tree 6ad3c4810bf2e7160cb8ccf6563e60cf9510d27b
parent a12ad515f8822d80e5a39ab634203efc1c629a21
author Antonio Barone <syntonyze@gmail.com> 1637788157 +0100
committer Antonio Barone <syntonyze@gmail.com> 1638265935 +0100

Fix error when deleting stacks using capacityProvider

Deleting stacks using ECS clusters having capacityProviders (i.e.
dual-primary and primary-replica recipes), fails with:

```
The Cluster cannot be deleted while Container Instances are active
or draining.
```

This is an issue that manifests itself as well via terraform [1] or CDK
[2].

Explicitly deleting the Autoscaling Groups _before_ the ECS cluster
deletion fixes the problem, since it ensures that no instances are
active or draining, as the error suggests.

This is safe to do, because prior to deleting the Autoscaling Groups,
every ECS service has already been destroyed, thus no instance is
actually running.

[1] https://github.com/hashicorp/terraform-provider-aws/issues/4852
[2] https://github.com/aws/aws-cdk/issues/14732
Bug: Issue 14698
Change-Id: I216307ef88bd7b7317706d2dc0a6a6e6fb367bd4

Change-Id: I27ece0f6971b157a474d91d7f3d9243dcff596e6