4. Creating Persistent Volume on Federation Host Cluster
- Mohammadreza Pasandideh
- Neha Jain
This step is optional and should be followed only if persistent storage for etcd is not disabled using "--etcd-persistent-storage=false" parameter during federation control plane deployment (Refer to step 6).
We intend to create 2 persistent volume (pv) with different "access mode" on the host cluster's Master node (i.e. kubefed-1 here).
#Assuming "enterprise" will be the federation cluster's context name, we pick a relevant name for the pv.
#Create a new file
ubuntu@kubefed-1:~# cat <<EOF > pv-volume1.yaml
kind: PersistentVolume
apiVersion: v1
metadata:
annotations:
volume.alpha.kubernetes.io/storage-class: "yes"
name: enterprise-apiserver-etcd-volume1
namespace: federation-system
labels:
app: federated-cluster
type: local
spec:
capacity:
storage: 11Gi
accessModes:
- ReadWriteMany
hostPath:
path: "/mnt/data"
EOF
ubuntu@kubefed-1:~# kubectl create -f pv-volume1.yaml
persistentvolume "enterprise-apiserver-etcd-volume1" created
ubuntu@kubefed-1:~#
#Create a new file
ubuntu@kubefed-1:~# cat <<EOF > pv-volume2.yaml
kind: PersistentVolume
apiVersion: v1
metadata:
annotations:
volume.alpha.kubernetes.io/storage-class: "yes"
name: enterprise-apiserver-etcd-volume2
namespace: federation-system
labels:
app: federated-cluster
type: local
spec:
capacity:
storage: 11Gi
accessModes:
- ReadWriteOnce
hostPath:
path: "/mnt/data"
EOF
ubuntu@kubefed-1:~# kubectl create -f pv-volume2.yaml
persistentvolume "enterprise-apiserver-etcd-volume2" created
ubuntu@kubefed-1:~#
#verify pv status is "Available"
ubuntu@kubefed-1:~# kubectl get pv --all-namespaces | grep enterprise
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
enterprise-apiserver-etcd-volume1 11Gi RWX Retain Available 35m
enterprise-apiserver-etcd-volume2 11Gi RWO Retain Available 45m
ubuntu@kubefed-1:~#
Mounting directory between site master node and federation master node
This step should be done on all sites that are part of kubernetes control plane (site-1 and site-2). It should be executed on each site's master node using root user. This will mount /dockerdata-nfs directory on each of site's master node.
|