Attached is a patch that will remove every reference to the files that I see in the tests. I think we should also remove the entirety of the bigtop-tests/test-artifacts/hive/src/main/resources/seed_data_files/ml-data directory, but doing so puts this patch over the size limit for Apache attachments, so I've left it out of my post.
While it would be nice to run the tests and make sure it all works before committing, I've been meaning to do that for a long time and just never get to it. Unless someone else is able to run them all quickly (I don't have a good environment to run them all in right now and I'm not super familiar with doing so), I propose we just drop the files we shouldn't be distributing, and if it fails on Jenkins we fix it when it fails. Any thoughts on that?