0920a991f9
The `--health-port` option is deprecated and `--http-endpoint` the new option to use. Since `--http-endpoint` additionally exposes metrics, the liveness probe is limited to only listen on localhost by default. With this only processes on the host network, like kubelet, can call the endpoint. Regular pods can no longer access the liveness probe via the node IP address. Signed-off-by: Tobi Nehrlich <tobi.nehrlich@amazee.io>
119 lines
15 KiB
Markdown
119 lines
15 KiB
Markdown
# Install CSI driver with Helm 3
|
|
|
|
## Prerequisites
|
|
- [install Helm](https://helm.sh/docs/intro/quickstart/#install-helm)
|
|
|
|
### Tips
|
|
- install csi snapshot controller: `--set externalSnapshotter.enabled=true`
|
|
- run controller on control plane node: `--set controller.runOnControlPlane=true`
|
|
- set replica of controller as `2`: `--set controller.replicas=2`
|
|
- Microk8s based kubernetes recommended settings(refer to https://microk8s.io/docs/nfs):
|
|
- `--set controller.dnsPolicy=ClusterFirstWithHostNet` with `--set node.dnsPolicy=ClusterFirstWithHostNet` -
|
|
external smb server cannot be found based on Default dns.
|
|
- `--set kubeletDir="/var/snap/microk8s/common/var/lib/kubelet"` - sets correct path to microk8s kubelet even
|
|
though a user has a folder link to it.
|
|
|
|
### install a specific version
|
|
```console
|
|
helm repo add csi-driver-nfs https://raw.githubusercontent.com/kubernetes-csi/csi-driver-nfs/master/charts
|
|
helm install csi-driver-nfs csi-driver-nfs/csi-driver-nfs --namespace kube-system --version v4.6.0
|
|
```
|
|
|
|
### install driver with customized driver name, deployment name
|
|
> only supported from `v3.1.0`+
|
|
- following example would install a driver with name `nfs2`
|
|
```console
|
|
helm install csi-driver-nfs2 csi-driver-nfs/csi-driver-nfs --namespace kube-system --set driver.name="nfs2.csi.k8s.io" --set controller.name="csi-nfs2-controller" --set rbac.name=nfs2 --set serviceAccount.controller=csi-nfs2-controller-sa --set serviceAccount.node=csi-nfs2-node-sa --set node.name=csi-nfs2-node --set node.livenessProbe.healthPort=39653
|
|
```
|
|
|
|
### search for all available chart versions
|
|
```console
|
|
helm search repo -l csi-driver-nfs
|
|
```
|
|
|
|
## uninstall CSI driver
|
|
```console
|
|
helm uninstall csi-driver-nfs -n kube-system
|
|
```
|
|
|
|
## latest chart configuration
|
|
|
|
The following table lists the configurable parameters of the latest NFS CSI Driver chart and default values.
|
|
|
|
| Parameter | Description | Default |
|
|
|---------------------------------------------------|------------------------------------------------------------|-------------------------------------------------------------------|
|
|
| `customLabels` | optional extra labels to k8s resources deployed by chart | `{}` |
|
|
| `driver.name` | alternative driver name | `nfs.csi.k8s.io` |
|
|
| `driver.mountPermissions` | default mounted folder permissions | `0`
|
|
| `feature.enableFSGroupPolicy` | enable [`fsGroupPolicy`](https://kubernetes.io/blog/2020/12/14/kubernetes-release-1.20-fsgroupchangepolicy-fsgrouppolicy/#allow-csi-drivers-to-declare-support-for-fsgroup-based-permissions) on a k8s 1.20+ cluster | `true` |
|
|
| `feature.enableInlineVolume` | enable inline volume | `false` |
|
|
| `feature.propagateHostMountOptions` | use the default host NFS mount configuration file [`/etc/nfsmount.conf`](https://man7.org/linux/man-pages/man5/nfsmount.conf.5.html) and/or the default host `/etc/nfsmount.d` mount configuration directory as source for mount options | `false` |
|
|
| `kubeletDir` | alternative kubelet directory | `/var/lib/kubelet` |
|
|
| `image.nfs.repository` | csi-driver-nfs image | `registry.k8s.io/sig-storage/nfsplugin` |
|
|
| `image.nfs.tag` | csi-driver-nfs image tag | `latest` |
|
|
| `image.nfs.pullPolicy` | csi-driver-nfs image pull policy | `IfNotPresent` |
|
|
| `image.csiProvisioner.repository` | csi-provisioner docker image | `registry.k8s.io/sig-storage/csi-provisioner` |
|
|
| `image.csiProvisioner.tag` | csi-provisioner docker image tag | `v4.0.0` |
|
|
| `image.csiProvisioner.pullPolicy` | csi-provisioner image pull policy | `IfNotPresent` |
|
|
| `image.livenessProbe.repository` | liveness-probe docker image | `registry.k8s.io/sig-storage/livenessprobe` |
|
|
| `image.livenessProbe.tag` | liveness-probe docker image tag | `v2.12.0` |
|
|
| `image.livenessProbe.pullPolicy` | liveness-probe image pull policy | `IfNotPresent` |
|
|
| `image.nodeDriverRegistrar.repository` | csi-node-driver-registrar docker image | `registry.k8s.io/sig-storage/csi-node-driver-registrar` |
|
|
| `image.nodeDriverRegistrar.tag` | csi-node-driver-registrar docker image tag | `v2.10.0` |
|
|
| `image.nodeDriverRegistrar.pullPolicy` | csi-node-driver-registrar image pull policy | `IfNotPresent` |
|
|
