[
https://issues.apache.org/jira/browse/KYLIN-4630?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17159659#comment-17159659
]
Shao Feng Shi commented on KYLIN-4630:
--------------------------------------
Beeline uses JDBC to connect with Hive server, not Kerberos. The third step is
to submit a MR job, not Hive, so it is different with the first 2 steps. You
can check that by running a word count MR job on Kylin node, to see whether the
kerberos has been configured properly.
> 构建cube的第三步骤Extract Fact Table Distinct Columns出现kerberos错误
> ----------------------------------------------------------
>
> Key: KYLIN-4630
> URL: https://issues.apache.org/jira/browse/KYLIN-4630
> Project: Kylin
> Issue Type: Bug
> Components: Job Engine
> Affects Versions: v3.0.0
> Reporter: xiaohuihui
> Priority: Major
> Labels: build
> Attachments: 2a4d463f478d998f14267f87e8fadb2.jpg,
> 6f90d9e0840e5574911bdb0116103c5.jpg, a02586af627d26c70a09ebefcb6c944.jpg,
> acdc89f6bdf23e397a4a2995582668e.jpg
>
>
> 在构建Cube的时候，第一步出现Peer indicated failure:GSS initiate
> fauled错误，然后我配置了kylin.properties中的Hive
> client由默认的cli更改为beeline方式连接，此时第一步通过，可以访问Hive数据库与创建临时中间表。但是到第三步时又出现Peer
> indicated failure:GSS initiate fauled错误，不明白的是既然第一步这个问题已经解决了，为什么第三部还会出现这个问题。
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
