P.S. PRs with fixes or improves are welcome ;-)

About

A Jitsi Meet Chart for Kubernetes with 2 replicas of HAproxy in peering. All Jitsi shard are in backend via server-template in configuration.

This chart will install and configure: HAProxy statefulset, Jicofo deployment, JVB statefulset (scalable), web interface, prosody XMPP server.

Architecture

The Jitsi Kubernetes namespace has the following architecture:

The setup shown above contains only a single shard (for visual clarity). Subsequent shards would be attached to the web service. A more detailed explanation of the system architecture with multiple shards can be found in architecture/architecture.md.

Install chart

To download and to install the Jitsi Meet Chart, make sure that you have the Helm CLI (v2+) installed and clone this repository on your machine.

To install the chart, in your terminal, go to the jitsi-helm repository and run the following command:

helm install -n jitsi ./ --wait

If you want to use your own custom watermark, first create configmap from your png file:

kubectl create configmap -n <namespace> watermark --from-file=watermark.png

It assumes that you have a Nginx Ingress controller and you use CertManager along with ACME issuer type for managing the HTTPS certificates.

Because --wait flag, the status will be given once Jisti Meet is ready.

Updating the chart

To update the chart, in your terminal, go to the jitsi-helm repository and run the following command:

helm upgrade jitsi ./ --wait

Delete the chart

To delete the chart, in your terminal, go to the jitsi-helm repository and run the following command:

helm delete --purge jitsi

Configuration

The following table lists the configurable parameters of the Jitsi Meet chart and their default values.

Parameter	Description	Default
shardCount	Number of shards	2
namespace	Namespace	jitsi
metacontrollerNamespace	Namespace of the metacontroller. Empty if not deployed.	metacontroller
haproxy.name	Haproxy statefulset name	jitsi/jicofo
haproxy.image	Docker image	haproxy:2.1
haproxy.ingressEnable	Enable ingress	true
haproxy.ingress.host	Ingress host	jitsi.domain.org
haproxy.ingress.tlsEnable	Enable TLS for ingress	true
jicofo.name	Jicofo deployment name	jicofo
jicofo.image	Jicofo docker image	jitsi/jicofo
jicofo.imagePullPolicy	Jicofo image pull policy	Always
jicofo.extraEnvs	Jicofo extra environment variables	[]
jvb.name	JVB statefulset name	jitsi/jvb
jvb.image	JVB docker image	jitsi/jvb
jvb.image.imagePullPolicy	JVB image pull policy	Always
jvb.replicas	JVB replica count	1
jvb.monitoringEnable	JVB exporter container	true
jvb.nodeportPrefix	JVB Node port prefix	30
jvb.extraEnvs	JVB extra environment variables	[]
prosody.name	Prosody deployment name	jitsi/prosody
prosody.image	Prosody docker image	jitsi/prosody
prosody.image.imagePullPolicy	Prosody image pull policy	Always
prosody.extraEnvs	Extra env var for prosody deployment	[]
prosody.extraVolumes	Additionnal volumes to the prosody deployment	[]
prosody.extraVolumeMounts	Additional volume mounts to the prosody deployment	[]
prosody.globalModules	Additional global modules to enable on prosody	[]
prosody.globalConfig	Additional global config parameters on prosody	[]
web.name	Web deployment name	web
web.image	Web docker image	jitsi/web
web.image.imagePullPolicy	Web image pull policy	Always
web.extraEnvs	Extra env var for web deployment	[]
web.extraVolumes	Additionnal volumes to the web deployment	[]
web.extraVolumeMounts	Additional volume mounts to the web deployment	[]
watermark	Watermark logo	true

Running two jitsi instances inside the same cluster

1. The first instance should be deployed with the metacontroller
2. The second instance should be deployed with specific settings :

• metacontrollerNamespace should be kept empty to avoid deploying a second metacontroller
• jvb.nodeportPrefix should use a different value from 30 to avoid ports conflicts