Mailing-List: contact common-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: common-issues@hadoop.apache.org
Date: Fri, 22 Feb 2013 19:48:15 +0000 (UTC)
From: "Aaron T. Myers (JIRA)" <jira@apache.org>
To: common-issues@hadoop.apache.org
Message-ID: <JIRA.12515596.1311817009607.325104.1361562495522@arcas>
In-Reply-To: <JIRA.12515596.1311817009607@arcas>
References: <JIRA.12515596.1311817009607@arcas>
Subject: [jira] [Commented] (HADOOP-7487) DF should throw a more reasonable
 exception when mount cannot be determined
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HADOOP-7487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13584592#comment-13584592 ] 

Aaron T. Myers commented on HADOOP-7487:
----------------------------------------

Patch looks pretty good to me. One suggestion: In the places in the test where you catch expected exceptions, use {{GenericTestUtils#assertExceptionContains}} to ensure that you're catching the exception you expect.
                
> DF should throw a more reasonable exception when mount cannot be determined
> ---------------------------------------------------------------------------
>
>                 Key: HADOOP-7487
>                 URL: https://issues.apache.org/jira/browse/HADOOP-7487
>             Project: Hadoop Common
>          Issue Type: Bug
>          Components: fs
>    Affects Versions: 0.23.0
>            Reporter: Todd Lipcon
>            Assignee: Andrew Wang
>              Labels: noob
>         Attachments: hadoop-7487-1.patch
>
>
> Currently, when using the DF class to determine the mount corresponding to a given directory, it will throw the generic exception "Expecting a line not the end of stream" if it can't determine the mount (for example if the directory doesn't exist).
> This error message should be improved in several ways:
> # If the dir to check doesn't exist, we can see that before even execing df, and throw a better exception (or behave better by chopping path components until it exists)
> # Rather than parsing the lines out of df's stdout, collect the whole output, and then parse. So, if df returns a non-zero exit code, we can avoid trying to parse the empty result
> # If there's a success exit code, and we still can't parse it (eg incompatible OS), we should include the unparseable line in the exception message.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira