multicast GRE tunnels at each of the 5 regional hub sites . The purpose of the GRE tunnels is to stream the multicast traffic from 4 of the hub sites to the main central hub site. A Windows server at the main hub site collects the multicast data and is suppose to display phone calling statistics for external user calls.
On the server (at the main hub site) we are seeing intermittent multicast traffic coming through. After checking the multicast traffic on the routers locally, the traffic is working fine for each of the local regional sites. However, we are seeing packet loss when the Multicast traffic is going over the GRE tunnels.
The initial configuration was setup as a fully meshed environment, i.e. each site had two tunnels connected to the other sites. But this was causing Multicast packets to drop because of RPF failure.
We overcame this problem by removing the fully mesh setup to only having a single tunnel from each of the hub sites to the central hub site. Now we are not seeing any RPF failures. Good.
The troubleshooting on the following site has been covered and exhausted:
We still are not seeing 100% multicast packets at the central site, only around 50% multicast traffic?
Could anyone shed some light if there is a way to check if the multicast packets are going over the GRE tunnels? Can you recommend a Windows based software or tool, router commands, etc so that multicast source to destination packets can be traced.
Any questions or suggestions welcome.