-
Notifications
You must be signed in to change notification settings - Fork 163
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
HelmRelease: CRDs of disabled subcharts get installed anyway #938
Comments
Ok turns out this is a bug upstream in Helm and never fixed helm/helm#8508 |
@MBrouns if your main chart contains no CRDs, you could skip all CRDs with: apiVersion: helm.toolkit.fluxcd.io/v2beta2
kind: HelmRelease
spec:
install:
crds: Skip
upgrade:
crds: Skip Besides this there is no way we can fix this issue in Flux. In Flux we use the Helm SDK to get the CRDs and the SDK returns all CRDs from all sub-charts before values are compiled. |
Thanks for that lightning-quick response! Unfortunately we do enable other subcharts in the mainchart that have CRDs, so I don't think that'll fix it for us. The weird thing is that if I do a helm install directly of the same manifests, the CRD is not installed. Only when I install the chart through flux it is. Are you suggesting there are differences between the helm cli and helm sdk? |
What version of the Helm CLI are you using? |
version.BuildInfo{Version:"v3.13.3", GitCommit:"c8b948945e52abba22ff885446a1486cb5fd3474", GitTreeState:"clean", GoVersion:"go1.20.11"} |
I don't see how the Helm CLI can do anything different, both the controller and the CLI call this code where all CRDs are merged from all sub-charts and applied: |
I agree that it's weird, but it does seem to be what's happening. Running Let me preface this by saying I'm not very familiar with helm's codebase, but I've done some digging to understand the issue better. I can confirm that when I put a breakpoint here https://github.com/helm/helm/blob/14d0c13e9eefff5b4a1b511cf50643529692ec94/pkg/action/install.go#L256 the length of The crd is initially in the
As far as I can tell though, Flux calls the same
helm-controller/internal/action/install.go Lines 58 to 60 in 9059e7f
|
I looked into it a bit further and took the very blunt approach of calling Can you confirm whether this makes sense? If so, I'll make the linked PR proper and add stuff like test cases |
Got bitten by this as well with splunk-otel-collector chart. A workaround with |
This impacted us as well, victoria metrics operator has dual approach for CRD management where by default it attempts to install CRD's as templated resources owned by the chart. This caused the HelmRelease to fail since CRD objects already exist but not managed by Helm. A workaround is to turn on crds.plain = true for this specific controller however this problem seems to be general in Flux's handling of CRD on install & upgrade. |
Hi, I encountered the same problem when installing grafana/loki: CRDs from the I can also confirm that this is not an issue when I install the same chart using the same values with the Helm CLI ( |
Describe the bug
Assume a helm chart with the following structure
where the mainchart's Chart.yaml says:
When I create a helm release with the following details (specifically note that the subchart is not enabled)
I still see the crd from the subchart created in the cluster. This is only the case for the crd in the subchart, not the resources defined in the subchartpod.yaml. This is not what I expect, and also not what happens when installing the helm chart directly into the cluster
I have tested this for all combinations of Kubernetes 1.26.8, 1.27.8 and 1.28.3 and flux 2.2.0 and 2.2.3
Running
helm install test . --set 'subchart.enabled=false'
directly does not install the subcharts crdSteps to reproduce
kubectl get crd
and observe the crd calledcrontabs.stable.example.com
being createdExpected behavior
I expect no crd to be created for disabled helm subcharts
Screenshots and recordings
No response
OS / Distro
MacOS running Minikube, but also on ranchers K8s
Flux version
v2.2.3
Flux check
► checking prerequisites
✔ Kubernetes 1.28.3 >=1.26.0-0
► checking version in cluster
✔ distribution: flux-v2.2.3
✔ bootstrapped: false
► checking controllers
✔ helm-controller: deployment ready
► ghcr.io/fluxcd/helm-controller:v0.37.4
✔ kustomize-controller: deployment ready
► ghcr.io/fluxcd/kustomize-controller:v1.2.2
✔ notification-controller: deployment ready
► ghcr.io/fluxcd/notification-controller:v1.2.4
✔ source-controller: deployment ready
► ghcr.io/fluxcd/source-controller:v1.2.4
► checking crds
✔ alerts.notification.toolkit.fluxcd.io/v1beta3
✔ buckets.source.toolkit.fluxcd.io/v1beta2
✔ gitrepositories.source.toolkit.fluxcd.io/v1
✔ helmcharts.source.toolkit.fluxcd.io/v1beta2
✔ helmreleases.helm.toolkit.fluxcd.io/v2beta2
✔ helmrepositories.source.toolkit.fluxcd.io/v1beta2
✔ kustomizations.kustomize.toolkit.fluxcd.io/v1
✔ ocirepositories.source.toolkit.fluxcd.io/v1beta2
✔ providers.notification.toolkit.fluxcd.io/v1beta3
✔ receivers.notification.toolkit.fluxcd.io/v1
✔ all checks passed
Git provider
GitHub
Container Registry provider
N/A
Additional context
If anyone's able to point me in the right direction I'd be happy to help fixing this
Code of Conduct
The text was updated successfully, but these errors were encountered: