This chapter explains how to operate your SGI Origin 3900 server in the following sections:
Before operating your server, familiarize yourself with the safety information in the following sections:
![]() | Caution: Observe all ESD precautions. Failure to do so can result in damage to the equipment. |
Wear an SGI-approved wrist strap when you handle an ESD-sensitive device to eliminate possible ESD damage to equipment. Connect the wrist strap cord directly to earth ground.
![]() | Warning: Before operating or servicing any part of this product, read the “Safety Information” in Appendix B. |
![]() | Warning: Keep fingers and conductive tools away from high-voltage areas. Failure to follow these precautions will result in serious injury or death. The high-voltage areas of the server are indicated with high-voltage warning labels. |
![]() | Caution: Power off the server only after the system software has been shut down in an orderly manner. If you power off the server before you halt the operating system, data may be corrupted. |
![]() | Note: A lithium battery is installed on the IO9 card located in the IX-brick. |
![]() | Warning: Only qualified SGI service personnel should replace the lithium battery. |
The system console enables you to perform the following activities:
Monitor your server by reading the server's status and error message information generated and displayed by the L1 and L2 controllers.
Enter L1 and L2 controller commands to monitor or change particular server functions. You can, for example, monitor the speed of fans for a particular brick. See your SGI L1 and L2 Controller Software User's Guide for descriptions of these commands.
Power on or power off individual bricks or all bricks (except for the storage modules) in your server.
If you connect a console that contains SGIconsole software, you can perform the functions listed above, manage multiple systems from one console, and have boundary scan capabilities. Boundary scan capabilities enable an SGI system support engineer (SSE) to test the interconnections of your server.
To connect a system console, follow these steps:
If your server has multiple L2 controllers, connect each L2 controller (via its Ethernet port) to an Ethernet hub.
Connect your console to the Ethernet hub or to the Console port of an L2 controller as shown in Figure 1-1. The connection of the system console to the Console port requires a special cable that you can purchase from SGI. This configuration also requires terminal emulator software (not available from SGI).
![]() | Note: You can also use SGIconsole to monitor your server. SGIconsole connects to multiple L2 controllers via an Ethernet hub. For instructions on how to connect SGIconsole to your server, see your SGIconsole Hardware Connectivity Guide. |
Besides adding a system console, you can add or replace the following hardware items on your Origin 3900 server:
Peripheral component interface (PCI) cards
Disk drives
The sections that follow discuss these activities in more detail.
![]() | Warning: You can add or replace only the items listed in this section. For your safety and for the protection of your server, contact your SGI system support engineer (SSE) to install any hardware items not listed in this section. |
![]() | Warning: Before installing, operating, or servicing any part of this product, read the “Safety Information” in Appendix B. |
The PCI-X based I/O system, an industry standard for connecting peripherals to a processor, is the primary I/O system for the Origin 3900 server. The hardware components of this I/O system are contained in IX- and PX- bricks. The IX-brick provides the base I/O functionality for the server and 12 PCI/PCI-X slots. One of the 12 slots is reserved for a system PCI card. If additional PCI slots are required beyond the 11 remaining slots in the IX-brick, PX-bricks are configured into the server. Each PX-brick provides an additional 12 PCI/PCI-X slots.
For instructions on how to install or remove PCI cards, see “Installing or Replacing PCI or PCI–X Card” in Chapter 14.
The IX-brick provides the system boot functions and has one or two low-profile SCSI disk drives. For instructions on how to install or remove these disk drives, see “Installing or Replacing Disk Drives in IX-brick ” in Chapter 14.
The TP900 is a SCSI-based disk storage module that provides JBOD (just a bunch of disks) mass storage. Each TP900 can contain up to 8 disk drives. For instructions on how to install or remove these disk drives, see “Installing or Replacing TP900 Drive Carrier Module” in Chapter 14.
The D-brick2 is a Fibre Channel-based disk storage enclosure that provides JBOD mass storage. Each D-brick2 can contain up to 16 disk drives. For instructions on how to install or remove these disk drives, see “Installing or Replacing D-brick2 Drive Carrier Module” in Chapter 14.
This section explains how to power on and power off individual bricks or your entire SGI Origin 3900 server, as follows:
You can power on and power off individual bricks or the entire server at the L2 controller touch display on the front door of rack 001 (the leftmost rack when viewing the front of the racks). Or if your server has a system console, you can power on and power off individual bricks or the entire server at the system console.
