* provider/aws: Increase AMI destroy timeout
Acceptance tests were timing out on AMI destroy, should alleviate the problem.
* Further increase timeout, cleanup test
* use function instead of printf
Fixes the two failing aws_ami acceptance tests, by polling until the AMI is deleted before exiting.
```
$ make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSAMI_basic'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/03/14 16:03:26 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSAMI_basic -timeout 120m
=== RUN TestAccAWSAMI_basic
--- PASS: TestAccAWSAMI_basic (43.98s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 43.989s
```
```
$ make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSAMI_snapshotSize'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/03/14 16:07:29 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSAMI_snapshotSize -timeout 120m
=== RUN TestAccAWSAMI_snapshotSize
--- PASS: TestAccAWSAMI_snapshotSize (28.64s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 28.654s
```
Previously the AMI creation accepted a static value for the AMI's block device's volume size.
This change allows the user to omit the `volume_size` attribute, in order to mimic the AWS API behavior, which will use the EBS Volume's size.
Also fixes a potential panic case when setting `iops` on the AMI's block device.
The `aws_ami` resource previously didn't have any acceptance tests, adds two acceptance tests and a full testing suite for the `aws_ami` resource, so further tests can be written, as well as expansion upon the other `aws_ami_*` acceptance tests
```
$ make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSAMI_'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/02/09 20:18:22 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSAMI_ -timeout 120m
=== RUN TestAccAWSAMI_basic
--- PASS: TestAccAWSAMI_basic (44.21s)
=== RUN TestAccAWSAMI_snapshotSize
--- PASS: TestAccAWSAMI_snapshotSize (45.08s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 89.320s
```
Previously this resource (and, by extension, the aws_ami_copy and
aws_ami_from_instance resources that share much of its implementation)
was handling correctly the case where an AMI had been recently
deregistered, and was thus still returned from the API, but not correctly
dealing with the situation where the AMI has been removed altogether.
Now we additionally handle the NotFound error returned by the API when
we request a non-existent AMI, and remove the AMI from the state in the
same way we do for deregistered AMIs.
This commit resolves the issue where lack of snapshot ID in the device mapping
section of the API response to DescribeImagesResponse would cause Terraform to
crash due to a nil pointer dereference. Usually, the snapshot ID is included,
but in some unique cases (e.g. ECS-enabled AMI from Amazon available on the
Market Place) a volume that is attached might not have it.
The API documentation does not clearly define whether the snapshot ID either
should be or must be included for any volume in the response.
Signed-off-by: Krzysztof Wilczynski <krzysztof.wilczynski@linux.com>
AWS provides three different ways to create AMIs that each have different
inputs, but once they are complete the same management operations apply.
Thus these three resources each have a different "Create" implementation
but then share the same "Read", "Update" and "Delete" implementations.
Radek Simko