[
https://issues.apache.org/jira/browse/BIGTOP-1339?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14059836#comment-14059836
]
Mark Grover commented on BIGTOP-1339:
-------------------------------------
I am sorry, [~rguo], I am not able to understand much of the proposals. I think
I will ask the same question [~rvs] asked before as to what is the problem you
are trying to solve?
And, your answer was:
{quote}
1. pseudo configuration in bigtop components is not consistent.
hadoop includes pseudo configuration and other components don't include pseudo
configuration.
{quote}
The terminology pseudo, as far as I know only applies to Hadoop. Because Hadoop
is the project that can be configured to run on multiple nodes (fully
distributed mode) or one node (pseudo distributed mode). Hive Server, Hive
Server 2 for example just run on one node. I am not sure what pseudo would
symbolize there. If you are referring to creating a new package that delivers
hive configuration that would work with Hadoop configured in a
pseudo-distributed mode (e.g. that config would set the namenode location to
localhost in hive-site.xml), that's definitely a reasonable request. But if I
recall correctly, installing hive and using it's default configuration along
with installation of hadoop-conf-pseudo makes hive (and hadoop) work out of the
box already. I have used that exact same idea in many of my demos of Bigtop in
the past (see
[here|https://github.com/markgrover/oscon-bigtop/blob/master/demo-setup.sh#L53],
for example). Did you find it to the contrary?
{quote}
2. It is easy to run basic/smoke test if pseudo configuration is default
configuration
{quote}
Agreed, but only for testing, we can't make pseudo configuration the default
for any component because there are other use-cases.
{quote}
3. pseudo configuration is not product configuration.
{quote}
Not sure what this means.
{quote}
4. It is easy for hadoop newbies to study hadoop. for example:
a. rpm -ivh *.rpm
b. run different examples like hadoop, hbase, oozie.
{quote}
I agree 100% That's why we have hadoop pseudo-distributed package. Hive, like I
said, before, with default configuration, works out of the box. I can't speak
for hbase or oozie.
If we need more convenience packages for configuration of hive/hbase/oozie to
work with Hadoop in pseudo-distributed configuration, I think that's a
reasonable request and if you'd like to contribute to create those additional
packages, the community would very much appreciate it.
> hadoop pseudo configuration can be default configuration
> --------------------------------------------------------
>
> Key: BIGTOP-1339
> URL: https://issues.apache.org/jira/browse/BIGTOP-1339
> Project: Bigtop
> Issue Type: Improvement
> Components: RPM
> Reporter: Guo Ruijing
>
> in existing implementation, hadoop-conf-pseudo-2.2.0.x86_64.rpm is to provide
> hadoop pseudo configuration. we may take hadoop pseudo configuration as
> hadoop default configuration and move hadoop pseudo configuration to
> core/hdfs/yarn rpm.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
