kuberbs

kuberbs

K8 deployment automatic rollback system

Blog post

Deploy kuberbs (without building from source)

If you just want to use KubeRBS (instead of building it from source yourself), please follow instructions in this section. You need a Kubernetes 1.10 or newer cluster. You can install Kuberbs on any Kubernetes cluster either on perm or in the cloud. Kuberbs supports metrics from DataDog or Stackdriver

GKE Specific Setup

You'll also need the Google Cloud SDK. You can install the Google Cloud SDK (which also installs kubectl) here.

Configure gcloud sdk by setting your default project:

gcloud config set project {your project_id}

Set the environment variables:

export GCP_REGION=us-central1
export GKE_CLUSTER_NAME=kuberbs-cluster
export PROJECT_ID=$(gcloud config list --format 'value(core.project)')

Create IAM Service Account and obtain the Key in JSON format

Create Service Account with this command:

gcloud iam service-accounts create kuberbs-service-account --display-name "kuberbs"

Create and attach custom kuberbs role to the service account by running the following commands:

gcloud iam roles create kuberbs --project $PROJECT_ID --file roles.yaml

gcloud projects add-iam-policy-binding $PROJECT_ID --member serviceAccount:kuberbs-service-account@$PROJECT_ID.iam.gserviceaccount.com --role projects/$PROJECT_ID/roles/kuberbs

Generate the Key using the following command:

gcloud iam service-accounts keys create key.json \
--iam-account kuberbs-service-account@$PROJECT_ID.iam.gserviceaccount.com

Create Kubernetes Secret

Get your GKE cluster credentials with (replace cluster_name with your real GKE cluster name):

gcloud container clusters get-credentials $GKE_CLUSTER_NAME
--region $GCP_REGION
--project $PROJECT_ID

None GKE Specific Setup

Create a Kubernetes secret by running:

kubectl create secret generic kuberbs-key --from-file=key.json --namespace=kube-system

Deploy Kuberbs

kubectl apply -f deploy/.

Create Configuration file

Change the example in examples\kuberbs-example.yaml to fit your needs

Here are the configuration options

watchperiod int - for how long to watch a deployment

metricssource string - currently we support stackdriver and datadog

for each namespace that you would like to watch you can have multiple deployments. Each deployments must have a name, a metric and the threshold per second.

Deploy your configuration file.

Deploy & Build From Source

You need a Kubernetes 1.10 or newer cluster. You also need Docker and kubectl 1.10.x or newer installed on your machine, as well as the Google Cloud SDK. You can install the Google Cloud SDK (which also installs kubectl) here.

Clone Git Repository

Make sure your $GOPATH is configured. You'll need to clone this repository to your $GOPATH/src folder.

git clone https://github.com/doitintl/kuberbs.git $GOPATH/src/kuberbs
cd $GOPATH/src/kuberbs 

Build kubeRBS's container image

Install go/dep (Go dependency management tool) using these instructions and then run

dep ensure

You can now compile the kuberbs binary and run tests

make

Build kuberbs's container image

Compile the kuberbs binary and build the Docker image as following:

make image

Tag the image using:

docker tag kuberbs gcr.io/$PROJECT_ID/kuberbs

Finally, push the image to Container Registry with:

For example

docker push gcr.io/$PROJECT_ID/kuberbs

GKE Specific Setup

Set Environment Variables

Replace us-central1 with the region where your GKE cluster resides and kuberbs-cluster with your real GKE cluster name

export GCP_REGION=us-central1
export GKE_CLUSTER_NAME=kuberbs-cluster
export PROJECT_ID=$(gcloud config list --format 'value(core.project)')

Create IAM Service Account and obtain the Key in JSON format

Create Service Account with this command:

gcloud iam service-accounts create kuberbs-service-account --display-name "kuberbs"

Create and attach custom kuberbs role to the service account by running the following commands:

gcloud iam roles create kuberbs --project $PROJECT_ID --file roles.yaml

gcloud projects add-iam-policy-binding $PROJECT_ID --member serviceAccount:kuberbs-service-account@$PROJECT_ID.iam.gserviceaccount.com --role projects/$PROJECT_ID/roles/kuberbs

Generate the Key using the following command:

gcloud iam service-accounts keys create key.json \
--iam-account kuberbs-service-account@$PROJECT_ID.iam.gserviceaccount.com

Create Kubernetes Secret

Get your GKE cluster credentials with (replace cluster_name with your real GKE cluster name):

gcloud container clusters get-credentials $GKE_CLUSTER_NAME
--region $GCP_REGION
--project $PROJECT_ID

None GKE Specific Setup

Create a Kubernetes secret by running:

kubectl create secret generic kuberbs-key --from-file=key.json --namespace=kube-system

Deploy kuberbs by running

kubectl apply -f deploy/.

Running kuberbs locally

Make sure you can access your cluster

APISERVER=$(kubectl config view --minify | grep server | cut -f 2- -d ":" | tr -d " ")

TOKEN=$(kubectl describe secret $(kubectl get secrets | grep ^default | cut -f1 -d ' ') | grep -E '^token' | cut -f2 -d':' | tr -d " ")

curl $APISERVER/api --header "Authorization: Bearer $TOKEN" --insecure

If you can access your cluster then you need to set up the credentials:

echo $TOKEN >token tr -d '\n' <token >t sudo cp t /var/run/secrets/kubernetes.io/serviceaccount/token ssh into one of the containers in your cluster and get /var/run/secrets/kubernetes.io/serviceaccount/ca.crt. copy that file to your local /var/run/secrets/kubernetes.io/serviceaccount/ca.crt

In the file pkg/clinetset/v1/rbs replace const RbsNameSpace = "kube-system" with const RbsNameSpace = "default"

References:

Event listening code was taken from kubewatch