1–3 of 3 posts
Of course bit flips can happen for same bit on 2 HDD at the same time but such cases potentially could be very very rare.
Depending on the implementation, reads can be spread across drives but no parity is checked.
No, in a RAID 1 configuration (of any number of drives) writes are written to all drives and each read comes from only 1 drive. Depending on the implementation, reads can be spread across drives but no parity is checked.
ZFS, on the other hand, stores the hash of every block separately from the block itself, and then uses those hashes to check every read. So it only reads the block from a single disk, but it hashes the read block and compares it against the expected hash. If there is an error of any kind the hashes won’t match and ZFS will increment the error counter for the drive and pool. It then examines the other mirrors to find the correct data. Once it has the correct data it overwrites the incorrect block.