Exercise 2: Wants= vs. Requires= for a Database Dependency — Possible Solution
====================================================================

-- With Wants= --
--
-- Per the chapter's own definition, Wants= is a soft dependency: if
-- the database fails to start, the service still proceeds to start
-- normally anyway, per the chapter's own wording, "if the wanted unit
-- fails to start, the wanting unit still proceeds normally." The
-- service would come up regardless of the database's own success or
-- failure -- which, for a service that genuinely CAN'T function
-- without its database, would likely mean the service starts
-- successfully from systemd's own point of view, but then fails or
-- errors out on its own once it actually tries to connect to a
-- database that was never actually running.

-- With Requires= --
--
-- Requires= is a hard dependency: per the chapter's own definition, if
-- the required unit (the database) fails to start, "systemd treats
-- this unit's own startup as a failure too." The service itself would
-- NOT be considered successfully started if its database dependency
-- failed -- systemd would report the service's own startup as having
-- failed, rather than letting it come up in a broken, database-less
-- state. Requires= additionally means that if the database is later
-- STOPPED while the service is running, systemd can stop the
-- dependent service too, per the chapter's own description.

-- Which is correct for a service that genuinely cannot function
without its database --
--
-- Requires= is the correct choice here, using the chapter's own
-- worked example directly -- the chapter itself names exactly this
-- case ("Requires= its own database if it literally cannot function
-- without one") as the appropriate use for the stricter dependency
-- type, since Wants= would let the service come up in a state where
-- it's reported as "running" by systemd while actually being
-- completely non-functional underneath.

WHY THIS WORKS AS AN ANSWER
------------------------------
This works through the concrete consequence of each choice
specifically for a database-dependent service (Wants= letting it come
up broken vs. Requires= correctly reporting failure), then explicitly
ties the conclusion back to the chapter's own stated example rather
than a generic restatement of the two definitions.