Unactionable Errors when running a very slimmed version of this helm chart
Summary
When using a slimmed down version of the chart, essentially the only item enabled is the Container Registry, this removes the central management of certificates. As this is desired, we have a negative impact when the certificates
init container runs and leaves some error messages that one could ignore as they are non actionable. For persons monitoring error logs spawning from their Kubernetes clusters closely, this may be a little misleading and leads to unnecessary churn as the Engineer investigates what has happened. Example errors:
rm: can't remove '/etc/ssl/certs/*': No such file or directory
and
WARNING: ca-certificates.crt does not contain exactly one certificate or CRL: skipping
Steps to reproduce
Spin up the helm chart ensuring ONLY the Container Registry is enabled.
Current behavior
Errors coming from the certificates
container that aren't actionable.
Expected behavior
No errors output. Whether that be because the certificates
init container is not started, or because the script this container uses handles this use case.
Versions
- Chart: 2.1.4
- Platform:
- Cloud: GKE
- Kubernetes: (
kubectl version
)- Client: 1.13.7
- Server: 1.13.7-gke.8
- Helm: (
helm version
)- Client: 2.14.2
- Server: tillerless