[
https://issues.apache.org/jira/browse/HADOOP-3579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12606832#action_12606832
]
Steve Loughran commented on HADOOP-3579:
----------------------------------------
I see why you'd like a group/instance model for this particular part of the
deployment (the queues), but I'm thinking of a generic hierarchical view, where
we'd have other sections for, say, hostname->IPAddress mappings, or even the
deployment parameters of a heartbeat MapReduce job you'd feed through
tasktrackers just to keep an eye on their health.
Equally importantly, I'm thinking of what to put in as a back end above and
beyond an XML format. Whether its JSON, LDAP or something else, a simple
hierarchy with no notion of type and no duplicates between child section
names and child properties wold be good. Its a prequisite to doing cross
references with notation such as ../queues/queue1/timeout and
/hosts/test/server and avoiding full XPath-like expressions such as:
/[EMAIL PROTECTED]"queues"]/[EMAIL PROTECTED]"queue"]/[EMAIL
PROTECTED]"timeout"/
While such expressions are wonderful in XSL, they are not what you want to
reliably refer to other bits in the tree.
> Support property groups in Hadoop configuration
> -----------------------------------------------
>
> Key: HADOOP-3579
> URL: https://issues.apache.org/jira/browse/HADOOP-3579
> Project: Hadoop Core
> Issue Type: Improvement
> Components: conf
> Reporter: Hemanth Yamijala
>
> Hadoop configuration is currently a list of key and value pairs. There are
> some use-cases to support configuring groups of related properties. There
> could also be multiple instances of such groups. The issue is for adding
> support for such configuration.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
