Open MTRNord opened 8 months ago
As per https://github.com/weaveworks/weave-gitops/issues/4194#issuecomment-1968455968 there's a good reason the account hasn't been paid for. (Weaveworks doesn't exist anymore :( )
Maybe we need to go to GitHub pages for Weave instead? Who maintains this repo at this time?
isn't there any other doc repo?
This also affects the helm chart repo right -> oci://ghcr.io/weaveworks/charts?
No, it would not affect the ghcr.io charts, @Sijoma
We've created a #flux-ecosystem channel on the CNCF Slack instance where this can be discussed, there are some competing interests and someone is likely to re-publish and re-host the Weave GitOps docs soon, (potentially even with their own new branding) but in the mean time it's still possible to read them in the repo, by perusing the "website" directory that contains the docusaurus sources:
https://github.com/weaveworks/weave-gitops/tree/main/website
I'll admit that it's a bit unfriendly as a user interface, but it is at least not MIA.
I think there is still a goal to avoid a fork, and we should try to make this place as hospitable as possible for vendors to avoid creating a bunch of disparate forks. But from a centralization perspective, without maintainers and without formal governance, it will be hard for individuals to contribute to the Weave GitOps repos without forking them.
Anyway @Sijoma what you said is true about the legacy Helm chart repository, if users were using the http-based Helm chart repository (legacy Helm repository) then it is currently out of service.
OCI charts should all still be available (but I have no idea how many of published charts were actually hosted in an OCI repo.)
I have heard from a few different people they are getting good results from a HelmRelease
backed by a GitRepository
who have been testing WGE on their own, without rebuilding everything, for what that's worth! I won't recommend that as a long-term solution, (someone needs to publish and sign their own releases, make themselves known in #flux-ecosystem)
Please direct people to the CNCF channel #flux-ecosystem where we'll share the link, and take your questions and complaints about this refurbished docs site, manually published
We created a temporary (ish?) home for the docs now, https://docs.microscaler.com
Please direct any and all questions about this to the #flux-ecosystem channel linked above. Please do not share this link widely outside of the thread, I cannot guarantee it will stay up in its current form, but am posting it now as you all are interested and it is responsive to the thread here. Thank you for any additional feedback that you can provide!
If there's anything broken or in need of update in the docs link above, please also report it but to keep this thread on-topic maybe you can report those issues on the separate repo.
We also already know about these issues:
The repo that I hosted it from is here: https://github.com/microscaler/gitops-docs (see also https://github.com/kingdonb/weave-gitops / https://github.com/microscaler/weave-gitops)
Am interested in recruiting maintainers who would want to keep this forked repo online long-term. I do not know if this link will stay good, I have hosted it on a personal account for now as I am personally interested in connecting Flux community and ecosystem with interested vendors and maintainers and/or customers, (and really want to see all of this fixed ASAP.)
I started to get 404 errors from the oci repository inside Flux. Without any changes. That's why I thought it was related aswell.
Doesn't seem to be down, but this might not be the same repository that you referenced (the way that GHCR handles sub-naming and chart vs. package is a little weird):
https://github.com/weaveworks/weave-gitops/pkgs/container/charts%2Fweave-gitops
Can you share the HelmRepository and HelmRelease that failed @Sijoma ?
Unfortunately, the 404 is gone since I requested a sync now.
apiVersion: source.toolkit.fluxcd.io/v1beta2
kind: HelmRepository
metadata:
annotations:
reconcile.fluxcd.io/requestedAt: "2024-04-03T18:21:19.285032791Z"
creationTimestamp: "2022-11-03T19:56:58Z"
generation: 3
labels:
kustomize.toolkit.fluxcd.io/name: infrastructure
kustomize.toolkit.fluxcd.io/namespace: flux-system
name: ww-gitops
namespace: flux-system
resourceVersion: "530069380"
uid: f381be98-685b-407a-a7e3-5559d5653c21
spec:
interval: 1h0m0s
provider: generic
timeout: 60s
type: oci
url: oci://ghcr.io/weaveworks/charts
status: {}
For right now, it's good to see that is still online. I wouldn't expect any of our OCI repos on GHCR to go down
But whoever is maintaining Weave GitOps into the future will probably opt to publish into their own (re-branded) chart repo. I'd hope also adopting cosign keyless verification ;)
I think it's true that HelmRepository
of type: oci
does not actually reconcile anything. So without the HelmRelease that goes with that HelmRepository
it would be meaningless to try to test. Anyway you said that it is working for you again now.
For right now, it's good to see that is still online. I wouldn't expect any of our OCI repos on GHCR to go down
Sorry, I meant the error message is gone from the status
but it still is NotReady
@Sijoma Right, that is a bug in Weave GitOps, it does not understand that HelmRepository of type: oci
does not have any status or reconciling, and type: legacy
does indeed reconcile. This is a mild idiosyncrasy that was added in a recent version of Flux, (it's not actually a problem with the HelmRepository. Just a side effect of this item I mentioned above):
I think it's true that HelmRepository of
type: oci
does not actually reconcile anything.
ref:
I've marked most of my comments as off-topic to not de-rail the original issue.
The https://gitops.weave.works docs page is down, I saw an issue somewhere reported that it was being taken down until the search could be brought back online (could someone link the issue where this is being tracked back here)
It seems like something expired on the webhost as its failing with an xml saying "The project to be billed is associated with a delinquent billing account." currently. Is that already known?