[
https://issues.apache.org/jira/browse/HIVE-10950?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14639739#comment-14639739
]
Daniel Dai commented on HIVE-10950:
-----------------------------------
+1. We can currently focus a subset of the test, so that we can get most done
in a short time. We shall run the whole set afterward to find all issues. Once
the tests are stabilized, we can select a typical subset to run on daily basis,
so that we don't increase the total runtime too much.
For issue #2, this is because tests does not recover from previous failures.
Once we hit the first failure, all the tests following it will fail due to the
inconsistent cache. However, if we run a specific test, the test may pass. The
number shown does not reflect the actual test passing rate. That's why I
recreate cluster every time. Once we fixed most of test failures, we can remove
the rebuild part.
> Unit test against HBase Metastore
> ---------------------------------
>
> Key: HIVE-10950
> URL: https://issues.apache.org/jira/browse/HIVE-10950
> Project: Hive
> Issue Type: Sub-task
> Components: Metastore
> Affects Versions: hbase-metastore-branch
> Reporter: Daniel Dai
> Assignee: Daniel Dai
> Fix For: hbase-metastore-branch
>
> Attachments: HIVE-10950-1.patch, HIVE-10950-2.patch,
> HIVE-10950.3.patch, HIVE-10950.4.patch
>
>
> We need to run the entire Hive UT against HBase Metastore and make sure they
> pass.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
