[ https://issues.apache.org/jira/browse/SLIDER-1143?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15395783#comment-15395783 ]
Jonathan Maron commented on SLIDER-1143: ---------------------------------------- I'm wondering about the possibility of storm being installed in the parent cluster as well as running on top of yarn. Would this potentially clobber the ZK data for the cluster-based Storm installation? Is there a way to limit the purge to the data associated with the YARN/Slider based Storm instance? > Clean Intermediate data (generated while apps running),when destry the app > -------------------------------------------------------------------------- > > Key: SLIDER-1143 > URL: https://issues.apache.org/jira/browse/SLIDER-1143 > Project: Slider > Issue Type: Bug > Components: core > Affects Versions: Slider 0.90.2 > Reporter: fengyongshe > Priority: Minor > Attachments: > SLIDER-1143-clean-intermediate-data-when-destroy-apps.patch > > > When create and run the apps,eg,hbase,storm,etc, We will create a zk > directory to support the service woriking. But if we want to destroy the > unused apps,we should destroy the directroy or data generated when running. > For example: > 1)create and start a storm service, created a zk path to hold the metadata > [zk: localhost:2181(CONNECTED) 2] ls / > [services, registry, storm] > zk: localhost:2181(CONNECTED) 1] ls /storm > [workerbeats, errors, supervisors, storms, assignments] > 2) destroy the stopped storm > slider destroy demo-storm --foce > [zk: localhost:2181(CONNECTED) 1] ls /storm > [workerbeats, errors, supervisors, storms, assignments] > 3) the zk path not deleted when destroy it -- This message was sent by Atlassian JIRA (v6.3.4#6332)