| `imagePullSecrets` | Specify docker-registry secret names as an array | [] (does not add image pull secrets to deployed pods) |
|
|
| `serviceAccount.create` | whether create service account of csi-nfs-controller | `true` |
|
|
| `rbac.create` | whether create rbac of csi-nfs-controller | `true` |
|
|
| `controller.replicas` | replica number of csi-nfs-controller | `1` |
|
|
| `controller.runOnMaster` | run controller on master node(deprecated on k8s 1.25+) |`false` |
|
|
| `controller.runOnControlPlane` | run controller on control plane node |`false` |
|
|
| `controller.dnsPolicy` | dnsPolicy of controller driver, available values: `Default`, `ClusterFirstWithHostNet`, `ClusterFirst` | `ClusterFirstWithHostNet` |
|
|
| `controller.defaultOnDeletePolicy` | default policy for deleting subdirectory when deleting a volume, available values: `delete`, `retain`, `archive` | `delete` |
|
|
| `controller.livenessProbe.host ` | the health check host for the liveness probe | `localhost` |
|
|
| `controller.livenessProbe.healthPort ` | the health check port for liveness probe | `29652` |
|
|
| `controller.logLevel` | controller driver log level |`5` |
|
|
| `controller.workingMountDir` | working directory for provisioner to mount nfs shares temporarily | `/tmp` |
|
|
| `controller.affinity` | controller pod affinity | `{}` |
|
|
| `controller.nodeSelector` | controller pod node selector | `{}` |
|
|
| `controller.tolerations` | controller pod tolerations | |
|
|
| `controller.priorityClassName` | controller pod priorityClassName | `system-cluster-critical` |
|
|
| `controller.resources.csiProvisioner.limits.memory` | csi-provisioner memory limits | 100Mi |
|
|
| `controller.resources.csiProvisioner.requests.cpu` | csi-provisioner cpu requests limits | 10m |
|
|
| `controller.resources.csiProvisioner.requests.memory` | csi-provisioner memory requests limits | 20Mi |
|
|
| `controller.resources.livenessProbe.limits.memory` | liveness-probe memory limits | 100Mi |
|
|
| `controller.resources.livenessProbe.requests.cpu` | liveness-probe cpu requests limits | 10m |
|
|
| `controller.resources.livenessProbe.requests.memory` | liveness-probe memory requests limits | 20Mi |
|
|
| `controller.resources.nfs.limits.memory` | csi-driver-nfs memory limits | 200Mi |
|
|
| `controller.resources.nfs.requests.cpu` | csi-driver-nfs cpu requests limits | 10m |
|
|
| `controller.resources.nfs.requests.memory` | csi-driver-nfs memory requests limits | 20Mi |
|
|
| `node.name` | driver node daemonset name | `csi-nfs-node`
|
|
| `node.dnsPolicy` | dnsPolicy of driver node daemonset, available values: `Default`, `ClusterFirstWithHostNet`, `ClusterFirst` |`ClusterFirstWithHostNet`
|
|
| `node.maxUnavailable` | `maxUnavailable` value of driver node daemonset | `1`
|
|
| `node.logLevel` | node driver log level |`5` |
|
|
| `node.livenessProbe.host ` | the health check host for the liveness probe | `localhost` |
|
|
| `node.livenessProbe.healthPort ` | the health check port for liveness probe |`29653` |
|
|
| `node.affinity` | node pod affinity | {} |
|
|
| `node.nodeSelector` | node pod node selector | `{}` |
|
|
| `node.priorityClassName` | node pod priority class name | `system-cluster-critical` |
|
|
| `node.tolerations` | node pod tolerations | |
|
|
| `node.resources.livenessProbe.limits.memory` | liveness-probe memory limits | 100Mi |
|
|
| `node.resources.livenessProbe.requests.cpu` | liveness-probe cpu requests limits | 10m |
|
|
| `node.resources.livenessProbe.requests.memory` | liveness-probe memory requests limits | 20Mi |
|
|
| `node.resources.nodeDriverRegistrar.limits.memory` | csi-node-driver-registrar memory limits | 100Mi |
|
|
| `node.resources.nodeDriverRegistrar.requests.cpu` | csi-node-driver-registrar cpu requests limits | 10m |
|
|
| `node.resources.nodeDriverRegistrar.requests.memory` | csi-node-driver-registrar memory requests limits | 20Mi |
|
|
| `node.resources.nfs.limits.memory` | csi-driver-nfs memory limits | 300Mi |
|
|
| `node.resources.nfs.requests.cpu` | csi-driver-nfs cpu requests limits | 10m |
|
|
| `node.resources.nfs.requests.memory` | csi-driver-nfs memory requests limits | 20Mi |
|
|
| `externalSnapshotter.enabled` | enable snapshot-controller | `false`
|
|
| `externalSnapshotter.name` | name of snapshot-controller | `snapshot-controller`
|
|
| `externalSnapshotter.controller.replicas` | replica number of snapshot-controller | 1
|
|
| `externalSnapshotter.resources.limits.memory` | snapshot-controller memory limits | 300Mi |
|
|
| `externalSnapshotter.resources.requests.cpu` | snapshot-controller cpu requests limits | 10m |
|
|
| `externalSnapshotter.resources.requests.memory` | snapshot-controller memory requests limits | 20Mi |
|
|
|
|
|
|
## troubleshooting
|
|
- Add `--wait -v=5 --debug` in `helm install` command to get detailed error
|
|
- Use `kubectl describe` to acquire more info
|