This section describes how to prepare to power on your server, and how to power on your server by using either of the following:
L2 controller touch display
System console
The L1 controller display, located on the front of each brick, should display L1 running once the power-on procedure starts.
![]() | Note: If you have a problem while powering on and an error message appears on your L2 controller touch display, your console, or the L1 controller display, see your online log files and, see “L1 Controller Error Messages” in Chapter 15 to learn what the error message indicates and how to resolve the problem. |
To prepare to power on your server, follow these steps:
Check to ensure that the cabling between the power distribution unit (PDU) and the wall power-plug receptacle is secure.
For each individual brick that you want to power on, make sure that the PWR (power) switch is set to the 1 (on) position, as shown in Figure 1-2. When the power switch is on and power is supplied to the brick, the L1 controller of the brick powers on. The 12-VDC LED of the brick illuminates green when 12 VDC is present.
If you plan to power on a D–brick2, or an entire rack of D–brick2s, see “Powering On D-brick2”.
To power on a TP900 storage module, set the power switch of each TP900 storage module to the 1 (on) position.
![]() | Note: The standard configuration of the TP900 storage module has one power supply, as shown in Figure 1-3. You can add a second power supply to the TP900 system to provide n+1 redundant power. The second power supply is identical to the first power supply and would be located in the lower-right section of the TP900 storage module. Both power switches must be in the 1 (on) position to provide redundant power. |
Make sure that the circuit breaker switches of the power distribution unit (PDU) and power distribution strip (PDS) shown in Figure 1-4 are in the on position.
This section describes how to power on individual bricks or the entire server from your L2 controller touch display (see Figure 1-5), which is located on the front door of rack 001.
For instructions on how to use the L2 controller touch display, see “L2 Controller ” in Chapter 8.
The home window, shown in Figure 1-6, displays the following items:
Firmware (xx.xx.x) version of the L2 controller.
Rack number (L2-001) of the L2 controller to which the L2 controller touch display is connected.
System serial number (Lxxxxxxx).
Server name in parentheses (firestorm).
Power status (Power: OFF) for the bricks designated in the destination (DEST:) field. For this example, all slots in all racks (r * s *), which amounts to 57 bricks, are powered off.
Four command buttons (Power UP, Power DOWN, RESET, and NMI) that when selected activate a confirmation window for the command. The command is sent to the bricks that are listed in the DEST: field after you okay the command from the confirmation window.
DEST: button that enables you to select the bricks that will receive a command. In addition, to the right of this button is an area that lists the current brick selection. For this example, all slots in all racks (r * s *), which amounts to 57 bricks, are selected to receive a command.
To power on selected bricks, a partition, or the entire server, follow these steps:
Touch the DEST: button in the home window and the destination selection window, shown in Figure 1-7 appears. Use this window to select the bricks or partitions that you want to power on.
The bricks are selected by their rack and slot/unit number, or by partition. Select ALL in the display if you want to power on all the bricks in the server. You can also select the bricks in all the slots of multiple racks, but you cannot select slots for multiple racks.
If you want to power on individual bricks, select the rack that contains the bricks that you want to power on from the Rack/Slot segment of the display. A box appears that lists the slot numbers for all of the bricks in the selected rack (see Figure 1-8).
Select the slot number of each brick that you want to power on. (Figure 1-8 shows slots 007, 015, 027, and 035 selected). The New DEST: field lists your selections. After you complete your selections, select Apply.
The home window shown in Figure 1-9 appears. The DEST: field of this window indicates that you want to power on the bricks that reside in slots 07, 15, 27, and 35 of rack 001.
If you want to power on a partition, select Partition from the destination selection window. The partition selection window, shown in Figure 1-10, appears.
You can select all partitions by selecting ALL on the partition selection window, or you can select a single partition or multiple partitions by selecting the individual partition numbers. Figure 1-11 shows partition 001 selected.
The New DEST: field shows p 1, which indicates that partition 001 was selected as the new destination. If you select Apply, your selection is confirmed and the home window, shown in Figure 1-12, appears. The DEST: field shows p 1 [2 Bricks], which indicates that the bricks in partition 1 are the new destination.
After you have selected the destination of the bricks you want to power on, select Power UP on the home window, and the power up confirmation window, shown in Figure 1-13, appears. This window indicates which bricks will receive the Power UP command. In this example, the window indicates that all slots (bricks) in all racks (r * s*) will be powered on. If you select OK, the power-up operation is confirmed, and the home window appears. Selecting Cancel stops the power-on operation, and the home window appears.
To power on your server at the system console, follow these steps:
At your console, switch to L2 mode by pressing Ctrl+T.
From the L2 prompt (L2>), power on an individual brick by entering the following command. (If you want to power on the entire server, proceed to step 3.)
L2> r <rack#> s <slot#> pwr u |
For example, to power on a Cx-brick in rack 1, slot 07, enter the following:
L2> r 1 s 7 pwr u |
The slot number is the lowest unit number that the brick occupies within a rack. For example, in Figure 1-14, the lowest Cx-brick is identified as U07. The next lowest Cx-brick is identified as U11, and so on.
If you want to power on several selected bricks from a rack at the same time, you must enter the rack number followed by the slot numbers of the bricks you want to power on. For example, to power on bricks in slots 7 and 11 of rack 4, enter the following command:
L2> r 4 s 7,11 pwr u |
If you want to power on the bricks that reside in the same slot in multiple racks, you must enter the number of the racks followed by the slot number of the bricks you want to power on. For example, to power on the bricks in slot 11 of racks 3 and 4, enter the following command:
L2> r 3, 4 s 11 pwr u
![]() | Note: To avoid problems with your server, do not try to power on multiple slots for multiple racks at the same time. |
If you want to power on the entire server, enter the following command:
L2> pwr u |
(The default setting for the pwr u command is all racks and all slots.)
From the L2 prompt, display the system configuration by entering the following command:
L2> config |
This command lists the bricks in the server and each brick's system controller address.
Before powering on the D-brick2, confirm that the following is true:
Drives are seated in the correct bays and blank plates are fitted in any empty bays.
Ambient temperature is within the specified range of 10 °C to 40 °C (50 °F to 104 °F).
The rack power distribution unit (PDU) and power distribution strip (PDS) are on.
To power on the D-brick2, follow these steps:
Connect an AC power cord to each PSU/cooling module.
Connect the AC power cords to the PDS.
Turn the power switch on each PSU/cooling module to the “on” position (1 = on, 0 = off). The location of the power switch on the D–brick2 PSU/cooling module is shown in Figure 1-15.
The green “PSU good” LED illuminates. Also, the “power on” LED on the ESI/ops panel of each module turns green when AC power is present.
If the “power on” LED on the ESI/ops panel does not illuminate, or if the amber “system/ESI fault” LED illuminates, verify that you followed all steps. For troubleshooting tips, see the SGI Total Performance 9100 (2Gb TP9100) Storage System User's Guide or contact your service provider.
You can power off individual bricks or your entire server from the L2 controller touch display (located on the front door of rack 001) or from the system console, as explained in the sections that follow.
![]() | Note: To verify that the power-off procedure is proceeding properly, make sure that the On/Off LEDs and the 48-VDC LEDs turn off and that your L1 controllers display that the server is powering off for each segment of the procedure. If you have a problem while powering off and an error message appears on your L2 controller touch display, your console, or the L1 controller display, see your online log files and the information in “L1 Controller Error Messages” in Chapter 15 to learn what the error message indicates and how to resolve the problem. |
If you are logged on to the server, log out.
If you are planning to power down a D-brick2 or an entire rack of D-brick2s, see “Powering Off D-brick2”.
This section describes how to use the L2 controller touch display to power off individual bricks or the entire server. If you have multiple racks whose L2 controllers are interconnected at an Ethernet hub, you can power off any brick in those racks or the entire server at the L2 controller touch display on the front door of rack 001.
For instructions on how to use the L2 controller touch display, see “L2 Controller ” in Chapter 8.
To power off selected bricks, a partition, or the entire server, follow these steps:
Select the DEST: button from the home window and the destination selection window, shown in Figure 1-17, appears. Use this window to select the bricks or partition you want to power off.
The bricks are selected by their rack and slot/unit number, or by partition. Select All in the display if you want to power off all the bricks in all the racks in the server. You can also select the bricks in all the slots of multiple racks, but you cannot select slots for multiple racks.
