Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-1824

Systemd service been deactivated in limited network environment

XMLWordPrintable

    • None
    • Agent Sprint 225, Agent Sprint 226
    • 2
    • Proposed
    • False
    • Hide

      None

      Show
      None

      Description of problem:

      When using the agent based instller to zero-touch provision the cluster. If the network bandwidth is low, and the assisted-service or the assisted-service fails to pull the docker image within the timeout. The create-cluster-and-infraenv, apply-host-config, and start-cluster-installation services will be deactivated due to dependency failed. The process will be blocked, and require enable & start the service manually.

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

      openshift-install 4.11.0
      built from commit 863cd1ea823559116e26de327705ed72ccdede8f
      release image quay.io/openshift-release-dev/ocp-release@sha256:300bce8246cf880e792e106607925de0a404484637627edf5f517375517d54a4
      release architecture amd64
      

      How reproducible:

      Install Openshift with agent based installer with local mirror.

      Steps to Reproduce:

      1.Stop the local registry or limit the network bandwidth to make assisted-service-pod.service or assisted-service.service fails to started within the 90s timeout.
      2.Start the local registry or mannully pull the image on the node0. 3.
      

      Actual results:

      When using the agent based instller to zero-touch pprovision  the cluster. If the network bandwidth is low, and the assisted-service or the assisted-service fails to pull the docker image within the timeout. The create-cluster-and-infraenv, apply-host-config, and start-cluster-installation services will be deactivated due to dependency failed. The process will be blocked, and require enable & start the service manually.

      Expected results:

      Provision start after the assisted-service started.

      Additional info:

      Given:
      assisted-service-pod.service requires assisted-service-db.service assisted-service.service
      assisted-service.service BindsTo=assisted-service-pod.service
      create-cluster-and-infraenv.service Requires=assisted-service.service and PartOf=assisted-service-pod.service
      apply-host-config.service Requires=create-cluster-and-infraenv.service
      start-cluster-installation.service Requires=apply-host-config.service
      Requires= "Configures requirement dependencies on other units. If this unit gets activated, the units listed here will be activated as well. If one of the other units gets deactivated or its activation fails, this unit will be deactivated."When assisted-service-pod.service starts, assisted-service-db.service and assisted-service.service also be started,
      Once assisted-service-pod.service fails to be started, assisted-service.service also fail to be started due to "BindsTo=assisted-service-pod.service".
      Then dependency failed for create-cluster-and-infraenv.service due to Requires=assisted-service.service which activation fails, Therefore it will be deactived.
      Then dependency failed for apply-host-config.service, due to Requires=create-cluster-and-infraenv.service which activation fails, Therefore it will be deactived.
      Then dependency failed for start-cluster-installation.service, due to Requires=apply-host-config.service which activation fails, Therefore it will be deactived.Then assisted-service-pod.service restarts, assisted-service.service and assisted-service-db.service restarts as well, since they are binded to assisted-service-pod.service.
      However, create-cluster-and-infraenv.service apply-host-config.service and start-cluster-installation.service was be deactivated, they requires to be activate mannully.Eventually, assisted-service started and hang with waitting for create infraenv. The provisioning is blocked.

              zabitter Zane Bitter
              eric_yue Eric Yue
              zhenying niu zhenying niu
              Votes:
              0 Vote for this issue
              Watchers:
              10 Start watching this issue

                Created:
                Updated:
                Resolved: