Prereqs
- The host VMs must be hardware version 15 or higher
- Install K8s on VM's must have disk.EnableUUID=1 set.
Step 1.
Create the Rancher cluster with the follow cloud_provider configuration.
cloud_provider:
name: vsphere
vsphereCloudProvider:
global:
insecure-flag: true # Or "false"
virtual_center:
vcenter.example.com:
datacenters: datacenter
port: '443'
user: username
password: password
workspace:
datacenter: /datacenter
default-datastore: datastore
folder: /datacenter/vm
resourcepool-path: /datacenter/host/cluster/Resources/resourcepool
server: vcenter.example.com
Step 2. clone this repo
git clone https://github.com/eroji/cns-installer.git
Step 3. Get your vCenter information from your VI admin. Edit the install.sh file to contain the correct settings.
VCENTER="<vcenter name or IP>"
VC_ADMIN="<vc admin>"
VC_PASS="<vc password>"
VC_DATACENTER="<vc datacentername>"
VC_NETWORK="<vc vm network name>"
Step 4.
# ./install.sh
Step 5. Create StorageClass referencing DataStore.
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
name: cns-vsan
annotations:
storageclass.kubernetes.io/is-default-class: "true" # Or "false" if not default
provisioner: csi.vsphere.vmware.com
parameters:
storagepolicyname: "vSAN Default Storage Policy" # Change this as needed
DatastoreURL: ds:///vmfs/volumes/vsan:<uuid>/
fstype: ext4
Remove all PVC's created with the Storage Class.
kubectl delete pvc <pvcname>
Then run the cleanup script.
./uninstall.sh
You can run kubectl get all --all-namespaces to verify
If the CSI driver for vSphere does not start, the Cloud Controller may not have untainted the nodes when it initilized. I have seen it work automatically and also had to run this:
NODES=$(kubectl get nodes -o name)
for n in $NODES
do
kubectl taint nodes $n node.cloudprovider.kubernetes.io/uninitialized=true:NoSchedule-
done
kubectl describe nodes | egrep "Taints:|Name:"