Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-3509

[WINC] Windows nodes name not matching hostname in GCP

    XMLWordPrintable

Details

    • Bug
    • Resolution: Done
    • Normal
    • 4.12.0
    • 4.12
    • Windows Containers
    • None
    • Moderate
    • 0
    • WINC - Sprint 228
    • 1
    • False
    • Hide

      None

      Show
      None
    • Done

    Description

      This is a clone of issue OCPBUGS-2028. The following is the description of the original issue:

      Description of problem:

      When getting the cluster nodes information, the Windows nodes identifier does not follow the same pattern as the rest of nodes in the cluster:
      
      $ oc get nodes
      NAME                                                         STATUS   ROLES                  AGE     VERSION
      jfrancoa-0510-sgstz-master-0.c.openshift-qe.internal         Ready    control-plane,master   4h18m   v1.25.0+3ef6ef3
      jfrancoa-0510-sgstz-master-1.c.openshift-qe.internal         Ready    control-plane,master   4h18m   v1.25.0+3ef6ef3
      jfrancoa-0510-sgstz-master-2.c.openshift-qe.internal         Ready    control-plane,master   4h18m   v1.25.0+3ef6ef3
      jfrancoa-0510-sgstz-windows-worker-a-kt2l5                   Ready    worker                 37m     v1.25.0-2492+3ef6ef3cfbd273
      jfrancoa-0510-sgstz-windows-worker-a-zd6l4                   Ready    worker                 35m     v1.25.0-2492+3ef6ef3cfbd273
      jfrancoa-0510-sgstz-worker-a-j5kxg.c.openshift-qe.internal   Ready    worker                 4h      v1.25.0+3ef6ef3
      jfrancoa-0510-sgstz-worker-b-xk4jn.c.openshift-qe.internal   Ready    worker                 4h      v1.25.0+3ef6ef3
      
      While the rest of the Linux nodes id match their hostname, this isn't the case for the Windows workers:
      
      $ oc get nodes  -o="jsonpath={.items[*].status.addresses[?(@.type=='Hostname')].address}"
      jfrancoa-0510-sgstz-master-0.c.openshift-qe.internal 
      jfrancoa-0510-sgstz-master-1.c.openshift-qe.internal 
      jfrancoa-0510-sgstz-master-2.c.openshift-qe.internal 
      jfrancoa-0510-sgstz-windows-worker-a-kt2l5.c.openshift-qe.internal 
      jfrancoa-0510-sgstz-windows-worker-a-zd6l4.c.openshift-qe.internal 
      jfrancoa-0510-sgstz-worker-a-j5kxg.c.openshift-qe.internal 
      jfrancoa-0510-sgstz-worker-b-xk4jn.c.openshift-qe.internal
      
      This pattern is also followed in other cloud providers like AWS in which the Windows nodes also match the Hostname field:
      
      [jfrancoa@localhost wmco]$ oc get nodes
      NAME                                         STATUS   ROLES                  AGE     VERSION
      ip-10-0-128-49.us-east-2.compute.internal    Ready    worker                 3h48m   v1.25.0-2492+3ef6ef3cfbd273
      ip-10-0-130-166.us-east-2.compute.internal   Ready    worker                 3h52m   v1.25.0-2492+3ef6ef3cfbd273
      ip-10-0-133-104.us-east-2.compute.internal   Ready    control-plane,master   4h34m   v1.25.0+3ef6ef3
      ip-10-0-143-248.us-east-2.compute.internal   Ready    worker                 4h28m   v1.25.0+3ef6ef3
      ip-10-0-165-239.us-east-2.compute.internal   Ready    control-plane,master   4h34m   v1.25.0+3ef6ef3
      ip-10-0-186-203.us-east-2.compute.internal   Ready    worker                 4h28m   v1.25.0+3ef6ef3
      ip-10-0-215-140.us-east-2.compute.internal   Ready    control-plane,master   4h34m   v1.25.0+3ef6ef3
      ip-10-0-217-40.us-east-2.compute.internal    Ready    worker                 4h28m   v1.25.0+3ef6ef3
      [jfrancoa@localhost wmco]$  oc get nodes  -o="jsonpath={.items[*].status.addresses[?(@.type=='Hostname')].address}"
      ip-10-0-128-49.us-east-2.compute.internal ip-10-0-130-166.us-east-2.compute.internal ip-10-0-133-104.us-east-2.compute.internal ip-10-0-143-248.us-east-2.compute.internal ip-10-0-165-239.us-east-2.compute.internal ip-10-0-186-203.us-east-2.compute.internal ip-10-0-215-140.us-east-2.compute.internal ip-10-0-217-40.us-east-2.compute.internal
      

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

      [jfrancoa@localhost openshift-tests-private]$ oc version
      Client Version: 4.11.0-0.ci-2022-06-09-065118
      Kustomize Version: v4.5.4
      Server Version: 4.12.0-0.nightly-2022-10-04-081353
      Kubernetes Version: v1.25.0+3ef6ef3
      [jfrancoa@localhost openshift-tests-private]$ oc get cm -n openshift-windows-machine-config-operator 
      NAME                                   DATA   AGE
      kube-root-ca.crt                       1      4h1m
      openshift-service-ca.crt               1      4h1m
      windows-machine-config-operator-lock   0      69m
      windows-services-7.0.0-01499ac         2      4h
      
      

      How reproducible:

      Always
      

      Steps to Reproduce:

      1. Deploy an IPI 4.12 OCP on GCP cluster
      2. Install WMCO 7.0.0
      3. Create couple of windows workers
      4. List the nodes once they join the cluster with: oc get nodes
      

      Actual results:

      Windows nodes do not contain the whol FQDN, while Linux nodes do
      

      Expected results:

      All nodes, Windows nodes included use the whole FQDN as identifier when listing the nodes
      

      Additional info:

      
      

      Attachments

        Activity

          People

            jvaldes@redhat.com Jose Valdes
            openshift-crt-jira-prow OpenShift Prow Bot
            Jose Luis Franco Arza Jose Luis Franco Arza (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: