Openshift集群创建DeploymentConfig后无法生成Deployer Pod求助

Got it, let's walk through the key troubleshooting steps for this issue where your DeploymentConfig's Replication Controller fails immediately with "unable to create deployer pod" and only shows a timeout event. Here's what to check step by step:

• Verify Deployer Service Account Permissions
  OpenShift uses a dedicated deployer pod (for Rolling strategy deployments) that runs under the namespace's default service account by default. If this SA lacks the necessary permissions to create pods or manage deployments, the deployer pod creation will fail silently (or only show a timeout).

  • Check the default SA's permissions:

    oc describe sa default -n example-test
    oc get rolebindings,clusterrolebindings -n example-test | grep default
    

  • Ensure the SA has at least edit permissions in the namespace or the specific system:deployer cluster role assigned.

• Inspect Cluster-Level Controller and API Server Logs
  The timeout event doesn't give details, so cluster component logs will likely reveal the root cause (like permission denials, resource shortages, or API errors):

  • Check the OpenShift Controller Manager logs:

    oc logs -n openshift-controller-manager -l app=controller-manager
    

  • Check the Kubernetes API Server logs:

    oc logs -n openshift-kube-apiserver -l app=kube-apiserver
    

  Search for entries related to example-1 or "deployer pod" to find specific error messages (e.g., Forbidden errors, Insufficient resources).

• Check Namespace Resource Quotas and Limits
  If your example-test namespace has ResourceQuotas or LimitRanges configured, hitting resource limits can block deployer pod creation without clear event logs:

  • List and describe resource quotas:

    oc get resourcequota -n example-test
    oc describe resourcequota <quota-name> -n example-test
    

  • Check LimitRanges for default container resource limits:

    oc get limitrange -n example-test
    oc describe limitrange <limit-name> -n example-test
    

  Ensure there's enough CPU, memory, or pod quota available to create the deployer pod.

• Test Deployer Image Pull and Manual Pod Creation
  The deployer pod uses the openshift/deployer image (or a version-specific tag depending on your OpenShift release). If this image can't be pulled, the pod creation will fail:

  • Try manually running a test deployer pod to see if it works:

    oc run test-deployer --image=openshift/deployer -n example-test
    

  • Check if the pod gets created, or if you see errors like ImagePullBackOff or ErrImagePull. If so, verify your image registry access and ImagePullSecrets for the namespace:

    oc get secrets -n example-test | grep pull
    oc describe sa default -n example-test | grep ImagePullSecrets
    

• Check Etcd Health and Resource Usage
  A degraded etcd cluster (e.g., high disk usage, latency, or failed members) can cause API requests to time out, leading to deployer pod creation failures:

  • Check etcd pod status:

    oc get pods -n openshift-etcd
    

  • Inspect etcd logs for errors or performance warnings:

    oc logs -n openshift-etcd -l app=etcd
    

  Also, verify etcd has enough disk space and isn't under heavy load.

• Validate DeploymentConfig Strategy Configuration
  Double-check your DeploymentConfig's strategy settings to ensure there's no misconfiguration that's blocking deployer pod creation:

  • Get the full DC YAML to inspect the strategy section:

    oc get dc example -o yaml -n example-test
    

  Ensure the strategy.type is Rolling (which uses a deployer pod) and that there are no invalid parameters in the strategy configuration.

If none of these steps uncover the issue, sharing the full YAML output of your DeploymentConfig and relevant cluster log snippets would help narrow things down further.

内容的提问来源于stack exchange，提问作者Hound