[
https://issues.apache.org/jira/browse/HDFS-7647?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14310191#comment-14310191
]
Arpit Agarwal commented on HDFS-7647:
-------------------------------------
bq. This poses a larger problem in that the client may expect changes to the
datanode state to be reflected in the output of LocatedBlock.getLocations(),
but they won't be because we create a new DatanodeInfoWithStorage object during
construction and return it in getLocations().
I think this is a test specific problem. Clients will never get access to the
NameNode object, instead they will have a replica constructed from the wire
message. The test case directly calls the FSNameSystem.getBlockLocations. The
test should be fixed to not poll the cached object but query a new one
periodically.
> DatanodeManager.sortLocatedBlocks() sorts DatanodeInfos but not StorageIDs
> --------------------------------------------------------------------------
>
> Key: HDFS-7647
> URL: https://issues.apache.org/jira/browse/HDFS-7647
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 2.6.0
> Reporter: Milan Desai
> Assignee: Milan Desai
> Attachments: HDFS-7647-2.patch, HDFS-7647-3.patch, HDFS-7647-4.patch,
> HDFS-7647.patch
>
>
> DatanodeManager.sortLocatedBlocks() sorts the array of DatanodeInfos inside
> each LocatedBlock, but does not touch the array of StorageIDs and
> StorageTypes. As a result, the DatanodeInfos and StorageIDs/StorageTypes are
> mismatched. The method is called by FSNamesystem.getBlockLocations(), so the
> client will not know which StorageID/Type corresponds to which DatanodeInfo.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
