v0.11.1
provider "helm" {
version = "~> 0.7.0"
//debug = true
install_tiller = false
service_account = "tiller"
namespace = "${local.namespace}"
kubernetes {
host = "${var.eks_endpoint}"
cluster_ca_certificate = "${base64decode(var.eks_ca_data)}"
token = "${data.external.aws-iam-authenticator.result.token}"
}
}
Error: Error refreshing state: 1 error(s) occurred:
* provider.helm: failed to configure kubernetes config: username/password or bearer token may be set, but not both
none
It should have used the provided token and completed the plan with success.
It acts like username and password along with token were provided.
terraform planRunning on AWS EKS.
I'm running into this as well.
Same issue with provider version 0.10.0
I've submitted a PR with a partial fix, but this also requires #304 to be merged so load_config_file = false is actually correctly detected.
Closing this. Feel free to reopen if the #307 doesn't fix it.
terraform version 0.12.12
helm provider version 0.10.4
The problem appeared, earlier with 0.10.2 version of helm-provider it doesn't exist.
P.S. The issue appears while working with Atlantis, on my local MacOS everything fine with 0.10.4
@rporres should I open a new ticket or you will re-open this one?
@vliubko what's your provider config? are you getting the same username/password or bearer token may be set error?
We've been running with two of these fixes for some time both locally and in Atlantis, and it worked fine for us.
@gordonbondon
1) Yes, error is the same. failed to configure kubernetes config: username/password or bearer token may be set, but not both
2) config:
provider "helm" {
version = "0.10.4"
tiller_image = "gcr.io/kubernetes-helm/tiller:v2.14.0"
service_account = var.service_account
kubernetes {
host = var.gke_cluster.primary_endpoint
cluster_ca_certificate = var.gke_ca_certificate
username = var.username
password = var.password
}
}
Adding load_config_file = false helped, thanks to @gordonbondon
Facing similar issue with:
terraform version = 0.12.15
and
provider "kubernetes" {
version = "1.10"
host = module.gke.gke_cluster_endpoint
username = module.gke.gke_cluster_user
password = module.gke.gke_cluster_password
client_certificate = base64decode(module.gke.gke_cluster_client_certificate)
client_key = base64decode(module.gke.gke_cluster_client_key)
cluster_ca_certificate = base64decode(module.gke.gke_cluster_ca_certificate)
load_config_file = false
}
Error: Failed to configure: username/password or bearer token may be set, but not both
on provider.tf line 18, in provider "kubernetes":
provider "kubernetes"
@avish1990 this is a known issue in kubernetes provider, not related to helm provider. You can see discussion and some workarounds here https://github.com/terraform-providers/terraform-provider-kubernetes/issues/679 ,
@gordonbondon Thanks! This workaround - https://github.com/terraform-providers/terraform-provider-kubernetes/issues/679#issuecomment-552119320 worked for me.
Most helpful comment
Adding
load_config_file = falsehelped, thanks to @gordonbondon