What would be the strategy / approach for Environment Management for Hadoop clusters? Are there any best practices / recommendations for moving across various stages of a Hadoop cluster setup - from development to testing to pre-deployment and production?
Any pointers please.
In my previous life, i would do the following:
Review release notes, documentation, behavior changes.
1. New release of Ambari and HDP, install a multinode cluster on my desktop using Ambari quick start guide https://cwiki.apache.org/confluence/display/AMBARI/Quick+Start+Guide
2. If i already am running a cluster abd new release is out, test upgrade using quick start guide. (Document all quirks).
3. Install/upgrade my dev physical cluster using same bits. (document all quirks).
4. Open a case with support for FYI. Upgrade production, take all experience from prior into consideration. In case of trouble, call support.
5. Document your experience on HCC.
@Artem Ervits I was looking from a slightly different perspective - Not only from setting up the cluster bits, but also managing the code, configurations etc across environments. Just to summarise, is there any best practice / recommendation from Hortonworks on how to ensure the environments are inline with customer expectations across development, testing, pre-production and production from cluster hardware / application code / service configuration etc perspectives.