Volume reference

How volumes work, from the disks to the bill. Prices are also on the product page.

Disks

A volume is a Standard or an Enterprise disk, shared or dedicated. The ids the API takes are in List disks.

The IOPS and throughput limits apply to reads and to writes each. Standard runs on Gen5 NVMe. Enterprise runs on enterprise NVMe with power-loss protection, which makes fsync faster, for databases and message queues. How fast a server can drive a disk also depends on its CPU, see what a server reaches.

Shared and dedicated

Both run on the same storage with the same limits. When the storage cluster runs short of resources, the performance of shared volumes is lowered first and dedicated volumes keep theirs. A volume switches between shared and dedicated in place; the data stays and the new price applies from the next hour. Standard and Enterprise live on different storage, so a volume cannot move between them: make a new one and copy the data.

Attaching

A volume attaches to one running server at a time, without a restart. The server sees it as the next disk, usually /dev/vdb, then /dev/vdc. The first time, make a filesystem on it and mount it; add it to /etc/fstab with nofail so the server boots without it too.

Unmount the volume inside the server before you detach it. A detached volume keeps its data and can go to another server. When a server is deleted, its volumes are detached and kept, unless you ask to delete them with it.

Growing

A volume grows while it is attached and in use, from its page or with the API, and cannot shrink. After it grows, grow the filesystem inside the server:

sudo resize2fs /dev/vdb      # ext4
sudo xfs_growfs /mnt/data    # xfs, by mount point

Snapshots and backups

A snapshot is a quick copy of the volume kept next to it, for a rollback or a new volume. A backup is an encrypted copy stored in another country, with a schedule you set on the volume or the account default. A new volume can be made from either. See Backups.

Billing

A volume is billed by the hour for its size, attached or not; a month is 720 hours. Snapshots and backups are billed apart, by what they store. Billing stops when the volume is deleted.

Durability

Every block of a volume is kept on more than one storage node, apart from the hosts the servers run on, so a failed disk or host does not take the data with it. Replicas do not protect from a deleted file or a broken filesystem; for that, keep snapshots and backups.

Limits

How many volumes an account can have and how large they can be depends on the account level; a new account has smaller ones until its first top-up, see Trial accounts. Your limits are on the Service Limits page in the Cloud Panel, where you can also ask for more.