Block Storage is commonly used by virtual servers as “volumes”, and behaves similar to disks attached to a physical server. They are expected to be used with filesystems or storage support inside the operating system of an instance. Every virtual server will have at least one block storage volume attached (referred to as the “root” volume) containing the operating system.
Block Storage volumes also back “Persistent Volumes” in our Managed Kubernetes services, so some of the concepts here apply to those as well.
The contents of a block storage volume can created as empty or from a source image or existing copy of another device. Details on how to create volumes from different sources are under this section of the documentation.
Block storage volumes can (with some exceptions) be detached or copied, snapshots and backups made of them, and can be attached to different servers as needed.
The block storage service implements techniques to reduce the impact of a physical failure, such as maintaining multiple copies of a block storage volume and regular checks to see these copies are consistent. However, you should always have backups of any block storage volumes, as failures can occur that affect data stored, and these features do not provide any means of recovery from deleted data inside the operating system. See Backups.
Block Storage volumes are tightly coupled to a specific location, and cannot be attached to servers in different locations. For storage that is distributed between regions, consider Object Storage instead.
Table of Contents: