river-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Shawn Ellis (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (RIVER-447) Leaked Executor Service Threads in LoadClass
Date Sat, 14 Jan 2017 15:35:26 GMT

    [ https://issues.apache.org/jira/browse/RIVER-447?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15822836#comment-15822836
] 

Shawn Ellis edited comment on RIVER-447 at 1/14/17 3:34 PM:
------------------------------------------------------------

I've attached a patch that is a hack which calls ExecutorService.shutdown() when it is being
removed. This prevents the threads from being leaked, but I realize that it's probably not
the best solution. Thanks for the help.



was (Author: ellisvelo):
I've attached a patch that is a hack which calls ExecutorService.shutdown() when it is being
removed. This prevents the threads from being leaked, but I realize that this probably not
the best solution. Thanks for the help.


> Leaked Executor Service Threads in LoadClass
> --------------------------------------------
>
>                 Key: RIVER-447
>                 URL: https://issues.apache.org/jira/browse/RIVER-447
>             Project: River
>          Issue Type: Bug
>          Components: net_jini_loader
>    Affects Versions: River_3.0.0
>         Environment: Linux with either JDK 1.7 or 1.8
>            Reporter: Shawn Ellis
>         Attachments: ExecutorShutdown.patch
>
>
> I am seeing an overall thread usage increase when using Apache River 3.0. I'm able to
reproduce the problem with both JDK 1.7 and 1.8. The issue is that LoadClass makes use of
a loaderMap that contains an Executor Service. After 10 seconds, the loaderMap will garbage
collect the Executor Service, but the Executor Service will not be shutdown. This leaves the
Executor Service thread still running and waiting for work.
>      How to Reproduce:
>      1. Start up an Apache River 3.0 instance
>      2. Have a client connect to the River instance
>      3. Wait 10 seconds
>      4. Have the client connect to the River instance a second time. The number
>         of threads will have increased.
>      The leaked threads have a stack trace similar to the one below.
>        "net.jini.loader.pref.PreferredClassLoader@7af8260a["httpmd://10.0.1.5:9070/reggie-dl.jar;sha=6c5b83e0caec74d5d4226dcd2c2311d29e81ac0a
httpmd://10.0.1.5:9070/jsk-dl.jar;sha=002bca7b77431ba20385d7ca5be8fa8ec1124a01"]_thread-0"
#30149 prio=5 os_prio=0 tid=0x00003fff68f79000 nid=0x5db9 waiting on condition [0x00003ffdc344d000]
>           java.lang.Thread.State: WAITING (parking)
>                at sun.misc.Unsafe.park(Native Method)
>                - parking to wait for  <0x00000000f2955ff0> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
>                at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
>                at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
>                at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442)
>                at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1067)
>                at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1127)
>                at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
>                at java.lang.Thread.run(Thread.java:745)



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message