Huawei V2 & V3 Server RAID Controller User Guide

RAID Card Type

Introduces RAID card types and the meaning of their naming conventions.

The following archived source catalogue lists RAID cards for the V2/V3-era guide; it is not a current support matrix. The subsequent compatibility and specification tables retain their original model columns, part numbers and footnotes.Table 1-1.

Different RAID card models support different servers and operating systems. For detailed information, please use theCompute Product Compatibility Query Assistant to obtain it.

Table 1-1 RAID Card Basic Parameters

Chip Type	Code	Requires RAID Key	Description
LSI SAS2208	02310QNM	Yes	RU320BC,LSI 2208 RAID Card,RAID0,1,10,5,50,6,60,512MB,16HDD,Board ID 0X22,Support Battery and SuperCap
	02310QNK	Yes	RU320BC,LSI 2208 RAID Card,RAID0,1,10,5,50,6,60,1GB,16HDD,Board ID 0X2b,Support Battery and SuperCap
	02310QNL	Yes	SR320BC Server Raid Controller - LSI 2208 - SAS/SATA 6G - RAID0,1,10,5,50,6,60 - 512MB - 16 HDD - Board ID 0X22-Support Battery and SuperCap
	02310QNJ	Yes	SR320BC,LSI 2208 RAID Card,RAID0,1,10,5,50,6,60,1GB,16 HDD,Board ID 0X2b,Support Battery and SuperCap
	02310SRU	Yes	SR420BC,LSI 2208 RAID Card,RAID0,1,10,5,50,6,60,1GB,32 HDD,Board ID 0X2b,Support Battery and SuperCap,All
	02310NKT	Yes	SR520C,LSI 2208 RAID Card,RAID0,1,10,5,50,6,60,1GB,32 HDD,CacheCade2.0,Board ID 0X2b,Only Support SuperCap,All
LSI SAS2308	03021ENL	No	SR120 SAS/SATA RAID Card,RAID0,1,1E,10,0 Cache(LSI2308),Board ID 0X21,2*2
	03021FTX	No	RU120 SAS/SATA RAID Card,RAID0,1,10,1E,0 Cache(LSI2308),Board ID 0X21,2*2
LSI SAS3008	03022CDE	No	SR130 SAS/SATA RAID Card,RAID0,1,10,1E,0 Cache(LSI3008),Board ID 0X2C,2*2
	03022QKU	No	RU130 SAS/SATA RAID Card,RAID0,1,10,1E,0 Cache(LSI3008),Board ID 0X2C,2*2
	03023QGH	No	RH2288 V3,BC61ESMNA,SR130-M(LSI3008),SAS/SATA HBA Card,Unsupport RAID,Board ID 0X29,IT work pattern, Support Out-of-Band Management,2*2
	03024JXJ	No	SR130 SAS/SATA RAID Card, RAID0, 1, 10, 1E, 0 Cache (LSI3008)-Board ID 0X27, customraid-2*2
	03023YLD	No	E9000,BC01ESMNB,RU130-iMR SAS/SATA HBA Card,Sideband Management Supported,0 Cache(LSI3008),Board ID 0X2E,2*2
LSI SAS3108	02310UUB	Yes	SR430C 1GB,BC1M05ESMQ,LSI 3108 RAID Card,RAID0,1,5,6,10,50,60,1GB,32HDD,Board ID 0X24,Support SuperCap
	02310UUA	Yes	SR430C 2GB,BC1M05ESML,LSI 3108 RAID Card,RAID0,1,5,6,10,50,60,2GB,32HDD,Board ID 0X2D,Support SuperCap
	02311GSF	Yes	SR630C 1GB,BC1M10ESMQ,LSI3108 RAID card-Board ID 0X24-RAID0,1,5,6,10,50,60-Support SuperCap(greater than 32 HDDs Chassis)
	02311GSG	Yes	SR630C 2GB,BC1M10ESML,LSI 3108 RAID CARD-Board ID 0X2d-RAID0,1,5,6,10,50,60-Support SuperCap(greater than 32 HDDs Chassis)
	02311DNY	Yes	RU430C 1GB,IT1M01ESMQ,LSI 3108 RAID Card,RAID0,1,5,6,10,50,60,1GB,32HDD,Board ID 0X24,Support SuperCap
	02311DPA	Yes	RU430C 2GB,IT1M01ESML,LSI 3108 RAID Card,RAID0,1,5,6,10,50,60,2GB,32HDD,Board ID 0X2D,Support SuperCap
	02311NLC	Yes	SR530C 1GB,BC1M12ESMQ,LSI 3108 RAID Card,Cachecade Key,RAID0,1,5,6,10,50,60,1GB,240HDD,Board ID 0X24,Support SuperCap
	02311NHK	Yes	SR530C 2GB,BC1M05ESMM,LSI 3108 RAID Card,Cachecade Key,RAID0,1,5,6,10,50,60,2GB,240HDD,Board ID 0X2D,Support SuperCap
	02311PTX	Yes	RU530C 1GB,IT1M07ESMQ,LSI 3108 RAID Card,CacheCade,RAID0,1,5,6,10,50,60,1GB,240HDD,Board ID 0X24,Support SuperCap
	02311PTY	Yes	RU530C 2GB,IT1M08ESML,LSI 3108 RAID Card,CacheCade,RAID0,1,5,6,10,50,60,2GB,240HDD,Board ID 0X2D,Support SuperCap
	02311PCJ	Yes	SR430C-M 1G,BC1M15ESMQ,LSI 3108 RAID Card,RAID0,1,5,6,10,50,60,1GB,32HDD,Support SuperCap and Out-of-Band Management,Board ID 0X24
	02311PCH	Yes	SR430C-M 2G,BC1M16ESML,LSI 3108 RAID Card,RAID0,1,5,6,10,50,60,2GB,32HDD,Support SuperCap and Out-of-Band Management,Board ID 0X2D
	02311RST	Yes	SR630C 1GB,BC1M10ESMQ,LSI3108 RAID card-Board ID 0X24-RAID0,1,5,6,10,50,60-Support SuperCap and Sideband Management(greater than 32 HDDs Chassis)
	02311RSU	Yes	SR630C 2GB,BC1M10ESML,LSI 3108 RAID CARD-Board ID 0X2d-RAID0,1,5,6,10,50,60-2GB-Support SuperCap and Sideband Management(greater than 32 HDDs Chassis)
	02311RSW	Yes	SR530C 1GB,BC1M12ESMQ,LSI 3108 RAID Card,Cachecade,RAID0,1,5,6,10,50,60,1GB,240HDD,Board ID 0X24,Support SuperCap and Sideband Management
	02311RSX	Yes	SR530C 2GB,BC1M05ESMM,LSI 3108 RAID Card,Cachecade,RAID0,1,5,6,10,50,60,2GB,240HDD,Board ID 0X2D,Support SuperCap and Sideband Management
LSI SoftRAID	05200449	No	Embedded OS,Embedded MegaRAID Software RAID for BIOS,Paper LABEL,Intel Grantley,Embedded RAID label optionROM, support RAID0/1/10
PM8060	03022JWU	No	SR425C SAS/SATA RAID Card,RAID0,1,5,6,10,1E,50,60,1G Cache,Support SuperCap,Board ID 0X23,2*2
	03023KWH	No	RH2288 V3,CN21ITRA,SP435(PMC 8060 PCIE Raid Card),2GB Cache,SuperCap,PCIE 3.0 X8-Vendor ID 9005-Device ID 028D-1,1*2
PM8068	03023DLX	No	RH2288 V3,BC11ESMS,SR130 SAS/SATA HBA RAID (PMC 8068) PCIE x8 FASTENING CARD, NONE CACHE, SUPPORTING RAID0,1,10,5,Board ID 0X20,2*2

The meaning of each RAID card naming suffix is asTable 1-2 shown.

Table 1-2 Meaning of RAID Card Naming Suffixes

Suffix	Supports RAID Group Creation	Supports Out-of-Band Management	Has Capacitor/Battery	Has Cache
IT	No	Yes	No	No
IR	Yes	No	No	No
iMR	Yes	Yes	No	No
MR	Yes	Yes	Yes	Yes

Technical Specifications

Introduces technical parameters related to RAID cards.

Basic parameters for each RAID card are asTable 2-1 shown.

Table 2-1 Basic Parameters

Features	LSI SAS2208	LSI SAS2308	LSI SAS3008IR	LSI SAS3008IT	LSI SAS3108	LSI SoftRAID	PM8060	PM8068
Interface Speed (Gbit/s)	6	6	12	12	12	6 (SATA drives only)	12	12
Number of Supported RAID Groups	64	2	2	NA	64	8	64	64
Maximum Number of Drives per RAID Group	16/32b	10	10	NA	32/240a	6	128	128
Maximum Total Number of Drives Across All RAID Groupsd	16/32b	14	14	NA	32/240a	6	240	240
Support for Drive Passthrough	No	Yes	Yes	Yes	Yes	No	Yes	Yes
Number of Drives Supported in Drive Passthrough Mode	NA	1000	1000	1000	32/240a	NA	240	240
Maximum Number of Supported Drives	16/32b	1000	1000	1000	32/240a	6	240	240
Support for CacheCade	No	No	No	No	Yese	No	No	No
Adjustable Cache Read/Write Ratio	NA	NA	NA	NA	No	NA	NA	NA
Maximum Number of SSDs Supported by CacheCade	NA	NA	NA	NA	32	NA	NA	NA
CacheCade Maximum Capacity	NA	NA	NA	NA	512GB	NA	NA	NA
RAID Levels Supported by CacheCade	NA	NA	NA	NA	0/1	NA	NA	NA
Hot Spare Support for CacheCade	NA	NA	NA	NA	Yes	NA	NA	NA
Support for FastPath	Yes (Version 107 and later)	No	Yes	Yes	Yes	No	No	No
Supports 4K drives	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes
Supported RAID levels	0/1/10/5/50/6/60	0/1/10/1E	0/1/10/1E	NA	0/1/10/5/50/6/60	0/1/10/5 (RAID 5 supported after adding a license)	0/1/10/5/50/6/60	0/1/10/5
Supports setting stripe size	Yes (can be set to 8KB/16KB/32KB/64KB/128KB/256KB/512KB/1024KB, default is 256KB)	No (default is 64KB)	No (default is 64KB)	NA	Yes (can be set to 64KB/128KB/256KB/512KB/1024KB, default is 256KB)	No (default is 64KB)	Yes (can be set to 64KB/128KB/256KB/512KB/1024KB, default is 256KB)	Yes (can be set to 8KB/16KB/32KB/64KB/128KB/256KB/512KB/1024KB, default is 256KB)
Supports battery power-loss protection	Yes (cache data protection time is determined by the battery operating mode, default 24h.)	No	No	No	No	No	No	No
Supports supercapacitor power-loss protection	Yes	No	No	No	Yes	No	Yes	No
Support for Global Hot Spare	Yes	Yes	Yes	NA	Yes	Yes	Yes	Yes
Support for Local Hot Spare	Yes	No	No	NA	Yes	Yes	Yes	Yes
Support for Emergency Hot Spare	Yes	No	No	NA	Yes	NA	NA	NA
Number of Supported Hot Spares	16/32b	2	2	NA	32/240a	6	2	2
Support for Online RAID Level Migration	Yes	No	No	No	Yes	No	Yes	Yes
Support for Online Capacity Expansion	Yes	No	No	No	Yes	No	Yes	Yes
Supports Automatic SMART Scan	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes
Supports Patrol Read	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes
Supports Consistency Check	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes
Supports Crypto Encryption	No	No	No	No	No	No	No	No
Supports Out-of-Band Management	No	No	No	Yes	Yesc	No	No	Yes
Cache Support	Yes	No	No	NA	Yes (1G/2G/4G)	No	No	No
Beeper Support	No	No	No	No	No	No	No	No
a: LSI SAS3108 supports two RAID Key specifications, supporting 32 drives and 240 drives respectively. For specific specifications, please refer toSupports RAID 0/1/5/6/10/50/60 inTable 9-4. b: LSI SAS2208 supports two RAID Key specifications, supporting 16 drives and 32 drives respectively. For specific specifications, please refer toSupports RAID 0/1/5/6/10/50/60 inTable 5-3. c: When the PCB version of LSI SAS3108 is .B or later, out-of-band management is supported; when the PCB version is .A, out-of-band management is not supported. d: For LSI SAS2308/LSI SAS3008 RAID cards, this value is the sum of hot spare disks and hard disks added to the RAID group; for RAID cards other than LSI SAS2308/LSI SAS3008, this value is the sum of hot spare disks, unconfigured good disks, and hard disks added to the RAID group. e: Please useComputing Product Compatibility Query Assistant to query and obtain the codes of RAID controllers that support the CacheCade function.

"NA" in the table indicates not applicable.

The performance and disk utilization of various RAID levels supported by the RAID card are shown inTable 2-2.

Table 2-2 RAID Level Comparison

RAID Level	Reliability	Read Performance	Write Performance	Disk Utilization
RAID 0	Low	High	High	100%
RAID 1	High	High	Medium	50%
RAID 5	Medium	High	Medium	(N-1)/N
RAID 6	Medium	High	Medium	(N-2)/N
RAID 10	High	High	Medium	50%
RAID 1E	High	Medium	Medium	50%
RAID 50	High	High	Higher	(N-M)/N
RAID 60	High	High	Higher	(N-M*2)/N
Note: N is the number of RAID member disks, and M is the number of RAID sub-groups.

The hard disk interface types supported by each RAID card are as shown inTable 2-3.

Table 2-3 Hard Disk Interface Types Supported by RAID Cards

Hard Disk Interface Type	SAS HDD	SATA HDD	SAS SSD	SATA SSD	NVMe PCIe SSD	M.2 SATA SSD
LSI SAS2208	Supported	Supported	Supported	Supported	Not supported	Not supported
LSI SAS2308	Supported	Supported	Supported	Supported	Not supported	Not supported
LSI SAS3008IR	Supported	Supported	Supported	Supported	Not supported	Not supported
LSI SAS3008IT	Supported	Supported	Supported	Supported	Not supported	Not supported
LSI SAS3108	Supported	Supported	Supported	Supported	Not supported	Not supported
LSI Softraid	Not supported	Supported	Not supported	Supported	Not supported	Not supported
PM8060	Supported	Supported	Supported	Supported	Not supported	Not supported
PM8068	Supported	Supported	Supported	Supported	Not Supported	Not Supported

The hard disk models supported for configuration by each RAID are subject to the compatibility list. For detailed information, please use theCompute Product Compatibility Query Assistant.

RAID-related Features

Hard Disk Groups and Virtual Disks

With the increasing volume of business in modern data centers, the amount of data running on a single server is also growing daily. When a single physical hard disk cannot provide sufficient capacity and security to support system operations, multiple hard disks must be combined in a specific manner to function as a single visible hard disk externally, thereby meeting actual requirements. A hard disk group combines a set of physical hard disks into a single entity presented externally and serves as the foundation for virtual disks.

A virtual disk is a logical block device allocated from one or more physical-disk groups. Capacity and redundancy depend on its actual RAID level and members; creating a virtual disk does not inherently provide redundancy or guarantee more capacity than one physical disk.

A virtual disk can be:

• A complete hard disk group.
• Multiple complete hard disk groups.
• Part of a hard disk group.
• Parts of multiple hard disk groups (each hard disk group contributes a portion to jointly form the virtual disk).

In the following descriptions:

• A hard disk group is generally referred to as "Drive Group" (abbreviated as "DG"), "Array", or "RAID Group".
• A virtual disk is generally referred to as "Virtual Drive", "Virtual Disk" (abbreviated as "VD"), "Volume", or "Logical Device" (abbreviated as "LD").

Fault Tolerance

Fault tolerance ensures data integrity and data processing capability when a hard disk error or failure occurs within the subsystem. The RAID controller implements this feature through redundant hard disk groups on RAID 1, 5, 6, 10, 50, and 60.

A two-member RAID 1 can tolerate the loss of one mirror member; RAID 5 can tolerate one failed member and RAID 6 two within the same set. These limits assume the surviving data and required metadata remain usable; redundancy is not a backup or a guarantee against corruption.

In the illustrated conventional RAID 10 layout, at most one failed member per two-disk mirror can be tolerated. RAID 50 tolerates at most one failed member per RAID 5 subgroup, and RAID 60 at most two per RAID 6 subgroup. The total number alone is insufficient: exceeding any subgroup’s limit can make the full virtual drive unavailable.

RAID 0 does not support fault tolerance. If a hard disk fails in a RAID 0 configuration, the RAID becomes invalid, resulting in data loss.

Fault tolerance can maintain service within the actual RAID level’s failure limits. A degraded array has reduced protection; the state does not certify data integrity or successful recovery.

Consistency Check

For RAID levels with redundancy features (RAID 1, 5, 6, 10, 50, 60), the RAID controller can perform consistency checks on the hard disk data within RAID groups. It verifies and calculates the hard disk data and compares it with the corresponding redundant data. If any inconsistency is detected, the controller attempts automatic repair and logs the error information.

Since RAID 0 lacks redundancy, it does not support consistency checks.

The archived source recommends at least one consistency check per month. Treat scheduling and any automatic repair policy as workload/controller-dependent maintenance settings rather than a universal requirement; review the actual vendor procedure and data-protection plan.

Hot Spare

The hot spare feature of the RAID controller is implemented via hot spare drives and emergency backup functions.

Hot Spare Drive

A hot spare drive is an independent hard disk within the hard disk system. When a hard disk in a RAID group fails, the hot spare drive automatically replaces the failed disk as a member disk and reconstructs the failed disk's data onto the hot spare drive.

In the RAID controller management interface or command-line tool, you can designate an idle drive with capacity greater than or equal to the RAID group member drives, and with the same media type and interface as the member drives, as the hot spare drive for the RAID group.

The RAID controller supports the following types of hot spare drives:

• Global Hot Spare: Shared among all configured RAID groups on the RAID controller. One RAID controller can be configured with one or more global hot spare drives. When any member disk in any RAID group fails, the global hot spare drive can automatically replace it.
• Local Hot Spare: Dedicated to a specific RAID group on the RAID controller. Each RAID group can be configured with one or more local hot spare drives. When any member disk in the specified RAID group fails, the local hot spare drive can automatically replace it.

Hot spare drives have the following characteristics:

• Hot spare drives are used only for RAID groups with redundancy features, including RAID 1, 5, 6, 10, 50, and 60.
• Hot spare drives are used only to replace failed disks on the same RAID controller.

Emergency Backup

The emergency backup function refers to the scenario where a member disk in a RAID group with redundancy functionality fails and no dedicated hot spare is specified. In this case, an idle disk under the RAID controller will automatically replace the failed member disk and perform reconstruction to prevent data loss.

Emergency backup requires that the capacity of the idle disk used for backup be greater than or equal to the capacity of the member disk, and the media type must match that of the member disk.

RAID Reconstruction

When a disk fails in a RAID group, the data on the failed disk can be reconstructed on a new disk using the data reconstruction function of the RAID controller. The data reconstruction function applies only to RAID levels with redundancy functionality: RAID 1, 5, 6, 10, 50, and 60.

The RAID controller supports automatic reconstruction of failed member disk data by a hot spare disk. When a usable hot spare disk is configured in the RAID group, the hot spare disk automatically replaces the failed disk and performs data reconstruction upon member disk failure. If no usable hot spare disk is available in the RAID group, data reconstruction can only be performed after replacing the disk with a new one. Once the hot spare disk begins data reconstruction, the failed member disk is marked as removable. If the system powers off during data reconstruction, the RAID controller will resume the data reconstruction task after the system restarts.

The source describes a 0–100 rebuild/reconstruction-rate setting. It controls the controller’s scheduling/resource allocation for background work; it is not a measured percentage of all host CPU resources or a guarantee that 100 consumes every CPU. Confirm the exact firmware semantics and balance rebuild time against foreground workload.

Use an eligible replacement with the required capacity, interface, sector format and supported media/protection properties. Existing foreign RAID metadata requires evaluation: preserve it if it may hold required array information, and clear it only when verified obsolete for the intended replacement. Do not erase or initialize original recovery members merely to make them selectable.

Hard Disk Status

Physical disks under the RAID controller have the statuses shownTable 3-1 in below.

Table 3-1 Physical Hard Disk Status

Status	Description
Available (AVL)	The disk may not be ready and is unsuitable for use in logical drives or hot spare pools.
Cfgincompat	Disk configuration is incompatible.
Copyback/Copybacking	A new disk is replacing a hot spare disk.
Diagnosing	The disk status is being diagnosed.
Degraded (DGD)	The disk is part of a logical drive and is in a degraded state.
Erase In Progress	The disk is undergoing erasure.
Failed (FLD)	This status appears when an unrecoverable error occurs on a disk in "Online" or "Hot Spare" status.
Fault	Disk failure.
Foreign	The disk contains external configuration.
Hot Spare (HSP)	Configured as a hot spare disk.
Inactive	The disk is a member of an inactive logical drive and can only be used after the logical drive is activated.
JBOD	The disk is used in pass-through mode.
Missing (MIS)	This status appears after a disk in "Online" status is removed.
Offline	The disk is a member of a virtual disk, cannot be used normally, and is in an offline state.
Online (ONL)	The disk is a member of a virtual disk, can be used normally, and is in an online state.
Optimal	The disk status is normal and is part of a logical drive.
Out of Sync (OSY)	The disk is part of an IR logical drive and is out of sync with other disks in that logical drive.
Predictive Failure	The controller reports a predictive-failure condition. Protect accessible data and follow the applicable vendor replacement procedure; the label is not a prediction of an exact failure time.
Raw (Pass Through)	Physical disk exposed through HBA/pass-through mode.
Raw Drive	The device is exposed as a raw physical drive.
Ready (RDY)	Applicable to RAID/Mixed mode for RAID controllers. Disks can be used to configure RAID groups. In RAID mode, disks in Ready status are not reported to the operating system (hereinafter referred to as "OS"). In Mixed mode, disks in Ready status are reported to the OS.
Rebuild/Rebuilding (RBLD)	The disk is performing data reconstruction to ensure the data redundancy and integrity of the virtual disk. During this process, the performance of the virtual disk may be affected to some extent.
Reconstructing	The disk is performing data reconstruction to ensure the data redundancy and integrity of the virtual disk. During this process, the performance of the virtual disk may be affected to some extent.
Spare	The disk is a hot spare disk and its status is normal.
Shield State	Temporary status of a physical hard disk during diagnostic operations.
Standby (SBY)	The device is not a hard disk device.
Unconfigured Good (ugood/ucfggood)	Hard disk status is normal, but it is not a member disk or hot spare disk of a virtual disk.
Unconfigured Bad (ubad)	"Unconfigured Good" state hard disk or an uninitialized hard disk. When an unrecoverable error occurs, this status is displayed.
Unknown	Hard disk status is unknown.
Unsupport	Hard disk specifications exceed the current specifications of the RAID controller.

Virtual disks created under the RAID controller have the followingTable 3-2 statuses shown below.

Table 3-2 Virtual Disk Status

Status	Description
Degrade/Degraded (DGD)	The virtual disk is degraded because one or more members failed or went offline. It may remain accessible within its RAID limits, with reduced redundancy; the amount of remaining protection depends on the level and failed-member distribution.
Delete fail	Deletion failed.
Deleting	Deleting this virtual disk.
Failed (FLD)	Virtual disk failure.
Fault	Virtual disk failure.
Formatting	Formatting in progress.
Inactive	Virtual disk is inactive and can be used only after activation.
Inc RAID	SMART or SATA extended commands are not supported.
Initializing (INIT)	Initializing.
Initialize fail	Initialization failed.
Interim Recovery	RAID group degraded because a faulty disk exists within the RAID group, reducing operation performance and potentially causing data loss. To resolve this issue, check whether the hard disk is properly connected or replace the faulty hard disk.
Max Dsks	The number of hard disks supported by the RAID group has reached the maximum limit; this hard disk cannot be added to the RAID array.
Missing (MIS)	Virtual disk lost.
Not formatted	Not formatted.
Not Syncd	Data on this hard disk is out of sync with data on other hard disks in the logical disk.
Normal	Virtual disk status is normal; all member disks are online.
Offline	This status appears when the number of failed or offline physical hard disks in the RAID group exceeds the maximum number of failed disks supported by the RAID group level.
Online (ONL)	Virtual disk is online.
Optimal	Virtual disk status is healthy; all member disks are online.
Okay (OKY)	Virtual disk is active and drives are operating normally. If the current RAID level provides data protection, user data will be protected.
Partial Degraded	This status appears as partial degradation when the number of failed or offline physical hard disks in the RAID group does not exceed the maximum number of failed disks supported by the RAID group level.
Primary	This hard disk is the primary disk in RAID 1 and is in normal status.
Ready for Rebuild	Preparing for rebuild.
Sanitizing	Virtual disk is performing data erasure.
Secondary	This hard disk is the secondary disk in RAID 1 and is in normal status.
Tool Small	Hard disk capacity is too small to be used as a hot spare disk for the current RAID group.
Wrg Intfc	Hard disk interface differs from the interfaces of hard disks in the current RAID group.
Wrg Type	This hard disk cannot be a member disk of this RAID group; it may be incompatible or faulty.
Write protect	The virtual disk is write-protected.

Virtual disk read/write policy

When creating a virtual disk, its data read/write policy must be defined to standardize the data read/write behavior during subsequent virtual disk operations.

Data read policy

In the configuration interface, this is generally referred to as "Read Policy". The RAID controller supports the following two data read policies:

• Prefetch mode: In the configuration interface, options such as "Always Read Ahead", "Read Ahead", and "Ahead" are typically available. After using this policy, when reading required data from the virtual disk, subsequent data is also read into the Cache simultaneously. When users subsequently access this data, it can be directly hit in the Cache, reducing disk seek operations, saving response time, and improving data read speed.
  Read prefetch does not itself require protected write-back cache. For write-back data, actual power-loss protection and firmware fallback determine exposure; an unprotected acknowledged write can be lost. The original repeats a protection claim under both read and write policy; the read-policy claim is corrected here.
• Non-prefetch mode: After using this policy, the RAID controller reads data from the virtual disk only upon receiving a data read command, without performing prefetch operations.

Data write policy

The source discusses four write-policy labels below. Their exact availability/naming is firmware-specific; distinguish configured policy from the current effective policy.

• Write Back: In the configuration interface, this is generally indicated by terms such as "Write Back". After using this policy, when writing data to the virtual disk, the data is written directly to the Cache. Once the accumulated data reaches a certain threshold, the RAID controller flushes the data to the virtual disk. This not only enables batch writing but also improves data write speed. After the controller Cache receives all transmitted data, it returns a data transmission completion signal to the host.
  Read prefetch does not itself require protected write-back cache. For write-back data, actual power-loss protection and firmware fallback determine exposure; an unprotected acknowledged write can be lost. The original repeats a protection claim under both read and write policy; the read-policy claim is corrected here.
• Write Through acknowledges a write after the disk subsystem receives its data. It avoids retaining acknowledged writes solely in unprotected controller write-back cache; it does not mean that all read caching or the drive’s own cache is disabled.
  Controller write-back protection is not required for the same acknowledgment behavior, but drive caches, device power-loss behavior and the rest of the storage stack still matter. Write Through is not a guarantee against every power-loss failure, and performance depends on workload.
• Write Back with BBU uses controller write-back while the required backup module is available and healthy and falls back to Write Through when protection is unavailable. The applicable battery/supercapacitor and effective-policy checks depend on the controller; use its actual status rather than assuming any installed capacitor is healthy.
• Forced Write Back: In the configuration interface, this is generally indicated by terms such as "Write Back Enforce" or "Always Write Back". This mode forces the use of "Write Back" mode when the RAID controller has no capacitor or the capacitor is damaged.
  If the server experiences an abnormal power-off and the capacitor is absent or in a charging state, the data written in the DDR (i.e., Cache) of the RAID controller will be lost. This mode is not recommended.

Data power-loss protection

Principle of power-loss protection

The speed of writing data to the RAID controller's high-speed cache is greater than the speed of writing data to the hard disk. During large-scale write operations on the server, the RAID controller's high-speed cache is used to improve system performance.

• Enabling the RAID controller's high-speed cache improves overall system write performance. When the server's write pressure decreases or the RAID controller's high-speed cache becomes full, data is then written from the RAID controller's high-speed cache to the hard disk.
• While enabling the RAID controller's high-speed cache improves write performance, it also increases the risk of data loss. In the event of unexpected power-off of the entire system, data in the RAID controller's high-speed cache will be lost.

To improve the overall system's high read/write performance and the safety of data in the RAID controller's high-speed cache, a supercapacitor can be configured for the RAID controller. The principle of the supercapacitor protection module is that in the event of an unexpected system power-off, the supercapacitor provides power to write the data in the RAID controller's high-speed cache permanently into the NAND Flash within the supercapacitor module.

Supercapacitor charge calibration

The archived source describes an automatic charge/calibration cycle for its backup module. Battery learn cycles and supercapacitor health checks vary by product and firmware; this sequence is not a universal maintenance procedure or an instruction to manually discharge a module.

The source’s described three-stage sequence is recorded below; confirm the exact backup-module manual before applying it to a particular card.

1. The RAID controller charges the supercapacitor to its maximum capacity.
2. Automatically start the calibration process to fully discharge the supercapacitor.
3. Recharge again until the maximum charge level is reached.

During a controller-managed learn/calibration or unavailable-protection state, effective write policy can fall back to Write Through, reducing write performance. Monitor the actual controller/module state and firmware guidance; do not assume every supercapacitor uses the same full-discharge cycle or duration.

Disk striping

When multiple processes access a single hard disk simultaneously, disk conflicts may occur. Most disk systems impose limits on the number of accesses (I/O operations per second) and data transfer rates (data volume transferred per second). When these limits are reached, subsequent processes requiring disk access must wait.

Striping is a technique that automatically balances I/O load across multiple physical disks. Striping technology divides a continuous block of data into multiple small parts and stores them separately on different disks. This allows multiple processes to access different parts of the data simultaneously without causing disk conflicts, and it maximizes I/O parallelism when sequential access to this data is required.

Disk striping divides disk space into multiple stripes according to a set size, and data modules are also divided according to the stripe size during writing.

For example, in a disk system composed of four member disks (such as RAID 0), the first data block is written to the first member disk, the second data block to the second member disk, and so on, asshown in Figure 3-1. Since multiple disks are written to simultaneously, system performance is greatly improved. However, this feature does not guarantee data redundancy.

Figure 3-1 Example of disk striping

Disk striping includes the following concepts:

• Strip width: The number of disks used in a disk group when performing striping. For example, for a disk group consisting of four member disks, the strip width is "4".
• Strip size of a disk group: The size of the data block written by the RAID controller simultaneously on all disks within a disk group.
• Strip size of a disk: The size of the data block written by the RAID controller on each individual disk.

Distinguish per-member strip size from full data-stripe size. In a four-disk RAID 0, four 64 KB strips form a 256 KB full stripe. The source’s separate arithmetic example of a 1 MB full stripe made of 64 KB strips requires 16 data strips; it is not the pictured four-drive layout. Parity layouts also distinguish data strips from parity strips.

Disk mirroring

RAID 1 and RAID 10 mirror data across the required members. A two-disk mirror maintains another copy after one member fails, provided that copy is usable. Reads may be served according to controller policy, while writes update the mirror members. Mirroring does not guarantee uninterrupted I/O in every failure or protect against deletion, corruption or all members failing.

Disk mirroring provides 100% full redundancy, but the cost is relatively high because each disk requires a backup disk during the mirroring process, asFigure 3-2 shown.

Figure 3-2 Disk mirroring example

External configuration

External configuration refers to configurations distinct from the current RAID controller's configuration. In the configuration interface, it is generally displayed as "Foreign Configuration" or similar terms.

External configurations typically appear under the following circumstances:

• A newly installed physical disk in the server contains RAID configuration information, which the RAID controller identifies as external configuration.
• After the server replaces its RAID controller, the new RAID controller identifies existing RAID information as external configuration.
• After hot-swapping a member disk of a RAID group, that member disk is marked as carrying an external configuration.

Detected external configurations can be handled based on the actual server situation. For example, if the RAID information carried by a newly inserted disk does not match the current usage scenario, this configuration can be deleted. After replacing the RAID controller, if you wish to use the previous configuration, you can import it to make it effective on the new RAID controller.

Disk power saving

Some controller/firmware and SAS/SATA HDD combinations support idle or hot-spare spin-down. The original says all SAS/SATA mechanical drives support it, which is too broad; confirm drive, enclosure and controller capability and any effect on spin-up/rebuild latency.

When the disk power-saving feature is enabled, idle disks and idle hot spare disks attached to the RAID controller are in power-saving mode. Disks in power-saving mode can be awakened when relevant operations occur (e.g., creating a RAID group, creating a hot spare disk, dynamic disk expansion, or hot spare reconstruction).

Disk pass-through

The source uses JBOD/pass-through for exposing eligible physical disks to the OS without first creating a RAID virtual drive. JBOD is not an unconditional guarantee of command transparency, error-free transmission or equivalence to every HBA mode.

After enabling the disk pass-through feature, the RAID controller can perform command passthrough on connected disks. Without configuring a RAID group, user commands can be directly passed through to the disks, facilitating access and control by upper-layer business software or management software.

For example, during the server operating system installation process, disks mounted under the RAID controller can be directly identified as installation disks. In contrast, with RAID controllers that do not support disk pass-through, only pre-configured virtual disks under the RAID controller can be found as installation disks during the OS installation process.

Common operations

Log in to the RAID controller management interface in EFI/UEFI mode (Brickland platform)

Operation scenario

The method to log in to the RAID controller management interface in EFI mode varies slightly across different BIOS platforms.

This section uses LSI SAS2208 as an example to explain how to log in to the RAID controller management interface in EFI mode on the Brickland platform.

Impact on the system

The server needs to be restarted, causing a temporary interruption of system services.

Prerequisites

Preconditions

You have remotely connected to the server's real-time desktop via a management tool (iBMC or iMana 200).

Operation steps

1. During server restart, press the "Delete" key according to the on-screen prompts to enter the BIOS configuration interface.
2. Switch to the "Advanced" tab.
   Display the list of currently existing RAID controllers, such asFigure 4-1 shown.
   Figure 4-1 RAID Card List
   
   
   The information displayed here varies slightly for different RAID card types.

Log in to the RAID card management interface in EFI/UEFI mode (Grantley platform)

Operation Scenario

The method to log in to the EFI mode RAID card management interface differs slightly across BIOS platforms.

This section uses LSI SAS2308 as an example to explain how to log in to the EFI mode RAID card management interface.

Impact on System

The server must be restarted, causing a service interruption.

Prerequisites

Preconditions

You have connected to the server's real-time desktop via the remote console of the management software (iBMC or iMana 200).

Data

None

Procedure Steps

1. During server startup, when the screen displaysFigure 4-2 press F11.
   A password input box appears.
   Figure 4-2 BIOS Boot Interface
   
2. Enter the password and press Enter.
   The Boot Manager interface opens, as shown inFigure 4-3.
   Startup messages may vary slightly for different products; follow the on-screen prompts.
   Figure 4-3 Boot Manager Interface
   
3. Press ESC.
   Return to the manager list, as shown inFigure 4-4.
   Figure 4-4 Manager List
   
4. Select Device Manager and press Enter.
   The Device Manager interface opens, as shown inFigure 4-5.
   
   Information displayed for different RAID card types may vary slightly.
   Figure 4-5 "Device Manager" interface
   

Configure LEGACY mode

Operation scenario

Users can set the server boot mode to Legacy mode based on actual operational requirements.

Impact on the system

The server must be restarted, causing a service interruption.

Prerequisites

Preconditions

Successfully connected to the server's real-time desktop via the remote console of the management software (iBMC or iMana 200).

