Skip to content
CloudOps
Newsletter
All prompts
AI for OpenStack Difficulty: Intermediate ClaudeChatGPT

Glance Multi-Store & Image Cache Design Prompt

Architect Glance multi-store backends and the image cache — Ceph RBD vs file vs Swift stores, store priorities, copy-image, and per-compute caching — so instance boots are fast and image storage is placed cost-effectively.

Target user
Operators designing Glance storage backends
Difficulty
Intermediate
Tools
Claude, ChatGPT

The prompt

You are a senior OpenStack operator who has designed Glance multi-store layouts that make VM boots near-instant on Ceph-backed clouds and keep image storage cheap and predictable.

I will provide:
- Current Glance store config (`glance-api.conf` `[glance_store]`, enabled_backends)
- Backend options available (Ceph RBD, file/NFS, Swift, S3)
- Whether Nova/Cinder also use Ceph (copy-on-write boot eligibility)
- Image catalog size, formats (qcow2/raw), and boot patterns
- Pain points (slow first boot, image duplication, store imbalance)

Your job:

1. **Choose the store layout** — recommend which backends to enable and the default store, justified by the cloud's Ceph usage. Emphasize: if Nova+Cinder use Ceph, images should be **raw** in Ceph RBD to enable copy-on-write cloning (instant boot, no download).

2. **Format policy** — when to enforce raw vs qcow2; how to convert on upload (`image_conversion`); the storage-vs-boot-speed tradeoff. Flag that qcow2 on Ceph defeats COW cloning.

3. **Multi-store mechanics** — `enabled_backends`, per-store priority, default store, and `glance image-create` store targeting. How `copy-image` replicates a public image across stores/sites.

4. **Image cache** — for non-COW backends, configure the Nova/Glance image cache (cache dir, max size, pruner/cleaner periodic tasks) so repeated boots skip re-download.

5. **Lifecycle & hygiene** — detect orphaned RBD images vs Glance DB, dedupe identical images, and reclaim space; manage public/community image visibility and signing.

6. **Placement economics** — put hot/golden images on fast Ceph, archival/rarely-booted images on cheaper stores; document the policy.

Output as: (a) recommended store layout + default store with justification, (b) format policy and conversion settings, (c) the `glance-api.conf` keys to set, (d) image-cache tuning for non-COW paths, (e) an orphan/dedup reclamation procedure.

Bias hard toward raw-on-Ceph COW cloning when the backend supports it — it is the single biggest boot-speed win.
Newsletter

Free: the DevOps AI Incident-Triage Cheat Sheet

Subscribe and we’ll send you the one-page cheat sheet — plus weekly AI prompts, automation ideas, and tool reviews for infrastructure engineers. One email a week. No spam, unsubscribe anytime.

  • AI Incident-Triage Cheat Sheet (PDF)
  • Access to 1,603 DevOps AI prompts
  • One practical workflow email per week