hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tsz Wo (Nicholas), SZE (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-5384) DataNodeCluster should not create blocks with generationStamp == 1
Date Wed, 04 Mar 2009 23:21:56 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-5384?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12678983#action_12678983
] 

Tsz Wo (Nicholas), SZE commented on HADOOP-5384:
------------------------------------------------

> NameNode can reject such blocks, but dieing does not seem correct.
Agree.  This is going to be fixed in HADOOP-5399.

Tested the patch manually.  It fixes the problem with DataNodeCluster.  Since the patch is
simple, I won't add new unit tests.

> DataNodeCluster should not create blocks with generationStamp == 1
> ------------------------------------------------------------------
>
>                 Key: HADOOP-5384
>                 URL: https://issues.apache.org/jira/browse/HADOOP-5384
>             Project: Hadoop Core
>          Issue Type: Bug
>          Components: test
>            Reporter: Tsz Wo (Nicholas), SZE
>            Assignee: Tsz Wo (Nicholas), SZE
>            Priority: Blocker
>             Fix For: 0.19.2, 0.20.0
>
>         Attachments: 5384_20090303.patch
>
>
> In DataNodeCluster.main(..), injected blocks are created with generationStamp == 1, which
is a reserved value but not a valid generation stamp.  As a consequence, NameNode may die
when those blocks are reported.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message