Can't install on CoreOS 1632.3.0

[mbaquer6]

Is this a BUG REPORT or FEATURE REQUEST?:

Uncomment only one, leave it on its own line:

/kind bug

/kind feature

What happened:
Hi,
I was trying to install on CoreOS 1632.3.0 but it fails because the regex can't deal with the "+" of the 1.8+ version.
I removed the script that checks the version to use the "tpr" or "crd" to use always "crd", the supported resource definition in Kubernetes 1.8 but it fails in phase 6:
What you expected to happen:
Finish the installation
How to reproduce it (as minimally and precisely as possible):
Try to install on Kubernetes 1.8+ (CoreOS 1632.3.0)

Anything else we need to know?:
This is the log of a daemonset in phase 6:

• echo '[INFO] successfully created vSphere.conf file at : /host/tmp/vsphere.conf'
• PHASE='[PHASE 6] Update Manifest files and service configuration file'
  [INFO] successfully created vSphere.conf file at : /host/tmp/vsphere.conf
• update_VcpConfigStatus vcp-daementset-5dh7m '[PHASE 6] Update Manifest files and service configuration file' RUNNING ''
• POD_NAME=vcp-daementset-5dh7m
• PHASE='[PHASE 6] Update Manifest files and service configuration file'
• STATUS=RUNNING
• ERROR=
• '[' RUNNING == FAILED ']'
• echo 'apiVersion: "vmware.com/v1"
  kind: VcpStatus
  metadata:
  name: vcp-daementset-5dh7m
  spec:
  phase: "[PHASE' '6]' Update Manifest files and service configuration 'file"
  status: "RUNNING"
  error: ""'
• retry_attempt=1
• kubectl apply -f /tmp/vcp-daementset-5dh7m_daemonset_status_update.yaml
• ls /host//etc/kubernetes/manifests/kube-apiserver.yaml
• '[' 1 -eq 0 ']'
• ls /host//etc/kubernetes/manifests/kube-controller-manager.yaml
• '[' 1 -eq 0 ']'
• ls /host//etc/systemd/system/kubelet.service.d/10-kubeadm.conf
• '[' 1 -eq 0 ']'
• IS_CONFIGURATION_UPDATED=false
• UPDATED_MANIFEST_FILE=/host/tmp/kube-controller-manager.json
• '[' yaml == yaml ']'
• UPDATED_MANIFEST_FILE=/host/tmp/kube-controller-manager.yaml
• '[' -f /host/tmp/kube-controller-manager.yaml ']'
• UPDATED_MANIFEST_FILE=/host/tmp/kube-apiserver.json
• '[' yaml == yaml ']'
• UPDATED_MANIFEST_FILE=/host/tmp/kube-apiserver.yaml
• '[' -f /host/tmp/kube-apiserver.yaml ']'
• '[' -f /host/tmp/vsphere.conf ']'
• cp /host/tmp/vsphere.conf /host//etc/vsphereconf/vsphere.conf
• '[' 0 -ne 0 ']'
• '[' -f /host/tmp/kubelet-service-configuration ']'
• '[' false == false ']'
• ERROR_MSG='No configuration change is observed'
• update_VcpConfigStatus vcp-daementset-5dh7m '[PHASE 6] Update Manifest files and service configuration file' FAILED 'No configuration change is observed'
• POD_NAME=vcp-daementset-5dh7m
• PHASE='[PHASE 6] Update Manifest files and service configuration file'
• STATUS=FAILED
• ERROR='No configuration change is observed'
• '[' FAILED == FAILED ']'
• echo '[ERROR] No configuration change is observed'
  [ERROR] No configuration change is observed
• echo 'apiVersion: "vmware.com/v1"
  kind: VcpStatus
  metadata:
  name: vcp-daementset-5dh7m
  spec:
  phase: "[PHASE' '6]' Update Manifest files and service configuration 'file"
  status: "FAILED"
  error: "No' configuration change is 'observed"'
• retry_attempt=1
• kubectl apply -f /tmp/vcp-daementset-5dh7m_daemonset_status_update.yaml
• exit 29
  [INFO] Done with all tasks. Sleeping Infinity.

Environment:

• Kubernetes version (use kubectl version): 1.9
• Cloud provider or hardware configuration: vSphere
• OS (e.g. from /etc/os-release): CoreOS 1632.3.0
• Kernel (e.g. uname -a): 4.14.19-coreos
• Install tools:
• Others:

Thanks.

[pshahzeb]

related to #429

[venilnoronha]

@mbaquer6 could you please post the exact versions of your kubectl and kubelet? You could use kubectl version and kubelet --version for the same.

[abrarshivani]

@mbaquer6 Can you please let us know whether you used kubeadm to configure your kubernetes cluster?
Also, can you check whether path "/etc/kubernetes/manifests/kube-apiserver.yaml" exists on master?

[venilnoronha]

@mbaquer6 we tried to test the enable-vcp-uxi script end-to-end, and it worked fine.

Following is our setup:

• vSphere v6.5
• ESXi v6.5
• Master VM OS: CoreOS 1576.5.0
• Node VM OS: CoreOS 1576.5.0
• Kubernetes v1.9.4
• Kubeadm v1.9.4

Also, following is the output produced by kubectl get VcpSummary --namespace=vmware -o json.

{
    "apiVersion": "v1",
    "items": [
        {
            "apiVersion": "vmware.com/v1",
            "kind": "VcpSummary",
            "metadata": {
                "annotations": {
                    "kubectl.kubernetes.io/last-applied-configuration": "{\"apiVersion\":\"vmware.com/v1\",\"kind\":\"VcpSummary\",\"metadata\":{\"annotations\":{},\"name\":\"vcpinstallstatus\",\"namespace\":\"\"},\"spec\":{\"nodes_being_configured\":\"0\",\"nodes_failed_to_configure\":\"0\",\"nodes_in_phase1\":\"0\",\"nodes_in_phase2\":\"0\",\"nodes_in_phase3\":\"0\",\"nodes_in_phase4\":\"0\",\"nodes_in_phase5\":\"0\",\"nodes_in_phase6\":\"0\",\"nodes_in_phase7\":\"0\",\"nodes_sucessfully_configured\":\"2\",\"total_number_of_nodes\":\"2\"}}\n"
                },
                "clusterName": "",
                "creationTimestamp": "2018-03-13T20:08:45Z",
                "generation": 0,
                "name": "vcpinstallstatus",
                "namespace": "",
                "resourceVersion": "5741",
                "selfLink": "/apis/vmware.com/v1/vcpinstallstatus",
                "uid": "5567ec52-26fa-11e8-9780-005056a1839f"
            },
            "spec": {
                "nodes_being_configured": "0",
                "nodes_failed_to_configure": "0",
                "nodes_in_phase1": "0",
                "nodes_in_phase2": "0",
                "nodes_in_phase3": "0",
                "nodes_in_phase4": "0",
                "nodes_in_phase5": "0",
                "nodes_in_phase6": "0",
                "nodes_in_phase7": "0",
                "nodes_sucessfully_configured": "2",
                "total_number_of_nodes": "2"
            }
        }
    ],
    "kind": "List",
    "metadata": {
        "resourceVersion": "",
        "selfLink": ""
    }
}

[mbaquer6]

Thanks for all your responses
@venilnoronha I deployed the cluster using Tectonic 1.8.4 using the CLI (Terraform), that installs Kuberentes 1.8.4, Container Linux by CoreOS stable (1632.3.0)
vSphere 6.0
I saw in the guides that the installation for the vsphere cloud provider on Kubernetes 1.9 is different from 1.8 when you do it manually. I tried that too but didn't have the /etc/kubernetes/manifests/kube-apiserver.yaml files, so I edited the kube-api-server running config to add the cloud provider lines and edit the kube-controller-manager too but didn't work either.

Client Version: version.Info{Major:"1", Minor:"9", GitVersion:"v1.9.1", GitCommit:"3a1c9449a956b6026f075fa3134ff92f7d55f812", GitTreeState:"clean", BuildDate:"2018-01-04T20:00:41Z", GoVersion:"go1.9.2", Compiler:"gc", Platform:"darwin/amd64"}
Server Version: version.Info{Major:"1", Minor:"8+", GitVersion:"v1.8.4+coreos.0", GitCommit:"4292f9682595afddbb4f8b1483673449c74f9619", GitTreeState:"clean", BuildDate:"2017-11-21T17:22:25Z", GoVersion:"go1.8.3", Compiler:"gc", Platform:"linux/amd64"}

[mbaquer6]

@abrarshivani I used Tectonic to configure the cluster.
/etc/kubernetes/manifests/kube-apiserver.yaml doesn't exists on the masters, there is just the pod checkpointer.json

[abrarshivani]

@mbaquer6 Where is your kube-api-server running config located?

[mbaquer6]

@abrarshivani I edit with kubectl edit daemonset kube-apiserver -n=kube-system

[abrarshivani]

@mbaquer6 Can you please share checkpointer.json? Also, do you know where are kubernetes manifests located on master?

[mbaquer6]

Looking at the files that Tectonic created, I see all the manifests in my machine but not in the master. I'll try copying them to the masters and try to reapply this tool and see what happens.