[
https://issues.apache.org/jira/browse/UIMA-3263?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jerry Cwiklik closed UIMA-3263.
-------------------------------
Resolution: Fixed
Modified code to use CachedThreadPool and to shutdown the pool when an agent's
child process terminates
> Fix Ducc Agent thread leak
> --------------------------
>
> Key: UIMA-3263
> URL: https://issues.apache.org/jira/browse/UIMA-3263
> Project: UIMA
> Issue Type: Bug
> Components: DUCC
> Reporter: Jerry Cwiklik
> Assignee: Jerry Cwiklik
> Fix For: 1.0-Ducc
>
>
> For every process the Agent launches, it instantiates a Metric Collector
> object which gathers process metrics like CPU, swap, and many others. The
> metric collection is driven by a Camel route at regular intervals. Once the
> process stops (or dies) the route is shutdown.
> Within Metric Collector object, each metric is collected in a separate thread
> managed by a thread pool. The main motivation is to collect metrics in
> parallel to improve performance.
> There are two separate problems in the current code:
> 1) The code uses a wrong type of thread pool, namely newFixedThreadPool. This
> pool has a fixed pool of threads which never expire even if idle. A better
> choice is the CachedThreadPool which creates threads on demand, reuses idle
> ones and kills those that are idle over 1 minute.
> 2) The thread pool is never shutdown when the process is shutdown. The code
> actually shuts down the Camel route which drives metric collection but its
> processor ( java object doing the metric collection) does not shutdown its
> thread pool causing the leak.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira