Bucket versioning

Bucket versioning is the capability to store the history of an object using versions. Each version is a complete copy of the object and occupies space in Object Storage. By using version control, you can protect your data from both unintentional user actions and application faults.

Versioning is enabled at the bucket level and applied to every object in the bucket.

For more information about how to enable versioning, see Managing bucket versioning.

• After you enable this feature, the version_id parameter is added to each uploaded object. It lets you manage specific object versions.

• Before versioning is enabled, each bucket object is assigned a version ID (version_id) equal to null.

  When versioning is paused, the version_id of existing objects no longer changes. Each new object is assigned a version ID of null. If a null version already exists, it's overwritten.

• When you overwrite an object version, a new object with the same ID and a randomly generated version_id is created.

  To access a previous version of the object, use the object ID and the desired version_id.

When you delete an object version, a delete marker is added to the version and it no longer takes up space.

You can remove unneeded or deleted object versions from a bucket manually or by setting up the object lifecycle.