Let's say the main RabbitMQ log file (such as /var/log/rabbitmq/rabbit@server001.log on Linux) contains the following.

=ERROR REPORT==== 14-Dec-2020::18:19:32 ===
Mnesia(rabbit@server001): ** ERROR ** mnesia_event got {inconsistent_database, running_partitioned_network, rabbit@server002}

The rabbitmqctl cluster_status command can be used to determine if the servers are missing from the cluster. In this example, server2 is not in running_nodes, meaning that server2 is not in the cluster.

~]# rabbitmqctl cluster_status
Cluster status of node rabbit@server1
[{nodes,[{disc,[rabbit@server1,rabbit@server2,rabbit@server3]}]},
 {running_nodes,[rabbit@server1]},
 {cluster_name,<<"rabbit@server1.example.com">>},
 {partitions,[]},
 {alarms,[{rabbit@server2,[]},{rabbit@server3,[]}]}]

The systemctl restart rabbitmq-server command can be used as a quick fix. In this scenario, issuing this command on server1 may temporarily resolve the issue.

There are 4 modes that can be used.

• ignore (this is the default mode)
• autoheal
• pause-minority
• pause-if-all-down

These modes can be set in the main RabbitMQ configuration file, like this.

[
  {rabbit, [
      {cluster_partition_handling, autoheal}
  ]},