<(CM) Support tip> Render issues that affect HPE RM 8.3.0

[Ethan_Wang]

Incident description :

Issue No.1

The following unique constraint violation error from Render transaction is received from time to time in Workgroup server log :

Error : XX User: XXXXXX -  The SQL transaction 'AddToRenderQueue' failed for user XXXXXX:XXXX: ORA-00001: unique constraint (XXXXXX.RNQRECURI) violated

Details: Extended SQL Error Information:

                Number                               = 0x80040e2f

                Source                  = OraOLEDB

                Description         = ORA-00001: unique constraint (XXXXXX.RNQRECURI) violated

                SQLState              =

                NativeError         = 1

SQL command: INSERT INTO TSRENDQUEU ( uri , rnqRecUri , rnqRenditionType , rnqPriority , rnqRequestDate , rnqRequestorUri , rnqProcessMachine ) VALUES ( ? , ? , ? , ? , ? , ? , ? )

Issue No.2

Failed jobs are automatically added back to Render queue for reprocessing.

Cause                       :  

These issues occur when the Select Type of Finalized is used in Render.

When Select Type of Finalized is used, render will be auto-polling render requests, and adding the request to the Render queue.

The root cause to the first issue is that when Render auto pooling interval is shorter than the timeframe for the queued jobs to be processed, and because the query used for polling fails to check the existence of the request in Render queue, it will attempt to add request for a record of which a request for the same type of rendition is already in the queue, which will trigger this error.

Root cause to the second issue is due to issue with the auto polling query which doesn’t filter out the failed queue jobs.

Resolution                :

These issues have been addressed in HPE CM 9.1, and HPE RM 8.3.0 Patch 4 Hotfix 1. The issues will also be addressed in a future patch for HPE RM 8.3.0.