[
https://issues.apache.org/jira/browse/BIGTOP-1106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13786297#comment-13786297
]
Roman Shaposhnik commented on BIGTOP-1106:
------------------------------------------
[~cos] agreed. Ideally we'd want to run 3 ZK quorum so that our tests
deployments are close to what's running in production. Unfortunately, our
puppet code is not sophisticated enough to let one easily implement that type
of topology.
That said, I really think that for the type of testing we do -- running one ZK
on a head node is more than enough.
> update our puppet deployment topology
> -------------------------------------
>
> Key: BIGTOP-1106
> URL: https://issues.apache.org/jira/browse/BIGTOP-1106
> Project: Bigtop
> Issue Type: Improvement
> Components: Deployment
> Affects Versions: 0.7.0
> Reporter: Roman Shaposhnik
> Assignee: Roman Shaposhnik
> Fix For: 0.7.0
>
> Attachments:
> 0003-BIGTOP-1106.-update-our-puppet-deployment-topology.patch
>
>
> Currently, our puppet deployment topology is putting a log of services on the
> hadoop_head_node. This puts unnecessary resource stress on it. I propose we
> start deploying a few services onto the gateway node itself.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
