[
https://issues.apache.org/jira/browse/HDFS-7782?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14380457#comment-14380457
]
Zhe Zhang commented on HDFS-7782:
---------------------------------
Thanks Jing for the quick feedback!
bq. We can first only provide a non-parallel and parallel reading of data
blocks.
Agreed. I think the 004 patch aligns with this plan. It has serial and parallel
modes without hedge or error handling.
bq. I think finally the client should be able to 1) read in parallel, and 2)
handle slow datanodes with data blocks, and 3) correctly handle possibly
missing data blocks. Maybe this finally should be all necessary, and users do
not need to use configuration to turn it on.
This is a great summary. In this long term vision do you think serial read will
still be a mode to support? If so I guess we still need a switch to configure
serial vs. parallel reads?
> Read a striping layout file from client side
> --------------------------------------------
>
> Key: HDFS-7782
> URL: https://issues.apache.org/jira/browse/HDFS-7782
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Li Bo
> Assignee: Zhe Zhang
> Attachments: HDFS-7782-000.patch, HDFS-7782-001.patch,
> HDFS-7782-002.patch, HDFS-7782-003.patch, HDFS-7782-004.patch
>
>
> If client wants to read a file, he is not necessary to know and handle what
> layout the file is. This sub task adds logic to DFSInputStream to support
> reading striping layout files.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
