On this page

Tigris Cloud Disk presents a Tigris bucket as a Linux block device. You format
it with any filesystem (ext4, XFS, btrfs) and run any application on it,
databases included. Capacity is effectively unlimited, you pay only for the data
you store, and the volume inherits the durability, replication, and access
controls of the bucket underneath.

Beta

Cloud Disk is in beta and under active development.

Everything that defines a disk — its geometry, its settings, its contents —
lives in the bucket. The disk you unmount on one machine is the disk you mount
on the next, including on another machine or in another region.

To work with disks, install the CLI below. The
[Architecture](/content/docs/cloud-disk/architecture/index.html) page covers how it works, and
[Kubernetes (CSI)](/content/docs/cloud-disk/kubernetes/index.html) covers provisioning disks as
PersistentVolumes.

## Install

Auto-detects CPU arch and package manager (Linux, amd64/arm64):

```bash
curl -fsSL https://cloud-disk.t3.tigrisfiles.io/install.sh | sh
```

## Quick start

One command creates the disk, formats it, and mounts it at
`/mnt/cloud-disk/<name>`. Get your access keys from the
[Tigris console](https://console.storage.dev/). Credentials are remembered, so
it's the only time they're needed:

```bash
export AWS_ACCESS_KEY_ID=tid_... AWS_SECRET_ACCESS_KEY=tsec_...

sudo -E cloud-disk create mydata -size 100G
```

From then on:

```bash
sudo cloud-disk umount mydata     # flushes everything to Tigris, unmounts

sudo cloud-disk mount mydata      # mount again — here or on another machine
```

On a new machine, export the credentials for the first command; after that
they're remembered there as well.

## Snapshots & forks

You can snapshot a disk while it's mounted and in use, then fork the snapshot
into an independent copy. Everything is uploaded to Tigris before the snapshot
is taken:

```bash
cloud-disk snapshot mydata create --name nightly    # prints the snapshot version

cloud-disk create mydata-test --parent mydata --snapshot <version> \
    --set mount-point=/mnt/cloud-disk/mydata-test
```

The fork starts with the parent's data and settings at that snapshot; the two
disks are fully independent afterwards.

## Settings

Settings are stored with the disk — set once, they apply wherever it's mounted:

```bash
cloud-disk config mydata set max-cache-size=8G flush-workers=16

cloud-disk config mydata show
```

See [Configuration & Tuning](/content/docs/cloud-disk/tuning/index.html) for the settings worth knowing.

## Status

```bash
cloud-disk status mydata
```

## Delete

Deletes the disk and all its data, after asking for confirmation:

```bash
cloud-disk delete mydata
```

## Next steps

- [Architecture](/content/docs/cloud-disk/architecture/index.html) — how a bucket becomes a block device, and
the durability trade-offs behind write-through and write-back.
- [Kubernetes (CSI)](/content/docs/cloud-disk/kubernetes/index.html) — provision disks as PersistentVolumes.
- [Configuration & Tuning](/content/docs/cloud-disk/tuning/index.html) — the settings worth knowing and when
to change them.