by six4withafro » Sat May 30, 2009 3:43 pm
i tried to do the maintenance through pgadmin, and encountered the same error at this point:
INFO: analyzing "public.game_players"
INFO: "game_players": scanned 3000 of 128881 pages, containing 103609 live rows and 0 dead rows; 3000 rows in sample, 4451077 estimated total rows
INFO: vacuuming "public.hand_histories"
ERROR: out of memory
DETAIL: Failed on request of size 531144000.
i clustered this db in the last 2 weeks, and was able to vacuum full/analyze then, so it's not a critical issue at this point, but it is awfully curious. there are only 4.5M hands in the db (as defined in the totals at the bottom of the page in the summary tab), and i know pt2 can handle more than that. perhaps i need to increase the maintenance work mem, or effective cache? my shared buffers is set at 1156MB and my effective cache is 1536MB, but the maintenance work mem is only 512MB. still, i've done this operation on larger dbs with less assigned maintenance mem in the past.