Replies: 1 comment 2 replies
-
Consider using GitHub Discussions and not issues for questions. You haven't provided any details as to how RabbitMQ was deployed, or what version is used, or why you decided to use an configuration the docs explicitly recommend against. But I guess in this case these are less important in this case. With MQTT 3.1, which RabbitMQ implements, every client gets its own queue. So there won't be a round-robin dispatch demonstrated in tutorial 2. MQTT also does not have the concept of a single active consumer like AMQP 0-9-1, AMQP 1.0 (to an extent) and RabbitMQ Stream protocol. MQTT 5 has a concept of SAC called "shared subscriptions". Its support will be coming to RabbitMQ in the next 6 or so months. In the meantime, you can use one of the protocols mentioned above. And please do not use two-node clusters. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Hi Expert,
Very apreciated for your answer on my question.
I deloyed the rabbitmq cluster(2 nodes) and a consume service(3 pods) in k8s env on my own pc.
In my testing, the message could be sent to the two rabbitmq with round roubin, but all the 3 pod of the consumer received the messages.
By the way, i used the mqtt protocol by enabling rabbitmq_mqtt protocol. How to ensure only one pod received the message and keep the consume order, Otherwise, it will lead to repeated consumption. Do we have the concept of group like kafka. Look forward for your response. Follwong is my code, connect the mqtt with the k8s service name :
Beta Was this translation helpful? Give feedback.
All reactions