[
https://issues.apache.org/jira/browse/HDFS-9869?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15246771#comment-15246771
]
Zhe Zhang commented on HDFS-9869:
---------------------------------
Thanks Rakesh for the update. Path LGTM overall. A few remaining minor issues:
# Chain deprecation below. Maybe we should combine them?
{code}
| dfs.replication.pending.timeout.sec |
dfs.namenode.replication.pending.timeout-sec |
| dfs.namenode.replication.pending.timeout-sec |
dfs.namenode.reconstruction.pending.timeout-sec |
{code}
# As a followup we should also deprecate other replication-related config keys.
# Renaming {{getExcessBlocksCount}} to {{getExtraBlocksCount}} doesn't seem
necessary. {{FSNamesystem}} is using "excess" anyway. Sorry my previous comment
pointed to this one. I just noticed the original name uses "blocks" instead of
"replicas". Maybe we can also keep the "excess" in {{excessRedundancyMap}}?
This matches better with the getter.
+1 after the above are addressed.
> Erasure Coding: Rename replication-based names in BlockManager to more
> generic [part-2]
> ---------------------------------------------------------------------------------------
>
> Key: HDFS-9869
> URL: https://issues.apache.org/jira/browse/HDFS-9869
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: erasure-coding
> Reporter: Rakesh R
> Assignee: Rakesh R
> Labels: hdfs-ec-3.0-must-do
> Attachments: HDFS-9869-001.patch, HDFS-9869-002.patch,
> HDFS-9869-003.patch, HDFS-9869-004.patch
>
>
> The idea of this jira is to rename the following entities in BlockManager as,
> - {{PendingReplicationBlocks}} to {{PendingReconstructionBlocks}}
> - {{excessReplicateMap}} to {{extraRedundancyMap}}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
