@Carl: it does indeed, and using standard RAID types as well (the one I did it on was running RAID5). As I say, it was fairly painless, though time-consuming, and the added space isn't available until all drives have been swapped. In fact, it's documented here.
Pretty sure my Synology at home will do it when using SHR ("Synology Hybrid RAID", similar to RAID5 or 6, but with their own twist), although I've never looked to see if it will do it under standard RAID5/6... and I would think something with its own proprietary RAID system, like Drobo, would as well. Edit: in fact, Synology documents the procedure as well.
Beyond that... no idea. I don't suppose it's that complicated a thing: if a system gives you the ability to create an array smaller than the maximum size of available disks, AND the ability to expand that array later, then it should be a simple matter to swap disk, rebuild, swap disk, rebuild, etc. etc. and then once all are swapped, expand the array into the new available space.
@Tiago: in one case, the unit behaved like it had power supply issues; upon sending it in, they said its firmware was corrupted, so they re-flashed it, which fixed it for a while... although it still had intermittent odd problems after that, that still felt like power issues to me. The painful part was that it took weeks of back-and-forth to even get to the point of RMA, usually because responses from them took hours or days... NOT good when one is trying to get help while on-site.
In two or three other cases, it's been a matter of the system burping on one drive in a RAID5 array, then on another before the array was completely rebuilt... granted, on at least one of those, it was running "green" drives as a cost-saving measure (against my specific objections)... since then the customer has also learned to eat the extra cost of using RAID6+spare to avoid such major problems caused by minor glitches.
In at least one of those cases, a QNAP tech was able to remote into the system and get the data back... but again, it was like pulling teeth to get to that point, and only happened after I'd left site.
As far as a metric for "good support" I'd say the main one is response time and accessibility... and this applies to support for ANYTHING!!! If I get a call to a site, find a problem, and need to contact support... I can't be sitting there for hours at $90 per waiting for someone to call me back. A phone number for contact (toll-free, ideally, of course) or fast email contact is crucial, not some web form whose submission goes into the ether with no indication of when you'll hear back.
I'll forgive number of attempts and even ability to solve, in most cases, because I know all too well, sometimes issues don't fully present themselves, or don't make themselves obvious. Some are intermittent, and of course, work fine when you DO get the support guy online. Heck, I'm sure we all regularly have instances where we get a problem call, get onsite, and find everything working fine, only to have it go on the fritz again five minutes after we've left.