Troubleshooting Slow Datastore Transfer in EIS
If transfer to Datastore is running very slowly (generally speaking transfer time for a scenario shouldn’t exceed an hour), look at the following possible causes:
- Have statistics been gathered on the MIPO schema recently?
With versions of MIPO >6.8, full schema stats are gathered in the post-processing step (after a scenario is loaded, for example) so this shouldn’t be an issue for MIPO. Statistics are gathered on the Datastore schema after a transfer is successful.
Prior to version 6.8 statistics were gathered with the “stale” option, meaning there had to be a noticeable change in the size of a table before Oracle would actually gather stats on that table.
- Are any indexes missing in the MIPO schema?
In one case, it was found that there were two indexes missing from the MIPO schema that caused the query behind the STOCKING_POINT transfer to run much longer than they should (>40 hours).
The indexes that were missing were: STOCKING_REPLENISH_PLAN_IDX1 and STOCKING_POINT_OUTPUT_IDX2.