You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I may not be following. Are you looking to use Spegel to fetch images from S3 or to store pulled images in S3? I am aware that there are people that use S3 as an upstream registry by storing layers to specific paths.
Could you expand on the use case for my edification?
sure , what I meant is to pull images from standard registry and then store it in s3 and let pods mount and read the images from s3 . Without having to consume the ephemeral storage of node disk
This is out of scope for Spegel, as it only intends to piggy back off of the CRI provider storage. Using a S3 bucket to store the layers to a degree defeats the purpose of Spegel, as it will now rely on an external storage provider. The goal is to keep Spegel as simple as possible and instead integrate well with other projects instead for these types of use cases.
I think you would be better off using a pull through mirror like Harbor instead. That way you could use Spegel and Harbor in conjunction with each other.
Describe the problem to be solved
Allow storing and mounting docker images to s3 buckets in k8s environments .
If an image is pulled it should be stored in s3 and pods will use the image data from the bucket
Proposed solution to the problem
can possibly use s3fs , where during this project initialization in cluster , the docker path in the node will be mounted to s3
The text was updated successfully, but these errors were encountered: