Exercise 1: What the Traditional Stack Requires That zpool Doesn't — Possible Solution
====================================================================

Explanation:

A traditional stack -- ext4 or Btrfs on top of LVM on top of mdadm
RAID -- requires each layer to be set up and managed as a genuinely
separate step, using separate tools. First mdadm would be used to
combine /dev/sdb and /dev/sdc into a redundant RAID array (a distinct
device, e.g. /dev/md0). Then LVM would be used to turn that array
into a physical volume, group it into a volume group, and carve out
one or more logical volumes from it. Only after both of those layers
exist would a filesystem (ext4 or Btrfs) finally be created on top of
a logical volume with mkfs. Three separate subsystems, three separate
tools, three separate places something can be misconfigured or fail.

Per the chapter, "zpool create tank mirror /dev/sdb /dev/sdc" collapses
all of that into a single command and a single system. ZFS's own pool
is built directly from the raw disks, with the mirroring (redundancy)
specified right there at creation time -- there is no separate mdadm
array to create first, and no separate LVM physical/volume group/
logical volume chain to set up afterward, because "a zpool is built
directly from raw disks or partitions, with redundancy... configured
at pool-creation time -- no separate mdadm or LVM layer needed
underneath at all." The dataset created afterward with "zfs create
tank/data" is immediately usable as a filesystem, with none of the
intermediate layers a traditional stack would require.

This is possible specifically because ZFS was designed from the start
as one integrated system managing volumes, redundancy, and the
filesystem together, rather than as a filesystem meant to sit on top
of whatever volume-management layer already existed underneath it --
which is the traditional Linux model ext4 and (mostly) Btrfs both
still follow.

WHY THIS WORKS AS AN ANSWER
------------------------------
This names the specific separate tools/layers (mdadm, then LVM's own
PV/VG/LV chain, then mkfs) the traditional stack genuinely requires,
and explains why ZFS's own single command replaces all of them, tying
back to the chapter's own explicit statement of the integrated model.