hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joep Rottinghuis (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HADOOP-7423) Document topology script requirements
Date Fri, 24 Jun 2011 05:12:47 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-7423?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13054241#comment-13054241
] 

Joep Rottinghuis commented on HADOOP-7423:
------------------------------------------

Isn't it true that even with IP addresses used in the topology script, one still has to make
sure that Hostname=>IP=>Hostname is always consistent?
We had a situation where some nodes were inserted in each rack all and the numeric part of
the names of the data nodes shifted by a few numbers (the domain also changed slighty). Even
though the IP addresses stayed the same, for some of the nodes an nslookup resulted in multiple
entries. We got some weird behavior.
In a brief glance through the code there seem to be several places where IP is mapped to name
and name is mapped to IP, or am I wrong in this?

> Document topology script requirements
> -------------------------------------
>
>                 Key: HADOOP-7423
>                 URL: https://issues.apache.org/jira/browse/HADOOP-7423
>             Project: Hadoop Common
>          Issue Type: Improvement
>          Components: documentation
>            Reporter: Eli Collins
>              Labels: newbie
>
> The topology script documentation is cluster_setup.xml is unclear. The topology script:
> # Only needs to handle IP addresses (not hostnames)
> # Needs to handle multiple arguments for caching to work effectively
> We should check in an example script or include an example one in the docs.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message