Remote desktops and virtual applications are critical, and growing, part of many businesses. Running applications in a non-persistent and virtual environment requires simplified management of user-profiles and Microsoft Office 365 applications. FSLogix by Microsoft was designed to improve the Virtual Desktop user experience and simplify the profile management in the Citrix and VMware VDI infrastructure.
Keeping the VDI infrastructure and applications available and resilient to infrastructure problems is crucial for keeping business running. Administrators that deploy FSlogix profile and Office 365 containers need to provide high availability for FSlogix profile containers.
When developing a highly available solution you need to consider two major paths and choose which one suits you better.
Option 1: Use FSlogix Cloud Cache
Cloud cache solves several conception issues. First, it provides a local copy of cached hot profile data of the container for the user profile, which reduces the network and CPU load of the file server. All read operations are performed locally, which significantly improves user experience. Using local cache also overcomes temporary outage of the infrastructure. All changes to profiles are stored locally first and then can be replicated to profile storage location. You can specify several profile locations that will automatically create infrastructure highly available in an active-active configuration.
While this sounds like a perfect solution, it has several major issues:
- You need to provide enough storage for the cache.
- Longer logon and logoff time due to a need to pre-load cache and flushing it back to the storage locations.
- Depending on the size of your FSlogix profile, Office 365 profile containers and network, it could translate into the significant delays that could affect user performance.
Option 2: Leverage VHD locations and replicate data between these locations by other solutions
You will get an active-passive high availability solution, but it won’t affect logon/logoff time and will not require any hardware updates. This approach may be also be preferred for roaming profiles, when you can implement a more complex replication strategy between on-premise file share servers in different locations and blob or file storage in the cloud.
It is still a preferred solution for some customers to achieve meaningful results now and have time to plan a proper deployment and improvement of the VDI, or when logon time is critical for user experience.
VHD/VHDX Replication options
The Distributed File System Replication (DFSR) service is an integrated Windows service that replicated data between data centers. It monitors changes to the file system and replicates only changed blocks between Windows servers. The DFS-R shows acceptable performance LAN and completely nonoperational over WAN. When you need to replicate FSLogix containers outside the office over the WAN, you should consider using other products. Another major issue of DFS-R is a lack of visibility and reporting for replication. DFS-R can significantly vary replication time, sometimes up to 50x or it can even completely stall without a clear way to detect or resolve the issue. It won’t be the best solution if you want to replicate VHD files geographically or to the cloud.
Robocopy is the file copy tool for Microsoft Windows. It provides basic copy operation between two locations. Because of its simplicity and availability, some people decide to use it for copying profile containers between storages. The obvious limitations of this approach are that you need to develop a set of scripts to make it robust (survive restart, report errors) to make it really work in the production environment. It can’t monitor folders for changes, so you need to periodically start it to perform copy operations. This delays time to replicate containers. Like DFS-R it is extremely slow and unpredictable when sending containers over the WAN. To learn more about Robocopy’s limitations, please read our Robocopy blog post.
Rsync is a well-known tool to effectively keep folders synchronized. It is very effective in detecting changes and making sure the folders are the same. It uses checksums to guarantee that files are identical after the transfer. While it is very effective in LAN, performance is significantly reduced over the WAN. You can’t use it to synchronize and replicate containers directly to blob storage in the cloud. You will also need additional scripts to make this work reliable in your production environment. Rsync can’t use real-time detection changes and needs a constantly rescan folder to detect changed files. To learn more about Rsync’s limitations, please read our Rsync blog post.
Resilio Connect is the best solution to replicate FSLogix containers for both Profile and Office containers. It monitors a folder for real-time changes and initiates replication in seconds. It is very effective both within an office (LAN) or outside the office (WAN). It can transfer containers to several destinations concurrently by leveraging peer-to-peer technology. It uses diff based transfer to transfer only changed blocks of the file. You can transfer directly to file and an object-based cloud of your choice: Azure, AWS, and Google Cloud or other. It also replicates NTFS permissions. Finally, it has a centralized management console that provides visibility, control, and notifications.
|Cloud, Blob, SMB||✔|