hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Robert Dyer (JIRA)" <j...@apache.org>
Subject [jira] [Created] (HBASE-7877) KeyPrefixRegionSplitPolicy and DelimitedKeyPrefixRegionSplitPolicy splits are not always optimal
Date Tue, 19 Feb 2013 06:19:12 GMT
Robert Dyer created HBASE-7877:
----------------------------------

             Summary: KeyPrefixRegionSplitPolicy and DelimitedKeyPrefixRegionSplitPolicy splits
are not always optimal
                 Key: HBASE-7877
                 URL: https://issues.apache.org/jira/browse/HBASE-7877
             Project: HBase
          Issue Type: Improvement
          Components: regionserver
    Affects Versions: 0.94.5, 0.96.0
            Reporter: Robert Dyer
            Priority: Minor


With KeyPrefixRegionSplitPolicy (and now DelimitedKeyPrefixRegionSplitPolicy), if a split
would break a group of keys it is modified to become the first key in the group's range. 
This is not always optimal.  If the distribution of keys are such that the group containing
the split has half the keys in the region, then no split will occur.

The best solution would be to compute both the current key group's first key (what the current
implementation does) as well as the next key group's first key and then choosing which of
the two is closest to the original split point.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message