On Fri, Dec 9, 2022 at 7:41 PM Olivier Lamy <ol...@apache.org> wrote:
> On Sat, 10 Dec 2022 at 00:48, Greg Stein <gst...@gmail.com> wrote: > >> On Fri, Dec 9, 2022 at 5:41 AM Olivier Lamy <ol...@apache.org> wrote: > >> > >> Hi > >> Are we seriously activating this gradle enterprise to every build and > >> so breaks a few builds without any notice? > > > > Apologies. We thought it would simply be an additional output, and would > not affect any builds. Gaven has turned it off, and we'll do some more > testing to find out why this happened. > > No worries. issues can happen but when something is touching builds, > it's probably better to announce it. > Understood, and normally we would. We honestly thought this was a "side effect" process that would operate while builds were processed. Gradle,Inc thought the same, too, and are now evaluating options. Again: goal is "nobody should have to change their build process, and we will get some build analysis features" > I don't understand why you guys INFRA do not talk first to the > opensource communities of Maven and/or Jenkins if you need data on > what is happening? > Again: we did not expect a problem. The expectation was that we'd get some data about builds, which we did not have before. That we could do some introspection. ... but primarily: we thought it would be *transparent* ... That didn't happen. But really. We do this all the time. We replace build nodes. Do we talk to the communities? Nah. Cuz upgrading a build node from 50G of disk to 400G of disk should not be a problem. And it never is. We upgrade our Jenkins controllers to new versions. Does that cause problems? Nope. We upgrade build agents from Ubuntu 18.04 to 22.04. No problem. We add latest Maven to the build nodes. Latest JDK. etc etc etc I am not going to fault anybody on the Infrastructure Team for not foreseeing this would break builds. Again: apologies that it did. Mistakes happen. >... Cheers, Greg InfraAdmin, ASF
