Troubleshooting
gke-gcloud-auth-plugin ENOENT: plugin not found
Glauco cannot run gke-gcloud-auth-plugin, aws, kubelogin or another exec plugin.
What you see
The main pane says Can't connect to cluster. The message names the context and the error, for example:
Glauco shows
gke-prod is unreachable. spawn gke-gcloud-auth-plugin ENOENT
ENOENT means Glauco could not find the program. Your kubeconfig asks Glauco to run an auth plugin, such as gke-gcloud-auth-plugin, aws, kubelogin or kubectl-oidc_login, and it is not on the PATH that Glauco sees.
Check that the plugin works
- Open a terminal.
- Find the plugin. Use the name from the error.
command -v gke-gcloud-auth-pluginGet-Command gke-gcloud-auth-pluginIf nothing is found, install the plugin:
- GKE:
gcloud components install gke-gcloud-auth-plugin - EKS: install the AWS CLI version 2, which provides
aws. - AKS:
az aks install-cli, which installskubelogin.
Then run kubectl get namespaces in the terminal. If this fails too, fix your kubeconfig or sign in again first, for example with gcloud auth login or aws sso login.
Fix it on macOS and Linux
When it starts, Glauco runs your login shell once and copies its PATH. So the plugin's folder must be on the PATH of a new login shell, not only in one open terminal.
- Add the plugin's folder to
PATHin your shell profile, for example~/.zshrc,~/.zprofile,~/.bashrcor~/.config/fish/config.fish. - Check it in a new terminal:
$SHELL -ilc 'command -v gke-gcloud-auth-plugin'. - Quit Glauco completely.
- Open Glauco again.
gke-gcloud-auth-plugin runs gcloud, so both folders must be on the PATH. If your shell takes more than 5 seconds to start, Glauco does not wait for it. Then make the profile faster, or put the full path to the plugin in the command field of your kubeconfig.Fix it on Windows
- Open Edit the system environment variables from the Start menu.
- Click Environment Variables….
- Add the plugin's folder to
Pathunder your user variables. - Quit Glauco completely.
- Open Glauco again.
Other connection errors
If Glauco gets no answer in 8 seconds, it says the context isn't responding. Check your VPN or credentials, or pick another cluster from the switcher. Check that your VPN is on and that kubectl works, then click Retry.
Still stuck? Contact us.