Re: problems with large table

On 13/09/2007, Alvaro Herrera <alvherre@commandprompt.com> wrote:
> Mike Charnoky wrote:
>
> > Alvaro: The cluster suggestion probably won't help in my case since data
> > in the table should already be naturally ordered by date.
>
> It's not helpful only for reordering, but also for getting rid of dead
> tuples.


Apart from creating a new table, indexing it, then renaming it to
original table -- is there an alternative to CLUSTER that doesn't
impose a painful ACCESS EXCLUSIVE lock on the table? We are on
Postgres 8.2.3 and have a heavy duty table that starts showing its
limits after a week or so. Autovacuum is on and working. FSM etc is
fine, maintenance_work_mem is 256MB. But cluster still takes upwards
of 30 minutes, which is unacceptable downtime for our web service.
Thanks for any tips!