That's what nearline SATA disks are for.
This is a very real problem for SSD's (we've had customers with pairs of mirrored SSD's wear out within a couple of days of each other after
2 years use). Much less so with hard drives, as long as you don't run them past their expected lifetimes (I always have to replace due to needing more space before they get to 4 years old).They have vastly different error handling.
A desktop drive is designed in the expectation it has the only copy of your data on it. When it gets a read error, it will try lots of things to try and get that data back, from simple rereading of the sector, through to reseeking from both ends of the disk, head wobble, and trying to stitch the sector back together from several different reads. That can take a long time, and if that fails, your operating system may ask the drive to try several more times to read the data, repeating the whole task over more times. It can be a minute or more before everything gives up trying to get your data back in the worst case.
OTOH, an Enterprise drive (including nearline drives) have the expectation that they're running in a RAID environment, and the last thing you want is for the drive to spend tens of seconds trying to get back a sector which the RAID can read in 4ms from another disk. So an Enterprise drive will "fail fast", to allow the RAID array as a whole to continue functioning without a long pause.
If you simply compare these two behaviours in a standalone fashion, then the desktop drive will appear to have a higher reliability, and the Enterprise drive to return more sector errors. But Enterprise reliability is more than just that of an individual disk - it's the reliability *and performance* of the RAID array as a whole, and ensuring that one faulty disk doesn't impact the whole array. If you use Enterprise disks in a desktop, they will seem to be unreliable, and if you use desktop disks in a RAID, the RAID will behave badly. They're each designed for specific use cases.