Jitsi clients unexpectedly getting "trying again after 30 seconds" error

Hello,

I’ve set up Jitsi on digital ocean, with one jitsi-meet instance and 2 jvb-only nodes as part of the same blade. I’ve confirmed my turn server is working by blocking port 10000 and seeing the video chat still work.

Lately, some of my users have been complaining that they are able to enter a room but after a couple of minutes get kicked out and get the “Unexpected error, will try again in X seconds” error message. Unfortunately, I have not been able to get console logs from them yet. I think, but I don’t know for sure, that these users are on a corporate network, which potentially explains why they are losing connectivity

On the server, I see many of these:

2020-08-27 19:18:08.723 INFO: [8132] [confId=ce3ace57219bee22 gid=206896 stats_id=Louisa-gCJ componentId=1 conf_name=939287 ufrag=brrt31egpa8on6 name=stream-4e50d314 epId=4e50d314 local_ufrag=brrt31egpa8on6] MergingDatagramSocket$SocketContainer.runInReaderThread#770: Failed to receive: java.net.SocketException: Socket closed

I’m assuming this is the cause, but I’m not sure. How can I go about debugging this?

Thank you very much for your help

1 Like

Did you block the outgoing UDP/10000 on the client side during testing?

Could you learn their client application? Chrome? Firefox? Mobile?
I experienced Chrome working most of the time.

Hello, thanks for the response.
I just enabled a firewall my client machine and it still seemed to work
I know that all clients are on Chrome.

Thank you!

A little remark: when blocking port 10000 to test turn connectivity, you must either test with 3 clients or disable p2p if you test with 2 clients. Your turn server can be just not working and a 2 persons meeting work flawlessly because p2p don’t use jvb anyway.

withdrew my previous post. my issue was in my coturn config - I added

cert=/etc/letsencrypt/live/turn.xxx.com/fullchain.pem
pkey=/etc/letsencrypt/live/turn.xxx.com/privkey.pem

but hadn’t removed the lines for the other certs, and I guess those were taking precedent.

However, I’m still getting reports of this behavior from some clients, and I don’t have a local repro. I’ll do my best to collect more information from them over the next few days

Okay I’ve got a bit more data.
Got on a video call with a client and set up a video chat with just him and me. So, a p2p call. It worked for a minute or two at first, but before long I got the “something went wrong; try again after 30 seconds” error.
I have client-side logs this time - the ones that look interesting are:

Strophe: request id 20.1 Error 0 happened

and

POST jitsi.xxx/http-bind?room=xxx net::ERR_NETWORK_CHANGED

so it looks like the network is changing underneath the chat, causing an interruption. However, we were on a Google Meet during that time, and it did not get interrupted. Further, this issue reliably reproduced within a couple of minutes, but we went on to meet.jit.si next, and the issue did not reproduce there.

Client was on a Chromebook (and I think this only occurs on Chromebook?), using Google Chrome, on a school network - likely behind a firewall.

Any ideas what could be causing this? I’m pretty stumped, but clearly meet.jit.si has solved this issue.

@damencho @emrah Would you have any ideas? Sorry to bother, and thanks