✨ adding helm chart for csi-driver #398
Conversation
|
I'd prefer that this PR lands in a way similar to how hccm did: by moving the current canonical Kustomize/vanilla manifests into the Helm chart, and then adding a build step somewhere that I can add some commits to this PR that do what I just described, if you like. If you're interested in doing it, then what I imagine is basically just Yes, I'm proposing we go back to how csi-driver k8s manifests looked before someone went in and made a mess of things ;) |
I tested this with the current manifests in the PR (and my review suggestions applied) and found that the labels are currently blocking this. The chart hardcodes the standardized labels in There were some other changes, but they are negligible IMO. Full DiffYour diff will probably be different, I already applied the changes I suggested in a review. $ dyff between <(kustomize build deploy/) <(helm template deploy/charts/hcloud-csi --set fullnameOverride=hcloud-csi --set pdb.create=false --set metrics.enabled=true --namespace kube-system )
_ __ __
_| |_ _ / _|/ _| between /proc/self/fd/11, nine documents
/ _' | | | | |_| |_ and /proc/self/fd/12, nine documents
| (_| | |_| | _| _|
\__,_|\__, |_| |_| returned 39 differences
|___/
(file level)
⇆ order changed
StorageClass/default/hcloud-volumes ServiceAccount/kube-system/hcloud-csi-controller
ServiceAccount/kube-system/hcloud-csi-controller StorageClass/default/hcloud-volumes
ClusterRole/default/hcloud-csi-controller ClusterRole/default/hcloud-csi-controller
ClusterRoleBinding/default/hcloud-csi-controller ClusterRoleBinding/default/hcloud-csi-controller
Service/kube-system/hcloud-csi-controller-metrics Service/kube-system/hcloud-csi-controller-metrics
Service/kube-system/hcloud-csi-node-metrics Service/kube-system/hcloud-csi-node-metrics
Deployment/kube-system/hcloud-csi-controller DaemonSet/kube-system/hcloud-csi-node
DaemonSet/kube-system/hcloud-csi-node Deployment/kube-system/hcloud-csi-controller
CSIDriver/default/csi.hetzner.cloud CSIDriver/default/csi.hetzner.cloud
(root level) (StorageClass/default/hcloud-volumes)
+ one map entry added:
reclaimPolicy: Retain
metadata (ServiceAccount/kube-system/hcloud-csi-controller)
+ one map entry added:
labels:
│ app.kubernetes.io/name: hcloud-csi
│ helm.sh/chart: hcloud-csi-1.0.0
│ app.kubernetes.io/instance: release-name
│ app.kubernetes.io/managed-by: Helm
│ app.kubernetes.io/component: controller
metadata (ClusterRole/default/hcloud-csi-controller)
+ one map entry added:
labels:
│ app.kubernetes.io/name: hcloud-csi
│ helm.sh/chart: hcloud-csi-1.0.0
│ app.kubernetes.io/instance: release-name
│ app.kubernetes.io/managed-by: Helm
│ app.kubernetes.io/component: controller
metadata (ClusterRoleBinding/default/hcloud-csi-controller)
+ one map entry added:
labels:
│ app.kubernetes.io/name: hcloud-csi
│ helm.sh/chart: hcloud-csi-1.0.0
│ app.kubernetes.io/instance: release-name
│ app.kubernetes.io/managed-by: Helm
│ app.kubernetes.io/component: controller
metadata.labels (Service/kube-system/hcloud-csi-controller-metrics)
- one map entry removed: + five map entries added:
app: hcloud-csi-controller app.kubernetes.io/name: hcloud-csi
helm.sh/chart: hcloud-csi-1.0.0
app.kubernetes.io/instance: release-name
app.kubernetes.io/managed-by: Helm
app.kubernetes.io/component: controller
spec.ports.metrics (Service/kube-system/hcloud-csi-controller-metrics)
- one map entry removed:
targetPort: metrics
spec.selector (Service/kube-system/hcloud-csi-controller-metrics)
- one map entry removed: + three map entries added:
app: hcloud-csi-controller app.kubernetes.io/name: hcloud-csi
