如何通过Helm从ConfigMap获取镜像仓库地址构建镜像路径？

The core issue here is that the image field in a Deployment is evaluated before the container starts, so you can't reference container environment variables (which are only available at runtime) to construct the image path. Here's how to solve this directly at the Helm template level:

Use Helm's lookup Function to Query the ConfigMap at Render Time #

Helm's lookup function lets you fetch existing Kubernetes resources from the cluster while rendering templates. This is perfect for your ArgoCD use case, as it lets you pull the registry IP directly from the ConfigMap when generating the Deployment manifest.

Updated Deployment Template #

apiVersion: apps/v1
kind: Deployment
metadata:
  name: {{ template "helm-guestbook.fullname" . }}
spec:
  template:
    metadata:
      labels:
        app: {{ template "helm-guestbook.name" . }}
        release: {{ .Release.Name }}
    spec:
      containers:
      - name: {{ .Chart.Name }}
        {{- if eq .Values.isOnPrem "true" }}
        {{- # Fetch the ConfigMap from the cluster }}
        {{- $registryConfig := lookup "v1" "ConfigMap" .Release.Namespace "docker-registry-config" }}
        {{- # Extract the registry IP from the ConfigMap data }}
        {{- $dockerRegistryIP := index $registryConfig.data "DOCKER_REGISTRY_IP" }}
        image: {{ printf "%s/%s:%s" $dockerRegistryIP .Values.image.repository .Values.image.tag }}
        {{- else }}
        image: "{{ .Values.image.repository }}:{{ .Values.image.tag }}"
        {{- end }}

Critical Notes for ArgoCD Compatibility #

1. Ensure ConfigMap Exists Before Deployment
   The lookup function will fail if the docker-registry-config ConfigMap doesn't exist in the target namespace when Helm renders the template. Fix this with ArgoCD's Sync Waves:

   • Add an annotation to your ConfigMap template to deploy it first:

     metadata:
       annotations:
         argocd.argoproj.io/sync-wave: "0"
     

   • Add a later sync wave to your Deployment to ensure it runs after the ConfigMap is created:

     metadata:
       annotations:
         argocd.argoproj.io/sync-wave: "1"
     

2. Handle First-Time Deployment Edge Case
   If you're deploying the ConfigMap and Deployment for the first time, the initial lookup might return nil (since the ConfigMap hasn't been created yet). Add a default value to avoid template errors:

   {{- $registryConfig := lookup "v1" "ConfigMap" .Release.Namespace "docker-registry-config" | default dict }}
   {{- $dockerRegistryIP := index $registryConfig.data "DOCKER_REGISTRY_IP" | default "fallback-registry.example.com" }}
   

   Replace fallback-registry.example.com with a suitable default if needed.

Alternative: ArgoCD Application-Level Parameter Injection #

If you prefer to handle this at the ArgoCD level, you can configure your ArgoCD Application to pull the registry value from the ConfigMap and pass it as a Helm parameter. However, this requires additional ArgoCD configuration and is less direct than using lookup.

For example, in your ArgoCD Application manifest:

spec:
  source:
    helm:
      parameters:
        - name: dockerRegistryIP
          valueFrom:
            configMapKeyRef:
              name: docker-registry-config
              key: DOCKER_REGISTRY_IP

Then reference {{ .Values.dockerRegistryIP }} in your Helm template. This approach works but requires exposing the parameter in your values.yaml, which you wanted to avoid.

Final Recommendation #

The lookup function method is the cleanest solution for your requirements: it avoids hardcoding, doesn't require user input, and integrates seamlessly with ArgoCD when using sync waves to manage deployment order.

内容的提问来源于stack exchange，提问作者rajiv chodisetti