giraph-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Maja Kabiljo (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (GIRAPH-297) Checkpointing on master is done one superstep later
Date Mon, 13 Aug 2012 09:03:37 GMT

    [ https://issues.apache.org/jira/browse/GIRAPH-297?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13433011#comment-13433011
] 

Maja Kabiljo commented on GIRAPH-297:
-------------------------------------

This also takes me back to the discussion about whether we should have master.compute() executed
before or after vertices (https://issues.apache.org/jira/browse/GIRAPH-259?focusedCommentId=13427285&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13427285).
In the BspServiceMaster.coordinateSuperstep() we first wait for workers to finish superstep
X, and then artificially increase the superstep and call master.compute() for superstep X+1,
after which we say we are done with superstep X. So the decision of compute order and our
implementation are inconsistent. Jan, can you give some comment on this?
                
> Checkpointing on master is done one superstep later
> ---------------------------------------------------
>
>                 Key: GIRAPH-297
>                 URL: https://issues.apache.org/jira/browse/GIRAPH-297
>             Project: Giraph
>          Issue Type: Bug
>            Reporter: Maja Kabiljo
>            Assignee: Maja Kabiljo
>
> On workers we store checkpoint X before compute() for superstep X are executed. On master
we do it after those compute() are executed and after master.compute() for superstep X+1.

--
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

        

Mime
View raw message