[
https://issues.apache.org/jira/browse/HIVE-14772?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15530892#comment-15530892
]
Anbu Cheeralan commented on HIVE-14772:
---------------------------------------
https://github.com/apache/hive/commit/ab605910e45cfa65b9895399f09194801f2cc091
Is this the solution that fixed the issue in master branch?
> NPE when MSCK REPAIR
> --------------------
>
> Key: HIVE-14772
> URL: https://issues.apache.org/jira/browse/HIVE-14772
> Project: Hive
> Issue Type: Bug
> Components: Hive
> Affects Versions: 2.1.0
> Environment: HiveRunner on OSX Yosemite
> Reporter: Per Ullberg
>
> HiveMetaStoreChecker throws NullPointerException when doing a MSCK REPAIR
> TABLE.
> The bug is here:
> {code}
> ...
> 18 package org.apache.hadoop.hive.ql.metadata;
> ...
> 58 public class HiveMetaStoreChecker {
> ...
> 408 if (!directoryFound) {
> 409 allDirs.put(path, null);
> 410 }
> ...
> {code}
> allDirs is a ConcurrentHashMap and those does not allow either key or value
> to be null.
> I found the bug while trying to port https://github.com/klarna/HiveRunner to
> Hive 2.1.0
> Implemented explicit test case that exposes the bug here:
> https://github.com/klarna/HiveRunner/blob/hive-2.1.0-NPE-at-msck-repair/src/test/java/com/klarna/hiverunner/MSCKRepairNPE.java
> Reproduce by cloning branch
> https://github.com/klarna/HiveRunner/tree/hive-2.1.0-NPE-at-msck-repair
> and run
> {code} mvn -Dtest=MSCKRepairNPE clean test{code}
> ( Does not work on windows :( )
> Looks like this email thread talks about the same issue:
> http://user.hive.apache.narkive.com/ETOpbKk5/msck-repair-table-and-hive-v2-1-0
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
