Utilize the ChangeResourceRecordSets to change the type of a record by
deleting and recreating with a new type.
As change batches are considered transactional changes, Amazon Route 53
either makes all or none of the changes in the batch request ensuring the
update will never be partially applied.
Documentation for the `aws_route_table` data source mentions that it supports a route table `id` as an argument, however it was missing from the actual provider code.
Adds in the missing provider code, adds a test, and updates the documentation to use `rtb_id` as the argument, instead of the more ambiguous `id`.
A VPC's main route table has an implicit subnet association, not an explicit subnet association. This caused a Terraform panic when using the `data_source_aws_route_table` resource to query the main route table for a VPC.
This fixes the Terraform panic, and allows the data lookup to complete successfully. Also added an acceptance test to verify the bugfix.
Fixes: #11134
The second step of the `TestAccDigitalOceanDroplet_ResizeWithOutDisk` acceptance test was regularly failing. Upon investigation it was found that the second step's Terraform configuration had omitted the `user_data` parameter, thus forcing Terraform to create a new droplet resource instead of updating the current resource.
```
-/+ digitalocean_droplet.foobar
disk: "20" => "<computed>"
image: "centos-7-x64" => "centos-7-x64"
ipv4_address: "138.197.0.55" => "<computed>"
ipv4_address_private: "" => "<computed>"
ipv6_address: "" => "<computed>"
ipv6_address_private: "" => "<computed>"
locked: "false" => "<computed>"
name: "foo" => "foo"
region: "nyc3" => "nyc3"
resize_disk: "true" => "false"
size: "512mb" => "1gb"
ssh_keys.#: "1" => "1"
ssh_keys.0: "5770472" => "5770472"
status: "active" => "<computed>"
user_data: "foobar" => "" (forces new resource)
vcpus: "1" => "<computed>"
```
This fixes the acceptance test by adding the missing `user_data` parameter.
request_path had Computed enabled which prevented updating it to an empty value
TF_ACC=1 go test ./builtin/providers/azurerm -v -run TestAccAzureRMLoadBalancerProbe -timeout 120m
=== RUN TestAccAzureRMLoadBalancerProbe_basic
--- PASS: TestAccAzureRMLoadBalancerProbe_basic (119.63s)
=== RUN TestAccAzureRMLoadBalancerProbe_removal
--- PASS: TestAccAzureRMLoadBalancerProbe_removal (122.50s)
=== RUN TestAccAzureRMLoadBalancerProbe_update
--- PASS: TestAccAzureRMLoadBalancerProbe_update (129.98s)
=== RUN TestAccAzureRMLoadBalancerProbe_duplicate
--- PASS: TestAccAzureRMLoadBalancerProbe_duplicate (115.22s)
=== RUN TestAccAzureRMLoadBalancerProbe_updateProtocol
--- PASS: TestAccAzureRMLoadBalancerProbe_updateProtocol (127.25s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/azurerm 614.657s
As brought up in #10174, our update_strategy property for instance group
managers in GCP would always be set to "RESTART" on read, even if the
user asked for them to be "NONE" in the config.
This adds a test to ensure that the user wishes were respected, which
fails until we check for update_strategy in the ResourceData before we
update it within the Read function. Because the update_strategy property
doesn't map to anything in the API, we never need to read it from
anywhere but the config, which means the ResourceData should be
considered authoritative by the time we get to the Read function.
The fix for this was provided by @JDiPierro in #10198 originally, but
was missing tests, so it got squashed into this.
* provider/aws: Save disabled ELB accesslogs to state
Save any explicitly disabled access_log to state. Do not save disabled
access_logs if they are not in the configuration.
* test that fails on master
Adds validation for the `type` parameter of an `aws_route53_record` resource.
This will allow Terraform to catch any user errors of a `type` parameter during a `terraform plan` instead of during a `terraform apply`.
Fixes: #11114
* provider/aws: New Resource - aws_codedeploy_deployment_config
* provider/aws: Adding acceptance tests for new
aws_codedeploy_deployment_config resource
* provider/aws: Documentation for the aws_codedeploy_deployment_config resource
* Update codedeploy_deployment_config.html.markdown
subnet_group
Fixes#11024
A change was introduced in 0.8.2 that allows db_instances to change
their db_subnet_group. Unfortunately, this caused an issue for
db_instances that were being restored from snapshot. The restore from
snapshot part of create calls the Update func whereas a normal create
calls the Read func
When calling the Update func, the db_instance was trying to go through a
db_subnet_group_name change and was failing for the following reason:
```
InvalidVPCNetworkStateFault: You cannot move DB instance _rds_instance_name_ to subnet group _subnet_group_name_. The specified DB subnet group and DB instance are in the same VPC. Choose a DB subnet group in different VPC than the specified DB instance and try again.
