cloudbox

Tooling for Terraform, Ansible, Kubernetes, AWS, Azure and Google Cloud in a Fedora based Docker/Podman image. The image can be used with the Visual Studio Code Remote - Containers extension and supports various in-container extensions like Azure Tools for Visual Studio Code Visual Studio Code Kubernetes Tools, Terraform Visual Studio Code Extension etc. When Visual Studio Code supports Fedora Toolbox this image might be replaced with a Toolbox setup with all Cloud-tools installed via Ansible.

cloudctl

All available options

Usage: cloudctl [OPTION]...

-b, --build                     Build image from Dockerfile
-f, --force                     Do not use cache when building the image. Only useful with build option.
-d, --docker                    Use Docker CLI instead of Podman
-e COMMAND, --exec COMMAND,
-e=COMMAND, --exec=COMMAND      Run COMMAND in runnning container 'cloudctl'
no option                       Create container 'cloudctl' and start a Bash session
-h, --help                      print help

Use

export CONTAINER_ALIAS='{ docker | podman }'

to set the Container Engine CLI for the current terminal session.

Build image

Build image using build cache.

cloudctl --build
cloudctl -b  # short version

Rebuild image not using the cache at all.

cloudctl --build --force
cloudctl -b -f  # short version

For more information about the build cache see Leverage build cache in the Docker Documentation.

Create container and run Bash session

cloudctl

Run COMMAND in container

# run kubectl in container
cloudctl --exec=kubectl
cloudctl -e kubectl  # short version

# run kubectl with args
cloudctl --exec 'kubectl version'
cloudctl -e='kubectl version'  # short version

Commands with spaces need to be quoted.

Run cloudctl as non-root user

If you do not want to preface cloudctl with sudo, there are options, see Manage Docker as a non-root user. However this impacts security in your system, see Docker Daemon Attack Surface. A better option is to use a rootless container tool like Podman. After basic setup of Podman in a Rootless environment, this can simply be done via alias docker=podman.

Project directory

The host user's ${HOME}/Projects directory is mounted into /home/${CLOUDBOX_USER}/Projects in the container.

If your current/working directory ${PWD} in your IDE or Terminal on the host machine is within that Projects folder, cloudctl changes into that directory on start.

SSH keys

The host user's ${HOME}/.ssh directory is mounted into /home/${CLOUDBOX_USER}/host_ssh in the container, and a symbolic link is created to the cloudbox user's .ssh directory.

If you want to use different ssh keys coming with your project, delete the link and create a new link to a ssh directory in your project:

[bastion]$ rm -f ~/.ssh
[bastion]$ ln -s ~/Projects/devOpsProject/ssh .ssh

Available tools

Terraform

[bastion]$ terraform version

Ansible

[bastion]$ ansible --version

Ansible is using Python 3 and has the following modules installed: dnspython, lxml, netaddr, pypsexec, pywinrm and pywinrm[credssp].

Kubernetes

[bastion]$ kubectl version

AWS Command Line Interface

[bastion]$ aws --version

Azure Command-Line Interface (CLI)

[bastion]$ az --version

Google Cloud Platform - Cloud SDK

gcloud command-line tool

[bastion]$ gcloud version

gsutil tool

[bastion]$ gsutil version

kubectl tool

See Kubernetes.

bq tool

[bastion]$ bq version

kubectl

Sample Shim (Bash script) to provide kubectl in your host system. Simply set a link to - or copy - the shim to a directory in your ${PATH}:

chmod +x kubectl
ln -s /usr/local/bin/kubectl kubectl

and tools in your host IDE, e.g. Visual Studio Code Kubernetes Tools, can find the Kubernetes binary.