[
https://issues.apache.org/jira/browse/HDFS-5648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13846104#comment-13846104
]
Tsz Wo (Nicholas), SZE commented on HDFS-5648:
----------------------------------------------
+1 patch looks good.
A minor comment: the type of volumes.volumes is already List<FsVolumeImpl>.
The for loop below should be "for(FsVolumeImpl v : : volumes.volumes)". Then
the cast in "((FsVolumeImpl) v).toDatanodeStorage()" is unnecessary.
{code}
+ for (FsVolumeSpi v : volumes.volumes) {
+ ArrayList<ReplicaInfo> finalizedList = finalized.get(v.getStorageID());
+ ArrayList<ReplicaInfo> ucList = uc.get(v.getStorageID());
+ blockReportsMap.put(((FsVolumeImpl) v).toDatanodeStorage(),
+ new BlockListAsLongs(finalizedList, ucList));
+ }
{code}
> Get rid of perVolumeReplicaMap
> ------------------------------
>
> Key: HDFS-5648
> URL: https://issues.apache.org/jira/browse/HDFS-5648
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: datanode
> Affects Versions: Heterogeneous Storage (HDFS-2832)
> Reporter: Arpit Agarwal
> Assignee: Arpit Agarwal
> Attachments: h5648.02.patch, h5648.08.patch, h5648.09.patch
>
>
> The perVolumeReplicaMap in FsDatasetImpl.java is not necessary and can be
> removed. We continue to use the existing volumeMap.
--
This message was sent by Atlassian JIRA
(v6.1.4#6159)
