On Windows, setup-one-machine-cluster hangs but works on Solaris; mqfailover requires Ctrl+C to cancel and then must be re-run. (ID 6195092) (Sun Java System Application Server Enterprise Edition 8.2 Release Notes)

Sun Java System Application Server Enterprise Edition 8.2 Release Notes

Previous: PointBase
Next: Documentation does not explicitly state that you need to create JMS resources before running the MQ Failover Sample Application following the asadmin deploy instructions. (ID 6198003)

On Windows, setup-one-machine-cluster hangs but works on Solaris; `mqfailover` requires `Ctrl+C` to cancel and then must be re-run. (ID 6195092)

Description

From install_dir\samples\ee-samples\failover\apps\mqfailover\docs\index.html, if you run the following commands:

Console 1

cd install_dir\samples\ee-samples asant start-mq-master-broker1

Console 2

cd install_dir\samples\ee-samples asant start-mq-cluster-broker1

Console 3

cd install_dir\samples\ee-samples asant start-mq-cluster-broker2

Console 4

cd install_dir\samples\ee-samples asadmin start-domain domain1

If you have already executed asant setup-one-machine-cluster-without-ha or asant setup-one-machine-cluster-with-ha for any other Enterprise Edition sample, then execute asant configure-mq otherwise execute asant setup-one-machine-cluster-and-configure-mq. In this case, the command appears to succeed:

start_nodeagent: [echo] Start the node agent cluster1-nodeagent 
[exec] Command start-node-agent executed successfully.

But then the system hangs indefinitely.

Solution

None at this time. This problem similarly affects all Enterprise Edition samples that use this ant target on Windows. A workaround is to Ctrl+C out of the hung process and then rerun it.