[
https://issues.apache.org/jira/browse/HAMA-515?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13221954#comment-13221954
]
Thomas Jungblut commented on HAMA-515:
--------------------------------------
bq. If this is for alternative barrier sync issue, probably you would like to
consider embedding a failure detector inside the bsp master so that master can
determine the status of GroomServer, which in turns harnesses bsp tasks.
No. It is basically to make it easier to share information between supersteps
without messaging.
For example my mindist search in graphs
(http://codingwiththomas.blogspot.com/2011/04/graph-exploration-using-apache-hama-and.html)
can save a sync when counters are synced and immediately accessible after
sync() in every task.
This can help us with the graph api, and there are several other things that
makes it easier.
This is not about fault tolerance, it is just about moving counters to be
handled with zookeeper.
> Counter should be synchronized by sync()
> ----------------------------------------
>
> Key: HAMA-515
> URL: https://issues.apache.org/jira/browse/HAMA-515
> Project: Hama
> Issue Type: New Feature
> Affects Versions: 0.4.0
> Reporter: Thomas Jungblut
> Priority: Minor
> Fix For: 0.5.0
>
>
> We should synchronize the counters in all tasks after a sync() call.
> Then someone can use it for flow control. E.G. graph related algorithms, to
> get the number of globally updated vertices.
> Two options to solve this:
> - Sync the whole stuff over RPC from the BSPMaster
> - Use Zookeeper for counter handling
--
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
