Hans Kristian Rosbach wrote:
dbmail=# explain ANALYZE select * from dbmail_messageblks;
QUERY PLAN
------------------------------------------------------------------------------------------------------------------
Seq Scan on dbmail_messageblks (cost=0.00..0.00 rows=1 width=58) (actual
time=0.000..190.000 rows=2502 loops=1)
Total runtime: 280.000 ms
(2 rows)
That looks like a reasonable time. I'm assuming this is on the new HDD.
Not sure why it estimates 1 row, but ends up with 2502 tho.
The estimates usually are off by a few %, but not that much.
This could point to a bug in pgsql query optimizer, but I'm not too
familiar with the 8.0 version yet.
Usually this points to a database that has either never been analyzed or
hasn't been in a long time. Gerrit, have you run analyzed against your
databases? If not, please run an analyze on your entire dbmail database.
