PRPC Caching DB connection string
During PRPC startup, it is caching DB connection sting which is blocking us to make the environment Hot-Hot or achieve HA. Our business wants the App to be up all the time but below are two scenarios are blocking us doing so. Please advise on how this can be addressed.
- The user has Oracle 11g RAC DB which has two DB nodes hosting App requests in parallel. For whatever reason, one of node goes down, the existing requests to that crashed node from a JVM are still trying to connect to the crashed node, until that DB connections timeout.
- Changes to DB URL in WebSphere require the user to restart the JVM or else it will not pickup latest value.
Steps to Reproduce
A third-party product issue
Here’s the explanation for the reported behavior:
Regarding first query :
Please make sure that all the configuration is as per the document below:
Make sure that connection cache name is unique. Because pega looks into data source URL for the connection and how it is handled for the RAC, the user must consult with Oracle.
Regarding the second query:
It seems to be ideal behaviour but it will most accurately answered by the Oracle , whether this is correct behaviour for RAC or we need some configuration changes.