In my infrastructure, I have two servers with Windows Server 2019 and Hyper-V installed. A SAN is directly connected to both servers via FC. The SAN provides three volumes to both servers: a volume for the quorum, a volume for VMs and a volume for data.
I plan to deploy a file service that is as much high available as my given infrastructure can go. Therefore - as I have two nodes - I want to deploy two file servers. This way, I can tolerate the failure of one whole server (host) or a failure of one virtual file server. With just one virtual file server (with HA enabled), I would just tolerate the failure of one host, but not a failure of the VM itsself.
I plan to use the data volume of my SAN to deploy a shared virtual hard disk, that both virtual file servers will use to provide the file shares.
Further more, I want that the users don't have to care which file server they access to access their files. \\FileSrv1\Data\README.md should be the same as \\FileSrv2\Data\README.md, but users should be able to access it like \\FS\Data\README.md. As far as I know, this is a typical use case for DFS. But I don't want two file servers that replicate their data, as I have a shared storage.
So my question is, can I use both - a shared storage for virtual file servers AND DFS to abstract the file access - in my scenario?
It turned out that I do not necessarily need DFS in my scenario. The guest failover cluster of the file servers already provides an abstraction layer to access the files, so I just called the clustered role
fs. In conclusion, files can always be accessed by\\FS\...no matter which of the file servers is active.