You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@nifi.apache.org by "ASF GitHub Bot (JIRA)" <ji...@apache.org> on 2018/11/01 15:30:02 UTC
[jira] [Commented] (NIFI-5770) Memory Leak in ExecuteScript
[ https://issues.apache.org/jira/browse/NIFI-5770?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16671756#comment-16671756 ]
ASF GitHub Bot commented on NIFI-5770:
--------------------------------------
Github user mattyb149 commented on the issue:
https://github.com/apache/nifi/pull/3117
+1 LGTM, ran unit tests and tried a Jython script in a live NiFi instance. Thanks for the improvement! Merging to master
> Memory Leak in ExecuteScript
> ----------------------------
>
> Key: NIFI-5770
> URL: https://issues.apache.org/jira/browse/NIFI-5770
> Project: Apache NiFi
> Issue Type: Bug
> Components: Extensions
> Reporter: Ed Berezitsky
> Assignee: Ed Berezitsky
> Priority: Major
> Labels: performance
> Attachments: 3117.patch, ExecuteScriptMemLeak.xml, jython_modules.zip
>
>
> ExecuteScript with Jython engine has memory leak.
> It uses JythonScriptEngineConfigurator class to configure jython execution environment.
> The problem is in the line:
> {code:java}
> engine.eval("sys.path.append('" + modulePath + "')");{code}
> There is no check if a module has already been added previously.
> As a result, with each execution (onTrigger), string value of module property is being appended, and never reset.
> Although InvokeScriptedProcessor uses the same engine configurator, memory leak is not reproducable in it,
> because ISP builds the engine and compile the code only once (and rebuilds every time any relevant property is changed).
> Attached:
> * template with a flow to reproduce the bug
> * simple python modules (to be unpacked under /tmp)
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)