Storage regions and availability zones¶

Storage is a core service and is available in all Binero clouds regions and availability zones. Below we outline the region and availability zone concepts for storage. If you are looking to roll out a geo redundant platform, understanding the base design principles will help you create a very resilient and highly available system as well as securing your data in the event of a hypothetical availability zone outage scenario.

Volumes¶

Volumes are local to an availability zone. Binero cloud does not stretch our storage platform(s) but use individual storage in each ends. This is to guarantee that the storage is not impacted in several zones at once in the event of some incident with the storage platform.

Binero cloud currently does not allow for replication of block storage between zones, this is mainly because using cluster file systems that take replication of block storage into account is usually less resilient than using a replication method connected to the application that stores data (for instance clustering of MySQL) - as well as more complex and with more dependencies.

When wanting to keep a fileshare up to date in more than a single availability zone, there are several techniques that will keep a filesystem synchronised, for instance DRBD or Rsyncd that can run above the platform layer. We would generally recommend using object storage for these scenarios, if possible though.

Moving a volume between availability zones¶

For security reasons, there is no network connectivity between the storage platforms of our availability zones. Because of this, there are two main options to move volumes:

• Create a backup and restore in the different zone.

• Create an image from a volume and setup an instance in a different zone based on the image.

Object storage¶

Object storage is available in all our availability zones. Replication is available using a separate endpoint that will replicate between europe-se-1a and europe-se-1b. When using replicated object storage, there are no special considerations (with regards to the actual replication) for the user as the platform manages the entire replication. The user should however take care to connect to the endpoint that is local to the application (or to europe-se-1a when connecting from the outside).

NVMe storage¶

NVMe storage is not replicated and is currently only available in the europe-se-1a availability zone.

Snapshots¶

All snapshots are local to the volume that the snapshot depends on. This means that snapshots are store in the availability zone where the volume they belong to are stored. When creating a volume from a snapshot, it will be save in the same availability zone as the snapshot. While you are able to provision an instance from a snapshot in a different availability zone, this provision task will fail.