On Thu, Jun 23, 2016 at 4:46 PM, Jon Ernster <[email protected]> wrote: > Check out: > > https://github.com/wangzw/hawq-devel-env > https://hub.docker.com/r/mayjojo/hawq-test/ > https://hub.docker.com/r/mayjojo/hawq-devel/
Great starting points! And perhaps something that we can coalesce with Jemish's efforts as well. More comments inline: > On Thu, Jun 23, 2016 at 3:13 PM, Jemish Patel <[email protected]> wrote: > >> I am working on configuring /installing HAWQ on a really bare-bones 4 >> container docker environment to make it easy for dev/test. >> >> The advantages for devs are that it is light weight and one can get to >> feature development or reproducing an issue , etc and abstracting away all >> the needed installation/configuration. >> >> Users could benefit as well as they can get a running HAWQ environment >> with ORCA, PXF, PL/python, PL/R and MADLib that they can use right away to >> prove out a theory or even simply try it out. >> >> Would this be of interest to anyone? I think this is super interesting. And in fact it reminds me a lot of the efforts that Bigtop is pursuing for all the Bigdata projects. In fact, quite a few projects in Hadoop ecosystem are leveraging Bigtop for that type of integration. See the following excellent presentation by Trend Micro on the work that is being done: http://schd.ws/hosted_files/apachebigdata2016/4e/Dockerized%20Hadoop%20Platform%20and%20Recent%20Updates%20in%20Apache%20Bigtop.pdf Recently Apache Apex leveraged Bigtop to create demo VM, etc. Zeppelin folks are also pretty active. So perhaps, one suggestion here would be to do this integration work in Bigtop itself. What do you guys think? Thanks, Roman.
