Preparing storage nodes

Note:

Procedures in this section are performed in specific cases. The following show those cases and physical servers to be installed.

Initial installation: Perform this operation for the storage nodes to be initially installed.
Storage node addition or storage node replacement: Perform this operation for the storage nodes to be added or replaced.
Spare node preparation: Perform this operation for spare nodes.

Before you begin

You must prepare physical servers used as storage nodes to meet the following configuration and number of storage nodes according to the user data protection method and the number of fault domains.

User data protection method ¹	Number of fault domains that can be specified ^1,2	Number of storage nodes that can be configured	Remarks
HPEC : 4D+1P	1	5 to 32
HPEC : 4D+2P	1	6 to 32
HPEC : 4D+2P	3	6, 12, 18, 24, 30	Storage nodes must be evenly spaced for each fault domain.
Mirroring : Duplication	1	3 to 32
Mirroring : Duplication	3	3, 6, 9, 12, 15, 18, 21, 24, 27, 30	Storage nodes must be evenly spaced for each fault domain.
1. For details about the user data protection method and fault domain, see Settings and functions for fault tolerance supported by VSP One SDS Block. 2. For details about the requirements for building a fault domain, see Settings and functions for fault tolerance supported by VSP One SDS Block.

Verify that a firmware version that can be used by VSP One SDS Block is applied on the physical servers to be used.

Note:
For details about the physical servers and firmware versions that can be used, see the VSP One SDS Block Bare Metal Hardware Compatibility Reference.
Record the mounting location of all drives (actual) to be added to the physical server in association with the drive SKU and the WWN or EUI on the label.

For the drive SKU, see the VSP One SDS Block Bare Metal Hardware Compatibility Reference.

The recorded information will be used when the drive is reduced or replaced. The information will be also used to confirm encryption key information in audit logs.
Record the mount positions of all FCHBAs to be added to the physical server and the WWPNs indicated on the labels, including their correspondence.

The recorded information will be used when an FCHBA is replaced.
Apply non-RAID configuration to use a user data drive.
When using disk controllers, connect the system drive and user data drive to separate disk controllers.
To distribute failure points and load, check the following points when connecting user data drives.
- When connecting user data drives to disk controllers:
  
  Verify the connection to the disk controller where only user data drives are connected, then install the user data drive in any available drive slot. However, if multiple disk controllers are available for connecting user data drives, connect the user data drives so that the number of user data drives connected to each disk controller is equal.
- When connecting user data drives with NVMe Drive Direct Attach:
  
  After verifying the wiring between each drive slot and the CPU, connect the user data drives so that the number of user data drives connected to each CPU is equal.
When you use a Smart array controller as a disk controller to which user data drives are connected, set the controller to HBA mode.
When you use a MegaRAID controller as a disk controller to which user data drives are connected, set the controller to JBOD mode.

CAUTION:

If the user data drives have existing partitions, installation of storage software might fail.

Use the drives after deleting their partitions as the user data drives.
It is possible to mount only one FCHBA in a storage node. If you mount multiple FCHBAs in a storage node and attempt port setting for each FCHBA, set up of VSP One SDS Block becomes unsuccessful.

Specify the default settings for each physical server according to the server vendor documentation.
Note:
For the following servers, see Server User Guide for the applicable server.
- Hitachi Advanced Server HA820
- Hitachi Advanced Server HA820 G2
- Hitachi Advanced Server HA810 G3
- Hitachi Advanced Server HA820 G3
- Hitachi Advanced Server HA810 G6
- Hitachi Advanced Server HA820 G6
Make network connections for each storage node or each spare node by referring to the following:
- Documentation of the server vendor
- Internode network requirements
- Control network requirements
- Compute network requirements (iSCSI)
- Compute network requirements (NVMe/TCP)
- Compute network requirements (FC)
- BMC network requirements
Connect the following networks:
- Controller network
- Compute network
- Internode network
- BMC network
Note:
- Note down the following information. For details about installation locations, see the VSP One SDS Block Bare Metal Hardware Compatibility Reference.
  - Physical location of an NIC (adapter) or FCHBA to be mounted on the physical server.
  - Network connection port number of an NIC (adapter) or FCHBA to be mounted on the physical server.
  The recorded information will be used when replacing an NIC (adapter) or FCHBA in the event of a failure.
- For the following servers, see Server User Guide for the applicable server.
  - Hitachi Advanced Server HA820
  - Hitachi Advanced Server HA820 G2
  - Hitachi Advanced Server HA810 G3
  - Hitachi Advanced Server HA820 G3
  - Hitachi Advanced Server HA810 G6
  - Hitachi Advanced Server HA820 G6
Specify the BMC-related settings according to the server vendor documentation.
Specify the following BMC-related settings:
- Setting the BMC network
- Synchronizing the time between the BMC and NTP server
- Setting a license
- Create a BMC user account
  
  CAUTION:
  - Using the BMC functionality, specify settings to synchronize the BMC time with that of the NTP server so that the BMC time matches the time of other components* comprising your storage cluster.
    
    Then, specify the settings so that the BMC time is applied to the physical servers as their system time. For details about the procedure, see Note described later.
  - When you use multiple NTP servers, also synchronize the time on NTP servers.
  - If the time is not synchronized between physical servers to be used as storage nodes or spare nodes, there is a risk of failure occurring such a trouble in configuring a storage cluster.
    
    By synchronizing the time, you can avoid the risk of problems due to the time gap, and identifying the cause of a problem (such as a failure) becomes easier.
  * The components include storage nodes, spare nodes, compute nodes, controller nodes, and network devices (switches).
Note:
- The following BMC-related operations are performed in initial installation. So, create a user that has permissions to perform them. For permissions required for BMC operations, see the server vendor documentation.
  - Operating the power supply
  - Changing the BMC settings
  - Changing the BIOS or UEFI settings
  - Changing the RAID configuration settings
  - Remote console function
  - ISO image mounting function
- For specific about how to set up Hitachi Advanced Server HA800 series, HA800 G2 series, HA800 G3 series, and HA800 G6 series, see the iLO User Guide.
- In the case that no NTP server exists, configure the NTP server according to Setting up an NTP server.
- For HITACHI iLO7 on the Hitachi Advanced Server HA800 G6 series, after configuring BMC time synchronization with an NTP server, select the pencil icon on the right under iLO Settings > Time > Overview and enable Propagate NTP Time to Host.
  
  For iLO6 and earlier version on the Hitachi Advanced Server HA800 series, HA800 G2 series, and HA800 G3 series, after configuring the BMC time to synchronize with an NTP server, enable Propagate NTP Time to Host in SNTP Settings.
  
  In addition, specify the BIOS or UEFI function settings to select UTC for the time data to be stored in the hardware clock (RTC).
  
  In the Hitachi Advanced Server HA800 series, HA800 G2 series, HA800 G3 series, and HA800 G6 series, you can specify this setting by selecting System Utilities > System Configuration > BIOS or Platform Configuration (RBSU) > Date and Time. For details about how to perform this setting, see the UEFI System Utility User Guide.
- For physical servers other than Hitachi Advanced Server HA800 series, HA800 G2 series, HA800 G3 series, and HA800 G6 series, configure the BMC time to be synchronized with the NTP server. In addition, configure the BIOS/UEFI function settings to select UTC for the time data to be stored in the hardware clock (RTC). For details about specific setting procedures, see the respective server vendor documentation.
- Installing and operating VSP One SDS Block requires an “iLO Advanced” license.
Configure the BIOS-related environment required for VSP One SDS Block on each storage node or spare node as follows.
- Hyperthreading (SMT): Disabled
- NUMA: Enabled
- Sub-NUMA: Disabled 
For details about the specific configuration and procedures for Hitachi Advanced Server HA800 series, HA800 G2 series, HA800 G3 series, and HA800 G6 series, see the server user’s guide and Supported Server Models in the VSP One SDS Block Bare Metal Hardware Compatibility Reference. For details about specific setting procedures for physical servers other than those servers, see the respective server vendor documentation.

CAUTION:
- If the above procedure is not correctly performed, the following problems might occur:
  - Configuration of the storage cluster fails.
  - The operation and performance of the system are adversely affected.
- To confirm or change the settings specified here after the storage cluster starts running, the storage cluster must be stopped and restarted. This procedure is mandatory.