Move ClusterResourceQuota to CRD

[mfojtik]

Goal of this change is to move ClusterResourceQuota from OpenShift API server to CRD so the admission plugins that run in kubernetes apiserver do not have dependency on OpenShift API resource. In case the OpenShift API server have downtime that means every request to Kubernetes API server is delayed.

This PR require openshift/cluster-config-operator#25 to land first, otherwise CRQ is broken.
This will also disable etcd storage for openshift clusterresourcequota so we don't accidentally write to a wrong bucket (and we can actually prove this works).

TODO:

• Fix TestIntegration/TestClusterQuota
• Fix TestIntegration/TestEtcd3StoragePath
• Fix test-cmd (need to create CRD)
• Fix appliedresourcequota

/cc @sttts
/cc @deads2k
/cc @enj
/cc @derekwaynecarr

[mfojtik]

/retest

[mfojtik]

/retest

[mfojtik]

/retest

[mfojtik]

/retest

[mfojtik]

unit test is flake
cmd looks like {\"message\":\"rolebindings.rbac.authorization.k8s.io \\\"system:image-pullers\\\" is forbidden: caches not synchronized\"}]} when creating project request
integration gives me python error

EDIT: cmd failed because the CRD was not installed and that caused the new project to fail create because cluster resource quota admission has not synced informer...

[mfojtik]

@deads2k test-cmd passed ! \o/

[mfojtik]

/retest

unit test flake

[mfojtik]

OK, looks like I got 1 green test-cmd pass but it is flaking again. Give the integration test failing on not being able to observe change in status for 30s, I suspect this might be related to either reconciliation controller being slow on updating status or something is messing up the status in storage?

[mfojtik]

/retest

[mfojtik]

/retest

[deads2k]

what an adventure.

/lgtm

[openshift-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: deads2k, mfojtik, soltysh

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:

• hack/OWNERS [deads2k,mfojtik]
• pkg/OWNERS [deads2k,mfojtik]
• test/integration/OWNERS [deads2k,mfojtik,soltysh]
• test/util/OWNERS [deads2k,mfojtik,soltysh]
• vendor/k8s.io/kubernetes/pkg/master/OWNERS [deads2k,mfojtik]
• vendor/k8s.io/kubernetes/staging/src/k8s.io/kube-aggregator/OWNERS [deads2k,mfojtik]

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

[mfojtik]

/retest

[sjenning]

https://bugzilla.redhat.com/show_bug.cgi?id=1698538 for TestNewManagerImplStartProbeMode test flake

[mfojtik]

/retest

[deads2k]

@mfojtik might be a failure to build

[mfojtik]

@deads2k
E0410 14:57:09.710110 13362 plugin_watcher.go:120] error dial failed at socket /tmp/volume/device_plugin236288530/device-plugin.sock, err: failed to dial socket /tmp/volume/device_plugin236288530/device-plugin.sock, err: context deadline exceeded when handling create event: "/tmp/volume/device_plugin236288530/device-plugin.sock": CREATE

looks more like an infra thing, it happens a lot in unit test job.

[mfojtik]

== RUN TestEventCreated E0410 15:15:36.490666 27410 horizontal.go:212] failed to compute desired number of replicas based on listed metrics for ReplicationController/test-namespace/test-rc: failed to get cpu utilization: no pods returned by selector while calculating replica count

/retest