No the drive does not need to bad
Would not recommend playing with a production server raid just to try things out. Strange thing can happen. Would definitely recommend you play with your raid before production. Panic never revived a server.
As an example of your question...
Few month ago, on a client's server, the Dell software informed me a drive was going south; the info stated the drive was not "bad" but it was only time before it would die. Anyway, call Dell and they sent another out, and I pull the "near bad" drive out and replaced it, hot. This system has a hot spare online, the controller rebuilds the info to the hotspare automatically.
Have not had a raid server without a hot spare for years so I am not sure but either you must reboot for the rebuild to start or you have to go into the array software and manually start the rebuild
Have replaced drives or "tested" arrays without powering down servers many times. Was on Adaptec's site a while back and they do not recommend pulling drives out hot. Recomendation is to down the server.
Mark your drives and slots, mark you cables to the raid, I use "whiteout" (non conductive) for the cables. When you have problem at 2 am, this will keep you from inserting drives into the wrong slots (disaster) or reversing the cables (disaster)