If you want to power off individual bricks, select the rack that contains the bricks that you want to power off from the Rack/Slot segment of the display. A box appears that lists the slot numbers for all of the bricks in the selected rack (see Figure 1-18).
Select the slot number of each brick that you want to power off. (Figure 1-18 shows slots 007, 015, 027, and 035 selected). The New DEST: field lists your selections. After you complete your selections, select Apply.
The home window shown in Figure 1-19 appears. The DEST: field of this window indicates that you want to power off the bricks that reside in slots 07, 15, 27, and 35 of rack 001.
If you want to power off a partition, select Partition from the destination selection window. The partition selection window, shown in Figure 1-20, appears.
You can select all partitions by selecting ALL on the partition selection window, or you can select a single partition or multiple partitions by selecting the individual partition numbers. Figure 1-21 shows partition 001 selected.
The New DEST: field shows p 1, which indicates partition 001 was selected as the new destination. If you select Apply, your selection is confirmed and the home window, shown in Figure 1-22, appears. The DEST: field shows p 1 [2 Bricks], which indicates that the two bricks in partition 1 are the new destination.
After you have selected the destination of the bricks you want to power off, select Power DOWN from the home window; the power down confirmation window, shown in Figure 1-23, appears. This window indicates which bricks will receive the Power DOWN command. In this example, the window indicates that all slots (bricks) in all racks (r * s*) will be powered off. If you select OK, the power-off operation is confirmed and the home window appears. Selecting Cancel stops the power-off operation and the home window appears.
To power off your server at the system console, follow these steps:
At your console, switch to L2 mode by pressing Ctrl+T.
From the L2 prompt (L2>), power off an individual brick by typing the following command. (If you want to power off the entire server, proceed to the next step.)
L2> r <rack#> s <slot#> pwr d |
For example, to power off a Cx-brick in rack 1, slot 07, type the following:
L2> r 1 s 7 pwr d |
The slot number is the lowest unit number that the brick occupies within a rack (see Figure 1-14).
If you want to power off several bricks from a rack at the same time, you must type the rack number followed by the slot numbers of the bricks you want to power off. For example, to power off bricks in slots 7 and 11 of rack 4, type the following:
L2> r 4 s 7,11 pwr d |
If you want to power off bricks that reside in the same location in multiple racks, you must type the number of the racks followed by the slot number of the bricks you want to power off for each rack. For example, to power off the bricks in slot 11 of racks 3 and 4, type the following:
L2> r 3, 4 s 11 pwr d
![]() | Caution: To avoid problems with your server, do not try to power off multiple slots for multiple racks at the same time. |
If you want to power off all the bricks in all the racks, type the following command:
L2> pwr d (The default setting for the pwr d command is all racks and all slots.) |
From the L2 prompt, display the brick configuration information by typing the following command:
L2> config |
This command lists all the bricks in the server and each brick's system controller address.
The L1 controller display for each brick should display Powered Down once the power down procedure completes.
Before powering off the D-brick2, confirm that the following has occurred:
All system users have been notified and are logged off.
Disk data has been backed up as appropriate.
To power off a D-brick2 enclosure, follow these steps:
Move the power switch on the rear of each PSU/cooling module to the “off” position (position I = on, O = off). The location of the power switch on the D–brick2 PSU/cooling module is shown in Figure 1-24.
Unplug the power cable from the PSU/cooling module(s) as appropriate. For example, you would follow this step when you are replacing a module.
If you are shutting down all of the D-brick2s in a rack, you may want to move the PDU breaker switch to the “off” position.
The LEDs on the back of the unit should turn dark a few seconds after you power off the PSU/cooling module.
You can monitor your SGI Origin 3900 server from the following sources:
L1 controller display - All bricks (except the TP900 and D-brick2) contain an L1 controller display that displays information about the brick. For example, you can see if the fans on a particular brick are operating properly.
If your server has a system console, you can view the status and error messages generated by the L1 and L2 controllers in your server. You can also use the system console to input L1 and L2 commands to manage and monitor your server.
If your server has SGIconsole, you can view the status and error messages generated by the L1 and L2 controllers in your server and you can use various software tools, such as VACM, Console Manager, and PCP, to manage and monitor your server. See the SGIconsole Start Here guide for descriptions of these tools and for references to other documents with information about these tools.