[
https://issues.apache.org/jira/browse/HBASE-9696?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13793033#comment-13793033
]
stack commented on HBASE-9696:
------------------------------
I +1'd 3.v3 over in RB as long as follow up issue is filed to address
outstanding review comments.
The differences in v3 are minor. My review comments from last night have been
answered or are nits to address in follow-on.
This patch makes us more resilient when move/split or move/merge are going on.
It is an incompatible change.
We have two no-dataloss runs over last night which is better than what we've
had at any time previous. The last verify actually failed but the second
verify passed -- no dataloss. Will work on making the first verify pass but
want to get this incompat change before RC.
[~enis] or [~jeffreyz] -- ok w/ you fellas if this gets committed?
> Master recovery ignores online merge znode
> ------------------------------------------
>
> Key: HBASE-9696
> URL: https://issues.apache.org/jira/browse/HBASE-9696
> Project: HBase
> Issue Type: Bug
> Components: master, Region Assignment
> Reporter: Jimmy Xiang
> Assignee: Jimmy Xiang
> Fix For: 0.98.0, 0.96.1
>
> Attachments: 0.96-9696_v3.2.patch, 0.96-9696_v3.3.patch,
> trunk-9696.patch, trunk-9696_v2.1.patch, trunk-9696_v2.patch,
> trunk-9696_v3.1.patch, trunk-9696_v3.2.patch, trunk-9696_v3.3.patch,
> trunk-9696_v3.patch, trunk-9696_v3.patch
>
>
> The online merge znode uses the new region to be created. When master
> restarts, the new region is still unknown if the merging is not completed.
> Therefore the znode is ignored, which should not. That means the two merging
> regions could be moved around. This could cause some data loss if we are not
> luck.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
