hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ishan Chhabra (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-9950) Row level replication
Date Tue, 12 Nov 2013 00:11:19 GMT

    [ https://issues.apache.org/jira/browse/HBASE-9950?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13819629#comment-13819629
] 

Ishan Chhabra commented on HBASE-9950:
--------------------------------------

[~stack] not sure yet. Since there is no notion of row level data in hbase storage, I would
have to create some special KVs that are stored for the row, which sounds very hacky.

[~apurtell] Replication scope is defined at the CF level, so I don't think Ill be able to
use it. I do need to plug in custom replication policy though if this is not a core feature.
There are no observers for replication, are there?

> Row level replication
> ---------------------
>
>                 Key: HBASE-9950
>                 URL: https://issues.apache.org/jira/browse/HBASE-9950
>             Project: HBase
>          Issue Type: Bug
>          Components: Replication
>            Reporter: Ishan Chhabra
>            Priority: Minor
>
> We have a replication setup with the same table and column family being present in multiple
data centers. Currently, all of them have exactly the same data, but each cluster doesn't
need all the data. Rows need to be present in only x out of the total y clusters. This information
varies at the row level and thus more granular replication cannot be achieved by setting up
cluster level replication. 
> Adding row level replication should solve this.



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message