[
https://issues.apache.org/jira/browse/HDFS-9719?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15188354#comment-15188354
]
Kai Zheng commented on HDFS-9719:
---------------------------------
Thanks [~umamaheswararao] for the review and great comments!
bq. I am seeing some variable names still like “toRecoverLen”. Can we take
chance in this patch to change them like toReconstructLen ?
Yes. I will fix the variable and check one more time for such instances.
bq. doReadMinimum : this method name looks to be wrong. Its actually reading
from minimum required sourced data nodes. But this name looks like it is
reading minimum data length/what?
The concern seems reasonable. I will rename: doReadMinimum =>
doReadMinimumSources, and the same, readMinimum => readMinimumSources.
bq. Also I assume StripedReader itself should handle multiple chunk/cell
readers.
Yeah, *Striped* should itself already be of the meaning of handling multiple
units, source datanodes or target datanodes.
bq. how about renaming class name like StripedReaders -> StripedReader and
StripedReader -> StripedChunkReader and the same comment applies for
StripedWriter*
Pretty good suggestions for more readable component names! Thanks a lot!
Will update the patch accordingly soon. ...
> Refactoring ErasureCodingWorker into smaller reusable constructs
> ----------------------------------------------------------------
>
> Key: HDFS-9719
> URL: https://issues.apache.org/jira/browse/HDFS-9719
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Kai Zheng
> Assignee: Kai Zheng
> Attachments: HDFS-9719-v1.patch, HDFS-9719-v2.patch,
> HDFS-9719-v3.patch, HDFS-9719-v4.patch, HDFS-9719-v5.patch, HDFS-9719-v6.patch
>
>
> This would suggest and refactor {{ErasureCodingWorker}} into smaller
> constructs to be reused in other places like block group checksum computing
> in datanode side. As discussed in HDFS-8430 and implemented in HDFS-9694
> patch, checksum computing for striped block groups would be distributed to
> datanode in the group, where data block data should be able to be
> reconstructed when missed/corrupted to recompute the block checksum. The most
> needed codes are in the current ErasureCodingWorker and could be reused in
> order to avoid duplication. Fortunately, we have very good and complete
> tests, which would make the refactoring much easier. The refactoring will
> also help a lot for subsequent tasks in phase II for non-striping erasure
> coded files and blocks.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
