hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hemanth Yamijala (JIRA)" <j...@apache.org>
Subject [jira] Commented: (MAPREDUCE-1075) getQueue(String queue) in JobTracker would return NPE for invalid queue name
Date Mon, 30 Nov 2009 06:47:20 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-1075?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12783552#action_12783552
] 

Hemanth Yamijala commented on MAPREDUCE-1075:
---------------------------------------------

I looked at this patch. I am a bit concerned that we interpret every IOException to mean the
queue does not exist. This may not be true if there are IOExceptions due to other cases like
network issues etc. Why can't we return null in relevant APIs and interpret that to mean queue
does not exist ? This pattern is similar to what we do if a job with a given id doesn't exist.
One issue I can think about this approach is that more checks may be necessary on the client.
But I feel it is cleaner. Thoughts ?

> getQueue(String queue) in JobTracker would return NPE for invalid queue name
> ----------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-1075
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1075
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>            Reporter: V.V.Chaitanya Krishna
>            Assignee: V.V.Chaitanya Krishna
>             Fix For: 0.21.0
>
>         Attachments: MAPREDUCE-1075-1.patch, MAPREDUCE-1075-2.patch, MAPREDUCE-1075-3.patch,
MAPREDUCE-1075-4.patch, MAPREDUCE-1075-5.patch
>
>


-- 
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