Uploaded image for project: 'OpenShift Hive'
  1. OpenShift Hive
  2. HIVE-1902

Infra node machineset creation fails to identify BYOVPC subnet

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Duplicate
    • Icon: Normal Normal
    • None
    • None
    • Quality / Stability / Reliability
    • False
    • Hide

      None

      Show
      None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      During investigation of #incident-osd-cluster-all-infra-unready , it was found that a BYOVPC cluster with a single customer-supplied subnet completed installation, but the infra node machineset on the cluster was unable to identify the subnet in which to create the infra nodes.

      The machinepool on hive had the correct configuration and subnetID. However, the machineset on cluster did not propagate the subnet id and was instead trying to identify the subnet via a tag `openshift-poc-w8rdr-private-us-east-1a`, which did not exist on the subnet, instead of the id.

      Manually adjusting the infra machineSet to use the subnetID allowed the nodes to be created and the cluster recovered on its own.

      This appears to have been the issue with a previous cluster provision failure [1] as well, but the cause was not fully understood at the time, it looks like.

      We think this could be resolved by having the machineSet controller better match `.spec.template.spec.providerSpec.value.subnet` in the machineSet with the `.spec.platform.aws.subnets` specified in the machinePool, but obviously defer to you folks who know more about how that works under the covers.

      [1] https://coreos.slack.com/archives/CCX9DB894/p1650655863315549?thread_ts=1650650880.642909&cid=CCX9DB894

      Cluster info:

      ID: 1sehv41od45s3h3c6idjde4maqa246oq
      External ID: 8aa1452c-c8ce-47c2-90dc-48594eaffacf
      Name: openshift-poc
      State: ready
      API URL: https://api.openshift-poc.0biu.p1.openshiftapps.com:6443
      API Listening: internal
      Console URL: https://console-openshift-console.apps.openshift-poc.0biu.p1.openshiftapps.com
      Masters: 3
      Infra: 2
      Computes: 3
      Product: rosa
      Provider: aws
      Version: 4.10.14
      Region: us-east-1
      Multi-az: false
      CCS: true
      Subnet IDs: [subnet-0a55bedf6a34db8f1]
      PrivateLink: true
      STS: true
      Existing VPC: true
      Channel Group: stable
      Cluster Admin: true
      Organization: Richard Fojta
      Creator: richard.fojta@sentinelone.com
      Email: richard.fojta@sentinelone.com
      Created: 2022-05-26T12:07:51Z
      Expiration: 0001-01-01T00:00:00Z
      Shard: https://api.hivep05ue1.pfj0.p1.openshiftapps.com:6443

              abutcher@redhat.com Andrew Butcher
              chcollin Chris Collins
              None
              None
              None
              None
              None
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

                Created:
                Updated:
                Resolved: