Create an Azure File Sync server endpoint

A server endpoint represents a specific location on a registered server, such as a folder on a server volume. A server endpoint must meet the following conditions:

  • A server endpoint must be a path on a registered server (rather than a mounted share). Network attached storage (NAS) is not supported.
  • Although the server endpoint can be on the system volume, server endpoints on the system volume may not use cloud tiering.
  • Changing the path or drive letter after you established a server endpoint on a volume is not supported. Make sure you are using a suitable path before creating the server endpoint.
  • A registered server can support multiple server endpoints, however, a sync group can only have one server endpoint per registered server at any given time. Other server endpoints within the sync group must be on different registered servers.
  • Multiple server endpoints can exist on the same volume if their namespaces are not overlapping (for example, F:\sync1 and F:\sync2) and each endpoint is syncing to a unique sync group.

This article helps you understand the options and decisions needed to create a new server endpoint and start sync. For this to work, you need to have finished planning for your Azure File Sync deployment and also have deployed resources needed in prior steps to creating a server endpoint.

Prerequisites

To create a server endpoint, you must first ensure that the following criteria are met:

  • The server has the Azure File Sync agent installed and has been registered. See Register/unregister a server with Azure File Sync for details on installing the Azure File Sync Agent.
  • Ensure that a Storage Sync Service has been deployed. See How to deploy Azure File Sync for details on how to deploy a Storage Sync Service.
  • Ensure that a sync group has been deployed. Learn how to Create a sync group.
  • Ensure that the server is connected to the internet and that Azure is accessible. Azure File Sync uses port 443 for all communication between the server and cloud service.

Create a server endpoint

To add a server endpoint, go to the newly created sync group and then select Add server endpoint.

Add a new server endpoint in the sync group pane

The Add server endpoint blade opens, enter the following information to create a server endpoint:

  • Registered server: The name of the server or cluster where you want to create the server endpoint.
  • Path: The Windows Server path to be synced as part of the sync group.
  • Cloud Tiering: A switch to enable or disable cloud tiering. With cloud tiering, infrequently used or accessed files can be tiered to Azure Files.
  • Volume Free Space: The amount of free space to reserve on the volume on which the server endpoint is located. For example, if volume free space is set to 50% on a volume that has only one server endpoint, roughly half the amount of data is tiered to Azure Files. Regardless of whether cloud tiering is enabled, your Azure file share always has a complete copy of the data in the sync group.
  • Initial download mode: An optional selection, that can be helpful when there are files in the Azure file share but not on the server. Such a situation can exist, for instance, if you create a server endpoint to add another branch office server to a sync group or when you disaster-recover a failed server. If cloud tiering is enabled, the default is to only recall the namespace, no file content initially. That is useful if you believe that rather user access requests should decide what file content is recalled to the server. If cloud tiering is disabled, the default is that the namespace will download first and then files will be recalled based on last-modified timestamp until the local capacity has been reached. You can however change the initial download mode to namespace only. A third mode can only be used if cloud tiering is disabled for this server endpoint. This mode avoids recalling the namespace first. Files will only appear on the local server if they had a chance to fully download. This mode is useful if for instance an application requires full files to be present and cannot tolerate tiered files in its namespace.

To add the server endpoint, select Create. Your files are now kept in sync across your Azure file share and Windows Server.

Cloud tiering section

When creating a new server endpoint, you can opt into the cloud tiering feature of Azure File Sync. The options in the Cloud tiering section can be changed later. However, different options in the following section are available based on having cloud tiering enabled or not for your new server endpoint.

Refer to the cloud tiering article that covers the basics, the policies, and best practices in detail.

Initial sync section

The Initial sync section is available only for the first server endpoint in a sync group. For any additional server endpoint, see the Initial download section.

There are two fundamentally different initial sync behaviors:

Merge

Authoritative upload

Merge is the standard option and selected by default. You should leave the selection on Merge unless for certain migration scenarios.

