The same problem as #25585, and the fix landed #26418, @erez-rabih ptal.

@bgrant0607 I am wondering whether we should create a cherry pick for 1.2, I have seen users ran into this bug several times.

@adohe I agree it looks like a duplicate of #25585. I'm closing this.

@adohe Yes, please create a cherrypick PR.

Hi! I'm facing the same issue now: adding a new env variable does not touch running pods, the new env varibale is not puplished down to the pods.

Client Version: version.Info{Major:"1", Minor:"10", GitVersion:"v1.10.2", GitCommit:"81753b10df112992bf51bbc2c2f85208aad78335", GitTreeState:"clean", BuildDate:"2018-04-27T09:22:21Z", GoVersion:"go1.9.3", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"10", GitVersion:"v1.10.2", GitCommit:"81753b10df112992bf51bbc2c2f85208aad78335", GitTreeState:"clean", BuildDate:"2018-04-27T09:10:24Z", GoVersion:"go1.9.3", Compiler:"gc", Platform:"linux/amd64"}

@danielloczi I think this has already be fixed, could u please show how to reproduce this? I would take a look.

cc @apelisse

Hi!

I have a set of environment variables configured for a deployment. The deployment is applied using kubectl apply -f deployment.yaml
After I edit the yaml file and add an extra env varibale, then apply it using the same command:
kubectl apply -f deployment.yaml
After that kubectl logs that the deployment is configured, but the pod not terminates, and the new env variable is not promoted to the pod level.

Example (original yaml):

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: vibe-backend-identityprovider-deployment
spec:
  replicas: 1
  template:
    metadata:
      labels:
        app: my-app
    spec:
      containers:
      - name: my-ctr
        image: localhost:5000/myimage
        ports:
        - containerPort: 5202
        env:
        - name: RepositoryUri
          value: "http://my-repo-service.default.svc.cluster.local:5203"
        - name: LogUri
          value: "http://my-messaging-service.default.svc.cluster.local:5201/log"

Edited yaml:

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: vibe-backend-identityprovider-deployment
spec:
  replicas: 1
  template:
    metadata:
      labels:
        app: my-app
    spec:
      containers:
      - name: my-ctr
        image: localhost:5000/myimage
        ports:
        - containerPort: 5202
        env:
        - name: RepositoryUri
          value: "http://my-repo-service.default.svc.cluster.local:5203"
        - name: LogUri
          value: "http://my-messaging-service.default.svc.cluster.local:5201/log"
        - name: MyNewEnv
          value: "something"

Update: I tried to reproduce it a couple of minutes ago, but I was unable. I trying to find the case, when it does not work.

Got this now in 1.11 on a DaemonSet.

Got this now in 1.11 on a DaemonSet.

Same for me

also having the issue with DaemonSet

Client Version: version.Info{Major:"1", Minor:"10", GitVersion:"v1.10.11", GitCommit:"637c7e288581ee40ab4ca210618a89a555b6e7e9", GitTreeState:"clean", BuildDate:"2018-11-26T14:38:32Z", GoVersion:"go1.9.3", Compiler:"gc", Platform:"darwin/amd64"}
Server Version: version.Info{Major:"1", Minor:"10+", GitVersion:"v1.10.13-eks-g484b8", GitCommit:"484b857e3134d55ac6373fea2f51798fefa0533f", GitTreeState:"clean", BuildDate:"2019-03-08T05:32:36Z", GoVersion:"go1.10.8", Compiler:"gc", Platform:"linux/amd64"}

why this issue is closed? on latest version of kubernetes i am facing the same issue.

updating the deployment.yaml with adding new env var , and then kubectl apply , is not updating the deployment/pod with new env var.

@iahmad-khan Would you mind coming-up with a minimal example that I would be able to build into a test or something?

I'd need to see:

• The exact sequence of event,
• How the object is impact on the server including the generation number.

Any fixes, seems like, I'm running commands, not in the correct order.

I updated env-config and deleted deployment and pods and all the services and created again with apply, still not picking new configs

Can you show the steps here with a minimal example, the commands that you run and how the yaml looks like after each step?

hitting this also sometimes, might be control plane &/ persistent problem

Once this occurs , it persists for me even i delete deployment & service to recreate both in following orders

1. Create a deployment & service exposing it
   (set some vars not yet existed on the deployment)
2. kubectl set env deploy/xxxxxx --containers=XXXX AAA=BBB
   3a)(check pods started &/ delete them to let replicaset recreate pods, see if vars exist )
   3b) [delete both deployment & service to further reproduce , still can't set]

OK Let's try to reproduce this:

$ kubectl create deployment --image nginx:latest nginx
deployment.apps/nginx created
$ kubectl set env deployment nginx --containers=nginx MY_ENV=1234
deployment.apps/nginx env updated
$ kubectl set env deployment nginx --containers=nginx MY_ENV_2=5678
deployment.apps/nginx env updated
$ kubectl get -o yaml deployment nginx | grep -A 5 env
      - env:
        - name: MY_ENV
          value: "1234"
        - name: MY_ENV_2
          value: "5678"
        image: nginx:latest
$ kubectl get pods nginx-7db6cbd99f-pwkxq -o yaml | grep -A 5 env 
  - env:
    - name: MY_ENV
      value: "1234"
    - name: MY_ENV_2
      value: "5678"
    image: nginx:latest

I still don't see what I should be looking at?