Give me a shout if unsure. I'm trying to put critical bug fixes and important perf improvements only on branch-2.0. I want to push out a 2.0.1 and have it be more stable and performant than 2.0.0 in the next week or so.
For example, here are some commits below [1] from last day or so. The first I'd say is not needed on branch-2.0. The two that follow are good because they help me w/ my perf-compare project. To be clear, no damage done (No need of 'apologies'!) I'll revert anything I think destabilizes but just a heads-up on what I'm thinking. Related, we should start the hbase3/hbase2.1 conversation one of these days. Would be sweet if we avoided a replay of the hbase2 experience in hbase3. Would be good too if we took up the Andrew Purtell suggestion of doing a "left-shift" in how we think about our version numbers; i.e. think of 3.0.0 as we would have thought of a 2.1.0. Thanks, St.Ack 1. commit d77989e9f7b51f058d09694e5109e3aec05e0596 Author: Mingdao Yang <[email protected]> Date: Sat May 5 22:26:12 2018 +0800 HBASE-20527 Remove unused code in MetaTableAccessor Signed-off-by: Chia-Ping Tsai <[email protected]> commit 81f69e585159840b622b9030e4eb6ebf35a7e6ae Author: Andrew Purtell <[email protected]> Date: Tue May 1 10:58:09 2018 -0700 HBASE-20513 Collect and emit ScanMetrics in PerformanceEvaluation commit 3c4fadae4257ff5078d2cc247efc938ea0d350ee Author: Andrew Purtell <[email protected]> Date: Thu May 3 16:25:17 2018 -0700 HBASE-20517 Fix PerformanceEvaluation 'column' parameter
