id: PG-08
category: compute
title: Disabled or down? Two fields
situation: >-
  `openstack compute service list` gives each `nova-compute` two independent fields. Each one is set,
  cleared and interpreted differently.

diagram:
  type: compare
  caption: Two fields, two questions
  columns:
    - kicker: "Status: enabled / disabled"
      title: May the scheduler place instances here?
      tone: ink
      points:
        - Set by an operator, or by the libvirt driver (reason `AUTO:`)
        - Instances already there are not affected
      verdict: A scheduling decision
    - kicker: "State: up / down"
      title: Does Nova believe the service is alive?
      tone: warn
      points:
        - From the heartbeat, or forced by an operator
        - "Once down: deletes become local, evacuate is allowed"
      verdict: A liveness belief
  note: No state stops a running instance. The danger is in what Nova allows once a service is down.

checks:
  - label: Both fields, reason, forced down
    hint: openstack compute service list --host <host> --long
  - label: Who disabled it?
    hint: "reason AUTO: = libvirt connection lost, it re-enables itself"
  - label: Is nova-compute alive on the host?
    hint: unit or container status, end of nova-compute.log
  - label: Evacuations that left this host
    hint: openstack server migration list --host <host> --type evacuation

caution:
  - level: destructive
    text: >-
      Never force down a compute that is not fenced. Nova then allows local deletes and evacuations of
      instances that are alive.
  - level: state-changing
    text: >-
      `--enable` makes the host schedulable at once. Read the reason first: ask whoever set it, and never
      enable over an `AUTO:` one.

lesson: Disabled is a scheduling decision. Down is a belief about liveness.

runbook:
  path: scheduling/compute-service-disabled-or-forced-down.md
  status: lab-verified
  release: "2026.1"

scope: >-
  Written from the Nova 2026.1 source and reproduced on a Kolla-Ansible lab (ML2/OVS, no Cinder). The Forced
  Down column needs compute API microversion 2.11 or later. Nova refuses to set a service back up while
  evacuation records from that host are still "done": nova-compute must be restarted there first, which
  destroys the local copies of the evacuated instances. Ironic is not covered.

linkedin:
  title: "Compute service: disabled and down are two different things"
  caption: |
    openstack compute service list shows two fields for every nova-compute. They are independent, and mixing them up is how instances get lost.

    Status (enabled / disabled) answers: may the scheduler place new instances here?
    It is a scheduling decision. Running instances are not affected.

    State (up / down) answers: does Nova believe the service is alive?
    It comes from the heartbeat, or from an operator who forced it down.

    The danger is in what Nova allows once a service is down: deleting an instance becomes a local delete (the guest runs on until nova-compute returns and destroys it), and evacuation is allowed.

    So never force down a compute that is not fenced. And before you re-enable one, read the disable reason: a reason starting with "AUTO:" was set by the libvirt driver and clears itself.
  hashtags: [OpenStack, Nova, CloudOperations, SRE, InfrastructureEngineering, OzarLand]
