Is there a way to cache the complete layers for copy work

[cowcomic]

With a poor network the copy work often failed A large image needs to download all the layers every time and it will fail every time. I need a way to cache the downloaded layer so that I don’t have to download it again when I copy it again after failure.

[mtrmac]

Thanks for reaching out.

A skopeo copy [--all --preserve-digests] docker://… dir: will create a local copy, which can then be pushed elsewhere. But for that, the whole image download needs to succeed, because a write to dir: wipes the previous contents.

Alternatively, you can run a temporary registry on localhost (maybe in a container); that one does work a layer at a time, and on a failure, restarts the copy from the failed layer (from start).

In recent versions, downloads from registries have a limited resume logic, so rare network interruptions of specific kinds don’t terminate the whole layer operation. But uploads do need to happen whole layers at a time, so if an upload is interrupted, it can only be restarted from scratch.

[cowcomic]

OK，i will try to use registry as a hub

[github-actions[bot]]

A friendly reminder that this issue had no activity for 30 days.

[mtrmac]

Please reopen if there are more questions.