Following a Database Server migration some Reporter extractions on some particular nodes fail.
Reporter restart does not fix the issue.
Following kind of errors appear in server.log:
[DATE] Process ABORTED. Status changed to ABORTED. Reason: java.util.concurrent.ExecutionException: java.sql.SQLException: ORA-00600: internal error code, arguments: [ktbdchk1: bad dscn], [], [], [], [], [], [], [], [], [], [], []ORA-06512: at "REPOWNER.INSERT_OR_UPDATE_RELAUNCH", line 81ORA-06512: at line 1
and
Caused by: java.util.concurrent.ExecutionException: java.sql.SQLRecoverableException: Closed Statement
Release : 6.9
Component : DOLLAR UNIVERSE
Issue due to a Corruption of Reporter Table's Indexes.
Solution consists on recreating the Corrupted Indexes in Oracle.
Additionally, implement the following changes to improve Reporter estability and performances:
1. In case Java 8 is being used, upgrade the ODBC Thin driver from ojdb6.jar to use either ojdbc8.jar or ojdbc7.jar corresponding to the Oracle Database version in uxsetenv
2. Increase Reporter memory allocation as it needs to be modified after each Reporter upgrade in file unistartrep.
Please replace 1024 by 4096
3. Modify the extraction settings to boost the Extractions Performances.
To do so, first try setting to false the extraction of the Labels, Audit Trail and Interventions if not used in Reports.
4. Remove from Extractions the Stopped nodes.