Updated on 2022-12-01 GMT+08:00

Volume

On-disk files in a container are ephemeral. When a container crashes and is then restarted, the files in the container will be lost. When multiple containers run in a pod, files often need to be shared between these containers. Kubernetes provides an abstraction to solve these two problems, that is, storage volumes. Volumes, as part of a pod, cannot be created independently and can only be defined in pods.

All containers in a pod can access its volumes, but the volumes must be attached and can be attached to any directory in the container.

The following figure shows how a storage volume is used between containers in a pod.

A volume will no longer exist if the pod to which it is attached does not exist. However, files in the volume may outlive the volume, depending on the volume type.

Volume Types

Kubernetes supports multiple types of volumes. The most commonly used ones are as follows:

  • emptyDir: an empty volume used for temporary storage
  • hostPath: a volume that mounts a directory of the host into your pod
  • ConfigMap and secret: special volumes that inject or pass information to your pod. For details about how to mount ConfigMaps and secrets, see ConfigMap and Secret.
  • persistentVolumeClaim: Kubernetes persistent storage class. For details, see PersistentVolume, PersistentVolumeClaim, and StorageClass.

emptyDir

emptyDir is an empty volume in which your applications can read and write the same files. The lifetime of an emptyDir volume is the same as that of the pod it belongs to. After the pod is deleted, data in the volume is also deleted.

Some uses of an emptyDir volume are as follows:

  • scratch space, such as for a disk-based merge sort
  • checkpointing a long computation for recovery from crashes

Example emptyDir configuration:

apiVersion: v1
kind: Pod
metadata:
  name: nginx
spec:
  containers:
  - image: nginx:alpine
    name: test-container
    volumeMounts:
    - mountPath: /cache
      name: cache-volume
  volumes:
  - name: cache-volume
    emptyDir: {}

emptyDir volumes are stored on the disks of the node where the pod is located. You can also set the storage medium to the node memory, for example, by setting medium to Memory.

volumes:
  - name: html
    emptyDir:
      medium: Memory

HostPath

hostPath is a persistent storage volume. Data in an emptyDir volume will be deleted when the pod is deleted, but not the case for a hostPath volume. Data in a hostPath volume will still be stored in the node path to which the volume was mounted. If the pod is re-created and scheduled to the same node, after a new hostPath volume is mounted, previous data written by the pod can still be read.

Data stored in hostPath volumes is related to the node. Therefore, hostPath is not suitable for applications such as databases. For example, if the pod in which a database instance runs is scheduled to another node, the read data will be totally different.

Therefore, you are not advised to use hostPath to store cross-pod data, because after the pod is rebuilt, it will be randomly scheduled to a node, which may cause inconsistency when data is written.

apiVersion: v1
kind: Pod
metadata:
  name: test-hostpath
spec:
  containers:
  - image: nginx:alpine
    name: hostpath-container
    volumeMounts:
    - mountPath: /test-pd
      name: test-volume
  volumes:
  - name: test-volume
    hostPath:
      path: /data