[
https://issues.apache.org/jira/browse/HBASE-24916?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17186617#comment-17186617
]
Wellington Chevreuil commented on HBASE-24916:
----------------------------------------------
{quote}
Suppose a cluster with two tables t1 and t2.
Table t1 with regions r1, r2, r3 and table t2 with regions s1,s2,s3
When a region hole is created by deleting s3 then hole has the region pair as
<s2, RegionInfo#UNDEFINED>
But when a region hole is created by deleting s1 then hole has the region pair
as <r3,s2>. Here both regions belong to different tables
This seems wrong. This hole's region pair should be <RegionInfo#UNDEFINED, s2>
{quote}
Thanks for the explanation, now it's clear to me. Apparently, the meta scan is
not differentiating regions for different tables.
> Region hole contains wrong regions pair when hole is created by first region
> deletion
> -------------------------------------------------------------------------------------
>
> Key: HBASE-24916
> URL: https://issues.apache.org/jira/browse/HBASE-24916
> Project: HBase
> Issue Type: Bug
> Components: hbck2
> Affects Versions: 2.2.3
> Reporter: Ajeet Rai
> Assignee: Mohammad Arshad
> Priority: Major
>
> HBCK2 Holes: Wrong region details in HBCK report ,if a region is missing in
> meta and start key is empty
> for hole, it reports two region where hole is present.
> Scenario where first region is missing for eg: in below scenario r1 is
> missing in meta
> start key end key region
> empty 10 r1
> 10 20 r2
> 20 30 r2
> 30 empty r3
>
> In our test hbck report contains two region where first region belongs to
> different table and second region is r2.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
