Rapid Storage Technology
Intel® RST, RAID
2055 Discussions

Intel RST RAID 1 Bit Rot ECC/Checksum Handling


TLDR: Can anyone confirm that RST is storing separate ECC/checksum data on each RAID1 disk as to the integrity of their blocks?



Backstory: I have some experience with ReFS and RAID1 ZFS on Linux that uses end to end checksums. RAID1 ZFS is such that if there is a block checksum failure on disk A but not disk B (or the inverse) then the data from the good checksum disk block read gets used and fixes the bad one on the alternate disk.


I read this Intel RST article today.




It's not 100% clear how RST determines which source has good data when there is a mismatch between the blocks of RAID1 disks. It just says the following.


"If the data on the mirror drive does not match the data on the source drive, the data on the mirror is overwritten with the data on the source."


Worst case scenario this means that if RAID1ArrayDisk0 is deemed the "source" and RAID1ArrayDisk1 is the "mirror" then if RAID1ArrayDisk0 has some block corruption, it will foolishly duplicate that corruption to the "mirror."

0 Kudos
0 Replies