G
George Potemkin
Guest
> Were you the only one who was connected to the database during the tests? What db log says? I'd like to recall my question. The data you sent me proves that new records were not added. Another possible explanation: AreDefrag.p program does not check all RPB slots in each data block. Otherwise it would run much slower. You scan a part of area just below HWM. The records recreated during defragmentation will be moved exactly here. Next run after defragmentation will read the same blocks but with new contents. And next run could see the records missed by its predecessor. Anyway the goal of AreaDefrag.p is just to /reduce/ the number of the fragmented records.
Continue reading...
Continue reading...