> For the complete documentation index, see [llms.txt](https://docs.vega365.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.vega365.com/redundancy-management-framework/user-manual/backfill/key-rmf-backfill-features.md).

# Key RMF Backfill Features

Peer-to-Peer Backfill

### Overview

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2F1SBWXkqjJRDlzkWUWxUp%2FRMF_P2P_BFill.png?alt=media&amp;token=27b2631c-0180-42c4-9185-1cb8ed22f1d1" alt="" width="375"><figcaption></figcaption></figure>

RMF Backfill restores video continuity across active-active redundant Milestone XProtect recorders after an outage. When a Primary or Secondary recorder goes offline- for maintenance, a network drop, a hardware failure, or any other reason- its peer keeps recording. Once the offline recorder returns, **Backfill copies the missed footage from the peer that stayed online**, so both sides of the redundant pair end up with the same continuous timeline.

The transfer happens **peer-to-peer between recording servers**, on a schedule and bandwidth budget you control from the Management Client. There is no external broker, no manual file copy, and no operator intervention required.

***

### How It Works

Backfill runs as a lightweight Windows service — **BackFillService** — installed on every XProtect recording server that participates in RMF. The Management Client manages the service centrally under **MIP Plug-ins → RMF → BackFill**.

When a recorder comes back online, its Backfill service:

1. Identifies the time window during which it was unreachable.
2. Looks up its **redundancy correspondence:** the same Primary↔Secondary mapping established during RMF configuration to determine which peer recorder holds the missed video for each of its cameras.
3. Requests the missing media from that peer over a dedicated port.
4. Writes the recovered video into local storage so playback, export, and evidence workflows behave as if the recorder was never down.

Because Backfill reuses RMF's existing redundancy correspondence, it needs no additional camera-by-camera mapping. If RMF knows where a camera's redundant copy lives, Backfill knows where to fetch it.

***

### The Flow Dashboard

The **Flow Dashboard** gives a live, at-a-glance view of every recorder participating in Backfill. Primary servers appear on the left, Secondary servers on the right, with animated flow lines connecting redundant peers. Each server card shows its online state, camera count, and whether Backfill is currently enabled.(`BACKFILL ON`).

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2F2maWtarhCankXN1Kgp3d%2F2026-04-14_12-31-46.png?alt=media&amp;token=eefacde2-4d63-4a1c-8046-1ab55389e33f" alt=""><figcaption></figcaption></figure>

When a recorder is offline, its card and flow line shift state so operators can see, without drilling into logs, exactly where redundancy is currently degraded.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FkHkZphPdPLk6EiTo8MKq%2FBackfill_1.PNG?alt=media&amp;token=48523a05-1891-4584-8a20-b5f970f2c4c8" alt=""><figcaption></figcaption></figure>

***

### The Configuration View

The **Configuration** tab lists every recording server in the deployment along with its current Backfill status: service state, uptime, number of backfills performed, and total data transferred.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FVd4CtKEPjhHrscwDwf36%2F2026-04-14_12-38-41.png?alt=media&amp;token=5b9a0c90-dc19-4f6e-b70e-ee793c16e3be" alt=""><figcaption></figcaption></figure>

From here, an administrator can:

* Enable or disable Backfill per recorder.
* Set the **listening port** and **target drive** for backfilled media.
* Define a **bandwidth throttling schedule** so that peer-to-peer transfers don't compete with peak corporate network traffic.
* Restart or refresh the service.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2F16glHrtSmYVgvUXrkLxN%2F2026-04-14_12-38-12.png?alt=media&amp;token=05248a96-9ae8-4e7e-8657-284426f08450" alt=""><figcaption></figcaption></figure>

#### Bandwidth Throttling

The throttling schedule accepts multiple time windows, each with its own MB/s ceiling. A typical pattern is generous overnight, restrictive during business hours, and moderate in the evening — letting Backfill catch up quickly when the network is quiet without disrupting daytime operations.

***

### Backfill History

Each recorder maintains a **Backfill History** showing the most recent transfer jobs: the time range that was recovered, when the request was issued, how much video was moved, and how long it took. The full history is available for download for audit and reporting.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2Fuz7LBNbNRr8IqShhTLeS%2FBackfill_2.PNG?alt=media&amp;token=a1a5d896-11e6-4831-9479-499daf305ad9" alt=""><figcaption></figcaption></figure>

### Redundancy Correspondence: What Backfill Supports

RMF lets administrators model redundancy in whatever way fits the deployment, and Backfill follows the same model.&#x20;

#### Server-to-Server Correspondence

All cameras on a Primary recorder are mirrored onto a single Secondary recorder. See [Mode Selection](https://docs.vega365.com/redundancy-management-framework/user-manual/configuration-dashboard/mode-selection.md).

* **Symmetric:** one Primary maps to one Secondary (1:1).

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FinTu5YjfImvlQKatV1IB%2Fone-to-one.png?alt=media&amp;token=b7e0b6d0-68a2-4032-b6cb-78eb5e0ad709" alt="" width="563"><figcaption></figcaption></figure>

* **Asymmetric:** many Primaries consolidate onto fewer Secondaries (N: M, N>M)

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FnZrEYlSmbb1AWnRAEE1E%2Fimage.png?alt=media&amp;token=f97e928f-8c83-42b5-99a2-424ac62e4e4a" alt="" width="563"><figcaption></figcaption></figure>

#### At-Will Correspondence

Cameras on Primary recorders are distributed across Secondary recorders in any user-defined arrangement, allowing maximum flexibility. See: [Mode Selection](https://docs.vega365.com/redundancy-management-framework/user-manual/configuration-dashboard/mode-selection.md)

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2F4YBgW4FdgTKAXWUD2fTl%2FAt-Will_BF.png?alt=media&amp;token=b7a800ee-858b-4d24-900b-2ab8ea3c9524" alt="" width="563"><figcaption></figcaption></figure>

Backfill supports **all** modes. When a recorder returns from an outage, its service consults the correspondence map and retrieves each camera's missing video from whichever peer holds it, even if that means contacting multiple peers to recover a single recorder.

***

### Asymmetric Retention Policies

A camera's storage profile on a Primary recorder does not have to match its storage profile on the Secondary. Retention periods, archive schedules, live/archive tier boundaries, and storage locations can all differ between the two sides.

Backfill handles this transparently. When a recorder returns from an outage, Backfill:

1. Fetches the missing media from whichever peer holds it, regardless of how that peer has tiered the footage.
2. Consults the **local** storage profile on the returning recorder.
3. Deposits each block of recovered video into the correct tier — **live**, **archive 1**, **archive 2**, and so on — exactly where it would have landed had the recorder never gone offline.

The only requirement is that the content still exists somewhere on the peer. The peer's retention policy, archive structure, and disk layout are irrelevant to the destination; Backfill always honors the **returning recorder's own** storage rules.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FlJED7saX40LZ0Bc4r6E7%2FSP_BF.png?alt=media&amp;token=0e3147cb-d025-43e8-9408-31742e30e0e3" alt="" width="563"><figcaption></figcaption></figure>

#### Example 1: Shorter Live Window on Secondary

A Primary keeps 7 days of live recording before archiving; the Secondary keeps only 2 days of live before archiving to colder storage. A Primary goes offline for 4 days. On return, it backfills from the Secondary. Some of the requested videos are still in the Secondary's live tier; others are already in the Secondary's archive. Backfill pulls both and deposits them into the **Primary's** live tier — because on the Primary, 4-day-old footage is still within the live window.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2Fh9S87qwpEFoP1d8uh4Jo%2FEx1_BF.png?alt=media&amp;token=405e6c59-c440-4d0d-bce1-cf50c4a55da7" alt="" width="563"><figcaption></figcaption></figure>

