more questions Re: database connection pool design

[al...@abnamro.com]

> The pool will get new connections on an as needed basis up to some
> maximum amount.  Since you need some connections available outside the
> pool, you would set the number used by the pool to a fraction
> of those available from the db.  The pool does not have a dynamic
> resizing of the maximum number of connections.

I see. Another thing that confuses me is that I observe some strange
behavior: in the log file it shows something like

(in + out = total): 4 + 8 = 12

Database connections are used by the application only for
short periods of time and then released back to the pool and records
like that appear in the log file when I know for sure that none of
the connections should be in use. And if I list open connections using my
db server utilities, it shows only 4 open connections for my application
which is consistent with what log file shows for available connections.

It looks like ConnectionPool.totalConnections contains wrong information
(it's supposed to be 4 as well) which is not good because
ConnectionPool.getConnection() won't establish any new connections if
totalConnections >= maxConnections. Could you please clarify what might be
going on?

Thanks and best regards,
Alexander Vorobiev



---------------------------------------------------------------------
To unsubscribe, e-mail: turbine-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: turbine-user-help@jakarta.apache.org

DB connection pooling

[Jonathan Carlson <jo...@yahoo.com>]

Somehow I feel I might get flamed on this, but I have to say it.  One 
thing to consider:  PoolMan (http://poolman.sf.net) may handle these 
issues better or differently than the current Turbine connection 
pooling.  It is licensed under the LGPL so it could be distributed with 
Turbine and then the Turbine developers could spend more time on 
developing functionality that would set Turbine apart even more than it 
already is instead of reinventing the wheel.

Please don't flame me.  I understand that some Apache types don't like 
anything GNU and that there is probably some good reasons to stick with 
the existing Turbine pooling mechanism.  I'm also not familiar at all 
with the quality of the PoolMan code, so my suggestion may not be worth 
anything.

If the only downside was the licensing difference, I'd go with not 
recreating the wheel.

For what it's worth,

Jonathan Carlson
joncrlsn@users.sf.net

alexander.vorobiev@abnamro.com wrote:

>>The pool will get new connections on an as needed basis up to some
>>maximum amount.  Since you need some connections available outside the
>>pool, you would set the number used by the pool to a fraction
>>of those available from the db.  The pool does not have a dynamic
>>resizing of the maximum number of connections.
>>
>
>I see. Another thing that confuses me is that I observe some strange
>behavior: in the log file it shows something like
>
>(in + out = total): 4 + 8 = 12
>
>Database connections are used by the application only for
>short periods of time and then released back to the pool and records
>like that appear in the log file when I know for sure that none of
>the connections should be in use. And if I list open connections using my
>db server utilities, it shows only 4 open connections for my application
>which is consistent with what log file shows for available connections.
>
>It looks like ConnectionPool.totalConnections contains wrong information
>(it's supposed to be 4 as well) which is not good because
>ConnectionPool.getConnection() won't establish any new connections if
>totalConnections >= maxConnections. Could you please clarify what might be
>going on?
>
>Thanks and best regards,
>Alexander Vorobiev
>
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: turbine-user-unsubscribe@jakarta.apache.org
>For additional commands, e-mail: turbine-user-help@jakarta.apache.org
>