hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Uma Maheswara Rao G (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-8096) DatanodeMetrics#blocksReplicated will get incremented early and even for failed transfers
Date Thu, 09 Apr 2015 06:15:12 GMT

    [ https://issues.apache.org/jira/browse/HDFS-8096?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14486802#comment-14486802
] 

Uma Maheswara Rao G commented on HDFS-8096:
-------------------------------------------

Thanks Vinay for the patch. Makesense to me for incrementing the metric count only when real
transfer happened. I agree for having test for intermediate failures would be difficult. Thanks
for having one positive test here.
+1


> DatanodeMetrics#blocksReplicated will get incremented early and even for failed transfers
> -----------------------------------------------------------------------------------------
>
>                 Key: HDFS-8096
>                 URL: https://issues.apache.org/jira/browse/HDFS-8096
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: datanode
>            Reporter: Vinayakumar B
>            Assignee: Vinayakumar B
>         Attachments: HDFS-8096-01.patch
>
>
> {code}    case DatanodeProtocol.DNA_TRANSFER:
>       // Send a copy of a block to another datanode
>       dn.transferBlocks(bcmd.getBlockPoolId(), bcmd.getBlocks(),
>           bcmd.getTargets(), bcmd.getTargetStorageTypes());
>       dn.metrics.incrBlocksReplicated(bcmd.getBlocks().length);{code}
> In the above code to handle replication transfers from namenode, {{DatanodeMetrics#blocksReplicated}}
is getting incremented early, since the transfer will happen in background. 
> And even failed transfers also getting counted.
> Correct place to increment this counter is {{DataTransfer#run()}}.



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

Mime
View raw message