Adjust Workspace.Next model classes to latest updates in the Workspace.Next vision

[garagatyi]

What does this PR do?

This PR rework Workspace.Next model classes in accordance with #9964 .
Also, it replaces kubsrv service with Apache web server that hosts Workspace.Next Feature/ChePlugin files. They are stored in YAML format instead of JSON (comparing to kubsrv).
Also added support of Che Server protocol and path, so it is possible to run Classic GWT IDE in Workspace Next now.
PR contains a dockerfile for building marketplace placeholder service and it is possible to replace content with other files, so developers can play with it during development.
Additional unit tests were also added.

What issues does this PR fix or reference?

#9964

Release Notes

Docs PR

[garagatyi]

ci-test

[codenvy-ci]

ci-test build report:
Build details
Test report
selenium tests report data
docker image: eclipseche/che-server:10246
https://github.com/orgs/eclipse/teams/eclipse-che-qa please check this report.

[garagatyi]

@sleshchenko @l0rd @ibuziuk @skabashnyuk Please, review the PR

[sleshchenko]

In my opinion, testChePlugin sounds like it should test ChePlugin, but in facts, it creates test ChePlugin object. It does not make much sense since it is test sources but maybe just createChePlugin would be clearer.

[garagatyi]

ok, will change it

[garagatyi]

done

[sleshchenko]

Should be feature.getSpec().getServices() method renamed to feature.getSpec().getFeature()?

[garagatyi]

I'm not going to update Feature related model as we are going to remove this entity from the model according to the discussion in workspace next walking skeleton #10123

[sleshchenko]

cheServiceReference -> chePluginReference
Please make sure that all usages of service is replaced with plugin

[garagatyi]

I'm not going to update Feature related model as we are going to remove this entity from the model according to the discussion in workspace next walking skeleton #10123

[sleshchenko]

OK, but it's not about feature model changing, but about variable name =)

[garagatyi]

done

[l0rd]

LGTM. I have 2 comments that we can discuss and address in separate issues:

• you are including the switch to apache for the marketplace (that was not part of the model adujstment) but I think it should live in a separate github repository
• you are basing everything on the kubectl deployment type instead of using helm whereas I think helm should be the default

[ibuziuk]

I wanted to test the PR on minishift by deploying che-server with feature-api in the same namespace and creating workspace via:

`curl 'http://che-mini-che.192.168.42.3.nip.io/api/workstribute=features:org.eclipse.che.che-theia/0.0.3' -H 'Content-Type: application/json;charset=UTF-8' --data-binary '{"environments":{"default":{"recipe":{"contentType":"application/x-yaml","type":"openshift","content":"---\nkind: List\nitems:\n-\n  apiVersion: v1\n  kind: Pod\n  metadata:\n    name: wsnextdemo\n  spec:\n    containers:\n      -\n        image: eclipse/ubuntu_jdk8\n        name: app"},"machines":{"wsnextdemo/app":{"attributes":{}}}}},"defaultEnv":"default","name":"workspace-next"}'

Workspace is started correctly but for some reason I do not have Theia server with URL to IDE:

{
  "links": {
    "self": "http://che-mini-che.192.168.42.3.nip.io/api/workspace/workspacekjyngxxri64tbduv",
    "ide": "http://che-mini-che.192.168.42.3.nip.io/che/workspace-next",
    "environment/outputChannel": "ws://che-mini-che.192.168.42.3.nip.io/api/websocket",
    "environment/statusChannel": "ws://che-mini-che.192.168.42.3.nip.io/api/websocket"
  },
  "attributes": {
    "features": "org.eclipse.che.che-theia/0.0.3",
    "updated": "1530720068026",
    "created": "1530719995093"
  },
  "namespace": "che",
  "id": "workspacekjyngxxri64tbduv",
  "temporary": false,
  "status": "RUNNING",
  "runtime": {
    "machines": {
      "wsnextdemo/app": {
        "attributes": {
          "memoryLimitBytes": "1073741824"
        },
        "servers": {},
        "status": "RUNNING"
      }
    },
    "warnings": [],
    "activeEnv": "default",
    "machineToken": "",
    "links": []
  },
  "config": {
    "defaultEnv": "default",
    "environments": {
      "default": {
        "machines": {
          "wsnextdemo/app": {
            "attributes": {},
            "servers": {},
            "volumes": {},
            "installers": [],
            "env": {
              "CHE_MACHINE_NAME": "wsnextdemo/app"
            }
          }
        },
        "recipe": {
          "type": "openshift",
          "content": "---\nkind: List\nitems:\n-\n  apiVersion: v1\n  kind: Pod\n  metadata:\n    name: wsnextdemo\n  spec:\n    containers:\n      -\n        image: eclipse/ubuntu_jdk8\n        name: app",
          "contentType": "application/x-yaml"
        }
      }
    },
    "name": "workspace-next",
    "projects": [],
    "commands": [],
    "links": []
  }
}

Any ideas what I'm doing wrong ?
Please, do not consider this comment as smth. that should prevent merging. I just getting acquainted with workspace next and might be doing smth. wrong

[garagatyi]

@l0rd I switched to Apache because didn't want to do model changes for kubsrv service in Golang. I think it is easier to support for the time being. Can you elaborate what exactly you would want to move to another repository?
I used kubectl since it is simpler for me, but yeah, we should support helm too.

[garagatyi]

@ibuziuk I answered in MM channel

[l0rd]

@garagatyi I would move the content of folder deploy/kubernetes/kubectl/wsnext/ to a distinct repository and setup the CI to build the Dockerfile. I will add that as a ws.next skeleton task.