[jira] Commented: (WHIRR-177) [HBase] Support embedded ZooKeeper

["Andrew Purtell (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/WHIRR-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12975147#action_12975147 ] 

Andrew Purtell commented on WHIRR-177:
--------------------------------------

Thinking when hacking the bash scripts was the cluster is not HA (no multimaster, no clustered namenode, etc.) so you can just colocate a single ZK instance on the HBase master node. If there were HA features, then it might make sense to choose a subset of slaves at random, on some kind of log scale: ~10 slaves = 3 ZK nodes, ~100 slaves = 5 ZK nodes, ~1000 slaves = 7 or 9 ZK nodes. 

> [HBase] Support embedded ZooKeeper
> ----------------------------------
>
>                 Key: WHIRR-177
>                 URL: https://issues.apache.org/jira/browse/WHIRR-177
>             Project: Whirr
>          Issue Type: Improvement
>    Affects Versions: 0.3.0
>            Reporter: Lars George
>            Assignee: Lars George
>             Fix For: 0.3.0
>
>
> This can be achieved like so:
> * Check if there is no "zk" role in the template -> set HBaseManagesZK = true
> * Emit a zoo.cfg from the HBase post-configure script next to the hbase-site.xml
> * Set the option in hbase-env.sh to indicate that HBase manages ZK:
> {code}HBASE_MANAGES_ZK=true{code}
> Question is how many nodes run ZK? All slaves? A subset?

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.