HCP with SAN-attached storage system hardware consists of:
- Nodes with internal storage (a typical starter system has four nodes). The nodes are numbered from 101 through 104 for a four-node system. The node numbers increase by one for each additional node.
The nodes in an HCP with SAN-attached storage system are HCP G11 servers.
- HCP S Series Nodes. The possible node models are:
- S11
- S31
- One or more SAN-attached storage arrays (a typical starter system has one array). Each storage array has one controller tray and one or more expansion trays. Arrays are not always included with the delivered HCP with SAN-attached storage system. You may need to supply them yourself.
Storage can be running storage or spindown storage. Running storage is storage on continuously spinning disks. Spindown storage is storage on disks that can be spun down and spun up as needed. All systems have running storage. Only some systems have spindown storage.
- Ethernet switches and cables for networking. The switches in
an HCP with SAN-attached storage can be for one or ten gigabyte back-end network configurations. The
possible switch models are:
- Supported one gigabyte switch:
- Brocade ICX 6430
- Supported ten gigabyte switches:
- Arista 7020
- Brocade VDX 6740
- Cisco Nexus 5548UP
- Cisco Nexus 5596UP
- Supported 25 gigabyte switch:
- 93180YC-FX3 NX-OS
- Additional infrastructure items such as a rack and Power Distribution Units (PDUs).
- Supported one gigabyte switch:
- Fibre Channel cables that connect the nodes to the storage arrays (Some
systems use Fibre Channel switches to connect nodes and the storage arrays). The possible
Fibre Channel switch models are:
- Brocade 6510
- Cisco MDS 9148S
An HCP system uses a back-end network, front-end network, and, in certain configurations, a management network. The isolated back-end network connects the HCP nodes to each other through one or two Ethernet switches, depending on your network configuration and switch model. Each node has a pair of bonded Ethernet ports for connecting to these switches. Node port locations vary, depending on the network configuration the node was constructed for.
Each node is configured with an additional pair of bonded Ethernet ports that allows external applications to access the system. The recommended setup includes either two independent Ethernet switches that connect these ports to the front-end network (that is, your corporate network) or one Ethernet switch with both HCP and the switch configured for active-active bonding.
The front-end network switches and the cables for connecting them to the HCP nodes are not included with the delivered HCP with SAN-attached storage system. The cables are customer supplied. You can use any reasonable Ethernet switch model for the front-end network, but you cannot use the back-end switches for the front-end network.
Each node also has an additional management network Ethernet port that allows for the creation of the management network. The management network segregates system and tenant administration, management API, SNMP, syslog, outgoing SMTP, and SSH traffic from the [hcp_system] network.
The following table describes the cables in this figure.
Cable | Connects from | Connects to |
---|---|---|
Red and blue Ethernet | Back-end network interface cards (NICs) in each node | Back-end switches |
Green and yellow Ethernet | Front-end NICs in each node | Front-end switches |
Purple Ethernet | Back-end switches | Each other |
Orange Fibre Channel | Each node | SAN array |
Black power | Each node | Two PDUs |
Each back-end switch | One PDU | |
Tray in the SAN-attached storage array | Two PDUs |