I agree with Ayush and Butao Zhang. Hive 4 introduces numerous new features, critical bug and security fixes, and significant performance improvements. Given that hive 4 has over 5000 commits compared to Hive 3, backporting changes from hive 4 to hive 3 would be a tedious task due to high possibility of merge conflicts. I believe it would be more productive to focus our active development efforts on the latest master, rather than continuing to maintain and release Hive 3.
I am +1 for marking 3.x EOL. Thanks! Simhadri G On Sat, Sep 28, 2024 at 1:29 PM Butao Zhang <butaozha...@163.com> wrote: > > Agreed with Ayush. I think we won't put much effort into 3.x, and even we > release a 3.x version, i think almost no one would like to maintain the 3.x > version. > IMHO, If we can guarantee full compatibility between the Hive4 and Hive3, > why we still try to release the 3.x version? > The most important thing, since Hive4 has lots of bug fix & improvment & > new features, we Hive community need to encourage people to use Hive4 and > listen to the feedback & continue to optimize Hive4, and then people can > always enjoy the up to date Hive4 version & this can also help the Hive > community grow in a positive way. > > I am +1 for marking 3.x EOL. > > > > Thanks, > Butao Zhang > ---- Replied Message ---- > From Ayush Saxena<ayush...@gmail.com> <undefined> > Date 9/28/2024 14:56 > To <dev@hive.apache.org> <dev@hive.apache.org> > Subject Re: [DISCUSS] Hive 3.x release > If I recall correctly, we were planning to call for a vote to mark 3.x > EOL after the 4.0.1 release. It seems no one is actively pursuing the > 3.x release line. Placing the burden of backporting security fixes, > managing the release, running tests (perf/tpcds/bugs), and validating > the release candidate on PMC members who have no interest in this line > seems unfair. This approach feels counterproductive. Even if someone > were to take on the effort, it's unlikely others would participate in > the vote or review the RC, especially since few active contributors > (myself included) have been involved with 3.x. Requiring three binding > votes seems impractical under these circumstances. > > I believe it's time to retire the 3.x line to address the recurring > security alerts we keep receiving. As far as I know, ending a release > line requires a majority vote from the PMC—more +1s than -1s. Based on > current sentiment, I believe most active PMC members would support > dropping 3.x. > > -Ayush > > On Sat, 28 Sept 2024 at 09:28, vihang karajgaonkar <vihan...@apache.org> > wrote: > > > Thanks for starting this thread. > > Last I checked, Aman Raj was driving the backports for branch-3. The > tracking ticket for this effort was > https://issues.apache.org/jira/browse/HIVE-26751 However, I don't see > recent activity on that ticket so I am not sure if they are actively > working on this anymore. If not, we should consider if we can make a > release out of branch-3 head. > > Thanks, > Vihang > > On Wed, Sep 25, 2024 at 2:57 PM Stamatis Zampetakis <zabe...@apache.org> > wrote: > > > Hi all, > > Hive 3.1.3 was released on 2022-06-17 and it is the most recent > release of the 3.x line. Two years have passed since then and we still > don't have a new 3.x release. Claiming that we have support for the > 3.x line and not getting releases out is hurting the project in > various ways. > > I know that there has been some backporting activity on branch-3 but > personally I have no visibility on what is pending and what is the ETA > for the next release. > > Is there a realistic plan that could get us a new release in the next > 1-2 months? > > Best, > Stamatis > > [1] https://lists.apache.org/thread/sxcrcf4v9j630tl9domp0bn4m33bdq0s > >
