Thanks Andrew. I recall that there was a discussion around this topic, but I had not followed up to the end so I did not know the discussion result...
So at least we could add a tool in create-release for cutting release branches. Andrew Purtell <[email protected]> 于2023年12月1日周五 01:16写道： > > Maven automates the change so the human does not edit the POM files by hand. > A PR to check this is wasteful of people’s time because a reviewer is just > signing off on a change made by a tool and won’t check the details (this is > human nature) and the RM is sitting around for who knows how long waiting for > the sign off (this is the actual problem) and as you know the process for > making a release is very lengthy already. > > Apache release requirements forbid use of Jenkins to generate release > artifacts. There are various conversations about this in the archives of > infrastructure@ or members@ or perhaps someone like Nick more directly > involved with build infrastructure has more context. Something to do with > automated pipelines can be hacked and having humans doing enough by hand that > they can really attest to the results provides more assurance. > > PRs are valuable and necessary most of the time because the changes are made > by humans and are nontrivial and hugely benefit from review by another > person. But PRs for signing off on changes made by tools are not. > > > On Nov 29, 2023, at 11:20 PM, 张铎 <[email protected]> wrote: > > > > That's the problem! Now we use $revision for better command line > > support when releasing 2.5.x and 2.5.x-hadoop3, so the command for > > bumping version is > > > > mvn versions:set-property -Dproperty=revision -DnewVersion=<newversion> > > > > I've been thinking of introducing a jenkins job for creating release > > for a long time, but haven't gotten enough time to do it since it is > > not easy for me as I'm not a jenkins expert... > > > > Andrew Purtell <[email protected]> 于2023年11月30日周四 10:45写道： > >> > >> It seems silly to do a review just for a change that modifies the project > >> version. I mean, it's trivial to "mvn versions:set > >> -DnewVersion=<newversion>", and surely RMs can be trusted to do this and > >> not mess it up. You might think the PR is fine, but it can introduce hours > >> if not days of delay in the release process, which takes long enough > >> already. Anyway glad we agree results produced by release tools don't > >> require signoff. > >> > >>> On Wed, Nov 29, 2023 at 5:13 PM 张铎(Duo Zhang) <[email protected]> > >>> wrote: > >>> > >>> Just for checking there is no mistake :) > >>> > >>> And if we can formalize the cutting release branch related operations > >>> in a release tool, I do not think we need the signoff, just running > >>> the automatic tool is fine, but now it still requires manual > >>> operations, so I think having a double check is better. > >>> > >>> Thanks. > >>> > >>> Andrew Purtell <[email protected]> 于2023年11月30日周四 02:45写道： > >>>> > >>>> I don't think RMs should need to get PR approvals for version changes in > >>>> the POMs. The create-release script normally does this automatically > >>>> anyway. Sure, this is renumbering branch-2, not a release branch, but who > >>>> is going to object? > >>>> > >>>> On Wed, Nov 