K
Q

Deleted Kubernetes Deployment but Node(docker) containers keep recreating themselves

10/19/2018

TLDR: Something keeps recreating containers with an image from my Kubernetes master machine and I cant figure out what!!?!?

I created a deployment(Web Project) and a service(HTTPS Service). The deployment created 3 replica sets of my app 'webProject'.

After I ran kubectl create -f webproject.yml, it spun everything up but then my Docker images got stuck somewhere during 'Rollout'.

So I kubectl delete deployments/webproject which then removed my deployments. I also removed the https service as well.

kubectl get pods
No resources found.

kubectl get deployments
No resources found.

kubectl get services
NAME         TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   10.96.0.1    <none>        443/TCP   3h38m

kubectl get nodes
NAME         STATUS   ROLES    AGE     VERSION
kubemaster100   Ready    master   3h37m   v1.12.1

As you can see it says there are no pods or worker nodes. So when I connect to the worker node to troubleshoot the images, I noticed that it still had containers running with my deployment name.

After I

docker stop 'container'
docker rm 'container'

docker ps
CONTAINER ID        IMAGE                  COMMAND                  CREATED             STATUS              PORTS               NAMES
259a77058d24        39825e5b6bcd           "/usr/sbin/apache2ct…"   22 seconds ago      Up 20 seconds                           k8s_webServer_webServer-deployment-7696fdd44c-dcjjd_default_fcf8fde0-d0c6-11e8-9f67-bc305be7abdb_2

They are instantly getting recreated again. Why?

-- Illuminati
docker
docker-compose
dockerfile
kubernetes

Similar Questions

Kubernetes Rest API - Exec (re)size
Cache results of auth_request module, keyed by username
Prometheus Pods restart in grafana
Is there a way to use Kubernetes RBAC to stop a user from creating an external loadbalancer on a service?

1 Answer

10/20/2018

So if you delete a node on Kubernetes it just deletes it from etcd where Kubernetes keeps its state. However, the kubelet is still running on your node and may hold a cache (not 100% sure about it). I would try:

systemctl stop kubelet

or

pkill kubelet

verify that is not running:

ps -Af | grep kubelet  # should not return anything.

Then stop and remove your container like you did initially.

-- Rico
Source: StackOverflow