Load Balancing between VideoBridges

i have setup one centralized xmpp server and several videobridges for my environment based on the info given by this video (https://jitsi.org/blog/tutorial-video-how-to-load-balance-jitsi-meet/). everything looks good except the load balancing between the videobridges. all conferences created using a particular videobridge and never assign to the others unless this one one goes down (then the 2nd videobridge can take over this role). i do not see big issue happened in jicofo, jvb or prosody log.

anyone can shed light on this? highly appreciated.

i can see these errors when starting the 2nd videobridge:

Apr 23, 2020 8:58:09 AM org.ice4j.ice.harvest.MappingCandidateHarvesters initialize
INFO: Initialized mapping harvesters (delay=511ms). stunDiscoveryFailed=false
Apr 23, 2020 8:58:09 AM org.xmpp.component.AbstractComponent handleIQError
INFO: (serving component ‘JitsiVideobridge’) IQ stanza of type error received:
Apr 23, 2020 8:58:09 AM net.java.sip.communicator.util.Logger warn
WARNING: Creating node failed with error. Continuing anyway.
Apr 23, 2020 8:58:09 AM net.java.sip.communicator.util.Logger info
INFO: PubSub node already exists (packetID=59K2n-7 nodeName=sharedStatsNode)
Apr 23, 2020 8:58:09 AM org.jitsi.utils.logging.LoggerImpl log
INFO: Using component socket: true
Apr 23, 2020 8:58:09 AM org.ice4j.ice.harvest.AbstractUdpListener
INFO: Initialized AbstractUdpListener with address 203.167.18.141:10000/udp. Receive buffer size 10485760 (asked for 10485760)
Apr 23, 2020 8:58:09 AM org.ice4j.ice.harvest.SinglePortUdpHarvester
INFO: Initialized SinglePortUdpHarvester with address 203.167.18.141:10000/udp
Apr 23, 2020 8:58:09 AM org.ice4j.ice.harvest.AbstractTcpListener addLocalAddresses
INFO: Not using link-local address /fe80:0:0:0:250:56ff:fe99:389%ens160 for TCP candidates.
Apr 23, 2020 8:58:09 AM org.jitsi.utils.logging.LoggerImpl log
INFO: Initialized TCP harvester on port 443, using SSLTCP:true
Apr 23, 2020 8:58:09 AM org.jitsi.utils.logging.LoggerImpl log
INFO: Performed a successful health check in 763ms. Sticky failure: false
Apr 23, 2020 8:58:09 AM org.xmpp.component.AbstractComponent handleIQError
INFO: (serving component ‘JitsiVideobridge’) IQ stanza of type error received:
Apr 23, 2020 8:58:09 AM net.java.sip.communicator.util.Logger warn
WARNING: Creating node failed with error. Continuing anyway.
Apr 23, 2020 8:58:09 AM net.java.sip.communicator.util.Logger info
INFO: PubSub node already exists (packetID=59K2n-10 nodeName=null)
Apr 23, 2020 8:58:09 AM net.java.sip.communicator.util.Logger error
SEVERE: Error received when publishing to the node: sharedStatsNode. Message: null. Condition: forbidden. For packet with id: 59K2n-10.

anybody can help on this?

Remove the pubsub config, it’s not needed

hi @Boris_Grozev

removed the config on all videobridges and jicofo?

hi @Boris_Grozev

if i remove the pubsub config from the 2nd VB and so on I’ll get the following then how can achieve the load balancing?

No pub-sub node mapped for xxx