RAID 1 with two 16 TB drives: what big drives change about a mirror

With big drives, the rebuild is the whole story. A mirror rebuild copies one full drive, which takes 28 hours at best on a quiet array. The same copy on a 4 TB drive takes about 7 hours, so this one takes 4 times as long.

During that copy the surviving drive is your only copy, and the rebuild reads all of it, 16 TB, from start to finish. An unreadable sector on that drive means data the copy can't reach.

At a spec rate of 1 in 10^14 bits, the chance of at least one unrecoverable read error in that read is at most about 72.2%. At 1 in 10^15 bits it's at most about 12.01%. Both are upper bounds, because a spec rate is a ceiling. Check your drive's datasheet for its rate, since the real figure may be better. On a 4 TB mirror, the same 1 in 10^14 figure gives at most about 27.39%, so the bigger drive carries more of this risk.

Three 8 TB drives in RAID 5 give the same 16 TB. Their rebuild also reads the same 16 TB, so the read error bound matches. The layouts differ in how the data is spread, not in what a rebuild reads.

A mirror copies mistakes as fast as it copies good data. A deleted file, a bad save or a ransomware-encrypted file lands on both drives in the same write. RAID isn't a backup, so this array needs a second copy on separate storage.

This layout suits a two bay NAS or a backup target.