Message "Bridge channel send: no opened channel" occurs from time to time

Dear All,
I already faced to this problem right after joining a room. The error message then occurs continuously.
In my case after specifying
org.ice4j.ice.harvest.NAT_HARVESTER_LOCAL_ADDRESS=
org.ice4j.ice.harvest.NAT_HARVESTER_PUBLIC_ADDRESS= in the sip-communicator.properties, I just make disappear this error message.
Now I see the same error message but it just happens for a while and disappears. I wonder whether it is due to some instantaneous network saturation somewhere (causing the disconnection). How can I debug the root reason of this problem please?
Many thanks

Is your JVB installed on the same server as your Jitsi Meet stack?

Dear @Freddie ,
I have several JVBs and one of them is installed in the same signaling server (nginx+jitsi meet + prosody + jicofo).
By the way I just figure out that I can even install nginx + jitsi-meet in a separated server from the one hosting jicofo + prosody. Can such topology cause some impact on the system? Is it likely the cause of the “no opened channel” ? If yes, why the error is just shown up from time to time ?
Many thanks

Sounds like loadbalancing is not configured properly. You won’t see the error when you’re using the JVB that’s installed on the same server with JMS, but when a meeting lands on any of the remote JVBs, then the error will show up.

Check your js console while you’re hosting a meeting with 3 participants. Share the log.

Dear Freddie,
In fact, the error message occurred from time to time in the same meeting (so hosting by the same JVB). That is why I don’t think it is related to the loadbalancing. I still vote for some network saturations somewhere. How can I confirm / debug such issue?
The fact that the error message happens not permanently so with 3 participants I dont see such issue. But for about 30 participants in one room, I did observe it sometimes (again not consistantly).
Many thanks