There are three equivalent forms for expressing "everyone" (including
anonymous) in IAM policies:
- "Principals": "*"
- "Principals": {"AWS": "*"}
- "Principals": {"*": "*"}
The more-constrained syntax used by our aws_iam_policy_document data
source means that the user can only express the latter two of these
directly. However, when returning IAM policies from the API AWS likes to
normalize to the first form, causing unresolvable diffs.
This fixes#9335 by handling the "everyone" case as a special case,
serializing it in JSON as the "*" shorthand form.
This change does *not* address the normalization of hand-written policies
containing such elements. A similar change would need to be made in
the external package github.com/jen20/awspolicyequivalence in order to
avoid the issue for hand-written policies.
This will allows us to filter a specific ebs_volume for attachment to an
aws_instance
```
make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSEbsVolumeDataSource_'✹
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2016/11/01 12:39:19 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v
-run=TestAccAWSEbsVolumeDataSource_ -timeout 120m
=== RUN TestAccAWSEbsVolumeDataSource_basic
--- PASS: TestAccAWSEbsVolumeDataSource_basic (28.74s)
=== RUN TestAccAWSEbsVolumeDataSource_multipleFilters
--- PASS: TestAccAWSEbsVolumeDataSource_multipleFilters (28.37s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws57.145s
```
tags were not being set in the read function
TF_ACC=1 go test ./builtin/providers/azurerm -v -run TestAccAzureRMSqlDatabase_basic -timeout 120m
=== RUN TestAccAzureRMSqlDatabase_basic
--- PASS: TestAccAzureRMSqlDatabase_basic (190.60s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/azurerm 190.719s
* Allow `active` state while waiting for the VPC Peering Connection.
This commit adds `active` as one of the valid states in which the VPC Peering
Connection can be when it being created.
Signed-off-by: Krzysztof Wilczynski <krzysztof.wilczynski@linux.com>
* Add more valid states.
Signed-off-by: Krzysztof Wilczynski <krzysztof.wilczynski@linux.com>
Fixes#9628Fixes#9298
When a route53_record alias is updated in the console, AWS prepends
`dualstack.` to the name. This is there incase IPV6 is wanted. It is
exactly the same without it as it is with it
In order to stop perpetual diffs, I introduced a normalizeFunc that will
that tke alias name and strip known issues:
* dualstack
* trailing dot
This normalize fun will continue to grow I'm sure
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSRoute53Record_' ✹
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2016/10/29 00:28:12 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSRoute53Record_ -timeout 120m
=== RUN TestAccAWSRoute53Record_basic
--- PASS: TestAccAWSRoute53Record_basic (124.64s)
=== RUN TestAccAWSRoute53Record_basic_fqdn
--- PASS: TestAccAWSRoute53Record_basic_fqdn (132.07s)
=== RUN TestAccAWSRoute53Record_txtSupport
--- PASS: TestAccAWSRoute53Record_txtSupport (134.07s)
=== RUN TestAccAWSRoute53Record_spfSupport
--- PASS: TestAccAWSRoute53Record_spfSupport (113.36s)
=== RUN TestAccAWSRoute53Record_generatesSuffix
--- PASS: TestAccAWSRoute53Record_generatesSuffix (112.62s)
=== RUN TestAccAWSRoute53Record_wildcard
--- PASS: TestAccAWSRoute53Record_wildcard (162.84s)
=== RUN TestAccAWSRoute53Record_failover
--- PASS: TestAccAWSRoute53Record_failover (126.18s)
=== RUN TestAccAWSRoute53Record_weighted_basic
--- PASS: TestAccAWSRoute53Record_weighted_basic (121.10s)
=== RUN TestAccAWSRoute53Record_alias
--- PASS: TestAccAWSRoute53Record_alias (118.14s)
=== RUN TestAccAWSRoute53Record_s3_alias
--- PASS: TestAccAWSRoute53Record_s3_alias (155.07s)
=== RUN TestAccAWSRoute53Record_weighted_alias
--- PASS: TestAccAWSRoute53Record_weighted_alias (235.41s)
=== RUN TestAccAWSRoute53Record_geolocation_basic
^[[C--- PASS: TestAccAWSRoute53Record_geolocation_basic (125.32s)
=== RUN TestAccAWSRoute53Record_latency_basic
--- PASS: TestAccAWSRoute53Record_latency_basic (122.23s)
=== RUN TestAccAWSRoute53Record_TypeChange
--- PASS: TestAccAWSRoute53Record_TypeChange (231.98s)
=== RUN TestAccAWSRoute53Record_empty
--- PASS: TestAccAWSRoute53Record_empty (116.48s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 2131.526s
```
Before this fix, I was getting the following by recreating the code in
```
~ aws_route53_record.alias
alias.1563903989.evaluate_target_health: "true" => "false"
alias.1563903989.name: "9828-recreation-106795730.us-west-2.elb.amazonaws.com." => ""
alias.1563903989.zone_id: "Z1H1FL5HABSF5" => ""
alias.318754017.evaluate_target_health: "" => "true"
alias.318754017.name: "" => "9828-recreation-106795730.us-west-2.elb.amazonaws.com"
alias.318754017.zone_id: "" => "Z1H1FL5HABSF5"
Plan: 0 to add, 1 to change, 0 to destroy.
