Uploaded image for project: 'Red Hat OpenStack Services on OpenShift'
  1. Red Hat OpenStack Services on OpenShift
  2. OSPRH-7413

Issues with an edpm-deployment with two nodesets

XMLWordPrintable

    • True
    • Hide

      waiting on reproducer and/or logs

      Show
      waiting on reproducer and/or logs
    • False
    • ?
    • ?
    • ?
    • ?
    • Important

      I am working on the rhoso18+bgp configuration/installation/job. It deploys both computes and networkers.

       

      I used a configuration that worked well (edpm deployment succeeded):

      • deploy networkers-nodeset
      • deploy networkers-deployment
      • deploy computes-nodeset
      • deploy computes-deployment

       

      I wanted to check whether deploying in parallel both nodesets would save some time, so I changed it to [1]:

      • deploy networkers-nodeset
      • deploy computes-nodeset
      • deploy edpm-deployment that includes both nodesets

       
       
      The first issue I found was that the `ssh-known-hosts` hosts service can only be included in one of the nodesets.

      2024-05-31T09:11:59Z ERROR Controllers.OpenStackDataPlaneDeployment OpenStackDeployment error for deployment {"controller": "openstackdataplanedeployment", "controllerGroup": "dataplane.openstack.org", "controllerKind": "OpenStackDataPlaneDeployment", "OpenStackDataPlaneDeployment": {"name":"edpm- deployment","namespace":"openstack"}, "namespace": "openstack", "name": "edpm-deployment", "reconcileID": "8a0b0b10-f186-4b6c-866e-f2503f47304c", "ObjectType": "*v1beta1.OpenStackDataPlaneDeployment", "ObjectNamespace": "openstack", "ObjectName": "edpm-deployment", "error": "global service ssh-known-hosts defined mu ltiple times"}

      I'm not sure if this limitation is correct. Shouldn't if be run on both nodesets?
       
      In any case, I removed it from the computes nodeset and applied the edpm-deployment again.
      Regarding `ssh-known-hosts`, it was run multiple times on both nodesets:

      [zuul@controller-0 architecture]$ oc -n openstack logs ssh-known-hosts-edpm-deployment-dgz8l | tail -7
      PLAY RECAP *********************************************************************
      compute-0                  : ok=10   changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      compute-1                  : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      compute-2                  : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      networker-0                : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      networker-1                : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      networker-2                : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      [zuul@controller-0 architecture]$ oc -n openstack logs ssh-known-hosts-edpm-deployment-dgz8l | tail -7
      Error from server (NotFound): pods "ssh-known-hosts-edpm-deployment-dgz8l" not found
      [zuul@controller-0 architecture]$ oc -n openstack logs ssh-known-hosts-edpm-deployment-8zvvg | tail -7
      PLAY RECAP *********************************************************************
      compute-0                  : ok=10   changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      compute-1                  : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      compute-2                  : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      networker-0                : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      networker-1                : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   
      networker-2                : ok=9    changed=4    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   

       
       
      The problem is that, apparently all the `openstackansibleee` pods finished successfully, but the deployment was not completed:

      NAME              STATUS   MESSAGE
      compute-nodes     False    Setup started
      networker-nodes   False    Setup started
      NAME              NODESETS                              STATUS   MESSAGE
      edpm-deployment   ["networker-nodes","compute-nodes"]   False    Deployment in progress

       
       
      Latest logs from the dataplane-operator pod:

      [zuul@controller-0 architecture]$ oc -n openstack-operators logs dataplane-operator-controller-manager-579ddd7898-gdlsn | tail
      2024-05-31T09:39:27Z    INFO    openstackdataplaneservice-resource      validate update {"name": "run-os"}
      2024-05-31T09:39:27Z    INFO    openstackdataplaneservice-resource      default {"name": "ssh-known-hosts"}
      2024-05-31T09:39:27Z    INFO    openstackdataplaneservice-resource      validate update {"name": "ssh-known-hosts"}
      2024-05-31T09:39:27Z    INFO    Controllers.OpenStackDataPlaneNodeSet   Skipping ensure service since it is not a service on this role  {"controller": "openstackdataplanenodeset", "controllerGroup": "dataplane.openstack.org", "controllerKind": "OpenStackDataPlaneNodeSet", "OpenStackDataPlaneNodeSet": {"name":"networker-nodes","namespace":"openstack"}, "namespace": "openstack", "name": "networker-nodes", "reconcileID": "28e86414-1c86-4964-9874-528225ee5404", "service": "swift"}
      2024-05-31T09:39:27Z    INFO    Controllers.OpenStackDataPlaneNodeSet   Skipping ensure service since it is not a service on this role  {"controller": "openstackdataplanenodeset", "controllerGroup": "dataplane.openstack.org", "controllerKind": "OpenStackDataPlaneNodeSet", "OpenStackDataPlaneNodeSet": {"name":"networker-nodes","namespace":"openstack"}, "namespace": "openstack", "name": "networker-nodes", "reconcileID": "28e86414-1c86-4964-9874-528225ee5404", "service": "telemetry"}
      2024-05-31T09:39:27Z    INFO    Controllers.OpenStackDataPlaneNodeSet   Skipping ensure service since it is not a service on this role  {"controller": "openstackdataplanenodeset", "controllerGroup": "dataplane.openstack.org", "controllerKind": "OpenStackDataPlaneNodeSet", "OpenStackDataPlaneNodeSet": {"name":"networker-nodes","namespace":"openstack"}, "namespace": "openstack", "name": "networker-nodes", "reconcileID": "28e86414-1c86-4964-9874-528225ee5404", "service": "update"}
      2024-05-31T09:39:27Z    INFO    openstackdataplaneservice-resource      default {"name": "validate-network"}
      2024-05-31T09:39:27Z    INFO    openstackdataplaneservice-resource      validate update {"name": "validate-network"}
      2024-05-31T09:39:27Z    INFO    Controllers.OpenStackDataPlaneNodeSet   RoleBinding networker-nodes - updated   {"controller": "openstackdataplanenodeset", "controllerGroup": "dataplane.openstack.org", "controllerKind": "OpenStackDataPlaneNodeSet", "OpenStackDataPlaneNodeSet": {"name":"networker-nodes","namespace":"openstack"}, "namespace": "openstack", "name": "networker-nodes", "reconcileID": "28e86414-1c86-4964-9874-528225ee5404"}
      2024-05-31T09:39:27Z    INFO    Controllers.OpenStackDataPlaneNodeSet   Deployment still running...     {"controller": "openstackdataplanenodeset", "controllerGroup": "dataplane.openstack.org", "controllerKind": "OpenStackDataPlaneNodeSet", "OpenStackDataPlaneNodeSet": {"name":"networker-nodes","namespace":"openstack"}, "namespace": "openstack", "name": "networker-nodes", "reconcileID": "28e86414-1c86-4964-9874-528225ee5404", "instance": {"apiVersion": "dataplane.openstack.org/v1beta1", "kind": "OpenStackDataPlaneNodeSet", "namespace": "openstack", "name": "networker-nodes"}}
      [zuul@controller-0 architecture]$ oc -n openstack-operators rsh dataplane-operator-controller-manager-579ddd7898-gdlsn date
      Fri May 31 10:04:03 UTC 2024

       
       
       [1] https://github.com/openstack-k8s-operators/architecture/blob/54cb4910013fb4a1b0fa9c2e85cfd3efe20d48ee/automation/vars/bgp.yaml#L97

            jpodivin@redhat.com Jiri Podivin
            eolivare Eduardo Olivares Toledo
            rhos-dfg-df
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

              Created:
              Updated:
              Resolved: