My objective is to be able to set up an IAM Role which can assume a role of a certain IAM user. After the creation of the role, I would like to come back later and modify this role by adding external IDs to establish a trust relationship. Let me illustrate with an example:
Let's say I want to create role:
resource "aws_iam_role" "happy_role" {
name = "happy-role"
assume_role_policy = data.aws_iam_policy_document.happy_assume_rule_policy.json
}
Let's also assume that happy_assume_role_policy looks something like:
data "aws_iam_policy_document" "happy_assume_role_policy" {
statement {
effect = "Allow"
actions = ["sts:AssumeRole"]
principals {
type = "AWS"
identifiers = [var.some_iam_user_arn]
}
}
}
Now, I will use the created role to create an external integration. But once I am done creating that integration, I want to go back to the role I originally created and modify it's assumed role policy. So now I want to add a condition to the assume role policy and make it look like:
data "aws_iam_policy_document" "happy_assume_role_policy" {
statement {
effect = "Allow"
actions = ["sts:AssumeRole"]
principals {
type = "AWS"
identifiers = [var.snowflake_iam_user_arn]
}
condition {
test = "StringEquals"
values = [some_integration.integration.external_id]
variable = "sts:ExternalId"
}
}
}
In other words, my workflow should be like:
Create role without assume conditions
Create an integration with that role
Take the ID from the created integration and go back to the created role and add a condition on it
Edit:
By "integration" I mean something like this. Once an Integration is created, there is an outputted ID, and then I need to take that ID and feed it back to the Assume Role I originally created. That should happen everytime I add a new integration.
I first tried to create two IAM roles, one for managing the integration creation, and another for managing the integration itself. That ran without circular reference errors; however, I was not able to establish a connection from the storage to the database, as it needs to be the same IAM Role creating and managing the integration.
This is what I have ended up doing (still not good for an accepted way to do it IMO). I created a role (with targeting) like:
resource "aws_iam_role" "happy_role" {
name = "happy-role"
assume_role_policy = data.aws_iam_policy_document.basic_policy.json
}
And used a basic assume role policy (without conditions). And then for the next run, I applied (without targeting) and it worked.
I followed the approach mentioned here, How to create a Snowflake Storage Integration with AWS S3 with Terraform?
As part of Storage Integration creation, just provide the role arn which is manually constructed without the resource is created. Terraform wont complain. Then create the role with the assume policy referring to the External Id and User ARN created by the Storage Integration
Related
I am trying to create an AWS VPC Endpoint Service (PrivateLink) where I can add Principals to those that already exist. Here is my current code
resource "aws_vpc_endpoint_service" "privatelink" {
provider = aws.customer
acceptance_required = true
network_load_balancer_arns = ["${aws_lb.nlb.arn}"]
}
resource "aws_vpc_endpoint_service_allowed_principal" "addition" {
provider = aws.customer
vpc_endpoint_service_id = aws_vpc_endpoint_service.privatelink.id
principal_arn = var.consumer_principal_arn
}
That works great for the one Principal specified in the variable but overwrites the existing Principal when I run it again with a different Principal. What I want is to append zero or more Principals to the list of existing Principals, each time I do a terraform apply. For example, the first time I run it, I specify Principal X. I run it again, specifying Principal Y. Now the list of allowed Principals is X and Y.
You would need to create multiple aws_vpc_endpoint_service_allowed_principal resources with each additional ARN. This way you can revoke principal(s) in the future without destroying other existing associations. Of course you can use for each loop and create aws_vpc_endpoint_service_allowed_principal resources with count and a list of principal ARNs. However, if you remove a principal from the list, associations for all the principals after the removed principal from the list will be recreated and the associations needs to be accepted again.
You can't edit the existing resource definition to add another principal. Terraform sees that as an update to the resource named "addition" and performs an update instead. Instead you need to add another aws_vpc_endpoint_service_allowed_principal resource.
I am currently having two (maybe conflicting) S3 bucket policies, which show a permanent difference on Terraform. Before I show parts of the code, I will try to give an overview of the structure.
I am currently using a module, which:
Takes IAM Role & an S3 Bucket as inputs
Attaches S3 Bucket policy to the inputted role
Attaches S3 Bucket (allowing VPC) policy to the inputted S3 bucket
I have created some code (snippet and not full code) to illustrate how this looks like for the module.
The policies look like:
# S3 Policy to be attached to the ROLE
data "aws_iam_policy_document" "foo_iam_s3_policy" {
statement {
effect = "Allow"
resources = ["${data. s3_bucket.s3_bucket.arn}/*"]
actions = ["s3:GetObject", "s3:GetObjectVersion"]
}
statement {
effect = "Allow"
resources = [data.s3_bucket.s3_bucket.arn]
actions = ["s3:*"]
}
}
# VPC Policy to be attached to the BUCKET
data "aws_iam_policy_document" "foo_vpc_policy" {
statement {
sid = "VPCAllow"
effect = "Allow"
resources = [data.s3_bucket.s3_bucket.arn, "${data.s3_bucket.s3_bucket.arn}/*"]
actions = ["s3:GetObject", "s3:GetObjectVersion"]
condition {
test = "StringEquals"
variable = "aws:SourceVpc"
values = [var.foo_vpc]
}
principals {
type = "*"
identifiers = ["*"]
}
}
}
The policy attachments look like:
# Turn policy into a resource to be able to use ARN
resource "aws_iam_policy" "foo_iam_policy_s3" {
name = "foo-s3-${var.s3_bucket_name}"
description = "IAM policy for foo on s3"
policy = data.aws_iam_policy_document.foo_iam_s3_policy.json
}
# Attaches s3 bucket policy to IAM Role
resource "aws_iam_role_policy_attachment" "foo_attach_s3_policy" {
role = data.aws_iam_role.foo_role.name
policy_arn = aws_iam_policy.foo_iam_policy_s3.arn
}
# Attach foo vpc policy to bucket
resource "s3_bucket_policy" "foo_vpc_policy" {
bucket = data.s3_bucket.s3_bucket.id
policy = data.aws_iam_policy_document.foo_vpc_policy.json
}
Now let's step outside of the module, where the S3 bucket (the one I mentioned that will be inputted into the module) is created, and where another policy needs to be attached to it (the S3 bucket). So outside of the module, we:
Provide an S3 bucket to the aforementioned module as input (alongside the IAM Role)
Create a policy to allow some IAM Role to put objects in the aforementioned bucket
Attach the created policy to the bucket
The policy looks like:
# Create policy to allow bar to put objects in the bucket
data "aws_iam_policy_document" "bucket_policy_bar" {
statement {
sid = "Bar IAM access"
effect = "Allow"
resources = [module.s3_bucket.bucket_arn, "${module. s3_bucket.bucket_arn}/*"]
actions = ["s3:PutObject", "s3:GetObject", "s3:ListBucket"]
principals {
type = "AWS"
identifiers = [var.bar_iam]
}
}
}
And its attachment looks like:
# Attach Bar bucket policy
resource "s3_bucket_policy" "attach_s3_bucket_bar_policy" {
bucket = module.s3_bucket.bucket_name
policy = data.aws_iam_policy_document.bucket_policy_bar.json
}
(For more context: Basically foo is a database that needs VPC and s3 attachment to role to operate on the bucket and bar is an external service that needs to write data to the bucket)
What is going wrong
When I try to plan/apply, Terraform shows that there is always change, and shows an overwrite between the S3 bucket policy of bar (bucket_policy_bar) and the VPC policy attached inside the module (foo_vpc_policy).
In fact the error I am getting kind of sounds like what is described here:
The usage of this resource conflicts with the
aws_iam_policy_attachment resource and will permanently show a
difference if both are defined.
But I am attaching policies to S3 and not to a role, so I am not sure if this warning applies to my case.
Why are my policies conflicting? And how can I avoid this conflict?
EDIT:
For clarification, I have a single S3 bucket, to which I need to attach two policies. One that allows VPC access (foo_vpc_policy, which gets created inside the module) and another one (bucket_policy_bar) that allows IAM role to put objects in the bucket
there is always change
That is correct. aws_s3_bucket_policy sets new policy on the bucket. It does not add new statements to it.
Since you are invoking aws_s3_bucket_policy twice for same bucket, first time in module.s3_bucket module, then second time in parent module (I guess), the parent module will simply attempt to set new policy on the bucket. When you perform terraform apply/plan again, the terraform will detect that the policy defined in module.s3_bucket is different, and will try to update it. So you end up basically with a circle, where each apply will change the bucket policy to new one.
I'm not aware of a terraform resource which would allow you to update (i.e. add new statements) to an existing bucket policy. Thus I would try to re-factor your design so that you execute aws_s3_bucket_policy only once with all the statements that you require.
Thanks to the tip from Marcin I was able to resolve the issue by making the attachment of the policy inside the module optional like:
# Attach foo vpc policy to bucket
resource "s3_bucket_policy" "foo_vpc_policy" {
count = var.attach_vpc_policy ? 1 : 0 # Only attach VPC Policy if required
bucket = data.s3_bucket.s3_bucket.id
policy = data.aws_iam_policy_document.foo_vpc_policy.json
}
The policy in all cases has been added as output of the module like:
# Outputting only the statement, as it will later be merged with other policies
output "foo_vpc_policy_json" {
description = "VPC Allow policy json (to be later merged with other policies that relate to the bucket outside of the module)"
value = data.aws_iam_policy_document.foo_vpc_policy.json
}
For the cases when it was needed to defer the attachment of the policy (wait to attach it together with another policy), I in-lined the poliicy via source_json)
data "aws_iam_policy_document" "bucket_policy_bar" {
# Adding the VPC Policy JSON as a base for this Policy (docs: https://registry.terraform.io/providers/hashicorp/aws/latest/docs/data-sources/iam_policy_document)
source_json = module.foor_.foo_vpc_policy_json # here we add the statement that has
statement {
sid = "Bar IAM access"
effect = "Allow"
resources = [module.s3_bucket_data.bucket_arn, "${module.s3_bucket_data.bucket_arn}/*"]
actions = ["s3:PutObject", "s3:GetObject", "s3:ListBucket"]
principals {
type = "AWS"
identifiers = [var.bar_iam]
}
}
}
Here is the terraform code I have used to create a service account and bind a role to it:
resource "google_service_account" "sa-name" {
account_id = "sa-name"
display_name = "SA"
}
resource "google_project_iam_binding" "firestore_owner_binding" {
role = "roles/datastore.owner"
members = [
"serviceAccount:sa-name#${var.project}.iam.gserviceaccount.com",
]
depends_on = [google_service_account.sa-name]
}
Above code worked great... except it removed the datastore.owner from any other service account in the project that this role was previously assigned to. We have a single project that many teams use and there are service accounts managed by different teams. My terraform code would only have our team's service accounts and we could end up breaking other teams service accounts.
Is there another way to do this in terraform?
This of course can be done via GCP UI or gcloud cli without any issue or affecting other SAs.
From terraform docs, "google_project_iam_binding" is Authoritative. Sets the IAM policy for the project and replaces any existing policy already attached. That means that it replaces completely members for a given role inside it.
To just add a role to a new service account, without editing everybody else from that role, you should use the resource "google_project_iam_member":
resource "google_service_account" "sa-name" {
account_id = "sa-name"
display_name = "SA"
}
resource "google_project_iam_member" "firestore_owner_binding" {
project = <your_gcp_project_id_here>
role = "roles/datastore.owner"
member = "serviceAccount:${google_service_account.sa-name.email}"
}
Extra change from your sample: the use of service account resource email generated attribute to remove the explicit depends_on. You don't need the depends_on if you do it like this and you avoid errors with bad configuration.
Terraform can infer the dependency from the use of a variable from another resource. Check the docs here to understand this behavior better.
It's an usual problem with Terraform. Either you do all with it, or nothing. If you are between, unexpected things can happen!!
If you want to use terraform, you have to import the existing into the tfstate. Here the doc for the bindind, and, of course, you have to add all the account in the Terraform file. If not, the binding will be removed, but this time, you will see the deletion in the tf plan.
The terraform structure of my project is:
iam/policies/policy1.tf
iam/roles/roles1.tf
policy1.tf includes several policies, etc.:
resource "aws_iam_policy" "policy1A" {
name = "..."
policy = "${data.template_file.policy1ATempl.rendered}"
}
roles1.tf includes several roles, etc.:
resource "aws_iam_role" "role1" {
name = ....
assume_role_policy = ....
}
Now I want to attach policy1A to role1. Since policy1 and role1 are not in the same folder, how do I attach them?
resource "aws_iam_policy_attachment" "attach1" {
name = "attachment1"
roles = ??? (sth. like ["${aws_iam_role.role1.name}"])
policy_arn = ??? (sth. like "${aws_iam_policy.Policy1.arn}")
}
You can't make references like this across directories because Terraform only works in one directory at a time. If you really do need to separate these into different folders (which I would recommend against here) then normally you would use data sources to retrieve information about resources that have already been created.
In this case though the aws_iam_policy data source is currently pretty much useless because it requires the ARN of the policy instead of the path or name. Instead you can construct it if you know the policy name as it fits a well known pattern. You obviously know the name of role(s) you want to attach the policy to and you know the name of the policy so that covers those things.
data "aws_caller_identity" "current" {}
resource "aws_iam_policy_attachment" "attach1" {
name = "attachment1"
roles = "role1"
policy_arn = "arn:aws:iam::${data.aws_caller_identity.current.account_id}:policy/policy1A"
}
This above example does briefly show off how to use data sources as it fetches the AWS account ID, using the aws_caller_identity data source, and dynamically uses that to build the ARN of the policy you are trying to attach.
I would echo the warning in the docs against using the resource aws_iam_policy_attachment unless you are absolutely certain with what you're doing because it will detach the policy from anything else it happens to be attached to. If you want to take a managed policy and attach it to N roles I would instead recommend using aws_iam_role_policy_attachment and it's like instead.
I want to attach one of the pre-existing AWS managed roles to a policy, here's my current code:
resource "aws_iam_role_policy_attachment" "sto-readonly-role-policy-attach" {
role = "${aws_iam_role.sto-test-role.name}"
policy_arn = "arn:aws:iam::aws:policy/ReadOnlyAccess"
}
Is there a better way to model the managed policy and then reference it instead of hardcoding the ARN? It just seems like whenever I hardcode ARNs / paths or other stuff like this, I usually find out later there was a better way.
Is there something already existing in Terraform that models managed policies? Or is hardcoding the ARN the "right" way to do it?
The IAM Policy data source is great for this. A data resource is used to describe data or resources that are not actively managed by Terraform, but are referenced by Terraform.
For your example, you would create a data resource for the managed policy as follows:
data "aws_iam_policy" "ReadOnlyAccess" {
arn = "arn:aws:iam::aws:policy/ReadOnlyAccess"
}
The name of the data source, ReadOnlyAccess in this case, is entirely up to you. For managed policies I use the same name as the policy name for the sake of consistency, but you could just as easily name it readonly if that suits you.
You would then attach the IAM policy to your role as follows:
resource "aws_iam_role_policy_attachment" "sto-readonly-role-policy-attach" {
role = "${aws_iam_role.sto-test-role.name}"
policy_arn = "${data.aws_iam_policy.ReadOnlyAccess.arn}"
}
When using values that Terraform itself doesn't directly manage, you have a few options.
The first, simplest option is to just hard-code the value as you did here. This is a straightforward answer if you expect that the value will never change. Given that these "canned policies" are documented, built-in AWS features they likely fit this criteria.
The second option is to create a Terraform module and hard-code the value into that, and then reference this module from several other modules. This allows you to manage the value centrally and use it many times. A module that contains only outputs is a common pattern for this sort of thing, although you could also choose to make a module that contains an aws_iam_role_policy_attachment resource with the role set from a variable.
The third option is to place the value in some location that Terraform can retrieve values from, such as Consul, and then retrieve it from there using a data source. With only Terraform in play, this ends up being largely equivalent to the second option, though it means Terraform will re-read it on each refresh rather than only when you update the module using terraform init -upgrade, and thus this could be a better option for values that change often.
The fourth option is to use a specialized data source that can read the value directly from the source of truth. Terraform does not currently have a data source for fetching information on AWS managed policies, so this is not an option for your current situation, but can be used to fetch other AWS-defined data such as the AWS IP address ranges, service ARNs, etc.
Which of these is appropriate for a given situation will depend on how commonly the value changes, who manages changes to it, and on the availability of specialized Terraform data sources.
I got a similar situation, and I don't want to use the arn in my terraform script for two reasons,
If we do it on the Web console, we don't really look at the arn, we search the role name, then attach it to the role
The arn is not easy to remember, looks like is not for human
I would rather use the policy name, but not the arn, here is my example
# Get the policy by name
data "aws_iam_policy" "required-policy" {
name = "AmazonS3FullAccess"
}
# Create the role
resource "aws_iam_role" "system-role" {
name = "data-stream-system-role"
assume_role_policy = jsonencode({
Version = "2012-10-17"
Statement = [
{
Action = "sts:AssumeRole"
Effect = "Allow"
Sid = ""
Principal = {
Service = "ec2.amazonaws.com"
}
},
]
})
}
# Attach the policy to the role
resource "aws_iam_role_policy_attachment" "attach-s3" {
role = aws_iam_role.system-role.name
policy_arn = data.aws_iam_policy.required-policy.arn
}