This article describes how to use Snapshots to capture and store filesystem state of a Pouta virtual machine.

Types of snapshots

There are two types of snapshots used in OpenStack: image (instance) snapshots and volume snapshots. Both snapshot types can be utilized when creating a new instance, image snapshots being the more common use case for this.

Snapshot types

Image snapshots hold the state of a given instance's root disk, and thus can typically be used as a bootable image. Please note that if the instance has an ephemeral disk, the contents of the ephemeral disk are not captured into the snapshot.

Image snapshots are managed under Compute | Images in the web UI. For instructions on creating image snapshots please see section Creating an image based on an existing image in the [next chapter].

Volume snapshots capture the state of a persistent volume. They can be used for example for one-time backup before an upgrade operation that touches volume contents. Volume snapshots are managed under Compute | Volumes in the web UI.

General Considerations

At the time of writing there is one [known issue] related to snapshots, to be fixed in a future IaaS upgrade.

We recommend powering off an instance before taking a snapshot of it. This is the best way to make sure the file system is captured in a consistent state.

Please note that snapshots may not always be the optimal method of getting an instance into a predefined state. For longer term cloud usage, configuration automation tools should be considered. These mostly replace using snapshots by starting from scratch then automatically setting up relevant software and configurations. Examples of these tools are Ansible and Puppet. By using these tools one can get more of the automatization benefits of the cloud, and get the benefit of latest security updates upon instance creation (as long as CSC built images are used).


Note to CSC: add links above to configuration automation tools