|
|
Using the Trade 6 application, we split the TradeEJB EJB modules and TradeWeb Web modules across two clusters. After configuring the backup cluster support and core group bridge service, we ran this simple test to verify that the configuration was working:
| 1. | Start all of the Trade clusters in the primary and backup cell. |
| 2. | Access the Trade Web site from a server hosting the Web modules in the primary cell: |
http://wascell02_servername:9080/trade
| 3. | Login and buy some stock. |
| 4. | Stop the primary Trade EJB cluster from the Administration Console. |
| 5. | Open a new browser and access the Trade Web site from a server in the primary cell. |
| 6. | Login and buy some stock. This should still succeed. If it fails, review the troubleshooting section. |
Additional tests include killing processes that you have assigned to be core group bridges. As long as there is one core group bridge, communication should continue. You can also kill the primary server processes or take down the machines they are running on (as long as the cluster with Web modules stays up) instead of a graceful stop.
To test failback, restart the primary servers. Stop the backup cluster to confirm that work is not being routed to it. If failback does not occur, verify that the backup cluster's information points to the cluster in the primary cell and there is an available core group bridge or the Deployment Manager is up. Review the troubleshooting section.
