Pivotal Knowledge Base

Follow

Pivotal Cloud Foundry® RabbitMQ node not starting

Environment

Product Version
Pivotal Cloud Foundry® RabbitMQ 1.5.x, 1.6.x

Symptom

RabbitMQ node within the cluster is not running.

Cause

RabbitMQ process has crashed for some reason but Erlang VM may still be running.

Resolution

The first step is to check if the rabbitmq-server process is still running on the affected node, to do this run the following monit [3] command:

monit summary

This will report if the Erlang process is still running as far as monit is concerned.

To test if the RabbitMQ process is running, you can run one of the following commands to see if it reports back the current status of the cluster. 

rabbitmqctl report
rabbitmqctl status

See "RabbitMQ CLI" section [1

 

If you cannot connect to RabbitMQ via the CLI then you should first try and restart the rabbitmq-server process using monit.

1. monit stop rabbitmq-server

2. Ensure the process has stopped using monit summary

3. monit start rabbitmq-server

For RabbitMQ for PCF, it is recommended that you do not start/stop RabbitMQ using rabbitmqctl , along with running others commands as outlined in the clustering plugin [2] as this is used by the service to control the cluster. 

[1] http://docs.pivotal.io/rabbitmq-cf/managing.html

[2] https://github.com/rabbitmq/rabbitmq-clusterer

[3] https://mmonit.com/monit/documentation/monit.html#Arguments

Comments

Powered by Zendesk