accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Josh Elser (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-3082) MAC processes still hang around when termination doesn't finish in 30s
Date Mon, 25 Aug 2014 05:31:57 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-3082?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14108757#comment-14108757
] 

Josh Elser commented on ACCUMULO-3082:
--------------------------------------

Oh, also noticed this in the junit log/report:

{noformat}
2014-08-24 04:50:36,381 [impl.MiniAccumuloClusterImpl] WARN : Unexpectedly had 2 task(s) remaining
in threadpool for execution when being stopped
{noformat}

Maybe we had a hard time destroying one of the processes and accidentally stopped the ES before
it could gracefully stop the other processes? Not sure.

> MAC processes still hang around when termination doesn't finish in 30s
> ----------------------------------------------------------------------
>
>                 Key: ACCUMULO-3082
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-3082
>             Project: Accumulo
>          Issue Type: Bug
>          Components: mini
>            Reporter: Josh Elser
>             Fix For: 1.5.2, 1.6.1, 1.7.0
>
>
> I added some support in ACCUMULO-2764 which executes the {{Process.destroy()}} and {{Process.waitFor()}}
that we call to stop the MAC processes with a timeout to prevent the Maven build from hanging
indefinitely.
> However, when we hit the timeout...
> {noformat}
> 2014-08-24 04:49:36,370 [impl.MiniAccumuloClusterImpl] WARN : Master did not fully stop
after 30 seconds
> java.util.concurrent.TimeoutException
>         at java.util.concurrent.FutureTask.get(FutureTask.java:201)
>         at org.apache.accumulo.minicluster.impl.MiniAccumuloClusterImpl.stopProcessWithTimeout(MiniAccumuloClusterImpl.java:755)
>         at org.apache.accumulo.minicluster.impl.MiniAccumuloClusterImpl.stop(MiniAccumuloClusterImpl.java:645)
>         at org.apache.accumulo.minicluster.impl.MiniAccumuloClusterGCTest.tea
> {noformat}
> The processes will hang around on the host system.
> Need to get:
> # Better diagnosis of why exactly the destroy and waitFor doesn't return
> # (try to?) Guarantee 100% cleanup of processes when {{MiniAccumuloCluster.stop()}} is
called.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message