[zparrish]

We just went through setting up RLM so that it could be accessed by one of our remote locations. In order for the request to succeed, it had to go through the following path:

RLM Host -> Managed Switch -> Firewall -> VPN Connection -> Firewall -> Managed Switch -> Client Host

It also had to route from one VLAN to another. Sounds like a recipe for disaster, right? That just sounds like fun to me

Based on the docs, there wasn't enough info to set this up correctly. In the end, we had to open up the following ports between the firewalls at both locations while the switches actually did the routing between VLANs. Also, make sure any host / client firewall isn't blocking these either:

TCP / UDP 5053 - This one is clearly noted in the docs
TCP / UDP 5045 - This one was implied in a few paths listed in the docs
TCP 45454 through 45474 - These were for network render support, and I suspect aren't really needed for RLM
TCP 49100 through 49500 - I could find no documented reason for these, but based on traffic monitoring, we determined these needed opened as well.

So, if anyone else is having trouble getting their RLM server to work and have no choice but to manage a firewall (or a multiple firewalls), try opening up TCP ports 49100 through 49500 from point to point. It worked for us!

Thanks all!