Skip to content

KubeStellar has multiple documentation versions to match its multiple releases.
Please make sure you are viewing the docs version which matches the release version of the code you are using!

Some known problems#

Here are some user and/or environment problems that we have seen.

For bugs, see the issues on GitHub and the release notes.

Wrong value stuck in hidden kflex state in kubeconfig#

The symptom is kflex ctx ... commands failing. See Confusion due to hidden state in your kubeconfig.

Kind clusters failing to work#

The symptom is kind cluster(s) that get created but fail to get their job done. See Potential Error with Kubestellar Installation related to Issues with Kind backed by Rancher Desktop.

Authorization fail for Helm fetching chart from ghcr#

The symptom is that attempting to instantiate the core Helm chart gets an authorization failure. See Authorization failure while fetching Helm chart from ghcr.io.

Missing results in a CombinedStatus object#

The symptom is a missing entry in the results of a CombinedStatus object. See Missing results in a CombinedStatus object.

Kind host not configured for more than two clusters#

This can arise when using kind inside a virtual machine (e.g., when using Docker on a Mac). The symptom is either a complaint from KubeStellar setup that sysctl fs.inotify.max_user_watches is only 155693 but must be at least 524288 or setup grinding to a halt. See Kind host not configured for more than two clusters.

Insufficient CPU for your clusters#

This can happen when you are using a docker-in-docker technique. The symptom is that setup stops making progress at some point. See Insufficient CPU for your clusters