Exercise 2: Why Large Drives Motivate a RAID 5 to RAID 6 Migration — Possible Solution
====================================================================

The specific risk:

Per the chapter, RAID 5's real weakness is rebuild time: "Replacing a
failed disk requires reading every remaining disk to reconstruct the
failed one's data -- on large modern drives this can take many hours,
and a second disk failure during that rebuild window loses the entire
array." At 18TB per drive, reconstructing a failed disk means reading
that same enormous amount of data back off every surviving disk in the
array, which is a genuinely long operation -- the chapter names this
risk as growing "directly with disk size, which is the genuine,
practical reason RAID 5 is increasingly discouraged on today's
largest drives." The bigger the drives, the longer the array spends
in this exposed, single-failure-tolerance-already-used-up state during
a rebuild, and the more real-world time there is for a second,
unrelated disk in the same array to fail during that same window --
which, on RAID 5, means total data loss for the entire array, not just
a further-degraded-but-still-alive state.

Why RAID 6 addresses it:

Per the chapter, "RAID 6... survives two simultaneous disk failures,"
and specifically "directly addresses RAID 5's own rebuild-window
vulnerability: even if a second disk fails during the rebuild after
the first, RAID 6 can still survive it." On RAID 6, the exact scenario
that destroys a RAID 5 array (a second failure while still rebuilding
from the first) is precisely the case RAID 6 was designed to survive,
since it can tolerate two failures rather than RAID 5's one. The
tradeoff the chapter names -- "write performance is worse than RAID 5...
and it needs at least four disks" -- is the real cost being accepted
in exchange for surviving exactly the failure mode that large,
slow-to-rebuild drives make meaningfully more likely.

WHY THIS WORKS AS AN ANSWER
------------------------------
This names the specific mechanism (rebuild time scaling with disk
size, increasing the real-world window of exposure to a second
failure) rather than a vague "RAID 6 is safer," and explains precisely
which RAID 5 failure scenario RAID 6's own second-failure tolerance
directly closes.