We're eagerly awaiting the release of LavinMQ 2.0, now in v2.0.0-rc.4. Once it's released, it will be available on CloudAMQP! This new version will bring advanced clustering features to enhance performance and reliability. Here’s a quick description of how LavinMQ clustering will be set up at CloudAMQP.
CloudAMQP will support different cluster configurations, including 1-node, 3-node, and 5-node setups. Each option is designed to offer different benefits:
- 1-node setup: Provides the highest performance with minimal latency and maximum throughput. This configuration delivers top performance but does not offer fault tolerance on its own. However, LavinMQ single nodes typically restart quickly, even with millions of messages in the queues.
- 3-node and 5-node setups: These configurations enhance fault tolerance and reliability. They are well-suited for production environments where high availability is the most important. These setups ensure that LavinMQ 2.0 remains operational even if one or more nodes goes down.
Minimal downtime during plan changes
LavinMQ 2.0 is built to handle plan changes and interruptions with minimal downtime. Switching roles from a follower to a leader or upgrading your plan may cause a brief delay. When a new leader is elected, the LavinMQ server needs to restart as it transitions from follower mode to leader mode. If there are thousands of queues, millions of messages, or a lot of data, this transition might take a bit longer.
However, adding more nodes to your cluster won’t cause downtime; your service will remain uninterrupted.
LavinMQ and etcd on CloudAMQP
LavinMQ relies on etcd for leader election and cluster management. At CloudAMQP, etcd is deployed on the same number of nodes as the number of LavinMQ servers. Even single-node clusters use etcd, making it simple to scale up from a single node to a multi-node setup as your needs changes.
LavinMQ is fully compatible with all Unix-based systems and runs on Ubuntu servers at CloudAMQP. It features automatic failover in all clusters, ensuring quick recovery if a server crashes. For those familiar with RabbitMQ clustering at CloudAMQP, transitioning to LavinMQ clustering is straightforward, as it offers a similar messaging experience.
Now, I will leave you to discover LavinMQ release candidate v2.0.0-rc.4 Happy to talk if you have any questions!
