创建Deployment时extensions/v1beta1版本正则验证失败求助

Hey there, let's break down exactly what's causing that frustrating regex error and get your Deployment up and running.

1. Root Cause of the Error #

That regex validation failure (quantities must match the regular expression '^([+-]?[0-9.]+)([eEinumkKMGTP]*[-+]?[0-9]*)$') is pointing to a simple but critical issue: your resource limit values aren't in a format Kubernetes understands.

Looking at your YAML, the resources.limits section uses placeholder variables like $SPARK_MASTER_MEMORY and $SPARK_MASTER_CPU—Kubernetes can't parse these as valid resource quantities. The mention of extensions/v1beta1 in the error is a red herring here; the real problem is unprocessed variables in your resource definitions.

2. Step-by-Step Fixes #

Replace Placeholder Variables with Valid Resource Values #

Kubernetes requires specific formats for CPU and memory quantities:

• Memory: Use units like Mi (mebibytes), Gi (gibibytes), M (megabytes), or G (gigabytes). Example: 1Gi, 512Mi
• CPU: Use whole numbers (for full cores) or m for millicores (1 core = 1000m). Example: 1, 750m

Update your resources block with concrete values:

resources:
  limits:
    memory: '1Gi'  # Replace with your actual master memory requirement
    cpu: '500m'    # Replace with your actual master CPU requirement

Optional (But Recommended): Update the API Version #

Your Kubectl client is v1.6.4, while the server is v1.5.7. In Kubernetes 1.6+, Deployment resources moved from extensions/v1beta1 to apps/v1beta1. While extensions/v1beta1 still works on v1.5, switching to apps/v1beta1 will make your YAML more future-compatible:

kind: Deployment
apiVersion: apps/v1beta1  # Updated from extensions/v1beta1
metadata:
  name: sf-spark-master
# ... rest of your YAML remains the same

Validate Variable Substitution (If Using CI/CD) #

If those $ variables are supposed to be replaced by a CI/CD pipeline (like Jenkins or GitLab CI), double-check that the pipeline is correctly substituting them before applying the YAML. You can test locally with a dry-run to catch issues early:

kubectl --kubeconfig=/var/go/.kube/mcc-pp-config --context=sam-mcc2-pp --namespace=sam-mcc2-pp apply -f k8s-config-sam-mcc2-pp/sf-spark-worker-deployment.yaml --record --dry-run

3. Final Verification #

Once you've updated the resource values (and optionally the API version), run your original apply command again—it should succeed without the regex error.

内容的提问来源于stack exchange，提问作者Ritabrata