[ZEPPELIN-323] Paragraph execution is very slow for remote Spark

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.6.0
• Component/s: Core
• Labels: performance

Environment

EC2

• Assignee: Unassigned
• Reporter: Samuel Alexander
• Created: 25/Sep/15 09:16
• Updated: 15/Oct/15 11:51

Description

I've built EC2 from github in one of EC2 instance. I've running my spark cluster running in EC2 instance as well.

I've configured zeppelin to use Spark cluster in EC2 by providing spark master URL in spark interpreter.

I loaded the sample "bank" table. Now the below sql code takes atleast 10 minutes to complete

%sql
select age, count(1) value
from bank
where age < ${maxAge=30}
group by age
order by age

Whereas when I use the local[*] spark it takes only few seconds.

Do I need to do any other configuration in zeppelin to execute the paragraph faster using external Spark cluster?