[
https://issues.apache.org/jira/browse/AMBARI-11890?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14583677#comment-14583677
]
Florian Barca commented on AMBARI-11890:
----------------------------------------
The cause for the failure seems to be that the HCat Client is installed prior
to any HDFS component. Normally, HDFS comes first, installs the Hadoop stack,
sets the environment vars, and then all the other components expect to find the
environment variables set. With blueprint, the order of installation seems to
be different of the normal one.
> [WinTP2] [Test] Cluster install via blueprint fails
> ---------------------------------------------------
>
> Key: AMBARI-11890
> URL: https://issues.apache.org/jira/browse/AMBARI-11890
> Project: Ambari
> Issue Type: Bug
> Components: ambari-agent, ambari-server
> Affects Versions: 2.1.0
> Environment: Windows
> Reporter: Florian Barca
> Assignee: Florian Barca
> Priority: Critical
> Fix For: 2.1.0
>
>
> Attempting a cluster installation according to the blueprint specs results in
> recoverable failure. The services are installed ok, but some of them fail to
> start. Hive seems to be particularly sensitive. Reinstallation from the UI is
> successful.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
