[
https://issues.apache.org/jira/browse/HDFS-8623?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14600163#comment-14600163
]
Jing Zhao commented on HDFS-8623:
---------------------------------
Thanks for working on this, Zhe. The patch looks good to me. Some minors:
# In {{hasMinStorage(BlockInfo)}} we can call {{hasMinStorage(BlockInfo, int)}}
# In {{computeReplicationWorkForBlocks}}, it will be better to call
{{getExpectedReplicaNum}} for computing
{{requiredReplication}}.{{getExpectedReplicaNum}} can be called instead of
{{getPreferredBlockReplication}} in several other places.
# For {{BlockInfoToAdd}}, it may not be necessary to have two getter functions
for its data fields.
# In {{processChosenExcessReplica}}, {{blockToInvalidate}} can be removed:
{code}
final Block blockToInvalidate = storedBlock;
addToInvalidates(blockToInvalidate, chosen.getDatanodeDescriptor());
{code}
> Refactor NameNode handling of invalid, corrupt, and under-recovery blocks
> -------------------------------------------------------------------------
>
> Key: HDFS-8623
> URL: https://issues.apache.org/jira/browse/HDFS-8623
> Project: Hadoop HDFS
> Issue Type: New Feature
> Affects Versions: 2.7.0
> Reporter: Zhe Zhang
> Assignee: Zhe Zhang
> Attachments: HDFS-8623.00.patch, HDFS-8623.01.patch,
> HDFS-8623.02.patch, HDFS-8623.03.patch
>
>
> In order to support striped blocks in invalid, corrupt, and under-recovery
> blocks handling, HDFS-7907 introduces some refactors. This JIRA aims to merge
> these changes to trunk first to minimize and cleanup HDFS-7285 merge patch so
> that it only contains striping/EC logic.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
