How to set up Tomcat active-active with PCS

An active-active Tomcat setup keeps multiple application instances available so incoming traffic can be balanced across nodes and a single host failure does not become an outage.

In a Pacemaker / Corosync cluster managed with pcs, a systemd unit can be registered as a cluster resource and monitored for health. Cloning that resource runs an instance on each eligible node, while the cluster tracks failures and restarts the unit when monitoring detects a problem.

Cluster control changes the normal service lifecycle, so the underlying Tomcat unit should remain disabled in systemd and only started through the cluster. Active-active does not replicate deployments or runtime state, so application files and configuration must be identical on every node and session state or uploads should use a shared backend or load balancer stickiness. Production clusters typically require fencing (STONITH) and a load balancer with health checks, since pcs does not distribute traffic.

1. Confirm the cluster is online with quorum.

   $ sudo pcs status
   Cluster name: clustername
   Cluster Summary:
     * Stack: corosync (Pacemaker is running)
     * Current DC: node-01 (version 2.1.6-6fdc9deea29) - partition with quorum
     * 3 nodes configured
     * 0 resource instances configured
   ##### snipped #####

   Related: How to check Pacemaker cluster status

2. Identify the Tomcat systemd service unit name.

   $ systemctl list-unit-files --type=service | grep -E '^tomcat.*\.service'
   tomcat10.service                             disabled        enabled

   Strip the trailing .service when using the systemd: resource type.

   Related: How to manage Pacemaker cluster services with systemctl in Linux

3. Disable the Tomcat unit outside the cluster so Pacemaker controls start and stop actions.

   $ sudo systemctl disable --now tomcat10.service

   Leaving Tomcat enabled in systemd can start it automatically at boot even when the cluster intends it to be stopped, causing conflicting state and confusing resource status.

   Related: How to manage Pacemaker cluster services with systemctl in Linux

4. Create the Tomcat service resource.

   $ sudo pcs resource create tomcat_service systemd:tomcat10 op monitor interval=30s

   Use systemd:tomcat10 when that unit is present.

   Related: How to create a Pacemaker resource

5. Clone the Tomcat service resource across nodes.

   $ sudo pcs resource clone tomcat_service clone-max=2 clone-node-max=1

   Omit clone-max to start an instance on every cluster node by default.

   Related: How to create a cloned resource in Pacemaker
   Related: How to set Pacemaker clone resource options

6. Verify the cloned resource status.

   $ sudo pcs status resources
     * Clone Set: tomcat_service-clone [tomcat_service]:
       * Started: [ node-01 node-02 ]

7. Confirm Tomcat is listening on the expected connector port on each node.

   $ sudo ss --listening --numeric --tcp --processes 'sport = :8080'
   State  Recv-Q Send-Q Local Address:Port Peer Address:PortProcess
   LISTEN 0      100                *:8080            *:*    users:(("java",pid=220191,fd=38))

   Replace 8080 with the port configured by the Connector in server.xml.

8. Configure the load balancer or DNS to distribute traffic across active nodes.

   Session-dependent applications require sticky sessions or an external session store to avoid user logouts when requests move between nodes.

   Related: How to plan traffic distribution for active-active PCS services

9. Run a failover test after distribution is updated.

   Related: How to run a Pacemaker failover test with PCS