[
https://issues.apache.org/jira/browse/MAPREDUCE-849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12744459#action_12744459
]
Vinod K V commented on MAPREDUCE-849:
-------------------------------------
These names look a lot cleaner. +1 for the overall direction. But, we should
also think of ways to continue doing this going forward even after this issue
gets committed.
While doing this, if we can create the corresponding java.lang.String property
names, ala HADOOP-3583 , and use them everywhere, it will be real good. For
e.g.,
{code}
static final String MAPREDUCE_CLUSTER_EXAMPLE_CONFIG_PROPERTY =
"mapreduce.cluster.example.config"
{code}
Also, I think usage of strings like _mapreduce.map.max.attempts_ and
_mapreduce.jobtracker.maxtasks.per.job_ should be discouraged in favour of
_mapreduce.map.max-attempts_ and _mapreduce.jobtracker.maxtasks-per-job_
respectively. Thoughts about this?
I am assuming that configuration related to sub-components should start with a
prefix of the parent component. For e.g., _mapred.healthChecker.script.args_
will be _mapreduce.tasktracker.healthChecker.script-args_ . Right?
> Renaming of configuration property names in mapreduce
> -----------------------------------------------------
>
> Key: MAPREDUCE-849
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-849
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Reporter: Amareshwari Sriramadasu
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
>
> In-line with HDFS-531, property names in configuration files should be
> standardized in MAPREDUCE.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
