[
https://issues.apache.org/jira/browse/HADOOP-11575?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14320442#comment-14320442
]
Naganarasimha G R commented on HADOOP-11575:
--------------------------------------------
Hi [~aw],
Initially even i was thinking about the same, but in log4j terms they call it
as log itself and interface definitions also finally narrow down to
{{LogFactory.getLog(String name)}
Also in some places in the code they have used the log created from one class
name is used in other classes also . For instance Log in the Name node class is
used in many other classes and one such class being :
DatanodeManager.registerDatanode
bq. NameNode.LOG.info("BLOCK* registerDatanode: " + nodeN);
Hence kept it as "Log Name" itself
bq.Back ticks. See Federation.md and search for dfsclusterheatlth.jsp for an
example.
bq. Should we consider deprecating this servlet and provide a REST API instead?
> Daemon log documentation is misleading
> --------------------------------------
>
> Key: HADOOP-11575
> URL: https://issues.apache.org/jira/browse/HADOOP-11575
> Project: Hadoop Common
> Issue Type: Improvement
> Components: documentation
> Reporter: Jagadesh Kiran N
> Assignee: Naganarasimha G R
> Attachments: CommandsManual.html, HADOOP-11575.20150213-1.patch,
> YarnCommands.html
>
>
> a. Execute the command
> ./yarn daemonlog -setlevel xx.xx.xx.xxx:45020 ResourceManager DEBUG
> b. It is not reflecting in process logs even after performing client level
> operations
> c. Log level is not changed.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
