[
https://issues.apache.org/jira/browse/AMBARI-15449?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Alexander Denissov updated AMBARI-15449:
----------------------------------------
Attachment: AMBARI-15449.branch_2_2.patch
patch applies to trunk as well
> HAWQ hdfs-client / output.replace-datanode-on-failure should be set to true
> by default
> --------------------------------------------------------------------------------------
>
> Key: AMBARI-15449
> URL: https://issues.apache.org/jira/browse/AMBARI-15449
> Project: Ambari
> Issue Type: Bug
> Reporter: Alexander Denissov
> Assignee: Alexander Denissov
> Priority: Minor
> Fix For: 2.2.2
>
> Attachments: AMBARI-15449.branch_2_2.patch
>
>
> On large cluster, replace-datanode-on-failure should be set to true, but on
> small clusters (developers environment or testing environment), it should be
> set to false, otherwise, if datanodes are overloaded, it will report error.
> This is the reason it was set to false by default earlier.
> Ambari should set it to true when cluster size > 4, otherwise set it to false.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
