With the recent BTFRS vs XFS cache issues, it has me rethinking my cache pool setup and I'd like to ask what my best options are
So I have an HP enterprise server with bay loaded with 8 SFF 200GB enterprise SSDs, and 1 consumer Samsung 1 TB SATA SSD. Currently I have the 8 small SSDs in a cache pool, and the 1 large as an unassigned disk. My main array is made out of spinning disks. The cache pool has seen an unusual amount of writes as I've moved to a more docker heavy system, rather than full server VMs. Also, having 8 small disks in the cache pool is pretty underwhelming with Raid 1.
Options:
1 .Set the small SSDs into a RAID 5 array in my storage controller, and present that to Unraid as 1 volume and convert to XFS. Continue using the unassigned disk for VMs
2. Convert the large disk to XFS, and make it the cache, and either have 8 unassigned SSDs or RAID them with storage controller as an unassigned disk
3. Convert the large disk to XFS and just not use the enterprise SSDs until the bug is fixed
4. Do a bunch of scripting to move docker files/directories around onto my unassigned disk, or move one of the 200GB SSDs to unassigned to be a sacrificial disk for Docker