[
https://issues.apache.org/jira/browse/HDFS-1639?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12998661#comment-12998661
]
Suresh Srinivas commented on HDFS-1639:
---------------------------------------
FSDataset currently has:
* FSVolumeSet - that manages one or more FSVolumes, one each for storage
directory configured.
* FSVolume - manages a storage directory. It manages the
<StorageDirectory>/current directory. It tracks
directories such as tmp, finalized, rbw and rur. It has FSDir for every
directory under finalized where blocks are
stored.
* ReplicasMap - manages a map of block ID to ReplicaInfo.
Proposed changes:
With the introduction block pool, change is needed to make FSDataset aware of
block pool.
FSVolume changes:
FSVolume will be changed to include one or more BlockPool objects in a map of
BlockPoolId <-> BlockPool.
Managing directories such as finalized, rbw, rur and tmp will be moved to
BlockPool. Also the FSDir management will be
moved to BlockPool.
ReplicasMap changes:
ReplicasMap will become from Map<BlockId, ReplicaInfo>, Map<BlockPoolId,
Map<BlockId,
ReplicaInfo>>.
> Add block pool management to FSDataset.
> ---------------------------------------
>
> Key: HDFS-1639
> URL: https://issues.apache.org/jira/browse/HDFS-1639
> Project: Hadoop HDFS
> Issue Type: New Feature
> Components: data-node
> Affects Versions: Federation Branch
> Reporter: Suresh Srinivas
> Assignee: Suresh Srinivas
> Fix For: Federation Branch
>
>
> FSDataset needs to changes to add the notion of block pools. More description
> in comments.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
