This chart will deploy zot (container registry) in the defined Namespace. Using kaniko in a CronJob we will push an image every 15 minutes until the PersistentVolume is full.
Every evening the images are deleted by another CronJob which frees up the space on the PersistentVolume.
We are deploying multiple resources (CronJobs, ServiceAccount, Secrets,ConfigMap) in the default "problem-patterns" Namespace (unless updated in the values.yaml).
All other components will adhere to the specified --namespace in the helm upgrade --install command.
In this demo we fill up a volume by constantly (by default: every 15 minutes) pushing images to the deployed container registry. Eventually the volume is full and the jobs that push the images will start to fail.
This can potentially raise an alert if the workload is monitored.
Unfortunately the used capacity cannot be checked using something as simple as kubectl.
The ability to check the actual usage might be limited depending on the storage provisioner. For cloud-based storage solutions (e.g., AWS EBS, GCP PD), you might need to use the cloud provider's monitoring tools to get detailed usage statistics.
If the above methods are not providing the necessary details, consider integrating a monitoring tool which can be configured to gather detailed storage metrics, including PVC usage.
After building the dependency for zot simply run a helm upgrade command from the root of this repository:
helm dependency build ./pvc-out-of-space
helm upgrade --install --dependency-update pvc-out-of-space ./pvc-out-of-space --namespace pvc-out-of-space --create-namespaceHelm uninstall will get rid of everything but the namespaces. Thus we need to issue a kubectl delete manually to finish the cleanup.
Note
The faultInjection.namespace will be left regardless. It is used by many of our charts, so we will not delete it.
Please remove it manually when you are sure it's not needed anymore.
helm uninstall pvc-out-of-space --namespace pvc-out-of-space
kubectl delete namespace pvc-out-of-spaceWe are using the zot helm chart as a dependency for this deployment. Any required configuration for this demo can be found in the values.yaml.
Please refer to the official documentation for configurations that you might need.