Hello:
We are investigating a highwater mark issue on one of our production databases.
Here is a quick
synopsis of the events:
1 Dataset stats: Capacity Entries Highwater Mark
600,000 371,263 371,263
2 A DBDELETE was performed removing ~100,000 records and placing them into a
flatfile.
We would anticipate that the highwater mark would remain unchanged.
Dataset stats: Capacity Entries Highwater Mark
600,000 271,263 371,263
3 A batch process was then executed which performed more DBDELETEs against
the dataset
removing an additional ~170,000 entries. At this point, we would also
anticipate that the
highwater mark would remain unchanged.
Dataset stats: Capacity Entries Highwater Mark
600,000 101,263 371,263
4 A final batch process was then executed, which performed DBPUTs of the
records removed
from Step 2 above, back into the dataset. Again, we would also anticipate
that the highwater
mark would remain, unchanged, however . . .
Dataset stats: Capacity Entries Highwater Mark
600,000 201,263 354,071
There was no execution of an Adager Repack or a DBGeneral Re-org or a complete
unload/load on
the dataset, but yet the highwater mark was lowered. What else could attribute
to this and would this
cause any performance degradation?
Thanks,
Stephen Bruno
Database Manager
Tufts Health Plan
Watertown, MA
888.880.8699 x2231
|