Hacker News Reader

Comment by edoceo | Hacker News Reader

https://www.weave.works/ "Operate and Manage Kubernetes easily" / "GitOps Automation for Kubernetes Stacks"

Probably most popular thing they built was FluxCD if you are confused where you have heard their name before.

Their company provided a ton of extra tooling and consulting around the product. Hopefully as FluxCD user, this doesn't impact the development of Flux.

Right. Flux was a handy little tool[1] that sync'd yaml manifests in git repos to live clusters. The concept was fascinating, and the tool was well done--small and efficient. Easy to learn.

In 2019, they announced they'd be "merging" with argocd[2]. It seems the merge never really took place, and after that they deprecated flux and announced flux2[3].

The sudden changes of course were a little confusing and perhaps not too well communicated.

1: https://github.com/fluxcd/flux 2: https://discuss.kubernetes.io/t/flux-cd-joins-forces-with-ar... 3: https://github.com/fluxcd/flux2

Sudden? We finally archived the repo at the end of 2022.

https://fluxcd.io/flux/migration/timetable/

I was hired to support Flux v1 two years after the events you described, in the beginning of 2021 to go on supporting Flux v1 until we could get everyone off the boat.

(I worked at Weaveworks until last month, and I'm still a Flux maintainer! Keep the Flux talk in Present tense please! ;-)

Yes, present tense is more appropriate. Best of luck to you on what's next.

People like the weave network policy controller too.. at least in the earlier days of kubernetes, their cni was pretty popular for awhile.

That was my first recollection -- after CoreOS' Flannel, I think the next two (or at least two of the earliest) overlay networks available for Kubernetes were Weave Net and Calico (whose core maintainers started Tigera). Flannel is still around and under active development despite CoreOS being long gone; hopefully the tools Weave employees maintained can keep healthy communities going around them.

I think people got turned off from them when it came to light they had invented their own cryptography without doing any vetting that it was actually secure. At least, that’s why we moved away from it as quickly as possible with our reasoning being: “if someone is foolish enough to roll their own crypto, what else are they doing foolishly?”

Despite having been the main maintainer of Weave Net for ~4 years, I have no idea what you are talking about. Weave Net used Google's NaCL library from day 1; later adding optional AES using the in-kernel implementation.

https://github.com/weaveworks/weave/blob/master/site/concept...

The key exchange was/is totally custom without using anything standard, though it uses standard concepts (DH, though it does some non-standard things with it), which *might* be secure, but as far as I know, never actually put to the test.

We were using the weave CNI on our kubernetes cluster. It started failing, so we called weaveworks to try to buy support. They said no, recommending a different CNI which would necessitate a disruptive cluster replacement. Once that was a necessity, all the other CNIs were on the table and we went with cilium, which has been ok with k8s 1.22 on CentOS 7, but which is now having undiagnosed trouble with k8s 1.24 on Red Hat 8. The messy process of figuring that out is underway.

I hope so too. Flux is the best Kubernetes Gitops toolkit IMO.

thanks, I didn't recognize the name

Reply on news.ycombinator.com