[ 
https://issues.apache.org/jira/browse/MAPREDUCE-6670?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15226555#comment-15226555
 ] 

Junping Du commented on MAPREDUCE-6670:
---------------------------------------

Thanks [~GergelyNovak] for the patch.
Agree that 1000 ms may not enough, given the sleep time in the worst case could 
reach to 900 ms.
{noformat}
    for (int i = 0; i < 9; i++) {
      values = cache.values();
      if (values.size() > maxSize) {
        Thread.sleep(100);
      } else {
        assertFalse("fileInfo1 should have been evicted",
          values.contains(fileInfo1));
        return;
      }
    }
{noformat}
However, 10 secs may not be necessary. Can we set to 5000 ms instead?

> TestJobListCache#testEviction sometimes fails on Windows with timeout
> ---------------------------------------------------------------------
>
>                 Key: MAPREDUCE-6670
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-6670
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: test
>    Affects Versions: 2.7.0, 2.8.0, 2.7.1, 2.7.2, 2.7.3
>         Environment: OS: Windows Server 2012
> JDK: 1.7.0_79
>            Reporter: Gergely Novák
>            Assignee: Gergely Novák
>            Priority: Minor
>         Attachments: MAPREDUCE-6670.001.patch
>
>
> TestJobListCache#testEviction often needs more than 1000 ms to finish in 
> Windows environment. Increasing the timeout solves the issue.



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

Reply via email to