No appropriate icecandidates on pro/enterprise in 2.19

I’m trying OV pro/enterprise again to benefit from MediaSoup beta.

I have a working OV CE instance on vSphere which I’ve cloned and added my pro credentials to. I have an newly created media node.

The OV pro server spins up fine and connects to the specified media node as entered in the .env with KMS value. All good in the logs.

When I connect a working openvidu-browser app to the pro stack, the console doesn’t show any correct remote icecandidates as it does with CE.

The media node and pro server both have separate public IPs, NATted behind a pfSense firewall. There is internal NAT reflection with split subnets to avoid any issues. This is the exact configuration as for CE instance (it’s a clone with same IPs).

The console for the pro stack just shows some local candidates.

Any ideas? At the moment I can’t do anything with OV pro/enterprise as no clients can send/receive video.

Many thanks,


What you mean by “NATted behind a pfSense firewall”?

Master node and media node have different public IPs? Every node needs a different set of ports opened (different between them). Please review on premises deployment.

Do you have those ports opened?

I recommend you to deploy OpenVidu PRO from scratch in new machines to avoid any issue with prev OpenVide CE config

Hi Micael,

Thanks for coming back to me.

Since I posted this, I provisioned a brand new Pro instance but with same IP at CE instance (CE instance switched off) so I can show it’s not a NAT issue for the server node. Same thing happens.

pfSense is providing a 1:1 NAT with 2x different public IPs, one for server node, one for media node. Each with the correct ports allowed.

I am connecting to these from within my LAN and so NAT shouldn’t be an issue.

On CE, the candidates list is massive and it reports all correct public IPs in IPv4 and IPv6, and all local IPs. Works great.

The Pro instance is configured exactly the same, same IPs. The media node is a fresh Ubuntu instance with a similar config to the server node. Both on same subnet internally.

When connecting with Pro running, I get none of the big list of candidates that CE reported, just some local ones of this machine. In the SDP report I can see a TCP candidate but it doesn’t get listed.

Can you please try with enterprise version?