[ https://issues.apache.org/jira/browse/SPARK-1896?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Tathagata Das resolved SPARK-1896. ---------------------------------- Resolution: Fixed Fix Version/s: 1.0.0 https://github.com/apache/spark/pull/846/ > MASTER masks spark.master in spark-shell > ---------------------------------------- > > Key: SPARK-1896 > URL: https://issues.apache.org/jira/browse/SPARK-1896 > Project: Spark > Issue Type: Bug > Reporter: Andrew Or > Fix For: 1.0.0 > > > For bin/spark-shell, the priority hierarchy is > {code} > MASTER > --master > spark.master (spark-defaults.conf) > {code} > Not sure if this was the intended behavior, but it was surprising to me when > I tried to run the shell on local mode, and it didn't take effect. Note that > this is inconsistent with running applications, where the hierarchy is > {code} > --master > spark.master (spark-defaults.conf) > MASTER > {code} > I think the MASTER is a good cluster default, but if the user explicitly > specifies a master we should obey it, like we do in normal applications. -- This message was sent by Atlassian JIRA (v6.2#6252)