SOA-P Cluster - Messages getting stuck in the ESB reply queues

Solution Verified - Updated -

Issue

After setting up a new SOA-P cluster environment with 2 nodes running on different machines, we start some high load scenarios to evaluate the capacity of our solution.

When running a single node no issues are found, but using 2 nodes alive one major issue happens: some of the messages delivered to the clustered ESB are getting stuck indefinitely on the reply queue for request-response services.

The clustered ESB services are called from a client component using the ServiceInvoker.deliverSync(..) method.

In the second case, a ResponseTimeoutException is retrieved to the client after reaching the timeout configured:

Exception: 
org.jboss.soa.esb.listeners.message.ResponseTimeoutException: No response received for service [sync:service], Told not to retry.
    at org.jboss.soa.esb.client.ServiceInvoker.post(ServiceInvoker.java:434)
    at org.jboss.soa.esb.client.ServiceInvoker.deliverSync(ServiceInvoker.java:221)

We think the consumers stop listening for some reason.

Environment

  • Red Hat JBoss SOA Platform
    • 5.3.1
  • JBoss Messaging

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase of over 48,000 articles and solutions.

Current Customers and Partners

Log in for full access

Log In