Troubleshooting apiservices Unavailability Due to Failed Discovery
This article is relevant for environments where Cilium CNI is in use.
If the errors or behavior described below occur in an environment where Cilium CNI is not in use, please create a support case.
Challenge
In Kubernetes clusters that use Cilium for CNI (Container Network Interface), Veeam Kasten for Kubernetes aggregated apiservices are unavailable due to failure in discovery. This causes the Kasten UI not to load, or display empty (null) values. This causes crucial Kasten APIs like applications, actions, and restore points, to be unusable as they rely on the aggregated apiservices.
The following error will be seen on the K10 dashboard along with HTTP 500 errors.
the server is currently unable to handle the request.
Checking the state of apiservices would show the following discoveryCheck failure message.
kubectl get apiservices | grep kasten v1alpha1.actions.kio.kasten.io kasten-io/aggregatedapis-svc False (FailedDiscoveryCheck)
Cause
In some Kubernetes distributions, Cilium CNI pods block connectivity to Kasten's aggregated apiservices by default.
Cilium is an Open Source security-focused CNI
Solution
Then, upgrade K10 to run the aggregated API service using a non-default port* and in hostNetwork mode.
*The default port for aggregated-apis service is 10250.
To report a typo on this page, highlight the typo with your mouse and press CTRL + Enter.