Not entirely. The D800 buffer design seems to work best with class 10 cards. I haven't tried mixing card classes - e.g., class 10 in one slot, class 6 in the other. Perhaps the problem you're experiencing is related to that - an empty slot misread by the camera? That would definitely be a bug. It's almost as if the camera gets periodically confused by the state of the unused slot and then slows down.
>Class 10 only means that card is capable of a minimum 10mb/sec >writing speed. But each manufacturer might be rating their >cards on top of this CLASS rating.
That's exactly what they're doing. The theoretical SD cardbus max is 310 MB/s, and once UHS-1 standards are settled we'll start seeing some really fast cards and camera slots that can handle the data. It's probably a D800 bug, and I hate to bring it up, but the issue of fake Sandisk and fake Lexar CF and SD cards has been raised in other threads over the past couple of years. But if you're sure your main card is genuine, a slow or poorly chipped fake isn't the problem.
>BTW, one of the posters here has 90mb/sec CF and 95mb/sec SD >with the same exact problem.
Problems occur, no doubt. Process of elimination should include confirming a genuine card as well as the rated write speed I think.
>So it is obvious to me that there is some kind of the bug. My >camera is one of the first ones and it was purchased in April >of last year.
Before sending the camera to Nikon for testing, maybe consider taking the camera back to your dealer (or the nearest dealer) and trying a couple of identical class 10/speed cards - CF and SD. Shoot in store for as long as it takes to either reproduce the problem or at produce some evidence that filling both slots with the same class/speed cards may eliminate the problem.