```
Adds region specific S3 bucket name validation. Currently all regions except for us-east-1 force a dns-compliant naming convention. Thus we cannot utilize the standard `SchemaValidateFunc` interface to validate an S3 bucket name.
This change creates a helper function outside of the schema validation interface so we can validate S3 bucket names for both naming conventions. At a later date, when the us-east-1 region is updated to conform to a dns-compliant naming scheme, we can refactor the `validateS3BucketName` function to fit the `SchemaValidateFunc` interface.
aws_api_gateway_integration_response
This continues the work carried out in #10696
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSAPIGatewayIntegrationResponse_'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/01/03 14:18:46 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v
-run=TestAccAWSAPIGatewayIntegrationResponse_ -timeout 120m
=== RUN TestAccAWSAPIGatewayIntegrationResponse_basic
--- PASS: TestAccAWSAPIGatewayIntegrationResponse_basic (57.33s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws57.352s
```
* Importing the OpsGenie SDK
* Adding the goreq dependency
* Initial commit of the OpsGenie / User provider
* Refactoring to return a single client
* Adding an import test / fixing a copy/paste error
* Adding support for OpsGenie docs
* Scaffolding the user documentation for OpsGenie
* Adding a TODO
* Adding the User data source
* Documentation for OpsGenie
* Adding OpsGenie to the internal plugin list
* Adding support for Teams
* Documentation for OpsGenie Team's
* Validation for Teams
* Removing Description for now
* Optional fields for a User: Locale/Timezone
* Removing an implemented TODO
* Running makefmt
* Downloading about half the internet
Someone witty might simply sign this commit with "npm install"
* Adding validation to the user object
* Fixing the docs
* Adding a test creating multple users
* Prompting for the API Key if it's not specified
* Added a test for multiple users / requested changes
* Fixing the linting
Fixes:#10902
AWS introduced a change to the Mount Target DNS Name to remove the
availability_zone from it -
https://aws.amazon.com/about-aws/whats-new/2016/12/simplified-mounting-of-amazon-efs-file-systems/
This was because there used to be a limit of 1 mount target per AZ -
this has been raised.
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSEFSMountTarget_'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/01/04 10:45:35 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSEFSMountTarget_ -timeout 120m
=== RUN TestAccAWSEFSMountTarget_importBasic
--- PASS: TestAccAWSEFSMountTarget_importBasic (236.19s)
=== RUN TestAccAWSEFSMountTarget_basic
--- PASS: TestAccAWSEFSMountTarget_basic (445.52s)
=== RUN TestAccAWSEFSMountTarget_disappears
--- PASS: TestAccAWSEFSMountTarget_disappears (228.31s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 910.044s
```
* vendor: update jen20/riviera to pull in endpoints change
* provider/auzrerm: support non public clouds
Ran tests below with ARM_ENVIRONMENT=german and changes the location to Germany
Central in the test config. The virtual network tests cover both Riviera (resource
groups) and the official SDK client.
TF_ACC=1 go test ./builtin/providers/azurerm -v -run TestAccAzureRMVirtualNetwork_ -timeout 120m
=== RUN TestAccAzureRMVirtualNetwork_importBasic
--- PASS: TestAccAzureRMVirtualNetwork_importBasic (81.84s)
=== RUN TestAccAzureRMVirtualNetwork_basic
--- PASS: TestAccAzureRMVirtualNetwork_basic (78.14s)
=== RUN TestAccAzureRMVirtualNetwork_disappears
--- PASS: TestAccAzureRMVirtualNetwork_disappears (78.45s)
=== RUN TestAccAzureRMVirtualNetwork_withTags
--- PASS: TestAccAzureRMVirtualNetwork_withTags (81.78s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/azurerm 320.310s
bool not string
Fixes: #11010
Adds a test to show cover the use-case that the OP suggested caused the
panic
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSDataSourceIAMServerCertificate_'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/01/03 22:39:21 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSDataSourceIAMServerCertificate_ -timeout 120m
=== RUN TestAccAWSDataSourceIAMServerCertificate_basic
--- PASS: TestAccAWSDataSourceIAMServerCertificate_basic (19.48s)
=== RUN TestAccAWSDataSourceIAMServerCertificate_matchNamePrefix
--- PASS: TestAccAWSDataSourceIAMServerCertificate_matchNamePrefix (1.95s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 21.454s
```
* provider:aws: Add support for updating aws_emr_cluster parameters
Fixes#10962
We now support the update of `visibile_to_add_users` and
`termination_protected` parameters
```
make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSEMRCluster_'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/01/03 18:45:20 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSEMRCluster_ -timeout 120m
=== RUN TestAccAWSEMRCluster_basic
--- PASS: TestAccAWSEMRCluster_basic (551.98s)
=== RUN TestAccAWSEMRCluster_terminationProtected
--- PASS: TestAccAWSEMRCluster_terminationProtected (545.52s)
=== RUN TestAccAWSEMRCluster_visibleToAllUsers
--- PASS: TestAccAWSEMRCluster_visibleToAllUsers (552.09s)
=== RUN TestAccAWSEMRCluster_tags
--- PASS: TestAccAWSEMRCluster_tags (598.91s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 2248.537s
```
* Update resource_aws_emr_cluster_test.go
Fixes: #10958
This wasn't implemented the first time around for some reason
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSEMRCluster_' 2 ↵ ✹ ✭
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/01/03 15:40:09 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSEMRCluster_ -timeout 120m
=== RUN TestAccAWSEMRCluster_basic
--- PASS: TestAccAWSEMRCluster_basic (529.36s)
=== RUN TestAccAWSEMRCluster_tags
--- PASS: TestAccAWSEMRCluster_tags (556.81s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 1086.197s
```
There is 1 breaking change. The OpsWorks Update API no longer supports
`tenancy` as an update. So that has to be marked as ForceNew: true
```
% make testacc TEST=./builtin/providers/aws
% TESTARGS='-run=TestAccAWSOpsworksInstance'
% 2 ↵ ✹ ✭
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
2017/01/02 17:11:29 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v
-run=TestAccAWSOpsworksInstance -timeout 120m
=== RUN TestAccAWSOpsworksInstance
--- PASS: TestAccAWSOpsworksInstance (99.89s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws99.911s
```
* Adding a missing property to the Consumer Groups doc
* Support for Event Hub Authorization Rules
* Documentation for Authorization Rules
* Missed a comment
* Fixing the `no authorisation rule` state
* Making the documentation around the Permissions more explicit / updating the import url
* Fixing up the tests
* Clearing up the docs
* Fixing the linting
* Moving the validation inside of the expand
* Fixing the indentation
* Adding the Container Registry SDK
* Implementing the container registry
* Enabling the provider / registering the Resource Provider
* Acceptance Tests
* Documentation for Container Registry
* Fixing the name validation
* Validation for the Container Registry Name
* Added Import support for Containr Registry
* Storage Account is no longer optional
* Updating the docs
* Forcing a re-run in Travis
The provider now consults the Providers API to detect which providers are already
registered and uses this call to test the credentials
A new option `skip_provider_registration` / `ARM_SKIP_PROVIDER_REGISTRATION` is
now available to opt out of provider registration entirely
Ensure that the `key_name` attribute is available to `aws_key_pair`
resource dependents, even when the attribute is not specifically
set (i.e., when `key_name_prefix` or automatic naming is performed).
Fixes#10983.
* Add 'aws_vpc_peering_connection' data source.
* Changes after code review.
* Add 'accepter' and 'requester' blocks to aws_vpc_peering_connection data source output attributes.
Computed
Fixes: #10887
In the Terraform documentation, we say the following:
```
metric_aggregation_type - (Optional) The aggregation type for the policy's metrics. Valid values are "Minimum", "Maximum", and "Average". Without a value, AWS will treat the aggregation type as "Average".
```
Therefore, because the parameter wasn't set as computed, we got a diff as follows:
```
~ aws_autoscaling_policy.scale
metric_aggregation_type: "Average" => ""
```
With the change in place, we get the correct behaviour
```
% terraform plan
[WARN] /Users/stacko/Code/go/bin/terraform-provider-aws overrides an internal plugin for aws-provider.
