[jira] Commented: (HDFS-1450) HDFS federation: Introduce block pool ID into FSDatasetInterface

["Boris Shkolnik (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HDFS-1450?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12921477#action_12921477 ] 

Boris Shkolnik commented on HDFS-1450:
--------------------------------------

DataBlockScanner: deleteBlocks([]) - 
 ExtendedBlock block = new ExtendedBlock();
     for ( Block b : blocks ) {
      block.set(poolId, b);
      deleteBlock(block);
}
These seems troublesome for me. First it reuses the same instance of EB. It is ok in this case, but if someone ever changes code below to use the EB reference it will stop working. But in more general case - shouldn't we just get the blockmap mapped from the poolid and look for a block in it?

Or is it planned for later? 

> HDFS federation: Introduce block pool ID into FSDatasetInterface
> ----------------------------------------------------------------
>
>                 Key: HDFS-1450
>                 URL: https://issues.apache.org/jira/browse/HDFS-1450
>             Project: Hadoop HDFS
>          Issue Type: New Feature
>    Affects Versions: Federation Branch
>            Reporter: Suresh Srinivas
>            Assignee: Suresh Srinivas
>             Fix For: Federation Branch
>
>         Attachments: HDFS-1450.patch
>
>
> FSDatasetInterface methods accept {{Block}}. Changing the interface methods to accept {{ExtendedBlock}}.

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