Re: Bad Query Plans on 10.3 vs 9.6

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: David Rowley <david(dot)rowley(at)2ndquadrant(dot)com>
Cc: Cory Tucker <cory(dot)tucker(at)gmail(dot)com>, "pgsql-general(at)postgresql(dot)org" <pgsql-general(at)postgresql(dot)org>
Subject: Re: Bad Query Plans on 10.3 vs 9.6
Date: 2018-03-29 13:47:01
Message-ID: 13699.1522331221@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

David Rowley <david(dot)rowley(at)2ndquadrant(dot)com> writes:
> On 29 March 2018 at 18:26, Cory Tucker <cory(dot)tucker(at)gmail(dot)com> wrote:
>> The plan on 9.6 v 10.3 are effectively identical except in 9.6 the planner
>> decides to use an index only scan on the primary key and in 10.3 it does a
>> sequential scan. The problem is the sequential scan is for a table of 75M
>> rows and 25 columns so its quiet a lot of pages it has to traverse.

> How certain are you that all the indexes match on each instance?

Another possibility is that 10.3 sees the index-only scan as too expensive
because it thinks most of the table isn't all-visible. Comparing
pg_class.relallvisible values might be informative.

regards, tom lane

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Cory Tucker 2018-03-29 14:21:06 Re: Bad Query Plans on 10.3 vs 9.6
Previous Message Alvar C.H. Freude 2018-03-29 10:50:25 Re: Question about buffers_alloc in pg_stat_bgwriter view for monitoring