Rod Taylor <rbt@rbt.ca> writes:
> On Thu, 2003-03-06 at 11:49, Tom Lane wrote:
>> The implementation I've had in mind for autovacuum would rely on FSM not
>> the pg_stats daemon.
> FSM as I understand it gets it information from vacuum, and has an space
> marker removed when an insert uses the tuple.
FSM doesn't, but easily could, keep track of the number of space
requests it gets. We could also add some code to use it to keep count
of tuple deletions. Unlike the pg_stats daemon, FSM doesn't necessarily
try to track every single table --- but it would track all the
heavily-used ones, which is plenty for AVD purposes.
I tend to think of pg_stats as a bit of a failed experiment. The
overhead is excessive for what it accomplishes. Maybe we should rethink
the communication mechanism ... although I have no immediate ideas for
what else to do.
regards, tom lane