[
https://issues.apache.org/jira/browse/TUSCANY-3842?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Simon Laws reassigned TUSCANY-3842:
-----------------------------------
Assignee: Simon Laws
> Node restart performance degrades as more restarts are performed.
> ------------------------------------------------------------------
>
> Key: TUSCANY-3842
> URL: https://issues.apache.org/jira/browse/TUSCANY-3842
> Project: Tuscany
> Issue Type: Bug
> Components: SCA Java Runtime
> Affects Versions: Java-SCA-2.0
> Environment: WinXP Sun JDK 1.6.0-20
> Reporter: Simon Laws
> Assignee: Simon Laws
>
> Ant has a Node unit test
> (http://svn.apache.org/repos/asf/tuscany/sca-java-2.x/trunk/modules/node-impl/src/test/java/org/apache/tuscany/sca/node/impl/PerflTest.java)
> which starts and stops a node repeatedly. If you watch what happens with
> increasing numbers of repetitions you'll see that the runtime slows and uses
> more memory.
> On initial inspection the DefaultImport is being added into each contribution
> on each restart in DeployerImpl.build hence the increasing memory usage. This
> raises the question though of why the code is doing all this work and each
> restart in the first place. Looking at the code in NodeImpl.start() there is
> a check that ensures that the contributions are only loaded once. But then it
> goes on to recreate the composite for the node. This will always be the same
> as the contributions are only loaded once.
> It looks like we can change the code in start() to only create
> domainComposite and compositeContext when the contributions are loaded.
> stop() will need changing so that these are not reset.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
