Cloud tiering is an optional feature of Microsoft Azure File Sync that ensures local file servers always have sufficient free space.
It also helps make sure that frequently accessed files are cached locally on a file server, while all other files are tiered (moved) to an Azure file share.

When a file is tiered, the File Sync file system filter replaces it locally with a pointer called a reparse point, which represents a URL to the file on an Azure file share.
When a user requests a tiered file, File Sync automatically recalls the file data from Azure Files without the user needing to know that the data is stored in Azure.

Configure Cloud Tiering

You can configure cloud tiering for each server endpoint.
If cloud tiering is disabled (default setting), all file data is synchronized from the cloud endpoint to the server endpoint.
After enabling cloud tiering, you can configure two policies, as described in the following table:

The volume free space policy always takes precedence over the date policy.
For example, if the date policy says files accessed in the last 7 days should remain locally, but the free space policy requires 20% free space, recent files will also be tiered if they exceed the 80% disk usage threshold.

If multiple server endpoints are on the same volume (each in a different sync group), the effective free space threshold is the highest among all endpoints.
For example, if one endpoint has a 20% policy and another has 30%, the volume will maintain at least 30% free space.

How Cloud Tiering Works

The File Sync system filter builds a namespace map on each server endpoint.
It monitors access attempts (read and write) over time, then assigns a heat score to each file based on frequency and recency of access.
A file accessed frequently and recently is considered hot, while a file rarely accessed and not opened for a long time is cold.

When the volume exceeds the free space threshold, File Sync moves the coldest files to Azure Files until the free space percentage is met.

With cloud tiering, files can be cached locally or moved to Azure.
Some file formats (such as media files or compressed ZIP files) can also be in a partially tiered state: only part of the file is cached locally.
This occurs when you open a tiered file and the format supports partial reads, allowing you to use the file without downloading it completely.

Recognize Tiered Files

Tiered files are accessed transparently, just like locally cached files.
You can identify them in several ways:

• Tiered files do not use local disk space because they are stored on an Azure file share.
  Regardless of their actual size, the size on disk is 0 bytes.
• In File Explorer, you can check the file properties, including the space it occupies on disk.
  For partially tiered files, the size on disk is greater than 0 bytes but less than the actual file size.
• The following attributes are set on tiered files: Offline, Reparse point, and Recall on data access.
  In File Explorer, you can view these in the Details tab or by adding the Attributes column.
• Tiered files have reparse pointers set.
  A reparse pointer is a special pointer used by the File Sync system filter.
  To check if a file has a reparse point, run the following command:

fsutil reparsepoint query <file-name>

fsutil reparsepoint query <file-name>

If the file has a reparse point, the output will include:

Reparse Tag Value: 0x8000001e

Reparse Tag Value: 0x8000001e

Manually Tier or Recall Files

When cloud tiering is enabled, files are automatically moved according to the defined policies.
You can manually force a file to be tiered, for example if you just saved a large file you won’t use for a long time.
Trigger the move with the PowerShell command:

PowerShell

Invoke-StorageSyncCloudTiering
Afficher plus de lignes

If a file is already tiered and you want to recall it, the easiest way is to open it (double-click or select in File Explorer).
You can also run the command:

Invoke-StorageSyncFileRecall

Invoke-StorageSyncFileRecall

This is especially useful if you want to recall multiple files at once.

Next unit: Migrate from DFSR to Azure File Sync