Retry ECS create/update when target group isn't yet attached

[tomelliff]

The ECS service can be attached to a load balancer's target group but at that point it may not yet be attached to a load balancer as the dependency graph Terraform generates normally will create the ECS service at the same time as the load balancer listener/listener rule.
This could be avoided by setting the ECS service 'depends_on' to include the load balancer listener/listener rule but this is not usable if the load balancer listener/listener rule is created in another module.

The target group in the acceptance test now gets the name from the load balancer, forcing a dependency between the load balancer and the target group.
Unfortunately it's not possible to use something from the listener to force the dependency as the listener already has a dependency on the target group.

Fixes #3495

[tomelliff]

Relevant acceptance tests:

$ make testacc TEST=./aws TESTARGS="-run=TestAccAWSEcsService_withAlb"
==> Checking that code complies with gofmt requirements...
TF_ACC=1 go test ./aws -v -run=TestAccAWSEcsService_withAlb -timeout 120m
=== RUN   TestAccAWSEcsService_withAlb
--- PASS: TestAccAWSEcsService_withAlb (278.82s)
PASS
ok  	github.com/terraform-providers/terraform-provider-aws/aws	278.837s
$ make testacc TEST=./aws TESTARGS="-run=TestAccAWSEcsService_healthCheckGracePeriodSeconds"
==> Checking that code complies with gofmt requirements...
TF_ACC=1 go test ./aws -v -run=TestAccAWSEcsService_healthCheckGracePeriodSeconds -timeout 120m
=== RUN   TestAccAWSEcsService_healthCheckGracePeriodSeconds
--- PASS: TestAccAWSEcsService_healthCheckGracePeriodSeconds (297.97s)
PASS
ok  	github.com/terraform-providers/terraform-provider-aws/aws	297.977s

[tomelliff]

pinging @bflad for reference following discussion in #3495

I'm not sure it's really great but the only way I can think of forcing a dependency chain that works otherwise since the 1.9.0 change is to put the listener port/protocol in the service name which is hacky as all hell and makes me very sad especially as that service name ends up being propagated all over the place into the AWS console, dashboards and log groups/streams. If the ECS service could be tagged then I would have slightly less objection about forcing an interpolation into a tag value there just because I can kinda sweep that under a rug but unfortunately that's not an option.

Otherwise I'm going to have to either remove the DRYness of my ecs-web-service module setting up a load balancer and calling the ecs-service module (which defaults to a worker based load balancer-less setup) so I repeat the ECS service stuff in both modules or collapse both modules together and litter every LB related resource with count = "${var.load_balanced_service ? 1 : 0}" which is much uglier and more confusing than I currently have.

[tomelliff]

Derp, helps if I actually stage the test changes before committing and pushing. Force pushed that back over with the test changes.

[tomelliff]

@bflad Just rebased it back on to master now, apologies for not spotting your message on Gitter sooner.

I've not re-ran the tests after the rebase but it looks fine from a glance. If you don't beat me to it I'll re-run the tests tomorrow morning when I get a chance.

[mikemac42]

Thanks for working on this fix. Been looking forward to this getting released, but the PR has been open for a while. Any chance this can get merged in soon?

[bflad]

Sorry this PR got lost in the shuffle. 😅 Let's get it in, thanks @tomelliff! 🚀

21 tests passed (all tests)
=== RUN   TestAccAWSEcsService_withEcsClusterName
--- PASS: TestAccAWSEcsService_withEcsClusterName (9.58s)
=== RUN   TestAccAWSEcsService_withFamilyAndRevision
--- PASS: TestAccAWSEcsService_withFamilyAndRevision (10.04s)
=== RUN   TestAccAWSEcsService_basicImport
--- PASS: TestAccAWSEcsService_basicImport (10.04s)
=== RUN   TestAccAWSEcsService_withUnnormalizedPlacementStrategy
--- PASS: TestAccAWSEcsService_withUnnormalizedPlacementStrategy (13.41s)
=== RUN   TestAccAWSEcsService_withReplicaSchedulingStrategy
--- PASS: TestAccAWSEcsService_withReplicaSchedulingStrategy (13.56s)
=== RUN   TestAccAWSEcsService_withPlacementConstraints_emptyExpression
--- PASS: TestAccAWSEcsService_withPlacementConstraints_emptyExpression (13.66s)
=== RUN   TestAccAWSEcsService_withPlacementConstraints
--- PASS: TestAccAWSEcsService_withPlacementConstraints (13.60s)
=== RUN   TestAccAWSEcsService_withDeploymentValues
--- PASS: TestAccAWSEcsService_withDeploymentValues (13.69s)
=== RUN   TestAccAWSEcsServiceDataSource_basic
--- PASS: TestAccAWSEcsServiceDataSource_basic (14.15s)
=== RUN   TestAccAWSEcsService_withServiceRegistries
--- PASS: TestAccAWSEcsService_withServiceRegistries (20.42s)
=== RUN   TestAccAWSEcsService_withDaemonSchedulingStrategy
--- PASS: TestAccAWSEcsService_withDaemonSchedulingStrategy (21.53s)
=== RUN   TestAccAWSEcsService_withServiceRegistries_container
--- PASS: TestAccAWSEcsService_withServiceRegistries_container (25.11s)
=== RUN   TestAccAWSEcsService_withRenamedCluster
--- PASS: TestAccAWSEcsService_withRenamedCluster (37.54s)
=== RUN   TestAccAWSEcsService_withARN
--- PASS: TestAccAWSEcsService_withARN (44.34s)
=== RUN   TestAccAWSEcsService_withPlacementStrategy
--- PASS: TestAccAWSEcsService_withPlacementStrategy (49.30s)
=== RUN   TestAccAWSEcsService_withLaunchTypeEC2AndNetworkConfiguration
--- PASS: TestAccAWSEcsService_withLaunchTypeEC2AndNetworkConfiguration (80.93s)
=== RUN   TestAccAWSEcsService_withIamRole
--- PASS: TestAccAWSEcsService_withIamRole (121.56s)
=== RUN   TestAccAWSEcsService_withLbChanges
--- PASS: TestAccAWSEcsService_withLbChanges (132.80s)
=== RUN   TestAccAWSEcsService_healthCheckGracePeriodSeconds
--- PASS: TestAccAWSEcsService_healthCheckGracePeriodSeconds (198.55s)
=== RUN   TestAccAWSEcsService_withAlb
--- PASS: TestAccAWSEcsService_withAlb (236.36s)
=== RUN   TestAccAWSEcsService_withLaunchTypeFargate
--- PASS: TestAccAWSEcsService_withLaunchTypeFargate (250.58s)

[mikemac42]

Thank you both!

[bflad]

This has been released in version 1.33.0 of the AWS provider. Please see the Terraform documentation on provider versioning or reach out if you need any assistance upgrading.

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thanks!