If you're happy with your current btrfs setup, no need to switch.
That said...I'd be curious what data you're storing on that btrfs raid and how mission-critical you consider it to be. Is this a homelab/testing NAS, or is this where you're storing the backups of your wedding photos?
> The simple redundancy RAID levels utilize different mirrors in a way that does not achieve the maximum performance. The logic can be improved so the reads will spread over the mirrors evenly or based on device congestion.
Which reads to me like a sugar-coated way of saying "if you have 2 mirrored drives, it'll try really hard to read only from the first one, rather than splitting the load evenly across the two".
> RAID56
> Some fixes went to 4.12, namely scrub and auto-repair fixes. Feature marked as mostly OK for now.
I personally don't trust a NAS filesystem that claims to be "mostly OK for now" for RAID 5/6 and throws away the biggest performance benefit of RAID10.
My own home NAS runs ZFS, and has for 5+ years, under various iterations of FreeBSD and Linux. I'm quite happy with it, but I'm also looking forward to bcachefs, once it hits the mainline kernel. Especially as the root-on-bcachefs story matures, since as happy as I am with ZFS on Linux, the impedance mismatch between ZFS and Linux bootloaders continues to make ZFS root not impossible, but more complicated than it ideally should be.
That said...I'd be curious what data you're storing on that btrfs raid and how mission-critical you consider it to be. Is this a homelab/testing NAS, or is this where you're storing the backups of your wedding photos?
From https://btrfs.wiki.kernel.org/index.php/Status#RAID1.2C_RAID... :
> RAID1, RAID10
> The simple redundancy RAID levels utilize different mirrors in a way that does not achieve the maximum performance. The logic can be improved so the reads will spread over the mirrors evenly or based on device congestion.
Which reads to me like a sugar-coated way of saying "if you have 2 mirrored drives, it'll try really hard to read only from the first one, rather than splitting the load evenly across the two".
> RAID56
> Some fixes went to 4.12, namely scrub and auto-repair fixes. Feature marked as mostly OK for now.
I personally don't trust a NAS filesystem that claims to be "mostly OK for now" for RAID 5/6 and throws away the biggest performance benefit of RAID10.
My own home NAS runs ZFS, and has for 5+ years, under various iterations of FreeBSD and Linux. I'm quite happy with it, but I'm also looking forward to bcachefs, once it hits the mainline kernel. Especially as the root-on-bcachefs story matures, since as happy as I am with ZFS on Linux, the impedance mismatch between ZFS and Linux bootloaders continues to make ZFS root not impossible, but more complicated than it ideally should be.