[
https://issues.apache.org/jira/browse/HBASE-12869?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14280801#comment-14280801
]
stack commented on HBASE-12869:
-------------------------------
Can purge vendor reference, np. Can call it something else: e.g.
"Implementation of ClusterManager where start scripts are in alternate
location". I believe the idea is to remove barriers running hbase-it on deploys
other than hand-crafted apache hadoop setups. Follow-ons I can see happening
would be ambari version (there was talk of such an animal at one time) and one
for bigtop.
> Add a ClouderaManager implementation of the ClusterManager interface
> --------------------------------------------------------------------
>
> Key: HBASE-12869
> URL: https://issues.apache.org/jira/browse/HBASE-12869
> Project: HBase
> Issue Type: New Feature
> Components: integration tests
> Reporter: Dima Spivak
> Assignee: Dima Spivak
>
> ChaosMonkey doesn't work on HBase/CDH clusters set up using Cloudera Manager
> because Cloudera Manager doesn't make use of hbase-daemon.sh (so killed
> RegionServers can't be started up again). To fix this, I want to implement
> ClouderaManagerClusterManager. I could just as well keep this internal (or
> only include it in Cloudera's distribution), but seeing as how it would open
> up more people across the community to run hbase-it, I'd like to commit it
> upstream. Anyone object?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
