Configure PM8222 RAID in the UEFI Firmware Utility

This archived OEM guide covers array creation, deletion, controller modes, spares, boot selection, controller properties and RAID 5 fault handling. All original English figures are retained unchanged. They include different controller/firmware examples and separate configuration cases, rather than one continuous verified installation. Protect required data and identify the exact controller, array and disks before a destructive operation. No BIOS, RAID or disk operation was executed while preparing this article.

1. RAID controller configuration

1.1 Open the UEFI controller utility

Press Del at the server logo to enter BIOS. In Advanced, select the entry naming the PM8222 controller. The original figure labels it INSPUR PM 8222-SHBA:

The main menu provides Controller Information, Configure Controller Settings, Array Configuration, Disk Utilities, Set Bootable Device(s) for Legacy Boot Mode and Administration. Using the UEFI configuration UI does not turn the separately labeled Legacy boot-device menu into UEFI OS boot order.

1.2 Inspect array and logical-drive status

Open Array Configuration > Manage Arrays > the intended array (Array A in the example) > List Logical Drives. Select the logical drive and open Logical Drive Details. The adjacent options include Edit Logical Drive and Delete Logical Drive:

Parameter	Meaning / recorded values
Status	Ok: healthy logical drive with its members online. Failed: failed/offline membership exceeds that layout’s tolerance. Interim Recovery: degraded operation requiring fault investigation and supported replacement/rebuild; remaining data availability is conditional.
Size / RAID Level	Recorded logical-drive capacity and layout. The example is RAID 1, 7.2 TiB (8 TB), with two SATA members.
Legacy Disk Geometry (C/H/S)	Legacy cylinders/heads/sectors geometry; source example 65535/255/32.
Strip Size / Full Stripe Size	Per-member and full data stripe, 256 KiB / 256 KiB in this RAID 1 example.
Number of Parity Groups (NPG)	Parity subgroups where applicable; for example, RAID 50 with two RAID 5 subgroups has NPG 2.
Logical Drive Label	User-defined logical-drive label.
Acceleration Method	Supported logical-drive acceleration policy; this example None.
Logical Drive Members	Member identity, port/box/bay, capacity and status; use these to correlate the actual disks.

1.3 Delete the selected array

Choose Array Configuration > Manage Arrays > the intended array > Delete Array. The next page warns that data on its logical drives will be lost. It also warns that deleting the only array restores default controller configuration. Submit Changes applies this destructive choice; check the exact target and required backups first:

The source says deletion succeeds, but the final supplied figure is the confirmation page, not an independently captured post-deletion inventory.

1.4 Create RAID 0

Choose Array Configuration > Create Array. Enable only the intended disks, select Proceed to next Form, set RAID Level to RAID 0, proceed again, then review label, strip/full stripe, size/unit and supported acceleration before Submit Changes:

The source enables CN0 / Box 1 / Bay 2 and leaves Bay 3 disabled. Its example logical drive is 7.277 TiB with 256 KiB/256 KiB and None. The final page reports Logical Drive Creation Successful. That records a controller configuration result, not OS installation or a data-integrity test. RAID 0 has no member-failure redundancy.

Parameter	Meaning / recorded values
RAID Level	Redundancy/striping level for the logical drive; choose only a level supported by the installed controller and selected members.
Logical Drive Name / Label	User-defined logical-drive name.
Strip Size / Full Stripe Size	Strip size is the per-member stripe unit; full stripe size is the combined data stripe across the logical drive. The source lists 16, 32, 64, 128, 256, 512 and 1024 KiB, with 256 KiB as its default. The utility calculates full stripe size after the strip size and layout are selected.
Size / Unit Size	Logical-drive capacity and its selected units. The source uses the maximum by default; allocate less if additional logical drives are to share the array. Preserve TiB/GiB versus decimal TB/GB as shown.
Acceleration Method	The source describes PM8222 as lacking an adjustable cache-acceleration policy; the shown value is None. This is not a statement that every controller named in the article has no memory or cache.

1.5 Create RAID 1

Repeat Create Array, enable the intended two members and select RAID 1. The example enables CN0 / Box 1 / Bays 0 and 1, with Bays 2 and 3 disabled. Review the same fields and Submit Changes:

The source shows 7.277 TiB, 256 KiB/256 KiB, None and the creation-success page.

1.6 Create RAID 5

Repeat Create Array, enable the three intended members and select RAID 5. The source enables Bays 0–2, leaves Bay 3 disabled, then sets the desired label, strip/full stripe, size/unit, initialization and acceleration before Submit Changes:

The example shows 14.554 TiB, 256 KiB/512 KiB, Parity Initialization Method Default and acceleration None, followed by the creation-success page. Default parity initialization can continue in the background; that page is not evidence that every initialization or subsequent workload test completed.

1.7 Create RAID 10

Enable all four intended members in Create Array and select the utility’s RAID 1+0 option. Review the same fields and submit the intended settings:

The example uses Bays 0–3, 14.554 TiB, 256 KiB/512 KiB and acceleration None, followed by the creation-success page. It is a separate configuration case from the RAID 0/1/5 examples.

1.8 Configure pass-through / JBOD modes for the actual backplane

The source names PM8204 when describing these modes even though this article’s main controller is PM8222. Verify support on the exact card and firmware; these mode descriptions do not identify every figure as PM8204:

Parameter	Meaning / recorded values
RAID	Expose configured RAID logical drives to the OS, rather than unconfigured pass-through disks.
Mixed	Expose both configured RAID logical drives and eligible pass-through disks.
HBA	Expose physical disks as pass-through devices and disable array creation. Existing arrays can prevent selecting this mode.

For the direct-attach backplane case, the source uses Configure Controller Settings > Modify Controller Settings, where Port CN0 Mode and Port CN1 Mode are visible:

Mixed permits RAID logical drives alongside pass-through disks; RAID does not expose unconfigured pass-through disks; HBA exposes physical disks without array creation. For an expander-backed case, the source instead uses Configure Controller Port Mode:

The newer Aptio 2.22.1282 / 2023 examples offer RAID/Mixed while an array exists, and also HBA in a later list. The source attributes that availability change to removing all arrays; it is not permission to delete data-bearing arrays just to enable HBA. The older 2.20.1275 direct-attach pictures and newer examples are different platform/firmware views. Ports attached to the same expander/multi-domain device require consistent modes; independently direct-attached paths can allow different modes, subject to the actual platform.

1.9 Assign, change and remove spare drives

Open Array Configuration > Manage Arrays, select the array to protect, and Manage Spare Drives. Select Assign Dedicated Spare or Assign Auto Replace Spare, enable the eligible intended disk, and submit the corresponding assignment. The example enables Bay 2 and leaves Bay 3 disabled:

Use the utility’s actual spare-type names. The source equates Dedicated with global and Auto Replace with local, which is misleading. A dedicated spare protects its assigned array or arrays; it is not automatically available to every array. An auto-replace spare is assigned to a particular array and becomes a permanent member after rebuilding a failed member. Confirm the actual associations and controller support. The Microchip maxView guide, sections 6.3–6.4, documents these meanings; its desktop wizard is a terminology reference, not a replacement for these OEM firmware menus.

The source’s result page reports Adding of Spare Successful. In Manage Spare Drives, Change Spare type to AutoReplace changes the association behavior; the following page offers changing back to Dedicated. Those choices are not merely renaming a global spare to local:

To remove an intended spare association, select Delete Spare Drives, enable the intended spare and submit that operation. The last source figure shows Bay 2 selected for deletion; it does not show the separate completed-removal result:

1.10 Set or clear Legacy boot-device preferences

Open Set Bootable Device(s) for Legacy Boot Mode. Select Bootable Logical Drive chooses a logical drive; Select Bootable Physical Drive chooses an eligible physical disk. Choose the intended drive, then Set as Primary Bootable Device or Set as Secondary Bootable Device:

The recorded success page requests a reboot to apply the change. Schedule that restart for the actual system; none was performed here. This menu applies to Legacy boot preferences even though it is displayed inside the UEFI setup utility.

Clear Bootable Device(s) lets you clear Primary, Secondary or All Bootable Devices, then confirm the selected clearing operation:

Clearing a preference is distinct from deleting the array or erasing its disks; still verify the intended boot path before changing it.

1.11 Controller information and settings

Parameter	Meaning / recorded values
Controller / Device ID / PCI Address	Model, device ID and assigned PCI address; the source shows INSPUR PM 8222-SHBA and device ID 28F.
Hardware Revision / Serial Number / WWID	Hardware revision and unique controller identifiers. WWID means World Wide Identifier. Do not infer another controller’s identity from this example.
Firmware version / release date	The recorded controller has firmware 1.98, dated Dec 20 2018 18:54:04.
UEFI Driver version / release date	The recorded driver is 1.3.5.2, dated Nov 27 2018 16:50:03.
Temperature	Controller temperature where reported. The source prose quotes 80 °C as abnormal and 105 °C as an upper limit, but this crop does not verify those limits. Use the exact card’s hardware/thermal specification and fault status, not these prose values as a universal operating target.
Controller Memory Module Size	The actual crop reports 128 MiB. The source elsewhere says PM8222 has no cache-acceleration policy; distinguish that from the visible controller-memory field. Memory presence does not establish protected write-back cache or adjustable acceleration.
Controller Mode / Supported Modes	Current and supported RAID, Mixed or HBA operation. Port CN0/CN1 capabilities and visible options depend on controller, firmware, existing arrays and backplane.

