> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.ibee.co.in/docs/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.ibee.co.in/docs/_mcp/server.

# Create a snapshot

> Capture a manual recovery point of a Cloud VM — root disk plus optional data volumes.

Manual snapshots produce an on-demand **Recovery Point** for a Cloud VM. They include the root disk and, optionally, attached data volumes. Use snapshots before risky changes, or as a clone source for new VMs via [restore](/docs/tools/snapshots/restore-a-snapshot).

## Steps

### 1. Open Snapshots

On the server detail page, open the **Snapshots** tab. The first time you visit, an info banner reminds you that **restoring a snapshot will stop the server briefly, roll back the disk, and start it again automatically.**

### 2. Add a label and (optional) description

Under **Create Snapshot**:

* **Label** (required) — short, human-readable name. Example: `pre-deploy-2026-05-09`.
* **Description** (optional) — context like the change you're about to make. Example: `Snapshot before upgrading Postgres to 17`.

Use clear, dated labels. Recovery points are also tagged with a system-generated identifier (`srp-20260509T162709Z`) but the label is what you'll see first in the list.

### 3. Pick a Capture Mode

Three modes are offered. Root volume is always captured; the mode controls what else is included.

| Mode                     | Capture                                                                    |
| ------------------------ | -------------------------------------------------------------------------- |
| **Root volume only**     | Boot disk only. Attached data volumes are skipped.                         |
| **All attached volumes** | Root disk plus every attached persistent data volume.                      |
| **Selective volumes**    | Root plus the data volumes you tick from the **Select Data Volumes** list. |

A green confirmation banner — **Root volume is always included** — appears regardless of mode.

### 4. (Selective only) Pick data volumes

For **Selective volumes**, the **Select Data Volumes** list appears with each attached volume's:

* Label (for example `testblockstorage`)
* Size and device path (for example `10 GB · vdc · /mnt/volume_<id>`)

Tick the boxes for the volumes to include. The header shows a running count (`1 selected`, `2 selected`, …).

### 5. Create the snapshot

Click **Create Snapshot** in the bottom-right of the form. The platform briefly captures the disks and produces a recovery point.

A success toast — **Snapshot set created** — confirms the operation. The new entry appears in **Recovery Points**.

## What you get

Each recovery point shows:

* **Snapshot** — your label, plus the system identifier (`srp-<timestamp>`).
* **Disks** — what was captured (`Root + 1 data disk`, etc.) with the provisioned disk size displayed for each volume.
* **Created** — date and time.
* **Storage** — total recovery storage used (provisioned size, not compressed size).
* **Actions** — **Details**, **Restore**, **Delete**.

Click **Details** to inspect the captured disks (root vs. data, target devices, filesystem types, and disk sizes) and copy mount instructions for any data volumes.

## Capture mode tips

* **Root volume only** is the fastest and smallest. Use when you're protecting only the OS install.
* **All attached volumes** is the safest default for production. Use when data and root must be recoverable together.
* **Selective volumes** is useful when one volume holds large, unimportant data (build cache, temp files) and you want to skip it.

## Best practices

* Use **descriptive, dated labels** so you can find the right recovery point later.
* **Quiesce the workload** if you need a transactionally consistent snapshot — flush database buffers (`pg_start_backup`, `FLUSH TABLES WITH READ LOCK`) before snapshotting, or stop the application briefly.
* **Test restore periodically** with **Create New VM** to confirm your snapshots are usable.
* **Delete old snapshots** you no longer need — recovery storage is billed.

## Verify

Switch to the **Recovery Points** section at the bottom of the Snapshots tab. The new row should match your label and show **Storage** (provisioned disk size) and **Created** values. If it shows `Loading snapshots…` or `No snapshots yet`, refresh after a few seconds.

## Troubleshooting

**Create button is disabled**
Enter a label — it's required.

**Snapshot stuck in progress**
Open the [Activity feed](/docs/infrastructure/cloud-vms/monitoring#activity-feed). A failed `VM snapshot_create` event will show with a message you can act on.

**Selective volumes list is empty**
The VM has no attached data volumes. Either switch to **Root volume only** or attach a [block storage volume](/docs/infrastructure/cloud-vms/attach-volumes) first.

## Related pages

* [VM snapshots overview](/docs/tools/snapshots)
* [Restore a snapshot](/docs/tools/snapshots/restore-a-snapshot)
* [Attach volumes](/docs/infrastructure/cloud-vms/attach-volumes)
* [VM backups](/docs/tools/backups)