Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-65961

BMH stuck in inspecting state with 404 errors inspection error in the ironic-agent logs

XMLWordPrintable

    • Incidents & Support
    • False
    • Hide

      None

      Show
      None
    • 1
    • None
    • None
    • None
    • Metal Platform 280
    • 1
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      Description of problem:

      Two BareMetalHost (BMH) nodes remain stuck in Inspecting.
      The ironic-agent repeatedly fails node lookup with a 404 Not Found error, even though the boot MAC address is included in the reported list. Inspection never completes.

      Sample logs:
       
       
      InsecureRequestWarning: Unverified HTTPS request is being made to host '<ip>'Failed looking up node with addresses '<mac address>,...' at https://<ip>:6385.Error 404: Resource could not be found.
      Two other BMH nodes with identical nmstateconfig structure discover successfully.


      Version-Release number of selected component:

      • ACM Hub: 2.14
      • Ironic + Ironic-Agent images verified as compatible with ACM 2.14

      How reproducible:

      Seen consistently on two affected hosts.
      Not reproduced on the other two hosts in the same environment.


      Steps to Reproduce:

      1. Register multiple BMHs in ACM 2.14 using Assisted Installer.
      1. Allow nodes to start inspection via Ironic Agent.
      1. Observe two nodes staying in Inspecting with repeated lookup failures.

      Actual results:

      • Ironic Agent returns a 404 during node lookup.
      • Hosts never complete hardware inspection.

        Expected results:

      • Ironic Agent should successfully match the host to a node in Ironic.
      • Inspection should complete and BMH should move to Ready.

      Additional info:

      • Issue is customer-blocking and high priority.
      • All four hosts use similar nmstateconfig. Only two fail.
      • Must-gather and journal logs are attached on the case.

              rhn-engineering-dtantsur Dmitry Tantsur
              rhn-support-mlele Mihir Lele
              None
              None
              Steeve Goveas Steeve Goveas
              None
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

                Created:
                Updated: