Docker Volumes

Data Persistence in Docker for Stateful Applications

  • Kindly refer the following articles if you are not aware of layer
  • By default, all the files created/modified by the containers reside on the writable layer of the container.
  • The problem with the writable layer of the container is "Writable layer will be deleted once the container is deleted"
  • To persist the data into the Docker Host (Machine on which Docker is installed) Docker provides the following options
    • Volumes:
      • Stored in the hostfile system managed by Docker (/var/lib/docker/volumes/ on Linux).
      • Non Docker Processes should not modify the file system
    • Bind Mounts:
      • Can be stored any where on the host system.
      • Non-Docker Processes can modify
    • tmpfs Mounts:
      • Store on the host systems memory, never written to file system Preview

Docker Volumes

  • Docker Volumes are preferred mechanisms for persisting data generated by and used by Docker Containers.
  • Advantages
    • Easier to backup or migrate that bind mounts
    • Can be managed by Docker CLI or the API
    • Work in both Linux & Windows Containers
    • Volume Drivers lets to store volumes on remote hosts or cloud providers, encryption can be added.
    • Volumes donot increase the size of the docker containers writable layer as the volumes’s contents exist outside of container lifecycle.

Creating and Managing Docker Volumes

  • Two Options Exist for the Creation of Docker Vlume
  • -v:
    • Consists of three fields separated by :.
    • fields order should be correnct
    • syntax: __-v <name of volume>:<path to mount in container>:<options>
  • Lets Experiment with volumes using -v and docker volume command
  • Create a docker volume called as my-tools. Execute the following commands on docker host
docker volume create my-tools

docker volume ls 
local               my-tools

docker volume inspect my-tools
        "CreatedAt": "2019-10-03T13:21:12Z",
        "Driver": "local",
        "Labels": {},
        "Mountpoint": "/var/lib/docker/volumes/my-tools/_data",
        "Name": "my-tools",
        "Options": {},
        "Scope": "local"
  • Mount the docker volume to alpine container using -v command
docker container run --name cont1 -d -v my-tools:/tools alpine sleep 1d
  • Execute inspect command & navigate to mount on container cont1 & execute other commands as shown below
docker container inspect cont1

"Mounts": [
                "Type": "volume",
                "Name": "my-tools",
                "Source": "/var/lib/docker/volumes/my-tools/_data",
                "Destination": "/tools",
                "Driver": "local",
                "Mode": "z",
                "RW": true,
                "Propagation": ""

docker container exec cont1 touch /tools/1.txt
docker container exec cont1 touch /tools/2.txt
docker container exec cont1 touch /tools/3.txt
docker container exec cont1 ls /tools

  • Lets mount the same volume on different container, in different path and see the contents of volume
docker container run --name cont2 -d -v my-tools:/mytools tomcat:8

docker container exec cont2 ls /mytools
  • –mount:
    • Initially was used only for docker services, now can be used for standalone containers as well
    • Syntax: –mount ‘type=<bind/volume/tmpfs>,source/src=<nameofvolume>,destination/dst/target=<mountpath in container>,
    • Refer Here for complete syntax
  • Lets use the same my-tools volume using –mount
  • Mount the docker volume to alpine container using –mount command
docker container run --name cont3 -d --mount 'type=volume,src=my-tools,dst=/tools' alpine sleep 1d
  • Execute the rest of commands as mentioned above

Docker Volume Drivers

  • Docker volume driver can be specified while creating a volume or when to start the container
  • Refer Here for Docker Volume Drivers
  • Refer Here for azure persistent volumes
  • Refer Here for aws persistent data volume


Other Way of Creating Docker Volume

  • In the Dockerfile use the VOLUME instruction.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

About continuous learner

devops & cloud enthusiastic learner