On 12/06/2011 09:00 PM, Tom Lane wrote:
> Mario Splivalo <mario.splivalo@megafon.hr> writes:
>> I have 8.4.8 on producion and 8.4.9 on test, could that explain the
>> difference in plans chosen?
>
> I'd wonder first if you have the same statistics settings on both.
> The big problem here is that the estimation of the join size is bad
> (8588 versus 0).
Just an update here. I did downgrade postgres on testbox to 8.4.8 and
now it's choosing bad plan there too.
So we upgraded postgres on production server and the bad plan went away.
We're preparing for upgrade to 9.1 now, we hope to offload some of the
SELECTs to the slave server, we'll see how that will work.
Thank you for your inputs!
Mario