[
https://issues.apache.org/jira/browse/HDFS-6581?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14140238#comment-14140238
]
Andrew Wang commented on HDFS-6581:
-----------------------------------
I'd like to also echo Colin's desire to see some positive benchmark numbers
before merging this. Since the point of this JIRA is to improve read/write
performance, it really seems like we should verify that some nice gains are
possible before adding this complexity. This is what we did for HDFS-4949, and
I did also ask for benchmarks in my comment about a month ago.
Arpit, I seem to remember you had a WIP short-circuit writes patch you were
working on a while ago. Do you have any numbers from that even? Colin wrote a
vectorized sum microbenchmark using zero-copy reads at HDFS-6287 you might be
able to adapt. This does dredge up the discussion about the design of
short-circuit writes, but maybe that's a bridge that has to be crossed.
> Write to single replica in memory
> ---------------------------------
>
> Key: HDFS-6581
> URL: https://issues.apache.org/jira/browse/HDFS-6581
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode
> Reporter: Arpit Agarwal
> Assignee: Arpit Agarwal
> Attachments: HDFS-6581.merge.01.patch, HDFS-6581.merge.02.patch,
> HDFS-6581.merge.03.patch, HDFS-6581.merge.04.patch, HDFS-6581.merge.05.patch,
> HDFS-6581.merge.06.patch, HDFSWriteableReplicasInMemory.pdf,
> Test-Plan-for-HDFS-6581-Memory-Storage.pdf
>
>
> Per discussion with the community on HDFS-5851, we will implement writing to
> a single replica in DN memory via DataTransferProtocol.
> This avoids some of the issues with short-circuit writes, which we can
> revisit at a later time.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
