Need to reboot JVMs if there is database issue for small time
When Oracle database (Pega database) does not work for a small interval of time, the application JVMs must be rebooted.
[ttp-nio-8070-exec-17] [STANDARD] [ ] [CM:01.01.02] (rnal.access.DBLockProviderImpl) ERROR xxx.your_webpage.com|127.0.0.1 x85xxx - Problem executing lock check
com.pega.pegarules.pub.database.ConnectionException: Problem executing lock check: code: 17410 SQLState: 08000 Message: No more data to read from socket
DatabaseException caused by prior exception: java.sql.SQLRecoverableException: No more data to read from socket
| SQL Code: 17410 | SQL State: 08000
Steps to Reproduce
Perform any action without restarting the application sever when the database stops working.
Connection validation is required at the application server level when the database server is unable to handle the failover by itself.
Here’s the explanation for the reported behavior:
The database must be able to handle the failover or the application server must be able to use valid connections.
Since Oracle RAC is already being used, use the information in 'Design and Deploy Tomcat Applications for Planned, Unplanned Database Downtimes and Runtime Load Balancing with UCP In Oracle Database RAC and Active Data Guard environments' Oracle White Paper for configurations that must be made when using Tomcat and Oracle RAC to support high availability.
This article discusses about using Universal Connection Pool (UCP) and fastConnectionFailoverEnabled (FCF), having the correct database driver, and other required JAR files to support high availability.
Validate the connection in the application server tier when the database does not provide support for high availability or failover. However, this is additional overhead on the application server.
Published August 29, 2018 - Updated October 8, 2020
Was this useful?
0% found this useful
Have a question? Get answers now.
Visit the Collaboration Center to ask questions, engage in discussions, share ideas, and help others.