^

Belits Computer Systems


TD-S448664A-G2-001 User's Manual

2. Installation.

2.1. General requirements.

The server should be installed horizontally in a 19" equipment rack or cabinet, or on a flat, horizontal surface. The server should not be tilted in any direction or installed in a non-horizontal position. In particular, server's left side (with the jacks and cables) should never be raised more than 7mm (slightly more than 1/4") above the right side while the server is running. Same applies to tilting the server toward its front or back.

The airflow is directed from the intake in front of the case to the exhaust on the back. Air intake and exhaust should not be obstructed. There are no air intake or exhaust openings on top or bottom surfaces of the server, so the server can be easily placed on shelves, or mounted on rails or brackets in a rack with other equipment, including the same type of servers. Even though the surface of the server's case is warmer than the outside air, it is not specifically designed to perform any heat dissipation.

To install the server in a 4-pole 19" equipment rack or cabinet, use M4 screws provided with the rails set to attach both rails on the sides of the unit. Rails specifically designed for this server have different lengths — left-side rail is shorter. This is made for increased convenience of easier access to the cable jacks in the niche. Other rails can be used as well as long as they have holes for mounting screws in proper positions. Additional holes may be drilled into rails to accommodate this type of servers.

Make sure that the screws do not protrude inside for more than 5mm or 3/16". Use rails and other rack-mounting hardware to complete the installation in the rack. After rails are securely connected to the rack, attach the power, network and console cables while the server is partially pulled out of the rack, and push the server into the rack. Different racks and cabinets may have different limits for horizontal movement of rails.

Alternatively, the server can be installed in an open 2-pole rack using brackets instead of rails. This configuration uses three brackets, two in the back, one in front (on the longer right side of the case), leaving an unobstructed access to the niche on the left side of the case.

When the server is placed on a table, or horizontal shelf, it should be supported over at least 2/3 of its length, and should not have any objects between the bottom of the server and the supporting surface.

The ambient air temperature should be in a range between 5 and 35°C (41 and 95°F), and humidity should be between 10% and 90% noncondensing. Lower temperature increases the reliability and longevity of equipment. Additional components installed in the servers or in the same racks may place additional requirements on temperature limits and required amount of airflow.

2.2. Power, console and network cables.

The server should be connected to AC power (100-240V, 60-50Hz) with a provided power cord, or any other power cord that fits into IEC 320 C13 jack, and provides reliable connection to power (rated for at least 4A for 200-240V, 8A for lower voltage) and grounding. No additional grounding connection is necessary. Network cables should be at least CAT5, either straight or crossover. Optionally provided console cable is compatible and interchangeable with Cisco 72-3383-01 cable.

2.3. Hardware and software configuration.

The servers configured with an operating system, can be installed and booted in their shipped configuration, and used standalone, or as a part of a larger networked setup. Servers shipped without an operating system should have the OS installed by the user.

If the server is shipped with any Linux or FreeBSD version, it is configured to use a serial console with 19200bps, 8N1, RTS/CTS flow control and usable with VT100, ANSI or similar terminal. Also VGA + PS/2 keyboard + mouse connectors on the back of the unit can be used to attach a monitor, keyboard and a mouse, serving as a console.

Unless stated otherwise in the enclosed configuration sheet, network is configured to use DHCP protocol to obtain an IP address. If the server is shipped without an operating system, Ethernet interfaces are configured to use PXE protocol for booting.

Power Management can be configured to immediately turn on the computer after the power cord is plugged in and after the power is restored in the case of power loss. Initially for safety reasons it is set to remain off when the power is applied until the power button on the front panel is pressed. If the server will be used in the environment with expected power losses, or will be shut down by the UPS with an expectation that it will automatically turn itself on after the power is restored, use BIOS setup to change this option.

2.4. Before the installation.

Make sure that the rack and mounting hardware, or a flat horizontal surface, are suitable for the server installation, and that the server will be securely placed in a horizontal position.

Check if power outlets are in the acceptable voltage/frequency range, have adequate grounding, can supply the required current, and are compatible with the provided power cords. If UPS is used, check if it provides sufficient output current to power the new server and existing equipment.

Determine if the server will be monitored/operated through a serial console (recommended for Linux or FreeBSD) or a VGA monitor with PS/2 keyboard. If serial port console should be used, configure a separate monitoring computer or a terminal server to match the parameters of connection in the enclosed configuration sheet. Attach the supplied DB9 to RJ-45 console cable to the monitoring computer or a terminal server, install and configure serial terminal software if necessary.

