We use Oracle 11.2.0 and it can handle more than 1000 connections.
Here is the stacktrace of the connection pool empty error that I receive in
tomcat logs
Caused by: org.apache.tomcat.jdbc.pool.PoolExhaustedException:
[http-apr-30690-exec-743] Timeout: Pool empty. Unable to fetch a connection
in 30 seconds, none available[size:100; busy:72; idle:0; lastwait:30000].
at
org.apache.tomcat.jdbc.pool.ConnectionPool.borrowConnection(ConnectionPool.java:667)
at
org.apache.tomcat.jdbc.pool.ConnectionPool.getConnection(ConnectionPool.java:182)
at
org.apache.tomcat.jdbc.pool.DataSourceProxy.getConnection(DataSourceProxy.java:128)
at
org.springframework.orm.hibernate3.LocalDataSourceConnectionProvider.getConnection(LocalDataSourceConnectionProvider.java:82)
at
org.hibernate.jdbc.ConnectionManager.openConnection(ConnectionManager.java:417)
On Thu, Jun 28, 2012 at 11:46 AM, chris derham <ch...@derham.me.uk> wrote:
> On Thu, Jun 28, 2012 at 12:38 PM, Yasser <yarafa...@gmail.com> wrote:
>
> > Yes. It does show that maxactive has reached 100. I also use splunk to
> get
> > the connection status at the oracle side.
> > What I dont understand is that Resin needs just 50 connections to handle
> > the same load. I am in the process of increasing the count to 300 and see
> > if that makes a difference. Oracle has the capacity to handle that many
> > connections.
> >
> > We hit an odd oracle exception that may be related. Can't explain it
> really well, but the essence of the bug in Oracle was that the
> tnslistener's connection count wasn't updated "regularly enough" when
> connections closed, so it could believe that the connection limit had
> exceeded when the db instance was still happy to accept more connections.
> As I recall this was a known Oracle issue, and was fixed by increasing the
> acceptable connection count in Oracle. Can you provide the exact exception
> from the logs that you hit?
>
> Also have you tried turning on connection pool logging to diagnose when
> actual physical connections are created?
>
> Chris
>