```
After this fix:
```
No changes. Infrastructure is up-to-date. This means that Terraform
could not detect any differences between your configuration and
the real physical resources that exist. As a result, Terraform
doesn't need to do anything.
When creating a CloudWatch Metric for an Application Load Balancer Target Group it is
neccessary to use the suffix of the ARN as the reference to the load
balancer TG . This commit exposes that as an attribute on the `aws_alb_target_group`
resource to prevent the need to use regular expression substitution to
make the reference.
Fixes#9410
When importing an azurerm_virtual_network that has no DNSServers,
terraform was throwing a panic as it was trying to dereference that list
of servers to set to state
This commit adds a simple check to make sure there are DNSServers before
dereferencing them
```
make testacc TEST=./builtin/providers/azurerm TESTARGS='-run=TestAccAzureRMVirtualNetwork_' 2 ↵ ✹
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2016/10/31 11:20:36 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/azurerm -v
-run=TestAccAzureRMVirtualNetwork_ -timeout 120m
=== RUN TestAccAzureRMVirtualNetwork_importBasic
--- PASS: TestAccAzureRMVirtualNetwork_importBasic (150.63s)
=== RUN TestAccAzureRMVirtualNetwork_basic
--- PASS: TestAccAzureRMVirtualNetwork_basic (122.90s)
=== RUN TestAccAzureRMVirtualNetwork_disappears
--- PASS: TestAccAzureRMVirtualNetwork_disappears (113.07s)
=== RUN TestAccAzureRMVirtualNetwork_withTags
--- PASS: TestAccAzureRMVirtualNetwork_withTags (139.56s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/azurerm526.168
```
Fixes#8679
The CallerReference attribute we passed to AWS in route53_health_checks
was `time.Now().Format(time.RFC3339Nano)`
When creating multiple resources with the Count meta-parameter, this was
causing issues as follows:
```
* aws_route53_health_check.healthstate.0: HealthCheckAlreadyExists: A different health check has already been created with the specified caller reference.
```
We have now exposed a new attribute called `reference_name` that can be set to pass multiple resources to the request
```
make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSRoute53HealthCheck_' 130 ↵ ✹
==> Cecking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2016/10/31 10:41:07 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSRoute53HealthCheck_ -timeout 120m
=== RUN TestAccAWSRoute53HealthCheck_importBasic
--- PASS: TestAccAWSRoute53HealthCheck_importBasic (17.08s)
=== RUN TestAccAWSRoute53HealthCheck_basic
--- PASS: TestAccAWSRoute53HealthCheck_basic (28.17s)
=== RUN TestAccAWSRoute53HealthCheck_withSearchString
--- PASS: TestAccAWSRoute53HealthCheck_withSearchString (28.07s)
=== RUN TestAccAWSRoute53HealthCheck_withChildHealthChecks
--- PASS: TestAccAWSRoute53HealthCheck_withChildHealthChecks (20.71s)
=== RUN TestAccAWSRoute53HealthCheck_IpConfig
--- PASS: TestAccAWSRoute53HealthCheck_IpConfig (16.09s)
=== RUN TestAccAWSRoute53HealthCheck_CloudWatchAlarmCheck
--- PASS: TestAccAWSRoute53HealthCheck_CloudWatchAlarmCheck (22.42s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 132.568s
```
The update of the test was causing a test failure - it was setting
desired_count to 1 when miz_size was set to 2 - this was causing a
perpetual diff in the test
Was failing due to using IAM user `test-name` as it was being used in
more than 1 place - this has been replaced by a random user and random
policy names now
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSUserPolicyAttachment_basic' 2 ↵ ✹
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2016/10/31 08:39:08 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSUserPolicyAttachment_basic -timeout 120m
=== RUN TestAccAWSUserPolicyAttachment_basic
--- PASS: TestAccAWSUserPolicyAttachment_basic (32.04s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 32.053s
```
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.
