As the Intel Rapid Storage Technology Enterprise GUI doesn't provide itself any detailed information about drives' health (e.g. S.M.A.R.T attributes). I have to use the external tools (e.g. SSD Toolbox, smartmontools, CrystalDiskoInfo etc.). There is no problem with the SATA based drives connected to PCH (even they are the RAID members) as such diagnostic data (including S.M.A.R.T attributes) are provided via the Common Storage Management Interface (CSMI) by Intel device driver.
However I've encountered with the problem in case of using NVMe SSD drives. Well, if such drives aren't members of RAID (i.e. VROC) the device driver provides the diagnostic data via the NVME Pass-Through protocol without problems. But if the NVMe drives are the members of VROC RAID (RAID1), the device driver supports the NVMe Pass-Through protocol for the first NVMe drive only (i.e. VMD port 3:0) - it behaves like the second drive doesn't exist at all! If I swap the NVMe drives in the M.2 slots so the previous second drive (i.e. in the VMD port 3:1 originally) has become the first drive (i.e. in the VMD port 3:0 now) the device driver again supports the NVMe Pass-Through protocol for such swapped NVMe drive connected to VMD port 3:0 only. So the problem is not caused by NVMe drives themselves.
I've done another test - On the same machine (based on C422/Basin Falls platform) I've tried to install RedHat Enterprise Linux (RHEL) 7.5 which has the driver support for VROC. I've realised that there is no problem in the RHEL environment, actually! Both NVMe drives' diagnostic data are provided corectly via the NVMe Pass-Through protocol (using several tools like smartmontools, nvme-cli etc.). So the problem is not hardware based but it seems that's the issue of iRSTe/VROC driver for Windows. I've tested the latest version of iRSTe (v. 126.96.36.1990) both in Windows 7 Enterprise x64 and Windows 10 Enterprise x64 but the issue mentioned above still persists :-(
As the diagnostic data of the NVMe drives (like Temperature, Total Bytes Written, Media Wearout etc.) are vital for the drive's health management I would like to ask you very much to check and fix the problem of the NVMe Pass-Through protocol to allow reporting the diagnostic data correctly for the all VROC RAID members (not only the first one).
Thank you very much