Does the disk rebuild speed in RAID6 depend on the number of disks in the array or not?

Question

RAID

I just can't find an answer to the question: does the disk rebuild speed in RAID6 depend on the number of disks in the array or not?
It is clear that the rebuild time is directly proportional to the capacity of the disk. And how is the speed related to the total number of disks in the array?

Answer 1

Theoretically, there is a weak dependence: the more disks, the slower. This is an average , i.e. if it's good to get confused, then you can choose such sets of disks from a large pile of disks to get the opposite. :)
I'll explain the essence: in order to restore some piece of data on one disk of the array, you need to read the corresponding pieces of data from the rest of the disks. But disk production is not perfect, and disks have a small spread in speed. Restoration of a piece of data will occur only after the slowest of the disks gives up the data. And the larger the group of disks, the greater (statistically) the spread in speed in this group. Those. on average, the slowest drive in a small drive group will be faster than the slowest drive in a large drive group.
This is pure mathematical statistics.

Answer 2

Not good. All disks will write to ONE disk, so the plug will be there. A big ride is bad because in order to rewrite just a block in 4k, you need to rewrite the whole stripe and pull ALL disks for recording. A rebuild depends only on the ability of the disk being restored.