Parameter	Meaning / recorded values
Modify Controller Settings	Supported controller policies, including physical-drive cache and per-port modes in these figures.
Clear Configuration	Destructive array/configuration and on-disk metadata removal.
Manage Power Settings	Power/performance and survival-mode policy.
BMC Settings	Controller management-bus address settings.
Configure Controller Port Mode	Supported RAID/Mixed/HBA controller/port mode.

The main menu’s Administration entry also groups functions such as firmware update, log collection and restoration of defaults. Those operations are not walked through or performed in this article.

Modify Controller Settings

Parameter	Meaning / recorded values
Unconfigured Physical Drive Write Cache State	Write-cache policy for non-array physical disks in RAID/Mixed mode: Default retains the device default, Enabled enables it, Disabled disables it.
HBA Physical Drive Write Cache State	Corresponding write-cache policy for HBA disks: Default / Enabled / Disabled.
Configured Physical Drive Write Cache State	Corresponding write-cache policy for array members in RAID/Mixed mode: Default / Enabled / Disabled.
Port CN0 Mode / Port CN1 Mode	Per-connector RAID/Mixed/HBA settings when the platform exposes them. Same-expander connections require consistent mode; direct-attach support can differ.
Submit Changes	Apply only the reviewed intended changes.

The older example shows unconfigured/configured cache Enabled, HBA cache Disabled and both ports Mixed. The newer earlier comparison picture shows all three cache settings Default and both ports RAID. These are different recorded choices, not one required default. The source UI itself warns that drive write cache can lose cached data on sudden power loss; check the actual protection and supported policy before enabling it.

Clear Configuration

Parameter	Meaning / recorded values
Delete All Array Configurations	Delete arrays/logical drives and reset the relevant configuration. Existing logical-drive data will be lost/inaccessible.
Delete configuration metadata on all physical drives	Remove applicable on-disk RAID metadata, including residual metadata on disks outside active arrays. Do not use it on a recovery set as a diagnostic shortcut.

Manage Power Settings

Parameter	Meaning / recorded values
Power Mode	Minimum Power lowers static settings, Balanced uses configuration-dependent settings, Maximum Performance favors maximum settings. The source example/default is Maximum Performance.
Survival Mode	Enabled/Disabled. When supported and triggered by temperature, the enabled policy can reduce power/performance to maintain operation. The UEFI example is Enabled, unlike the separate Legacy guide’s Disabled picture. It does not establish safe operation above the exact hardware’s limits.
Submit Changes	Apply the reviewed policy; a reboot may be requested to optimize the selected power mode, as the UI help notes.

BMC Settings

The field is an 8-bit hexadecimal I2C slave address, with UI help listing 0xD0–0xFF; the example is de (0xde). The source describes the encoded address byte as seven address bits plus a read/write bit. Do not confuse it with a seven-bit address value or an Ethernet address. Retain the documented platform default unless the applicable procedure requires a change; Submit Changes applies an intended change.

Configure Controller Port Mode

The final example has current and requested mode Mixed. The source calls Mixed the default, but availability and defaults must be checked on the actual controller. Existing RAID arrays can block HBA selection. Submit Changes applies a reviewed supported transition; it is not a way to preserve arrays while bypassing that restriction.

2. Fault recovery and limits

2.1 One failed RAID 5 member

The source describes RAID 5 entering a degraded state after a single member goes offline and rebuilding after an appropriate healthy replacement is recognized. Inspect the fault, connection, exact member and current redundancy before replacement; reseating a drive is not a universal first action. Monitor the actual rebuild and final logical-drive state. Remaining accessibility and successful recovery depend on the other members and the fault; no fault/rebuild result screenshots are supplied in this section.

2.2 Multiple failed RAID 5 members

The source describes two or more failed members in the same RAID 5 set causing an offline/unavailable logical drive. Preserve the fault evidence and obtain qualified recovery assistance if required; do not force a failed array online, initialize it, clear metadata or recreate its layout as a recovery shortcut. A spare does not increase the RAID level’s simultaneous-failure tolerance.

References

• https://www.ieisystem.com/qa/raid/14541.html
• https://ww1.microchip.com/downloads/aemDocuments/documents/adaptec/documents/msm_smart_7_24_users_guide.pdf

The original OEM article is the source for its menu path, examples and model-specific claims. The linked English Microchip materials support terminology and related behavior; they cover different product/management versions and are not a claim of exact PM8222/PM8204 firmware equivalence. Check the installed OEM version’s documentation for actual support and defaults.