[
https://issues.apache.org/jira/browse/HBASE-12233?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16582845#comment-16582845
]
Andrew Purtell commented on HBASE-12233:
----------------------------------------
bq. We'd try and avoid mistakes of the past regards stuff like key scheme and
wacky comparator to interpret it.
Do you mean just for new root or new root and new splittable meta both? Presume
the latter. Seems difficult to have a single-region meta facade for older
clients, if it points to a table with new rules they can't parse. It might be
better in the long run to split out meta handling into separate code, keep the
old stuff as "MetaPre3.0Compat" (but less ugly in the naming, of course), write
up a new implementation, and then switch between them given runtime detection
of whether a root region exists or not. Could be as simple as checking if a
znode to root exists. This could mostly be contained to the client, but I am
not saying refactor in the server too wouldn't be useful (it might)
> Bring back root table
> ---------------------
>
> Key: HBASE-12233
> URL: https://issues.apache.org/jira/browse/HBASE-12233
> Project: HBase
> Issue Type: Sub-task
> Reporter: Virag Kothari
> Assignee: Virag Kothari
> Priority: Major
> Attachments: HBASE-12233.patch
>
>
> First step towards splitting meta.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