If you did not expect to see this message you will need to remove the old plugin.
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.launch: Refreshing state... (ID: config)
aws_autoscaling_group.asg: Refreshing state... (ID: asg)
aws_autoscaling_policy.scale: Refreshing state... (ID: scale)
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.
```
* Documentation
* Scaffolding Consumer Groups
* Linting
* Adding a missing </li>
* User MetaData needs to explicitly be optional
* Typo
* Fixing the test syntax
* Removing the eventHubPath field since it appears to be deprecated
* Added a Complete import test
* WIP
* Updating to use SDK 7.0.1
* Calling the correct method
* Removing eventhubPath and updating the docs
* Fixing a typo
This commit modifies the GNUmakefile used for starting PostgreSQL for
testing purposes to prefer the version installed via Homebrew, followed
by the version installed via MacPorts. POSTGRES, PSQL and INITDB may be
overriden as Make variables if neither of these package management
systems is in use.
applied.
This is only a problem when the ROLE is removed before the revokation of
the schema's policies on the doomed PostgreSQL ROLE. Comment out the
tests for now to get unit tests to pass. Using hard-coded values that
exist outside of the Terraform unit test work.
removing a PostgreSQL role.
Add manual overrides if this isn't the desired behavior, but it should
universally be the desired outcome except when a ROLE name is reused
across multiple databases in the same PostgreSQL cluster, in which case
the `skip_drop_role` is necessary for all but the last PostgreSQL
provider.
We just check these for syntax rather than exact value because the
underlying archive library is free to evolve its exact encoding over time
as long as the result is semantically equivalent, and we don't want these
tests to break when we build on different versions of Go.
A new create_timeout attribute was added that had some backwards
incompatibilities, and as per discussion in #10823, it was determined we
could make upgrading to 0.8.x easier by fixing them, without really
losing any functionality.
Because create_timeout is not something stored or transmitted to the
API, it's not something we need a ForceNew on. Also, because an update
wouldn't result in an API call, we can add a state migration to avoid a
false positive diff that requires people to plan and apply but doesn't
actually make an API call.
The subsets for backend address pools and inbound nat rules weren't being hashed
properly as part of the ip_configuration hash which caused multiple
ip_configurations to be expanded and sent to the API with conflicting names
TF_ACC=1 go test ./builtin/providers/azurerm -v -run TestAccAzureRMNetworkInterface -timeout 120m
=== RUN TestAccAzureRMNetworkInterface_basic
--- PASS: TestAccAzureRMNetworkInterface_basic (160.24s)
=== RUN TestAccAzureRMNetworkInterface_disappears
--- PASS: TestAccAzureRMNetworkInterface_disappears (157.00s)
=== RUN TestAccAzureRMNetworkInterface_enableIPForwarding
--- PASS: TestAccAzureRMNetworkInterface_enableIPForwarding (156.86s)
=== RUN TestAccAzureRMNetworkInterface_multipleLoadBalancers
--- PASS: TestAccAzureRMNetworkInterface_multipleLoadBalancers (185.87s)
=== RUN TestAccAzureRMNetworkInterface_withTags
--- PASS: TestAccAzureRMNetworkInterface_withTags (1212.92s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/azurerm 1872.960s
This commit adds some basic tests for block device functionality. It also
expands the existing block device documentation as well references the
new "volume attach" resources for further block storage functionality.
This commit makes the openstack_blockstorage_volume resources better able
to handle volume creation errors upon resource creation. The cause of this
change is because there could be some storage backend error that happens
during storage provisioning that won't manifest in an "err" but will set
the volume's status to "error". We now check for a status of "error" and
propagate the error up the stack.
* Implementing Redis Cache
* Properties should never be nil
* Updating the SDK to 7.0.1
* Redis Cache updated for SDK 7.0.1
* Fixing the max memory validation tests
* Cleaning up
* Adding tests for Standard with Tags
* Int's -> Strings for the moment
* Making the RedisConfiguration object mandatory
* Only parse out redis configuration values if they're set
* Updating the RedisConfiguration object to be required in the documentaqtion
* Adding Tags to the Standard tests / importing excluding the redisConfiguration
* Removing support for import for Redis Cache for now
* Removed a scaling test
* Drop alias from state file if missing from lambda.