app.kubernetes.io/instance: release-name
app.kubernetes.io/component: controller
metadata.labels (Service/kube-system/hcloud-csi-node-metrics)
- one map entry removed: + five map entries added:
app: hcloud-csi app.kubernetes.io/name: hcloud-csi
helm.sh/chart: hcloud-csi-1.0.0
app.kubernetes.io/instance: release-name
app.kubernetes.io/managed-by: Helm
app.kubernetes.io/component: node
spec.ports.metrics (Service/kube-system/hcloud-csi-node-metrics)
- one map entry removed:
targetPort: metrics
spec.selector (Service/kube-system/hcloud-csi-node-metrics)
- one map entry removed: + three map entries added:
app: hcloud-csi app.kubernetes.io/name: hcloud-csi
app.kubernetes.io/instance: release-name
app.kubernetes.io/component: node
metadata (Deployment/kube-system/hcloud-csi-controller)
+ one map entry added:
labels:
│ app.kubernetes.io/name: hcloud-csi
│ helm.sh/chart: hcloud-csi-1.0.0
│ app.kubernetes.io/instance: release-name
│ app.kubernetes.io/managed-by: Helm
│ app.kubernetes.io/component: controller
spec (Deployment/kube-system/hcloud-csi-controller)
+ one map entry added:
strategy:
│ type: RollingUpdate
spec.selector.matchLabels (Deployment/kube-system/hcloud-csi-controller)
- one map entry removed: + three map entries added:
app: hcloud-csi-controller app.kubernetes.io/name: hcloud-csi
app.kubernetes.io/instance: release-name
app.kubernetes.io/component: controller
spec.template.metadata.labels (Deployment/kube-system/hcloud-csi-controller)
- one map entry removed: + five map entries added:
app: hcloud-csi-controller app.kubernetes.io/name: hcloud-csi
helm.sh/chart: hcloud-csi-1.0.0
app.kubernetes.io/instance: release-name
app.kubernetes.io/managed-by: Helm
app.kubernetes.io/component: controller
spec.template.spec (Deployment/kube-system/hcloud-csi-controller)
+ two map entries added:
securityContext:
│ fsGroup: 1001
initContainers:
spec.template.spec.containers (Deployment/kube-system/hcloud-csi-controller)
⇆ order changed
- csi-attacher, csi-resizer, csi-provisioner, hcloud-csi-driver, liveness-probe
+ csi-attacher, csi-resizer, csi-provisioner, liveness-probe, hcloud-csi-driver
spec.template.spec.containers.csi-attacher (Deployment/kube-system/hcloud-csi-controller)
+ two map entries added:
resources:
│ limits: {}
│ requests: {}
imagePullPolicy: IfNotPresent
spec.template.spec.containers.csi-resizer (Deployment/kube-system/hcloud-csi-controller)
+ two map entries added:
resources:
│ limits: {}
│ requests: {}
imagePullPolicy: IfNotPresent
spec.template.spec.containers.csi-provisioner (Deployment/kube-system/hcloud-csi-controller)
+ two map entries added:
resources:
│ limits: {}
│ requests: {}
imagePullPolicy: IfNotPresent
spec.template.spec.containers.hcloud-csi-driver (Deployment/kube-system/hcloud-csi-controller)
+ one map entry added:
resources:
│ limits: {}
│ requests: {}
spec.template.spec.containers.hcloud-csi-driver.image (Deployment/kube-system/hcloud-csi-controller)
± value change
- hetznercloud/hcloud-csi-driver:latest
+ docker.io/hetznercloud/hcloud-csi-driver:2.2.0
spec.template.spec.containers.hcloud-csi-driver.imagePullPolicy (Deployment/kube-system/hcloud-csi-controller)
± value change
- Always
+ IfNotPresent
spec.template.spec.containers.hcloud-csi-driver.livenessProbe (Deployment/kube-system/hcloud-csi-controller)
+ one map entry added:
successThreshold: 1
spec.template.spec.containers.liveness-probe (Deployment/kube-system/hcloud-csi-controller)
+ one map entry added:
resources:
│ limits: {}
│ requests: {}
spec.template.spec.containers.liveness-probe.imagePullPolicy (Deployment/kube-system/hcloud-csi-controller)
