You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hbase.apache.org by "Ted Yu (JIRA)" <ji...@apache.org> on 2011/09/24 01:25:26 UTC

[jira] [Created] (HBASE-4474) Use variable length format to store the memstoreTS

Use variable length format to store the memstoreTS
--------------------------------------------------

                 Key: HBASE-4474
                 URL: https://issues.apache.org/jira/browse/HBASE-4474
             Project: HBase
          Issue Type: Sub-task
            Reporter: Ted Yu
             Fix For: 0.92.0


HBASE-4344 introduced memstoreTS for KeyValues.

The following suggestion was from Kannan:
We should consider using variable length format to store the memstoreTS on disk. Also, at the start of the flush, we can probably prune most of these timestamps to 0 since only the ones that are higher than the current read point for all active scanners need to be maintained at the fine grain level. So like often times, for a majority of the KVs, we might be able to just write a 0. And hence, storing in varying width format would be an even bigger win.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Resolved] (HBASE-4474) Use variable length format to store the memstoreTS

Posted by "Ted Yu (Resolved) (JIRA)" <ji...@apache.org>.

     [ https://issues.apache.org/jira/browse/HBASE-4474?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Ted Yu resolved HBASE-4474.
---------------------------

    Resolution: Duplicate

This has been covered in the latest patch for HBASE-2856
                
> Use variable length format to store the memstoreTS
> --------------------------------------------------
>
>                 Key: HBASE-4474
>                 URL: https://issues.apache.org/jira/browse/HBASE-4474
>             Project: HBase
>          Issue Type: Sub-task
>            Reporter: Ted Yu
>             Fix For: 0.92.0
>
>
> HBASE-4344 introduced memstoreTS for KeyValues.
> The following suggestion was from Kannan:
> We should consider using variable length format to store the memstoreTS on disk. Also, at the start of the flush, we can probably prune most of these timestamps to 0 since only the ones that are higher than the current read point for all active scanners need to be maintained at the fine grain level. So like often times, for a majority of the KVs, we might be able to just write a 0. And hence, storing in varying width format would be an even bigger win.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira