Run a JAR file on Amazon EMR which I created using Hadoop 2 7 5

Question

I found that Amazon EMR has changed it's hadoop version from 2.7.3 to 2.8.3 and there is no option for 2.7.5. Now, I am using custom jar which is created using hadoop 2.7.5. Is there any way to run that JAR on Amazon EMR? The cluster is terminating with error while I am trying to run that jar using existing EMR version (either with hadoop 2.7.3 or 2.8.3) which is logical.

Answer 1

I suggest you recompile the code or downgrade back to 2.7.3, which should be able to run most 2.7.5 code
Amazon keeps all its EMR versions intact. All you need to do while lauching a EMR job is specify the release lable. E.g. for 2.7.3, I use emr-5.11.0 . Here is detailed info :

https://docs.aws.amazon.com/emr/latest/ReleaseGuide/images/emr-releases-5x.png