[
https://issues.apache.org/jira/browse/HBASE-11059?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14025868#comment-14025868
]
Jeffrey Zhong commented on HBASE-11059:
---------------------------------------
{quote}
Tell us more please. Do it in HBASE-11165?
{quote}
The basic idea in high-level is that master only coordinates the assignment
requests without really being involved in the actual assignment work flow(it's
better that if we have FATE like library to back it up) and we can do it later
if we're agreed.
1) Master create a new state:"assigning", reason and new destination RS for a
region in meta
2) Master sends an assign request to the new RS
3) The new RS( and the current RS hosting the region) will assume the duty to
do the whole assignment and update meta states along with assignment including:
prepare-closing, closing, closed, opening & opened. This needs region server to
talk to each other in order to communicate info that current ZK notifications
carry and report status back to master by RPC if it fails the assignment.
4) Master monitors un-assigned & failed open regions periodically
Thanks.
> ZK-less region assignment
> -------------------------
>
> Key: HBASE-11059
> URL: https://issues.apache.org/jira/browse/HBASE-11059
> Project: HBase
> Issue Type: Improvement
> Components: master, Region Assignment
> Reporter: Jimmy Xiang
> Assignee: Jimmy Xiang
> Fix For: 0.99.0
>
> Attachments: hbase-11059.patch, hbase-11059_v2.1.patch,
> hbase-11059_v2.patch, zk-less_am.pdf
>
>
> It seems that most people don't like region assignment with ZK (HBASE-5487),
> which causes many uncertainties. This jira is to support ZK-less region
> assignment. We need to make sure this patch doesn't break backward
> compatibility/rolling upgrade.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
