Recently migration completion of Oracle DB to another machine, but before migrating, the DBA already cloned the DB so it has an exact same DB as the old one. The migration was successful, the data_engine can insert the data to DB and we can see the historical data in UMP. After migration of Oracle DB some strange behavior from data_engine. Sometimes it stopped insert data to DB and causing the data_engine queue to grow. However checked on the data engine log (log level 5), see no error there. Then if we found this error happened, we tried to restart the data_engine probe and it become normal again.
But after some times, the problem comes back again. have checked with DBA, and they found no error in DB when the problem happened. And actually we have custom probe that also insert data to the same DB, and that custom probe still insert data normally to the DB, while this problem happened, so it appears that the problem only happened in data_engine side.
Environment
CA UIM 8.51 and Oracle R11 as the DB
Resolution
The local Oracle DBA team set the "expire time" parameter which hasn't been set before , after this setting the data_engine has been stable.
Additional Information
Actually there was an error message:
Nov 22 10:41:39:379 [879164] de: data_engine [QoS] [QoSData] - Oracle_BulkInsert::commitDataNoMax [QoSData] data_engine [QoS] status: -1 OCI_ERROR - ORA-03114: not connected to ORACLE
Suggestion from CA to reduce the number of threads-- connecting processes from UIM end by using a smaller value for thread_count_insert.