Volumes
A volume is durable block storage you can attach to a server. It lives independently of the server's boot disk, so you can detach it, move it to another server, resize it or back it up without touching the machine itself. Manage volumes under Storage → Volumes in the console.
What a volume is for
- Keep data that must outlive a server — detach it before you delete a machine and attach it to a new one.
- Grow storage independently of compute.
- Back data up to your own S3 or rclone destination and restore it later.
A volume is created in a region and can only be attached to a server in that same region.
Create a volume
- Open Storage → Volumes and click Create volume.
- Give it a name, choose a size in GB and pick a region.
- The console shows the live per-GB monthly rate and an estimated monthly cost as you change the size.
- Click Create.
Volumes are provisioned lazily: a fresh volume becomes available immediately, and the underlying storage is materialized on a node the first time you attach it.
Size limits
The minimum volume size is 10 GB. New volumes are formatted with ext4 by default. Billing is per GB of provisioned size.
Attach and detach
Attach connects a volume to a running server:
- Open the volume's menu and choose Attach.
- Pick one of your running servers and confirm.
The volume and the target server must be in the same region. Once attached, the volume shows the server it's bound to and a mount point inside the guest.
Detach releases the volume from its server (menu → Detach). The data stays on the volume — you can attach it elsewhere afterwards.
WARNING
You must detach a volume before you can delete or restore it. A volume that is attached (or mid-attach/detach) cannot be removed, because that would leave a dead disk on the guest.
Resize
Volumes can only grow — you cannot shrink them.
- Open the volume's menu and choose Resize.
- Set a new size larger than the current one (up to the platform maximum) and confirm.
The estimated monthly cost updates to reflect the new size. The filesystem is grown to match the new capacity.
Backups
Volume backups are written to a backup destination you control, and you can run them on demand or on a schedule.
Backup destinations
Click Destinations on the Volumes page to manage where backups are stored. Two destination types are supported:
- S3 — an S3-compatible bucket. Provide the endpoint, bucket, access key ID and secret access key (region, prefix and provider are optional).
- rclone — an rclone remote. Provide the rclone config and the remote name (with an optional path).
TIP
Credentials are encrypted at rest and are only sent to the node that performs the backup or restore.
Run a backup
From a volume's menu choose Backup, select a destination and start it. The backup runs in the background and moves through pending → running → completed (or error).
List and delete backups
Choose Backups on a volume to see its history — each entry shows status, size, destination and when it was created. Deleting a backup removes both the catalog entry and, best-effort, the object in the remote destination.
Restore
From the Backups list, restore a completed backup back into the volume.
WARNING
The volume must be detached before you restore, and only backups in the completed state can be restored. Restoring overwrites the volume's current contents.
Scheduled backups
Open a volume's backup settings to configure an automatic schedule:
- Enable the schedule and pick a destination.
- Choose a frequency — daily or weekly.
- Set a retention count — how many recent backups to keep (default 7). Older backups beyond this count are pruned automatically.
The console shows the next and last run times for the schedule.
Using a volume inside the guest
When a volume is attached it is exposed to the guest as a block device and mounted at the mount point shown in the console. If you ever need to work with it manually, the standard Linux flow applies:
# Find the device
lsblk
# Format it (only for a brand-new, empty volume — this erases data)
sudo mkfs.ext4 /dev/vdb
# Mount it
sudo mkdir -p /mnt/data
sudo mount /dev/vdb /mnt/data
# Make it persist across reboots (use the device UUID from `blkid`)
echo 'UUID=<uuid> /mnt/data ext4 defaults,nofail 0 2' | sudo tee -a /etc/fstabWARNING
Only run mkfs on a brand-new, empty volume. Formatting a volume that already holds data will erase it.