hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chia-Ping Tsai (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-18992) Comparators passed to the Memstore's flattened segments seems to be wrong
Date Thu, 12 Oct 2017 13:24:00 GMT

    [ https://issues.apache.org/jira/browse/HBASE-18992?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16201927#comment-16201927
] 

Chia-Ping Tsai commented on HBASE-18992:
----------------------------------------

+1

bq. I was also thinking that we dont make in memory compaction for the META regions. 
The default memstore policy for META table is {{BASIC}}, because
# The default policy in {{HStore}} is {{BASIC}}
# [FSTableDescriptors|https://github.com/apache/hbase/blob/master/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FSTableDescriptors.java#L147]
don't set policy to meta table descriptor

The interesting thing is why no failure is caused by this issue? The meta data is too small
to trigger any in-memory compaction?
 

> Comparators passed to the Memstore's flattened segments seems to be wrong
> -------------------------------------------------------------------------
>
>                 Key: HBASE-18992
>                 URL: https://issues.apache.org/jira/browse/HBASE-18992
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 2.0.0-alpha-3
>            Reporter: ramkrishna.s.vasudevan
>            Assignee: ramkrishna.s.vasudevan
>             Fix For: 2.0.0-beta-1
>
>         Attachments: HBASE-18992.patch
>
>
> While doing HBASE-18945 found that the comparators passed to the CellChunkImmutableSegment
and CellArrayImmutableSegment are sometimes wrong. We pass the comparator that we got from
the Store but when doing reinitializeCellSet we always use CellComparator.COMPARATOR.
> From code I don think we are stopping the META region to also go through CompactingMemstore.
If we do that then this JIRA is invalid. But considering that is not the case we should fix
it and pass the right comparator.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message