[
https://issues.apache.org/jira/browse/HADOOP-3271?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eric Yang resolved HADOOP-3271.
-------------------------------
Resolution: Invalid
The pure parser function still works by using local file system. Job
Information can be summarized from job log, the master index feature is not
necessary.
> JobHistory.log and DefaultJobHistory.MasterIndex are missing in 0.17 dev
> ------------------------------------------------------------------------
>
> Key: HADOOP-3271
> URL: https://issues.apache.org/jira/browse/HADOOP-3271
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Environment: Red Hat Enterprise Linux AS release 4 (Nahant Update 5),
> JVM 1.5.0_15, hadoop 0.17.0-dev-version 3
> Reporter: Eric Yang
> Priority: Blocker
>
> There used to be JobHistory.log (MasterIndex in Default HistoryParser class)
> in 0.15.3 to 0.16.2 which provides the high level overview of the job
> history. For HADOOP-2178, this information is no longer being provided by
> the DefaultHistoryParser class. In addition, the parser class can only read
> job log files from dfs, not from an input stream.
> For performance monitoring, we have programs that depends on the pure parser
> class and JobHistory.log to extract job information. Please continue to
> provide a pure parser class and equivalent function of the JobHistory.log.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
