Here is how I do it.
2 x Windows Box : One is my "Main" PC, and the other for "Backup").
Each runs a copy of "Drive Pool" that lets you:
- Group std HDDs of any size into a Pool and presents it to windows as a single blob
- You can select folders / files to be "duplicated" if you want redundancy (though I use the backup server for this)
- All my drives in these pools are formatted as ReFS (each drive is still visible as a "normal" drive if you want - no RAID
)
- I have an NVME drive in each pool that acts like a high speed write cache, and new media is then automatically moved to std HDDs (archive drives)
- the two PC's are connected by 10gbe so back is quick and easy even for large volumes.
- for backup I use Free File Sync to check for differences between the two, and only if I'm happy then do I commit the change to the backup PC (eg helps with incorrect deletes etc)
- As I need to increase capacity, I buy new large HDD (eg 16tb), put that in my MAIN PC's pool, move data off one of the older drives are then redeploy those to the backup server.
- I share the MAIN Media pool using UNC so all MC Library Server clients all the same path as the MAIN Library
It works, is simple, has redundancy, and importantly is high performance if you Library Server needs to transcode for DLNA Clients etc. It's can even be "cheap" if you have enough SATA Ports and Drive bays in your PCs (though I'm using IBM m1015 in flashed to IT Mode for extra sata ports & SAS Expanders).
I've tried lots of other options:
- Cheap NAS Box: Terribly unreliable and poor throughput
- RAID Controller in a PC: CHKDSK and my Raid Controller had a fight one boot up. Scrambled the entire array. Never Again. Took weeks to re-rip everything.