You can choose to either use the specific pieces below to only clean them up to reinstall or re-test those pieces in your cluster, or follow the cluster cleanup instructions to destroy all EKS/AWS/cluster resources.
- To uninstall all Redis resources in all clusters and associated services (local and federated) and reset back to a clean namespace, run the uninstall script:
bash redis-ha/uninstall-redis.sh
- To remove all policies and tiers and allow all traffic in the cluster, delete the policies based on your cluster context:
kubectl delete -f federated-policy/cluster-1-policy
kubectl delete -f federated-policy/cluster-2-policy
- Depending on cluster context, delete the namespaces and the default namespace
nginx
application:kubectl delete -f demo-apps/01-namespaces.yaml
kubectl delete -f demo-apps/02-namespaces.yaml
kubectl delete -f demo-apps/40-nginx-deploy.yaml
kubectl delete -f federated-svc/nginx-federated.yaml
- This will also clean up any associated services of type
LoadBalancer
and the NLBs in AWS which is necessary before attempting to destroy the EKS clusters.
- To uninstall the cluster-mesh in the clusters and clean up associated secrets and objects, run the script:
bash teardown-federation-overlay.sh
-
Clean up any remaining services of type
LoadBalancer
in all clusters (if they still exist after deleting Redis HA or the other apps) usingkubectl delete svc -n <namespace> <svc-name>
-
Delete VPC peering connection/s:
-
Get the peering id:
PEER_ID=$(aws ec2 describe-vpc-peering-connections --region <your-region-code> --query "VpcPeeringConnections[0].VpcPeeringConnectionId" --output text)
-
Delete the peering connection/s:
aws ec2 delete-vpc-peering-connection --region <your-region-code> --vpc-peering-connection-id $PEER_ID
-
-
Clean up the EKS clusters using eksctl:
- Use
eksctl get cluster --region=<your-region>
to get the clusters that were spun up as part of this repo. - Then use
eksctl delete cluster --region=<your-region>
to delete the cluster. Repeat this for all clusters.
- Use