[
https://issues.apache.org/jira/browse/HADOOP-10139?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13886183#comment-13886183
]
Arpit Agarwal commented on HADOOP-10139:
----------------------------------------
This is a really useful update and should take care of a common complaint on
the mailing lists.
My comments:
# I think the pseudo-distributed instructions result in the job being executed
‘locally’ instead of being submitted to yarn so it’s not a complete example. We
need minimal yarn configuration to demonstrate job execution on yarn.
# {{bin/start-dfs.sh}} should be {{sbin/start-dfs.sh}}. Same with
{{start-yarn.sh}}
# {{HADOOP_PREFIX}} seems to be necessary, else {{HADOOP_CONF_DIR}} gets
initialized to /etc/hadoop and this breaks the DataNode, for one. We should
document setting {{HADOOP_PREFIX}}.
# Typo: _quickly perform single operations using Hadoop_; _single_ should be
_simple_?
# Windows setup is quite different from Linux and the steps given here would be
confusing for someone attempting to install on Windows. Can you please add a
link to the Windows setup wiki page
https://wiki.apache.org/hadoop/Hadoop2OnWindows and clarify that the
instructions given here are for Linux only? Users tend to take the Linux
instructions and attempt to run them in Cygwin which never works as expected.
Nitpick comments, not that important:
# Instead of "Format a new distributed-filesystem” perhaps we can just say
“Format the filesystem”.
# It may help readability if steps in the execution section were a numbered
list. What do you think?
# It would be nice if the bullet points after "Now you are ready to start your
Hadoop cluster in one of the three supported modes” were links to subsequent
sections, not a big deal.
> Single Cluster Setup document is unfriendly
> -------------------------------------------
>
> Key: HADOOP-10139
> URL: https://issues.apache.org/jira/browse/HADOOP-10139
> Project: Hadoop Common
> Issue Type: Improvement
> Components: documentation
> Affects Versions: 2.2.0
> Reporter: Akira AJISAKA
> Assignee: Akira AJISAKA
> Attachments: HADOOP-10139.2.patch, HADOOP-10139.patch
>
>
> The document should be understandable to a newcomer because the first place
> he will go is "setup a single node".
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
