Starting HiveMQ in cluster mode with persistent data

When a HiveMQ node is started in cluster mode it will automatically move the contents of its existing data folder to the cluster-backup folder, effectively starting up the node without any data.
This is done on purpose. The logic behind this is that a running cluster always has the more relevant and consistent data than a restarted node. Even if the data present on the node is only a few minutes old it is likely that the persisted data for MQTT clients has changed a lot.

During specific scenarios such as maintenance windows or disaster recovery it can be useful or necessary to start up a HiveMQ node in cluster mode with persistent data.

Instructions

HiveMQ provides a specific run script called recovery.sh that can be utilized for starting a HiveMQ node in cluster mode with persistent data

Startup via using the start up script

Execute the the recovery.sh file in the HIVEMQ_HOME/bin folder.

Start up HiveMQ as a service (init.d)

Open the HiveMQ init.d start script
```
vi /etc/init.d/hivemq
```

Change the startup to using recovery.sh instead of ~~run.sh~~ at line 19

...
        failure $"HiveMQ start"
        echo; else
        echo -n Starting HiveMQ server
        su $HIVEMQ_USER -c "$HIVEMQ_HOME/bin/recovery.sh >/dev/null 2>&1 &"
        ### Create the lock file ###
        touch /var/lock/subsys/hivemq
        success $"HiveMQ server started"
...

Start the HiveMQ Service
```
/etc/init.d/hivemq start
```
Make sure to revert the changes from step 2

Start up HiveMQ as a service (Debian based, using init.d)

Open the HiveMQ init.d start script
```
vi /etc/init.d/hivemq
```