On 11/08/2013 05:41 AM, si24 wrote:
> Ok now I seem to have gotten it working by changing the store port numbers to
> pgbouncer because its using port no 5432 and postgres is now using 6432.
>
> So now it seems to be giving me the server details that is being used when
> the map gets run:
>
> pgbouncer=# show servers;
> type | user | database | state | addr | port | local_addr |
> local_port | connect_time | request_time | ptr | link
>
------+----------+----------+-------+-----------+------+------------+------------+---------------------+---------------------+----------+------
> S | postgres | manifold | idle | 127.0.0.1 | 6432 | 127.0.0.1 |
> 62261 | 2013-11-08 15:08:37 | 2013-11-08 15:08:48 | 0197da20 |
> S | postgres | manifold | idle | 127.0.0.1 | 6432 | 127.0.0.1 |
> 62259 | 2013-11-08 15:08:37 | 2013-11-08 15:08:47 | 0197d930 |
> (2 rows)
>
> and connections have gone down from 100 to 5 when I check on pgadmin for the
> connections.
>
> Is it because of putting the default pool size from 20 - 2 ( I was trying to
> test something).
>
> Is meant to show that its idle or that its being used on the server part on
> the top
In order to get answers you will need to provide more and consistent
information. For instance in the configure file you sent earlier default
pool size was 100, then you mention it was 20 and now is 2. Also what is
the store and how does it fit into the scheme of things? More generally
what is the scheme of things? In other words detail how your
applications(s), server(s) are laid out relative to each other when it
comes to the flow of information. As to what the above means it would
depend on what 'when the map gets run' means. Some of that could be
found out by tailing the Postgres log directly and seeing what is
happening at the same time.
--
Adrian Klaver
adrian.klaver@gmail.com