commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mark Thomas (JIRA)" <j...@apache.org>
Subject [jira] Resolved: (POOL-145) Number of tested objects in eviction runs of GenericKeyedObjectPool
Date Fri, 19 Jun 2009 16:00:09 GMT

     [ https://issues.apache.org/jira/browse/POOL-145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Mark Thomas resolved POOL-145.
------------------------------

    Resolution: Fixed

Fixed in trunk and will be in 1.5.2 onwards

> Number of tested objects in eviction runs of GenericKeyedObjectPool
> -------------------------------------------------------------------
>
>                 Key: POOL-145
>                 URL: https://issues.apache.org/jira/browse/POOL-145
>             Project: Commons Pool
>          Issue Type: Bug
>    Affects Versions: 1.5.1
>         Environment: Operating System: All
> Platform: All 
>            Reporter: Klaus Kopruch
>            Priority: Minor
>             Fix For: 1.5.2
>
>
> Hi, 
> if numTestsPerEvictionRun is set to n and there are less than n idle objects in the pool,
the evictor thread will still make n tests on these objects in a round-robin manner, i.e.
idle objects can be tested more than once per eviction run. As this also includes validity
tests if testWhileIdle is enbabled and validity tests may be time-consuming, this is a rather
unwanted behavior.
> Instead of testing getNumTests() objects, the pool should test min(_pool.size(),getNumTests())
objects per eviction run.
> [~thomas@schuerger.com] already reported this bug for GenericObjectPool (see POOL-6).
I found out now, that this bug still exists for GenericKeyedObjectPool. Apparently, POOL-6
only solved the bug in GenericObjectPool.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message