Jean-David Beyer wrote:
[snip]
2007-11-14 12:00:31 EST DEBUG: analyzing "public.vl_in"
2007-11-14 12:00:31 EST DEBUG: "vl_in": scanned 2001 of 2001 pages,
containing 183983 live rows and 52 dead rows; 3000 rows in sample, 183983
estimated total rows
2007-11-14 12:00:31 EST DEBUG: analyzing "public.vl_cf"
2007-11-14 12:00:31 EST DEBUG: "vl_cf": scanned 1064 of 1064 pages,
containing 134952 live rows and 89 dead rows; 3000 rows in sample, 134952
estimated total rows
2007-11-14 12:00:31 EST DEBUG: analyzing "public.vl_as"
2007-11-14 12:00:31 EST DEBUG: "vl_as": scanned 1732 of 1732 pages,
containing 134952 live rows and 120 dead rows; 3000 rows in sample, 134952
estimated total rows
2007-11-14 12:00:31 EST DEBUG: analyzing "public.vl_ranks"
2007-11-14 12:00:31 EST DEBUG: "vl_ranks": scanned 1485 of 1485 pages,
containing 188415 live rows and 162 dead rows; 3000 rows in sample, 188415
estimated total rows
2007-11-14 12:00:31 EST DEBUG: analyzing "public.vl_mi"
2007-11-14 12:00:31 EST DEBUG: "vl_mi": scanned 1325 of 1325 pages,
containing 134952 live rows and 191 dead rows; 3000 rows in sample, 134952
estimated total rows
2007-11-14 12:00:31 EST DEBUG: analyzing "public.vl_li"
2007-11-14 12:00:31 EST DEBUG: "vl_li": scanned 1326 of 1326 pages,
containing 134952 live rows and 218 dead rows; 3000 rows in sample, 134952
estimated total rows
What does vacuum verbose have to say about this situation?
It is possible that analyze is not getting the number of dead rows right?
Does analyze, followed by vacuum verbose give the same dead row counts?
Sorry for lots of questions, I'm just throwing ideas into the mix.
Russell.
---------------------------(end of broadcast)---------------------------
TIP 7: You can help support the PostgreSQL project by donating at
http://www.postgresql.org/about/donate
