Document keyvault post terraform manual steps

This is to document and configure the post-terraform commands necessary
for k8s hosts in the vmss to access the keyvault through flexvol.
This commit is contained in:
Rob Gil 2020-01-13 12:05:52 -05:00
parent 53cf42103e
commit 8416c18258

View File

@ -90,6 +90,8 @@ terraform apply
Check the output for errors. Sometimes the syntax is valid, but some of the configuration may be wrong and only rejected by the Azure API at run time. If this is the case, fix your mistake, and re-run.
# After running TF (Manual Steps)
## VM Scale Set
After running terraform, we need to make a manual change to the VM Scale Set that is used in the kubernetes. Terraform has a bug that is not applying this as of `v1.40` of the `azurerm` provider.
In order to get the `SystemAssigned` identity to be set, it needs to be set manually in the console.
@ -99,6 +101,38 @@ Navigate to the VM Scale Set for the k8s cluster you're managing (in the console
![SystemAssigned Identity](images/system-assigned.png)
_Just click the `Status` to `On`_
## KeyVault Policy
There is a bug (missing feature really) in the `azurerm` terraform provider which exposes the wrong `object_id/principal_id` in the `azurerm_kubernetes_cluster` output. The `id` that it exposes is the `object_id` of the cluster itself, and _not_ the Virtual Machine Scale Set SystemAssigned identity. This needs to be updated manually after running terraform for the first time.
To update, just edit the `keyvault.tf`. Set the `principal_id` to the `object_id` of the Virtual Machine Scale set. This can be found in the Azure portal, or via cli.
```
az vmss list
```
In that list, find the scale set for the k8s cluster you're working on. You'll want the value of `principal_id`.
The error looks like the following
```
Warning FailedMount 8s (x6 over 25s) kubelet, aks-default-54410534-vmss000001 MountVolume.SetUp failed for volume "flask-secret" : mount command failed, status: Failure, reason: /etc/kubernetes/volumeplugins/azure~kv/azurekeyvault-flex
volume failed, Access denied. Caller was not found on any access policy. r nCaller: appid=e6651156-7127-432d-9617-4425177c48f1;oid=f9bcbe58-8b73-4957-aee2-133dc3e58063;numgroups=0;iss=https://sts.windows.net/b5ab0e1e-09f8-4258-afb7-fb17654bc5
b3/ r nVault: cloudzero-dev-keyvault;location=eastus2 InnerError={code:AccessDenied}
```
Final configuration will look like this.
**keyvault.tf**
```
module "keyvault" {
source = "../../modules/keyvault"
name = var.name
region = var.region
owner = var.owner
environment = var.environment
tenant_id = var.tenant_id
principal_id = "f9bcbe58-8b73-4957-aee2-133dc3e58063"
}
```
# Shutting down and environment
To shutdown and remove an environment completely as to not incur any costs you would need to run a `terraform destroy`.