I'm wondering whether we should remove hbase coverage from core tests
and only run against hbase in exhaustive tests.

This would not remove coverage in absolute terms, but detecting a
newly-introduced regression against hbase functionality would take
longer and might require more work to pin down the exact commit.

Does anybody believe this would place an extra burden on the
development process that would outweigh the benefits (faster test data
loads and test execution)?

Reply via email to