Обсуждение: Managing connections

C K wrote:
> Dear Friends,
> I want to know if I open a table to view the data a new connection is
> opened. If I open another table/execute a query does a another new
> connection is opened? As per manual, I know that for a database
> cluster and the same port a postgresql instance runs. But if we run
> multiple SQL statements from same client machine with same username,
> does postgresql creates a new instance?
How can connections to the db are handled is determined by the
application program that accesses the database. Some application
programs create a connection for each user and re-use it each time the
user accesses the database. The app is also responsible for closing the
connection when the user signs off.

Other apps create a pool of connections and lease a connection on demand
to each operation that accesses the db. When the operation is complete,
the connection is returned to the pool.  This is the most efficient way
to use connections and I would recommend this approach for just about
any program.

We created our own connection pool program because we wanted a
connection pool that we could use to access any db and because we
considered it to be a key technology.  It's a good thing that we did
because we started out using MS SQLServer for our SaaS applications and
migrated to Postgres.  Migration was a breeze.

Postgresql offers a connection pool which I have not used but appears to
be highly regarded. You can find info about it here:
http://pgpool.projects.postgresql.org/

Some apps are not well designed and don't close connections or un-lease
connections when they should. I call this phenomena "leaking
connections". You can detect this by having everyone sign off the
application and see how many connections are still open.

Cheers,
HH
> If yes, how to manage memory? (As per my observations, normally each
> postgresql instance takes 5-10 MB. So if we have to manage
> simultenious 1000 connections, it will require 1000*5MB=5GB + system
> memory requirements (approx. 1 GB) = total 6 GB (min.) Does this is
> correct?)
> Please give the details. I am using Win XP Pro with PostgreSQL 8.3.
> Thanks
> CPK
>
> --
> Keep your Environment clean and green.
>
> --------------------------------
>
>
>


--
H. Hall
ReedyRiver Group LLC
http://www.reedyriver.com

C K wrote:
> Thanks for this information. We are migrating our application from
> mysql to postgresql and use MS Access/ODBC for front end.
Mystery explained. Access is notorious for opening a connection for
everything. Visual Basic used to do the same thing. I think the dot net
stuff now uses a connection pool.
> When I tested on windows, PG starts it's new instance for each
> application that uses ODBC to connect.
Remember it's the app that calls for the connection through ODBC.
Postgres simply responds to the call.
> Even we open multiple tables from a single file, it use only one
> instance, but when another application with same username even from
> same machine is started, another PG instance is started. Now we are
> testing for other applications and impact on memory.
>
> Regards,
> CPK
>
>
> On Thu, Aug 7, 2008 at 3:09 PM, H. Hall <hhall1001@reedyriver.com
> <mailto:hhall1001@reedyriver.com>> wrote:
>
>     C K wrote:
>
>         Dear Friends,
>         I want to know if I open a table to view the data a new
>         connection is opened. If I open another table/execute a query
>         does a another new connection is opened? As per manual, I know
>         that for a database cluster and the same port a postgresql
>         instance runs. But if we run multiple SQL statements from same
>         client machine with same username, does postgresql creates a
>         new instance?
>
>     How can connections to the db are handled is determined by the
>     application program that accesses the database. Some application
>     programs create a connection for each user and re-use it each time
>     the user accesses the database. The app is also responsible for
>     closing the connection when the user signs off.
>
>     Other apps create a pool of connections and lease a connection on
>     demand to each operation that accesses the db. When the operation
>     is complete, the connection is returned to the pool.  This is the
>     most efficient way to use connections and I would recommend this
>     approach for just about any program.
>
>     We created our own connection pool program because we wanted a
>     connection pool that we could use to access any db and because we
>     considered it to be a key technology.  It's a good thing that we
>     did because we started out using MS SQLServer for our SaaS
>     applications and migrated to Postgres.  Migration was a breeze.
>
>     Postgresql offers a connection pool which I have not used but
>     appears to be highly regarded. You can find info about it here:
>     http://pgpool.projects.postgresql.org/
>
>     Some apps are not well designed and don't close connections or
>     un-lease connections when they should. I call this phenomena
>     "leaking connections". You can detect this by having everyone sign
>     off the application and see how many connections are still open.
>
>     Cheers,
>     HH
>
>         If yes, how to manage memory? (As per my observations,
>         normally each postgresql instance takes 5-10 MB. So if we have
>         to manage simultenious 1000 connections, it will require
>         1000*5MB=5GB + system memory requirements (approx. 1 GB) =
>         total 6 GB (min.) Does this is correct?)
>         Please give the details. I am using Win XP Pro with PostgreSQL
>         8.3.
>         Thanks
>         CPK
>
>         --
>         Keep your Environment clean and green.
>
>         --------------------------------
>
>
>
>
>
>     --
>     H. Hall
>     ReedyRiver Group LLC
>     http://www.reedyriver.com
>
>
>
>
> --
> Keep your Environment clean and green.
>
> --------------------------------
>
>
>


--
H. Hall
ReedyRiver Group LLC
http://www.reedyriver.com