support helm repo add

[weironz]

support helm repo add, like this

$ helm repo add seaweedfs-operator https://seaweedfs.github.io/seaweedfs-operator/helm
Error: looks like "https://seaweedfs.github.io/seaweedfs-operator/helm" is not a valid chart repository or cannot be reached: failed to fetch https://seaweedfs.github.io/seaweedfs-operator/helm/index.yaml : 404 Not Found

i find seaweedfs and seaweedfs-csi-driver is support this

helm repo add seaweedfs https://seaweedfs.github.io/seaweedfs/helm
helm repo add seaweedfs-csi-driver https://seaweedfs.github.io/seaweedfs-csi-driver/helm

[egg82]

Mostly resolved by #97 (and I tested a few minutes ago)

helm repo add seaweedfs-operator https://seaweedfs.github.io/seaweedfs-operator/helm
helm template  seaweedfs-operator seaweedfs-operator/seaweedfs-operator

Note that a few values had to change:

1. image.tag needed to be set to 1.0.2 - the default 0.0.1 will never work, as the image of that version is actually v0.0.1 (and also is two years old)
2. webhook.enabled needed to be set to false as the seaweedfs-operator-webhook-server-cert secret isn't created (I'm assuming it uses cert-manager, but that part doesn't seem to be working) - NOTE: seaweedfs-operator-create-webhook-certificates-xxx does spin up later and seems to create the secret, so after disabling this I suggest re-enabling it. A seaweedfs-operator-patch-mutating-webhook-xxx is also created which will fail in a CrashLoopBackOff state until the webhook is re-enabled.
3. serviceMonitor.enabled needed to be set to false as it required the ServiceMonitor CRD which I don't have

Here's my FluxCD config, in a directory called seaweedfs-operator:

kustomization.yaml

apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization
resources:
  - seaweedfs-operator-namespace.yaml
  - seaweedfs-operator-helmrepository.yaml
  - seaweedfs-operator-helmrelease.yaml

seaweedfs-operator-namespace.yaml

apiVersion: v1
kind: Namespace
metadata:
  name: seaweedfs-operator

seaweedfs-operator-helmrepository.yaml

apiVersion: source.toolkit.fluxcd.io/v1
kind: HelmRepository
metadata:
  name: seaweedfs-operator
  namespace: seaweedfs-operator
spec:
  interval: 1h
  url: https://seaweedfs.github.io/seaweedfs-operator/helm

seaweedfs-operator-helmrelease.yaml

apiVersion: helm.toolkit.fluxcd.io/v2
kind: HelmRelease
metadata:
  name: seaweedfs-operator
  namespace: seaweedfs-operator
spec:
  interval: 1h
  chart:
    spec:
      chart: seaweedfs-operator
      sourceRef:
        kind: HelmRepository
        name: seaweedfs-operator
        namespace: seaweedfs-operator
  values:
    webhook:
      enabled: false
    serviceMonitor:
      enabled: false
    image:
      tag: "1.0.2"

[chrislusf]

Plan to create a PR for this?

[egg82]

What, specifically, would this PR change?

[chrislusf]

Maybe adjust the default values for values.yaml?

[egg82]

I could see serviceMonitor.enabled being set to false and image.tag being updated by default, but the webhook is weird because it's a catch-22. The seaweedfs-operator-webhook-server-cert secret is created by the seaweedfs-operator-create-webhook-certificates-xxx pod, but that pod only spins up after the seaweedfs-operator-xxxx deployment is running, but that deployment uses the secret so it hangs forever.

not having the webhook enabled brings CrashLoopBackoff for the pods that run after (I assume some sort of ConfigMap edit and validation?), and having it enabled causes the deployment to never run in the first place.

The other two could be changed, though, yes.