Oh, wait a minute. IIRC we have a consensus to use the jira page as release note instead of committing a file to the repo? The problem is that, sometimes we just missed something in CHANGES.md or RELEASENOTES.md, no actually code change, but since we have committed these files to our git repo, we have to sink the RC and generate a new one.
IIRC for making the two alpha releases for 3.0.0, I have already started to link the jira page instead of CHANGES.md and RELEASENOTES.md. Maybe we could just remove these steps from the create release scripts? But maybe for branch-2.4 we still need to use these steps... Thanks. Andrew Purtell <andrew.purt...@gmail.com> 于2022年6月2日周四 22:56写道： > I just noticed the release notes issue on the branch last evening. Let’s > add the spotless application to the release note generation step in > create-release at least or it will keep happening. The RM does not get a > chance to intervene. The script uses the RM credentials to commit > immediately after generating the files. > > > On Jun 2, 2022, at 5:31 AM, 张铎 <palomino...@gmail.com> wrote: > > > > Oh, I think we should include HBASE-27023, it fixes the license issue > after > > we apply the spotless formatter where it will add the license header > > automatically. > > > > And also, it seems that the generated RELEASENOTE.md will introduce some > > lines ending with whitespace, then it will cause pre commit build > failure. > > > > I think we should run a spotless:apply before committing the files in the > > release scripts, and maybe we should also add spotless:check in the mvn > > verify stage so it will also fail the mvn install command, which will let > > developers pay more attention on it. > > > > Let me open an issue to land these changes. > > > > 张铎(Duo Zhang) <palomino...@gmail.com> 于2022年6月2日周四 11:14写道： > > > >> OK, got it. Thanks for the clarification~ > >> > >> Andrew Purtell <andrew.purt...@gmail.com> 于2022年6月2日周四 11:12写道： > >> > >>> I don’t think a flaky test is cause to fail the release. So these all > >>> have a fix version of 2.5.1. I mentioned them so people could be aware > of > >>> these known issue when evaluating their own findings. Hope that helps. > >>> > >>>>> On Jun 1, 2022, at 8:11 PM, 张铎 <palomino...@gmail.com> wrote: > >>>>> > >>>>> On flakyness, if it is a big problem on promoting the RC, I think we > >>> can > >>>> ignore the test first(instead of removing it). > >>>> > >>>> Can open a follow on issue to make it stable. > >>>> > >>>> Thanks. > >>>> > >>>> Andrew Purtell <apurt...@apache.org> 于2022年6月1日周三 23:25写道： > >>>> > >>>>> Makes sense and if nobody has a concern we can do that. > >>>>> > >>>>> Also you wanted to support the change to the RSGroup API and put back > >>> the > >>>>> unit test (and fix it to not flake), correct, which seems fine to me > >>> too. > >>>>> > >>>>>> On Tue, May 31, 2022 at 10:54 PM 张铎(Duo Zhang) < > palomino...@gmail.com > >>>> > >>>>>> wrote: > >>>>>> > >>>>>> The ReplicationLoadSink change is done in HBASE-26490, this is the > >>> commit > >>>>>> > >>>>>> > >>>>>> > >>>>> > >>> > https://github.com/apache/hbase/commit/b5b286d793f00ec8b9ed02f51fd9324e46f29c86 > >>>>>> > >>>>>> The constructor is marked as IA.Private so I do not think we need to > >>>>>> restore it back. > >>>>>> > >>>>>> Thanks. > >>>>>> > >>>>>> Andrew Purtell <apurt...@apache.org> 于2022年6月1日周三 13:08写道： > >>>>>> > >>>>>>> This is not a VOTE. > >>>>>>> > >>>>>>> HBase 2.5.0RC0 is ready for testing, other evaluation, and > >>>>> consideration > >>>>>> of > >>>>>>> compatibility concerns. > >>>>>>> > >>>>>>> The release notes need to be rebased on the latest 2.4.x release, > >>>>> 2.4.12. > >>>>>>> Fix > >>>>>>> versions on relvant JIRAs must be cleaned up first. This is in > >>>>> progress. > >>>>>> I > >>>>>>> also plan to try our new audit tool for matching git history and > fix > >>>>>>> versions. > >>>>>>> > >>>>>>> There are known flaky unit tests and minor issues. You can find > them > >>> by > >>>>>>> searching for the '2.5.1' fix version, or this URL: > >>>>>>> > >>>>>>> https://s.apache.org/hbase-v2.5.1 > >>>>>>> > >>>>>>> The compatibility report, based on 2.4.12, can be found here: > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>> > >>>>> > >>> > https://dist.apache.org/repos/dist/dev/hbase/2.5.0RC0/api_compare_2.4.12_to_2.5.0RC0.html > >>>>>>> > >>>>>>> An initial response to some of the findings can be found on > >>>>> HBASE-27081, > >>>>>>> along > >>>>>>> with additional discussion. Further discussion is probably > necessary. > >>>>>>> > >>>>>>> The tag to be evaluated is 2.5.0RC0: > >>>>>>> > >>>>>>> https://github.com/apache/hbase/tree/2.5.0RC0 > >>>>>>> > >>>>>>> This tag currently points to git reference 2da2dd917 . > >>>>>>> > >>>>>>> The release files, including signatures, digests, as well as > >>> CHANGES.md > >>>>>>> and RELEASENOTES.md included in this RC can be found at: > >>>>>>> > >>>>>>> https://dist.apache.org/repos/dist/dev/hbase/2.5.0RC0/ > >>>>>>> > >>>>>>> Maven artifacts are available in a staging repository at: > >>>>>>> > >>>>>>> > >>>>>> > >>> > https://repository.apache.org/content/repositories/orgapachehbase-1486/ > >>>>>>> > >>>>>>> Artifacts were signed with the 0xD5365CCD key which can be found > in: > >>>>>>> > >>>>>>> https://downloads.apache.org/hbase/KEYS > >>>>>>> > >>>>>>> The second release candidate, 2.5.0RC1, will be made available for > >>>>>>> evaluation > >>>>>>> after discussion and any necessary changes and fixes have settled. > >>>>>>> > >>>>>>> Best, > >>>>>>> > >>>>>>> Your 2.5 Release Manager > >>>>>>> > >>>>>> > >>>>> > >>>>> > >>>>> -- > >>>>> 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 > >>>>> > >>> > >> >
