These are artifacts of building Spark 1.6.3 with Hadoop 2.7.3
Spark 1.6.3 doesn't have hadoop-2.7 profile. The provided patch is to add that profile in.
Build commands are:
./make-distribution.sh --skip-java-test --name without-hadoop --tgz -Phadoop-2.7 -Phive -Phive-thriftserver -Pyarn -Phadoop-provided
./make-distribution.sh --skip-java-test --name hadoop-2.7.3 --tgz -Phadoop-2.7 -Phive -Phive-thriftserver -Pyarn