[
https://issues.apache.org/jira/browse/HAMA-744?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13600890#comment-13600890
]
Edward J. Yoon commented on HAMA-744:
-------------------------------------
Hmm, now I remember why I didn't use messaging system.
The problem of use of messaging system is that every message must be sent with
its partition ID to destination peer, because the number of tasks is not equal
to the number of desired partitions. So, we need something like
GraphJobMessage[1].
Another one thing is, as I question before, we cannot inject partitioning phase
as a first superstep if we can't specify the different number of tasks per
supersteps.
1.
http://svn.apache.org/repos/asf/hama/trunk/graph/src/main/java/org/apache/hama/graph/GraphJobMessage.java
> Errors on generated random data
> -------------------------------
>
> Key: HAMA-744
> URL: https://issues.apache.org/jira/browse/HAMA-744
> Project: Hama
> Issue Type: Bug
> Affects Versions: 0.6.1
> Reporter: Edward J. Yoon
> Assignee: Edward J. Yoon
>
> See http://markmail.org/message/vflxymwe3ao2bsr4
> Summary:
> * Error occurs on generated data by sym-mat gen or fastgen
> * HAMA-704 requires sorted input.
> I'm not sure whether partitioned input data need to be Sorted by default. It
> should be configurable, or Should we change the VerticesInfo?
--
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
