fix(ci): correct core-cli auth + S3 backend config for CI runner (iterate)

.forgejo/workflows/terraform-apply.yml

@@ -80,8 +80,6 @@ jobs:
       - name: Terraform Init
         working-directory: terraform
         env:
-          # MinIO S3 backend — use internal DNS (storage ns) instead of external hostname
-          # Runner pod blocked from external DNS; internal DNS is routable
           AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
           AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
           TF_VAR_kubeconfig_path: /tmp/kubeconfig
@@ -89,9 +87,15 @@ jobs:
           TF_SKIP_REGION_VALIDATION: "true"
           TF_SKIP_REQUESTING_ACCOUNT_ID: "true"
         run: |
-          # Override S3 endpoint to use internal cluster DNS
-          export TF_CLI_ARGS_init="-backend-config=endpoints.s3=http://minio.storage.svc.cluster.local:9000"
-          terraform init
+          terraform init \
+            -backend-config="bucket=terraform-state" \
+            -backend-config="key=homelab/terraform.tfstate" \
+            -backend-config="region=us-east-1" \
+            -backend-config="endpoint=http://minio.storage.svc.cluster.local:9000" \
+            -backend-config="access_key=$AWS_ACCESS_KEY_ID" \
+            -backend-config="secret_key=$AWS_SECRET_ACCESS_KEY" \
+            -backend-config="skip_credentials_validation=true" \
+            -backend-config="use_path_style=true"
 
       - name: Pull Terraform State
         working-directory: terraform