  • When joining a server location, in most scenarios either the server location or the cloud share is empty. In these cases, Merge is the right behavior and will lead to expected results.
  • When both locations contain files and folders, the namespaces will merge. If there are files or folder names on the server that also exist in the cloud share, there will be a sync conflict. Conflicts are automatically resolved.

    Within the Merge option, there is a selection to be made for how content from the Azure file share will initially arrive on the server. This selection has no impact if the Azure file share is empty. You can find more details in the upcoming paragraph: Initial download

Authoritative upload is an initial sync option reserved for a specific migration scenario. Syncing the same server path that was also used to seed the cloud share with for instance Azure Data Box. In this case, the cloud and the server locations have mostly the same data but the server is slightly newer. Users kept making changes while Data Box was in transport. This migration scenario then calls for updating the cloud seamlessly with the changes on the server (newer) without producing any conflicts. So the server is the authority of the shape of the namespace and Data Box was used to avoid large-scale initial upload from the server. Server authoritative upload enables a zero-downtime adoption of the cloud, even when an offline data transport mechanism was used to seed the cloud storage.

A server endpoint can only succeed provisioning with the authoritative upload option, when the server location has data in it. This block is to protect from accidental misconfigurations. Authoritative upload works like RoboCopy /MIR. This mode mirrors source to target. The source is the AFS server and the target is the cloud share. Authoritative upload will shape the target in the image of the source.

  • New or updated files and folders will be uploaded from the server.
  • Files and folders that don't exist on the server (anymore) will be deleted from the cloud share.
  • Metadata-only changes to files and folders on the server will be efficiently moved to the cloud share as metadata-only updates.
  • Files and folders might exist on server and the cloud share. But some files or folders may have changed their parent directory on the server since the seeding of the Azure file share. These files and folders will be purged from the cloud share and uploaded again. Because of this, it's best to avoid restructuring your namespace at a larger scale during a migration.

Initial download section

The Initial download section is available for the second and any more server endpoints in a sync group. The first server endpoint in a sync group has extra options that relate to migration with Azure Data Box. These options don't apply if this server endpoint isn't the first one in your sync group.

Note

Selecting an initial download option has no impact if the Azure file share is empty.

As part of this section, a choice can be made for how content from the Azure file share will initially arrive on the server:

  • Namespace only
    Will bring only the file and folder structure from the Azure file share to the local server. None of the file content is downloaded. This option is the default if you previously enabled cloud tiering for this new server endpoint.
  • Namespace first, then content
    For a faster availability of the data, your namespace is brought down first, independent of your cloud tiering setting. Once the namespace is available on the server, the file content is then recalled from the cloud to the server. Recall happens based on the last-modified timestamp on each file. If there is not enough space on the server volume, the remaining files will remain tiered files. This option is the default if you did not turn on cloud tiering for this server endpoint.
  • Avoid tiered files
    This option will download each file in its entirety before the file shows up in the folder on server. This option avoids a tiered file to ever exist on the server. A namespace item and file content are always present at the same time. Avoid this option if fast disaster recovery from the cloud is your reason for creating a server endpoint. If you have applications that require full files to be present, and cannot tolerate tiered files in their namespace, this is ideal. This option is not available if you are using cloud tiering for your new server endpoint.

Once you selected an initial download option, you cannot change it after you confirm to create the server endpoint. How files appear on the server after initial download finishes, depends on your use of the cloud tiering feature and whether or not you opted to proactively recall changes in the cloud. The latter is a feature useful for sync groups with multiple server endpoints in different geographic locations.

  • Cloud tiering is enabled
    New and changed files from other server endpoints will appear as tiered files on this server endpoint. These changes will only come down as full files if you opted for proactive recall of changes in the Azure file share by other server endpoints.
  • Cloud tiering is disabled
    New and changed files from other server endpoints will appear as full files on this server endpoint. They will not appear as tiered files first and then recalled. Tiered files with cloud tiering off are a fast disaster recovery feature and appear only during initial provisioning.

Next steps

There's more to discover about Azure file shares and Azure File Sync. The following articles will help you understand advanced options and best practices. They also provide help with troubleshooting. These articles contain links to the Azure file share documentation where appropriate.