I (finally) got Z2JH running on a 5 node cluster pulling images from a private docker registry using the helm chart.
I have 4 images defined in the profilelist in the chart and but neither the latest tag or a specific version tag will refresh the node image when updated/pushed in the registry and a new version is applied to the config.
The node logs show “registry.lab.myserver.net:5000/all-spark-notebook-custom already present on machine” and I have confirmed that the continuous image puller is running on all nodes.
I have tried deleting the user pod and restarting the deployments but the node image is not updated. The only time it seems to have refreshed is when all nodes are tainted and the cluster is restarted.
Looking for suggestion on where to start troubleshooting or clarification on when to expect node images to refresh.