[PYTHON] [Docker context] ~ How to access docker in remote environment from VScode ~

Introduction

Access the docker environment in the remote environment using VScode. (The image of the connection is as shown in the figure.) I have been developing with jupyter-lab without knowing this method, but it was very difficult to modify the py file and edit the web code. .. (It can be easily realized just by using docker context ...) It's a very useful method, so please give it a try!

If you find it helpful, I would appreciate it if you could use LGTM m (_ _) m

table of contents

1. ssh connection settings
2. docker context settings
3. Connect to remote docker
4. Return to local docker

1. ssh connection settings

On the client side, enter the ssh connection conditions in the configuration file

~/.ssh/config

Host home-ubuntu
  HostName home-ubuntu.mydns.jp
  IdentityFile ~/.ssh/home_ubuntu
  User shota
  Port 5504
  TCPKeepAlive yes
  IdentitiesOnly yes

• When setting the private key from the ssh server construction, refer to the URL below. https://qiita.com/s-inoue-git/items/fdd7aac2996297424a1f

2. docker context settings

What is docker context? (Quoted from the document below)

A single Docker CLI can have multiple contexts. Each context contains endpoint and security information needed to manage different clusters and nodes. You can easily set and switch between those contexts with the docker context command. Once you've set the context, you can easily switch between contexts by simply running the top-level command docker context use .

Create and switch the context from the following command

# 1.Add context
$ docker context create home-ubuntu --docker "host=ssh://shota@home-ubuntu"
# home-ubuntu
# Successfully created context "home-ubuntu"

# 2.Switching context
$ docker context use home-ubuntu
# home-ubuntu

# 3.Confirm context
$ docker context ls
# NAME                TYPE                DESCRIPTION                               DOCKER ENDPOINT               KUBERNETES ENDPOINT   ORCHESTRATOR
# default             moby                Current DOCKER_HOST based configuration   # unix:///var/run/docker.sock                         swarm
# home-ubuntu *       moby

If you check the image and container with the docker command, you can see that it is switched to docker on the host side

$ docker ps
# CONTAINER ID   IMAGE                     COMMAND       CREATED       STATUS       PORTS     NAMES
# 451ead9000bf   tensorflow-from-scratch   "/bin/bash"   2 hours ago   Up 2 hours             tensorflow-from-scratch

3. Connect to remote docker

Install the following two extensions in VScode

If you click the extended icon of remote connection and click the window mark or folder mark, you can see the docker conteainer on the host side that you switched earlier. Successful access!

4. Return to local docker

When returning the target to the docker to be used locally

$ docker ps
# CONTAINER ID   IMAGE                     COMMAND       CREATED       STATUS       PORTS     NAMES
# 451ead9000bf   tensorflow-from-scratch   "/bin/bash"   2 hours ago   Up 2 hours             tensorflow-from-scratch

#When returning the target
$docker context use default

$ docker ps -a
#CONTAINER ID   IMAGE            COMMAND   CREATED       STATUS                        PORTS     NAMES
#0b0edc2e9df2   conda-init-env   "bash"    2 hours ago   Exited (255) 44 minutes ago             conda-init-env
#↑ Different container is displayed

that's all! If you find it helpful, I would appreciate it if you could use LGTM m (_ _) m