On Wed, Jun 9, 2010 at 12:04 AM, Greg Smith wrote:
> Mark Rostron wrote:
>>
>> - It was necessary to take a site outage and perform a “vacuum full
>> analyze” on the table
>>
>> - Following this, the query plan reverted to the more efficient btree
>> lookup
>>
>> Clearly, the garbage buildup resul
Mark Rostron wrote:
- It was necessary to take a site outage and perform a “vacuum full
analyze” on the table
- Following this, the query plan reverted to the more efficient btree
lookup
Clearly, the garbage buildup resulting from transaction activity on
the table is the villain here.
-
On Tue, Jun 8, 2010 at 8:58 PM, Mark Rostron wrote:
> I am looking for some specific information regarding optimizer behavior.
>
> We recently experienced a situation where a query that was previously using
> a btree lookup (efficient) SWITCHED to using seqscan/hash lookup.
OK, before we go any f
I am looking for some specific information regarding optimizer behavior.
We recently experienced a situation where a query that was previously using a
btree lookup (efficient) SWITCHED to using seqscan/hash lookup.
My questions would be:
- Under what circumstances is the optimizer like
