Chang server:stop behavior and added server:delete

[l0rd]

With this PR the command server:stop:

• calls Che server shutdown endpoint (/api/system/stop?shutdown=true)
• scales down the deployment to zero (as well as keycloak and postgres if needed)

In this PR is also included a new command: server:delete. It rapidly purges all Che resources (deployments, pods, services, routes, cm, pvc etc...) to start from scratch.

Some other changes included in this PR:

• Automatically find out if it's Kubernetes or OpenShift (currently only stop and delete commands)
• server:stop is the first command that supports authenticated Che API (user has to provide the access token as a param, I would like to create a new auth:token command that would help with that)
• Added 2 end2end tests for minikube and minishift: I would like to setup a circleCi job that use them for PR validation
• Added an option to use verbose Listr renderer for all commands because it's better for automated tests

[l0rd]

Fixes #71 and #56

[benoitf]

Hello,

I tried on minikube

server:start then server:stop and now server:start but it's stuck at

  ❯ ✅  Post installation checklist
    ❯ Che pod bootstrap
      ⠦ scheduling
        downloading images

[benoitf]

kubectl get all --namespace che

$ kubectl get all --namespace che
NAME               TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)             AGE
service/che-host   ClusterIP   10.103.228.218   <none>        8080/TCP,8087/TCP   15m

NAME                  READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/che   0/0     0            0           15m

NAME                             DESIRED   CURRENT   READY   AGE
replicaset.apps/che-5959c547b6   0         0         0       15m

[l0rd]

@benoitf you are right. I had not made the changes needed to support this scenario in this PR: server:start never check if che server was already deployed or not.

The only way to restart Che is kubectl scale deployment che --replicas=1. But that doesn't make this PR consistent so I will try to look how much changes are needed to support your scenario. If it needs too much work I would rather do it in another PR.

[benoitf]

ok, I was not able to test on minishift as each time it pulls all images it takes me at least one hour :-)

[azatsarynnyy]

checked on minishift - works well

[l0rd]

I am going to merge this PR since fixing server:start properly doesn't look so easy. I have opened #77 to track that.