This was quite likely caused byt the cassandra.input.split.size_in_mb being misinterpreted as bytes, defaulting to 64 bytes, which produced crazy small splits. Just published 1.4.0-M2 with this bug fixed.
Stefano Ortolani July 29, 2015 at 10:55 AM
Edited
I can confirm that b1.4 (1.4.0-M2-SNAPSHOT) works.
Stefano Ortolani July 29, 2015 at 10:42 AM
I am on it. Are you on IRC by any chance?
Piotr Kołaczkowski July 29, 2015 at 10:40 AM
Edited
Can you try with b1.4 head? There shouldn't be much difference between b1.4 head and master now.
I am trying to run below code with spark 1.4.0 and its not working
build file :
Console Log :
The same code is working with spark 1.3.1