I've been upgrading some of my disks with larger ones, and ran a preclear on the new disks before adding to the array (mainly to check for failures), then when adding to the array, replacing an old (smaller) disk, it still needs to write to the whole new drive capacity.

Proposal: If unraid detects a preclear signature on the new disk, the replacement only takes as long as writing the original disk's capacity to the new drive.

E.G. replacing a 4TB disk with a 16TB disk, if the 16TB disk is precleared, then there's only 4TB of data to copy to the new disk, with the remaining space already being zeroed (due to preclear), therefore parity will already be valid. This would then take 1/4 the time to add.

I second this. Replacing an existing disk in an array with a larger one should take the time of copying the data on the replaced disk, not the new (blank) one.

Scenario: I originally set up my array with 16T drives. I upgraded the Parity to 28T, now I'm upgrading data drives to 28T as they are affordable.

If the disk contains a maximum of 16T of data, why spend the time copying 28T?

This makes a big difference as we get to these larger drives that take multiple days to fully copy.

The related question is why if the disk is 50% full and is replaced with one of the same size, why does the full capacity of the drive have to be copied?