[
https://issues.apache.org/jira/browse/HBASE-2697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12895095#action_12895095
]
HBase Review Board commented on HBASE-2697:
-------------------------------------------
Message from: [email protected]
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://review.cloudera.org/r/484/#review646
-----------------------------------------------------------
I only got to first file and didn't even finish that... will do more later
branches/0.90_master_rewrite/BRANCH_TODO.txt
<http://review.cloudera.org/r/484/#comment2424>
Why not let it be on the heartbeat? We have to have a handler for case
where it does not make it across anyways? Or is it because there is no longer
a mechanism for taking action on heartbeat other than the passing of server
load?
branches/0.90_master_rewrite/BRANCH_TODO.txt
<http://review.cloudera.org/r/484/#comment2425>
Disable/Enable has to be sloppy and allow splits to come in. You could set
a disabling flag in zk but there'd be a race between the "point of no return",
the updating of meta that parent is offlined, and checking flag in zk.
branches/0.90_master_rewrite/BRANCH_TODO.txt
<http://review.cloudera.org/r/484/#comment2426>
Do the old ones works still?
branches/0.90_master_rewrite/BRANCH_TODO.txt
<http://review.cloudera.org/r/484/#comment2427>
Nah
- stack
> Implement new open/close logic in handlers and stop using heartbeats for
> open/close messages
> --------------------------------------------------------------------------------------------
>
> Key: HBASE-2697
> URL: https://issues.apache.org/jira/browse/HBASE-2697
> Project: HBase
> Issue Type: Sub-task
> Components: ipc, master, regionserver
> Reporter: Jonathan Gray
> Assignee: Jonathan Gray
> Priority: Critical
> Fix For: 0.90.0
>
>
> This issue is doing the meat of what HBASE-2485 is about and continues what
> was started in HBASE-2694 after some code cleanup to make life easier.
> This deals with no longer piggybacking messages from Master to RegionServers
> on heartbeat responses and instead sending direct unsolicited messages. This
> also deals with moving the open/close logic fully into handlers and removing
> the existing open/close code on both the RS and M sides. There may also be
> some changes to the master in-memory state of regions in transition. The new
> load balancer will probably be introduced with this issue but not fully
> integrated yet.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
