[AWS] AWS throws InvalidGroup.Duplicate when user submits a lot of parallel launch

[Michaelvll]

To reproduce:
seq 10 | xargs -P 16 -n 1 bash -c 'sky launch -c test-$0 --cpus 2 --use-spot --down -i0 -y &'

Observed this with error

ot error(s) in AWS:[sky.exceptions.NoClusterLaunchedError] Failed to create security group. Error: botocore.exceptions.ClientError: An error occurred (InvalidGroup.Duplicate) when calling the CreateSecurityGroup operation: The security group 'sky-sg-root-xxx' already exists for VPC 'vpc-xxxxxxx'

[cg505]

Does this need to be a new VPC that you haven't used skypilot on before?

[Michaelvll]

I don't think it is a new VPC. The issue happens for only some of the launches. This issue happens for some of the tests with pytest tests/test_smoke.py --aws.