Configure RabbitMQ Clusters for Partition Recovery

When two or more nodes in a RabbitMQ cluster lose contact for a period of time, network partitioning can occur that prevents the cluster from functioning as designed. Configure RabbitMQ clusters for partition recovery to prevent this scenario.

You may skip this task if there is only one RabbitMQ server in the customer's environment.

  1. Open a text editor.
  2. In the blank text file, enter only one of the following lines of text depending on the number of RabbitMQ nodes in your system:
    • If you have 2 RabbitMQ nodes in your system, enter:

      cluster_partition_handling = autoheal

      OR

    • If you have 3 or more RabbitMQ nodes in your system, enter:

      cluster_partition_handling = pause_minority

  3. Save the file as rabbitmq.conf in the directory [Drive]:\RabbitMQ.
  4. Open an administrative command prompt and change to the RabbitMQ sbin directory similar to the following example (the path to the sbin folder may vary depending on your configuration):

    C:\Program Files (x86)\RabbitMQ Server\rabbitmq_server-<version number>\sbin\

  5. Run the following commands:

    1. Rabbitmq-service.bat stop

    2. Rabbitmq-service.bat remove

    3. Rabbitmq-service.bat install

    4. Rabbitmq-service.bat start

  6. Repeat this task on all RabbitMQ nodes in the cluster.

 

<< Previous Task | Configure RabbitMQ Clusters