I have done many diags and found the following:
Samsung EVO Pro 250gb is not a suitable ssd for raid 1 on these chipsets.
Daily I did verification checks and came up with less than 100 errors on the raid 1. No matter what caching was changed the errors continued. This issue is more than likely a driver issue and how it controls the ram on the drives. Worse still is that when a disk was disconnected the system would hang due to there still being inconsistencies between the 2 ssd's
I attached 2x 250gb Plextor SSD's and no longer have any verification errors on the Raid 1.Broke the array and the system still runs.
What is dissapointing is that my Raid 5 volume of 3 seagate 2tb mechanical disks continues to have less than 50 verification errors. This should not be the case as the same disks are used at many of my clients sites with different chipsets with no errors.
Like the days before IRST where disks would continuously fall offline I would expect intel to formulate better drivers to resolve such issues. The older chipsets worked as do I expect the new chipsets to work with stable drivers.