helix-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "dafu (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (HELIX-30) ZkHelixManager.carryOverPreviousCurrentState() should use a special merge logic
Date Wed, 27 Mar 2013 01:27:14 GMT

     [ https://issues.apache.org/jira/browse/HELIX-30?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

dafu resolved HELIX-30.
-----------------------

    Resolution: Fixed

https://git-wip-us.apache.org/repos/asf?p=incubator-helix.git;a=commitdiff;h=1e36a882bb8c17bfd0538b9496e8609b88482395
                
> ZkHelixManager.carryOverPreviousCurrentState() should use a special merge logic
> -------------------------------------------------------------------------------
>
>                 Key: HELIX-30
>                 URL: https://issues.apache.org/jira/browse/HELIX-30
>             Project: Apache Helix
>          Issue Type: Bug
>    Affects Versions: 0.6.0-incubating
>            Reporter: Shi Lu
>            Assignee: dafu
>             Fix For: 0.6.1-incubating
>
>
> We have a potential race condition for ZkHelixManager.carryOverPreviousCurrentState().
> It happens after addLiveInstance(), so it is possible that the participant has received
some message from controller, updated current state then the currentstate is overwritten by
the carry over.
> when we carryover, we should provide a merge logic such that only partition keys that
does not exist on the existing current state are updated. This will solve the race condition
issue.

--
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