Massive Conference - crash while switching publisher

,

In your case calls yo OV server are executed and user does not recibe Stream or directly OV server ignores the request?

We have some similar issues reported in #459 but not sure It is the same issue.

Our topology has 2-3 publishers and 25 subscribers (only conectes) that becomes in publishers (publish calle after minutes conected) when raise hand process os accepted. In this proccess usually one of new publishers cannot be henares/viewed by other publishers or subscribers.

I don’t think that my problem is the same as #459, but yes i have the same senario as you, with 2-4 publishers and up to 40 subsctibers that must raise hand to became publishers. And in this proccess one of new publishers has frozen media to couple publishers or subscribers, not all.

Hi, we just done first test in house with over 100 devices connected to same network different AP’s, server its running in Frankfut (AWS), to make everything work we change resolution to 120p. I seams everything work fine. Switching between publisher went smoothly without delay or crashing, (t3.xlarge) max 30% CPU usage.

We have some big issues with OV signaling server, not all streams events are received in all participants…

All should have 4 subscribers but one receives 4, another 3, another 2…

@dovalin can you share OpenVidu logs when some signaling problems appear? We need them to understand why messages are being lost.

After switching multiple publisher, a new participant publisher got this error when try to publish and other participants cant see and hear:

Complete Log here: https://drive.google.com/file/d/1QAh3jhe8y8eLU0N0JAjVGOCaQE7AG4ci/view?usp=sharing

@micael.gallego you have the logs and a new report above

I also receive “JsonRpcClient is disconnected from WebSocket server” error on my servers (they are all on Premises, VPS or Dedicated, CPU is enough, I constantly monitor CPU and other resources). I’m primarily using OV 2.19 and I just upgraded one of the servers to the latest version, which makes the problem worse.

Could anyone explain me the meaing of this error? Does this mean that KMS crashed? It it recoverable state or not?