To test the highly available Gateway group scenario, you configure, then run the installation verification test (IVT) supplied with CICS® Transaction Gateway.
Install the IVT application ECIIVT.ear. For more information see JCA resource adapter installation verification test (IVT).
Configure the IVT application to use the highly available Gateway group GroupA, by specifying the resource reference.
Select WebSphere enterprise applications from the Applications > Application types section of the navigation menu. From the Enterprise Applications dialog, Preferences section, click ECIIVT from the table of applications.
From the Enterprise Applications configuration panel for ECIIVT, References section, click Resource references.
The Resource references dialog shown in Figure 1 allows you to specify the name of the JNDI (Java™ Naming Directory Interface) to be used by the application. Click Browse to display the Available resources dialog shown in Figure 2.
Select ECI-XA-GROUPA resource and click Apply. The admin console returns you to the Resource references dialog. The Target Resource JNDI Name now has the value eis/CICSGroupA shown in Figure 3.
You have now configured the ECIIVT application to use the highly available Gateway group GroupA. Save the changes to the master configuration.
Start the IVT application in WebSphere® Application Server and start the gateway daemons CTGA1 and CTGA2.
Run the ECI IVT several times to send ECI requests to CICS, using different combinations of available Gateway daemons and CICS servers each time. If at least one of the CTGA1, CTGA2 Gateway daemons and one of the CICSA1, CICSA2 servers are available, the request will succeed.
For information on how to run the IVT see JCA resource adapter installation verification test (IVT).
If an error message indicates that the connection to CICS Transaction Gateway failed, this is usually because a Gateway daemon has been shut down. The managed connections in the connection factory pool are reused but if a managed connection was connected to a Gateway daemon that is no longer running, it attempts to reuse that Gateway daemon, resulting in the error.
The solution is to ensure that the purge policy for the connection factory pool is set so that the entire connection pool is purged not just the single connection. Reset the purge policy then run the IVT again. When you have done this, if a managed connection cannot connect, all managed connections to that Gateway daemon are deallocated and the IVT should run successfully.
This error is related to the fact that high availability uses TCP/IP port sharing. The error can occur if you shut down a Gateway daemon, but should not occur if you shut down a CICS server.