Various computers may have no built-in serial ports, or a monitoring computer may have all its serial ports used for other purposes. In this case a USB to serial converter can be used instead of built-in serial port.

Software on the monitoring computer may be any of the following:

Linux, FreeBSD Solaris:

cu
Minicom
C-Kermit

running in a terminal such as:
xterm
GNOME Terminal
KDE Konsole
or on the text console.

Microsoft Windows (all versions):

HyperTerminal
ProComm Plus
TeraTerm
Kermit 95
Telix
ZTERM

MacOS X:

ZTerm
cu
Minicom
C-Kermit

MS-DOS (all versions):

Telix
ProComm Plus
MS-DOS Kermit

Details and functionality, OS and hardware support may vary between those programs, versions and their configurations.

Choose the serial port where the cable is connected, and configure serial terminal software according to configuration sheet. If serial terminal supports “Direct Connection” and “Modem” modes, choose “Direct Connection”.

Fig. 5

Fig. 5: Serial console

Configure the network to provide an IP address through DHCP protocol, or to accept a host with parameters listed in the configuration sheet. If the preconfigured parameters are unacceptable for existing configuration, or DHCP server is not installed on the network, server can be booted without a network, and configured manually from the console. If the server is preconfigured with an operating system, and you do not intend to reinstall it on the first boot-up, do not configure PXE boot image for it in your DHCP server configuration.

DHCP server can be provided by various software and networking equipment. The most common DHCP server software is:

Linux, FreeBSD, Solaris:

ISC DHCP Server

Microsoft Windows (various versions):

Windows DHCP Server
ISC DHCP Server (with Cygwin)
Vicomsoft DHCP Server

MacOS X:

OSX DHCP Server
ISC DHCP Server
IPNetRouterX
Vicomsoft DHCP Server

Cisco IOS (various Cisco hardware):

Cisco IOS DHCP Server

If DHCP is used, a particular IP address can be assigned to the server using its MAC address listed in the configuration sheet. Particular configuration options necessary for that are described in the DHCP server's documentation. ISC DHCP Server uses the following format for its DHCP entries in dhcpd.conf file:

group {
  server-name "serverbox-1";
  option domain-name "example.com"
  option domain-name-servers serverbox-1.example.com;
  host serverbox-2 {
    hardware ethernet 00:0C:76:AD:4C:16;
    fixed-address 192.168.10.10;
  }
  host serverbox-3 {
    hardware ethernet 00:0C:76:AD:4C:18;
    fixed-address 192.168.10.11;
  }
}

This example (that should be located at the host serverbox-1.example.com) includes the configuration of the group of two servers that have fixed IP addresses on a private subnet. The MAC addresses of both computers are in “hardware ethernet” statement, and are exactly in the same format as listed in the configuration sheet. Please note that this is a fragment of a configuration file, and the complete network configuration should include a proper subnet definition, DNS configuration, etc. Particular details of this configuration usually have to correspond to the set of addresses provided by ARIN or allocated by hosting or Internet service provider, existing network architecture, etc. Isolated servers, or private subnets not connected to the Internet directly, should use private IP address space in the ranges:

10.0.0.0 - 10.255.255.255
172.16.0.0 - 172.31.255.255
192.168.0.0 - 192.168.255.255

as described in RFC 1918.

Fig. 6 - Fig. 10 show examples of various possible network configurations. Configuration on Fig. 6 and Fig. 7 is a typical setup used in a lab, Fig. 8 is the minimal setup (no hubs, switches or external network connectivity), Fig. 9 is a small office setup (with a single appliance is responsible for both connectivity and network services), Fig. 10 is an example of a mid-size office network.

Fig. 6

Fig. 6: Photo of a typical configuration with a monitoring computer

Fig. 7

Fig. 7: Typical configuration with a monitoring computer

Fig. 8

Fig. 8: Minimal configuration

Fig. 9

Fig. 9: Small office configuration example

Fig. 10

Fig. 10: Mid-size office configuration example

In all those examples the monitoring computer is connected to the server through Ethernet network and through a serial console (in mid-size office example consoles are handled by a terminal server, accessible from a system administrator's workstation over the network). Minimal setup lacks DHCP server or external network connectivity — user has to configure the server with a static IP address, and use the monitoring computer to upload additional software and updates on the server. Typical and small office configurations use existing DHCP server for automatic network configuration, and Internet or media on other computers for software installation and updates. The network in a mid-size office example can be configured in various ways, with some servers configured statically, running DHCP for the rest of the network, or all servers relying on a DHCP server on a router.