This commit fixes an issue where if you remove a AWS Lambda, the corresponding alias for that Lambda is also deleted.
* Added missing imports.
* Removed non-local reference to constant.
* vendor: update github.com/Ensighten/udnssdk to v1.2.1
* ultradns_tcpool: add
* ultradns.baseurl: set default
* ultradns.record: cleanup test
* ultradns_record: extract common, cleanup
* ultradns: extract common
* ultradns_dirpool: add
* ultradns_dirpool: fix rdata.ip_info.ips to be idempotent
* ultradns_tcpool: add doc
* ultradns_dirpool: fix rdata.geo_codes.codes to be idempotent
* ultradns_dirpool: add doc
* ultradns: cleanup testing
* ultradns_record: rename resource
* ultradns: log username from config, not client
udnssdk.Client is being refactored to use x/oauth2, so don't assume we
can access Username from it
* ultradns_probe_ping: add
* ultradns_probe_http: add
* doc: add ultradns_probe_ping
* doc: add ultradns_probe_http
* ultradns_record: remove duplication from error messages
* doc: cleanup typos in ultradns
* ultradns_probe_ping: add test for pool-level probe
* Clean documentation
* ultradns: pull makeSetFromStrings() up to common.go
* ultradns_dirpool: log hashIPInfoIPs
Log the key and generated hashcode used to index ip_info.ips into a set.
* ultradns: simplify hashLimits()
Limits blocks only have the "name" attribute as their primary key, so
hashLimits() needn't use a buffer to concatenate.
Also changes log level to a more approriate DEBUG.
* ultradns_tcpool: convert rdata to schema.Set
RData blocks have the "host" attribute as their primary key, so it is
used by hashRdatas() to create the hashcode.
Tests are updated to use the new hashcode indexes instead of natural
numbers.
* ultradns_probe_http: convert agents to schema.Set
Also pull the makeSetFromStrings() helper up to common.go
* ultradns: pull hashRdatas() up to common
* ultradns_dirpool: convert rdata to schema.Set
Fixes TF-66
* ultradns_dirpool.conflict_resolve: fix default from response
UltraDNS REST API User Guide claims that "Directional Pool
Profile Fields" have a "conflictResolve" field which "If not
specified, defaults to GEO."
https://portal.ultradns.com/static/docs/REST-API_User_Guide.pdf
But UltraDNS does not actually return a conflictResolve
attribute when it has been updated to "GEO".
We could fix it in udnssdk, but that would require either:
* hide the response by coercing "" to "GEO" for everyone
* use a pointer to allow checking for nil (requires all
users to change if they fix this)
An ideal solution would be to have the UltraDNS API respond
with this attribute for every dirpool's rdata.
So at the risk of foolish consistency in the sdk, we're
going to solve it where it's visible to the user:
by checking and overriding the parsing. I'm sorry.
* ultradns_record: convert rdata to set
UltraDNS does not store the ordering of rdata elements, so we need a way
to identify if changes have been made even it the order changes.
A perfect job for schema.Set.
* ultradns_record: parse double-encoded answers for TXT records
* ultradns: simplify hashLimits()
Limits blocks only have the "name" attribute as their primary key, so
hashLimits() needn't use a buffer to concatenate.
* ultradns_dirpool.description: validate
* ultradns_dirpool.rdata: doc need for set
* ultradns_dirpool.conflict_resolve: validate
tags were documented, just not implemented
TF_ACC=1 go test ./builtin/providers/azurerm -v -run TestAccAzureRMDnsZone_ -timeout 120m
=== RUN TestAccAzureRMDnsZone_importBasic
--- PASS: TestAccAzureRMDnsZone_importBasic (89.04s)
=== RUN TestAccAzureRMDnsZone_basic
--- PASS: TestAccAzureRMDnsZone_basic (92.91s)
=== RUN TestAccAzureRMDnsZone_withTags
--- PASS: TestAccAzureRMDnsZone_withTags (105.88s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/azurerm 287.912s
* provider/pagerduty: Allow 'team_responder' role for pagerduty_user resource
* Change unit test to exercise 'team_responder' and reformat
* Update the test fixture to use the 'team_responder' role
* provider/aws: Support eu-west-2
This is the new London region - we don't have access yet but several
enquiries have come from customers who do.
* provider/aws: Support eu-west-2 region
* Update hosted_zones.go
Peter McAtominey