Chapter 6. Configuring JBoss EAP for Microsoft Azure

Starting JBoss EAP as a Service at Boot on Red Hat Enterprise Linux

If you are starting JBoss EAP as a service on Red Hat Enterprise Linux in Microsoft Azure, you must change the following configuration of the JBoss EAP service to ensure that it starts after the Azure virtual machine host name has been initialized.

Open /etc/systemd/system/ and add a dependency on waagent.service:

Replace the following line:

with: waagent.service

Load Balancing with mod_cluster

As a result of mod_cluster advertising not being supported in Azure, if you are configuring a JBoss EAP load balancing environment, you must ensure that all balancers and workers are bound to IP addresses that are accessible on your internal Microsoft Azure virtual network.


Balancers and workers must be bound to the actual internal Azure virtual IP address, for example, and not

Additionally, because of the unavailability of mod_cluster advertising, if you are using JBoss EAP as an Undertow mod_cluster proxy load balancer, you must configure each worker node to use a proxy list containing each balancer’s IP address and port, as shown below:

/socket-binding-group=standard-sockets/remote-destination-outbound-socket-binding=proxy-one:add(host="BALANCER_IP_ADDRESS", port="BALANCER_PORT")

ActiveMQ Artemis High Availability

JBoss EAP messaging high availability using Artemis shared stores is not supported in Microsoft Azure. To configure JBoss EAP messaging high availability in Azure, you must use a replicated journal with the AZURE_PING JGroups discovery protocol using a TCP JGroups stack. For more information on using a replicated journal, see Data Replication in Configuring Messaging for JBoss EAP.

After configuring the TCP JGroups stack with AZURE_PING, you must add the JGroups stack to the discovery-group and broadcast-group jgroups-cluster configurations in the messaging-activemq subsystem. For details, see Server Discovery in Configuring Messaging for JBoss EAP.