Re: Yet another abort-early plan disaster on 9.3

On 29 September 2014 22:54, Josh Berkus <josh(at)agliodbs(dot)com> wrote:
> On 09/26/2014 01:06 AM, Simon Riggs wrote:
>> On 23 September 2014 00:56, Josh Berkus <josh(at)agliodbs(dot)com> wrote:
>>
>>> We've hashed that out a bit, but frankly I think it's much more
>>> profitable to pursue fixing the actual problem than providing a
>>> workaround like "risk", such as:
>>>
>>> a) fixing n_distinct estimation
>>> b) estimating stacked quals using better math (i.e. not assuming total
>>> randomness)
>>> c) developing some kind of correlation stats
>>>
>>> Otherwise we would be just providing users with another knob there's no
>>> rational way to set.
>>
>> I believe this is a serious issue for PostgreSQL users and one that
>> needs to be addressed.
>>
>> n_distinct can be fixed manually, so that is less of an issue.
>
> It's an issue for the 99.8% of our users who don't know what n_distinct
> is, let alone how to calculate it. Also, changing it requires an
> exclusive lock on the table. Of course, you and I have been over this
> issue before.

In 9.4 you'll be able to set n_distinct using only a Share Update
Exclusive lock.

So that's no longer a problem.

The quality of the n_distinct itself is an issue, but with no current
solution, but then that is why you can set it manually,

--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services