Upgrade Execution Hang During enable_rowkey.SQL – Resolution [Apps10 to 24R2]

Hello everyone,

• I wanted to share an issue we encountered during an upgrade from IFS Apps10 to IFS Cloud 24R2SU3, which might help others facing a similar challenge.

Issue Description:

The upgrade process was hanging during the execution of enable_rowkey.SQL. Additionally, the _deploy.log file showed no updates beyond a certain point, and the process appeared stalled for about 1 day.

Root Cause:

We discovered that the Container Database (CDB) had the parameter job_queue_processes set to 0. This setting effectively disables all background job processing in the entire multitenant database, including all pluggable databases (PDBs), causing the upgrade jobs to stall indefinitely.

Resolution Steps:

1. Connect to the CDB as the SYS user.

2. Set a non-zero value for job_queue_processes in the CDB. For example;

ALTER SYSTEM SET job_queue_processes = 10 SCOPE=BOTH;

3.) Monitor the _enable_rowkey.log file - the process should start progressing gradually.

4.) The upgrade then completed successfully without further issues.

Notes:

• If the job_queue_processes parameter is set to 0 in the CDB root (CDB$ROOT), it will effectively disable all DBMS_JOB and Oracle Scheduler jobs in the entire multitenant database, including all pluggable databases (PDBs), regardless of their individual job_queue_processes settings.

• job_queue_processes controls the maximum number of job worker processes for DBMS_JOB and Oracle Scheduler jobs in the database.

• In a multitenant environment:

  • The CDB root's job_queue_processes acts as a global cap.

  • If it is set to 0, no jobs can run anywhere in the CDB or any PDB, even if PDBs have non-zero values.

  • Oracle recommends setting the CDB root job_queue_processes to at least twice the number of open containers to avoid starvation of jobs across PDBs.

BR!

/Dilshan