> For the complete documentation index, see [llms.txt](https://docs.silk.us/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.silk.us/learn/data-management-concepts/snapshots.md).

# Snapshots

Learn how Silk snapshots capture point-in-time data, how scheduling and retention affect recovery points, and how to restore or access snapshot data.

A Silk snapshot captures the state of a volume group's data at a specific point in time. Use snapshots to create recovery points before planned changes, protect data on a regular schedule, and provide point-in-time data for testing, troubleshooting, and other secondary workloads.

Silk creates snapshots instantly, without affecting the performance of the source volumes. Snapshot creation doesn't depend on the number or size of the volumes being captured.

For examples of how teams use snapshots, views, and clones to support development, analytics, AI, and other secondary workloads, see [Cloud Efficiency with Copy Data Management](/learn/data-management-concepts/cloud-efficiency-with-copy-data-management.md).

## How snapshots work

Silk creates snapshots at the volume group level. A volume group snapshot captures all the volumes and clones in that group at the same point in time and preserves the order of writes across them. Place related volumes in the same volume group when you need to capture them together, such as when an application stores its data, logs, and configuration on separate volumes. If an application's data spans multiple volume groups, create a cross-volume-group snapshot to capture those groups at the same time.

Snapshots share unchanged data blocks with the source. Silk uses a redirect-on-write approach: when a host writes new data, the DataPod writes it to a new location and preserves the original blocks for the snapshot. A snapshot consumes additional capacity only for data that changes on the source after the snapshot is taken, and that data benefits from the same data reduction as the source volumes.

Snapshots can't be mapped to hosts directly. To work with snapshot data on a host, create a view or a clone. For details, see [Restore or access snapshot data](#restore-or-access-snapshot-data).

{% hint style="warning" %}
**Before you rely on snapshots**

* The default independent volume groups, `INDEPENDENT_DEDUP_VOLUMES` and `INDEPENDENT_NON_DEDUP_VOLUMES`, don't support snapshots. Place volumes that need snapshot protection in a dedicated volume group.
* Moving a volume to another volume group deletes the volume's existing snapshots.
* Each volume group supports up to 40 snapshots and views combined. For other limits, see [Silk Platform Configuration Limits](/learn/references/silk-platform-configuration-limits.md).
  {% endhint %}

## Snapshot consistency

Each snapshot is either crash-consistent or application-consistent. In the SDP Management Console, the **Snapshots & Replication > Snapshots** page shows the consistency type of each snapshot.

* A **crash-consistent** snapshot captures data as it exists on the volumes at the moment of capture, including any writes still in progress at the application level. This is the default. An application or database that starts from crash-consistent data may need to perform its own recovery.
* An **application-consistent** snapshot is taken while no I/O is in progress. Before creating one, pause or quiesce application writes to the volume group.

For databases, Silk Echo coordinates operations between the database host and the DataPod, so you can create application-consistent database snapshots without managing I/O manually. For details, see [Silk Echo Overview](/learn/products-and-core-concepts/silk-product-fundamentals/silk-echo-overview.md).

## Snapshot scheduling and retention

Create a manual snapshot when you need a recovery point before a planned change, such as an application upgrade. Use snapshot schedulers to create recovery points at regular intervals. You can assign one scheduler to several volume groups, and a volume group can use several schedulers. For example, a volume group might use one scheduler for hourly snapshots and another for daily snapshots. Silk names scheduled snapshots using the volume group name, the scheduler name, and a UTC timestamp.

The **scheduler** controls when Silk creates snapshots. The **retention policy** controls how long Silk keeps them. Retention policies apply per volume group: if two volume groups use the same policy, each group retains its own snapshots independently. Review both when planning how far back you need to recover.

Silk includes the following default retention policies. You can also create custom policies based on a retention period or a number of snapshots to keep.

