In this part we are going to create an error, and we are going to see how we can notice it in our toolset.
- Let's begin by watching the pods, run the following command from CloudShell:
watch kubectl get pod
Before starting the commit, make sure CloudShell, Grafana and Kubernetes Dashboard, if not go back to step 7 and open them from the last point where we finished the previous part.
- Now it's time to commit the broken image in the repository
i. Open oracle-gitops-workshop repository in your GitHub webpage
ii. Go to clusters/default/flux-system/hello-kubernetes.yaml file
iii. Сlick on the pencil to edit the file
iv. Under the values section and broken image repository and click on Commit changes
apiVersion: helm.toolkit.fluxcd.io/v2beta1
kind: HelmRelease
metadata:
name: hello-kubernetes
namespace: flux-system
spec:
interval: 1m
timeout: 1m
chart:
spec:
chart: ./charts/hello-kubernetes
sourceRef:
kind: GitRepository
name: oracle-gitops-workshop
interval: 1m
targetNamespace: default
values:
message: This is new one
image:
repository: not-exists
valuesFrom:
- kind: ConfigMap
name: myconfig
• Note you can copy & paste the values to the GitHub editor
- Open the following 3 tabs:
Grafana
Kubernetes Dashboard
Helm upgrade failed: timed out waiting for the condition CloudShell
output:
NAME READY STATUS RESTARTS AGE
default-hello-kubernetes-58df79b9f9-tp2kz 0/1 ImagePullBackOff 0 6m50s
default-hello-kubernetes-784d569bb5-5nfpk 1/1 Running 0 37m
default-hello-kubernetes-784d569bb5-5s66w 1/1 Running 0 37m
The more tools you have the more insights about the deployment issue you will get, if you want alerts you can configure it in Grafana, if you want to see failures, you can see the events in the Kubernetes Dashboard. After we see the broken change, we can now revert to fix this. Alternatively, before reverting this change as mentioned in the next step, you can revert to your previous commit before you broke it.
- Revert change of broken image repository
i. Open oracle-gitops-workshop repository in your GitHub webpage
ii. Go to clusters/default/flux-system/hello-kubernetes.yaml file
iii. Сlick on the pencil to edit the file
iv. Under the values remove image section and click on Commit changes
apiVersion: helm.toolkit.fluxcd.io/v2beta1
kind: HelmRelease
metadata:
name: hello-kubernetes
namespace: flux-system
spec:
interval: 1m
timeout: 1m
chart:
spec:
chart: ./charts/hello-kubernetes
sourceRef:
kind: GitRepository
name: oracle-gitops-workshop
interval: 1m
targetNamespace: default
values:
message: This is new one
valuesFrom:
- kind: ConfigMap
name: myconfig
• Note you can copy & paste the values to the GitHub editor
- Now it's time to open the 3 tabs to review the change and everything works fine. Open the following 3 tabs:
Grafana
Kubernetes Dashboard
CloudShell
output:
NAME READY STATUS RESTARTS AGE
default-hello-kubernetes-784d569bb5-5nfpk 1/1 Running 0 42m
default-hello-kubernetes-784d569bb5-5s66w 1/1 Running 0 42m
- Now let's commit an invalid value to break the helm release with remediation Verify that you are in watch in the CloudShell:
watch kubectl get pod
- Commit broken image repository
i. Open oracle-gitops-workshop repository in your GitHub webpage
ii. Go to clusters/default/flux-system/hello-kubernetes.yaml file
iii. Сlick on the pencil to edit the file
iv. Under the values section and broken image repository and click on Commit changes
apiVersion: helm.toolkit.fluxcd.io/v2beta1
kind: HelmRelease
metadata:
name: hello-kubernetes
namespace: flux-system
spec:
interval: 1m
timeout: 1m
chart:
spec:
chart: ./charts/hello-kubernetes
sourceRef:
kind: GitRepository
name: oracle-gitops-workshop
interval: 1m
targetNamespace: default
values:
message: This is new one
image:
repository: not-exists
upgrade:
remediation:
remediateLastFailure: true
valuesFrom:
- kind: ConfigMap
name: myconfig
• Note you can copy & paste the values to the GitHub editor
Grafana
Kubernetes Dashboard
CloudShell
output:
NAME READY STATUS RESTARTS AGE
default-hello-kubernetes-58df79b9f9-kn2sw 0/1 ImagePullBackOff 0 24s
default-hello-kubernetes-784d569bb5-5nfpk 1/1 Running 0 53m
default-hello-kubernetes-784d569bb5-5s66w 1/1 Running 0 53m
Don't worry about the broken pod, we configured remediation in the YAML and the deployment will heal itself.
Excellent, now you have provisioned oke-day2 tools, and you have more eyes on your system.
Continue to Part 9 Sealed Secrets
If you want to return to the workshop homepage: