[
https://issues.apache.org/jira/browse/HBASE-11288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17111521#comment-17111521
]
Andrew Kyle Purtell commented on HBASE-11288:
---------------------------------------------
bq. IMHO ideally the registry should've been served out of the regionservers.
No, because the registry addresses must be explicitly discoverable, at a
manageable scale for that purpose. By design we already had to publish the
location of ZK quorum peers, so we replace this with a requirement to publish
the master locations. Nobody's service discovery logic needs to change. In both
cases, a string containing a short list of network endpoints. And like before,
the client can choose any at random to contact.
Number of masters will be up to orders of magnitude less than regionservers. So
we can't directly serve the registry data out of regionservers and call it a
day. Masters will serve as the front end for the registry, at least. On the
other hand, from the client perspective, we don't care where registry data is
located. HBASE-18095 was concerned with publishing the location of meta, which
is to date still managed as znode data. Future additions to configuration data
served by the registry may have their data hosted somewhere else. As long as
the masters know where to go and what to do to serve answers to client
inquiries, it's fine.
> Splittable Meta
> ---------------
>
> Key: HBASE-11288
> URL: https://issues.apache.org/jira/browse/HBASE-11288
> Project: HBase
> Issue Type: Umbrella
> Components: meta
> Reporter: Francis Christopher Liu
> Assignee: Francis Christopher Liu
> Priority: Major
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
