You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@nifi.apache.org by "Matt Burgess (Jira)" <ji...@apache.org> on 2021/10/27 17:48:00 UTC

[jira] [Updated] (NIFI-4946) nifi-spark-bundle : Adding support for pyfiles, file, jars options

     [ https://issues.apache.org/jira/browse/NIFI-4946?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Matt Burgess updated NIFI-4946:
-------------------------------
    Affects Version/s:     (was: 1.6.0)
               Status: Patch Available  (was: Open)

> nifi-spark-bundle : Adding support for pyfiles, file, jars options
> ------------------------------------------------------------------
>
>                 Key: NIFI-4946
>                 URL: https://issues.apache.org/jira/browse/NIFI-4946
>             Project: Apache NiFi
>          Issue Type: New Feature
>          Components: Extensions
>         Environment: Ubuntu 16.04, IntelliJ
>            Reporter: Mageswaran
>            Priority: Major
>         Attachments: nifi-spark-options.png, nifi-spark.png
>
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Adding support for submitting PySpark based Sparks jobs (which is normally structured as modules) over Livy on existing "ExecuteSparkInteractive" processor.
> This is done by reading file paths for pyfiles and file and an option from user whether the processor should trigger a batch job or not.
> [https://livy.incubator.apache.org/docs/latest/rest-api.html]
>  *Current Work flow Logic ( [https://github.com/apache/nifi/pull/2521 )|https://github.com/apache/nifi/pull/2521]*
>  * Check whether the processor has to handle code or submit a Spark job
>  * Read incoming flow file
>  ** If batch == true
>  *** If flow file matches Livy `batches` JSON response through `wait` loop
>  **** Wait for Status Check Interval
>  **** Read the state
>  **** If state is `running` route it to `wait` or if it  is `success` or `dead` route it accordingly
>  *** Else
>  **** Ignore the flow file
>  **** Trigger the Spark job over Livy `batches` endpoint
>  **** Read the state of the submitted job
>  **** If state is `running` route it to `wait` or if it  is `success` or `dead` route it accordingly
>  ** Else:
>  *** Existing Logic to handle `Code`
>  
> !nifi-spark-options.png!
> !nifi-spark.png!
>  
> Thanks.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)