What you're describing is something of a "serialistic support nightmare".
Just reality when you're doing such stuff for a living
My conclusion: Stop dealing with this part directly. We cooperate now with partners that sell a lot more RAIDs and focus on three vendors. One relies on Infortrend (and EasyNAS in the meantime), the other uses stock SuperMicro gear combined with Open-E (this rocks, ZFS on Linux clusters, somewhat expensive configs with zmirrors shared accross at least two cluster nodes but insanely high IO performance, both sequential and especially random IO).
These guys ran in all possibles troubles already somewhere else and are the ones who tell you stuff like 'don't use these HDDs with this firmware version unless you adjust this hidden setting on the controller to this undocumented value since otherwise everything will fail after n hours of operation'.
IMO you need either a lot of experiences with the RAID implementation you use (includes hardware and software) or you test through every situation you can imagine (most probably you can not as novice). Or you skip the whole idea and focus on backup strategies instead
And again: it's only about availability so unless you run a business or have to serve small children the redundant RAID modes are useless anyway