Mailing-List: contact issues-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hive.apache.org
Date: Tue, 1 Nov 2016 16:07:59 +0000 (UTC)
From: =?utf-8?Q?Sergio_Pe=C3=B1a_=28JIRA=29?= <jira@apache.org>
To: issues@hive.apache.org
Message-ID: <JIRA.13016925.1478014327000.145235.1478016479647@Atlassian.JIRA>
In-Reply-To: <JIRA.13016925.1478014327000@Atlassian.JIRA>
References: <JIRA.13016925.1478014327000@Atlassian.JIRA> <JIRA.13016925.1478014327147@arcas>
Subject: [jira] [Commented] (HIVE-15102) Hiveptest is killing nodes where IP
 is reused after previous node termination
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Tue, 01 Nov 2016 16:08:01 -0000


    [ https://issues.apache.org/jira/browse/HIVE-15102?page=3Dcom.atlassian=
.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D1562=
5821#comment-15625821 ]=20

Sergio Pe=C3=B1a commented on HIVE-15102:
------------------------------------

It could be. This is one of 2 issues found. The other is due to a possible =
thread leak on tomcat. However, the thread leak has been there for a while =
without causing other issues. I need to dig more on this one, and try to re=
produce it.

For more context, we recently moved the ptest server from Amazon to Google =
Compute. Amazon was giving us different IPs all the time, but GCE is giving=
 us only IPs on the range of 10.0.0.x; so only 255 possible values, and reu=
sing those is happening too often.

> Hiveptest is killing nodes where IP is reused after previous node termina=
tion
> -------------------------------------------------------------------------=
----
>
>                 Key: HIVE-15102
>                 URL: https://issues.apache.org/jira/browse/HIVE-15102
>             Project: Hive
>          Issue Type: Bug
>          Components: Hive
>    Affects Versions: 2.2.0
>            Reporter: Sergio Pe=C3=B1a
>            Assignee: Sergio Pe=C3=B1a
>         Attachments: HIVE-15102.1.patch
>
>
> NO PRECOMMIT TESTS
> The Hiveptest framework has a background thread that runs every hour, and=
 attempts to kill zombie nodes that are not being used by the test executio=
n anymore.=20
> These killed nodes are kept in a list of terminated nodes, and next time =
the background thread is executed, it will attempt to kill all those nodes =
again because Hiveptest consider them as zombie nodes.
> The problem is that cloud providers can give you the same IP numbers for =
new nodes, and when the background thread runs, it will kill those nodes th=
at may still be in used by Hiveptest.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)