Good morning,

We are facing an issue in the production environment. We wanted to know, how the pr_sys_statusnodes table is being populated?

Quick Background of issue:

Pega version: 7.1.8 CPM version: 7.1.3 pega call version: Proprietary information hidden

Websphere version: Proprietary information hidden

1. Production environment has three nodes[node1, node2 and node3] mapped with two clusters. Cluster1 is configured with node1,node2 which are for business users login, where load balancer is configured and cluster2 is configured with node3 which is for batch requestors [Agents/Listeners]
2. Suddenly, one of the node [node1/node2] goes down in the cluster1 with the following exception,

2016-05-31 20:40:07,221 [        SCDCA0000130] [  STANDARD] [                    ] [                    ] (   internal.mgmt.PREnvironment) ERROR   - com.hazelcast.spi.exception.TargetNotMemberException: Not Member! target:Address[10.130.1.xx4]:5701, partitionId: 0, operation: com.hazelcast.map.impl.operation.QueryPartitionOperation, service: hz:impl:mapService

Issue resolution approach:

1. We have raised a support request [ SR-A23918 -Error 503: PegaRULES server not available] for the above issue. Below is the solution suggested,
2. Create dynamic system setting or set <env name="cluster/hazelcast/interface" value="169.124.255.*" /> in prconfig.xml. we have created dynamic system setting as below,

DSS - cluster/hazelcast/interface

Value – 53.90.146.xx1-xx3

Even after we have set the dynamic system setting, we have observed that in the pr_sys_statusnodes table pyclusteraddress column is being populated with backup ip address instead of application server ip address. so, please help to make us understand from which configuration setting of websphere the pyclusteraddress column is getting populated or how the pr_sys_statusnodes table is getting populated.

Thank you.