I also think this should go in to branch-2 only for 2.6.0. 2.5.0 has been held up long enough. There are a few things in flight still but they seem like they can fit into a time box of this week. The new TLS stuff is going to take longer than that to harden and bake.
Related, there is no particular reason to expect a long drawn out time in between 2.5.0 and 2.6.0 like we had with 2.4.0 and 2.5.0. As soon as 2.5.0 is out, we could begin work on 2.6.0 if that is where the center of gravity shifts. Or stuff like the TLS RPC can bake in branch-2 for a while. Whatever the situation demands. On Tue, Jul 12, 2022 at 11:56 AM Viraj Jasani <[email protected]> wrote: > FWIW, I agree with Duo, we should aim to land this great TLS work on > branch-2 but not on 2.5 and let 2.5.0 release make progress. Also, having > TLS on 3.0.0/2.6.0 would anyways be compelling reason for us to move > forward with those releases, try it out at scale etc but at least 2.5.0 has > already waited long enough with Netty and SimpleRpcServer issues in SASL. > > Are you fine with this plan, Bryan? > > > On Tue, Jul 12, 2022 at 3:37 AM 张铎(Duo Zhang) <[email protected]> > wrote: > > > This one is too big. > > > > I prefer we land it on master and branch-2, but not branch-2.5. > > > > The 2.5.0 release has already been delayed a lot, adding new big feature > > may cause new problems... > > > > Thanks. > > > > Bryan Beaudreault <[email protected]> 于2022年7月12日周二 > > 17:37写道： > > > > > Thanks! One other -- Andor and others have been doing some great work > on > > > native TLS for HBase in > > https://issues.apache.org/jira/browse/HBASE-26666. > > > This would be nice to include as a big new feature in 2.5.0. We have > more > > > work we can do to add things like TLS-based client auth, etc, which > could > > > land in future releases. As it stands, I've deployed it in a real > cluster > > > and verified that encryption works end-to-end. > > > > > > Andor is just finishing up a couple tests, which he is hoping to look > at > > > today. It already got a bunch of reviews a couple months ago, but if > > anyone > > > is looking to give it a last look now is the time. > > > > > > On Mon, Jul 11, 2022 at 9:34 PM Andrew Purtell <[email protected]> > > > wrote: > > > > > > > Agreed, lgtm. Thanks for bringing it to our attention. > > > > > > > > On Mon, Jul 11, 2022 at 6:26 PM Bryan Beaudreault > > > > <[email protected]> wrote: > > > > > > > > > https://issues.apache.org/jira/browse/HBASE-27149 > > > > <https://issues.apache.org/jira/browse/HBASE-27149> > > > > is also patch available, > > > > > if anyone has time. It's a relatively simple one I think, and helps > > > avoid > > > > > scanner accumulation under heavy load. > > > > > > > > > > On Mon, Jul 11, 2022 at 9:18 PM Andrew Purtell < > [email protected]> > > > > > wrote: > > > > > > > > > > > We have landed several fixes for referencing counting concerns, > > with > > > > one > > > > > > pending, and the SimpleRpcServer problem has also been addressed. > > > > > > Therefore, I would like to roll 2.5.0RC1 by the end of this week. > > > Time > > > > to > > > > > > get the train moving again. > > > > > > > > > > > > There is one blocker filed by Duo on HBASE-27189 which has an > > > > associated > > > > > PR > > > > > > we can apply. I put the issue into Patch Available state and am > > > > prepared > > > > > to > > > > > > merge the PR. After this one there are six other issues in Patch > > > > > Available > > > > > > state: HBASE-27097 (simplerpcserver fix), HBASE-27188 (metrics > > > > > > improvement), HBASE-27186 (metrics improvement), HBASE-27161 > (test > > > > > > improvement), HBASE-27153 (tracing improvement), HBASE-27088 > > > > (integration > > > > > > test improvement). I hope we can get these all landed in the next > > > > couple > > > > > of > > > > > > days. > > > > > > > > > > > > Best, > > > > > > Andrew > > > > > > > > > > > > > > > > > > > > > > > -- > > > > Best regards, > > > > Andrew > > > > > > > > Unrest, ignorance distilled, nihilistic imbeciles - > > > > It's what we’ve earned > > > > Welcome, apocalypse, what’s taken you so long? > > > > Bring us the fitting end that we’ve been counting on > > > > - A23, Welcome, Apocalypse > > > > > > > > > > -- Best regards, Andrew Unrest, ignorance distilled, nihilistic imbeciles - It's what we’ve earned Welcome, apocalypse, what’s taken you so long? Bring us the fitting end that we’ve been counting on - A23, Welcome, Apocalypse
