On Thu, Aug 1, 2013 at 4:50 PM, BladeOfLight16 wrote:
> On Thu, Aug 1, 2013 at 10:25 AM, Sandeep Gupta
> wrote:
>>
>> @Jeff : Thanks for pointing this out. Turns out that was the case.
>>
>> @Tom: Thank you for the reference to random_page_cost parameters. It would
>> be very useful for us. Would
On Thu, Aug 1, 2013 at 10:25 AM, Sandeep Gupta wrote:
> @Jeff : Thanks for pointing this out. Turns out that was the case.
>
> @Tom: Thank you for the reference to random_page_cost parameters. It would
> be very useful for us. Would go through the rest of the documentation as
> well.
>
I can't sa
@Jeff : Thanks for pointing this out. Turns out that was the case.
@Tom: Thank you for the reference to random_page_cost parameters. It would
be very useful for us. Would go through the rest of the documentation as
well.
On Wed, Jul 31, 2013 at 3:55 PM, Tom Lane wrote:
> Sandeep Gupta writes:
Sandeep Gupta writes:
> details regarding buffer usage:
> [ 100% buffer hit rate ]
Your database is evidently fully cached in memory. If that's the
operating mode you expect, you need to change the planner's cost
parameters, in particular reduce random_page_cost to equal seq_page_cost.
There is
details regarding buffer usage:
for database 1:
QUERY
PLAN
---
Aggregate (cos
Hi Pavel,
Yes. The postgresql.conf is exactly the same. The have the same index and
clustering and are on the same compute node as well but running on
different ports.
-Sandeep
On Wed, Jul 31, 2013 at 3:14 PM, Pavel Stehule wrote:
> Hello
>
> do you have same configuration?
>
> Regards
>
>
Hello
do you have same configuration?
Regards
Pavel
2013/7/31 Sandeep Gupta :
> I have two postgres instances each with a database of same schema. The
> dataset in both is ''same'' but for randomness i.e. both contain two tables
> pc(did) and tc(pid, did) that have almost
> same number of row
I have two postgres instances each with a database of same schema. The
dataset in both is ''same'' but for randomness i.e. both contain two
tables pc(did) and tc(pid, did) that have almost
same number of rows and have been generate from same distribution.
However the query plan for the join turn
