Unable to delete EC2 instance for ECS(EC2 Container Service)

4
votes

As Amazon ECS being launched recently, I tried to start a sample with it. It looks like ECS will create a EC2 instance as a "cluster" for ECS, whose name is like "ECS Instance - EC2ContainerService-default-e40c3e79-1d27-4512-a790-c6782b45b15a". However, I cannot find a way to delete this EC2 instance when I don't want to serve my container any more.

When I terminate this instance through web console, another instance with the same name(but different instance id) will show up later. It's quite annoying and has already charged my some money. Here's the screenshot of EC2 web console:

screenshot of EC2 web console

where there's always a running instance.

Here's my screenshot of ECS web console, where I have deleted all the cluster.

enter image description here enter image description here

At this point, I can't find any documentation or blog on this issue. :(

3
amazon-web-servicesamazon-ec2amazon-ecs
It sounds like you have an auto scaling group respawning the instance. In the EC2 console navigation pane, under Auto Scaling, click Auto Scaling Group and see what you find there. – Michael - sqlbot
@Michael-sqlbot Thanks! That it is. The instance is registered under Auto Scale Group. When I deleted that group, the instance is gone :) – eval

3 Answers

13
votes

Note

If your instance is maintained by an Auto Scaling group or AWS CloudFormation stack, terminate the instance by updating the Auto Scaling group or AWS CloudFormation stack; otherwise, the Auto Scaling group will recreate the instance after you terminate it.

http://docs.aws.amazon.com/AmazonECS/latest/developerguide/deregister_container_instance.html

Auto Scaling groups can be found in the the EC2 console navigation pane: under Auto Scaling, click Auto Scaling Group.

0
votes

If you have created the instance using Elastic Beanstalk. Then you should follow the below link to terminate this permanently.

https://docs.aws.amazon.com/elasticbeanstalk/latest/dg/using-features.terminating.html

0
votes

As pointed out by Michael - sqlbot, the AWS Auto Scaling mechanisms are likely re-creating the instances to meet the initial configuration demands (e.g. ECS clusters).

One of the options to prevent this from happening is to update the auto-scaling group to 0 instances by issuing,

aws autoscaling update-auto-scaling-group --auto-scaling-group-name modify-this-with-auto-scaling-group-name --min-size 0 --max-size

and to further, if the auto-scaling group is not required, delete the auto-scaling group, by issuing,

aws autoscaling delete-auto-scaling-group --auto-scaling-group-name modify-this-with-auto-scaling-group-name

Hope it helps!