[
https://issues.apache.org/jira/browse/HBASE-22941?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16934030#comment-16934030
]
Duo Zhang commented on HBASE-22941:
-----------------------------------
OK found another problem when backporting to branch-2.1, seems we have
HBASE-22777 in place, but in MasterRpcServices.mergeTableRegions we still have
a assert which assume that there are only two regions to merge...
> MetaTableAccessor.getMergeRegions() returns parent regions in random order
> --------------------------------------------------------------------------
>
> Key: HBASE-22941
> URL: https://issues.apache.org/jira/browse/HBASE-22941
> Project: HBase
> Issue Type: Bug
> Affects Versions: master
> Reporter: István Tóth
> Assignee: István Tóth
> Priority: Major
> Fix For: 3.0.0, 2.3.0, 2.2.1, 2.1.7
>
>
> The multi-region merge patch has changed the behaviour of the
> MetaTableAccessor.getMergeRegions() method. Earlier, the parent regions were
> returned in the same order that they were passed to mergeRegion().
> The patch has changed this to return the regions in random order. To be more
> precise, the parent regions are stored in random order in Meta.
> I propose storing and returning the parent regions in ascending order.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
