Dmitry, I think that developer convenience is of much smaller priority here than user experience. Moreover, we are not going to unfreeze AI 2.5 scope completely. Instead, we propose to add BLT fix to AI 2.5, but disallow anything else. This way most contributors will continue working with master as usual.
On Sat, Apr 28, 2018 at 5:36 PM, Dmitry Pavlov <dpavlov....@gmail.com> wrote: > Hi Vladimir, > > The fact that we will postpone the release will somehow allow users to > receive fixes faster? What if we will fix (BL)AT it in 2.6 instead? > > It will be additional issue for all developers to maintain 2.5 & master > branches. > > Sincerely, > Dmitriy Pavlov > > сб, 28 апр. 2018 г. в 17:28, Vladimir Ozerov <voze...@gridgain.com>: > > > Igniters, we have a problem. > > > > Baseline topology was introduced in AI 2.4. Looks like it have two > serious > > flaws which confuse our users a lot: > > 1) No automatic rebalance by default > > 2) Topology misalignment between persistent and in-memory caches > > > > These are really *critical* problems which severely affects virtually > every > > user. We agreed to return automatic re-balance by default and place > > in-memory and persistent caches under the same umbrella. > > > > I propose to postpone AI 2.5 until it is fixed. From what I heard it > should > > take about a month. > > > > Please share your thoughts. > > > > Vladimir. > > >
