| From: | Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com> |
|---|---|
| To: | Phoenix Kiula <phoenix(dot)kiula(at)gmail(dot)com> |
| Cc: | pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: REINDEX takes half a day (and still not complete!) |
| Date: | 2011-03-19 05:30:15 |
| Message-ID: | AANLkTikfa2dFTc1QoV2b6=aggiDUrT8OznRaqc8RR3T5@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
On Fri, Mar 18, 2011 at 9:07 PM, Phoenix Kiula <phoenix(dot)kiula(at)gmail(dot)com> wrote:
> autovacuum = on
> autovacuum_vacuum_cost_delay = 20
>
> vacuum_cost_delay = 20
> vacuum_cost_limit = 600
>
> autovacuum_naptime = 10
also, if vacuum can't keep up you can increase the vacuum cost limit,
and lower the cost delay. Anything above 1ms is still quite a wait
compared to 0. And most systems don't have the real granularity to go
that low anyway, so 5ms is about as low as you can go and get a change
before 0. Also, if you've got a lot of large relations you might need
to increase the max workers as well.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | phb07 | 2011-03-19 09:00:29 | Performance on AIX |
| Previous Message | Scott Marlowe | 2011-03-19 04:58:17 | Re: REINDEX takes half a day (and still not complete!) |