hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Suresh Srinivas (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HDFS-567) Two contrib tools to facilitate searching for block history information
Date Wed, 02 Sep 2009 00:03:32 GMT

    [ https://issues.apache.org/jira/browse/HDFS-567?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12750188#action_12750188
] 

Suresh Srinivas commented on HDFS-567:
--------------------------------------

# README: Add an example for the mapreduce utility and BlockForensics utility in the README
# README: should {{sleep time between requests}} be {{sleep time between namenode query for
corrupt blocks}}.
# BlockForensics.RunHadoopCmd() - is there a significance in starting the method name with
caps?
# BlockForensics.RunHadoopCmd() - in case hadoop_home is not defined, should the command fail
by printing System.err message? Also it might be good idea to check this early in main() itself.
# BlockForensics.main() - when sleep time parse fails, a message is printed indicating the
default value is being used. Can you please print the default value in that message?
# BlocksLog and BlockForensics - could you please add comments to the class and brief description
of the functionality


> Two contrib tools to facilitate searching for block history information 
> ------------------------------------------------------------------------
>
>                 Key: HDFS-567
>                 URL: https://issues.apache.org/jira/browse/HDFS-567
>             Project: Hadoop HDFS
>          Issue Type: New Feature
>          Components: tools
>    Affects Versions: 0.21.0
>            Reporter: Bill Zeller
>            Assignee: Bill Zeller
>            Priority: Minor
>             Fix For: 0.21.0
>
>         Attachments: hdfs-567-1.patch, hdfs-567-2.patch, hdfs-567-3.patch, hdfs-567-4.patch
>
>   Original Estimate: 5h
>  Remaining Estimate: 5h
>
> Includes a java program to query the namenode for corrupt replica information at some
interval. If a corrupt replica is found, a map reduce job is launched that will search (supplied)
log files for one or more block ids. The mapred job can be used independently of the java
client program and can also be used for arbitrary text searches. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message