What is the effort required for migrating from Hadoop 0.20.2 to 0.20.205 and from 0.20.2 to 1.0.1?

Question

I was looking to migrate my EMR implementation from an older version to the latest versions because I am primarily facing a lot of issues.

My current implementation uses Hadoop 0.20.2.

I wanted to understand how much effort in terms of code change would be required for migrating from 0.20.2 to -

1. 0.20.205
2. 1.0.1

Are the APIs very different and require a lot of recoding? Any basic idea would be highly helpful.

Answer 1

0.20.205 was just renamed to 1.0 so it is esentially the same release. The APIs have hardly any difference. 1.0 is similar to 0.20.2 with append & security features which basically means it supports HBase integration and can be used in enterprises.

We ported our jobs running on EMR on 0.20.2 to directly run on 1.0. All our jobs, whether they were using the new or old API did not have a single issue but ran correctly without us having to change anything. So I believe you should not face any issues.