CBR_TS_NO_INDEX_FOUND_FOR_INDEXATION_ID when re-indexing content

This has recently been a common problem as more and more customers upgrade P8 to current release(s).  Existing content stored in the FileNet Content Engine objectstore does not index when a P8 Administrator creates an index job. Newly added content to the same objectstore does index for full text search without any errors.  When re-indexing content already stored in a P8 objectstore using an index job initiated using FileNet Enterprise Manager (FEM), or Administration Console Content Engine (ACCE), one of two things typically occurs;

1. A row is never inserted into the IndexRequests table and thus nothing is even attempted to index
2. A row is inserted into the IndexRequests table but it is in an indefinite pending state

A typical exception found in either the p8_server_error.log or p8_server_trace.log is;

ERROR ObjectStore:"Objectstore name here"; IndexJob index class attempt failed; attempt #12\ncom.filenet.api.exception.EngineRuntimeException: FNRCB0044E: CBR_TS_NO_INDEX_FOUND_FOR_INDEXATION_ID: The IBM Content Search Services index is missing for objects with indexation ID {GUID of non existant text collection here}.\n at com.filenet.engine.cbr.spi.CSEIndexAreaCache.getTextSearchIndex(CSEIndexAreaCache.java:342)\n at com.filenet.engine.cbr.spi.CSEIndexRequester.doSubmitRequest(CSEIndexRequester.java:271)\n at com.filenet.engine.cbr.CBREngineIndexer.submitRequest(CBREngineIndexer.java:111)\n at com.filenet.engine.cbr.CBRIndexer.submitRequest(CBRIndexer.java:2923)\n at com.filenet.engine.cbr.CBRIndexer.processSingleItemToIndex(CBRIndexer.java:3272)\n at com.filenet.engine.cbr.CBRIndexer.processJob(CBRIndexer.java:757)\n at com.filenet.engine.cbr.CBRIndexer.unitOfWork(CBRIndexer.java:578)\n at com.filenet.engine.tasks.ScheduledBackgroundTask.execute(ScheduledBackgroundTask.java:252)\n at com.filenet.engine.tasks.BackgroundTask.safeExecute(BackgroundTask.java:244)\n at com.filenet.engine.tasks.ScheduledObjectStoreTask.safeExecute(ScheduledObjectStoreTask.java:51)\n at com.filenet.engine.tasks.BackgroundTask$BackgroundTaskPriviledgedExceptionAction.run

This can occur if the P8 environment previously used Legacy Content Search Engine (LCSE) and when upgrading to either Content Engine version 5.1 or version 5.2.x with IBM Content Search Services (CSS) the indexed text collections were not migrated to IBM Content Search Services.  It can also occur if documents from an object store previously configured with Legacy Content Search Engine (LCSE) were imported into Content Engine 5.1 or 5.2.x still containing the old invalid indexation_id values from an LCSE configuration.

Query the indexation_id column in the DocVersion table, specifically for the GUID reported in the p8_server_error.log or p8_server_trace.log. The GUID is part of the overall error as reported above (The IBM Content Search Services index is missing for objects with indexation ID {GUID of non existant text collection here}). There are a certain amount of rows associated with this GUID, and the indexation_id value for these rows has to be updated to show null in order to resolve this.

How to resolve this?

Using FileNet Enterprise Manager (FEM) or Administration Console Content Engine (ACCE), delete any index-job(s) in that are currently pending indefinitely.

1. Query the objectstore DocVersion table for the number of rows affected by the GUID no longer associated with any text collections;

   select count(*) from DocVersion where indexation_id = 'GUID from P8 log here'

2. Disable CBR Dispatcher in FileNet Enterprise Manager (FEM) or Administration Console Content Engine (ACCE).

3. Take a database backup of at minimum the DocVersion table, but preferably the entire object store database, especially if this is a production server.

4. Using a database query utility, execute an update statement that will nullify the indexation_id value from the rows returned in step 1.  Never remove the rows themselves, only the value so that the indexation_id column for these rows show as null.

5. Re-enable CBR Dispatcher in FileNet Enterprise Manager (FEM) or Administration Console Content Engine (ACCE).

6. Re-run the index-job previously canceled in step 1. It should now terminate normally.