AWS Beanstalk环境重建后CloudFormation更新失败求助

Let’s work through each of your problems step by step to get your Rails app back on track—this messy situation is totally recoverable:

1. Fix CloudFormation Stack and Environment Drift First #

When you clicked Rebuild Environment in the console, you created a mismatch between your CloudFormation stack's desired configuration and the actual environment state (drift). This is why your update-stack commands are failing—CloudFormation can’t reconcile the conflicting settings.

Resolve the drift first:

• Head to the CloudFormation console, find your stack, and check the Drift Detection tab to see exactly which resources (like SQS, load balancer) are out of sync.
• Use the CLI to get a detailed drift report:

  aws cloudformation detect-stack-drift --stack-name YOUR_STACK_NAME
  

• Then force the stack to align with your template using:

  aws cloudformation update-stack --stack-name YOUR_STACK_NAME --template-body file://your-template.yml --parameters YOUR_PARAMETERS --override-existing-template
  

• If drift is severe, consider spinning up a temporary stack with your updated template, verifying it works, then migrating your app to it (a last-resort option).

2. Resolve Bundler and Ruby Version Conflicts #

Your deployment failures come from two version mismatches:

a. Bundler Version Gap #

The Beanstalk environment uses Bundler 1.16.0, but your Gemfile.lock requires 1.17.3. Fix this by adding a Beanstalk config file:
Create .ebextensions/bundler.config with:

commands:
  update_bundler:
    command: "gem install bundler -v 1.17.3"

Commit this file to your codebase so it deploys with your next update.

b. Ruby Version Mismatch #

Your Gemfile specifies Ruby 2.6.5, but your environment is set to 2.6.6. Pick one of these fixes:

• Update your Gemfile to match the environment:

  ruby '2.6.6'
  

  Run bundle update locally to regenerate Gemfile.lock, commit both files, and redeploy.
• Or, adjust your CloudFormation template to use a Beanstalk platform version tied to Ruby 2.6.5 (though this defeats your original upgrade goal).

Double-check your CloudFormation template uses the correct platform ARN for Ruby 2.6.6 on Amazon Linux 2—something like arn:aws:elasticbeanstalk:us-east-1::platform/Ruby 2.6 running on 64bit Amazon Linux 2/2.11.7 (adjust the region as needed).

3. Fix ElasticSearch Environment Variable #

The old ELASTICSEARCH_HOST persists because the rebuild didn’t pull the new value from your CloudFormation template. Fix this:

• Verify your template includes the correct variable in the Beanstalk environment settings:

  Resources:
    YourBeanstalkEnv:
      Type: AWS::ElasticBeanstalk::Environment
      Properties:
        # Other properties...
        OptionSettings:
          - Namespace: aws:elasticbeanstalk:application:environment
            OptionName: ELASTICSEARCH_HOST
            Value: YOUR_NEW_ES_INSTANCE_ADDRESS
  

• Re-run the update-stack command with the override flag (from step 1) to force the environment variable update.
• As a temporary fix, manually set the variable via the Beanstalk console (Configuration > Software > Environment Properties), then re-run the CF update to make it permanent.

4. Force CloudFormation to Apply Changes #

If all else fails:

1. Delete or recreate drifted resources to match your template (start with non-critical ones like test SQS queues).
2. Run the update command with a flag to bypass empty changeset errors:

   aws cloudformation update-stack --stack-name YOUR_STACK_NAME --template-body file://your-template.yml --parameters YOUR_PARAMETERS --no-fail-on-empty-changeset
   

3. If the deployment hook still fails, SSH into a Beanstalk instance, manually run gem install bundler -v 1.17.3 and bundle install to resolve the Bundler issue temporarily, then retrigger the CF update.

内容的提问来源于stack exchange，提问作者Puneet Pandey