Currently image-pull-progress-deadline is set to 2m. For larger images or slow pulls from busy registries, this needs to be increased. This results in ImagePullBackOff when the cluster is upgraded and many images are pulled at the same time.

By increasing it to at least 5m, this would solve pulling images when a cluster is upgraded and restarted at the same time.

The issue is when the cluster ends up trying to pull e.g large images simultaneously, all end up timing out.

The issue happens both with Docker registry and Gitlab registry. I tested with images ~2GB, and pulling 5 simultaneously on a droplet takes about 4-5 minutes.

These answers are provided by our Community. If you find them useful, show some love by clicking the heart. If you run into issues leave a comment, or add your own answer to help others.

×
1 answer

That’s a good point! I think that you could submit a bug report via your DO control panel -> go to Kubernetes and on the bottom right there should be a blue horn button that you could click and submit the bug/feature report. That way it would go to the right department.

Submit an Answer