Mailing-List: contact hdfs-dev-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hdfs-dev@hadoop.apache.org
Message-ID: <2099871201.1251927872734.JavaMail.jira@brutus>
Date: Wed, 2 Sep 2009 14:44:32 -0700 (PDT)
From: "Konstantin Boudnik (JIRA)" <jira@apache.org>
To: hdfs-dev@hadoop.apache.org
Subject: [jira] Created: (HDFS-591) Namenode should mark the replica as
 corrupted if a DN has reported different length of a Complete block
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

Namenode should mark the replica as corrupted if a DN has reported different length of a Complete block
-------------------------------------------------------------------------------------------------------

                 Key: HDFS-591
                 URL: https://issues.apache.org/jira/browse/HDFS-591
             Project: Hadoop HDFS
          Issue Type: Bug
          Components: name-node
            Reporter: Konstantin Boudnik


it has been found that NameNode is updating the length of a Complete block when its reported by a DataNode. As demonstrated by BlockReport.messWithBlocksLen() test ([see HDFS-551's patch|https://issues.apache.org/jira/secure/attachment/12418437/HDFS-551.patch]), when the length of blocks is changed and a block report is forced for the set of the blocks NameNode updates these lengths in memory, instead of marking such replicas as corrupted.

It seems to be a wrong behavior.

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