proper logical order when deploying multiple resources using cloudformation

dataking

Hey community,

 

I am using bitbucket-pipelines.yml to deploy a cloudformation using the atlassian/aws-cloudformation-deploy:0.12.0 pipe and my lambda function to s3 using the atlassian/aws-s3-deploy:1.1.0 pipe.

I have all of my resources being created in the cloudformation template. The resources include an S3 bucket, and lambda function which is referencing the zip file code in the S3 bucket.

I keep running into issues because I cant ( or at least I dont think I can) put the S3 pipe before the cloudformation template because the bucket would not be created, and if I put the S3 pipe after the cloudformation pipe it will fail because the lambda function cloudformation step tries to reference the zipped code in the S3 bucket, which the S3 pipe is supposed to upload.

Any recommendations? I could potentially have separate cloudformation templates for each resource and create a new step in the bitbucket-pipelines.yml, but it seems a bit more challenging to manage all the files, but maybe not?

Any suggestions?

Igor Stoyanov

@dataking  hi. Thanks for your question. Try this to split logic:

step:
  script:
    - pipe: atlassin/cloudformation-deploy:0.12.0
      <create s3 bucket from template>
    - pipe: atlassian/aws-s3-deploy:1.1.0
      <deploy to s3>
    - pipe: atlassin/cloudformation-deploy:0.12.0
      <create lambda>

Regards, Igor