helix-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Zhen Zhang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HELIX-531) Extend controller to support multi-dimensional mapping
Date Wed, 29 Oct 2014 19:00:34 GMT

    [ https://issues.apache.org/jira/browse/HELIX-531?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14188795#comment-14188795
] 

Zhen Zhang commented on HELIX-531:
----------------------------------

i think we can put multiple states together also. for example, if we want to do backup on
slave partitions, we are now using two resources: one is the actual db, and the other is a
backup job which has target resource of the actual db; e.g.

{noformat}
// db partition
MyDB_1: SLAVE

// backup task for this partition
BackupTask_1 (for MyDB_1): START
{noformat}

Using multi-dimension mapping, we can potentially do:
{noformat}
MyDB_1: SLAVE, START_BACKUP
{noformat}



> Extend controller to support multi-dimensional mapping
> ------------------------------------------------------
>
>                 Key: HELIX-531
>                 URL: https://issues.apache.org/jira/browse/HELIX-531
>             Project: Apache Helix
>          Issue Type: Sub-task
>            Reporter: Zhen Zhang
>
> Instead of calculating best-possible mappings on a single dimension, controller pipeline
should be able to calculate best-possible mappings on multiple dimensions given state-machines
on each dimension and dependency among different dimensions. 
> AC
> - Controller pipeline supports transition on multiple dimensions.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message