HostPath Volume

Simply add hostpath to volumes section in the deployment definition:


      volumes:
- name: local-vol
hostPath:
path: {{ .Values.persistentVolume.path }}
type: DirectoryOrCreate


Example:

kind: Deployment
...
    spec:
      containers:
      - name: registry
...		
        volumeMounts:
        - mountPath: /var/lib/registry
          name: local-vol
          subPath: registry/data
      volumes:
      - name: local-vol
        hostPath:
          path: {{ .Values.persistentVolume.path }}
          type: DirectoryOrCreate
...


HostPath Storage using Persistent Volume 

This is the simplest and best approach for bare metal deployments when a network file system is not available. 


Define storage class and make it the default

helm-template.yaml
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: hostpath-storage
  annotations:
    storageclass.kubernetes.io/is-default-class: "true"
provisioner: keystone/hostpath-storage
volumeBindingMode: Immediate
reclaimPolicy: Retain


Remove default status from other storage classes

> kubectl patch storageclass <STORAGE_CLASS> -p '{"metadata": {"annotations":{"storageclass.kubernetes.io/is-default-class":"false"}}}'


Create Persistent Volumes 

This example creates 10 persistent volumes using helm

helm-template.yaml
{{- $root := . -}}
{{range $i, $e := until 10}}
apiVersion: v1
kind: PersistentVolume
metadata:
  name: pv-{{ $i }}
spec:
  capacity:
    storage: {{ $root.Values.persistentVolume.size }}
  volumeMode: Filesystem
  accessModes:
  - ReadWriteOnce
  persistentVolumeReclaimPolicy: Retain
  storageClassName: hostpath-storage
  hostPath:
    path: /var/pv/pv-{{ $i }}
    type: DirectoryOrCreate
  nodeAffinity:
    required:
      nodeSelectorTerms:
      - matchExpressions:
        - key: kubernetes.io/hostname
          operator: NotIn
          values:
          - master
---
{{end}}


Make a Claim using Default Storage Class

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: my-claim
spec:
  storageClassName: ""
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 2Gi

Local Storage using Persistent Volume and Claim 


We can use disk space on a node by defining a PersistentVolume (see below) and then making a claim against that volume by specifying the storageclass name in the PersistentVolumeClaim.

  • Only one claim can be made against a volume.
  • File path (local.path) must exist for the volume to be usable.
  • USE HOSTPATH STORAGE since it will create the folders for you.


apiVersion: v1
kind: PersistentVolume
metadata:
  name: local-storage
spec:
  capacity:
    storage: 10Gi
  # volumeMode field requires BlockVolume Alpha feature gate to be enabled.
  volumeMode: Filesystem
  accessModes:
  - ReadWriteOnce
  persistentVolumeReclaimPolicy: Persist
  storageClassName: local-storage
  local:
    path: /var/k8s/LOCAL_STORAGE
  nodeAffinity:
    required:
      nodeSelectorTerms:
      - matchExpressions:
        - key: kubernetes.io/hostname
          operator: In
          values:
          - k8sworker1
          - k8sworker2
          - k8sworker3
          - docker-for-desktop


Make a claim by specifying the storage class

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: local-storage-claim
spec:
  storageClassName: local-storage
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 3Gi


Make any required folders on the worker nodes:

ssh k8sworker1

sudo mkdir -p /var/k8s/LOCAL_STORAGE

Repeat for all nodes requiring local storage.


Using Default Storage Class with Prebuilt Persistent Volumes


We can create a storage class for our local-storage and use it as default storage. The only issue with doing this with local-storage is that we need to pre-build all of the persistent volumes. Since only 1 claim can be made against a volume, we will need to make a few.


apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: local-storage
  annotations:
    storageclass.kubernetes.io/is-default-class: "true"
provisioner: keystone/local-storage
volumeBindingMode: Immediate
reclaimPolicy: Retain
---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: local-storage-1
spec:
  capacity:
    storage: 2Gi
  volumeMode: Filesystem
  accessModes:
  - ReadWriteOnce
  persistentVolumeReclaimPolicy: Retain
  storageClassName: local-storage
  local:
    path: /var/pv1
  nodeAffinity:
    required:
      nodeSelectorTerms:
      - matchExpressions:
        - key: kubernetes.io/hostname
          operator: NotIn
          values:
          - master
---
...REPEAT UNTIL HAPPY ...


To use the default storage of the cluster, you just need to create a claim and specify "" for storageClassName.


kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: local-storage-claim
spec:
  storageClassName: ""
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 3Gi


GlusterFS

Gluster-kubernetes is a project to provide Kubernetes administrators a mechanism to easily deploy GlusterFS as a native storage service onto an existing Kubernetes cluster. 

See https://github.com/gluster/gluster-kubernetes

References

  • No labels