Cluster Platform 280/3 Introduction

This chapter describes the following information:

• Cluster Platform 280/3 System Overview
• Software Components
• Hardware Components
• Connectivity
• Power and Heating Requirements

---

Cluster Platform 280/3 System Overview

The Cluster Platform 280/3 system is a self-sustained platform, integrated through the Sun Cluster technology, with shared, mirrored Fibre-Channel Arbitrated Loop (FC-AL) storage to support highly available applications. You can use this two-node cluster system to implement a highly available file server, web server, mail server, or an Oracle® database server. Sun Cluster 3.0 provides global file systems, global devices, and scalable services. These features allow independent cluster nodes, running independent Solaris operating environment instances, to run distributed applications while providing client access through a single IP address.

---

Note - This cluster platform only provides a basic cluster environment. You must install and configure your data services after you have customized and configured the system.

---

FIGURE 1-1 is a basic block diagram of the system, which includes a two-node cluster with shared, mirrored storage, a terminal concentrator, and a management server. An Ethernet hub provides connection between the cluster nodes, management server, and Sun StorEdge T3 arrays. The interfaces on both nodes are cabled to provide multiple cluster interconnects between nodes.

The management server is the repository for the operating environment, software, and patches. The management server provides access to the cluster console and functions as a JumpStart/Flash server (installation server) for the cluster nodes. For advanced system monitoring, the management server includes Sun Management Center software to monitor the cluster nodes.

The Cluster Platform does not have a monitor. You access the Cluster Platform using a system on the network that you provide (shown as your local system in FIGURE 1-1). The local system is simply used as a remote display to the management server. You do not need to dedicate any particular system as the local system. It can be any system that is convenient at the time.

 FIGURE 1-1 Cluster Platform 280/3 System Overview

---

Software Components

Once the Cluster Platform 280/3 system configuration is complete, the following software components are installed:

Management Server

• Solaris 8 10/01 operating environment with patches
• Apache 1.3 web server software
• Solstice DiskSuite 4.2.1 with patches, and data services
• Sun Cluster 3.0 12/01 and recommended patches
• Cluster Platform 280/3 node operating system software
• Sun Management Center 3.0 Update 4 software with Help desk
• AnswerBook2 software
• Sun Enterprise Explorer 3.5 software
• JumpStart/Flash server
• Sun StorEdge Component Manager 2.2 software
• Highly Available (HA) NFS application software
• SunRay Server 1.3 software

Cluster Nodes

• Solaris 8 10/01 operating environment
• Sun Cluster 3.0 12/01 software
• Solstice DiskSuite 4.2.1 or VERITAS Volume Manager 3.1.1 software¹
• Sun Management Center 3.0 Update 2 agent software

^{1 You must obtain two VERITAS license keys (one for each node) if you intend to use the VERITAS Volume Manager product on your Cluster Platform. You can obtain your VERITAS license and keys from Sun Microsystems, or from VERITAS Corporation. These license keys are not needed if you plan to use Solaris DiskSuite instead of VERITAS Volume Manager.}

---

Note - To see the collection of patches that were installed on each cluster node, look in the /jumpstart/Patches/8_recommended directory of the management server.

---

---

Note - Recovery disks are included with the Cluster Platform 280/3 system. The two disks included are the Cluster Platform 280/3 Mini Root CD and Cluster Platform 280/3 Recovery DVD. See To Use the Cluster Recovery Utility for recovery details.

---

---

Hardware Components

Your factory-integrated Cluster Platform 280/3 system includes the following hardware components:

• Consolidation expansion cabinet: One 72-inch Sun StorEdge expansion cabinet which includes two power sequencers, one front door option, one rear door, and one fan assembly
• Management server: One Netra T1 AC200 server configured with one 500 MHz UltraSPARC IIe CPU, 1-Gbyte of memory, two internal mirrored 18-Gbyte boot disks, and a DVD-ROM drive
• Terminal concentrator: One 8-port 10BASE-T terminal concentrator preinstalled in a Sun StorEdge expansion cabinet
• Cluster nodes: Two Sun Fire 280R servers, with two 750-MHz or 900-MHz UltraSPARC III CPUs, 2-Gbytes of memory, two mirrored 36-Gbyte boot disks, and a DVD-ROM drive
• Cluster node storage: Two Sun StorEdge T3 arrays, each with 1-Gbyte controller board cache and nine 36-Gbyte disks or 72.4-Gbyte disk drives.
• Four internal PCI cards per node; two Sun Quad FastEthernet PCI cards per node installed in PCI1 and PCI2 slots; two FC-AL PCI host adapter cards per node installed in PCI3 and PCI4 slots
• Two FC-AL StorEdge hubs, with three Gigabit Interface Converters (GBICs) installed in each hub, pre-cabled to the cluster nodes and to the Sun StorEdge T3 disk arrays
• One 3COM 12-port Ethernet hub for the internal administrative network
• All required mounting brackets, power cords, and cables.

Cluster Platform Component Location

FIGURE 1-2 shows how the Cluster Platform 280/3 is arranged in the expansion cabinet. The hardware components are placed in the dual-power sequenced 72-inch expansion cabinets.

System components are cabled in the expansion cabinet to provide power and connections between the hardware components.

---

Note - The expansion cabinet arrangement complies with weight distribution, cooling, electromagnetic interference (EMI), and power requirements.

---

 FIGURE 1-2 Cluster Platform Expansion Cabinet Placement

---

Connectivity

The Cluster Platform 280/3 provides maximum connectivity in your enterprise network using the following network implementations:

• Production network connectivity--You connect the cluster nodes to your main network using the Sun Fire Server 280R qfe0 connectors. Network automatic failover (NAFO) is provided on separate Sun Quad FastEthernet controllers (qfe4) to avoid a single point of failure (SPOF).
• Administration network connectivity--Once you connect the management server and the terminal concentrator to your administration network, the management server has the connectivity needed to give you access to the administration software (Sun Management Center, Component Manager, and SunRay server software). This network configuration isolates the administrative functions from your production network. If your environment does not have an administration network, you can connect the management server and the terminal concentrator to your production network, maintaining the same cluster platform functionality, without the network isolation.
• Internal administration network--The management server, cluster nodes, and cluster storage are connected to each other through the built-in Ethernet hub. Internal administration connections are configured at the factory. This network is referred to as the internal administration network to distinguish it from the administration network (described above).
• Cluster interconnect--The cluster connectivity between nodes uses Ethernet patch cables (no Ethernet switch required), with multiple qfe 100BASE-T ports (qfe1 and qfe5) on two separate controllers to avoid a controller SPOF.

Cluster Platform 280/3 System Cabling

The Cluster Platform 280/3 system is shipped with the servers, hubs, and each of the arrays already connected in the cabinet. You do not need to cable the internal system components. The next chapter describes how to connect the Cluster Platform to your enterprise network.

For factory configured cable connections, see Appendix B.

 

 

 FIGURE 1-3 Cluster Platform Network Connections

---

Power and Heating Requirements

The Cluster Platform 280/3 hardware must have two dedicated AC breakers. The cabinet should not share these breakers with other, unrelated equipment. The system requires two L30-R receptacles for the cabinet, split between two isolated circuits. For international installations, the system requires two Blue 32AIEC309 (international) receptacles.

---

Note - To eliminate any single point of failure (SPOF), you must provide two independent primary power sources. For details on power sources and power sequencers, refer to the Sun StorEdge Expansion Cabinet Installation and Service Manual. This document is supplied to you in hard copy format.

---

If the cabinet is installed on a raised floor, cool conditioned air should be directed to the bottom of the expansion cabinet through perforated panels.

The 72-inch cabinet in the Cluster Platform 280/3 consumes power and dissipates heat as shown in TABLE 1-1.

TABLE 1-1 Power and Heat Requirements for Cluster Platform 280/3

72-Inch Cabinet	Maximum Power Draw	Heat Dissipation
1	2320 W	7965 BTU/hr.