Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Message-ID: <28556016.32331277353729987.JavaMail.jira@thor>
Date: Thu, 24 Jun 2010 00:28:49 -0400 (EDT)
From: "Jonathan Gray (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Subject: [jira] Commented: (HBASE-2700) Handle master failover for regions
 in transition
In-Reply-To: <27617847.33701276033580816.JavaMail.jira@thor>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HBASE-2700?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12882032#action_12882032 ] 

Jonathan Gray commented on HBASE-2700:
--------------------------------------

Context of above is just a quick first go at what a master will do when it is a failed-over-to master.

The ZK-based region transitions definitely break with the BT paper.  This direction came out of discussions about the best way to persist master state over in HBASE-2485.

I'm generally very in favor of moving the master away from the basescanner/heartbeat way of doing things towards zk and rpc.  Assignment is untenable in the current style and master failover does not currently work.  This design makes this stuff a lot easier imo, not to mention faster.

We can address specifics as patches come in, but in general is there opposition to the design of using zk for regions in transition and assignment rather than the basescanner and heartbeats?

> Handle master failover for regions in transition
> ------------------------------------------------
>
>                 Key: HBASE-2700
>                 URL: https://issues.apache.org/jira/browse/HBASE-2700
>             Project: HBase
>          Issue Type: Sub-task
>          Components: master, zookeeper
>            Reporter: Jonathan Gray
>            Priority: Critical
>             Fix For: 0.21.0
>
>
> To this point in HBASE-2692 tasks we have moved everything for regions in transition into ZK, but we have not fully handled the master failover case.  This is to deal with that and to write tests for it.

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