You can use the ResUsageSps table to:
- Get a historical view of workload behavior for utilities and SQL operations.
- Determine the number of workload requests that are using AMP Worker Task. See the NumRequests column in TASM: Monitor WD Columns for details.
- Examine queue wait and service time numbers to find backed up queries and allocation groups.
- Determine which workload is responsible for I/O skew.
- Monitor CPU usage managed by the Priority Scheduler.
- Identify the percent of CPU being used by different workloads.
Teradata recommends that you use ResSpsView to access the data rather than accessing the ResUsageSps table directly. For more information, see Resource Usage Views.
If you are using the ResUsageSps table on SLES 10, this allows you to validate the relative weights given to workloads.
For a complete description of the SLES 10 Priority Scheduler and its components, see “Priority Scheduler (schmon)” chapter in Utilities.
For a complete description of the SLES 11 Priority Scheduler and its components, see Carrie Ballinger, Priority Scheduler for Linux SLES 11: Teradata Database 14.10, Teradata Database Orange Book 541-0008867.
|If you are running SLES 10 or earlier ...||If you are running SLES 11 or later ...|
|and using TASM, each WD is equivalent of one PGId.||and using TASM, each WD is equivalent of one Priority Scheduler workload definition ID (pWDid).|
|and table logging is enabled on ResUsageSps, a row is written to the database once for every triplet of Vproc Type (VprType), PG ID, and Performance Period ID (VprType, PGId, PPId) on each node in the system for each log interval.||and table logging is enabled on ResUsageSps, a row is written to the database once for every pWDid and VprType in the system for each log interval.|
For more information on the PGId and pWDid columns, see the RowIndex1 column in ResUsageSps Miscellaneous Housekeeping Columns.