Open SanjinOsmanbasic opened 4 years ago
Duplicate of #403
@Zakspeed - or is the #403 a duplicate of this one since it was created later? Why not #343? Or some other with the similar issue?
But I think that this issue gives a little bit more information on infrastructure that I am using ( mainly NGINX proxy for Ingress ).
And Im also looking for a universal fix that will allow me to spawn multiple Jitsi instances in an automated way on k8s, rather than just opening one port just for one instance.
@saghul - Can I ask you why did you close this? The problem is not solved.
https://github.com/jitsi/docker-jitsi-meet/pull/349 should have fixed it. Have you tried the images published yesterday?
Unfortunately it did not.
When I started setting up everything i noticed that namespace is missing on secrets and then i removed namespace completely from all YAML files.
The issue is still present unfortunately.
I'll reopen, but note that K8S support is self-serving.
I understand that - Im just hoping that someone with a similar setup/ or more knowledge will see the issue and get involved.
Thanks for reopening
@iGoogledYouYesterday have you tried this helm chart? https://github.com/jitsi/docker-jitsi-meet/tree/k8s-helm/k8s its working quite well for me. Still trying to get it working with jvb load balancing.
Hey all,
I have been trying to setup this awesome software on my Google cloud console using kubernetes files provided in the example directory.
The only speciality is that I'm using a nginx ingress controller for HTTPS and certificate generation. I have managed to create correct Ingress references and the HTTP traffic is now routed trough NGINX-ingress controller.
Im still using NodePorts to route the trafic for UDP to each node.
All was well - the web UI is working - until i try to connect a third person When the third person joins in i get a lot of errors in the console:
Logger.js:154 2020-04-07T14:13:30.663Z [JitsiConference.js] <e.sendMessage>: Failed to send E2E ping request or response. 2020-04-07T14:14:01.516Z [modules/RTC/BridgeChannel.js] <e.value>: Bridge Channel send: no opened channel.
I have tried adding PUBLIC_URL parameter to web, also tried changing the value of DOCKER_HOST_ADDRESS to my internal node address, then to my LoadBalancer IP address - and finally the IP that was displayed in the NodePorts section in Deployment.
None of it changed the situation - I may be confused what should be inserted as a DOCKER_HOST_ADDRESS.
Anyone has any ideas what should I try next?
If you need any more info - please post it - ill gladly provide it.