aws-samples / ecs-blue-green-deployment Goto Github PK

What is the command in step 3 to run bin?

3. Run bin/deploy

Thanks for the reference example. The post provides a Blue/green strategy of using different ports and listeners to control the traffic between blue and green. In enterprises like ours, we use products for authentication and they can work only in certain ports like 443. What could be the blue/green strategy in such cases?
We have an apache container running on 443 linked to a springboot container that is running on 8080. Only port 443 can be used since the tool (Cleartrust) used for authentication require that port.

Hi,

I realise this is a fairly old demo now but I was surprised when it didn't work out of the box.

I tried to create the stack on the fargate branch in eu-west-2 so had to update the TemplateURL fields to:

https://${TemplateBucket}.s3.eu-west-2.amazonaws.com/templates/vpc.yaml

This meant the stack creation could progress further, but still failed to create the DeploymentPipeline stack with the following error for the LoadBalancer resource:

S3 error: Access Denied For more information check http://docs.aws.amazon.com/AmazonS3/latest/API/ErrorResponses.html

I then added s3:* to the inline policy within CloudFormationExecutionRole, but I am still seeing the above error.

Can anyone verify that they can get this stack working?

Thank you

David

The code in this repo has been superseded by https://github.com/aws-containers/ecs-workshop-blue-green-deployments with an accompanying tutorial resource at https://ecsworkshop.com/blue_green_deployments

As described in my posts at https://forums.aws.amazon.com/thread.jspa?threadID=274249 and https://stackoverflow.com/questions/48917361/autoscaling-blue-green-ecs-clusters, we built a blue-green cluster on Terraform inspired by this sample, but we have not been able to get autoscaling working. We have been able to get EC2 instances added to the cluster, but the Blue and Green services do not spin up additional tasks to make use of these instances.

It would be great if this sample were expanded to provide working code for autoscaling.

The documentation indicates

If you discover bugs while in Production, you can revert to previous application version by clicking and approving the change again.

Does this imply you need to revert code and re-commit to have the pipeline build again, or is there a way to simply flip the blue/green deployments back to the previous version. The later would be preferable. In this case would a seperate pipeline which skips the code build step and only performs the flip operation be required?

Hi there,

cloning this repo and following the instructions on your workshop page (https://cicd-pipeline-cfn-ecs-fargate-bluegreen.workshop.aws/en/cicd/script.html) results in a stack creation error of the DeploymentPipeline stack with the following error:

"Resource handler returned message: "The runtime parameter of python2.7 is no longer supported for creating or updating AWS Lambda functions. We recommend you use the new runtime (python3.9) while creating or updating functions. (Service: Lambda, Status Code: 400, Request ID: 96df5a45-faa2-4641-9b6f-31b401622c1d, Extended Request ID: null)" (RequestToken: 1ae75b77-6c07-1dc6-0fc9-1483a5033e45, HandlerErrorCode: InvalidRequest)"

It seems that the used runtime for the lambda (python 2.7) is not supported by cloudformation anymore

First of all, thank you for creating this project, which gave me a great starting point!

Instead of using Github as the source of my pipeline, we build the container images in another system and push the images to ECR. Therefore, we need CodePipeline to pick up changes in ECR and run deploy the latest image.