no matching manifest for linux/amd64 in the manifest list entries

[diogosimao]

Environment information

• MWDB version (from /about): v2.6.0
• Installation method:
  • [ ] mwdb.cert.pl service
  • [ ] From PyPi (pip install mwdb-core)
  • [ ] From docker-compose
  • [x] Other (docker hub image pull)
• Plugins installed: None

Behaviour the bug (what happened?)

An error occurs when que command docker pull certpl/mwdb:v2.6.0 is executed

Expected behaviour

I expected a succefull pull output like:

v2.6.0-dev1: Pulling from certpl/mwdb-web Digest: sha256:a58ffb737802549082df2fa633c84282674a0419294629456bfc28e3f16d6e33 Status: Image is up to date for certpl/mwdb-web:v2.6.0-dev1 docker.io/certpl/mwdb-web:v2.6.0-dev1

Reproduction Steps

Executing docker pull certpl/mwdb:v2.6.0

Screenshots

Additional context

[psrok1]

Thanks for bug report! And uh, that looks weird. Somehow stable images tagged as v2.6.0 are completely empty on Docker Hub

I will inspect the Github Actions workflow if it's our fault or we just triggered some temporary bug while publishing the release.

[psrok1]

@diogosimao: Ok, I pushed correct images for the v2.6.0 release to Docker Hub and tested locally. Would you like to check if it works for you as well?

I completely don't know what happened but our pipeline has started producing empty images that were pushed to Docker registry. I turned off all caching for release pipeline and it looks to work correctly right now (#526).

Maybe it's somehow related with another GHA cache issue that has broken our pipeline recently (https://github.com/CERT-Polska/mwdb-core/pull/516). Unfortunately Github Actions are not very stable yet, so these problems happen from time to time. I need to investigate further whether it's my fault (e.g. by violating some not very well-documented constraint that GHA cache should not be shared across jobs) or it's just a known bug.

[diogosimao]

Awesome :D Works like a charm.