terraform-provider-helm: Getting Wrong Error: chart [...] version [...] not found in [...] repository

Terraform, Provider, Kubernetes and Helm Versions

Terraform version: 0.14.8
Provider version: 2.0.3
Kubernetes version: 1.19.1
Helm version: 3.5.2

Affected Resource(s)

  • helm_release

Terraform Configuration Files

terraform {
  required_providers {
    helm = {
      source  = "hashicorp/helm"
      version = "2.0.3"
    }
  }
}

provider "helm" {
  kubernetes {
    config_path = "~/.kube/config"
  }
}

resource "helm_release" "nginx_ingress" {
  name       = "nginx-ingress"
  repository = "https://kubernetes.github.io/ingress-nginx"
  chart      = "nginx-ingress"
  version    = "3.24.0"
  namespace  = "kube-system"

  values = [
    file("helm/nginx-ingress/values.yml")
  ]
}

Debug Output

For full log see:

I cannot find any useful information there but:

2021-03-13T03:19:40.079Z [DEBUG] plugin: starting plugin: path=.terraform/providers/registry.terraform.io/hashicorp/helm/2.0.3/linux_amd64/terraform-provider-helm_v2.0.3_x5 args=[.terraform/providers/registry.terraform.io/hashicorp/helm/2.0.3/linux_amd64/terraform-provider-helm_v2.0.3_x5]
2021-03-13T03:19:40.081Z [DEBUG] plugin: plugin started: path=.terraform/providers/registry.terraform.io/hashicorp/helm/2.0.3/linux_amd64/terraform-provider-helm_v2.0.3_x5 pid=13458
2021-03-13T03:19:40.082Z [DEBUG] plugin: waiting for RPC address: path=.terraform/providers/registry.terraform.io/hashicorp/helm/2.0.3/linux_amd64/terraform-provider-helm_v2.0.3_x5
2021-03-13T03:19:40.147Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: configuring server automatic mTLS: timestamp=2021-03-13T03:19:40.147Z
2021-03-13T03:19:40.223Z [DEBUG] plugin.terraform-provider-helm_v2.0.3_x5: plugin address: address=/tmp/plugin426193591 network=unix timestamp=2021-03-13T03:19:40.223Z
2021-03-13T03:19:40.223Z [DEBUG] plugin: using plugin: version=5
2021-03-13T03:19:40.301Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [DEBUG] [resourceDiff: nginx-ingress] Start: timestamp=2021-03-13T03:19:40.300Z
2021/03/13 03:19:40 [WARN] Provider "registry.terraform.io/hashicorp/helm" produced an invalid plan for helm_release.nginx_ingress, but we are tolerating it because it is using the legacy plugin SDK.
    The following problems may be the cause of any confusing errors from downstream operations:
      - .atomic: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .lint: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .verify: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .wait: planned value cty.True does not match config value cty.NullVal(cty.Bool)
      - .cleanup_on_fail: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .replace: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .timeout: planned value cty.NumberIntVal(300) does not match config value cty.NullVal(cty.Number)
      - .dependency_update: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .disable_webhooks: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .force_update: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .disable_crd_hooks: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .skip_crds: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .disable_openapi_validation: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .recreate_pods: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .reuse_values: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .create_namespace: planned value cty.False does not match config value cty.NullVal(cty.Bool)
      - .max_history: planned value cty.NumberIntVal(0) does not match config value cty.NullVal(cty.Number)
ty.NullVal(cty.Bool)
      - .reset_values: planned value cty.False does not match config value cty.NullVal(cty.Bool)
helm_release.nginx_ingress: Creating...
2021/03/13 03:19:40 [DEBUG] EvalApply: ProviderMeta config value set
2021/03/13 03:19:40 [DEBUG] helm_release.nginx_ingress: applying the planned Create change
2021-03-13T03:19:40.367Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [DEBUG] setting computed for "metadata" from ComputedKeys: timestamp=2021-03-13T03:19:40.366Z
2021-03-13T03:19:40.367Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [DEBUG] [resourceReleaseCreate: nginx-ingress] Started: timestamp=2021-03-13T03:19:40.367Z
2021-03-13T03:19:40.367Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [DEBUG] [resourceReleaseCreate: nginx-ingress] Getting helm configuration: timestamp=2021-03-13T03:19:40.367Z
2021-03-13T03:19:40.367Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [DEBUG] [INFO] GetHelmConfiguration start: timestamp=2021-03-13T03:19:40.367Z
2021-03-13T03:19:40.367Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [DEBUG] Using kubeconfig: /home/runner/.kube/config: timestamp=2021-03-13T03:19:40.367Z
2021-03-13T03:19:40.367Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [INFO] Successfully initialized kubernetes config: timestamp=2021-03-13T03:19:40.367Z
2021-03-13T03:19:40.367Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [DEBUG] [INFO] GetHelmConfiguration success: timestamp=2021-03-13T03:19:40.367Z
2021-03-13T03:19:40.367Z [INFO]  plugin.terraform-provider-helm_v2.0.3_x5: 2021/03/13 03:19:40 [DEBUG] [resourceReleaseCreate: nginx-ingress] Getting chart: timestamp=2021-03-13T03:19:40.367Z
2021/03/13 03:19:40 [DEBUG] helm_release.nginx_ingress: apply errored, but we're indicating that via the Error pointer rather than returning it: chart "nginx-ingress" version "3.24.0" not found in https://kubernetes.github.io/ingress-nginx repository

