Script seems to deliver random results

[B45man]

I'm on Windows 10, latest update. Got the latest version of both VirtualBox and Vagrant installed. The scripts don't work on Windows, so I converted them to powershell:

vagrant up --provider virtualbox

echo '######################## WAITING TILL ALL NODES ARE READY ########################'
sleep 60

echo '######################## INITIALISING K8S RESOURCES ########################'
#chmod 400 certs/id_rsa
vagrant ssh cluster1-master1 -c "sudo su - root -c 'kubectl apply -f /vagrant/k8s/init.yaml'"
vagrant ssh cluster1-master1 -c "mkdir -p .kube ; sudo cp /root/.kube/config ./.kube/config ; sudo chown vagrant:vagrant .kube/config"
sleep 20

echo '######################## ALL DONE ########################'

And

vagrant destroy --force
rm -r .vagrant
rm -r tmp
mkdir tmp
New-Item tmp/.gitignore

However, the script already seems to fail at the first vagrant up --provider virtualbox command. Sometimes installing kubectl seems to fail according to the logs, sometimes the second node can't be instantiated, it's a new random bug every time. The errors are very random. Does this script store state somewhere in a way that can affect subsequent runs?

[wuestkamp]

As of now Windows isn't supported as the init scripts are all written for bash. If you figure this out and make this work with Windows feel free to create a PR, that would be much appreciated :)

Does this script store state somewhere in a way that can affect subsequent runs?

Vagrant iteself stores state in a local .vagrant folder. I suggest you delete this one and try again to spin it up

Make sure to wait till all nodes area actually ready before applying the k8s resources from init.yaml.

[B45man]

My apologies, I appear to be having issues with Vagrant and VirtualBox on Windows. Your scripts are fine, sorry for implying otherwise.

The scripts I posted above are fine. It's the call to vagrant up --provider virtualbox that always seems to fail at different steps. Either steps time out, fail to connect through SSH, packages fail to install, or otherwise. And 1 in 5 times the scripts work as intended. I am assuming it's just Vagrants low quality on Windows platforms, as I have yet to have 100% success with any Vagrant file. I'm going to try Vagrant on Ubuntu instead to see if my luck changes.

[garyttt]

The scripts do work on Windows10 home edition 20H2 update with latest version of Vagrant (2.2.14) and VirtualBox (6.1), instead of installing GIT for Windows for running the "./up.sh" script, you may install RedHat CYGWIN64 to run "./up.sh".

Many kubernetes self-learners may be using Docker Desktop with built-in kubernetes engine enabled, i.e. single-node cluster.

For Windows 10 environment, I have found Docker Desktop and Vagrant-VirtualBox created k8s cluster are mutually exclusive.

If Docker Desktop is running, the "./up.sh' will fail at 'SSH auth method: private key' stage with the following message.

cluster1-master1: SSH auth method: private key

Timed out while waiting for the machine to boot. This means that Vagrant was unable to communicate with the guest machine within the configured ("config.vm.boot_timeout" value) time period.

The solution is to stop Docker Desktop, in Windows 10 start run Command Prompt as administrator and then reboot PC bcdedit /set hypervisorlaunchtype off

This will fix the private key setup issue.

On the other hand, if Docker Desktop is to be run:

The solution is to stop VirtualBox VMs, in Windows 10 start run Command Prompt as administrator and then reboot PC bcdedit /set hypervisorlaunchtype auto