Storage
Durable storage for workloads that must not lose data. Block volumes follow an instance, shared file storage is mounted by many instances at once, and both are replicated across all three Availability Zones.
Information
Storage on UNIQCloud is what separates a server from the data it holds. A virtual machine's own disk is disposable — it exists only as long as the instance does. Anything a business depends on belongs on storage that outlives the machine, can be moved to another one, and is replicated so that a hardware failure is an operational event rather than a data-loss event.
Two families are available through Managed IaaS: block storage, which behaves like a disk attached to a single instance, and shared file storage, which many instances mount at the same time. For unstructured data such as documents, media, and archives, see Object Storage instead.
Outlives the instance
A persistent volume exists independently of any virtual machine. Rebuild or replace the server and reattach the data, without a restore.
Replicated across zones
Volumes are held on CEPH and replicated across all three local Availability Zones, so a rack or zone failure does not take the data with it.
Performance you choose
Block Storage offers three performance classes, so an archive and a transactional database are not paid for at the same rate.
Shared where it is needed
NFS shares let several instances read and write the same file system — the usual answer for clustered applications and shared content.
Storage options
Managed IaaS provides two storage options. Choose based on how the application uses its data: NFS when several instances need the same files, or Block Storage when one instance needs a durable disk with a defined performance level. Both options are replicated across the three local Availability Zones.
| Choose by requirement | NFS | Block Storage |
|---|---|---|
| Access model | One shared file system for multiple instances | A durable virtual disk for one instance |
| Best for | Shared application content, uploads, and clustered services | Operating system data, databases, and application disks |
| Performance choice | Shared file access through the NFS service | Select iops1, iops2, or iops3 for the workload |
| Data lifecycle | Independent of the instances that mount it | Independent of the instance it is attached to |
| Snapshots | Yes | Yes |
NFS
NFS is shared file storage. Multiple instances can mount the same share and read or write the same directory structure, which makes it useful for shared application content, clustered services, and user uploads. A share can be expanded as the application grows and can be protected with snapshots.
Choose NFS when the business needs a common file system and simple access from multiple servers. It is not a replacement for a high-performance disk attached to one database server, and it should not be used as a general-purpose backup archive when Object Storage is a better fit.
Block Storage
Block Storage provides a persistent virtual disk for a virtual machine. The volume continues to exist when an instance is rebuilt or replaced, so data can be moved between servers without restoring it from backup. It is the natural choice for operating system data disks, databases, and applications that need consistent disk performance.
Block volumes are attached to one instance at a time, can be resized, and can be snapshotted for backup and recovery. Select the IOPS class according to the workload's access pattern and business impact: lower-cost storage is suitable for infrequent access, while higher IOPS reduces waiting time for demanding applications.
Ephemeral disk
The local disk included with a virtual machine flavor is not one of the two managed storage options. It survives a guest operating system reboot but is permanently deleted when the instance is terminated. Use it only for temporary data, caches, or workloads that can recreate their data.
Volume classes
Block Storage provides three classes of Cinder volumes, each optimised for a different IOPS requirement. All three are replicated across every local Availability Zone; the choice is between performance and cost.
iops1 · 500 IOPS
Choose the lowest-cost class for development, logs, and low-intensity data.
iops2 · 1,000 IOPS
Use the balanced class for most application and business workloads.
iops3 · 4,000 IOPS
Reserve the highest performance for transactional databases and latency-sensitive services.
| Specification | iops1 | iops2 | iops3 |
|---|---|---|---|
| IOPS | 500 | 1 000 | 4 000 |
| Business positioning | Lowest cost for infrequent access | Balanced cost and performance | Fast response for critical workloads |
| Typically used for | Development, logs, and low-intensity data | Most application and business workloads | Transactional databases and latency-sensitive services |
| Replication | Replicated across all local Availability Zones | ||
| Snapshots | Yes | ||