Quoting Doug McNaught <[EMAIL PROTECTED]>:
> "Chris Stokes" <[EMAIL PROTECTED]> writes:
>
> > >The REINDEX is needed because VACUUM doesn't free up index space in
> > >some circumstances. 7.4 (currently in late beta) will fix this.
> >
> > Sorry Doug,
> >
> > Yes I am doing a vacuum regularly
Dennis Gearon <[EMAIL PROTECTED]> writes:
> Is ist possible for a trigger to fork off a 'reindex' command and the
> trigger terminates leaving the reindex running?
Even if possible, ick.
I'd say the best thing to do would to have such a trigger set a flag
(value in a one-row table) and have a cr
Vivek Khera wrote:
"DM" == Doug McNaught <[EMAIL PROTECTED]> writes:
DM> The index bloat problem can occur when you have a an indexed SERIAL
DM> column whose value always increases, and you delete older rows on a
DM> regular basis. VACUUM recycles the row storage, but the index never
> "DM" == Doug McNaught <[EMAIL PROTECTED]> writes:
DM> The index bloat problem can occur when you have a an indexed SERIAL
DM> column whose value always increases, and you delete older rows on a
DM> regular basis. VACUUM recycles the row storage, but the index never
DM> shrinks.
I get this
"Chris Stokes" <[EMAIL PROTECTED]> writes:
> >The REINDEX is needed because VACUUM doesn't free up index space in
> >some circumstances. 7.4 (currently in late beta) will fix this.
>
> Sorry Doug,
>
> Yes I am doing a vacuum regularly - in fact, when the speed of the
> DB becomes slower, the va
"Chris Stokes" <[EMAIL PROTECTED]> writes:
> WE currently use PG for the local database on our POS workstations
> in shops. Data is then extracted and sent back to head office. Also
> product updates etc are regularly sent to the shops. I have noticed
> over time that the shop PG databases get slo
