Skip to main content
Version: 1.17 (current)

Troubleshooting database cluster issues

If your database cluster isn’t behaving as expected, start with these checks:

  1. Check cluster status.

    kubectl describe databasecluster <cluster-name> -n <your namespace>
  2. Check the operator logs.

    kubectl logs -n <operator-namespace> deployment/everest-operator-controller-manager

The table below summarizes frequent issues, their possible causes, and how to resolve them.

IssuePossible CauseHow to Check / FixResolution
Cluster stuck in creating or initializingInsufficient CPU, memory, or nodes available- Inspect pod events:
kubectl describe pod <pod-name> -n <namespace>
kubectl describe pod -l app.kubernetes.io/instance=<db-cluster-name> -n <namespace>

- Check node resource allocation:
`kubectl describe nodes
grep -A5 "Allocated resources"<br /><br />- Monitor live usage:<br />kubectl top nodes<br />kubectl top pods -A`
Cluster in error stateMisconfigured storage class or insufficient resources- Check available storage classes:
kubectl get storageclass

- Inspect pod events:
kubectl describe pod <pod-name> -n <namespace>
kubectl describe pod -l app.kubernetes.io/instance=<db-cluster-name> -n <namespace>

- Check node resource allocation:
kubectl describe nodes
Update the storage class in the cluster spec or request more resources.
Backup or restore failsInvalid backupStorageName or missing/incorrect S3 credentials- Inspect backup:
kubectl describe dbbackup <backup-name> -n <namespace>
- Verify secrets:
kubectl get secrets -n <namespace>
Fix the BackupStorage configuration and update S3 credentials secret.
Monitoring not workingMissing or invalid PMM credentials- Check monitoring config:
kubectl get monitoringconfig -n <namespace>
kubectl describe monitoringconfig <name> -n <namespace>
Recreate the PMM secret with valid base64-encoded credentials.
Proxy not startingWrong proxy type or insufficient resources- Check proxy pod status:
kubectl get pods -n <namespace>
kubectl describe pod <proxy-pod> -n <namespace>
Correct the proxy type in spec and/or increase CPU/memory requests.