#### Example 2: Longer Retention on Secondary

A Secondary site is configured as the long-term retention tier: 90 days total, with 7 days in live and 83 days in archive. The Primary keeps only 30 days, all in live. A Secondary recorder goes offline for **three weeks**. When it returns, it backfills the 21-day gap from the Primary. On the Primary, all 21 days are in live; on the Secondary's profile, the most recent 7 days are in live storage, and the older 14 days are in the archive. Backfill deposits the recovered video accordingly, **routing each day of footage to the correct local tier based on the Secondary's own retention rules**, not the Primary's.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FFTE6xl6JVovZ6uIzC0yV%2FEx2_BF.png?alt=media&amp;token=5d239fb0-7b99-47e8-a6f6-fe90a0573cde" alt="" width="563"><figcaption></figcaption></figure>

#### Why This Matters

Symmetric storage profiles force customers to over-provision the cheaper site or under-retain at the more expensive one. Backfill's profile-aware deposit logic removes that constraint: **each site keeps the storage profile that makes sense for its role**, and redundancy still works correctly across the mismatch.

***

### Architectures Supported

Backfill is designed to work across every architecture RMF itself supports:

* **Single-site** — recorders share one XProtect deployment with a clustered Management Server / SQL.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FZDMbsUUjwTScJEGvnDew%2FRMF_Single_Site.png?alt=media&amp;token=f32b7548-c61b-4377-90b5-ffff8f4df149" alt=""><figcaption></figcaption></figure>

* **Federated** — recorders belong to a Parent site and one or more Child sites.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2Fkn7gyXNOCzkmDoumrBMM%2FRMF_Federated.png?alt=media&amp;token=efa78ee4-000a-4f72-8476-c0644343be0a" alt=""><figcaption></figcaption></figure>

* **Independent sites** — each site is its own XProtect deployment, joined only by RMF.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FezRo5o1GpTmF6RDzNzuX%2FRMF_Independent.png?alt=media&amp;token=0e4976c2-5a29-456f-ae4a-4163e605ccde" alt=""><figcaption></figcaption></figure>

***

### XProtect Variants Supported

It also works across every XProtect product tier:

* XProtect Express+
* XProtect Professional+
* XProtect Expert
* XProtect Corporate

***

### Limitations

#### Platform Support Dependency

RMF Backfill is not a native Milestone XProtect feature. It reads from and writes to the XProtect media database independently of XProtect's recording pipeline. Because of this, **changes to XProtect's internal media database format — which Milestone may make at any time without notice — can break Backfill's ability to read, transfer, or deposit video.**

Vega Systems works to keep Backfill compatible with each XProtect release, but the media database format is not a published API. If a breaking change occurs, a fix may depend on collaboration with Milestone. There is no guarantee that every format change can be accommodated, or that a fix can be delivered.

#### Edge Recording Exclusion

Assign cameras that support edge recording and cameras that rely on RMF Backfill to **separate recorder pairs**. Assign edge-recording cameras to a dedicated redundant pair with Backfill disabled; assign Backfill cameras to their own dedicated redundant pair with Backfill **enabled**. The two groups must not share a recorder.

Backfill is unnecessary for edge-recording cameras. When a recorder in the pair comes back online after an outage, it independently retrieves the missing content directly from each camera's onboard storage. No peer-to-peer transfer between recorders is involved.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FkaMJzN1gqcRU5oeS8aOX%2FEdge_Cams_and_Backfill_Cams_Separate_Recorder_Pairs.png?alt=media&amp;token=5e74c727-179f-4fe2-977e-f117f73d3c67" alt="" width="563"><figcaption></figcaption></figure>

#### Recording Server Encryption

XProtect recorders must **not** be configured to encrypt content before writing to disk. The Backfill application cannot transfer encrypted content. Ensure that media database encryption is disabled on all recorders participating in Backfill.

<figure><img src="https://73860689-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FK95oxUATgzuZZ2VhRoDw%2Fuploads%2FDFN6evEJfbzDb6h216vH%2Fdo_not_encrypt.png?alt=media&amp;token=5ecdd80e-c385-4b95-a55d-f42ebd13ea1c" alt="" width="563"><figcaption></figcaption></figure>
