I can't find these hosted anywhere (but they may be?) https://github.com/apache/spark/tree/a57bbf2ee02e30053e67f62a0afd6f525bba5c66/resource-managers/kubernetes/docker/src/main/dockerfiles/spark
Another (better?) option would be to assume a docker image where spark has been built, and then just build mahout and ship from that. Just my .02 On Wed, Dec 11, 2019 at 10:36 PM Andy Palumbo - [mahout] [slack] < 0-14-0-rele...@mahout.mailclark.ai> wrote: > Also not a docker expert here, But I'm not sure if we want to Necessarily > build Spark, whichi is why I'd posted the second link to the branch that id > started the night before friday's meeting. I forgot to explain my > thinking .. or rather the question that or got me started writing two.. Do > we want to install spark necessarily? > > I guess to put it simply, should we just assume that if a person is > Installing Mahout, that they already have Spark setup? > > We also have the case, though, where whe we want people to have erything > working out of the box, and a p[re-configurted docker and spark image gould > be nice for that. > > JUST FYI- these are the resons that I'd pasted 2 different files onto the > ticket.
