Question about architecture: Longhorn V1 and V2 (SPDK) data disks (not backups) create a network mirror (raid1), usually 3-way. Is there any mechanism for check-summing and checking data and metadata errors during standard read (as in ZFS or BTRFS)? And, thanks to this, is there a way to self-repair RAID 1 replica data (or even better, some kind of "zpool scrub ...")? I only found vague information about SPDK, that it might be possible to use an "NVMe DIF (Data Integrity Field) checksum" on NVME disks. Can "NVMe DIF" be used or is it used for virtualized SPDK vdev and SPDK raid1 to checking and repairing data and metadata?