Docker Volume Manager

WARNING

Docker 1.7 introduced some changes that broke functionality in this project. Docker 1.8 introcuded yet more changes that broke functionality here. This has always been intended as a way to prototype this functionality for Docker itself. Being that native docker volumes support is planned for Docker 1.9 and the maintainence burden this project brings, I am no longer going to maintain this. I will happily accept pull requests to fix issues, but will not provide any new fixes. Sorry for the inconveniece, and thank you for helping out with this project!

This is a tool for managing your Docker volumes.

Basically, since volumes are not yet first-class citizens in Docker they can be difficult to manage. Most people tend to have extra volumes laying around which are not in use because they didn't get removed with the container they were used with.

You can run this tool remotely just as you do with the Docker CLI. It reads DOCKER_HOST or you can specify a host using the same syntax as with Docker, with -H unix:///path/to/sock.sock or --host unix:///path/to/sock.sock. This also works with TCP endpoints

Use this to see all your volumes, inspect them, export them, or clean them up.

The primary goal of this project is to spec out UI/API for inclusion in Docker. Some of the implementation is hacky since I specifically wanted the CLI for this to act just like the Docker CLI, that is that it doesn't need to be running on the host or have access to the host's filesystem for it to work.

Caveats

The export function is horribly inefficient, for a couple of reasons:

The tools is inteded to be used remotely, so there is no direct access to the host FS, and as such the volumes or container filesystems.
The docker cp command, and the corresponding APIs, do not support volumes. For instance you cannot do docker cp jolly_torvalds:/path/to/volume like you can for things not in volumes. Well, you can, but it won't be the data in the volume... it will be the data at that location from the container's FS

To work around these issues the export function actually copies data from a volume into a container's FS, then uses the docker cp APIs to pull it.

Installation

You can download a pre-built binary from the releases section. Or you can use the docker image as such:

docker run -v /var/run/docker.sock:/var/run/docker.sock cpuguy83/docker-volumes list

Building

You can use the provided Dockerfile.build which will compile a binary for you or build yourself.

git clone https://github.com/cpuguy83/docker-volumes.git
cd docker-volumes
docker build -t docker-volumes -f Dockerfile.build .
docker run --name docker-volumes docker-volumes
docker cp docker-volumes:/docker-volumes ./

By default when compiling from the Dockerfile it will compile for linux/amd64. You can customize this using environment variables as such:

docker run --name docker-volumes -e GOOS=darwin -e GOARCH=amd64 docker-volumes

This would make a binary for darwin/amd64 (OSX), available for docker cp at the same location as above.

Alternatively, if you already have golang installed on your system you can compile it yourself:

git clone https://github.com/cpuguy83/docker-volumes.git
cd docker-volumes
go get
go build

Usage

Commands:

list - Lists all volumes on the host
inspect - Get details of a volume, takes ID or name from output of list
rm - Removes a volume. A volume is only removed if no containers are using it
export - Creates an archive of the volume and outputs it to stdout. You can optionally pause all running containers (which are using the requested volume) before exporting the volume using --pause
import - Import a tarball generated by the export command from stdin to a specified container. Be default it will import it into the same directory path the volume existed on (eg, if it came from /data, it will put it into /data) You can optionally specify a different volume path, but a volume must exist at that path already or you will get an error

NAME:
   docker-volumes - The missing volume manager for Docker

USAGE:
   docker-volumes [global options] command [command options] [arguments...]

VERSION:
   1.0.0

AUTHOR:
  Brian Goff - <cpuguy83@gmail.com>

COMMANDS:
   list		List all volumes
   inspect	Get details of volume
   rm		Delete a volume
   export	Export a as a tarball. Prints to stdout
   import	Import a tarball produced by the export command the specified container
   help, h	Shows a list of commands or help for one command

GLOBAL OPTIONS:
   --host, -H '/var/run/docker.sock'	Location of the Docker socket [$DOCKER_HOST]
   --help, -h				show help
   --version, -v			print the version

Examples

docker-volumes list

+--------------+-----------------------------+---------------------------------------------------------------------------------------------------+
  f2247e6839b1 | romantic_thompson:/data     | /mnt/sda1/var/lib/docker/vfs/dir/f2247e6839b1ea7ded4123bdc2790e184e6ab6ee3b614fbcb1a72e7bad40e90c
  452484414b40 | romantic_thompson:/moreData | /mnt/sda1/var/lib/docker/vfs/dir/452484414b407c7823357b2cac6812d4e0aa47ca13bd91c68fbe8626760e1adb
  d40880431eb5 | focused_brattain:/data      | /mnt/sda1/var/lib/docker/vfs/dir/d40880431eb5f49a36bba5f5dd5500ae5fc85f9d8d8e4253a7b434302750dead
  f92b748ca057 | insane_feynman:/data        | /mnt/sda1/var/lib/docker/vfs/dir/f92b748ca05768688b41703c2b011520cba7dc2a58acdf10007a83e6c17c5084

# exports volume at /data and sends into ./foo.tar
docker-volumes export insane_feynman:/data > foo.tar

# export and also pause each container using that volume, unpauses when export is finished
docker-volumes export --pause insane_feynman:/data > foo.tar

# pipe in foo.tar and import to the insane_feynman container at the same /data path
cat foo.tar | docker-volumes import insane_feynman

# pipe in foo.tar and import to the romantic_thompson container at the /moreData path
cat foo.tar | docker-volumes import romantic_thompson /moreData

# export from focussed_brattain and pipe directly into the import for insane_feynman
docker-volumes export focused_brattain:/data | docker-volumes import insane_feynman

# export focussed_brattain and pipe into jolly_torvalds at a remote docker instance
docker-volumes export focused_brattain:/data | docker-volumes -H tcp://1.2.3.4:2375 jolly_torvalds