[
https://issues.apache.org/jira/browse/HDFS-7743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14310192#comment-14310192
]
Jing Zhao commented on HDFS-7743:
---------------------------------
bq. It might make sense if BlockInfo became a common super-class for various
impls like contiguous, sparse, etc.
yes, this is the case for our current erasure coding implementation. I guess
the motivation has been clearly stated in the description. Please also see
HDFS-7285 and HDFS-7716 for design details of EC.
bq. If yes, I'd please request refraining since I'm extremely close to posting
a reference-free block triplets replacement and would like to avoid unnecessary
merge issues.
Cool. Then we're both working on some big features that requires later merge.
But instead of just mentioning here, could you please open a jira and post your
patch there? I would be very happy to help you do the merging.
> Code cleanup of BlockInfo and rename BlockInfo to BlockReplicationInfo
> ----------------------------------------------------------------------
>
> Key: HDFS-7743
> URL: https://issues.apache.org/jira/browse/HDFS-7743
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: namenode
> Reporter: Jing Zhao
> Assignee: Jing Zhao
> Priority: Minor
> Attachments: HDFS-7743.000.patch, HDFS-7743.001.patch,
> HDFS-7743.002.patch
>
>
> In the work of erasure coding (HDFS-7285), we plan to extend the class
> BlockInfo to two subclasses: BlockReplicationInfo and BlockGroupInfo
> (HDFS-7716). To ease the HDFS-EC branch syncing with trunk, this jira plans
> to rename the current BlockInfo to BlockReplicationInfo in trunk. In the
> meanwhile, we can also use this chance to do some minor code cleanup.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
