[
https://issues.apache.org/jira/browse/SPARK-10647?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14791387#comment-14791387
]
Josh Rosen commented on SPARK-10647:
------------------------------------
The spark.deploy.zookeeper.* properties are used by the standalone mode's HA
recovery features
(https://spark.apache.org/docs/latest/spark-standalone.html#high-availability).
I think the correct fix here is to update the Mesos code to use
spark.deploy.mesos.zookeeper.dir
(https://github.com/apache/spark/pull/5144/files#diff-3c5e5516915ada1d89f10000259de069R97).
We should also update the Mesos documentation to mention these configurations,
since they don't appear to be documented anywhere.
[~tnachen], I'm going to assign the doc updates and bugfixes to you.
> Rename property spark.deploy.zookeeper.dir to spark.mesos.deploy.zookeeper.dir
> ------------------------------------------------------------------------------
>
> Key: SPARK-10647
> URL: https://issues.apache.org/jira/browse/SPARK-10647
> Project: Spark
> Issue Type: Improvement
> Components: Mesos
> Reporter: Alan Braithwaite
> Priority: Minor
>
> This property doesn't match up with the other properties surrounding it,
> namely:
> spark.mesos.deploy.zookeeper.url
> and
> spark.mesos.deploy.recoveryMode
> Since it's also a property specific to mesos, it makes sense to be under that
> hierarchy as well.
--
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
