Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-17050

Supportability of baremetal IPI clusters installed using Assisted Installer via ACM

XMLWordPrintable

    • Quality / Stability / Reliability
    • False
    • Hide

      None

      Show
      None
    • None
    • Moderate
    • No
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      Description of problem:

      The baremetal OCP clusters installed using Assisted Installer via ACM may end in a configuration which is a mix between IPI and UPI, where the KNI infra is installed (coredns, haproxy and keepalived pods), but actually the nodes are ignited more like it is UPI and are either associated to unmanaged baremetalhosts (master nodes) or not associated with any baremetalhost (worker nodes). (see additional info section)
      
      Our understanding is that the hosts are going to be managed by ACM and the components of the BareMetal Operator are going to be installed with the cluster but actually not being used.
      Since there is not an official way to achieve these results manually, our goal is to clarify the supportability of this solution, our opinion is that since this is a widely used and consolidated way of deploying the clusters, it should be considered as supported by Red Hat.
      
      Our questions to engineering:
      - can this installation type raise any supportability issues, especially on charge of the BareMetal Operator components?
      - can the BareMetal Operator be safely deemed inactive in these kind of installations?

      Version-Release number of selected component (if applicable):

       

      How reproducible:

      by installing a baremetal cluster using Assisted Installer via ACM.

      Steps to Reproduce:

      Install a baremetal cluster using Assisted Installer via ACM
      

      Actual results:

       

      Expected results:

       

      Additional info:

      We had some discussions with both AI and metal engineering that we are reporting here for completeness:
      
      Discussion thread in the #forum-ocp-assisted-installer:
      https://redhat-internal.slack.com/archives/CUPJTHQ5P/p1690365152338019
      
      Discussion thread in the #forum-ocp-metal-platform:
      https://redhat-internal.slack.com/archives/CFP6ST0A3/p1690295392066059
      
      -----------
      
      Sample of resulting install-config: (oc get cm -n kube-system cluster-config-v1 -o yaml):
      ~~~
      data:
        install-config: |
          apiVersion: v1
          baseDomain: ocp4.example.com
          compute:
          - architecture: amd64
            hyperthreading: Enabled
            name: worker
            platform: {}
            replicas: 0
          controlPlane:
            architecture: amd64
            hyperthreading: Enabled
            name: master
            platform: {}
            replicas: 3
          metadata:
            creationTimestamp: null
            name: ocp-app
          networking:
            clusterNetwork:
            - cidr: 1.2.0.0/16
              hostPrefix: 23
            machineNetwork:
            - cidr: 10.0.0.128/26
            networkType: OpenShiftSDN
            serviceNetwork:
            - 1.12.0.0/16
          platform:
            baremetal:
              apiVIP: 10.0.0.189
              externalBridge: baremetal
              hosts:
              - bmc:
                  address: ""
                  disableCertificateVerification: false
                  password: ""
                  username: ""
                bootMACAddress: 00:25:b5:aa:bb:cc
                bootMode: legacy
                hardwareProfile: unknown
                name: master-0.ocp4.example.com
                role: master
              - bmc:
                  address: ""
                  disableCertificateVerification: false
                  password: ""
                  username: ""
                bootMACAddress: 00:25:b5:aa:bb:cd
                bootMode: legacy
                hardwareProfile: unknown
                name: master-1.ocp4.example.com
                role: master
              - bmc:
                  address: ""
                  disableCertificateVerification: false
                  password: ""
                  username: ""
                bootMACAddress: 00:25:b5:aa:bb:ce
                bootMode: legacy
                hardwareProfile: unknown
                name: master-2.ocp4.example.com
                role: master
              ingressVIP: 10.0.0.190
              libvirtURI: qemu:///system
              provisioningNetwork: Disabled
              provisioningNetworkInterface: ""
          publish: External
          pullSecret: ""
      ~~~
      
      Sample of bmh resources:
      ~~~
      NAME                        STATUS       STATE       CONSUMER                 BMC   HARDWARE_PROFILE   ONLINE   ERROR   AGE
      master-0.ocp4.example.com   discovered   unmanaged   ocp-app-zsmfk-master-0                            true             270d
      master-1.ocp4.example.com   discovered   unmanaged   ocp-app-zsmfk-master-1                            true             270d
      master-2.ocp4.example.com   discovered   unmanaged   ocp-app-zsmfk-master-2                            true             270d
      ~~~
      
      Sample of machine resources:
      ~~~
      NAME                     PHASE     TYPE   REGION   ZONE   AGE    NODE                        PROVIDERID                                                                                              STATE
      ocp-app-zsmfk-master-0   Running                          270d   master-0.ocp4.example.com   baremetalhost:///openshift-machine-api/master-0.ocp4.example.com/b91c0424-7404-4d95-ae53-614167a86ded   unmanaged
      ocp-app-zsmfk-master-1   Running                          270d   master-1.ocp4.example.com   baremetalhost:///openshift-machine-api/master-1.ocp4.example.com/f58cecd5-acc6-40f3-a08a-58ee14b95459   unmanaged
      ocp-app-zsmfk-master-2   Running                          270d   master-2.ocp4.example.com   baremetalhost:///openshift-machine-api/master-2.ocp4.example.com/9aeae70a-456e-4ee4-96d8-454ffb9bff36   unmanaged
      ~~~
      

              rhn-engineering-dtantsur Dmitry Tantsur
              fcristin1@redhat.com Francesco Cristini
              None
              None
              Pedro Jose Amoedo Martinez Pedro Jose Amoedo Martinez
              None
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

                Created:
                Updated:
                Resolved: