[
https://issues.apache.org/jira/browse/AMBARI-17111?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15320750#comment-15320750
]
Hudson commented on AMBARI-17111:
---------------------------------
FAILURE: Integrated in Ambari-trunk-Commit #5033 (See
[https://builds.apache.org/job/Ambari-trunk-Commit/5033/])
AMBARI-17111 ATS needs a dynamic dependency on Tez and Spark (dsen) (dsen:
[http://git-wip-us.apache.org/repos/asf?p=ambari.git&a=commit&h=4bfeb41d1ef777f9f65fc6b65e21b9ab541a5264])
* ambari-server/src/main/resources/stacks/HDP/2.3/services/YARN/metainfo.xml
> ATS needs a dynamic dependency on Tez and Spark
> -----------------------------------------------
>
> Key: AMBARI-17111
> URL: https://issues.apache.org/jira/browse/AMBARI-17111
> Project: Ambari
> Issue Type: Bug
> Components: stacks
> Affects Versions: 2.2.0
> Reporter: Dmytro Sen
> Assignee: Dmytro Sen
> Priority: Blocker
> Fix For: 2.4.0
>
> Attachments: AMBARI-17111.patch
>
>
> Today, ATS can be installed without Tez, which is fine.
> However, when you add Tez, the client also needs to be installed on the host
> that has ATS.
> Without it, the Tez View will not work and be unable to display detailed
> information.
> A similar problem exists for Spark, in which YARN shouldn't require Spark,
> but installing Spark should require reconfiguring YARN and installing Spark
> client on the ATS box.
> This has existed from the time that Spark was added in Ambari 2.2.
> STR:
> Install a 3 node cluster with HDP 2.3 or higher.
> Install ATS on host 1
> After installation, add Tez service, and only install the Tez client on host 3
> The ATS host should also install Tez client.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
