I am trying to specify an alternative path to the kubeconfig:
provider "helm" {
kubernetes {
config_path = "~/path/to/config"
}
}
but I am seeing this:
Error running plan: 1 error(s) occurred:
* provider.helm: error creating tunnel: "Get http://localhost:8080/api/v1/namespaces/kube-system/pods?labelSelector=app%3Dhelm%2Cname%3Dtiller: dial tcp [::1]:8080: getsockopt: connection refused"
Terraform v0.9.11
I think you should be using terraform 0.10.0 or grater.
@calind You might be correct. Just double checked the requirements.
although, I don't see any logic for handling that here: https://github.com/mcuadros/terraform-provider-helm/blob/master/helm/provider.go#L195-L243
I am fixing it right now. Thanks for reporting it.
@mcuadros I see Terraform 0.10 as a requirement. Unfortunately, Terragrunt does not work with that version of TF yet. I am seeing some strange things with this provider and I don't know if it could be due to the fact that I am using TF 0.9.11. When I create charts, I get errors that some resources already exist (a ConfigMap for example). I am also seeing that a TF plan thinks it needs to install charts that already have been created successfully. Has something changed in TF 0.10 that would cause this? I can create another issue if I need to.
The provider was created for TF 0.10, but I must to say that is completely untested from my side, I never used. I am starting to use this days.
The logic is not very complex, so I don't see why should fail with the plans, anyways as I said, let me use it first.
I am closing this, if you have more issues, can you open another issue. Thanks!
@mcuadros This issue still exists. #8 fixed the issues I have in the comment, but not the config_path issue.
https://github.com/mcuadros/terraform-provider-helm/releases/tag/v0.2.0
Can you use this version, I just committed the fix.
Ahh yes. Thank you for this provider! Has really fit into our workflow nicely.