Error: chart "nginx-ingress" version "3.24.0" not found in https://kubernetes.github.io/ingress-nginx repository

  on main.tf line 16, in resource "helm_release" "nginx_ingress":
  16: resource "helm_release" "nginx_ingress" {


2021-03-13T03:19:40.404Z [WARN]  plugin.stdio: received EOF, stopping recv loop: err="rpc error: code = Unavailable desc = transport is closing"
2021-03-13T03:19:40.405Z [DEBUG] plugin: plugin process exited: path=.terraform/providers/registry.terraform.io/hashicorp/helm/2.0.3/linux_amd64/terraform-provider-helm_v2.0.3_x5 pid=13458
2021-03-13T03:19:40.405Z [DEBUG] plugin: plugin exited
Error: Terraform exited with code 1.
Error: Process completed with exit code 1.

Steps to Reproduce

  1. Use https://kubernetes.github.io/ingress-nginx as a helm_release repository with chart nginx-ingress and version 3.24.0
  2. See https://github.com/pathob/tf_helm/blob/main/main.tf
  3. See https://github.com/pathob/tf_helm/blob/main/.github/workflows/test.yml
  4. See https://github.com/pathob/tf_helm/runs/2095542326?check_suite_focus=true

Expected Behavior

  1. Terraform should find the chart like e.g. with using repository https://helm.nginx.com/stable and chart nginx-ingress and version 0.8.0 (but there is something else odd with this chart);
  2. See https://github.com/pathob/tf_helm/blob/working/main.tf
  3. See https://github.com/pathob/tf_helm/blob/working/.github/workflows/test.yml
  4. See https://github.com/pathob/tf_helm/runs/2093139142?check_suite_focus=true

Actual Behavior

Terraform cannot find the chart like mentioned in the expected behaviour, although I’m using the same chart / version / repo with Helm / Helmfile and the chart with that version in that repo definetly exists, see https://kubernetes.github.io/ingress-nginx/index.yaml

Workaround

I’m currently using the full chart URL as the chart without repo and version like so: https://github.com/kubernetes/ingress-nginx/releases/download/helm-chart-3.24.0/ingress-nginx-3.24.0.tgz

Important Factoids

I think none.

References

I think none.

Community Note

  • Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
  • If you are interested in working on this issue or have submitted a pull request, please leave a comment

About this issue

  • Original URL
  • State: open
  • Created 3 years ago
  • Reactions: 25
  • Comments: 18

Most upvoted comments

I was having the same issue trying to install the consul. Usually to do that manually I run:

helm repo add hashicorp https://helm.releases.hashicorp.com
helm install hashicorp/consul

so the naive approach for me was to use following resource configuration:

resource "helm_release" "consul" {
  name             = "consul"
  repository       = "https://helm.releases.hashicorp.com"
  chart            = "hashicorp/consul"
}

It fails with the same error everyone else is having so I tried to “bruteforce” it and found the way to make it work:

resource "helm_release" "default" {
  name             = "consul"
  repository       = "https://helm.releases.hashicorp.com"
  chart            = "consul"
}

The more I think of it the more sense it makes. When I run helm repo ls it returns:

NAME                            URL
hashicorp                       https://helm.releases.hashicorp.com

so to install the chart from that specific repo I have to prefix my chart name with a repo name, hashicorp/consul, while Terraform Helm provider does not need that prefix.

+23
trolleksii on Nov 10, 2021

I was getting the same error. My work around was to do away with the repository and version arguments.

