While this is likely true at the consumer level, it is not true at the enterprise level. Raid6 is used extensively, and the file hash is stored in metadata, separately. The files are walked periodically to detect and repair corruption. The same is done for parity. This has to be done with zfs too, via zfs scrub. I just don't think you can directly compare zfs with raid, as zfs is a filesystem and therefore implements some of the same things that other companies end up building on top of raid with their own custom filesystems.
I did not realize there were RAID implementations that did block level checksums.
ZFS is both a file system and a block layer. You would use RAID + ext4 or ZFS. The block layer parts of ZFS are comparable to RAID and are superior in many ways, including the most important: checksums.
Comments
While this is likely true at the consumer level, it is not true at the enterprise level. Raid6 is used extensively, and the file hash is stored in metadata, separately. The files are walked periodically to detect and repair corruption. The same is done for parity. This has to be done with zfs too, via zfs scrub. I just don't think you can directly compare zfs with raid, as zfs is a filesystem and therefore implements some of the same things that other companies end up building on top of raid with their own custom filesystems.
I did not realize there were RAID implementations that did block level checksums.
ZFS is both a file system and a block layer. You would use RAID + ext4 or ZFS. The block layer parts of ZFS are comparable to RAID and are superior in many ways, including the most important: checksums.