[
https://issues.apache.org/jira/browse/SAMZA-123?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14032804#comment-14032804
]
Martin Kleppmann commented on SAMZA-123:
----------------------------------------
I just had a thought: am I right in thinking that this mechanism can also be
used to implement "broadcast streams", i.e. a stream for which *all* partitions
are consumed by *all* task instances in a job? (Previously discussed
[here|http://mail-archives.apache.org/mod_mbox/samza-dev/201402.mbox/%3ccf310c72.23bee%[email protected]%3E],
for example.) It seems that this would simply require a SSPGrouper which puts
the SSPs for the broadcast stream in the set of SSPs for each of the task
instances. Is that right?
> Move topic partition grouping to the AM and generalize
> ------------------------------------------------------
>
> Key: SAMZA-123
> URL: https://issues.apache.org/jira/browse/SAMZA-123
> Project: Samza
> Issue Type: Sub-task
> Components: container
> Affects Versions: 0.6.0
> Reporter: Jakob Homan
> Assignee: Jakob Homan
> Attachments: SAMZA-123-2.patch, SAMZA-123-3.patch,
> SAMZA-123-design-doc.md, SAMZA-123-design-doc.pdf, SAMZA-123-draft.patch
>
>
> Currently the AM sends a set of all the topics and partitions to the
> container, which then groups them by partition and assigns each set to a task
> instance. By moving the grouping to the AM, we can assign arbitrary groups to
> task instances, which will allow more partitioning strategies, as discussed
> in SAMZA-71.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
