Recently while working with a customer on HACP Extended Edition, we encountered a situation where the customer’s pages were not working as expected in a cluster environment. If they closed the browser sessions, while having multiple HACP EE sessions active, the sessions were not getting disconnected. They could see a few LU’s still active on their mainframe. On further debugging the issue, we found that some of the session ID information sent in the close URL request had incorrect data, due to which sessions were not getting closed. Because of this connect and disconnect requests were going to the wrong network nodes. Finally, using the logs provided by the customer and using the information from WAS, we narrowed down the issue to some missing settings in IBM WebSphere Application Server in a cluster environment. Below are the steps that need to be followed while setting up HACP EE in WAS environment. Step1:- Select “Replication domains” from WAS admin console. Step 2:- Create a new replication domain by clicking on New. Give any name to the cluster e.g. “newcluster”, enable “Entire domain” and click Apply, Save and then OK. You can verify that your replication domain has got created Step 3:-Now navigate to WebSphere application servers. We had setup 3 members in the cluster. You can setup as many as needed. Step 4:- Click on one of the members e.g. here on mem1. Then click on Session management and then on Custom Properties. Add a new general property “HttpSessionCloneId”. Assign any value e.g. “test123”. Click Apply and then Save Step 5:- Click on the same member as in Step 4. Then click on Session management -> Distributed environment settings. On this page, select “Memory-to-memory replication”. Now click on “Memory-to-memory replication” to open its properties. Step 6:- Click on another member, e.g. here mem2. Then click on Session management. Click on Custom Properties. Add a new general property “HttpSessionCloneId”. Assign a value e.g. “test1234”. Click Apply then Save and Ok. Please note that “HttpSessionCloneId” should be assigned a unique value for each application cluster member node. Step 7:- Click on same member as in Step 6, Then click on Session management -> Distributed environment settings. On this page, select “Memory-to-memory replication”. Now click on “Memory-to-memory replication” to open its properties. In the “Replication domain” drop-down select the one that we newly created in Step 3. Set the replication mode to “Both client and the server”. Click apply, then Save and Ok. Repeat Step 6 and Step 7 for all members in your web application servers. Step 8:- Go to Web servers. Click on Stop, then start. Click on Generate Plug-in. Click on Propagate Plug-in. Step 9:- Now navigate to WebSphere enterprise applications and Install HACP EE following the regular install procedure of HACP EE. After installing HACP EE application go to Session management ->Custom properties. Add “SessionRewriteIdentifier” with value SESSIONID. Now under Enterprise Applications, select HACPEE_EAR. Then click on Session management -> Distributed environment settings. On this page, select “Memory-to-memory replication”. Now click on “Memory-to-memory replication” to open its properties. In the “Replication domain” drop-down select the one that we newly created in Step 3. Set the replication mode to “Both client and the server”. Click apply, then Save and Ok. After all these settings, restart the cluster set up. In our set up we restarted “newcluster” Once these settings are completed, users can access their HACP EE URL in the network environment, without facing any disconnect issues.
Kavya Bhat L3 Support Engineer – IBM Host On-Demand
0 Comments
Leave a Reply. |