How to debug docker push to google cloud artifacts registry error "manifest unknown: Requested-CodePudding

We push images from gitlab-ci to google artifacts registry in order to deploy to google cloud run and it suddenly stopped working.

the script is

- echo $SERVICE_ACCOUNT_KEY > ./____keyfile.json
- gcloud auth activate-service-account --key-file="./____keyfile.json"
- gcloud auth configure-docker europe-west1-docker.pkg.dev
- docker pull $DOCKER_IMAGE:$DOCKER_IMAGE_TAG # pulls the previously build image that was pushed to gitlab's own registy
- docker tag $DOCKER_IMAGE:$DOCKER_IMAGE_TAG europe-west1-docker.pkg.dev/project-id/repository/$DOCKER_IMAGE_NAME:$DOCKER_IMAGE_TAG
- docker push europe-west1-docker.pkg.dev/project-id/repository/$DOCKER_IMAGE_NAME:$DOCKER_IMAGE_TAG

the last push shows:

The push refers to repository [europe-west1-docker.pkg.dev/project-id/repository/some-name]
bcf45157cf76: Preparing
151948c565a1: Preparing
5455690f232d: Preparing
65f565da952a: Preparing
66b19dc4e200: Preparing
e796b82d6587: Preparing
dcd9c7cf433b: Preparing
1f1a410f80c7: Preparing
b45078e74ec9: Preparing
e796b82d6587: Waiting
dcd9c7cf433b: Waiting
1f1a410f80c7: Waiting
b45078e74ec9: Waiting
151948c565a1: Layer already exists
65f565da952a: Layer already exists
bcf45157cf76: Layer already exists
66b19dc4e200: Layer already exists
5455690f232d: Layer already exists
e796b82d6587: Layer already exists
dcd9c7cf433b: Layer already exists
b45078e74ec9: Layer already exists
1f1a410f80c7: Layer already exists
manifest unknown: Requested entity was not found.

unfortunatly this error message "manifest unknown: Requested entity was not found." is very generic and does not show what went wrong. Is there a way to show more verbose logs?

CodePudding user response：

Altough I did not manage to get more information from docker push, the error seems to have been caused by some outage on google's side in the specific region (in that case europe-west1). Switching the region resolved the issue.

CodePudding user response：

I Asked GCP Support about this issue. Here is their response :

Thank you for your patience. The product team got back to us and they were able to confirm that this is indeed an issue that affected the region you mentioned. They have notified me that the cause of this issue has since been rolled back. Can you please confirm if this has been fixed for you, or whether you are still observing the same behavior ?

It's Fixed for me