Error could not download chart: looks like URL is not a valid chart repository or cannot be reached

[AErmie]

Description

I'm deploying an EKS cluster using the EKS Blueprint version 4.24.0. When the pipeline runs, several errors are reported, specifically about applying helm charts.

Note: This is identical to the error I reported here: Error: Could not download chart cluster_proportional_autoscaler with module CoreDNS, though CoreDNS does deploy.

When I check the URLs that the module is trying to interact with, the OPA Gatekeeper URL throws a 404 error.

It throws a different error for the KubeCost module: could not download chart: mkdir /.cache: permission denied

But the helm chart URL for External-Secrets is functional: https://charts.external-secrets.io/. It just seems like the module is incorrectly including the ending quote (") as part of the URL, and therefore cannot reach it.

• ✋ I have searched the open/closed issues and my issue is not listed.

Versions

• Module version [Required]: v1.24.0

• Terraform version: 1.3.9

• Provider version(s):
  • hashicorp/local v2.3.0 (signed by HashiCorp)
  • hashicorp/aws v4.56.0 (signed by HashiCorp)
  • hashicorp/helm v2.9.0 (signed by HashiCorp)
  • gavinbunney/kubectl v1.14.0 (self-signed, key ID AD64217B5ADD572F)
  • hashicorp/kubernetes v2.18.1 (signed by HashiCorp)
  • hashicorp/random v3.4.3 (signed by HashiCorp)
  • terraform-aws-modules/http v2.4.1 (self-signed, key ID B2C1C0641B6B0EB7)
  • hashicorp/null v3.2.1 (signed by HashiCorp)
  • hashicorp/tls v4.0.4 (signed by HashiCorp)
  • hashicorp/cloudinit v2.3.2 (signed by HashiCorp)
  • hashicorp/time v0.9.1 (signed by HashiCorp)
  • hashicorp/http v3.2.1 (signed by HashiCorp)

Reproduction Code [Required]

locals {
  cluster_version= "1.25"
}

module "eks_blueprints_kubernetes_addons" {
  source = "github.com/aws-ia/terraform-aws-eks-blueprints//modules/kubernetes-addons?ref=v4.24.0"

  eks_cluster_id = module.eks_blueprints.eks_cluster_id

  # EKS Addons
  enable_external_secrets = true
  enable_gatekeeper = true
  enable_kubecost = true 
}

Steps to reproduce the behavior:

1. terraform init
2. terraform plan
3. terraform apply

Expected behaviour

I expect the EKS add-ons to be fully deployed.

Actual behaviour

The terraform apply fails and causes the pipeline to error out.

Terminal Output Screenshot(s)

Additional context

This is being run in an Azure Pipeline, using Azure DevOps (online, not server). Also, this is not using self-managed build services, but rather, uses the Microsoft-provided cloud build servers.

[askulkarni2]

@AErmie we do not see this issue. It seems like an issue in your local setup. I am not familiar with how Azure Pipeline works but in general running helm repo update command is known to get rid of these type of error.

[AErmie]

Thanks @askulkarni2, but in this case/context, I don't think running helm repo update is an option, because applying these charts is triggered/controlled through the use of the EKS Blueprint addons module.

Unless is it your suggestion to simply run helm repo update as an independent step before running terraform apply?

[bryantbiggs]

just FYI - this is a provider level issue, not something we can resolve in a module level implementation hashicorp/terraform-provider-helm#630

[askulkarni2]

Unless is it your suggestion to simply run helm repo update as an independent step before running terraform apply?

Yes exactly.