pig-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel Dai (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (PIG-1865) BinStorage/PigStorageSchema cannot load data from a different namenode
Date Wed, 20 Apr 2011 23:51:05 GMT

    [ https://issues.apache.org/jira/browse/PIG-1865?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13022550#comment-13022550
] 

Daniel Dai commented on PIG-1865:
---------------------------------

There is no test case added. It requires two clusters and is hard to do in unit test. Manually
tested and it works.

> BinStorage/PigStorageSchema cannot load data from a different namenode
> ----------------------------------------------------------------------
>
>                 Key: PIG-1865
>                 URL: https://issues.apache.org/jira/browse/PIG-1865
>             Project: Pig
>          Issue Type: Bug
>    Affects Versions: 0.7.0, 0.8.0, 0.9.0
>            Reporter: Vivek Padmanabhan
>            Assignee: Daniel Dai
>             Fix For: 0.9.0
>
>         Attachments: PIG-1865-1.patch
>
>
> BinStorage/PigStorageSchema cannot load data from a different namenode. The main reason
for this is that, in the getSchema method , they use org.apache.pig.impl.io.FileLocalizer
to check whether the exists, but the filesystem in HDataStorage refers to the natively configured
dfs.
> The test case is simple :
> a = load 'hdfs://<nn2>/input' using BinStorage();
> dump a;
> Here if I specify -Dmapreduce.job.hdfs-servers, it should have worked , by pig still
takes the fs from fs.default.name so to make it work i had to override  fs.default.name in
pig command line.
> Raising this as a bug since the same scenario works with PigStorage.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message