I recently learned about Docker, so let's store the notes I took and some configuration files in here.
To obtain the list of commands available in docker, simply type docker in the
command line. To obtain more information from a certain command, use the
--help flag.
docker <command> --help```
To download something, type the command below. Where image stands for the
image's name you need to download.
```bash
docker pull imageTo see the images available use:
docker imagesTo run an image you need to specify the image to execute and the tag assigned to the image. If no message pops-up it is because the process has just started and it is waiting.
docker run <image>:<tag>The terminal where this command is executed now executes the container, and
therefore it cannot be used. Unless the flag -d is added to the command.
docker run -d <image>:<tag>Furthermore, to specify the port redirection between the host and the container,
use the -p host-port:container-port flag. By default it uses ports 8080:80.
Several ports can be mapped. To do so, simply add another -p flag.
docker run -p <hp>:<cp> <image>:<tag>
docker run -p <hp>:<cp> -p <hp>:<cp> ... <image>:<tag>Not to just trust the names given by default (harder to remember), you can give
the container your own name by simply using the --name flag.
docker run --name <name>To host some simple static content execute the command:
docker run -v <path-to-content>/<path-suggested-container-developer>:<permissions>You are also able to copy the contents of a container to another. By simply
including the --volumes-from flag.
docker run --volumes-from <container>To see the running containers, run any of the following commands:
docker container ls
docker psTo stop a container, you need to use the CONTAINER ID displayed after
executing the previous command; or simply use the NAME of the container.. The
execution should return the CONTAINER ID/NAME from the stopped contianer.
docker stop <container-id/name>This means that the container has been stopped, by niot deleted. The same
container can be started again by typinng the CONTAINER ID or the NAME.
docker start <container-id/name>To actually delete a container, instead of stopping it, it must be removed.
Regard that a running container cannot be removed. To force it use the -f
flag.
docker rm <container-id/name>
To delete ALL container, command redirection is suggested:
docker rm $(docker ps -aq).
To execute in interactive mode run:
docker exec -it <container> <command>In order to create an image, it is neede d a file that specifies the different
options of that image. That file is called Dockerfile. The Dockerfile must
be created under the root directory.
Go to the
official documentation to
know what to include in the Dockerfile file.
To build that image, execute the command build
docker build <path-to-dockerfile>The image should be displayed after docker image ls.
Pretty much like a .gitignore file.
When building an image, docker recovers the cached steps (unchanged steps that
have already been computed). But, after a certain step is detected as modified
(i.e.: include new source code, should change ADD step) all the following
steps need to be re-executed.
To take advantage of this feature, some steps can be altered or included to avoid that re-execution of consuming resources (i.e.: re-install packages).
Each image usually has multiple tags available. By default, the latest tag is
the one selected for a newly downloaded image.
Other important tag is the alpine tag. It is a light-weight, simple and secure
linux distribution with a small image size. In Docker's images repository it is
given the syntax for every specify image.
Tags allow to control image version. It also avoidsbreaking changes. It is safe.
When creating an image with the same name and tag, the previous tag is
overrided, and it is displayed as <nonw> <none> when showing the list of
images downloaded (docker image ls).
Just execute thw following command:
docker tag <old-repository>:<old-tag> <new-repository>:<new-tag>With latest, a new image will be created. With other images, the command acts as a rename of the image. It allows to have version control.
Highly scalable server side application that stores and lets you distribute Docker images. Used ub CD/CI Pipeline (Continous Delivery/Continous Integration). Run your applications.
There are public and private registries.
Public registries:
- Docker Hub (allows public and private).
- quay.io
- Amazon ECR
To create a new repository, go to hub.docker.com , sign in and create a new repository.
docker push <username>/<repository>:<tag>DISCLAIMER: The username or repository should be part of the image name.
Meaning that the format of the repository column when docker image ls is done
should have that format.
Sometimes, you need to inspect/debug the content of a container. To do so, use:
docker inspect <id/name>Some useful piece of information, are logs. They provide information on traffic, usage, etc.
docker logs <id/name>Using the -f flag leaves a process that blocks the terminal and that is
updated on every log interaction.
docker logs -f <id/name>To directly jump into a container is by simply typing the exec command.
docker exec -it <id/name> /bin/<cmd>