You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@samza.apache.org by "Chris Riccomini (JIRA)" <ji...@apache.org> on 2014/10/24 23:21:37 UTC

[jira] [Updated] (SAMZA-444) Provide a Samza job data model for job coordinator

     [ https://issues.apache.org/jira/browse/SAMZA-444?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Chris Riccomini updated SAMZA-444:
----------------------------------
    Attachment: SAMZA-444-0.patch

Attaching patch. RB at:

https://reviews.apache.org/r/27156/

Changes:

# Wrote JobModel, ContainerModel, and TaskModel, which provide a data model for Samza jobs.
# Wrote Jackson JSON serdes for the data model.
# Wrote a JobCoordinator to build and serve the data model.
# Addressed late feedback from SAMZA-438.
# Updated SamzaContainer, ProcessJob, ThreadJob, and YARN AM to use JobCoordinator.
# Deleted a bunch of obsolete Util methods that were replaced by JobCoordinator.

Looking for feedback on data model naming/package space, etc. I kept everything in core for now, but I think the data model might eventually end up in samza-api, so it'd be good to get feedback on it. I kept the Jackson/JSON encoding totally separate from the core data model so that we don't have to suck a Jackson dependency into samza-api.

> Provide a Samza job data model for job coordinator
> --------------------------------------------------
>
>                 Key: SAMZA-444
>                 URL: https://issues.apache.org/jira/browse/SAMZA-444
>             Project: Samza
>          Issue Type: Sub-task
>          Components: container, yarn
>    Affects Versions: 0.9.0
>            Reporter: Chris Riccomini
>            Assignee: Chris Riccomini
>             Fix For: 0.9.0
>
>         Attachments: SAMZA-444-0.patch
>
>
> We currently have no data model for Samza jobs. As a result, we have really hacky code in JsonHelpers to handle (de)serializing job/container/task information between the job coordinator and its SamzaContainers.
> We should provide a proper data model for Samza jobs, which includes a Job, Container, and Task. The objects should contain relevant data that the job coordinator and SamzaContainer need. These classes should be written in Java, and use proper Jackson annotations to allow for easy (de)serialization.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)