Attach a Daytona volume to share large persistent data across multiple sandboxes

Documented steps

1. Create the volume: Python `daytona.volume.create('volume-name')`; TypeScript `await daytona.volume.create('volume-name')`. The CLI can create and mount in one step with `daytona create --volume my-volume:/home/daytona/volume`.
2. Capture the returned volume's `id` — mounts reference the volume by id, not name.
3. Build a mount descriptor: `VolumeMount(volume_id=volume.id, mount_path='/home/daytona/data', subpath='users/alice')`, using `subpath` for per-tenant isolation inside one volume.
4. Validate the mount path: it must be absolute, must not be root, must contain no relative components or consecutive slashes, and must not target system directories such as /proc, /sys, /dev, /boot, /etc, /bin, /sbin, /lib or /lib64.
5. Pass the mount(s) in the `volumes` parameter of the sandbox creation call so the volume is attached at start.
6. Mount the same volume id into a second sandbox — at the same or a different path — to share data between them.
7. Expect writes to become visible to other sandboxes immediately, since volumes are FUSE-based mounts.
8. Keep database files, SQLite, and anything needing block-level semantics on the sandbox's own disk rather than on the volume.
9. Stay within the per-organization cap of 100 volumes and reuse volumes with subpaths instead of creating one per task. Official docs: https://www.daytona.io/docs/en/volumes, https://www.daytona.io/docs/en/warm-pools