Data

None

Procedure steps

1. Restart the server. When the screen displays as shown inFigure 4-6 press the "Delete" key.
   Enter the BIOS configuration interface.
   Figure 4-6 BIOS restart interface
   
2. In the "Boot" tab of the BIOS, set the RAID card boot mode to "LEGACY" mode.
   • On the Grantley platform, set "Boot Type" to "Legacy Boot Type", as shown inFigure 4-7.
     Figure 4-7 Configuring LEGACY mode on the Grantley platform
     
   • On the Brickland platform, first set "Boot mode select" to "LEGACY", then set "Storage" to "Legacy only", as shown inFigure 4-8 andFigure 4-9.
     Figure 4-8 Configuring LEGACY mode on the Brickland platform
     
     Figure 4-9 Setting the "Storage" parameter
     
3. Press "F10" to save the configuration.
   An operation confirmation dialog box appears.
4. Select "Yes" and press "Enter".
   The server automatically restarts.

Configure EFI/UEFI mode

Operation Scenario

Users can set the server boot mode to EFI/UEFI mode based on actual operation requirements.

Impact on System

The server must be restarted, causing an interruption of system services.

Prerequisites

Prerequisite Conditions

You have connected to the server's real-time desktop via the remote console of the management software (iBMC or iMana 200).

Data Requirements

None

Procedure

1. Restart the server. When the screen displayingFigure 4-10 appears, press the "Delete" key.
   Enter the BIOS configuration interface.
   Figure 4-10 BIOS Restart Interface
   
2. In the BIOS "Boot" tab, set the RAID controller boot mode to EFI/UEFI mode.
   • On the Grantley platform, set "Boot Type" to "UEFI Boot Type", as shown inFigure 4-11.
     Figure 4-11 Configuring UEFI Mode on the Grantley Platform
     
   • On the Brickland platform, first set "Boot mode select" to "UEFI", then set "Storage" to "UEFI only", as shown inFigure 4-12 andFigure 4-13.
     Figure 4-12 Configuring UEFI Mode on the Brickland Platform
     
     Figure 4-13 Setting the "Storage" Parameter
     
3. Press "F10" to save the configuration.
   An operation confirmation dialog box appears.
4. Select "Yes" and press "Enter".
   The server restarts.

RAID Controller Driver Download and Installation

Download the RAID Controller Driver

1. Click the following link to log in to the Computing Component iDriver page.
   • Enterprise Users:Computing Component iDriver
   • Carrier Users:Computing Component iDriver
2. Download the driver compatibility table.
   1. Click the target "Version and Patch Number" to enter the corresponding page.
   2. Scroll to the bottom and download the "Computing Component iDriver X.X.X Driver Compatibility Table" for the corresponding architecture.
      • Enterprise Network Users: Download from "Version Information".
      • Carrier Users: Download from "Version Documents".
3. Determine the compatibility relationship between RAID controller drivers and the OS using the driver compatibility table.
   1. Open the driver compatibility table.
   2. Click to enter the corresponding operating system sheet.
      It mainly contains the following information:
      • System Version: The operating system version compatible with the driver.
      • External Driver Version: The name of the driver file package under Computing Component iDriver.
      • Driver File: The name of the image containing the driver file. Unpack "External Driver Version" to obtain it.
      • Onboard ISO Driver contain Files: The driver installation package or compressed package. Unpack or mount the image file shown in "Driver File" to obtain it.
      • Card Name: The model number of the card.
      • Driver Version: The driver version.
      • FW Version: The compatible firmware version.
      • Chip: The chip model.
      • VID/DID/SVID/SSID: Quad-tuple information.
      • Driver Name: The driver name.
      • Model: The card type.
      • Chip Type: The chip type.
      • Remarks: Notes.
      • iDriver Version: The iDriver version where the file is located.
   3. Filter the controller card keyword by "Card Name".
   4. Based on the operating system version, determine the iDriver version and the name of the driver file package where the driver file is located.
4. Download the RAID controller driver.
   1. Enter the iDriver version obtained from the driver compatibility table.
   2. Download the driver software package name obtained from the driver compatibility table.
   3. Unpack the driver package to see the RAID controller drivers that can be installed under the specified operating system.
      RAID controller drivers come in two formats:
      • img file: the archived source uses this for loading an applicable driver during OS installation. Actual packaging and installer support depend on OS and package version; the extension alone does not establish its use.
      • iso: the source describes post-install driver-media packages, with the two examples below. An ISO can have other uses, including installer-accessible virtual media; follow the actual package readme rather than inferring timing solely from the extension.
        • The standalone ".iso" file is an independent driver package for a single RAID controller card.
        • The "onboard_driver_xxx.iso" file is a driver package for all cards compliant with the current server compatibility.

Install RAID Card Driver

There are currently two ways to obtain the RAID card driver installation method:

• The "readme.txt" in the driver package briefly introduces the server board driver installation method.
• "Huawei Server Operating System Installation Guide".

Manage RAID Card via iBMC

The source states that iBMC 2.32 introduced out-of-band basic query/configuration for applicable LSI SAS3008 and SAS3108 hardware. Availability still depends on the card/PCB, firmware and platform; this is not support for every RAID controller on every later iBMC version.

For detailed operation information, please refer toiBMC User Guide.