v3.0.1 consuming huge amount of memory with pre-existing v.2.53.1 tfstate

[javespru]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritise this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritise the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform (and AzureAD Provider) Version

Affected Resource(s)

• azuread_3.0.1
• terraform 1.9.7

Terraform Configuration Files

# Copy-paste your Terraform configurations here - for large Terraform configs,
# please use a service like Dropbox and share a link to the ZIP file. For
# security, you can also encrypt the files using our GPG public key: https://keybase.io/hashicorp

My terraform configuration files consist of 181 groups with an average of 30 users per group.
It worked ok with v.2.53.1 of the provider in a VM with 16 GB.

Debug Output

N/A

Panic Output

N/A

Expected Behavior

After upgrading to provider 3.0.1 I added up just 1 user to any group, run

1. terraform init --upgrade
2. terraform apply

And the user should be provisioned.

Actual Behavior

With my original VM where I ran previously (v.2.53.1) the given configuration I got a gRPC error related to lack of memory (I verified with the top command)
If I increase the VM RAM to 128 GB, it takes a while, but in the end, the VM's memory is exhausted.

Steps to Reproduce

Having a huge configuration of groups and users 180+, migrate the provider from v.2.53.1 to v.3.0.1

1. terraform apply

Important Factoids

N/A

References

[javespru]

My provider file is case is relevant
_terraform {
required_version = ">= 1.5.1, <= 1.9.9"

required_providers {
azurerm = {
source = "hashicorp/azurerm"
version = "4.3"
}
azuread = {
source = "hashicorp/azuread"
version = "3.0.1"
}

}

backend "azurerm" {
subscription_id = "XXXXXX"
resource_group_name = "easa-comm-rgp-01-seed"
storage_account_name = "tfseedcom22uqwswrh"
container_name = "d4s-seed-common"
key = "d4s-infra/011-group-users.tfstate"
}

}

provider "azurerm" {
features {}
subscription_id = "XXXXXX"
resource_provider_registrations = "none"
}

provider "azuread" {
tenant_id = "YYYYYYYY"
}_

[javespru]

I tested with a fresh new tfstate v3.0.1 and 1 group with 2 users, and it exhausted 16 GB RAM

test.tf

resource "azuread_group" "group-test-01" {
  display_name = "delete-group-test-1212"

  security_enabled = true

  owners = ["27478bc9-e081-4144-9063-xxxx"]
  members = [
    "27478bc9-e081-4144-9063-yyy",
    "250a11b2-0b4b-4b24-adf2-yyy"
  ]
}

error message

Error: Request cancelled
│ 
│   with azuread_group.group-test-01,
│   on test.tf line 1, in resource "azuread_group" "group-test-01":
│    1: resource "azuread_group" "group-test-01" {
│ 
│ The plugin.(*GRPCProvider).ApplyResourceChange request was cancelled.

[cohendiego]

We are noticing this exact same behavior with this latest version of azure_ad provider.

[donjuanmon]

Same here, only azuread_group is displaying this behavior for us.