AWS SAM Deploy Missing option '--image-repository' or '--image-repositories'

Question

Hi all, I am trying to deploy to Lambda using `aws-sam-deploy` pipe and got an "Missing option '--image-repository' or '--image-repositories'" error when packaging. However, I have included "image_repositories" inside my `samconfig.toml` file and set SAM_CONFIG to `samconfig.toml` file. I can build and deploy to AWS Lambda using Docker images from local without any problems with the command `sam build` and then `sam deploy`. So I do not think there's anything wrong with my `template.yaml` or `samconfig.toml` What steps am I missing? Any help would be greatly appreciated. Thank you!! HCL bitbucket-pipeline error message: INFO: Using default authentication with AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY. INFO: Start package... SAM CLI now collects telemetry to better understand customer needs. You can OPT OUT and disable telemetry collection by setting the environment variable SAM_CLI_TELEMETRY=0 in your shell. Thanks for your help! Learn More: https://docs.aws.amazon.com/serverless-application-model/latest/developerguide/serverless-sam-telemetry.html Usage: sam package [OPTIONS] Try 'sam package -h' for help. Error: Missing option '--image-repository' or '--image-repositories' ✖ Failed SAM package. Relevant part of my pipeline (monorepo, so I have to cd into the specific directory where `samconfg.toml` and `template.yaml` resides: script: - apt-get update - cd project/function - pipe: atlassian/aws-sam-deploy:1.2.0 variables: AWS_ACCESS_KEY_ID: $AWS_ACCESS_KEY_ID AWS_SECRET_ACCESS_KEY: $AWS_SECRET_ACCESS_KEY AWS_DEFAULT_REGION: $AWS_DEFAULT_REGION S3_BUCKET: 'lambda-function-bucket' STACK_NAME: 'MyLambdaFunction' CAPABILITIES: ['CAPABILITY_IAM'] SAM_TEMPLATE: 'template.yaml' SAM_CONFIG: "samconfig.toml" WAIT: 'true' WAIT_INTERVAL: 60 And here's my samconfig.toml: ``` version = 0.1 [default] [default.deploy] [default.deploy.parameters] stack_name = "MyLambdaFunction" s3_bucket = "lambda-function-bucket" s3_prefix = "s3-prefix/" region = "us-east-1" capabilities = "CAPABILITY_IAM" image_repositories = ["MyLambdaFunction=12345678910.dkr.ecr.us-east-1.amazonaws.com/my-lambda-function"] confirm_changeset = false

Han_yhwrnv · Answer

Solved using AWS SAM CLI directly. @Daniel Coppell

FYI, deployment time is highly dependent the size of the docker image, but I find it faster than my local machine probably from higher bandwidth connection to AWS

- step:
  name: CD
  caches:
    - pip
  script:
    - apt-get update
    - pip install awscli
    - pip install aws-sam-cli
    - aws configure --profile cicd set AWS_ACCESS_KEY_ID ${AWS_ACCESS_KEY_ID}
    - aws configure --profile cicd set AWS_SECRET_ACCESS_KEY ${AWS_SECRET_ACCESS_KEY}
    - sam build --profile cicd
    - sam deploy --profile cicd
  services:
    - docker

Daniel Coppell · Answer

"I suspect you would prefer aws-sam-deploy pipe, where you have samconfig and you would like this samconfig to be support for packaging and deploying."

This would be an excellent feature. I haven't been able to use the aws-sam-deploy pipe thus far because it's useless to me without accepting a samconfig toml file for the deploy command. Currently I have to use custom images with the AWS SAM CLI installed which are rather large and increase my build time.

If there's an issue link or a roadmap available with this feature on it I'd love a link so I can come back and check periodically to see if this feature has been added.

I wonder how anyone using this pipe is avoiding getting errors without the

--no-fail-on-empty-changeset option available? CI/CD pipelines surely hit the "Error: No changes to deploy. Stack stack-name is up to date'." error all the time?

Han_yhwrnv · Answer

@Halyna Berezovska

Correct. I would like to have alignment between both manual `sam deploy` and CD via aws-sam-deploy pipe.

AWS just released and AWS SAM Pipelines that generates support for serveless applications, image or otherwise. https://aws.amazon.com/blogs/compute/introducing-aws-sam-pipelines-automatically-generate-deployment-pipelines-for-serverless-applications/. Unfortunately, it only supports Jenkins, GitHub, and GitLab.

When can I expect Bitbucket to reach feature parity with this?

HCL