What If One RabbitMQ VM Fails to Be Restarted When a Cluster RabbitMQ Instance Is Being Restarted?

Restarting a RabbitMQ instance will restart only the RabbitMQ processes instead of VMs on which the instance runs.

For a cluster RabbitMQ instance, if the RabbitMQ process fails to be restarted on one VM, the instance will be still in the Running state after the restart and a message will be displayed indicating that some nodes are faulty. A RabbitMQ daemon process runs on each VM and periodically checks whether the RabbitMQ process exists. If the RabbitMQ process does not exist, the RabbitMQ process will be automatically started.

If a RabbitMQ instance exception lasts for more than 1 minute, an alarm will be reported, and HUAWEI CLOUD technical support will take follow-up actions.