This message was deleted.
# atlantis-community
s
This message was deleted.
r
This is the output when I run the same command on the Atlantis POD.
Copy code
atlantis-0:/atlantis-data/repos/rtmuller/azure-lab/9/default/atlantis3$ /usr/local/bin/terraform init -input=false -upgrade

Initializing the backend...

Initializing provider plugins...
- Finding hashicorp/azurerm versions matching "~> 3.60.0"...
- Using previously-installed hashicorp/azurerm v3.60.0

Terraform has been successfully initialized!
Backend configuration file (TF)
Copy code
terraform {
  required_providers {
    azurerm = {
      source  = "hashicorp/azurerm"
      version = "~>3.60.0"
    }
  }
  backend "azurerm" {
    resource_group_name  = "Engineering"
    storage_account_name = "xxxx"
    container_name       = "tfstate-demo-azure"
    key                  = "atlantis03.tfstate"
  }
}
provider "azurerm" {
  features {}
  client_id       = "xxxx"
  client_secret   = "xxxxx"
  tenant_id       = "xxxxx"
  subscription_id = "xxxxx"
  storage_use_azuread  = true
}
g
That’s probably because while logging into the pod you are setting the $PATH correctly from your interactive session. But the Atlantis session from the PR doesn’t know where to find the
az
command.
Check what you have on your PATH and set that on an env variable on the Atlantis container.
r
I got it, but I didn't install the az command on the POD.
Copy code
atlantis-0:/atlantis-data/repos/rtmuller/azure-lab$ az
bash: az: command not found
I'm trying to figure out the differences between the PR and the POD execution.
g
That might be something related to the azurerm terraform provider. Are you using the same user as the Atlantis service running on the pod?
r
The provider looks good because removing the backend configuration to use a local backend work well.
Is something related to the backend azurerm
but again, if I run the terraform init and plan from the POD works fine.
🤯
I've changed to using UserAssigned with all necessary permissions, and it worked. Thx.
🙌 1