k8s-git-autodeploy

This tool can help you manage the yaml (JSON) files in Git and automatically deploy them to the k8s cluster, so as to realize file to version control, facilitate team management and avoid manual deployment.

Premise

• Make sure that you can access the k8s cluster network
• you can operate the k8s cluster by serviceaccount in pod

Installation and Getting Started

• Clone the code
• In the project root catalog, run mvn clean package
• Run docker build -t [imageName]:[imageTag] .

then you can run this image in your local or in k8s cluster.These methods are described below.

K8S cluster(recommended)

• deploy this image to your k8s cluster as a deployment
• create a service for the deployment

Local by docker (not recommended, just for testing)

• Run docker run -e [some envs] [imageName]:[imageTag]
• Run docker exec -it [container id] bash
• Configure the config file to $HOME/.kube/
• Test whether the kubectl command can obtain the k8s cluster information

Local by java (not recommended, just for testing)

• Configure the config file to $HOME/.kube/
• Test whether the kubectl command can obtain the k8s cluster information
• Run java -jar [your jar]

Two modes

1. Git webhook: through the web hook function of git, the modification request is sent to the specified URL.
2. Timing git pull: set the pull interval to realize automatic deployment

Env Sets

Name	Required	Settable Value	Mode	Description
REMOTE_REPO_URL	Y			repo http address
BRANCH	Y			your branch name(master is recommended)
LOCAL_PATH	Y			git file path in your local or pod
USERNAME	Y			repo username
PASSWORD	Y			repo password
AUTODEPLOY_TYPE	Y	webhook/timing		use mode
TIME_INTERVAL_OF_PULL	N		timing	time interval of pull
NEED_PROXY	N			if need proxy
HTTP_PROXY	N			proxy address
HTTP_PORT	N			proxy port

Attention

For git modification, only add, modify and delete are supported now, because I consider that rename and copy are not very necessary for managing files.

Todo

• add deployment.yaml and service.yaml
• retry