----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/40576/#review108352 -----------------------------------------------------------
docs/src/site/sphinx/admin/Installation.rst (line 24) <https://reviews.apache.org/r/40576/#comment167774> "any number of machines". no need for 'arbitrary' docs/src/site/sphinx/admin/Installation.rst (line 29) <https://reviews.apache.org/r/40576/#comment167775> Copy the Sqoop artifact to the machine where you want to run Sqoop server. The Sqoop server acts as a Hadoop client, therefore Hadoop libraries (Yarn, Mapreduce, and HDFS jar files) and configuration files (...) must be available on this node. You do not need to run any Hadoop related services - running the server on a "gateway" node is perfectly fine. docs/src/site/sphinx/admin/Installation.rst (line 35) <https://reviews.apache.org/r/40576/#comment167776> "or later" To install the Sqoop server, decompress the tarball (in a location of your choosing) and set the newly created forder as your working directory docs/src/site/sphinx/admin/Installation.rst (line 52) <https://reviews.apache.org/r/40576/#comment167780> In this chunk of documentation we seem to describe a more complicated configuration pattern before describing the simpler one. I would reverse this. The Sqoop server uses environment variables to find Hadoop libraries. If the environment variable ``$HADOOP_HOME`` is set, Sqoop will look for jars in the follwing locations: ``$HADOOP_HOME/share/hadoop/common``, ``$HADOOP_HOME/share/hadoop/hdfs``, ``$HADOOP_HOME/share/hadoop/mapreduce`` and ``$HADOOP_HOME/share/hadoop/yarn``. You can specify where the Sqoop server should look for the common, hdfs, mapreduce, and yarn jars indepently with the ``$HADOOP_COMMON_HOME``, ``$HADOOP_HDFS_HOME``, ``$HADOOP_MAPRED_HOME`` and ``$HADOOP_YARN_HOME`` environment variables. docs/src/site/sphinx/admin/Installation.rst (line 69) <https://reviews.apache.org/r/40576/#comment167782> "Sqoop server requires further configuration on Hadoop side" is not needed. I would split this section into two subsections: "HDFS Proxy Users" and "YARN Whitelist". docs/src/site/sphinx/admin/Installation.rst (line 71) <https://reviews.apache.org/r/40576/#comment167783> does this always need to be enabled. could there be use cases (such as moving data for s3 to an rdbms) where the cluster does not need to be configured to allow impersonation? docs/src/site/sphinx/admin/Installation.rst (line 97) <https://reviews.apache.org/r/40576/#comment167785> I would describe the use case here in a little bit more detail. We have a few different ways of loading jars into sqoop 2. Perhaps we should move this out of the installation documentation into its own page to keep initial installation simple. docs/src/site/sphinx/admin/Installation.rst (line 118) <https://reviews.apache.org/r/40576/#comment167786> "for easier execution" docs/src/site/sphinx/admin/Installation.rst (line 124) <https://reviews.apache.org/r/40576/#comment167787> lets be consistent for shell commands vs binaries. "The remainder of the Sqoop 2 documentation assumes that the shell commands are in your ``$PATH``." I would remove the last sentence, I think the point has been already made clear. docs/src/site/sphinx/admin/Installation.rst (line 129) <https://reviews.apache.org/r/40576/#comment167788> I don't think the first sentence is necessary. I think that is made clear by our documentation showing the start step after this step. docs/src/site/sphinx/admin/Installation.rst (line 135) <https://reviews.apache.org/r/40576/#comment167789> Rename this section "Repository Initialization" docs/src/site/sphinx/admin/Installation.rst (line 138) <https://reviews.apache.org/r/40576/#comment167793> "The metadata repository needs to be initialized. The tool-upgrade can be used..." docs/src/site/sphinx/admin/Installation.rst (line 144) <https://reviews.apache.org/r/40576/#comment167794> "for first execution" is not necessary docs/src/site/sphinx/admin/Installation.rst (line 156) <https://reviews.apache.org/r/40576/#comment167795> "the follwing command" docs/src/site/sphinx/admin/Installation.rst (line 162) <https://reviews.apache.org/r/40576/#comment167796> Remove "similarly" add "the" docs/src/site/sphinx/admin/Installation.rst (line 168) <https://reviews.apache.org/r/40576/#comment167797> "daemons" and "ports" do not need to be plural. docs/src/site/sphinx/admin/Installation.rst (line 173) <https://reviews.apache.org/r/40576/#comment167799> Put the first sentence as a note at the bottom of the section. I think the documentation is clearer if we first state what we will be accomplishing and then resolving common misconceptions after. docs/src/site/sphinx/admin/Installation.rst (line 179) <https://reviews.apache.org/r/40576/#comment167800> "documentation for Sqoop shell" docs/src/site/sphinx/admin/Tools.rst (line 60) <https://reviews.apache.org/r/40576/#comment167801> "the same file" docs/src/site/sphinx/admin/Tools.rst (line 67) <https://reviews.apache.org/r/40576/#comment167802> Perhaps we should make it clear that we are not actually downloading and installing a new version of sqoop 2 as a side node? Really this is a "repository upgrade tool". - Abraham Fine On Nov. 25, 2015, 7:40 p.m., Jarek Cecho wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/40576/ > ----------------------------------------------------------- > > (Updated Nov. 25, 2015, 7:40 p.m.) > > > Review request for Sqoop. > > > Bugs: SQOOP-2708 > https://issues.apache.org/jira/browse/SQOOP-2708 > > > Repository: sqoop-sqoop2 > > > Description > ------- > > I've updated the installation guide and incoporated changes that were missing. > > > Diffs > ----- > > docs/src/site/sphinx/admin/Installation.rst 9d56875 > docs/src/site/sphinx/admin/Tools.rst fb0187a > > Diff: https://reviews.apache.org/r/40576/diff/ > > > Testing > ------- > > > Thanks, > > Jarek Cecho > >
