メインコンテンツまでスキップ

Clone a Warehouse

Cloning creates a new warehouse from a Lightweight Snapshot backup. VeloDB Cloud provisions a new warehouse and restores the snapshot into it, giving you a copy of the source warehouse at the point the snapshot was taken. Use a clone when you want a separate copy of a warehouse without affecting the source.

Cloning uses the same restore action described in Restore from a backup, with New Warehouse selected as the restore target. This page covers the warehouse-provisioning side of that flow. Because cloning restores a full-warehouse snapshot, it is available only with Lightweight Snapshot, not Export Backup.

Note:

Lightweight Snapshot is in private preview. To use it, contact VeloDB Cloud Support to join the allowlist.

When to clone a warehouse​

Clone a warehouse when you want to:

  • Create a development, test, or staging copy from production data.
  • Validate an upgrade, schema change, or data reload against real data before applying it to the source warehouse.
  • Run a migration dry run or hand a point-in-time copy to another team.

For in-place recovery of the source warehouse, restore to the current warehouse instead. See Restore from a backup.

BYOC clone infrastructure and storage considerations​

You can clone both SaaS and BYOC warehouses. However, cloning a BYOC warehouse has infrastructure and storage considerations. This section explains those considerations.

Note:

For a BYOC warehouse, a cloned warehouse runs in the same cloud infrastructure as the upstream (source) warehouse. VeloDB Cloud does not provision new infrastructure for the clone.

A BYOC clone reuses the upstream warehouse's infrastructure in two ways:

  • Network: The clone runs in the same VPC and region as the upstream warehouse. It does not create a new VPC.
  • Object storage: The clone is created from the snapshot using copy-on-write, so it references the same data files as the snapshot instead of copying them, and it writes its own new files to the same bucket as the upstream warehouse. The clone does not create a new bucket. For the storage model, see How snapshots and clones use storage.

This has three consequences to plan for:

  • Isolation: A clone shares network and storage infrastructure with the upstream warehouse, so it is not isolated at the infrastructure level. Its compute runs on its own clusters and incurs its own compute service fee.
  • Storage billing: For a BYOC warehouse, the clone's data files are stored in your own bucket, and your cloud provider bills you for the physical data stored. Because copy-on-write stores files shared with the source once, the clone does not add duplicate storage. Storage is metered differently for SaaS warehouses. See Storage for snapshots and clones.
  • Cleanup: Because the clone shares data files and one bucket with the upstream warehouse and the snapshot, deleting a warehouse does not immediately reclaim shared files. See Delete a clone safely below.

Prerequisites​

Before you clone:

  • A Lightweight Snapshot backup of the source warehouse exists and is in Success status. You can only clone from a successful backup.
  • You have a warehouse name and an admin password ready for the new warehouse.
  • Lightweight Snapshot is in private preview. To use it, contact VeloDB Cloud Support to join the allowlist.

Clone a warehouse from a snapshot​

To clone a warehouse, follow these steps:

  1. Log in to the VeloDB Cloud console.
  2. In the upper-left corner, select the source warehouse.
  3. In the left navigation pane, under MORE, click Backups.
  4. On the Backups tab, find the snapshot backup in Success status. In the action column, click the actions menu, then click Restore from Backup.
  5. In the Restore Backup dialog, for Restore to, select New Warehouse.
  6. In Warehouse Name, enter a name for the cloned warehouse.
  7. In Admin Password, enter the admin password for the cloned warehouse.
  8. (Optional) Expand Advanced Settings to adjust the Zone, Compute, and Cache for the cloned warehouse. For a BYOC clone, the zone options are limited to the upstream warehouse's infrastructure.
  9. Review the Compute Service Fee shown for the cloned warehouse. The clone runs on its own clusters and is billed separately from the source warehouse.
  10. Click Confirm. VeloDB Cloud creates the new warehouse and starts a restore task that loads the snapshot into it.

Track the restore task in the Restores tab. After it completes, verify the cloned warehouse's databases, tables, row counts, and representative queries, and confirm the connections and permissions your applications need.

Delete a clone safely​

A BYOC clone shares data files and one object storage bucket with its upstream warehouse and the snapshot it was cloned from. Because storage uses copy-on-write, deleting a warehouse does not immediately reclaim the files it shared. A data file is removed only after no warehouse references it and every snapshot that references it is removed.

Warning:

  • Deleting a clone frees only the new files that clone owned. Files it shared with the source warehouse or the snapshot remain until those also stop referencing them.
  • Do not empty or delete the shared bucket while any warehouse or snapshot still references files in it. Emptying the bucket removes data for every warehouse that shares it, and this cannot be undone.

Clear the bucket only after you delete the last warehouse that uses it and remove the snapshots that reference its files. For the provider-specific teardown steps, see Delete a Warehouse.

See also​