Drive is not certified on Dell machine in RAID setup

It means the drive isn't one of the models blessed by that particular software's "recommended hardware" list. Dell would rather you purchase the hardware directly from them rather than at somewhat of a markup.

I wouldn't worry about it, so long as the drives are working.

Ok. But since the hadware is bought from Dell it should be certified for that RAID controller?

It's possible Dell sent you a new drive which isn't in the list of recommended hardware that the software checks it against. You're right though if you bought it directly from Dell it should show up as certified. I don't think it's going to affect the performance of the drive, but it might be worth a phone call to them to see if they have an explanation. Perhaps a software update to the raid software will resolve it.

I am sorry, but Frosty's answer is grossly incorrect. You risk DATA LOSS.  
Dell uses disks with specialized firmware that changes the timing for error recovery.   Everything will work fine UNTIL your disk goes into deep recovery.  The PERC controller allows approx 7 seconds for a disk to return from deep recovery until it thinks the disk failed.   Consumer-class disks can easily take more than 30 seconds.  The enterprise class disks with Dell firmware return in typically 3 seconds.

So what is the suggestion? Do I need firmware update on the disks?

You can not obtain a firmware update which will do the job you need.  The firmware on a disk will reject a new image that has a different vendor/product ID.  

Now you CAN purchase an enterprise-class disk from seagate and you will be fine.  The LSI controllers that are rebranded as PERC/6s are qualified on them.  Go here, and choose any disk with proper physical dimensions that says the disk is OK to use behind a RAID controller and you will be fine.

http://www.seagate.com/www/en-us/products/enterprise

But since this machine is bought from Dell and they made all the configuration they should obviously have either told me this or configured the system with discs that where certified? Specially if you mean that using disks that are not certified can result in data loss.

And you say it works fine "UNTIL your disk goes into deep recovery". What you mean by this? When does the disks go into deep recovery and will they both do that at the same time? There is no problem if only one disk fail so I can manage to replace that one before the other does the same.

P.S.  It is perfectly OK to use the 'cuda disks on those controllers IF you configure them as non-RAID (JBOD) disks.  Doing so disables the error recovery aspect of the firmware and makes them "safe".   You could even use windows-based mirroring and then you would also be safe.

Note, you just can't disable the RAID and switch to software-based raid w/o a full backup, reconfigure, restore.  If you attempt to do that in-place, then you will suffer 100% data loss.

Thanks for excellent info. Really liked your article about "Disk drive reliability overview" !

I doubt Dell will replace disk for some certified ones if they cost 2-3 times more. I called support and they suspected that it had to be something wrong with one of the disks and said I had to run diagnostics, with no results of course.

I'm not going to buy new disks as this is a workstation with no cirtical data and the RAID setup on D:\ was a "nice to have feature". The employee is responsible to save all work on the server and not just locally. But as always this does not always happens so I thought that a RAID setup on project work would give an extra security.

The disk are empty at the moment (as the OS and programs are running from the SSD) so I can change whatevere I without data loss. What would be the best set up in my instance? You say I might can configure them as non-RAID (JBOD) disks, how do I do this? The disk name is "ST31500341AS" with firmware CC4G.

If not is it safer to keep them as they are in RAID1 or rather use just one disk without the RAID controller and save the RAID controller for something else with certified disks?

Just go to the BIOS of the controller, tell it that the 2 disks are to be presented as non-RAID (i.e, JBOD).   Boot, and windows will see the 2 individual drives.  Go to disk administrator, highlight both of them, and configure as RAID-1.

That will work just fine.

(The reason diagnostics pass is that there is nothing wrong with the disks.  It is an incompatibility/timing issue.  This is what throws people off.   Separately everything works fine, but together, they don't.  If you had the right software you could easily reproduce this problem by injecting errors in the right locations then just issuing reads.

it is NOT safer to keep them in RAID1 when the CONTROLLER does the RAID1.  The correct, approved, and best-practice is to let windows and ONLY windows do the RAID1.  Windows7 has a quite robust and well-written RAID1 design that works just fine with consumer class disks.

PowerEdgeTech is right, but in this case the single-disk RAID0 is acceptable and is effectively JBOD.  It carves out some metadata, but the root cause, the issue with recovery will be out of the picture.