id: PG-40
category: storage
title: Volume stuck in creating?
situation: >-
  The request is accepted, but the volume never becomes `available`: `error` within seconds, or `creating`
  for hours.

diagram:
  type: branches
  head: [Status and host of the volume, The cause is]
  split: 48
  rows:
    - see: "`error`, no host"
      means: No backend passed
      sub: Capacity, volume type, zone, backend down
      tone: stop
    - see: "`error`, a host is set"
      means: The backend failed
      sub: Driver error, or no space to convert the image
      tone: stop
    - see: "`creating`, its cinder-volume `down`"
      means: The service stopped
      sub: "At start, it sets these volumes to `error`"
      tone: warn
    - see: "`creating`, no host, services up"
      means: The request was lost
      sub: Check RabbitMQ first
      tone: warn

checks:
  - label: Status, chosen backend, type
    hint: openstack volume show <id> -c status -c os-vol-host-attr:host -c type
  - label: Cinder's own explanation
    hint: openstack --os-volume-api-version 3.3 volume message list
  - label: Are the Cinder services up?
    hint: openstack volume service list
  - label: What each pool has left
    hint: openstack volume backend pool list --long

caution:
  - level: destructive
    text: >-
      Forcing the status to `error` while a backend is still creating the volume makes the record lie, and
      can leave an orphan on the storage.
  - level: know
    text: >-
      A volume that was never `available` holds no user data, but it counts in the quota until it is
      deleted.

lesson: "No host: the scheduler. A host set: the backend. Cinder's message, or its log, says the rest."

runbook:
  path: volumes/volume-stuck-in-creating-or-error-after-create.md
  status: lab-verified
  release: "2026.1"

scope: >-
  Written from the Cinder 2026.1 source and Kolla-Ansible 2026.1; four causes reproduced on an all-in-one
  lab. A large image copied slowly is a further case: wait, and follow the cinder-volume log. A request
  refused at once (quota, invalid type) creates no volume and is not this case. Volumes created by Nova for a
  boot from volume, backups, snapshots and encrypted volumes are not covered.

linkedin:
  title: "Volume stuck in creating: one field tells the scheduler from the backend"
  caption: |
    A volume is created. The request is accepted. And it never becomes available: "error" within seconds, or "creating" for hours.

    One field splits the problem in two, the host Cinder chose for the volume:
    → error, no host: no backend passed the scheduler. Capacity, a volume type that matches no backend, the zone, a backend that is down.
    → error, a host is set: the backend was chosen and its driver failed.
    → creating, and that host's cinder-volume is down: the service stopped with the request in hand.
    → creating, no host, every service up: the request was lost. Look at RabbitMQ.

    And Cinder often explains itself in one line: openstack --os-volume-api-version 3.3 volume message list.

    The shortcut to avoid is forcing the status while a backend is in fact still working on the volume.
  hashtags: [OpenStack, Cinder, Storage, CloudOperations, SRE, OzarLand]
