tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From sasidhar prabhakar <sasidhar1...@gmail.com>
Subject Re: DBCP abandoned trace - unable to understand the leak
Date Wed, 10 Nov 2010 09:50:06 GMT
When I get this problem, I tried the query in DB manually

by this query *select count(*) from v$process;*
*
*
The count some times very less, like if total connections are 200 it shows *
*
some times 60,40,162 like this.


On Wed, Nov 10, 2010 at 3:20 AM, Mark Thomas <markt@apache.org> wrote:

> On 10/11/2010 09:02, Pid wrote:
> > On 04/11/2010 12:04, sasidhar prabhakar wrote:
> >> dataSource = ConnectionUtil.getDataSource();
> >> }
> >
> > Is the class you posted the only DAO?  Could the leak be from another
> class?
> >
> > Can you post ConnectionUtil.java?
>
> Given the SQL seen so far and that some queries take longer than 30s to
> complete, my money is on the the app trying to process more long running
> queries in parallel then the pool has connections available.
>
> With a low time-out (30s), the pool was abandoning the connections.
>
> With a long time-out (300s), the pool was becoming exhausted.
>
> If this analysis is correct, the fix is to address the root cause of the
> long running queries. Unless you are lucky and there is one poorly
> performing query, the chances are the application and/or database have
> architectural issues that will require significant work to put right.
> Web applications should not routinely be running queries as part of
> request processing that take in excess of a second or so to run.
>
> Mark
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
> For additional commands, e-mail: users-help@tomcat.apache.org
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message