Yes, you can try it, though I doubt that will 100% work. Have a look at the "hadoop 3" JIRAs and PRs still in progress on master.
On Fri, Apr 5, 2019 at 1:14 PM Anton Kirillov <akirillov.mail...@gmail.com> wrote: > > Marcelo, Sean, thanks for the clarification. So in order to support Hadoop 3+ > the preferred way would be to use Hadoop-free builds and provide Hadoop > dependencies in the classpath, is that correct? > --------------------------------------------------------------------- To unsubscribe e-mail: dev-unsubscr...@spark.apache.org