29, 2023 at 5:38 AM 张铎(Duo Zhang) <[email protected]> > >>> wrote: > >>>> > >>>>> Better to open an umbrella issue for tracking the task, and open a > >>>>> sub-task for each step. > >>>>> > >>>>> For bumping version to 2.7.0-SNAPSHOT, still better to open an PR as > >>>>> it is a code commit. > >>>>> > >>>>> Others are all good. Thanks Byran for taking care of this! > >>>>> > >>>>> Bryan Beaudreault <[email protected]> 于2023年11月29日周三 21:24写道： > >>>>>> > >>>>>> Here's my planned approach. Please let me know if this looks good: > >>>>>> > >>>>>> # Create new branch > >>>>>> git checkout branch-2 > >>>>>> git pull upstream branch-2 > >>>>>> git status # check all is good > >>>>>> git checkout -b branch-2.6 > >>>>>> git push upstream branch-2.6 > >>>>>> > >>>>>> # Update snapshot version on branch-2 > >>>>>> git checkout branch-2 > >>>>>> Edit root pom.xml, change properties.revision to 2.7.0-SNAPSHOT > >>>>>> git commit, "Renumber to 2.7.0-SNAPSHOT after branching branch-2.6" > >>> with > >>>>>> self-sign-off > >>>>>> git push upstream branch-2 > >>>>>> > >>>>>> After that I will start looking through JIRA for fixVersion 2.6.0 > >>> issues, > >>>>>> to make sure there's no inconsistencies. *Is there anything I need > >>> to do > >>>>> to > >>>>>> set up nightly builds for the new branch? *We should probably let > >>> nightly > >>>>>> builds run for a few days before starting RC. > >>>>>> > >>>>>> Prior to starting the RC process, I will try deploying the newly > >>> minted > >>>>>> 2.6-snapshot onto a test cluster in my environment. I notice that > >>> Andrew > >>>>>> did a bunch of load and chaos testing for the 2.5.0 release. I will > >>> try > >>>>>> running these locally, but we have not ever run ITBLL, etc in my > >>>>> company's > >>>>>> environment so that might take some time to set up. > >>>>>> > >>>>>> On Wed, Nov 22, 2023 at 8:50 AM Bryan Beaudreault < > >>>>> [email protected]> > >>>>>> wrote: > >>>>>> > >>>>>>> Sorry for the delay. I just sent a new thread to the list to alert > >>> to > >>>>>>> creating branch-2.6 next week. > >>>>>>> > >>>>>>> On Tue, Nov 21, 2023 at 3:35 AM 张铎(Duo Zhang) < > >>> [email protected]> > >>>>>>> wrote: > >>>>>>> > >>>>>>>> Bump again :) > >>>>>>>> > >>>>>>>> Are we ready to cut branch-2.6? > >>>>>>>> > >>>>>>>> Bryan Beaudreault <[email protected]> 于2023年10月26日周四 > >>> 21:22写道： > >>>>>>>>> > >>>>>>>>> Just as an update here, I plan to do this but have been busy > >>> with > >>>>>>>> internal > >>>>>>>>> work. I'll try to carve out time in the next 1-2 weeks. > >>>>>>>>> > >>>>>>>>> On Thu, Oct 19, 2023 at 8:50 PM 张铎(Duo Zhang) < > >>>>> [email protected]> > >>>>>>>> wrote: > >>>>>>>>> > >>>>>>>>>> Just send a notice to the dev mailing list, and then use the > >>>>> normal > >>>>>>>>>> git way to create a new branch called branch-2.6 from > >>> branch-2, > >>>>> push > >>>>>>>>>> it to github, done :) > >>>>>>>>>> > >>>>>>>>>> Bryan Beaudreault <[email protected]> 于2023年10月20日周五 > >>>>> 07:53写道： > >>>>>>>>>>> > >>>>>>>>>>> I can give it a shot :) > >>>>>>>>>>> > >>>>>>>>>>> I’ve started looking at the release docs, but don’t see any > >>>>> info on > >>>>>>>> how > >>>>>>>>>> to > >>>>>>>>>>> create a new release branch. Is there a process for