hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Phil Yang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-15408) MiniCluster's master crash on initialization and unittest timeout
Date Mon, 07 Mar 2016 04:58:40 GMT

    [ https://issues.apache.org/jira/browse/HBASE-15408?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15182585#comment-15182585
] 

Phil Yang commented on HBASE-15408:
-----------------------------------

I find a wait logic in JVMClusterUtil:
{code}
if (System.currentTimeMillis() > startTime + maxwait) {
        String msg = "Master not initialized after " + maxwait + "ms seconds";
        Threads.printThreadInfo(System.out,
          "Thread dump because: " + msg);
        throw new RuntimeException(msg);
      }
{code}
We will throw the Exception after 200 seconds if the master crashes. It may be too long because
the tests marked on @Category(MediumTests.class) will be timeout after 180 seconds. I think
we can decrease it from 200 to 30?


> MiniCluster's master crash on initialization and unittest timeout
> -----------------------------------------------------------------
>
>                 Key: HBASE-15408
>                 URL: https://issues.apache.org/jira/browse/HBASE-15408
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Phil Yang
>         Attachments: fail.txt
>
>
> These days there are many tests timeout on build.apache.org. I have no log on timeout
tests but I find a possible reason: master crash on initialization and minicluster will log
"No master found; retry" forever. The crash is caused by OutOfOrderScannerNextException.



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

Mime
View raw message