When configuring a network, make sure that if servers use static IP addresses, they are compatible with network setup and do not conflict with IP addresses of other devices. If there is an address conflict, disconnect the network cable from the server, fix the problem from the console, apply all changes, verify that the new parameters are correct, then reconnect the server. If servers use DHCP, make sure that there are no multiple conflicting DHCP servers on the network, and that DHCP server is configured to provide IP addresses and other network parameters that match the intended position of the server on existing network. Use specific MAC address to IP address mapping on DHCP server if necessary.

If the server is configured to use DHCP, yet the network lacks DHCP server, boot the server, wait for the timeouts while the server is configuring its IP address or starting network services. Ignore all network-related errors and delays caused by unusable name resolution. Once the server is booted up, login from the console and configure a static IP address, netmask, gateway, DNS server, etc., then reboot the server or manually restart all network-related services.

Servers can be installed without any console, however in that situation various kinds of troubleshooting may become difficult or impossible because the server becomes accessible only after its network interfaces are configured, and remote login services are started. If console is available, system administrator can easily verify that the server is booted up, check and modify network configuration, and run various kinds of diagnostic utilities.

If serial console is used, yet the server does not seem to communicate with a terminal when powered up, it may be a result of console redirection misconfiguration or BIOS configuration reset. Turn server off, connect PS/2 keyboard and VGA-compatible monitor, then power it up again and use that console for the initial configuration. Once console redirection is configured, boot the server with a serial console attached, or with both VGA + keyboard and serial console.

2.5. Installation and server boot-up.

Place the server into a rack, cabinet or on a flat surface. Attach power, Ethernet and console (or VGA + keyboard) cables.

Start all necessary services and perform all configuration on the monitoring computer (if any) and network servers (if they are used in this configuration).

Fig. 11

Fig. 11: Serial console program started on the monitoring computer

Press and release the "Power" button on the server's front panel. Five blue LEDs in the logo and LCD backlight should turn on, fans and hard drives should start spinning, producing a noise. After passing a POST sequence, a BIOS banner will be displayed, followed by the PXE startup screen or system boot-up.

BIOS setup screen that can be accessed by pressing F2 while the startup screen is displayed.

Fig. 12

Fig. 12: BIOS startup screen on a serial console

Note: In this manual all BIOS banners, prompts and configuration screens are shown on a 80x25 characters console while all other screens are on a 80x24 console. This is done to show BIOS screens with all lines that they are supposed to show on a 80x25 VGA monitor. In real-life setups the console is usually in a 80x24 characters mode, so the last line of the some BIOS screens is missing. BIOS menus are designed not to rely on that line being visible. If VGA + keyboard are used as a console, everything uses a 80x25 screen.

More messages and configuration prompts may be displayed while additional ROMs initialize some other hardware and/or attempt to boot from it.

Fig. 13

Fig. 13: Network ROM messages and prompts

If network ROM and PXE booting is enabled, network interface will be configured using DHCP, and a boot image will be downloaded from a TFTP server. If hard drive is configured as the primary boot device, system will boot from a local hard drive.

If GRUB bootloader is used (standard setup for Linux and FreeBSD configuration on TD-44 series servers), a GRUB banner and menu, if enabled, will be displayed. Choose the system option to boot, or wait for timeout to expire, to boot the default system configuration.

Fig. 14

Fig. 14: GRUB menu

From this point the behavior is OS-dependent. If the server is pre-configured with Linux or FreeBSD, system messages will be displayed, following the system boot-up process. As various services will start up, network interfaces and LCD panel will be activated. If the server is configured to use DHCP, and no DHCP server is present on the network, a delay may occur while the DHCP client will be waiting for a response. If MTA (mail transfer agent) is configured, and DNS server is not reachable, another delay may happen while MTA is waiting for the name resolution. After completing the boot-up sequence, the login prompt will be displayed on the console.

Fig. 15

Fig. 15: Login prompt

Prev Home Up Next


Navigation

Prev Home Up Next PDF