<table data-search="false"><thead><tr><th width="283.703857421875">Retention policy</th><th>Retention period</th></tr></thead><tbody><tr><td><strong>Best_Effort_Retention</strong> (default)</td><td>As long as capacity allows under the applicable capacity policy</td></tr><tr><td><strong>Test_and_Dev</strong></td><td>As long as capacity allows</td></tr><tr><td><strong>Backup</strong></td><td>7 days</td></tr><tr><td><strong>CDP</strong> (Continuous Data Protection)</td><td>1 day</td></tr><tr><td><strong>Clone_Retention</strong></td><td>1 hour (the default for snapshots created as clone sources)</td></tr><tr><td><strong>Replication_Retention</strong></td><td>Latest two snapshots created by external APIs</td></tr><tr><td><strong>External_Replication_Retention</strong></td><td>Latest two snapshots created by the replication engine during each RPO cycle</td></tr></tbody></table>

{% hint style="info" %}
Snapshots with dependent views or clones are locked and can't be deleted, regardless of their retention policy.
{% endhint %}

### How capacity affects snapshots

Retention policies aren't the only way snapshots are removed. To keep source volumes writable, capacity policies can automatically delete snapshots when a DataPod or volume group runs low on capacity. Under the default system capacity policy:

* At **90%** usage (Low free capacity), Silk deletes older snapshots. Views are retained.
* At **96%** usage (Capacity stress), Silk can't create new snapshots or views and deletes all existing snapshots and views except mapped views.

To keep an important snapshot through these states, select **Prevent auto-deletion by Capacity Policies** when you create it. The snapshot still follows its retention policy. Capacity thresholds are configurable for volume groups.

{% hint style="info" %}
**Account for snapshots when planning capacity**

Frequent writes and longer retention increase the amount of data your snapshots preserve. Review your expected write activity and retention needs when estimating DataPod capacity.

For details, see [Plan DataPod Capacity](/deploy/plan-your-deployment/plan-datapod-capacity.md).
{% endhint %}

## Restore or access snapshot data

Choose how to use a snapshot based on your goal. You can return the source volumes to an earlier point in time, expose the whole volume group's captured data through a view, or create a writable copy of a specific volume with a clone.

<table><thead><tr><th width="254.0740966796875">Goal</th><th>Use</th></tr></thead><tbody><tr><td>Return source volumes to an earlier point in time</td><td><strong>Restore the snapshot.</strong> A restore writes the snapshot data back to its base volumes and overwrites their current data.</td></tr><tr><td>Work with point-in-time data from an entire volume group</td><td><strong>Create and map a view.</strong> A view exposes a volume group snapshot to hosts as writable volumes. Changes made through the view don't affect the snapshot or the source volumes.</td></tr><tr><td>Create a writable copy of a specific volume</td><td><strong>Create a clone.</strong> A thin clone consumes capacity only as data is written to it. A full volume clone is available immediately and becomes fully independent of its source once Silk finishes copying the data in the background. Clones don't count toward the volume group's snapshot and view limit.</td></tr></tbody></table>

{% hint style="danger" %}
**Restoring a snapshot overwrites current data**

A restore replaces the data on the base volumes with the snapshot's data. Before restoring, stop any applications that use the volumes and, if needed, unmount their file systems on the host. During the restore, Silk creates a snapshot named `pre_restore_[index_number]`.
{% endhint %}

To learn the procedures for creating, scheduling, restoring, and deleting snapshots, see the [SDP Management Console User Guide](https://support.silk.us/sys/document/preview/Silk-Cloud-Platform-User-Guide-240223011600..pdf) (requires Silk Support Portal access).

## Snapshots and replication

Silk replication is snapshot-based: it asynchronously transfers volume group snapshots from one DataPod to another. Local snapshots are stored on the same DataPod as their source volumes, so use replication when you need recovery points on another DataPod, in another availability zone, or in another region for disaster recovery.

## Related resources

{% content-ref url="/pages/yH4ng7yl0HUw4u0AuGya" %}
[Volumes and Volume Groups](/learn/data-management-concepts/volumes-and-volume-groups.md)
{% endcontent-ref %}

{% content-ref url="/pages/bE7CIdjb9q4cwFW2CkWX" %}
[Cloud Efficiency with Copy Data Management](/learn/data-management-concepts/cloud-efficiency-with-copy-data-management.md)
{% endcontent-ref %}

{% content-ref url="/spaces/7UwK9S9vo5qr2qGPjCUv/pages/ZUkR7xovQDMAbI7eiZ8b" %}
[Replication and Recovery](/manage/protect-your-environment/replication-and-recovery.md)
{% endcontent-ref %}

{% content-ref url="/pages/y4QzejE28yq2J0hdbKOc" %}
[Silk Platform Configuration Limits](/learn/references/silk-platform-configuration-limits.md)
{% endcontent-ref %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.silk.us/learn/data-management-concepts/snapshots.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
