In issue was found where refreshing a Workflow Monitor of a very large Workflow sporadically causes the JCP to generate a forced trace and to become unresponsive.
In such a situation you would find a line in the JCP log like follows:
20210305/105106.409 - 2217 U00045001 The database connection with session id '358' is no longer valid.
20210305/105106.411 - 2217 U00045014 Exception 'java.sql.SQLRecoverableException: "Closed Connection"' at 'oracle.jdbc.driver.PhysicalConnection.getAutoCommit():1828'.
20210305/105106.411 - 2217 UCUDB32 RBCK RET 0001 HSTMT: 0000000000000358 VALUE: 0000000000000000 ALL: 0.00100 DB: 0.00000 ODBC: 0.00000 UDB: 0.00000
20210305/105106.412 - 2217 U00045000 Database connection with session ID '358' was closed.
20210305/105106.413 - 2217 UCUDB32 CLST RET 0000 HSTMT: 0000000000000358 VALUE: 0000000000000000
20210305/105106.417 - 2217 U00003434 Server routine 'ErtEstimationResource$$Lambda$282/0x000000084065f440/rest-transaction' required '0' minutes and '21' seconds for processing.
20210305/105106.419 - 2217 EXIT ErtEstimationResource$$Lambda$282/0x000000084065f440 TIME: 21684ms
20210305/105106.421 - 2217 UCUDB32 RBCK RET 0001 HSTMT: 0000000000000358 VALUE: 0000000000000000 ALL: 0.00000 DB: 0.00000 ODBC: 0.00000 UDB: 0.00000
20210305/105106.422 - 2217 U00045014 Exception 'java.sql.SQLException: "Unable to rollback session 358 because the connection is already closed."' at 'com.automic.database.impl.DBConnectionImpl.rollback():241'.
20210305/105106.423 - 2217 U00003620 Routine 'com.automic.database.api.DBException' forces trace because of error.
Release : 12.3
Component : AUTOMATION ENGINE
The issue is caused by a 20 seconds timeout, which stops the ERT API call, if it takes longer than that. This causes the DB connection to be forcibly closed.
The issue has been confirmed to be a bug and will be fixed with the Automation.Engine component of
* Automation Engine V 21.0 - available
* Automation Engine V12.2.10 - available
* Automation Engine V12.3.7 - available