that? > >>>>>>>>>>> > >>>>>>>>>>> On Thu, Oct 19, 2023 at 1:55 PM Andrew Purtell < > >>>>> [email protected] > >>>>>>>>> > >>>>>>>>>> wrote: > >>>>>>>>>>> > >>>>>>>>>>>> I second this nomination! > >>>>>>>>>>>> > >>>>>>>>>>>> On Thu, Oct 19, 2023 at 6:39 AM 张铎(Duo Zhang) < > >>>>>>>> [email protected]> > >>>>>>>>>>>> wrote: > >>>>>>>>>>>> > >>>>>>>>>>>>> OK, good. So we are good to go I think. > >>>>>>>>>>>>> And we need a release manager for 2.6.x, do you have any > >>>>>>>> interest in > >>>>>>>>>>>>> being a release manager for 2.6.x? :) > >>>>>>>>>>>>> > >>>>>>>>>>>>> Bryan Beaudreault <[email protected]> > >>> 于2023年10月19日周四 > >>>>>>>> 21:22写道： > >>>>>>>>>>>>>> > >>>>>>>>>>>>>> Hello. I'm not aware of any major blockers at the > >>> moment. > >>>>> In > >>>>>>>> fact > >>>>>>>>>> we > >>>>>>>>>>>>>> recently finished our migration and are fully using > >>> 2.6's > >>>>> TLS > >>>>>>>> in > >>>>>>>>>> all of > >>>>>>>>>>>>> our > >>>>>>>>>>>>>> production. > >>>>>>>>>>>>>> > >>>>>>>>>>>>>> On Thu, Oct 19, 2023 at 12:07 AM 张铎(Duo Zhang) < > >>>>>>>>>> [email protected]> > >>>>>>>>>>>>>> wrote: > >>>>>>>>>>>>>> > >>>>>>>>>>>>>>> Hi Bryan, do we still have any blocker issues for > >>>>> cutting > >>>>>>>>>> branch-2.6 > >>>>>>>>>>>>>>> and starting to release 2.6.0? > >>>>>>>>>>>>>>> > >>>>>>>>>>>>>>> Thanks. > >>>>>>>>>>>>>>> > >>>>>>>>>>>>>>> Andrew Purtell <[email protected]> 于2023年9月26日周二 > >>>>> 01:02写道： > >>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>> Thanks Duo. Following up on HBASE-28042. > >>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>> On Sun, Sep 24, 2023 at 7:53 PM 张铎(Duo Zhang) < > >>>>>>>>>>>> [email protected] > >>>>>>>>>>>>>> > >>>>>>>>>>>>>>> wrote: > >>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>> Some updates here. > >>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>> The tcnative module is almost done, only some > >>> small > >>>>>>>> concerns > >>>>>>>>>> on > >>>>>>>>>>>>> how to > >>>>>>>>>>>>>>>>> deal with the license and notice files. > >>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>> There is still a blocker issue HBASE-28081, but > >>>>> anyway > >>>>>>>> we > >>>>>>>>>> have > >>>>>>>>>>>>> already > >>>>>>>>>>>>>>>>> found the root cause is HBASE-28042, so I think > >>> it > >>>>> will > >>>>>>>> be > >>>>>>>>>> fixed > >>>>>>>>>>>>> soon. > >>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>> All other issues have been resolved. > >>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>> Thanks. > >>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>> Andrew Purtell <[email protected]> > >>> 于2023年9月22日周五 > >>>>>>>> 02:21写道： > >>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>> Both sound good, thanks. > >>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>> On Tue, Sep 19, 2023 at 7:46 PM 张铎(Duo Zhang) > >>> < > >>>>>>>>>>>>> [email protected] > >>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>> wrote: > >>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>> We'd better get a new thirdparty release > >>> first, > >>>>>>>> update > >>>>>>>>>>>>> branch-2 to > >>>>>>>>>>>>>>> use > >>>>>>>>>>>>>>>>>>> the new thirdparty release, and then make > >>> the > >>>>> 2.6.0 > >>>>>>>>>> release, > >>>>>>>>>>>>> since > >>>>>>>>>>>>>>> the > >>>>>>>>>>>>>>>>>>> new netty release includes the improvements > >>>>>>>> contributed > >>>>>>>>>> by > >>>>>>>>>>>>> Bryan. > >>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>> https://github.com/netty/netty/pull/13551 > >>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>> The current blocking issues are > >>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>> HBASE-28075, where we want to shade netty > >>>>> tcnative > >>>>>>>> as > >>>>>>>>>> well, > >>>>>>>>>>>> and > >>>>>>>>>>>>>>> Bryan > >>>>>>>>>>>>>>>>>>> said he already had a workable solution. > >>>>>>>>>>>>>>>>>>> HBASE-22138, where we want to also > >>> relocated the > >>>>>>>> proto > >>>>>>>>>> files > >>>>>>>>>>>> in > >>>>>>>>>>>>>>> shaded > >>>>>>>>>>>>>>>>>>> protobuf, there is no technical problem now > >>> and > >>>>> I've > >>>>>>>>>> opened a > >>>>>>>>>>>>> PR > >>>>>>>>>>>>>>> using > >>>>>>>>>>>>>>>>>>> the 4.1.5-SNAPSHOT version of > >>> hbase-thirdparty, > >>>>>>>>>>>>>>>>>>> https://github.com/apache/hbase/pull/5411. > >>>>> There is > >>>>>>>>>> another > >>>>>>>>>>>>>>> problem > >>>>>>>>>>>>>>>>>>> that we used to exclude CanIgnoreReturnValue > >>>>>>>> annotation > >>>>>>>>>> to > >>>>>>>>>>>>> ignore > >>>>>>>>>>>>>>> some > >>>>>>>>>>>>>>>>>>> error prone errors, but I think we should > >>> add it > >>>>>>>> back > >>>>>>>>>> and fix > >>>>>>>>>>>>> the > >>>>>>>>>>>>>>>>>>> error prone errors, instead of ignoring it, > >>>>> should > >>>>>>>> not be > >>>>>>>>>>>> very > >>>>>>>>>>>>>>> hard to > >>>>>>>>>>>>>>>>>>> do. > >>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>> All the above issues have a clear path to > >>>>> resolve, > >>>>>>>> so I > >>>>>>>>>> think > >>>>>>>>>>>>> we > >>>>>>>>>>>>>>> will > >>>>>>>>>>>>>>>>>>> not delay too much. > >>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>> Thanks. > >>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>> Bryan Beaudreault <[email protected]> > >>>>>>>> 于2023年9月20日周三 > >>>>>>>>>>>>> 02:51写道： > >>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>> I hate to keep delaying, but it might be > >>> nice > >>>>> to > >>>>>>>> wait > >>>>>>>>>> for > >>>>>>>>>>>>>>>>>>>> > >>>>> https://issues.apache.org/jira/browse/HBASE-28065. > >>>>>>>> It > >>>>>>>>>> was > >>>>>>>>>>>> a > >>>>>>>>>>>>>>> pretty > >>>>>>>>>>>>>>>>> big > >>>>>>>>>>>>>>>>>>> (if > >>>>>>>>>>>>>>>>>>>> long standing) bug for us, and given the > >>>>> solution > >>>>>>>> is > >>>>>>>>>> ready > >>>>>>>>>>>> ( > >>>>>>>>>>>>>>>>>>>> https://github.com/apache/hbase/pull/5384) > >>> we > >>>>>>>> might as > >>>>>>>>>>>> well > >>>>>>>>>>>>> get > >>>>>>>>>>>>>>> it > >>>>>>>>>>>>>>>>> in > >>>>>>>>>>>>>>>>>>> our > >>>>>>>>>>>>>>>>>>>> next releases. > >>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>> On Tue, Sep 19, 2023 at 1:37 PM Andrew > >>>>> Purtell < > >>>>>>>>>>>>>>> [email protected]> > >>>>>>>>>>>>>>>>>>> wrote: > >>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>> Thanks Duo. > >>>>>>>>>>>>>>>>>>>>> I do not need help. > >>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>> I have been waiting to ensure more > >>> should do > >>>>>>>> items > >>>>>>>>>> do not > >>>>>>>>>>>>>>> appear. > >>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>> I think we are good now? > >>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>> On Mon, Sep 18, 2023 at 12:31 AM 张铎(Duo > >>>>> Zhang) < > >>>>>>>>>>>>>>>>> [email protected]> > >>>>>>>>>>>>>>>>>>>>> wrote: > >>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>> HBASE-28061 has been merged. > >>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>> Do you need any help on the release > >>>>> processing > >>>>>>>>>> Andrew? > >>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>> I could offer my help. > >>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>> Thanks. > >>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>> 张铎(Duo Zhang) <[email protected]> > >>>>>>>>>> 于2023年9月14日周四 > >>>>>>>>>>>>> 15:55写道： > >>>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>>> If you have not cut the RC yet, > >>> please > >>>>> also > >>>>>>>>>> include > >>>>>>>>>>>>>>>>> HBASE-28061 in > >>>>>>>>>>>>>>>>>>>>>>> 2.5.6, as it could solve the > >>> problem for > >>>>>>>> running > >>>>>>>>>> pre > >>>>>>>>>>>>>>> compiled > >>>>>>>>>>>>>>>>> 2.5.x > >>>>>>>>>>>>>>>>>>>>>>> tarball and also our published > >>>>> artifacts in > >>>>>>>> maven > >>>>>>>>>>>> repo > >>>>>>>>>>>>>>> against > >>>>>>>>>>>>>>>>>>> hadoop > >>>>>>>>>>>>>>>>>>>>>>> 3.3.x. > >>>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>>> The hive community needs this as > >>> they > >>>>> want > >>>>>>>> to > >>>>>>>>>> upgrade > >>>>>>>>>>>>>>> hbase and > >>>>>>>>>>>>>>>>>>> hadoop > >>>>>>>>>>>>>>>>>>>>>>> dependencies. > >>>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>>> Thanks. > >>>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>>> Andrew Purtell <[email protected] > >>>> > >>>>>>>>>> 于2023年8月30日周三 > >>>>>>>>>>>>>>> 06:48写道： > >>>>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>>>> I am back from vacation. Slowly > >>>>> clearing > >>>>>>>> my > >>>>>>>>>> inbox. > >>>>>>>>>>>>> 1,545 > >>>>>>>>>>>>>>>>> unread, > >>>>>>>>>>>>>>>>>>> it's > >>>>>>>>>>>>>>>>>>>>>> been > >>>>>>>>>>>>>>>>>>>>>>>> a while. > >>>>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>>>> It seems we are ready to cut > >>>>> branch-2.6 > >>>>>>>> for > >>>>>>>>>>>>> stabilizing > >>>>>>>>>>>>>>> for > >>>>>>>>>>>>>>>>>>> 2.6.0. > >>>>>>>>>>>>>>>>>>>>> Any > >>>>>>>>>>>>>>>>>>>>>>>> volunteers to RM 2.6.0? Otherwise > >>> I'm > >>>>>>>> happy to > >>>>>>>>>> do > >>>>>>>>>>>> it, > >>>>>>>>>>>>>>>>> starting > >>>>>>>>>>>>>>>>>>> next > >>>>>>>>>>>>>>>>>>>>>> week, > >>>>>>>>>>>>>>>>>>>>>>>> probably 9/5 or 9/6. > >>>>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>>>> Some good changes have landed in > >>>>>>>> branch-2.5. > >>>>>>>>>> RC0 > >>>>>>>>>>>> for > >>>>>>>>>>>>>>> 2.5.6 by > >>>>>>>>>>>>>>>>>>> Friday > >>>>>>>>>>>>>>>>>>>>>> 9/1. > >>>>>>>>>>>>>>>>>>>>>>>> > >>>>>>>>>>>>>>>>>>>>>>>> -- > >>>>>>>>>>>>>>>>>>>>>>>> Best regards, > >>>>>>>>>>>>>>>>>>>>>>>> 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 > >>>>>>>>>>>>>>> > >>>>>>>>>>>>> > >>>>>>>>>>>> > >>>>>>>>>>>> > >>>>>>>>>>>> -- > >>>>>>>>>>>> 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 > >>> > >> > >> > >> -- > >> 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
