[
https://issues.apache.org/jira/browse/OOZIE-1880?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dénes Bodó resolved OOZIE-1880.
-------------------------------
Resolution: Cannot Reproduce
This ticket seems abandoned. If you feel that this still an issue in Oozie 5
then please reopen.
> Change of NameNode DNS entry not picked up by oozie
> ---------------------------------------------------
>
> Key: OOZIE-1880
> URL: https://issues.apache.org/jira/browse/OOZIE-1880
> Project: Oozie
> Issue Type: Bug
> Affects Versions: 4.0.1
> Environment: Amazon OS, CentOS, Amazon EMR Hadoop (Apache Hadoop 2.2)
> Reporter: Corby Wilson
> Priority: Minor
>
> We run dynamic clusters. I.e. we terminate the cluster after a run.
> But we keep the oozie server up and running so we can get to the logs for
> auditing and splunk ingestion.
> When a new cluster is started we update a DNS server with the new IP address
> of the head node which contains the NameNode and YARN server.
> Sometimes the oozie server seems to cache the old IP address and when a new
> workflow is submitted, the old IP is used and we get a connection timeout
> problem,
> To fix this we have to restart the oozie server.
> Here's the kicker, this behavior seems sporadic. Sometimes the oozie server
> picks up the new IP address and sometimes it doesn't,
> I'm available for more info,
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
