Cannot fetch a row from OLE DB provider during cube processing job

Products

• IT Analytics

Issue/Introduction

An IT Analytics (ITA) cube processing job fails after running for several hours and the following error is logged in the SQL Server Agent log:

Error: OLE DB or ODBC error: Cannot fetch a row from OLE DB Provider.

Environment

Release : 2.9.1

Component : Cubes

Cause

Possible causes included insufficient permissions to read cube data sources, an external connection timeout on the Analysis Server, or a remote connection timeout on the SQL database server. Other causes include processing job scheduling conflicts, a sub-optimal topology, and inadequate resources for the size of the environment.

Resolution

Ensure the Remote query timeout setting on the SQL server is set to 0 (disabled). The ExternalCommandTimeout, ExternalConnectionTimeout, and ServerTimeout settings on the SQL Server Analysis Services (SSAS) server should be set to 360000.

Per Appendix D (Oracle Database Client Provider) of the IT Analytics Data Loss Prevention Server Administrator Guide, ensure the OraOLEDB.Oracle provider option Allow inprocess is enabled, and set the OLE DB driver's Chunk and Fetch sizes in the registry to 200 and 500, respectively. After enabling Allow inprocess, restart the SQL Server service (MSSQL). Restart Windows if the Chunk or Fetch sizes are modified.

Evaluate whether cube processing jobs are scheduled to run too close together, resulting in resource conflicts. Consider consolidating cube processing to reduce the number of times individual dimensions are reprocessed (e.g., create a processing job that includes all Endpoint cubes), and process cubes only as often as reports and views are expected to update.

Finally, if all settings are configured as prescribed and scheduling is optimized, evaluate whether the processor, memory, and storage resources allocated to the ITA environment - and the environment topology - are sufficient for the volume of incidents being processed. See the attached white paper Sizing and Scalability Recommendations for IT Analytics Data Loss Prevention Pack.