Frequent Issues & Troubleshooting This page lists common problems CIRRUS users run into , and quick pointers to the right fix or reference guide. Each issue links back to the detailed documentation so we don’t duplicate entire procedures.
Application Deployment Symptom Quick Fix Docker build fails in CI or locally Check your Dockerfile base image and dependencies. See the container build tips in create containers . Argo CD shows Sync Failed Run helm lint locally and verify required values. See Helm chart deployment in adding applications . Template files placed outside templates/ directory Move all Kubernetes manifests (Deployment, Service, Ingress, etc.) into the chart’s templates/ folder - Argo CD and Helm only render files inside this directory.
Authentication & Access Symptom Quick Fix Cannot docker login to Harbor Regenerate your CLI secret and retry. Full steps in image management . JupyterHub server won’t spawn Verify CIT credentials are valid and you’re on VPN. See jupyterhub .
Symptom Quick Fix Image pulls are very slow Ensure you’re pulling from hub.k8s.ucar.edu, not Docker Hub. Pod OOM-killed Increase resources.limits.memory in your Helm chart (see adding applications ).
GitHub Actions Symptom Quick Fix Workflow stuck in Queued Confirm your runner scale set is healthy; see runner scale sets . “Access denied” pushing image in workflow Make sure the PAT stored in OpenBao still has write:packages scope.
When in Doubt If a problem persists:
Consult the full documentation page linked above. Search Slack/Jira for existing reports. Create a ticket with logs, steps to reproduce, and screenshots. For support levels and response times, see our service level agreements .