[
https://issues.apache.org/jira/browse/SPARK-14538?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Davies Liu updated SPARK-14538:
-------------------------------
Assignee: (was: Davies Liu)
> Increase the default stack size of spark shell
> ----------------------------------------------
>
> Key: SPARK-14538
> URL: https://issues.apache.org/jira/browse/SPARK-14538
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Reporter: Davies Liu
>
> The default stack size of 64-bit JVM is 1M, the new SQL parser may fail as
> StackOverflow if the SQL query is very long. Also the anaylizer/optimizer
> very deep stacks, it may also fail on complicated query.
> We should increase the default stack size for Spark (at least for driver) to
> 2M or 4M.
> The down side of this is that we will increase the memory usage for each
> thread. Right now, a fresh new spark-shell (with HiveContext) take 814M on my
> mac, it increased to 886M if it change the stack size to 4M (about 20+
> threads).
> cc [~rxin]
--
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
