CODEWITHSUNDEEP
cluster-computingactivemq-artemis
MF1
MF1

Reputation: 23

Apache Artemis do not synchronize retain message in cluster mode <message-load-balancing>STRICT</message-load-balancing>

I have an issue with an Artemis MQTT Cluster (two nodes). The cluster work well but only the retain message isn't synchronized between my two cluster nodes.

I looked a bit around and found this suggestion:

<redistribution-delay>0</redistribution-delay>

Unfortunately this does not work for me.

To understand and clarify this, I have looked in the Artemis documentation which states:

This option only will work for ON_DEMAND message-load-balancing.

However, I want to use STRICT mode.

Currently I am using apache-artemis-2.11.0 running on two Ubuntu 18 nodes. I use Java version 1.8.0_111 on both nodes.

I configured the artemis cluster and ha-policy in colocated mode:

<ha-policy>
   <replication>
      <colocated/>
   </replication>
</ha-policy>

At the cluster-connections section I use the <message-load-balancing>STRICT</message-load-balancing> to build an active-active cluster.

I am a bit confused because everything is working fine except the retain message sync. Do you have any suggestions what my problem could be?

Upvotes: 0

Views: 254

Answers (1)

Justin Bertram
Justin Bertram

Reputation: 34998

The documentation already indicates what your problem is. You're using STRICT as the message-load-balancing rather than ON_DEMAND. Message redistribution won't work if you use STRICT.

Also, it's worth noting that colocated configurations will be deprecated soon so I would recommend against using them.

Upvotes: 1

Related Questions