🌱 update controller runtime and cluster-api to newest version #1623
Conversation
|
@cgroschupp I created a PR based on yours. 🌱 update controller runtime and cluster-api to newest version by guettli · Pull Request #1628 · syself/cluster-api-provider-hetzner I get a strange linting error in my PR I guess we need to update from golangci-lint 1.x to 2.x Does |
|
@cgroschupp thank you for your PR. Currently the CI for Linting and Tests fail. I updated your PR in my PR (which is based on your version): 🌱 update controller runtime and cluster-api to newest version by guettli · Pull Request #1628 · syself/cluster-api-provider-hetzner I was able to fix the golangci-lint error by syncing the dependencies to the version used by cluster-api (k8s and cel). Currently we are unsure how to proceed. The linter complains about a lot of small things. Either we work on that, or you do it. Do you have time during the next days to work on the PR? |
|
@cgroschupp thank you for your initial PR. I will continue with it here: #1628 |
|
@guettli Sorry, I was on holiday. Now 'make lint' is running. see here: https://github.com/cgroschupp/cluster-api-provider-hetzner/tree/update-controller-runtime |
|
Have you merged the changes of my PR into yours? |
|
I don't have write access to your branch. I added some commits to my branch. |
|
@cgroschupp please base on my branch. Tell me if there are reasons against doing it. Then please continue on your branch. Ping me, when you are ready. Then I will merge your branch into mine, so that we run the e2e tests. I am sorry, it is a bit complicated, because currently e2e tests are not available for external PRs. |
|
reopening. |
…yself#1635) build(tools): make helm and hcloud downloads platform-agnostic Replace hardcoded `linux-amd64` architecture with dynamic `$(go env GOOS)-$(go env GOARCH)` to support multiple platforms and architectures during tool installation. Signed-off-by: Tomas Dabašinskas <[email protected]>
…yself#1637) * 🌱 Add SkipCreatingHetznerSecretInWorkloadCluster to hetznercluster_controller SkipCreatingHetznerSecretInWorkloadCluster indicates whether the Hetzner secret should be created in the workload cluster. By default the secret gets created, so that the ccm (running in the wl-cluster) can use that secret. If you prefer to not reveal the secret in the workload cluster, you can set this to value to false, so that the secret is not created. Be sure to run the ccm outside of the workload cluster in that case, e.g. in the management cluster. Closes syself#1636
|
@guettli rebase is done. Two tests are currently failing. I need some time to find the problem. Perhaps you have an idea. |
|
@cgroschupp The good news: These two errors are stable. You have them, they are in CI, and I have them locally. I have a look at it now. |
|
@cgroschupp BTW, I started a TODO list at the top of my PR: 🌱 update controller runtime and cluster-api to newest version by guettli · Pull Request #1628 · syself/cluster-api-provider-hetzner |
…nto update-controller-runtime
|
@cgroschupp it would be great, if you could have a look at this deprecation: remove |
Context: kubernetes-sigs/cluster-api#10784 We'll maintain them here for the time being, until we have conditions replacing these errors.
|
@guettli 427cb76 this is only a temporary solution. In v1beta2, the FailureReason and FailureError fields should no longer be used; conditions should be used instead. We can probably remove the SetFailure calls completely, but this would constitute a breaking change and should be implemented when cluster-api-provider-hetzner switches to v1beta2. See proposal: https://github.com/kubernetes-sigs/cluster-api/blob/main/docs/proposals/20240916-improve-status-in-CAPI-resources.md |
|
@cgroschupp thank you. I pulled your changes to my branch to trigger CI. |
What this PR does / why we need it:
Updated controller-runtime to 0.21.0 and cluster-api to 1.10.3.
To get the newest from controller-runtime, for example the priority queue and support k8s 1.32 and 1.33