Fisrtly, the mdadm warning you got should show you which disc(s) were degraded. Could you cut-and-paste that into your question, along with your current
Secondly, write errors to an md device should be logged under syslog. Can you find anything with
grep sda /var/log/messages? You will probably need to run through
hd[a-d] in the grep to be sure to catch all likely discs.
smartctl -a /dev/sda should give you health check information on /dev/sda, and similarly for your other HDDs, if they're SMART-aware, as most modern drives are.
If you can't get something out of that, it's probably not very failing!
Later edit: Marcog, sorry, my bad, I missed the "during boot" bit about your warning. I really should read more carefully. I agree with you that it sounds like a disc not being correctly detected. I do recommend the
smartctl route, though; it can be used with
-t to force one of a suite of tests (see the man page), many of which are suitable for use on a live disc (contains data and is mounted). I do these every few months just to try to avoid unpremeditated failures