| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Herodotos Herodotou <hero(at)cs(dot)duke(dot)edu> |
| Cc: | Jeff Davis <pgsql(at)j-davis(dot)com>, Emmanuel Cecchet <manu(at)asterdata(dot)com>, Greg Stark <gsstark(at)mit(dot)edu>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>, Eric Friedman <Eric(dot)Friedman(at)asterdata(dot)com>, John Cieslewicz <John(dot)Cieslewicz(at)asterdata(dot)com>, Dheeraj Pandey <Dheeraj(dot)Pandey(at)asterdata(dot)com>, "nedyalko(at)cs(dot)duke(dot)edu" <nedyalko(at)cs(dot)duke(dot)edu> |
| Subject: | Re: Join optimization for inheritance tables |
| Date: | 2009-09-26 16:28:15 |
| Message-ID: | 22334.1253982495@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Herodotos Herodotou <hero(at)cs(dot)duke(dot)edu> writes:
> On Tue, Sep 22, 2009 at 8:06 PM, Jeff Davis <pgsql(at)j-davis(dot)com> wrote:
>> I think you mean that the planning time is in milliseconds, not seconds.
> The planning time is actually in seconds.
This is exactly why I think this is a dead-end approach. Trying to do
theorem proving from an unstructured collection of constraints simply
cannot scale to hundreds of partitions, at least not if you want
reasonable planning performance. There are other aspects of our current
partitioning approach that don't scale either, eg the complexity of the
insert redirection triggers. We could handle standard cases where
there's a simple partitioning rule with far less overhead than this,
if we had an explicit model of the partitioning rule inside the system.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Dan Colish | 2009-09-26 18:28:59 | Re: Hot Standby on git |
| Previous Message | Tom Lane | 2009-09-26 16:19:47 | Re: recovery is stuck when children are not processing SIGQUIT from previous crash |