[ 
https://issues.apache.org/jira/browse/WICKET-2300?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Igor Vaynberg resolved WICKET-2300.
-----------------------------------

    Resolution: Won't Fix
      Assignee: Igor Vaynberg

this is caused by resource polling being enabled during development mode. you 
can turn this off in settings getresourcesettings.setpollfrequency(null).

this is caused by websphere/jdk combination where url connections into jarred 
resources are not being properly cleaned up. there is nothing really we can do 
about it, you just cant use resource polling on websphere.

> Memory leak when development is active
> --------------------------------------
>
>                 Key: WICKET-2300
>                 URL: https://issues.apache.org/jira/browse/WICKET-2300
>             Project: Wicket
>          Issue Type: Bug
>    Affects Versions: 1.3.5
>         Environment: WebSphere Platform 6.1 [WBI 6.1.2.0 o0824.03] [ND 
> 6.1.0.17 cf170821.07]  running with process name widCell\widNode\server1 and 
> process id 11399
> Host Operating System is Linux, version 2.6.18-128.1.6.el5.centos.plus
> Java version = J2RE 1.5.0 IBM J9 2.3 Linux x86-32 j9vmxi3223ifx-20080616 (JIT 
> enabled)
> J9VM - 20080613_20226_lHdSMr
> JIT  - 20080130_0718ifx5_r8
> GC   - IFIX2008_05, Java Compiler = j9jit23, Java VM name = IBM J9 VM
>            Reporter: christophe bouhon
>            Assignee: Igor Vaynberg
>
> Scenario:
> -put the application in development mode
> -deploy the ear
> -hit a few random pages of my application (just display)
> -memory start to rise at a rate > 1Mb/sec up to 2Gb and more
> -at the end the process is killed by the os due to the memory usage
> Comments:
> - only changing the mode from development to deployment is enough to make the 
> memory leak behavior not happen
> - it does not seem to be a heap leak because:
> -- the java heap memory setting was set to 512Mb max
> -- the heap size was low (250Mb or so) while process size was > 2Gb: seen via 
> vm dump (kill -QUIT) and also via profiler (jProfiler)
> however triggering gc at regular interval seems to influence the behavior...

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

Reply via email to