How long does it take to retrieve data from Glacier storage?

Data objects archived in Amazon Glacier vaults are not directly available. Instead, data is stored on Amazon S3 and developers move selected items to Glacier. Within Glacier storage, users can retrieve only a directory -- a list of the archive's contents -- not the archived data itself.

To retrieve data stored in Amazon Glacier, the desired objects must be temporarily restored to Simple Storage Service (S3); archived data is accessible from the temporary S3 copy. Restores can take three to five hours, and the copy is only accessible for the amount of time a developer stipulates when making a restore request. AWS charges for S3 and Glacier storage while the archive copy is available.

While the Glacier data restoration process seems confusing, it's similar to legacy tape-based backup and restoration paradigms. IT teams can only view and manage Glacier objects through Amazon S3, and object restores occur through the Amazon S3 console. Admins can also use the console or an S3 API to track the progress of a restoration.

Consider Glacier storage costs and terms of use to get the biggest benefits from the service. Data that requires prompt access should typically remain online in S3 or shift to a low-cost online option such as S3 infrequent access.