Migrate Cloudflare R2 bucket: Tenant-scoped object prefixes · Ample Computer
Migrate Cloudflare R2 bucket: Tenant-scoped object prefixes
Move object storage from Cloudflare R2 bucket to Ample, one component at a time. Destination verified on Ample: every tenant's objects stored under tenants/<id>/ in a private bucket with an ownership row per object, reads refused for keys outside the caller's prefix or not owned (cross=forbidden), and a prefix-bound listing returning only that tenant's objects (own=1). Source procedure: Inventory the bucket with rclone size or the R2 dashboard. Move separately: R2 public buckets become published Ample buckets; Workers bindings and lifecycle rules move into the app. Cutover: Switch the app's S3_* environment to the Ample bucket with a redeploy, verify reads and writes, keep R2 read-only until confirmed. Rollback: keep the source untouched until you confirm; nothing at the source is changed or deleted by this guide.
Verified November 5, 2026
What it runs on
- S3-compatible object storage verified
Buckets speak the S3 API with SigV4 credentials Ample Computer issues;ample bucket upload-urlreturns a presigned upload URL. PutObject and GetObject are checked by canary.
You need
- Authorized access to the Cloudflare R2 bucket source and its export tooling
- An inventory of every component in scope and out of scope
- A validated backup or copy before any cutover
- An Ample account token with servers:write, buckets:write
Exact tested configuration
build npm run build --if-present
install npm install
runtime node
sizes-1vcpu-1gb
start npm run start
template node-22
Steps
Inventory the source. List what Cloudflare R2 bucket provides beyond the component you are moving. Plan these separately: R2 public buckets become published Ample buckets; Workers bindings and lifecycle rules move into the app.
Source step 1. Inventory the bucket with
rclone sizeor the R2 dashboard.Source step 2. Copy objects with
rclone syncfrom the R2 S3-compatible endpoint to the Ample bucket endpoint using the issued credentials (path-style).Source step 3. Verify a sample of objects by size and checksum after the copy.
Create the destination bucket and copy. Create the bucket, copy with rclone or the S3 CLI in path-style mode using the issued credentials, then pass the credentials to the app as encrypted S3_* variables.
$ ample bucket create --name <bucket-name>
Validate before cutover. Run the app's own checks and, for data, compare counts and checksums; the example checks are the pattern self-tests (/p/tenant-scoped-object-prefixes).
Cut over. Switch the app's S3_* environment to the Ample bucket with a redeploy, verify reads and writes, keep R2 read-only until confirmed.
Rollback. Point DNS or configuration back to the source. The source was never modified; deletion is a separate, user-executed step after validation.
How it was checked
✓ destination app responds on its public URL· http get/ → “ample canary express patterns”
✓ tenant-scoped-object-prefixes check from the destination example· http get/p/tenant-scoped-object-prefixes → “see the pattern fixture checks”
✓ data or object counts and checksums match the source· manual → “operator comparison before cutover”
Evidence
- November 5, 2026 · canary run
Destination side verified: the Express pattern fixture deployed on Ample (npm install, npm run build --if-present, npm run start on the node-22 template) and its checks passed (every tenant's objects stored under tenants/<id>/ in a private bucket with an ownership row per object, reads refused for keys outside the caller's prefix or not owned (cross=forbidden), and a prefix-bound listing returning only that tenant's objects (own=1)). The source-side export from Cloudflare R2 bucket uses the vendor's standard tooling.
Tested examples
- Express pattern fixture (destination). Verified destination basis: tenant-scoped object prefixes.
What it doesn’t cover
- A step-by-step guide: each stage is an ordinary
ample deploy,ample databaseorample bucketcommand you run. - The destination side was verified on Ample with the pattern fixture; the source-side steps use Cloudflare R2 bucket's standard tooling.
- Plan separately: R2 public buckets become published Ample buckets; Workers bindings and lifecycle rules move into the app.
- Verified on the node-22 template at s-1vcpu-1gb.
- Buckets speak the standard S3 API; copy objects with rclone or the AWS CLI in path-style mode.