You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hive.apache.org by "Vihang Karajgaonkar (JIRA)" <ji...@apache.org> on 2018/12/03 20:10:00 UTC

[jira] [Commented] (HIVE-20997) Make Druid Cluster start on random ports.

    [ https://issues.apache.org/jira/browse/HIVE-20997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16707727#comment-16707727 ] 

Vihang Karajgaonkar commented on HIVE-20997:
--------------------------------------------

Thanks [~bslim] for taking this up. Not sure if you already planned for this but we need not-only random ports but also a retry if the druid cluster doesn't come up with port busy errors since it could happen that by the time the druid cluster tries to start on the random port, it is used by some other process like miniHS2 or metastore.

> Make Druid Cluster start on random ports.
> -----------------------------------------
>
>                 Key: HIVE-20997
>                 URL: https://issues.apache.org/jira/browse/HIVE-20997
>             Project: Hive
>          Issue Type: Sub-task
>            Reporter: slim bouguerra
>            Assignee: slim bouguerra
>            Priority: Major
>
> As of now Druid Tests will run in a Single batch.
> To avoid timeouts we need to support batching of tests.
> As suggested by [~vihangk1] it will be better to start the Druid tests setups on a totally random ports.
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)