On Tue, Sep 11, 2007 at 10:24:58AM -0600, Kevin Kempter wrote:
> I let the cron vacuum run for 14 days. in that 14 days the time it takes to
> vacuum the table grew from 1.2hours directly after the rebuild to > 8hours
> last nite.
Sounds to me like daily isn't enough, and that your FSM is too sm
Kevin Kempter wrote:
> It's difficult to try and vacuum this table during the day as it seems to
> begin blocking all the other queries against the database after some time.
Vacuum can generate so much I/O that it overwhelms all other
transactions, but it shouldn't block other queries otherwise
Hi List;
I've recently started cleaning up some postgres db's which previous to my
recent arrival had no DBA to care for them.
I quickly figured out that there were several tables which were grossly full
of dead space. One table in particular had 75G worth of dead pages (or the
equivelant in o
