> This gets stated on perhaps an annual basis, so perhaps we should start > having these conversations on wiki to avoid the repetition. I didn't state this in my original email: my goal is to come to a consensus and codify it in the wiki going forward.
> * or do the two version thing and not bother with overlap. To state it back to make sure I understand: In this scenario, you're advocating for us supporting 2 LTS per major but not bothering with overlap. That would put non-containerized bare metal operators in a scenario where they needed to have 2 JDK's in their environment and start the new version of C* w/the new JDK vs. old w/old. This would require the "install all the JDK's you need access to" paradigm which almost all linux distros support quite seamlessly. And for containerized environments, this is a no-op since JDK is bundled. vs. if we require overlap, we'll basically have at worst a year delay between an LTS JDK release and a C* release that supports it, assuming we release C* the day before the new JDK goes LTS. That accurate? Assuming we're up to date on JDK support and we release C* yearly, that doesn't seem *too* bad a tax to pay to me to get the flexibility of having JDK and C* version upgrades decoupled for operators. And as with many other topics, if we had the habit to cut an alpha from trunk once a quarter and were quick about integrating new JDK support, we'd cut that wait down from 12 months to 3. After talking that through, I have a slight preference for the "couple them, consecutive releases must share 1 JDK they support, and we cut alphas". The "cut alphas" thing we've discussed before but didn't hit a consensus and document so that kind of pollutes this thread. If it came down to either: 1. Support latest 2 LTS JDK on any given C* release, no overlap. Operators deal with it. 2. Require 1 overlapping JDK version support on consecutive releases. Users may have to wait a year for new JDK features ... I'm neutral. Both have pros and cons. On Tue, May 20, 2025, at 12:06 PM, Brandon Williams wrote: > On Tue, May 20, 2025 at 10:59 AM Jon Haddad <j...@rustyrazorblade.com> wrote: > > > There is also that recommendation that I keep on hearing - don’t do C* > > > major upgrade and JDK upgrade simultaneously. I believe that was one of > > > the reasons for overlap too > > > > There's no practical reason for this today. Maybe in the Java 6 or 8 days, > > sure. But now, it's a useless requirement. > > If I'm going to encounter a strange bug after upgrading, I'd like the > surface area to be limited to one of C* or the JVM if possible. > > Kind Regards, > Brandon >