± value change
- Always
+ IfNotPresent
metadata.labels (DaemonSet/kube-system/hcloud-csi-node)
- one map entry removed: + five map entries added:
app: hcloud-csi app.kubernetes.io/name: hcloud-csi
helm.sh/chart: hcloud-csi-1.0.0
app.kubernetes.io/instance: release-name
app.kubernetes.io/managed-by: Helm
app.kubernetes.io/component: node
spec (DaemonSet/kube-system/hcloud-csi-node)
+ one map entry added:
updateStrategy:
│ type: RollingUpdate
spec.selector.matchLabels (DaemonSet/kube-system/hcloud-csi-node)
- one map entry removed: + three map entries added:
app: hcloud-csi app.kubernetes.io/name: hcloud-csi
app.kubernetes.io/instance: release-name
app.kubernetes.io/component: node
spec.template.metadata.labels (DaemonSet/kube-system/hcloud-csi-node)
- one map entry removed: + five map entries added:
app: hcloud-csi app.kubernetes.io/name: hcloud-csi
helm.sh/chart: hcloud-csi-1.0.0
app.kubernetes.io/instance: release-name
app.kubernetes.io/managed-by: Helm
app.kubernetes.io/component: node
spec.template.spec (DaemonSet/kube-system/hcloud-csi-node)
+ two map entries added:
securityContext:
│ fsGroup: 1001
initContainers:
spec.template.spec.containers (DaemonSet/kube-system/hcloud-csi-node)
⇆ order changed
- csi-node-driver-registrar, hcloud-csi-driver, liveness-probe
+ csi-node-driver-registrar, liveness-probe, hcloud-csi-driver
spec.template.spec.containers.csi-node-driver-registrar (DaemonSet/kube-system/hcloud-csi-node)
+ two map entries added:
resources:
│ limits: {}
│ requests: {}
imagePullPolicy: IfNotPresent
spec.template.spec.containers.hcloud-csi-driver (DaemonSet/kube-system/hcloud-csi-node)
+ one map entry added:
resources:
│ limits: {}
│ requests: {}
spec.template.spec.containers.hcloud-csi-driver.image (DaemonSet/kube-system/hcloud-csi-node)
± value change
- hetznercloud/hcloud-csi-driver:latest
+ docker.io/hetznercloud/hcloud-csi-driver:2.2.0
spec.template.spec.containers.hcloud-csi-driver.imagePullPolicy (DaemonSet/kube-system/hcloud-csi-node)
± value change
- Always
+ IfNotPresent
spec.template.spec.containers.hcloud-csi-driver.livenessProbe (DaemonSet/kube-system/hcloud-csi-node)
+ one map entry added:
successThreshold: 1
spec.template.spec.containers.liveness-probe (DaemonSet/kube-system/hcloud-csi-node)
+ one map entry added:
resources:
│ limits: {}
│ requests: {}
spec.template.spec.containers.liveness-probe.imagePullPolicy (DaemonSet/kube-system/hcloud-csi-node)
± value change
- Always
+ IfNotPresent
|
|
@batistein Would it be okay for you if I take over this PR and make the requested changes myself? |
|
Sure I'm a bit lacking time at the moment! |
|
No worries :) You already spent quite the time getting this up initially. Just want to get this over the finish line and close one item in my backlog 👍 (And maybe clean up my own cluster configuration while i am at it..) |
|
This PR is continued in #500. Thanks a lot @batistein for contributing the initial version :) |
What this PR does:
Things to discuss
Currently, the helm chart is also available in the following repository: https://github.com/syself/charts. While it's possible to use the remote variant of the chart from there, we would like to contribute this helm chart to the "upstream" repository. However, this repository do not have a helm chart publishing process in place.
For the time being, I have left the chart as it is in our helm chart repository so that the community can use a hosted version. I would suggest to revisit this in the future once a publishing process for the helm chart in the csi-driver repository is implemented.