db-torque-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ian Ragsdale <...@SKYLIST.net>
Subject Re: Connection leak
Date Fri, 24 Sep 2004 20:24:50 GMT
We recently had a similar problem.  Here's what we did, which appeared 
to alleviate the problem without addressing the underlying connection 
leak:

# don't wait indefinitely for a connection
torque.dsfactory.scimitar.pool.maxWait = 5000

# allow up to 25 connections in the pool (up from 8)
torque.dsfactory.scimitar.pool.maxActive = 25

# remove connections that have been abandoned
torque.dsfactory.scimitar.pool.removeAbandoned = true

This should postpone running out of connections by adding more and give 
the pool time to reclaim abandoned connections.  There is also a 
setting to have it track where connections are leaked from, at the 
expense of a little overhead when opening connections:

logAbandoned = true

More configuration settings are here:

http://jakarta.apache.org/commons/dbcp/configuration.html

On Sep 24, 2004, at 3:06 PM, Husek, Paul wrote:

> Hello,
>
> I'm using Torque v3.1 with commons-dbcp connection pooling.  I'm 
> getting
> what I believe is a connection leak.  After a certain time of my 
> application
> being up and responsive it just hangs.
>
> Is there anyway I can verify this?  E.g. Can I see how many 
> connections are
> left in the pool, how many have been taken and not returned, etc? In 
> some
> places I do a Torque.getConnection() but may not be returning them.  
> Can I
> cause them to be reaped automatically if I forget to return them?
>
>
>
>
>
> Paul
>
>
>


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


Mime
View raw message