@Erich Schubert: why not trying to package Hadoop in Debian?



As a follow-up on your blog post, where you complain about the state of Hadoop. First, I couldn’t agree more with all you wrote. All of it! But why not trying to get Hadoop in Debian, rather than only complaining about the state of things?


I have recently packaged and uploaded Sahara, which is OpenStack big data as a service (in other words: running Hadoop as a service on an OpenStack cloud). Its working well, though it was a bit frustrating to discover exactly what you complained about: the operating system cloud image needed to run within Sahara can only be downloaded as a pre-built image, which is impossible to check. It would have been so much work to package Hadoop that I just gave up (and frankly, packaging all of OpenStack in Debian is enough work for a single person doing the job… so no, I don’t have time to do it myself).

OpenStack Sahara already provides the reproducible deployment system which you seem to wish. We “only” need Hadoop itself.