* provider/google Document MySQL versions for second generation instances
Google Cloud SQL has first-gen and second-gen instances with different
supported versions of MySQL.
* provider/google Increase SQL Admin operation timeout to 10 minutes
Creating SQL instances for MySQL 5.7 can take over 7 minutes,
so the timeout needs to be increased to allow the
google_sql_database_instance resource to successfully create.
This commit adds an ability to modify the `AutoMinorVersionUpgrade` property of the
Replication Group (which is enabled by default) accordingly.
Signed-off-by: Krzysztof Wilczynski <krzysztof.wilczynski@linux.com>
* Adding private gateway and static route resource to cloudstack provider
Testing the private gateway and static route resource requires a ROOT
account in Cloudstack
* changes requested by reviewer
Fixes#9654
Before the fix, I created an ASG with a schedule on it. Went to the AWS
console and deleted the schedule. A terraform plan looked as follows:
```
% terraform plan
See https://www.terraform.io/docs/internals/internal-plugins.html
Refreshing Terraform state in-memory prior to plan...
The refreshed state will be used to calculate this plan, but
will not be persisted to local or remote state storage.
aws_launch_configuration.foobar: Refreshing state... (ID:
terraform-test-foobar5)
aws_autoscaling_group.foobar: Refreshing state... (ID:
terraform-test-foobar5)
aws_autoscaling_schedule.foobar: Refreshing state... (ID: foobar)
Error refreshing state: 1 error(s) occurred:
* aws_autoscaling_schedule.foobar: Unable to find Autoscaling
* Scheduled Action: []*autoscaling.ScheduledUpdateGroupAction(nil)
```
After the fix:
```
terraform plan 1 ↵
Refreshing Terraform state in-memory prior to plan...
The refreshed state will be used to calculate this plan, but
will not be persisted to local or remote state storage.
aws_launch_configuration.foobar: Refreshing state... (ID: terraform-test-foobar5)
aws_autoscaling_group.foobar: Refreshing state... (ID: terraform-test-foobar5)
aws_autoscaling_schedule.foobar: Refreshing state... (ID: foobar)
The Terraform execution plan has been generated and is shown below.
Resources are shown in alphabetical order for quick scanning. Green resources
will be created (or destroyed and then created if an existing resource
exists), yellow resources are being changed in-place, and red resources
will be destroyed. Cyan entries are data sources to be read.
Note: You didn't specify an "-out" parameter to save this plan, so when
"apply" is called, Terraform can't guarantee this is what will execute.
+ aws_autoscaling_schedule.foobar
arn: "<computed>"
autoscaling_group_name: "terraform-test-foobar5"
desired_capacity: "0"
end_time: "2018-01-16T13:00:00Z"
max_size: "0"
min_size: "0"
recurrence: "<computed>"
scheduled_action_name: "foobar"
start_time: "2018-01-16T07:00:00Z"
Plan: 1 to add, 0 to change, 0 to destroy.
```
Tests run as expected:
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSAutoscalingSchedule_' 2 ↵ ✹
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2016/10/27 17:45:19 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSAutoscalingSchedule_ -timeout 120m
=== RUN TestAccAWSAutoscalingSchedule_basic
--- PASS: TestAccAWSAutoscalingSchedule_basic (140.94s)
=== RUN TestAccAWSAutoscalingSchedule_disappears
--- PASS: TestAccAWSAutoscalingSchedule_disappears (179.17s)
=== RUN TestAccAWSAutoscalingSchedule_recurrence
--- PASS: TestAccAWSAutoscalingSchedule_recurrence (186.72s)
=== RUN TestAccAWSAutoscalingSchedule_zeroValues
--- PASS: TestAccAWSAutoscalingSchedule_zeroValues (167.73s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 674.530s
```
* provider/aws: data source for AWS Security Group
* provider/aws: add documentation for data source for AWS Security Group
* provider/aws: data source for AWS Security Group (improve if condition and syntax)
* fix fmt
* provider/scaleway: fix scaleway_volume_attachment with count > 1
since scaleway requires servers to be powered off to attach volumes to, we need
to make sure that we don't power down a server twice, or power up a server while
it's supposed to be modified.
sadly terraform doesn't seem to sport serialization primitives for usecases like
this, but putting the code in question behind a `sync.Mutex` does the trick, too
fixes#9417
* provider/scaleway: use mutexkv to lock per-resource
following @dcharbonnier suggestion. thanks!
