Restore objects with Batch Operations - Amazon Simple Storage Service
Differences when restoring from S3 Glacier and S3 Intelligent-TieringOverlapping restoresLimitations

Restore objects with Batch Operations

You can use Amazon S3 Batch Operations to perform large-scale batch operations on Amazon S3 objects. The Restore operation initiates restore requests for the archived Amazon S3 objects that are listed in your manifest. The following archived objects must be restored before they can be accessed in real time:

  • Objects archived in the S3 Glacier Flexible Retrieval or S3 Glacier Deep Archive storage classes

  • Objects archived through the S3 Intelligent-Tiering storage class in the Archive Access or Deep Archive Access tiers

Using a Restore (S3InitiateRestoreObjectOperation) operation in your S3 Batch Operations job results in a RestoreObject request for every object that's specified in the manifest.

Important

The Restore job only initiates the request to restore objects. S3 Batch Operations reports the job as complete for each object after the request is initiated for that object. Amazon S3 doesn't update the job or otherwise notify you when the objects have been restored. However, you can use S3 Event Notifications to receive notifications when the objects are available in Amazon S3. For more information, see Amazon S3 Event Notifications.

When you create a Restore job, the following arguments are available:

ExpirationInDays

This argument specifies how long the S3 Glacier Flexible Retrieval or S3 Glacier Deep Archive object remains available in Amazon S3. Restore jobs that target S3 Glacier Flexible Retrieval and S3 Glacier Deep Archive objects require that you set ExpirationInDays to 1 or greater.

Important

Don't set ExpirationInDays when creating Restore operation jobs that target S3 Intelligent-Tiering Archive Access and Deep Archive Access tier objects. Objects in S3 Intelligent-Tiering archive access tiers aren't subject to restore expiration, so specifying ExpirationInDays results in a RestoreObject request failure.

GlacierJobTier

Amazon S3 can restore objects by using one of three different retrieval tiers: EXPEDITED, STANDARD, and BULK. However, the S3 Batch Operations feature supports only the STANDARD and BULK retrieval tiers. For more information about the differences between the retrieval tiers, see Understanding archive retrieval options.

For more information about the pricing for each tier, see the Requests & data retrievals section on the Amazon S3 pricing page.

Differences when restoring from S3 Glacier and S3 Intelligent-Tiering

Restoring archived files from the S3 Glacier Flexible Retrieval or S3 Glacier Deep Archive storage classes differs from restoring files from the S3 Intelligent-Tiering storage class in the Archive Access or Deep Archive Access tiers.

  • When you restore from S3 Glacier Flexible Retrieval or S3 Glacier Deep Archive, a temporary copy of the object is created. Amazon S3 deletes this copy after the value that you specified in the ExpirationInDays argument has elapsed. After this temporary copy is deleted, you must submit an additional restore request to access the object.

  • When restoring archived S3 Intelligent-Tiering objects, do not specify the ExpirationInDays argument. When you restore an object from the S3 Intelligent-Tiering Archive Access or Deep Archive Access tiers, the object transitions back into the S3 Intelligent-Tiering Frequent Access tier. After a minimum of 90 consecutive days of no access, the object automatically transitions into the Archive Access tier. After a minimum of 180 consecutive days of no access, the object automatically moves into the Deep Archive Access tier.

  • Batch Operations jobs can operate either on S3 Glacier Flexible Retrieval and S3 Glacier Deep Archive storage class objects or on S3 Intelligent-Tiering Archive Access and Deep Archive Access storage tier objects. Batch Operations can't operate on both types of archived objects in the same job. To restore objects of both types, you must create separate Batch Operations jobs.

Overlapping restores

If your S3InitiateRestoreObjectOperation job tries to restore an object that's already in the process of being restored, S3 Batch Operations proceeds as follows.

The restore operation succeeds for the object if either of the following conditions is true:

  • Compared to the restoration request already in progress, this job's ExpirationInDays value is the same and its GlacierJobTier value is faster.

  • The previous restoration request has already been completed, and the object is currently available. In this case, Batch Operations updates the expiration date of the restored object to match the ExpirationInDays value that's specified in the in-progress restoration request.

The restore operation fails for the object if any of the following conditions are true:

  • The restoration request already in progress hasn't yet been completed, and the restoration duration for this job (specified by the ExpirationInDays value) is different from the restoration duration that's specified in the in-progress restoration request.

  • The restoration tier for this job (specified by the GlacierJobTier value) is the same or slower than the restoration tier that's specified in the in-progress restoration request.

Limitations

S3InitiateRestoreObjectOperation jobs have the following limitations:

  • You must create the job in the same Region as the archived objects.

  • S3 Batch Operations doesn't support the EXPEDITED retrieval tier.

For more information about restoring objects, see Restoring an archived object.