[hibernate-issues] [Hibernate-JIRA] Commented: (HSEARCH-492) FSMasterDirectoryProvider and FSSlaveDirectoryProvider don't shutdown the Executor

[Sanne Grinovero (JIRA)]

    [ http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-492?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=36200#action_36200 ] 

Sanne Grinovero commented on HSEARCH-492:
-----------------------------------------

So assuming we add the configuration parameter, should this always be configured to a reasonably large amount of time to guarantee the the copy finishes correctly? In that case I could better not have a timeout and wait for whatever time it takes to finish.

If on the other side our standpoint is that it isn't dangerous to interrupt the copy, that means we can interrupt the copy whenever we like, then we should be interrupting immediately; I believe the copy process was designed to be safe enough, and a program start would start with a fresh copy anyway.

> FSMasterDirectoryProvider and FSSlaveDirectoryProvider don't shutdown the Executor
> ----------------------------------------------------------------------------------
>
>                 Key: HSEARCH-492
>                 URL: http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-492
>             Project: Hibernate Search
>          Issue Type: Bug
>          Components: directory provider
>    Affects Versions: 3.1.1.GA
>         Environment: Java 1.6.0
>            Reporter: Sylvain Charton
>            Priority: Minor
>             Fix For: 3.2.0.CR1
>
>
> I'm using Hibernate Search clustering feature in a stand alone Java application (not a webapp). 
> My problem is the application never stops: many threads are still living.
> Both FSMasterDirectoryProvider and FSSlaveDirectoryProvider use a Single Thread Executor to perform the CopyTask. 
> When the FSSlaveDirectoryProvider.stop() is called, the timer is stopped, but not the Executor in the TimerTask. 
> Consequence is the process can't stop, because some threads are still living, endlessly. 
> This happens after the first CopyDirectory task was executed. 
> My fix was to copy this class in my code, and replace the Executor in TriggerTask class by an ExecutorService in FSSlaveDirectoryProvider class. 
> The ExecutorService is initialized in the start() method:
> {code:Java}this.executor = Executors.newSingleThreadExecutor() ;{code}
> And properly shut down in the stop() method:
> {code:java}
> executor.shutdown();
> boolean terminated = false ;
> try{
>   terminated = executor.awaitTermination(20, TimeUnit.SECONDS);
> } catch (InterruptedException e1) {
>   terminated = false ;
> }
> if (!terminated) {
>   executor.shutdownNow() ;
> }
> {code}

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira