Troubleshoot the local trial
When an install step fails, the installer shows a red row, such as docker not answering, with the fix below it. This
page lists those rows, why they happen, and how to check the fix worked.
Try this first
-
Run the installer again. It continues where it stopped and keeps finished downloads.
-
See where the install stands:
testkube status local -
If your shell says
testkube: command not found, use the full path,~/.testkube/bin/testkube, or add~/.testkube/binto yourPATH. -
If the last step failed, the installer saved a diagnostics file in
~/.testkube/logs/, namedinstall-<date>-<time>.txt.
Docker
docker not found
Docker is not installed, or not on your PATH. Install Docker, start
it, then run the installer again.
Check: docker info prints details about the server.
docker not answering, or not reachable
Docker is installed but not running, or still starting. Start Docker, wait until it is ready, then run the installer again.
Check: docker info prints details about the server, not an error.
docker permission denied
On Linux, your user is not allowed to use Docker yet. Add it to the docker group:
sudo usermod -aG docker $USER
newgrp docker
Check: docker info works without sudo. Then run the installer again.
docker not found in WSL
On Windows, Docker Desktop must share Docker with your WSL distribution. In Docker Desktop, open Settings > Resources > WSL integration, turn on your distribution, then run the installer again.
docker not supported yet
The installer needs Docker. Podman is not supported yet. Install Docker and run the installer again.
Not enough CPU or memory
The cpu or memory row shows a warning, or the last step fails with can't start: Docker doesn't have enough memory.
Testkube needs at least 4 CPUs and 6 GB of memory for Docker.
- Docker Desktop: open Settings > Resources and raise CPUs and Memory.
- Colima:
colima stop && colima start --cpu 4 --memory 6. - Docker on Linux: Docker uses the machine's own CPU and memory. Free some up, or use a bigger machine.
Then run the installer again.
inotify too low
On Linux and WSL, low file-watch limits can make Testkube crash at random. Raise them:
sudo sysctl fs.inotify.max_user_instances=512 fs.inotify.max_user_watches=524288
To keep them after a restart, add the same two settings to /etc/sysctl.conf.
Network
network can't reach a site
The installer downloads tools, charts, and images from the sites it lists. If you use a proxy, set HTTPS_PROXY in your
shell and in the Docker proxy settings, then run the installer again. A firewall must allow those sites.
Docker Hub's download limit was reached
The last step fails with can't download its image and mentions the Docker Hub limit. Docker Hub limits downloads per
network for anonymous users. Finished downloads are kept. Wait a few hours or switch networks, then run the installer
again. See Docker Hub usage limits.
kind, helm, or kubectl could not download, or the download was damaged
Check your network, proxy, or firewall, then run the installer again. If the download is damaged every time, a proxy may be changing downloads.
Cluster
cluster port taken
Another program started using one of Testkube's ports while the cluster was being created. Close that program, then run the installer again. On a first install, the installer moves to free ports by itself.
cluster name already taken
A kind cluster named testkube exists that this installer didn't create. If you don't need it, delete it, then run the
installer again:
~/.testkube/bin/kind delete cluster --name testkube
cluster could not start
The cluster container exists but does not start. If a port is in use, close that program and run the installer again. Otherwise, remove the cluster and run the installer again:
testkube uninstall local
Testkube
testkube gave up after a few minutes
A Testkube service did not become ready. The installer names the service and the reason below the row:
- still downloading its image: your connection is slow. Finished downloads are kept. Run the installer again.
- can't download its image: check your network, proxy, or firewall, then run the installer again.
- can't start: Docker doesn't have enough memory: see Not enough CPU or memory.
- keeps crashing or failed: this looks like a Testkube problem. Run the installer again, and if it fails again, get help.
testkube timed out after 15 minutes
The install took too long, usually because downloads are slow. Finished downloads are kept. Run the installer again.
testkube old data can't be opened, or saved passwords can't be read
~/.testkube/data holds data from an earlier install, but its passwords are lost or damaged. Restore
~/.testkube/secrets.json from a backup, or start over.
License
license cannot reach license.testkube.io
Testkube checks your license at license.testkube.io and cannot run without it. Check your network, proxy, or firewall,
then run the installer again.
license not valid, or expired
Copy the key again from your trial email. If it has expired, sign up for a new trial.
license no key given
Without a terminal to type the key in, the installer needs it up front:
curl -fsSL https://trial.testkube.io | TESTKUBE_LICENSE=<key> bash
Start over
To remove the trial and everything it stored, then install from scratch:
testkube uninstall local --delete-data
curl -fsSL https://trial.testkube.io | TESTKUBE_LICENSE=<key> bash
--delete-data deletes your test results and settings. You cannot undo this.
Get help
If the install still fails, contact us and attach the diagnostics file from
~/.testkube/logs/.