hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Todd Lipcon (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-1923) Bulk incremental load into an existing table
Date Fri, 28 May 2010 05:43:38 GMT

    [ https://issues.apache.org/jira/browse/HBASE-1923?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12872859#action_12872859

Todd Lipcon commented on HBASE-1923:

Regarding the logseqid issue, I think I like the following solution:

- we insert no seqid at all in the HFile when we create it
- instead, we insert a special meta field that specifies that this is a bulk load result file
(perhaps give the timestamp or the MR job ID or whatever)
- change the Map<Long, StoreFile> inside HRegion to be a List<StoreFile>, making
necessary changes so that existing functionality still works (maybe need to add logseqid as
a member of storefile?)
- ensure that RS startup handles hfiles correctly when they have no seqid (in terms of log

The hard part is that last bit - where in the StoreFile list do bulk load files end up? Do
we always put them "on top"? I think with our new gets-as-scans implementation this will be
fine, but we need to think whether it will cause issues with compaction, etc. Does anyone
have some thoughts here?

> Bulk incremental load into an existing table
> --------------------------------------------
>                 Key: HBASE-1923
>                 URL: https://issues.apache.org/jira/browse/HBASE-1923
>             Project: HBase
>          Issue Type: New Feature
>          Components: client, mapred, regionserver, scripts
>    Affects Versions: 0.21.0
>            Reporter: anty.rao
>            Assignee: Todd Lipcon
>         Attachments: hbase-1923-prelim.txt
> hbase-48 is about bulk load of a new table,maybe it's more practicable to bulk load aganist
a existing table.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message