I have just run a seqscan query on a fairly large table
(1.8Mrows, ca 40MB). I have a 256MB machine, so I expected
the postgres process to take up approximately that amount of
memory. On the other hand, observing top, I noticed postgres
was consistently below 4000KB. What does that mean? Isn't
postgres supposed to cache most of the database in memory,
so as to limit disk access?
Alex