Creating cluster member removes some configuration

Question

I'm using WAS ND and want to have dmgr profile with federated managed profile app.

I am creating cluster using:

AdminTask.createCluster('[-clusterConfig [-clusterName %s -preferLocal true]]' % nameOfModulesCluster)

Next, I'm configuring my WAS instance, queues, datasources, jdbc, JMS Activation Specs, factories etc.

By the time I want to create cluster member, I'm displaying:

print("QUEUES: \n" + AdminTask.listSIBJMSQueues(AdminConfig.getid('/ServerCluster:ModulesCluster/')))
print("JMS AS: \n" + AdminTask.listSIBJMSActivationSpecs(AdminConfig.getid('/ServerCluster:ModulesCluster/')))

And it returns all queues I've created earlier. But when I'm calling

AdminTask.createClusterMember('[-clusterName %(cluster)s -memberConfig [-memberNode %(node)s -memberName %(server)s -memberWeight 2 -genUniquePorts true -replicatorEntry false] -firstMember [-templateName default -nodeGroup DefaultNodeGroup -coreGroup DefaultCoreGroup -resourcesScope cluster]]' %    {'cluster': nameOfCluster, 'node': nameOfNode, 'server': nameOfServer})
AdminConfig.save()

configuration displayed earlier is... gone. Some configuration (like datasources) is still displayed in ibm/console, but queues and jms as are not. The same print is displaying nothing, but member is added to cluster.

I can't find any information using google. I've tried AdminNodeManagement.syncActiveNodes(), but it won't work since I'm using

/opt/IBM/WebSphere/AppServer/bin/wsadmin.sh -lang jython -conntype NONE -f global.py

and AdminControl is not available.

What should I do in order to keep my configuration created before clustering? Do I have to sync it somehow?

Answer 1

This is the default behavior and is due to the -resourcesScope attribute in the createClusterMember command. This attribute determines how the server resources are promoted in the cluster, while adding the first cluster member.

Valid options for resourcesScope are :

Cluster: moves the resources of the first cluster member to the cluster level. The resources of the first cluster member replace the resources of the cluster. (is the default option)

Server: maintains the server resources at the new cluster member level. The cluster resources remain unchanged.

Both: copies the resources of the cluster member (server) to the cluster level. The resources of the first cluster member replace the resources of the cluster. The same resources exist at both the cluster and cluster member scopes.

Since you have set "-resourcesScope cluster" in your createClusterMember command, all configuration created at cluster scope are being deleted/replaced by the empty configurations of the new cluster member.

So, for your configurations to work, set "-resourcesScope server", such that the cluster configurations are not replaced by the cluster member configurations.

AdminTask.createClusterMember('[-clusterName %(cluster)s -memberConfig [-memberNode %(node)s -memberName %(server)s -memberWeight 2 -genUniquePorts true -replicatorEntry false] -firstMember [-templateName default -nodeGroup DefaultNodeGroup -coreGroup DefaultCoreGroup -resourcesScope server]]' %    {'cluster': nameOfCluster, 'node': nameOfNode, 'server': nameOfServer})
AdminConfig.save()

Refer "Select how the server resources are promoted in the cluster" section in https://www.ibm.com/support/knowledgecenter/en/SSAW57_8.5.5/com.ibm.websphere.nd.doc/ae/urun_rwlm_cluster_create2_v61.html for more details.