[ https://issues.apache.org/jira/browse/CAMEL-13033?focusedWorklogId=179802&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-179802 ]
ASF GitHub Bot logged work on CAMEL-13033: ------------------------------------------ Author: ASF GitHub Bot Created on: 30/Dec/18 16:38 Start Date: 30/Dec/18 16:38 Worklog Time Spent: 10m Work Description: fvaleri commented on pull request #2694: CAMEL-13033: Memory leak in ReactiveHelper class URL: https://github.com/apache/camel/pull/2694 It seems that, in some cases, we are never polling from the back list of work in org.apache.camel.support.ReactiveHelper.Work and this leads to OOM error. This is a test case (try to set numOfRecords=1000000 and maxWaitTime=20000): https://github.com/fvaleri/camel/tree/bigxml-split-example/examples/camel-example-bigxml-split I tried to set the "main" parameter to true in the schedule call inside ReactiveHelper.scheduleLast. If I understand correctly, this avoids the accumulation of after processor work from CamelInternalProcessor by executing it in the route's main execution thread. After this change I see the after processor TRACE log (CamelInternalProcessor line 232) and I have similar test results as with 2.23.0, even if I have to increase maxWaitTime a little for the 1.3GB test. I'm not completely aware of the implications of this change in other parts of the code, so I need a double check from you. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking ------------------- Worklog Id: (was: 179802) Time Spent: 10m Remaining Estimate: 0h > Memory leak in ReactiveHelper class > ----------------------------------- > > Key: CAMEL-13033 > URL: https://issues.apache.org/jira/browse/CAMEL-13033 > Project: Camel > Issue Type: Bug > Components: camel-core > Affects Versions: 3.0.0 > Reporter: fvaleri > Priority: Major > Fix For: 3.0.0 > > Attachments: Screenshot 2018-12-28 at 7.18.11 PM.png > > Time Spent: 10m > Remaining Estimate: 0h > > I was doing a performance test using the Splitter EIP and after reaching good > results with 2.23.0 release I tried the same test with 3.0.0-SNAPHOST > discovering a memory leak located in the ReactiveHelper class (back > LinkedList). It's at the routine engine level and a OOM error should be > triggered with every load test, but I can share my test if it's needed. -- This message was sent by Atlassian JIRA (v7.6.3#76005)