hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Konstantin Boudnik (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HDFS-1502) TestBlockRecovery triggers NPE in assert
Date Wed, 01 Dec 2010 00:25:11 GMT

    [ https://issues.apache.org/jira/browse/HDFS-1502?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12965502#action_12965502
] 

Konstantin Boudnik commented on HDFS-1502:
------------------------------------------

Looking somewhat more into this I see in a debugger that {{reply.getNumBytes()}} is 6000 {{newBlock.getNumBytes()}}
is only 5000 (as set in the test class). So they are different which causes assertion to happen.

{noformat}
        if(rState == bestState) {
          minLength = Math.min(minLength, r.rInfo.getNumBytes());
          participatingList.add(r);
        }
{noformat}

The difference is set by the code above, apparently. I am not that well familiar with recovery
process in particular, but it seems like there are some issues with this part of the code.
Thoughts?


> TestBlockRecovery triggers NPE in assert
> ----------------------------------------
>
>                 Key: HDFS-1502
>                 URL: https://issues.apache.org/jira/browse/HDFS-1502
>             Project: Hadoop HDFS
>          Issue Type: Bug
>    Affects Versions: 0.22.0
>            Reporter: Eli Collins
>            Priority: Minor
>             Fix For: 0.22.0
>
>         Attachments: HDFS-1502.patch
>
>
> {noformat}
> Testcase: testRBW_RWRReplicas took 10.333 sec
>         Caused an ERROR
> null
> java.lang.NullPointerException
>         at org.apache.hadoop.hdfs.server.datanode.DataNode.syncBlock(DataNode.java:1881)
>         at org.apache.hadoop.hdfs.server.datanode.TestBlockRecovery.testSyncReplicas(TestBlockRecovery.java:144)
>         at org.apache.hadoop.hdfs.server.datanode.TestBlockRecovery.testRBW_RWRReplicas(TestBlockRecovery.java:305)
> {noformat}
> {noformat}
>         Block reply = r.datanode.updateReplicaUnderRecovery(
>             r.rInfo, recoveryId, newBlock.getNumBytes());
>         assert reply.equals(newBlock) &&
>                reply.getNumBytes() == newBlock.getNumBytes() :
>           "Updated replica must be the same as the new block.";    <----- line 1881
> {noformat}
> Not sure how reply could be null since  updateReplicaUnderRecovery always returns a newly
instantiated object.

-- 
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