Restore a volume snapshot

Create a new volume from a completed volume snapshot, in the snapshot's zone.

Last updated 18 November 2026View as Markdown

A restore always creates a new volume; it never writes into an existing one. Replace the ids with your own.

Important

The new volume is billed for its size from creation until you delete it.

Prerequisites

  • Permission for ec2:CreateVolume, and ec2:DescribeSnapshots on the snapshot (else 404 "Snapshot not found").
  • In the console, also ec2:DescribeVolumes to see the new volume on Volumes, which opens after the restore (without it the page says "Access denied").
  • The snapshot is completed, else 409 "Snapshot … is …" with its state. In the console, Restore to new volume is unavailable until then.
  • Room in the volumes and volume storage quotas.
  • Any tags your organization requires, else 400 "Missing required tags: …".

Restore

  1. Open Compute, then Volume snapshots, and choose the snapshot in the list.
  2. Choose Restore to new volume.
  3. For Name, keep the suggested name (the snapshot's name, or its volume's id, followed by -restored) or enter another.
  4. For Size (GiB), keep the snapshot's size or enter more.
  5. Choose Restore. Volumes opens.

The volume is created in the snapshot's zone when you name none; another zone is refused with 409 "Snapshot … is in …; a volume restored from it is created in the same zone." In Terraform, set availability_zone_id to the snapshot's zone when it is not de1a. A size below the snapshot's is refused with 400 "The volume must be at least … GiB, the snapshot's size". kmsKeyId does not apply (400): a volume restored from an encrypted snapshot is encrypted with the snapshot's key and opens like the original; one restored from an unencrypted snapshot is not encrypted.

Result

The volume stays creating (Restoring n % in the console) until every piece is written and checked against its hash, then turns available; nothing formats it. If the restore fails, the volume turns error. Measured on 2026-11-12: restoring a snapshot of a 50 GiB volume holding 6.6 GiB into a new 50 GiB volume took 73 seconds. A bigger volume keeps the filesystem's size: grow it on the instance (resize2fs; for an encrypted volume, see Grow or rename a volume).

Did this page help you?