commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Phil Steitz (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (DBCP-334) BasicDataSource does not consider the method "setSoftMinEvictableIdleTimeMillis" of CommonsPool.GenericObjectPool
Date Mon, 02 May 2011 02:02:03 GMT

     [ https://issues.apache.org/jira/browse/DBCP-334?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Phil Steitz resolved DBCP-334.
------------------------------

    Resolution: Fixed

Fixed in trunk (r1098458) and 1.4 branch (r1098459).

Thanks for the patch!

> BasicDataSource does not consider the method "setSoftMinEvictableIdleTimeMillis" of CommonsPool.GenericObjectPool
> -----------------------------------------------------------------------------------------------------------------
>
>                 Key: DBCP-334
>                 URL: https://issues.apache.org/jira/browse/DBCP-334
>             Project: Commons Dbcp
>          Issue Type: Bug
>    Affects Versions: 1.2.1, 1.2.2, 1.3, 1.4
>            Reporter: Alberto Mozzone
>            Priority: Critical
>             Fix For: 1.3.1, 1.4.1
>
>
> The BasicDataSource should have the "softMinEvictableIdleTimeMillis" (get and set) as
the GenericObjectPool, to avoid destroying idle connections below the "minIdle" value. In
this case, connections are dropped and immediately recreated with no reason. The "minEvictableIdleTimeMillis"
should not be considered because there are ways to validate idle connections before they are
used.
> The class should have these lines of code, besides those existing:
> // property
> 	private long softMinEvictableIdleTimeMillis = GenericObjectPool.DEFAULT_SOFT_MIN_EVICTABLE_IDLE_TIME_MILLIS;
> // getter
> 	/**
> 	 * Returns the {@link #softMinEvictableIdleTimeMillis} property.
> 	 * @return the value of the {@link #softMinEvictableIdleTimeMillis} property
> 	 * @see #softMinEvictableIdleTimeMillis
> 	 */
> 	public synchronized long getSoftMinEvictableIdleTimeMillis() {
> 		return softMinEvictableIdleTimeMillis;
> 	}
> // setter
> 	/**
> 	 * Sets the {@link #softMinEvictableIdleTimeMillis} property.
> 	 * @param softMinEvictableIdleTimeMillis the minimum amount of time an object
> 	 * may sit idle in the pool 
> 	 * @see #softMinEvictableIdleTimeMillis
> 	 */
> 	public synchronized void setSoftMinEvictableIdleTimeMillis(final long softMinEvictableIdleTimeMillis)
{
> 		this.softMinEvictableIdleTimeMillis = softMinEvictableIdleTimeMillis;
> 		if (connectionPool != null) {
> 			connectionPool.setSoftMinEvictableIdleTimeMillis(softMinEvictableIdleTimeMillis);
> 		}
> 	}
> // creator of the pool
> 	@Override protected void createConnectionPool() {
> 		super.createConnectionPool();
> 		connectionPool.setSoftMinEvictableIdleTimeMillis(softMinEvictableIdleTimeMillis);
> 	}
> I implemented myself a "NotSoBasicDataSource" class, inheriting from "BasicDataSource"
and used it in a single test. It worked: idle connections do not decrease under (and re-grow
to) "minIdle".
> Please, consider this feature, because it can save resources.
> I don't know what is the behavior of the Tomcat DBCP, but it seems to me it has the same
problem.
> I'm waiting your opinions.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message