I'm trying to setup a RabbitMQ cluster, the purpose is to have working all RabbitMQ queues even when one node is down. I use 3 nodes as the VirtulBox machines, all are running on Ubuntu Server 24.04. RabbitMQ 3.12.1

Nodes have the following IPs and hostnames that are set up in /etc/hosts on all nodes:

192.168.88.11 node1
192.168.88.12 node2
192.168.88.13 node3

On the node1 I added a host like this:

sudo rabbitmqctl add_vhost myhost --default-queue-type quorum

And I can see the queue type = quorum in the web management interface.

On node2 and node3 I executed this command:

sudo rabbitmqctl join_cluster rabbit@node1

and then the cluster_status command returns the following:

Running Nodes

rabbit@node1
rabbit@node2
rabbit@node3

So it seems everything is working. I could connect to the cluster by IP/host of any node and it works.

But the problem appears when any node becomes down (I just disable the network on the virtual machine as Bridged -> Not attached). And all remaining nodes hang. The cluster_status command hangs on the

Cluster status of node rabbit@node3 ...

line. Connections from clients are not being accepted (timeout error). The web management interface also does not respond.

So how to fix this? I need the cluster continue working when one node becomes down, I don't say about 2 nodes down, but at least with 1 unavailable node it should work. That's why I need a cluster.