Instead of this

resource "helm_release" "nginx_ingress" {
  name       = "nginx-ingress"
  repository = "https://kubernetes.github.io/ingress-nginx"
  chart      = "nginx-ingress"
  version    = "3.24.0"
  namespace  = "kube-system"

  values = [
    file("helm/nginx-ingress/values.yml")
  ]
}

I retrieved the install link from Artifacthub and supplied that to the chart argument

resource "helm_release" "nginx_ingress" {
  name       = "nginx-ingress"
  chart      = "https://github.com/kubernetes/ingress-nginx/releases/download/helm-chart-3.24.0/ingress-nginx-3.24.0.tgz"
  namespace  = "kube-system"

  values = [
    file("helm/nginx-ingress/values.yml")
  ]
}

Also note that the version is part of the install link, so no need to specify the version argument as well. More about the chart argument here.

+11
aidan-melen on May 19, 2021

I’ll double post instead of edit because I think it may be useful to other people.

When using the format:

resource "helm_release" "prometheus_stack" {
  name       = "ps"
  namespace  = "monitoring"
  repository = "https://prometheus-community.github.io/helm-charts"
  chart      = "prometheus-community/kube-prometheus-stack"
  version    = "18.0.13"
}

It fails with the mentioned error:

Error: failed to download "https://github.com/prometheus-community/helm-charts/releases/download/kube-prometheus-stack-18.0.13/kube-prometheus-stack-18.0.13.tgz" at version "18.0.13" (hint: running `helm repo update` may help)

But, if you manually add your repo with helm:

helm repo add prometheus-community https://prometheus-community.github.io/helm-charts

And change your code to rely on your local list:

resource "helm_release" "prometheus_stack" {
  name       = "ps"
  namespace  = "monitoring"
  chart      = "prometheus-community/kube-prometheus-stack"
  version    = "18.0.13"
}

It works perfectly. So it seems that whatever is the problem, is when terraform tries to add the repository.

I was to just post this comment, but I decided to try another thing. Just to see if I could reproduce the problem, I deleted the prometheus-community repository from the ~/.config/helm/repositories.yaml file. And then I tried again with the first method:

resource "helm_release" "prometheus_stack" {
  name       = "ps"
  namespace  = "monitoring"
  repository = "https://prometheus-community.github.io/helm-charts"
  chart      = "prometheus-community/kube-prometheus-stack"
  version    = "18.0.13"
}

And it worked as expected.

Then, again to see if I can reproduce some problem, I added the repo again to the ~/.config/helm/repositories.yaml file:

helm repo add prometheus-community https://prometheus-community.github.io/helm-charts

And applied again with the last format, and it still worked as expected.

So it seems that messing with the ~/.config/helm/repositories.yaml file can fix the problem, even when theoretically terraform doesn’t use it, since all the information is in the code.

My conclusion is, I know nothing.

+1
alexppg on Sep 29, 2021

Hello @pathob ! My workaround is to downgrade to 1.3.2 https://registry.terraform.io/providers/hashicorp/helm/1.3.2/docs/data-sources/repository and using helm_repository data resource.

EDIT: Ignore this post… this was working locally because Terraform Helm was using my local Helm configuration which has the configuration for ChartMuseum. Check this post instead https://github.com/hashicorp/terraform-provider-helm/issues/707#issuecomment-844548333

👎

I can confirm that this solution worked for me. Using a private ChartMuseum repository with basic HTTP auth and it does not with the provider versions

terraform: 0.12.31
helm: 2.1.2

Not working

helm provider: 2.1.2

helm_release xxx {
  repository          = "https://chartmuseum.domain.com/"
  repository_username = data.aws_secretsmanager_secret_version.chartmuseum_credentials["username"].secret_string
  repository_password = data.aws_secretsmanager_secret_version.chartmuseum_credentials["password"].secret_string
}

Working

helm provider: 1.3.2

helm_release xxx {
  repository = data.helm_repository.chartmuseum.metadata[0].name
}

data helm_repository chartmuseum {
  name     = "chartmuseum"
  URL      = "https://chartmuseum.domain.com/"
  username = data.aws_secretsmanager_secret_version.chartmuseum_credentials["username"].secret_string
  password = data.aws_secretsmanager_secret_version.chartmuseum_credentials["password"].secret_string
}
+1
carlosjgp on Jun 1, 2021

Hello @pathob ! My workaround is to downgrade to 1.3.2 https://registry.terraform.io/providers/hashicorp/helm/1.3.2/docs/data-sources/repository and using helm_repository data resource.

+1
serhiiromaniuk on Apr 13, 2021
Read more comments on GitHub
← terraform-provider-helm: First helm release always succeeds and doesn't wait for all pods running
terraform-provider-helm: Helm Chart for "stable/instana-agent" not working with terraform 0.13 →