Docker course

docker login - to authenticate

docker build -t <dockerhub name>/<image name> . docker build -t nolik2000/docker-r-studio .

docker images

docker run

docker run --rm -p 8787:8787 nolik2000/docker-r-studio

docker run --rm -p 8787:8787 -e USER=myself -e PASSWORD=guest nolik2000/docker-r-studio

--using option -v to map computer and container file system

docker run --rm -p 8787:8787 -e USER=myself -e PASSWORD=guest -v /Users/anastasiia_kostiv/Documents/Udemy_Course_Shared:/home/myself/r-studio nolik2000/docker-r-studio

-- using vim editor to create and edit file

--command chmod +x filename to make file executable

vim Run_RStudio - creating executable file (to start writing press i), to save file press esc, then :wq (w - write, q - quite)

chmod +x Run_RStudio - make file executable

docker login

docker push nolik2000/docker-r-studio

docker save nolik2000/docker-r-studio > docker-r-studio.tar

docker images

docker rmi 03703b737ba3 - remove by IMAGE ID OR By name

docker load --input [name of the archive file]

docker load --input docker-r-studio.tar

docker container ls OR docker ps

Execute bash script in container: docker ps

docker exec -it [CONTAINER ID] bash

! We need to save the state of that image while container is running, so go to terminal:

docker ps

docker commit -m "added package name" [CONTAINER ID] [IMAGE NAME]

docker commit -m "added package name" 6130ce9edade nolik2000/docker-r-studio - now image updated

docker push <dockerhubuser>/<imagename>

docker push nolik2000/docker-r-studio

docker tag nolik2000/docker-r-studio nolik2000/docker-r-studio:Version0.9

docker images

docker push nolik2000/nolik2000/docker-r-studio:Version0.9

--we can link both Docker Hub and GitHub accounts

--change in the code will rebuild the image. Automatically!

--the image locally is not obviously the same is not in Docker Hub, Dockerhub has a newer image because it was built on the Docker Hub Servers, so good to do pull in this case

docker pull nolik2000/docker-r-studio

--We have built docker image directly on Dockerhub via connection GitHub with DockerFile

-- It is usually recommended when we already sure that build will work

docker pull nolik2000/docker-r-h2o

##Test container

docker images

run exec file - inside you can use R bash commands like sd(5), search() etc

docker ps -a

Dockerfile contains sets of instructions to build an image

Command docker build . will build the image on the local computer

Once the image is built we can launch the container with a command docker run ...

If necessary we can update the image from running container by executing a command docker commit ... (from another terminal)

Image can be stored (published) on the Docker Hub docker push repoID/imageID

Image can be saved locally docker save imageID > imageID.tar

Other users could get the image from Docker Hub with a command docker pull repoID/imageID

... or restore it from the archive with docker load --input imageID.tar

Sometimes we can forget to launch our containers by specifying --rm key or there may be intermediate images with a failed build. Those objects may take some space (e.g. 10 GB) so it is a good practice to delete those. The easiest way to delete those is to execute command:

docker system prune

##Note on Docker Compose Typically, in order to run multi-container applications, a dedicated method exists. This is called docker-compose. As this method per se is quite complicated to grasp it will not be covered in this section. Instead, a separate section on docker compose will be published to wrap up the course.

The idea will be to develop a fully functional multi-container application like this:

-Shiny App

-Plumber API

-Database

-Network

These will be assembled and started with one docker compose file.

Stay tuned, keep learning and don't forget to have some fun with this course!

docker network create -d bridge my-net

docker network ls

docker container stop Dev API

docker network rm my-net

docker network inspect my-net