Support for Documentation On Disabling containerd To Fix 403 Error On Build

[tparikka]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Description

After a lot of searching to address a build error with the docker_image resource I came across #534 and it addressed the issue instantly. I would like to add a note about this to image.md if appropriate to help users understand that disabling containerd is required for docker_image to function correctly, at least on some platforms. I have access to Windows, Linux, and macOS environments and can test to see which if them are impacted to include this in the doc if there are differences.

New or Affected Resource(s)

• image.md documentation

Potential Terraform Configuration

None to speak of.

References

• failed to read downloaded context: failed to load cache key: invalid response status 403 #534

[tparikka]

Opened a PR for this at #665. First time opening a PR on GitHub or for a F/OSS project in a while, if I misunderstood anything from the CONTRIBUTING.md let me know and I'll get it fixed ASAP.

[garysassano]

In addition to updating the Terraform provider docs, which I agree is necessary, I believe BuildKit should offer guidance on resolving the issue, rather than showing a cryptic error message.

[tparikka]

The issue in buildkit has been present since May it sounds like. We could definitely open a second PR to revert the documentation change and keep it in draft format until they're able to fix that issue and containerd works again (documented at moby/moby#47717).

[rumpl]

👋 I just opened a PR in moby that seems to fix this when the containerd image store is enabled moby/moby#49178

[tparikka]

I'm inclined to hold on merging this change then, because if we already have a PR to fix the root cause I don't see a need to rush out a docs update only to have to revert it later.