This chapter describes the function and physical components of the administrative server that runs the SGI Management Center software (also known as the head node) in the following sections:
For purposes of this chapter “head node” is used as a catch-all phrase to describe the stand-alone server that acts as a management infrastructure controller. The specialized functions this server performs within the system primarily include:
Administration and management
Compute node provisioning functions
Under certain circumstances, head nodes provide services, such as:
Fabric management
Login
Batch
I/O gateway (storage)
The head node is connected to the interconnect network and also to the “outside world”, typically via the local area network (LAN). The head node is the point of submittal for all MPI application runs in the cluster. An MPI job is started from the head node and the sub-processes are distributed to the cluster compute nodes from the head node. The main process on the head node will wait for the sub-processes to finish.
The stick enclosures each containing two slices (compute nodes) are identical computing systems that run the primary processes of MPI applications. These compute nodes are connected to each other through high-speed switches that support the interconnect network.
The network interconnect components are Gigabit Ethernet and InfiniBand. The MPI messages are passed across this network between the processes. This compute node network does not connect directly to the “outside world” because mixing external and internal cluster network traffic could impact application performance.
User interfaces include the Compute Cluster Administrator, the Compute Cluster Job Manager, and a Command Line Interface (CLI). Management services include job scheduling, job and resource management, Remote Installation Services (RIS), and a remote command environment.
The head node is the cluster system's primary interface to the “outside world”, typically a local area network (LAN). The server's control panel features are shown in Figure 5-2.
Table 5-1. System Head Node control panel functions
Functional feature | Functional description |
|---|---|
Universal information LED | This multi-color LED blinks red quickly, to indicate a fan failure and blinks red slowly for a power failure. A continuous solid red LED indicates a CPU is overheating. This LED will be on solid blue or blinking blue when used for UID (Unit Identifier). |
System over temp LED | This red LED lights up to indicate a system over temperature condition. |
NIC 2 Activity LED | Indicates network activity on LAN 2 when flashing green. |
NIC 1 Activity LED | Indicates network activity on LAN 1 when flashing green. |
Disk activity LED | Indicates drive activity when flashing. |
Power LED | Indicates power is being supplied to the server's power supply units. |
Reset button | Pressing this button reboots the server. |
Power button | Pressing the button applies/ removes power from the power supply to the server. Turning off power with this button removes main power but keeps standby power supplied to the system. |
The head node server is an AMD based system that offers the following technical features:
Four 8-core or 12-core AMD 6100 series Opteron processors
32 DIMM slots (eight per processor) that support DDR3 1333/1066/800 MHz SDRAM
Six SATA or SAS disk drives (optional RAID is supported) see Figure 5-3
Two low-profile PCI Express 2.0 x16 slots and two low-profile PCI Express 2.0 x8 slots
Standard PS/2 mouse and keyboard ports and two gigabit Ethernet ports
Integrated Matrox G200 video controller providing a 16MB DDR2 graphics interface through the system VGA connector
Optional redundant power supply (80+ Gold rated)
For more information on operating, troubleshooting or upgrading your Prism XL head node server, see the SGI Rackable H2106-G7 System User's Guide, (P/N 007-5683-00x). For information on using the cluster administrative software (SGI Management Center) see the SGI Management Center System Administrator's Guide, (P/N 007-5642-00x).