* provider/scaleway: cleanup waitForServerState signature
* provider/scaleway: store serverID in var
* provider/scaleway: correct imports
* provider/scaleway: increase timeouts
* Improve messaging when storage account isn't found.
* Add client for finding resources when you don't know it's resource group.
* Add function to find Storage Account resource group name.
* Use the storage account resource group, not the virtual machine's resource group when deleting VHDs.
* Add description of storage account ID for clarity.
* Improve VHD deletion test when storage account is in different resource group.
* Use common function for ID parsing of storage account.
* Add AWS Prefix List data source.
AWS Prefix List data source acceptance test.
AWS Prefix List data source documentation.
* Improve error message when PL not matched.
* provider/scaleway speedup server deletion
using `terminate` instead of `poweroff` leads to a faster shutdown
fixes#9430
* provider/scaleway: extract server shutdown code
This test was previously passing but the limit appears to have been reduced
since.
TF_ACC=1 go test ./builtin/providers/azurerm -v -run TestAccAzureRMServiceBusTopic_enableParti -timeout 120m
=== RUN TestAccAzureRMServiceBusTopic_enablePartitioning
--- PASS: TestAccAzureRMServiceBusTopic_enablePartitioning (377.14s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/azurerm 377.235s
* Updated docker container documentation
Feedback from ticket #9350 indicated that documentation was out of date
renamed `hosts_entry` to `host`
added correct type information to *Extra Hosts* section.
Refs: 9350
* Fixes for docker_container host object
Feedback from ticket #9350 updated codebase so it reflects the requirements from docker in regards to `host` which is `Required` and not optional.
It now accurately reflects the docker requirements and the terraform documentation.
Test results
> Bear in mind the failure it is because my laptop doesnt support memory swap. So this test will always fail.
Changing the Schema from `optional` to `required` made no difference to the tests.
make testacc TEST=./builtin/providers/docker/
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2016/10/14 15:04:40 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/docker/ -v -timeout 120m
=== RUN TestAccDockerRegistryImage_basic
--- PASS: TestAccDockerRegistryImage_basic (4.57s)
=== RUN TestAccDockerRegistryImage_private
--- PASS: TestAccDockerRegistryImage_private (6.22s)
=== RUN TestProvider
--- PASS: TestProvider (0.00s)
=== RUN TestProvider_impl
--- PASS: TestProvider_impl (0.00s)
=== RUN TestAccDockerContainer_basic
--- PASS: TestAccDockerContainer_basic (7.16s)
=== RUN TestAccDockerContainer_volume
--- PASS: TestAccDockerContainer_volume (7.37s)
=== RUN TestAccDockerContainer_customized
--- FAIL: TestAccDockerContainer_customized (18.99s)
testing.go:265: Step 0 error: Check failed: Check 2/2 error: Container has wrong memory swap setting: -1
Please check that you machine supports memory swap (you can do that by running 'docker info' command).
=== RUN TestAccDockerImage_basic
--- PASS: TestAccDockerImage_basic (2.58s)
=== RUN TestAccDockerImage_private
--- PASS: TestAccDockerImage_private (2.70s)
=== RUN TestAccDockerImage_destroy
--- PASS: TestAccDockerImage_destroy (30.00s)
=== RUN TestAccDockerImage_data
--- PASS: TestAccDockerImage_data (5.93s)
=== RUN TestAccDockerNetwork_basic
--- PASS: TestAccDockerNetwork_basic (0.24s)
=== RUN TestAccDockerVolume_basic
--- PASS: TestAccDockerVolume_basic (0.05s)
FAIL
exit status 1
FAIL github.com/hashicorp/terraform/builtin/providers/docker 85.816s
Makefile:47: recipe for target 'testacc' failed
make: *** [testacc] Error 1
Refs: 9350
* govendor: update go-cloudstack dependency
* Separate security groups and rules
This commit separates the creation and management of security groups and security group rules.
It extends the `icmp` options so you can supply `icmp_type` and `icmp_code` to enbale more specific configs.
And it adds lifecycle management of security group rules, so that security groups do not have to be recreated when rules are added or removed.
This is particulary helpful since the `cloudstack_instance` cannot update a security group without having to recreate the instance.
In CloudStack >= 4.9.0 it is possible to update security groups of existing instances, but as that is just added to the latest version it seems a bit too soon to start using this (causing backwards incompatibility issues for people or service providers running older versions).
* Add and update documentation
* Add acceptance tests
Christoph Blecker