[
https://issues.apache.org/jira/browse/CHUKWA-123?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12698441#action_12698441
]
Jerome Boulon commented on CHUKWA-123:
--------------------------------------
-1 I don't understand what this patch tries to solve
Also could you be more concrete.
What's made you not confident with the current implementation?
Everything is configured from one single file hadoop-metrics.properties instead
of an hidden log4.properties file embedded in chukwa-hadoop-client.jar.
You can take a look at conf/hadoop-metrics.properties.template, I've put some
info on how to use it at the same time of the Jira.
Example:
mapred.class=org.apache.hadoop.chukwa.inputtools.log4j.Log4JMetricsContext
mapred.period=60
mapred.directory=/tmp/
mapred.uuid=true
The UUID flag will add a timestamp at the end of the filename. So you need to
use it when there's more than one process, running on the same machine and
outputting the same kind of metrics like jvm or mapred.
> regression on chukwa-hadoop-client.jar
> --------------------------------------
>
> Key: CHUKWA-123
> URL: https://issues.apache.org/jira/browse/CHUKWA-123
> Project: Hadoop Chukwa
> Issue Type: Bug
> Components: input tools
> Affects Versions: 0.1.2
> Environment: Redhat 5.1, Java 6
> Reporter: Eric Yang
> Assignee: Eric Yang
> Priority: Blocker
> Fix For: 0.1.2
>
> Attachments: CHUKWA-123.patch
>
>
> There is currently no metrics file being generated by Log4jMetricsContext
> plugin.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
