Add status tracking with loading spinners, clean up logging

[BenTheElder]

• Adds a --loglevel flag for setting the logrus log level
• moves the default log level to warn
• adds a nice loading spinner system that tracks phases of creating the cluster for friendlier output while waiting for the cluster to boot

TODO (followup):

• log kubeadm output at trace level (?) after upgrading logrus, or log it to a debug log file
• give helpful output when kubeadm init fails
• add preflight checks for docker, previous clusters, etc. add environment diagnostics #39
• possibly just replace node.Run with RunQ

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: BenTheElder

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [BenTheElder]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[BenTheElder]

cc @Katharine @munnerz

This gives shiny output like:

$ kind create cluster --name=asdf
Creating cluster 'kind-asdf' ...
 ✓ Ensuring node image (kindest/node:v1.11.3@sha256:855562266d5f647b5770a91c76a0579ed6487eb5bf1dfe908705dad285525483) 🖼
 ✓ [kind-asdf-control-plane] Creating node container 📦
 ✓ [kind-asdf-control-plane] Fixing mounts 🗻
 ✓ [kind-asdf-control-plane] Starting systemd 🖥
 ✓ [kind-asdf-control-plane] Waiting for docker to be ready 🐋
 ✓ [kind-asdf-control-plane] Starting Kubernetes (this may take a minute) ☸
Cluster creation complete. You can now use the cluster with:

export KUBECONFIG="/Users/bentheelder/.kube/kind-config-asdf"
kubectl cluster-info

[BenTheElder]

trying to record this with asciinema is amusing https://asciinema.org/a/kNsHMXcXpRAj4z0lUqM8l4CRV

[BenTheElder]

2x speed gif:

[BenTheElder]

Rendered: https://github.com/BenTheElder/kind/tree/ux-improvements

[neolit123]

nice 👍

[munnerz]

This is awesome 😄

My main comment on this, is making sure it's still possible/easy to debug when failures do happen (or if I just like to see lots of log output 😄)

[munnerz]

It'd be awesome if we can make --loglevel=debug also print all of these messages, for debug purposes.

From what I can tell, with this patch there will no longer be any way to debug failures.

Also, if a particular step fails can we make it print out the stdout/stderr of that failed command?

[BenTheElder]

I was thinking maybe --loglevel=trace should do this after we upgrade logrus, but debug actually sounds good.

Definitely we should print out failed command output. Especially for kubeadm.

[BenTheElder]

I think what we really want is:

• when critical things fail, inform the user and tell them what they can do about it
• if the log level is cranked high enough, log even "silent" command output on exit

[BenTheElder]

Lots of log output will still mostly work with this, you can set --loglevel across all commands now.
FWIW, I've done quite a bit of kind debugging now, and the output from the commands we call has not been helpful. I'll be PRing a kind command for obtaining useful debugging info in a follow up. #69 (comment)

[munnerz]

/lgtm