Hi Colin, The idea isn't to hold up 4.0.0 any longer; in fact, it's the opposite. If things slip a bit with 3.8.0 and we take, for example, an extra month to deliver it (or even to cut the branch), I don't think this should necessarily translate to an extra month of latency between now and 4.0.0, given exactly what you mention about the major changes we plan to include in 4.0.0 (which consist more of dropping support for existing things than adding support for new things).
If we want to avoid confusion, we could say something like "no later than 3 to 4 months after the 3.8 branch is created". Frankly though, I think it's unnecessary to specify an exact timeline for 4.0 in this KIP, since nothing in the proposal actually diverges from the usual time-based release plan we follow. The only necessary part seems to be to state that 4.0 will directly follow 3.8 (as opposed to 3.9, 3.10, etc.). But perhaps I'm missing something? Cheers, Chris On Mon, Jan 8, 2024 at 2:38 PM Colin McCabe <cmcc...@apache.org> wrote: > On Mon, Jan 8, 2024, at 09:05, Chris Egerton wrote: > > Hi Josep, > > > > Thanks for the KIP! +1 (binding). > > > > One small nit: I don't think we necessarily have to hold ourselves to > > releasing 4.0.0 "3 to 4 months after 3.8 branch is created" (quoting the > > timeline section of the KIP). IMO it's fine to leave some wiggle room for > > the 4.0.0 release without codifying a timeline in this KIP. Maybe > something > > like "some time after 3.8 branch is created" would be sufficient? > Anyways, > > not a huge thing, I'm sure we'll all give 4.0.0 the flexibility it needs > > with the understanding that this KIP is more focused on 3.8.0 than 4.0.0. > > > > Hmm... I don't see any obstacles in the path of releasing 4.0 after the > traditional 4 months of development. Keep in mind, we're removing things > from the code (the ability to support JDK8, ZooKeeper mode, etc.), not > adding things. We already support JDK11 so saying that it's the minimum is > a very quick change. Similarly, we already support KRaft so saying that > it's the only mode should be a pretty quick change. > > Also, we added a note that "the timeline is very rough" to KIP-833 and it > caued all kinds of confusion. So overall I'd prefer to leave the language > about 4.0 unchanged. > > best, > Colin > > > > > Cheers, > > > > Chris > > > > On Mon, Jan 8, 2024 at 11:41 AM Greg Harris <greg.har...@aiven.io.invalid > > > > wrote: > > > >> Thanks Josep for leading the KIP and building consensus on 3.8! > >> > >> +1 (binding) > >> > >> Greg > >> > >> On Sun, Jan 7, 2024 at 11:45 PM Josep Prat <josep.p...@aiven.io.invalid > > > >> wrote: > >> > > >> > Hi all, > >> > > >> > Thanks for your comments, > >> > I reworded some parts of the KIP to express that: > >> > - The KIP is to agree that we need at least one more minor version in > the > >> > 3.x series > >> > - Explicitly saying that the list of KIPs is not exhaustive and that > if > >> > some are not done, we would need another minor version > >> > - Which are the KIPs/Features the community identified that should be > >> > present in a 3.x version so they can safely migrate to a potential 4.0 > >> > version > >> > - The timeline of 4.0 (starting after branching, not after release) > >> > - Wording is now focusing more on the need to have at least another > minor > >> > release in 3.x to enable and ease migration to a potential 4.0 version > >> > > >> > I always mention potential in terms of 4.0 as we don't know what will > be > >> in > >> > there yet, and this KIP's scope is not meant to define this. > >> > > >> > Best, > >> > > >> > On Fri, Jan 5, 2024 at 10:46 PM Ismael Juma <m...@ismaeljuma.com> > wrote: > >> > > >> > > I agree with Colin. Also, 4.0 would be started after 3.8 is branched > >> (not > >> > > after 3.8.0 is released). The rest looks good. > >> > > > >> > > Thanks! > >> > > > >> > > Ismael > >> > > > >> > > On Fri, Jan 5, 2024 at 11:27 PM Colin McCabe <cmcc...@apache.org> > >> wrote: > >> > > > >> > > > Hi, > >> > > > > >> > > > Thanks for calling the vote, Josep. > >> > > > > >> > > > I re-checked this, and saw something that we missed updating. One > >> thing > >> > > we > >> > > > talked about earlier is that KIP-966 is actually not required for > >> 3.8. > >> > > What > >> > > > is required is some way of enabling unclean leader election by > >> default in > >> > > > KRaft. (Which could be KIP-966, or something else). Please revise > >> this. > >> > > > > >> > > > best, > >> > > > Colin > >> > > > > >> > > > On Fri, Jan 5, 2024, at 02:50, Anton Agestam wrote: > >> > > > > +1 from me. > >> > > > > > >> > > > > Den fre 5 jan. 2024 kl 10:33 skrev Josep Prat > >> > > > <josep.p...@aiven.io.invalid>: > >> > > > > > >> > > > >> Hi all, > >> > > > >> > >> > > > >> I'd like to start a vote on KIP-1012: > >> > > > >> > >> > > > >> > >> > > > > >> > > > >> > https://cwiki.apache.org/confluence/display/KAFKA/KIP-1012%3A+The+need+for+a+Kafka+3.8.x+release > >> > > > >> > >> > > > >> Discussion thread is here: > >> > > > >> > https://lists.apache.org/thread/kvdp2gmq5gd9txkvxh5vk3z2n55b04s5 > >> > > > >> > >> > > > >> Thanks! > >> > > > >> > >> > > > >> --- > >> > > > >> Josep Prat > >> > > > >> Open Source Engineering Director, aivenjosep.p...@aiven.io | > >> > > > >> +491715557497 | aiven.io > >> > > > >> Aiven Deutschland GmbH > >> > > > >> Alexanderufer 3-7, 10117 Berlin > >> > > > >> Geschäftsführer: Oskari Saarenmaa & Hannu Valtonen > >> > > > >> Amtsgericht Charlottenburg, HRB 209739 B > >> > > > >> > >> > > > > >> > > > >> > > >> > > >> > -- > >> > [image: Aiven] <https://www.aiven.io> > >> > > >> > *Josep Prat* > >> > Open Source Engineering Director, *Aiven* > >> > josep.p...@aiven.io | +491715557497 > >> > aiven.io <https://www.aiven.io> | < > >> https://www.facebook.com/aivencloud> > >> > <https://www.linkedin.com/company/aiven/> < > >> https://twitter.com/aiven_io> > >> > *Aiven Deutschland GmbH* > >> > Alexanderufer 3-7, 10117 Berlin > >> > Geschäftsführer: Oskari Saarenmaa & Hannu Valtonen > >> > Amtsgericht Charlottenburg, HRB 209739 B > >> >