[
https://issues.apache.org/jira/browse/SPARK-18041?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15608380#comment-15608380
]
sudheesh k s commented on SPARK-18041:
--------------------------------------
We need to kill spark stream applications before submitting the same
application again(all kafka partitions will be handled by first one so second
one just sits idle). We were checking on json and was killing any running app.
We wont be knowing submitted driver id, and dint find a way to find out a way
to list out running drivers in cluster.
Now multiple applications were getting submitted to cluster but with different
class names.
To avoid killing others we thought it will be good idea to have class name in
json.
> activedrivers section in http:sparkMasterurl/json is missing Main class
> information
> -----------------------------------------------------------------------------------
>
> Key: SPARK-18041
> URL: https://issues.apache.org/jira/browse/SPARK-18041
> Project: Spark
> Issue Type: Improvement
> Components: Spark Submit
> Affects Versions: 1.6.2
> Reporter: sudheesh k s
> Priority: Minor
>
> http:sparkMaster_Url/json gives the status of running applications as well as
> drivers. But it is missing information like, driver main class.
> To identify which driver is running on driver class information is needed.
> eg:
> "activedrivers" : [ {
> "id" : "driver-20161020173528-0032",
> "starttime" : "1476965128734",
> "state" : "RUNNING",
> "cores" : 1,
> "memory" : 1024
> } ],
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
