[
https://issues.apache.org/jira/browse/HIVE-14964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15609380#comment-15609380
]
Zoltan Haindrich commented on HIVE-14964:
-----------------------------------------
I've also seen these various results ;)
I don't know if the test batches are stable(always the same for the same input)
or random/heuristic(changes from execution to execution dynamically)?
i've seen a few cases where an earlier test have caused a later one to
fail...these are sometimes just leftover tables...
But there are some tests which sometimes do very wierd things at the jvm
level...(like clearing all System.env variables) - this may make the whole jvm
from that point on pretty hard to be reuseable to execute other tests (in case
surefire is configured to do so - I don't know what's the current setting of
surefire/jvm forking in precommit tests, this only applies if the jvm is reused)
Particulary this jar loader have replaced the classloader - and added a few
extra classes - if there are some library clashes because of this; that may
wreck havock in some other test (again only if it executes inside the same jvm)
I will look into these in the next few days...
> Failing Test: Fix TestBeelineArgParsing tests
> ---------------------------------------------
>
> Key: HIVE-14964
> URL: https://issues.apache.org/jira/browse/HIVE-14964
> Project: Hive
> Issue Type: Sub-task
> Components: Tests
> Reporter: Jason Dere
> Assignee: Zoltan Haindrich
> Attachments: HIVE-14964.1.patch, HIVE-14964.2.patch
>
>
> Failing last several builds:
> {noformat}
> org.apache.hive.beeline.TestBeelineArgParsing.testAddLocalJar[0] 0.12
> sec 12
>
> org.apache.hive.beeline.TestBeelineArgParsing.testAddLocalJarWithoutAddDriverClazz[0]
> 29 ms 12
> org.apache.hive.beeline.TestBeelineArgParsing.testAddLocalJar[1] 42 ms
> 12
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
