Onboard Intel VROC SATA/sSATA RAID in UEFI: Arrays, Spares and Rebuilds
These archived examples use a server BIOS UEFI configuration interface. The original screenshots are already in English and remain unchanged. Confirm the actual controller, drive mapping, firmware, boot mode and supported OS driver. Changing boot/storage modes or entering setup through a restart can interrupt an installed system.
Creation and deletion can destroy existing data; identify the intended available drives and protect required data first. The article contains separate examples, rather than one uninterrupted sequence in which all arrays coexist. A controller’s successful-operation or Normal/Optimal state alone does not verify completed initialization, filesystem access, OS boot or backups.
1. Enable the appropriate UEFI and SATA RAID configuration
Open Advanced → CSM Configuration. The source keeps CSM Support Enabled while setting Boot Mode to UEFI Mode and Network, Storage, Video OPROM Policy and Other PCI devices to UEFI. Do not interpret the screenshot as disabling CSM entirely:
Under Chipset, open PCH SATA Configuration and PCH sSATA Configuration for the controllers being used. The source selects RAID in SATA Mode Options and sSATA Mode Options, then saves and restarts before reentering BIOS. Do not change a working installation’s storage mode without its matching driver and boot plan:
2. View a volume
Under Advanced, choose the appropriate Intel VROC SATA Controller or Intel VROC sSATA Controller for the connected disks. This is the onboard SATA/sSATA path, separate from VMD NVMe RAID. Select the intended volume to view its details and members:
The detail example is Volume0 RAID0, 128 KB, 14158.86 GB, Normal, Bootable Yes and 512-byte blocks. Members are ports 0 and 1, each a 7452.04 GB ST8000NM0055-1RM112. These are the captured configuration values, not a promise of identical usable capacity on every pair of drives.
3. Delete the verified volume
Open the controller, select the target volume and choose Delete. The source’s target is the preceding RAID0 case. Its confirmation warns ALL DATA ON VOLUME WILL BE LOST and says deleting resets the disks to non-RAID. Choose Yes only for the intended destructive operation:
4. Create RAID0
Open the correct controller and choose Create RAID Volume. Select RAID0(Stripe), then mark the intended available disks with X. This onboard implementation requires at least two members for the illustrated RAID0; it does not provide the source’s desired single-disk RAID0 case:
The source selects ports 0 and 1, leaving other disks unselected. Review the complete form: it shows 128 KB strips and 14158.86 GB. Choose Create Volume only after verifying the final level, members and capacity. This branch ends on that command, without a separate new RAID0 result capture:
5. Create RAID1
Repeat the controller → Create RAID Volume path, select RAID1(Mirror) and mark the two intended drives. The example again chooses ports 0 and 1, each 7452.04 GB; these are separate array examples, not simultaneous use of those members by every level:
Review the 7079.42 GB capacity and choose Create Volume. The following list shows Volume0 RAID1(Mirror), 7079.42 GB, Normal, with the other drives still non-RAID:
6. Create RAID5
Repeat the controller/Create RAID Volume navigation. The open RAID-level menu picture still has an intermediate RAID10 field and Volume1 name; the final RAID5 form instead shows Volume0 and three X marks on ports 0, 1 and 2. Use the final form rather than combining intermediate selections:
The final settings show 64 KB strips, 14158.86 GB and RWH Policy Disable. Do not describe that policy as enabled write-hole protection. Choose Create Volume when the intended settings and data disposition are confirmed. The result shows Intel VROC SATA Driver 6.2.0.1034 and Volume0 RAID5(Parity), 14158.86 GB, Normal:
7. Create RAID10
Open the same controller/Create RAID Volume path and select RAID10(RAID0+1), as this firmware labels the menu option. That literal UI label does not establish identical failure behavior for every RAID10 and RAID0+1 topology. The final form selects four members on ports 0 through 3:
Review 64 KB strips and 14158.86 GB, then choose Create Volume. The result shows Volume0 RAID10(RAID0+1), 14158.86 GB, Normal. Its capacity matches the RAID0/RAID5 examples but its level and member set differ:
8. Mark the intended spare
The spare branch returns to a RAID5 Normal case. Select the available non-RAID drive on port 3, ST8000NM0055-1RM112, 7452.04 GB, then Mark as Spare. Do not select the separate 447.13 GB Intel SSD merely because it is also unassigned:
The confirmation explicitly says marking the disk as a spare removes all data on it. Choose Yes only for the verified suitable spare. The source’s last picture is this confirmation, without a separate post-confirmation Spare-state view:
9. A member goes offline
The source’s 14158.86 GB RAID5 example changes to Degraded. After the supported replacement/spare process it shows Rebuilding, then Normal in a later capture. Rebuilding is still in progress; verify completion and actual data access before considering recovery finished. This redundant-array example is not a rule that a failed RAID0 member can be rebuilt:
10. Multiple-member failure
The final image shows the RAID5 volume Failed. Preserve the drive identities and failure history. For required data, use a verified backup or obtain qualified recovery support before new writes, initialization or configuration clearing. An offline volume cannot simply provide an ordinary backup merely because the source says to seek a recovery company:
