SPARC Enterprise™ M8000/M9000 Servers Service Manual Manual Code C120-E330-09EN Part No.
Copyright 2007, 2010 FUJITSU LIMITED, 1-1, Kamikodanaka 4-chome, Nakahara-ku, Kawasaki-shi, Kanagawa-ken 211-8588, Japan. All rights reserved. Sun Microsystems, Inc. provided technical input and review on portions of this material. Sun Microsystems, Inc.
Copyright 2007, 2010 FUJITSU LIMITED, 1-1, Kamikodanaka 4-chome, Nakahara-ku, Kawasaki-shi, Kanagawa-ken 211-8588, Japon. Tous droits réservés. Entrée et revue tecnical fournies par Sun Microsystems, Incl sur des parties de ce matériel. Sun Microsystems, Inc. et Fujitsu Limited détiennent et contrôlent toutes deux des droits de propriété intellectuelle relatifs aux produits et technologies décrits dans ce document.
Contents Preface 1. xvii Safety and Tools 1.1 Symbols 1–1 1–1 1.1.1 Text Conventions 1–1 1.1.2 Prompt Notations 1–2 1.1.2.1 1–2 1.1.3 Environmental Requirements for Using This Product 1.1.4 Conventions for Alert Messages 1.1.4.1 1.2 Command syntax Precautions 1–3 Alert Messages in the Text Operating Environment of the Product 1.2.2 Maintenance 1.2.3 Conversion and Reuse of This Product Tools Required for Maintenance 1.
2. Product Overview and Troubleshooting 2.1 2.2 2.3 System Views 2–1 2.1.1 SPARC Enterprise M8000 Server 2.1.2 SPARC Enterprise M9000 Server (Base Cabinet) 2.1.3 SPARC Enterprise M9000 Server (Expansion Cabinet) Labels 2–3 System Name Plate Label, Rating Label, ID Label (Japan) or EZ Label (besides Japan), and Standard Label 2–11 2.2.2 Labels About Handling Operator Panel 2–15 2–16 2.3.1 Operator Panel Location 2–16 2.3.2 Appearance and Operations 2.3.3 LED 2.3.
2.8 Using the Troubleshooting Commands 2.8.1 Using the showlogs Command 2.8.2 Using the fmdump Command 2.8.3 Using the fmadm Command 2.8.4 2.9 2–34 2–35 2–35 2.8.3.1 fmadm config Command 2–35 2.8.3.2 fmadm faulty Command 2–36 2.8.3.3 fmadm repair Command 2–36 Using the fmstat Command 2–37 Traditional Solaris Troubleshooting Commands 2.9.1 2.9.2 iostat Command 2–38 2.9.1.1 2–38 2.9.3 2–41 2–42 Options 2–43 2–43 2–44 Options ps Command 2.9.6.1 2.9.7 Options ping Command 2.
4. FRU Removal Preparation 4.1 Types of Replacement Procedures 4.1.1 4.2 4–3 Disconnecting a FRU from a Domain 4–3 4.2.1.2 Disconnecting a PCI card 4–4 4.2.2 Disconnecting and Replacing a FRU 4–5 4.2.3 Configuring a FRU into a Domain 4–6 4.2.3.1 Configuring CMU/IOU 4–6 4.2.3.2 Configuring a PCI card Confirming the Hardware Hot Replacement 4.3.2 Confirming the Hardware 4–12 4–15 4–18 4.4.1 Powering the Server Off 4–18 4.4.2 Powering the Server On 4–19 4.4.2.
4.5.3 6. 4–27 4.5.3.1 SPARC Enterprise M8000 Server Single-Phase Power Feed 4–28 4.5.3.2 SPARC Enterprise M8000 Server Single-Phase and Dual Power Feed 4–29 4.5.3.3 SPARC Enterprise M8000 Server Three-Phase Power Feed 4–30 4.5.3.4 SPARC Enterprise M9000 Server Single-Phase Power Feed 4–31 4.5.3.5 SPARC Enterprise M9000 Server Single-Phase and Dual Power Feed 4–32 4.5.3.6 SPARC Enterprise M9000 Server Three-Phase Power Feed 4–33 4.6 Emergency Switch-Off 4.
6.2.1 SPARC64 VII CPU Modules Added to a New Domain 6.2.2 SPARC64 VII Processors Added to an Existing Domain 7. 8. 9. x 6–8 6.2.2.1 Preparing to Add SPARC64 VII Processors to an Existing Domain 6–8 6.2.2.2 Adding a SPARC64 VII CPU Module to a Domain Configured With SPARC64 VI 6–10 6.2.2.3 Upgrading a SPARC64 VI CPU Module to SPARC64 VII on an Existing Domain 6–11 6.3 Active Replacement and Hot Replacement 6.4 Cold Replacement 6.5 CPU and DIMM Replacement 6–12 6–23 6–29 6.5.
10. 11. 12. 13. 14. 15. Operator Panel Replacement 10–1 10.1 Overview of the Operator Panel 10.2 Cold Replacement 10–4 XSCF Unit Replacement 11–1 10–1 11.1 Overview of the XSCFU 11–1 11.2 Active Replacement and Hot Replacement 11.3 Cold Replacement 11–5 11–12 Hard Disk Drive (HDD) Replacement 12.1 Overview of the HDD 12.2 Active Replacement 12.3 Cold Replacement 12–1 12–1 12–3 12–5 PCI Slot Device Replacement 13–1 13.1 Overview of PCI Slot Devices 13.
16. 17. 18. 19. 20. 21. 22. xii Clock Control Unit Replacement 16.1 Overview of the CLKU 16.2 Cold Replacement 16–1 16–4 Crossbar Unit Replacement 17–1 17.1 Overview of XBUs 17–1 17.2 Cold Replacement 17–4 AC Section Replacement 18–1 18.1 Overview of ACSs 18–1 18.2 Cold Replacement 18–5 DDC Replacement 16–1 19–1 19.1 Overview of the DDC 19.2 Active Replacement and Hot Replacement 19.3 Cold Replacement 19–7 Backplane Replacement 20–1 20.1 Overview of the BP 20.
23. 24. Switch Backplane Replacement 23.1 Overview of SWBPs 23.2 Cold Replacement 23–1 23–5 Addition and Deletion of a RDPF Option, Power Cabinet, and M9000 Expansion Cabinet 24–1 24.1 25. 23–1 Addition and Deletion of Rack-mountable Dual Power Feed 24.1.1 Overview of RDPF 24.1.2 Addition and Deletion 24–2 24–2 24–3 24.1.2.1 Addition 24–3 24.1.2.2 Deletion 24–13 24.2 Addition and Deletion of Power Cabinet 24–13 24.
B. Components B–1 B.1 CPU Memory Board Unit B.2 CPU Module B.3 Memory B–7 B.4 I/O Unit B–8 B.5 Hard Disk Drive B.6 PCI Cassette B.7 IOU Onboard Device Card B.8 Link Card (External I/O Expansion Unit Connection Card) B.9 Crossbar Unit B.10 Clock Control Unit B.11 XSCF Unit B.12 CD-RW/DVD-RW Drive Unit B.13 Tape Drive Unit B.14 Operator Panel B.15 Sensor Unit B.16 Power Supply Unit B.17 AC Section B.18 FAN Unit B.19 Power Cabinet B.
C.1 Serial Port C–1 C.2 UPC Port C–2 C.3 USB Port C–2 C.4 Connection Diagram for Serial Cable D. UPS Controller C–3 D–1 D.1 Overview D–1 D.2 Signal Cable D.3 Configuration of Signal Lines D.4 Power Supply Conditions D–1 D.4.1 Input Circuit D.4.2 Output Circuit D.5 UPS Cable D.6 Connections D.7 UPC Port D–2 D–4 D–4 D–4 D–5 D–6 D–7 E.
xvi SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
Preface This manual a maintenance manual for the SPARC Enterprise™ M8000/M9000 servers. The manual explains basic operations and detailed replacement procedures for field-replaceable units (FRUs), which are components that can be replaced at the customer's site. As a rule, Fujitsu certified service engineers on the SPARC Enterprise M8000/M9000 servers. However, another person such as the system administrator may perform some of the simple work, under the direction of one of the service engineers.
Structure and Contents of This Manual This manual is organized as described below: ■ PART I Basic Information for Maintenance and Troubleshooting Provides notes on handling the SPARC Enterprise servers and rules about operation and descriptions, and it also describes the required tools for maintenance. ■ Chapter 1 Safety and Tools: Provides notes on handling the SPARC Enterprise servers and rules about operation and descriptions, and it describes the required tools for maintenance.
■ Chapter 9 Power Supply Unit (PSU) Replacement: Explains the replacement procedures for a power supply unit (PSU). ■ Chapter 10 Operator Panel Replacement: Explains the replacement procedures for the operator panel. ■ Chapter 11 XSCF Unit Replacement: Explains the replacement procedures for an XSCF unit. ■ Chapter 12 Hard Disk Drive (HDD) Replacement: Explains the replacement procedures for a hard disk drive (HDD).
■ Chapter 24 Addition and Deletion of a Rack-mountable Dual Power Feed Option, Power Cabinet, and M9000 Expansion Cabinet: Explains the replacement procedures for rack-mountable dual power feed(RDPF). ■ Chapter 25 Addition and Deletion of CMU, DIMM, IOU, HDD, PCI Cards and TAPEU: Explains the procedures for adding a unit to the SPARC Enterprise M8000/M9000 servers and deleting a unit from the SPARC Enterprise M8000/M9000 servers.
SPARC Enterprise M8000/M9000 Servers Documentation The manuals listed below are provided for reference.
xxii Book Titles Manual Codes External I/O Expansion Unit Product Notes C120-E456 SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers Glossary C120-E514 SPARC Enterprise /PRIMEQUEST Common Installation Planning Manual C120-H007 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
1. Manuals on the Web The latest versions of all the SPARC Enterprise Series manuals are available at the following websites. Global Site http://www.fujitsu.com/sparcenterprise/manual/ Japanese Site http://primeserver.fujitsu.com/sparcenterprise/manual/ Note – Product Notes are available on the website only. Please check for the most recent update on your product. 2. Documentation CD For the Documentation CD, please contact your local sales representative.
Product Handling Maintenance Caution – Certain tasks in this manual should only be performed by a certified service engineer. User must not perform these tasks. Incorrect operation of these tasks may cause electric shock, injury, or fire.
Emission of Laser Beam (Invisible) Caution – The main unit and high-speed optical interconnect cabinet contain modules that generate invisible laser radiation. Laser beams are generated while the equipment is operating, even if an optical cable is disconnected or a cover is removed. Do not look at any light-emitting part directly or through an optical apparatus (e.g., magnifying glass, microscope).
Limitations and Cautions Power Control and Operator Panel Mode Switch When you use the remote power control utilizing the RCI function or the automatic power control system (referred to below as APCS), you can disable this remote power control or the APCS by switching to Service mode on the operator panel. Disabling these features ensures that you do not unintentionally switch the system power on or off during maintenance. Note system power off with the APCS cannot be disabled with the mode switch.
Fujitsu Welcomes Your Comments If you have any comments or requests regarding this document, or if you find any unclear statements in the document, please state your points specifically on the form at the following URL. For Users in U.S.A., Canada, and Mexico: http://www.computers.us.fujitsu.com/www/support_servers.shtml?support /servers For Users in Other Countries: SPARC Enterprise contact http://www.fujitsu.com/global/contact/computing/sparce_index.
xxviii SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
PA RT I Basic Information for Maintenance and Troubleshooting Part I provides maintenance information, explains methods of problem analysis, troubleshooting, and basic operations for replacing FRUs.
CHAPTER 1 Safety and Tools This chapter provides notes on handling the high-end server descriptions and rules about operation, and it lists the tools required for maintenance. This information is explained in the following sections: ■ ■ ■ ■ Section 1.1, Section 1.2, Section 1.3, Section 1.4, “Symbols” on page 1-1 “Precautions” on page 1-4 “Tools Required for Maintenance” on page 1-5 “Antistatic Precautions” on page 1-6 1.1 Symbols 1.1.
1.1.2 Fonts/symbols Meaning Example AaBbCc123 The names of commands, files, and directories; on-screen computer output. This font represents the example of command input in the frame. XSCF> showuser -P User Name: jsmith Privileges: useradm auditadm Italic Indicates the name of a reference manual. See the SPARC Enterprise M3000/M4000/M5000/M8000/M 9000 Servers XSCF User’s Guide . "" Indicates names of chapters, sections, items, buttons, or menus.
1.1.3 Environmental Requirements for Using This Product This product is a computer that is intended to be used in a computer room. For details on the operational environment, see the SPARC Enterprise M8000/M9000 Servers Site Planning Guide. 1.1.4 Conventions for Alert Messages This manual uses the following conventions to show alert messages, which are intended to prevent injury to the user or bystanders as well as property damage, and important messages that are useful to the user.
1.2 Precautions The following notes must be observed in maintenance work: 1.2.1 Operating Environment of the Product Use the SPARC Enterprise in the correct operating environment. The SPARC Enterprise are assumed to be used in a computer room. For details of the operating environment, see the SPARC Enterprise M8000/M9000 Servers Site Planning Guide. 1.2.2 Maintenance The work listed below is to be performed by authorized service engineers.
1.2.3 Conversion and Reuse of This Product If this product is converted or a used article of this product is overhauled for the purpose of reuse, unexpected injury on users and bystanders or damage to their property may result. 1.
Caution – Be sure to wear an antistatic wrist strap when handling components. Place removed components on an antistatic mat. Failure to do so may result in serious damage or injury. 1.4 Antistatic Precautions During normal operations, all components mounted in the server, including the dummy (filler) units, are properly grounded through the chassis. Prior to performing maintenance, ensure that any static electricity is discharged from the FRUs to be inserted and the person performing the maintenance.
FIGURE 1-1 Conductive Mat 2. Connect an antistatic wrist strap clip to a server grounding port. See Section 1.4.2, “Grounding Port Connection Locations” on page 1-11. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat. Failure to take these antistatic measures might result in serious damage.
3. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. FIGURE 1-2 Antistatic Wrist Strap Showing the Metallic Underside metallic surface 1.4.1.1 Removing Static Electricity on a CMU and an IOU 1. Prior to mounting a new CMU or IOU, place it on the grounded antistatic mat.
2. Touch the metallic chassis for 5 or more seconds with your bare hand wearing the antistatic wrist strap. (See FIGURE 1-3 or FIGURE 1-4) You cannot remove static electricity by touching the label. FIGURE 1-3 Metallic Chassis (CMU) Touch for 5 or more seconds. FIGURE 1-4 Metallic Chassis (IOU) Touch for 5 or more seconds.
3. Touch each of the designated points on the guide blocks for 5 or more seconds with your bare hand wearing the antistatic wrist strap. (See FIGURE 1-5 or FIGURE 1-6) FIGURE 1-5 Guide Block (CMU) Touch for 5 or more seconds. FIGURE 1-6 Guide Block (IOU) Touch 5 or more seconds.
1.4.2 Grounding Port Connection Locations FIGURE 1-7 to FIGURE 1-10 show the locations of the grounding port on each server. The grounding port can be used for the grounding of the antistatic wrist strap and the antistatic mat. If the type of clip does not securely fit the grounding port, it can also be connected to the grounding wire for the door. For details, see Section 5.3, “How to Remove a Door” on page 5-3.
1-12 FIGURE 1-8 M8000 Grounding Port Connection Locations for the Wrist Strap Clip and the Antistatic Mat (Rear View) FIGURE 1-9 M9000 Grounding Port Connection Locations for the Wrist Strap Clip and the Antistatic Mat (Front View) SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 1-10 M9000 Grounding Port Connection Locations for the Wrist Strap Clip and the Antistatic Mat (Rear View) Chapter 1 Safety and Tools 1-13
1-14 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 2 Product Overview and Troubleshooting This chapter provides information that is required in troubleshooting. This information is explained in the following sections: ■ ■ ■ ■ ■ ■ ■ ■ ■ 2.1 Section 2.1, Section 2.2, Section 2.3, Section 2.4, Section 2.5, Section 2.6, Section 2.7, Section 2.8, Section 2.
The front and rear views of FIGURE 2-1, FIGURE 2-2, FIGURE 2-4, FIGURE 2-5, FIGURE 2-7, and FIGURE 2-8 include names and abbreviations for field-replaceable units (FRUs). Components that are mounted inside the system are shown FIGURE 2-3, FIGURE 2-6, and FIGURE 2-9. The abbreviations are used in messages and the like. If multiple FRUs of the same type are mounted, the number sign # and a sequential number is added to their names to distinguish them from one another.
2.1.
FIGURE 2-2 Rear View - M8000 ACS FAN_B IOU Air Filter 2-4 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 2-3 Internal View - M8000 PSUBP_B PSUBP_A FANBP_C BP_A MEDBP SWBP FANBP_C Chapter 2 Product Overview and Troubleshooting 2-5
2.1.
FIGURE 2-5 Rear View - M9000 (Base Cabinet) FAN_A CMU IOU Air Filter Chapter 2 Product Overview and Troubleshooting 2-7
FIGURE 2-6 Internal View - M9000 (Base Cabinet) FANBP_B PSUBP_A FANBP_A BP_B MEDBP SWBP 2-8 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
2.1.
FIGURE 2-8 Rear View - M9000 (with the Expansion Cabinet) FAN_A CMU IOU Air Filter 2-10 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 2-9 Internal View - M9000 (with the Expansion Cabinet) FANBP_B PSUBP_A FANBP_A BP_B MEDBP SWBP 2.2 Labels 2.2.1 System Name Plate Label, Rating Label, ID Label (Japan) or EZ Label (besides Japan), and Standard Label The important labels affixed on this server are shown in FIGURE 2-10 and FIGURE 2-11. The actual description on the labels may differ from FIGURE 2-10 and FIGURE 2-11.
■ The ID label or EZ label is affixed on the front door of the server, and it includes the model name and serial number, both of which are written on the system name plate label.
FIGURE 2-10 M8000 Label Location System Name Plate Label Front Rear Standard label Chapter 2 Product Overview and Troubleshooting 2-13
FIGURE 2-11 M9000 Label Location System Name Plate Label Front Rear Standard label 2-14 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
2.2.2 Labels About Handling The labels shown below, which are affixed on the high-end server, provide field engineers with important information on component removal and mounting. Caution – Never peel off the labels.
■ 2.3 Removing an I/O unit (IOU) Operator Panel The operator panel controls the high-end server power. The operator panel is usually locked with a key to prevent the server from being mistakenly powered off through an operator error during system operation. Before starting maintenance work, ask the system administrator to unlock the operation panel. 2.3.1 Operator Panel Location FIGURE 2-12 indicates the location of the operator panel (OPNL) of the high-end servers.
FIGURE 2-12 Operator Panel Location (at the Front of M8000) OPNL 2.3.2 Appearance and Operations The operator panel can be used while the front door of the server is closed. Field engineers, and the system administrator use the operation panel to check the operating state of the server and to perform system power operations. To check the operating state of the server, look at the LEDs.
FIGURE 2-13 2.3.3 Operator Panel LED TABLE 2-1 lists the states of the server that are displayed with the LEDs on the operator panel. The blinking period is one second (frequency of 1 Hz). Besides the states listed in TABLE 2-1, the operator panel also displays various states of the server using combinations of the three LEDs. TABLE 2-2 indicates the states that are usually displayed in the course of operation from the power-on to power-off of the high-end server.
TABLE 2-1 LED State Display by the LEDs (Operator Panel) (Continued) Name Light color Description of function and state CHECK Amber Indicates the operating status of the SPARC Enterprise server. Off Normal state. Otherwise, this indicates that the main line switches were switched off or a power failure occurred. Blinking (*1) Indicates that the operator panel is the maintenance target device. Lit Indicates that the server cannot be started.
2.3.4 Switch The operator panel has the mode switch, which sets the operation mode, and the POWER switch, which is used to power on and off the system. TABLE 2-3 Switch Switches (Operator Panel) Name Description of function Mode This key switch is used to set an operation mode for the server. Insert the special key that is under the customer’s control, to switch between modes.
TABLE 2-4 Meanings of the Mode Switch Function Mode Switch Locked Locked Inhibition of Break Signal Reception Enabled. Reception of the break signal can be enabled or disabled for each domain using setdomainmode. Disabled Power On/Off by power switch Only power on is enabled Enabled 2.4 Determining Which Diagnostics Methods To Use When an error occurs, a message is often displayed on the monitor. Use the flowcharts in FIGURE 2-14 to find the correct methods for diagnosing problems.
FIGURE 2-14 Diagnostic Method Flow Chart Start OS panic occurred or there’s an error on performance YES e-mail sent or not by XSCF mail function? NO Check OS console and XSCF console for error information displayed NO Execute showlogs or fmadm on XSCF to display the fault information Check /var/adm/messages on Solaris OS FMA message? YES Is there error message on XSCF console? NO YES Use fmadm ? YES NO Execute fmadm to display fault information NO Message ID available? Write down the displa
2.5 Checking the Server and System Configuration Before and after maintenance work, the state and configuration of the server and components should be checked and the information saved. For recovery from a problem, conditions related to the problem and the repair status must be checked. The operating conditions must remain the same before and after maintenance. A functioning server without any problems should not display any error conditions. For example: 2.5.
TABLE 2-5 Commands for Checking Hardware Configuration and Status Command Description showhardconf Displays the system layer that includes a faulty component. showstatus Displays the status of a component. This command is used to check only a faulty component. showboards Displays the use status of individual devices and resources. showdcl Displays domain configuration information (hardware resource information). showfru Displays device setting information.
■ The software configuration and version can be checked in the Solaris Operating System. Refer to the Solaris OS documentation for more information. ■ The firmware configuration and versions can be checked from the maintenance terminal using XSCF Shell commands. Refer to the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide for more detailed information. Check the software and firmware configuration information with assistance from the system administrator.
2.5.2.2 Checking the Firmware Configuration Login authority is required to check the firmware configuration. The following procedure for these checks can be made from the maintenance terminal:. 1. Log in to the XSCF. 2. Execute the version command. XSCF> version The version command will print the firmware version information to the screen. See the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide for more detailed information. 2.5.
You can find more detailed descriptions of Solaris OS Predictive Self-Healing at the website below: http://www.sun.com/bigadmin/features/articles/selfheal.html Predictive self-healing is an architecture and methodology for automatically diagnosing, reporting, and handling software and hardware fault conditions. This new technology lessens the time required to debug a hardware or software problem and provides the administrator and technical support with detailed data about each fault. 2.6.
TABLE 2-8 Predictive Self Healing Message Output displayed Description Nov 1 16:30:20 dt88-292 EVENT-TIME: Tue Nov 1 16:30:20 PST 2005 EVENT-TIME: the time stamp of the diagnosis. Nov 1 16:30:20 dt88-292 PLATFORM: SUNW,A70, CSN: -, HOSTNAME: dt88-292 PLATFORM: A description of the server encountering the problem. Nov 1 16:30:20 dt88-292 SOURCE: eft, REV: 1.13 SOURCE: Information on the Diagnosis Engine used to determine the fault.
TABLE 2-9 Command Commands for Checking the Monitoring Output Operand showlogs(8) console 2.6.3 Description XSCF firmware collects console logs of console messages that were output through the XSCF. This command collects all the console messages displayed to users. monitor Logs the messages displayed in the message window of the BUI/CLI. panic Saves as panic logs the console logs that are logged when a reset is received after a panic notification.
Commands for Checking the Messaging Output TABLE 2-10 Command Operand Description showlogs env Collects the temperature history log. The SPARC Enterprise server environmental temperature data and power status are collected at a 10-minute interval. The data is stored for a maximum of six months. power Collects the log of power events and reset events. The target range covers the SPARC Enterprise server, External I/O Expansion units, and UPSs.
that has caused the error indicate that an error has occurred. The LEDs on the operator panel (back) indicate the status of the operator as a single unit. However, some FRUs like DIMMs do not have LEDs. Whether a FRU without LEDs is in the normal state can be checked by executing the XSCF Shell commands showhardconf and ioxadm from a maintenance terminal. For details of the commands, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF Reference Manual. 2.7.
TABLE 2-11 LED Display That Should Be Checked When a FRU Is Replaced (Common) (Continued) LED Display and meaning CHECK (amber) Indicates either that the unit contains an error or that the unit is a target device for replacement. Lit Indicates that an error has been detected in the hardware of the FRU. (For an HDD, the LED is lit according to the instruction from the software or middleware.) Blinking (*1) Indicates that the FRU is to be replaced. Off Indicates that the state of the FRU is normal.
TABLE 2-12 Status Display of LEDs Defined Individually for Each FRU (Continued) LED FRU Type Display Meaning XSCFU and IOU (display part for LAN) ACT Lit (green) Indicates that communication is being performed through the Ethernet port (LAN port). Off Indicates that no communication is being performed through the Ethernet port (LAN port). Lit (amber) Only for an IOU: Indicates that the communication speed is 1G bps. Lit (green) Indicates that the communication speed is 100M bps.
TABLE 2-12 Status Display of LEDs Defined Individually for Each FRU (Continued) LED FRU Type Display Meaning PSU: power supply unit POWER Lit (green) Indicates that the power to the system is turned on and being supplied. Blinking (green) Indicates that power is being supplied to the PSU, but the PSU is not turned on. Off Indicates that power is not being supplied to the PSU. Lit (amber) Indicates that an error occurred in the PSU. Maintenance can be performed.
■ ■ ■ console message log panic message log IPL message log XSCF> showlogs error Dated: Mar 30 12:45:31 JST 2005 Code: 00112233-44556677-8899aabbcceeff0 Status: Alarm Component: PSU#1,PSU#2 Msg: ACFAIL occurred (ACS=3)(FEP type = A1) Date: Mar 30 17:45:31 JST 2005 Code: 00112233-44556677-8899aabbcceeff0 Status: Faulted Component: PSU#1,PSU#2,* Msg: ACFAIL occurred (ACS=3)(FEP type = A1) XSCF>0 2.8.
# fmadm config MODULE cpumem-diagnosis cpumem-retire eft fmd-self-diagnosis io-retire syslog-msgs 2.8.3.2 VERSION 1.5 1.0 1.13 1.0 1.0 1.0 STATUS active active active active active active DESCRIPTION UltraSPARC-III/IV CPU/Memory Diagnosis CPU/Memory Retire Agent eft diagnosis engine Fault Manager Self-Diagnosis I/O Retire Agent Syslog Messaging Agent fmadm faulty Command The fmadm faulty command can be used mainly to identify the status of faulty components.
Note – The error information remains in the Solaris OS even when replacement of the faulty component is completed. Identify the UUID by executing the fmadm faulty command, and reset the error information by executing the fmadm repair command with the UUID specified. 2.8.4 Using the fmstat Command The fmstat command can report statistics associated with the Solaris Fault Manager. The fmstat command shows information about DE performance.
2.9.1 iostat Command The iostat command iteratively reports terminal, drive, and tape I/O activity, as well as CPU utilization. 2.9.1.1 Options TABLE 2-13 describes options for the iostat command and how those options can help troubleshoot the server. TABLE 2-13 Options for iostat Option Description How It Can Help No option Reports status of local I/O devices. A quick three-line output of device status.
The following example shows output for one iostat command. # iostat -En c0t0d0 Soft Errors: 0 Hard Errors: 0 Transport Errors: 0 Model: ST3120026A Revision: 8.01 Serial No: 3JT4H4C2 Size: 120.03GB <120031641600 bytes> Media Error: 0 Device Not Ready: 0 No Device: 0 Recoverable: 0 Illegal Request: 0 c0t2d0 Soft Errors: 0 Hard Errors: 0 Transport Errors: 0 Vendor: LITE-ON Product: COMBO SOHC-4832K Revision: O3K1 Serial No: Size: 0.
The following example shows output for the prtdiag command in verbose mode.
==================== Hardware Revisions ==================== System PROM revisions: ---------------------OBP 4.24.4 2007/11/05 10:27 =================== Environmental Status =================== Mode switch is in UNLOCK mode =================== System Processor Mode =================== SPARC64-VII mode # 2.9.3 prtconf Command Similar to the show-devs command run at the ok prompt, the prtconf command displays the devices that are configured.
2.9.3.1 Options TABLE 2-15 describes options for the prtconf command and how those options can help troubleshooting. TABLE 2-15 Options for prtconf Option Description How It Can Help No option Displays the device tree of devices recognized by the OS. If a hardware device is recognized, then it is probably functioning properly. If the message "(driver not attached)" is displayed for the device or for a sub-device, then the driver for the device is corrupt or missing.
pseudo-console, instance #0 nvram (driver not attached) pseudo-mc, instance #0 pseudo-mc, instance #1 pseudo-mc, instance #4 cmp (driver not attached) core (driver not attached) cpu (driver not attached) cpu (driver not attached) (The rest is omitted.) 2.9.4 netstat Command The netstat command displays the network status. 2.9.4.1 Options TABLE 2-16 describes options for the netstat command and how those options can help troubleshooting.
The following example shows output for the netstat -p command. # netstat -p Net to Device -----bge0 bge0 bge0 bge0 bge0 bge0 bge0 bge0 bge0 bge0 bge0 bge0 bge0 bge0 2.9.5 Media Table: IPv4 IP Address ---------------------phatair-46 ns-umpk27-02-46 moreair-46 fermpk28a-46 fermpk28as-46 kayakr matlock toronto2 tocknett mpk28-lobby efyinisedeg froggy d-mpk28-46-245 224.0.0.0 Mask Flags Phys Addr ----------------- ----- --------------255.255.255.255 08:00:20:92:4a:47 255.255.255.255 08:00:20:93:fb:99 255.
2.9.5.1 Options TABLE 2-17 describes options for the ping command and how those options can help troubleshooting. TABLE 2-17 Options for ping Option Description How It Can Help hostname The probe packet is sent to hostname and returned. Verifies that a host is active on the network. -g hostname Forces the probe packet to route through a specified gateway. By identifying different routes to the target host, those individual routes can be tested for quality.
2.9.6 ps Command The ps command lists the status of processes. Using options and rearranging the command output can assist in determining the resource allocation. 2.9.6.1 Options TABLE 2-18 describes options for the ps command and how those options can help troubleshooting. TABLE 2-18 Options for ps Option Description How It Can Help -e Displays information for every process. Identifies the process ID and the executable. -f Generates a full listing.
2.9.7 prstat Command The prstat utility iteratively examines all active processes and reports statistics based on the selected output mode and sort order. The prstat command provides output similar to the ps command. 2.9.7.1 Options TABLE 2-19 describes options for the prstat command and how those options can help troubleshooting. TABLE 2-19 Options for prstat Option Description How It Can Help No option Displays a sorted list of the top processes that are consuming the most CPU resources.
2-48 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 3 Periodic Maintenance Caution – The information in this chapter and subsequent chapters is intended for service engineers. Persons other than the authorized field engineers should use this information only for reference and should not actually perform the work. This chapter explains the maintenance work that must be performed regularly regardless of whether a problem has occurred. The actual work is limited to preventing dust in the environment from creating pollution.
Each tape drive unit used for operation must be cleaned once every 24 hours of operation. Even tape drive units not used for operation must be cleaned once every month. Although cleaning work can be performed in either hot or cold system maintenance mode, the SPARC Enterprise server power must be on when a cleaning cassette is used. The cleaning procedure is as follows. 1. If a tape cassette has been inserted in the tape drive unit, remove it from the unit. 2.
Note – One concern about cleaning the air filters while power to the server is on is that dislodged dust may be sucked inside the system when the air filters are pulled out. Therefore, gently and slowly pull them out. Complete the cleaning as quickly as possible. Caution – If you must use a vacuum cleaner for this work, use it outside the computer room. Do not use it inside the computer room. Using a vacuum cleaner inside the computer room may result in a server failure.
FIGURE 3-1 Removing Air Filters (Example for the M8000) Fixing bracket (x3) 3. Pull out all of the air filters. 4. Use a cleaner to remove dust from the air filters. Attach a brush to the tip of the cleaner, and clean both sides of the filters. 5. Restore each air filter to its original location and orientation, which means the knob is on the side closest to you and the arrow on the label points up (the latticework faces upward). 6.
FIGURE 3-2 Removing of Air Filters (Example Using the M9000 Base Cabinet) Fixing bracket (x3) 3.3 Cleaning an Air Filter (I/O Unit) Each I/O unit has two air filters. Clean them at the same time that the air filters of the server are cleaned. Caution – If you must use a vacuum cleaner for this work, use it outside the computer room. Do not use it inside the computer room. Using a vacuum cleaner inside the computer room may result in a server failure. The cleaning procedure is as follows. 1.
FIGURE 3-3 Removing of Air Filters (I/O Unit) 3. Use a vacuum cleaner to remove dust from the air filter. 4. After the cleaning is completed, follow the removal procedure in reverse order to mount it.
CHAPTER 4 FRU Removal Preparation This chapter explains the required basic operations for replacing components, in the following sections: ■ ■ ■ ■ ■ ■ ■ Section 4.1, Section 4.2, Section 4.3, Section 4.4, Section 4.5, Section 4.6, Section 4.
■ Section 4.6, “Emergency Switch-Off” on page 4-34 4.1 Types of Replacement Procedures 4.1.1 FRU Replacement The three types of replacement procedures explained below are supported for FRU replacement. Choose the most suitable replacement procedure according to the customer's system environment. ■ Active replacement A target FRU is operated while the Solaris OS of the domain to which the FRU belongs is operating. The target FRU is operated by using Solaris OS commands or XSCF commands.
4.2 Active Replacement In active replacement the Solaris OS must be configured to allow the component to be replaced. Active replacement has four stages: ■ ■ ■ ■ Section 4.2.1, Section 4.2.2, Section 4.2.3, Section 4.2.
4.2.1.2 Disconnecting a PCI card Caution – Before you remove the PCI cassette, make sure that there is no activity on the card in the cassette. Caution – In the PCI cassette part, when removing cables such as LAN cable, if your finger can't reach the latch lock of the connector, press the latch with a flathead screwdriver to remove the cable. Forcing your finger into the clearance can cause damage to the PCI card. 1. From the Solaris OS use the cfgadm command to get the component status: # cfgadm AP_ID ..
4.2.2 Disconnecting and Replacing a FRU 1. From the XSCF Shell prompt, use the replacefru command: XSCF> replacefru ---------------------------------------------------------------Maintenance/Replacement Menu Please select a type of FRU to be replaced. 1. CMU/IOU(CPU Memory Board Unit/IO Unit) 2. FAN(Fan Unit) 3. PSU(Power Supply Unit) 4. XSCFU(Extended System Control Facility Unit) 5.
Diagnostic tests for FAN_A#0 have started. [This operation may take up to 2 minute(s)] (progress scale reported in seconds) 0..... 30..... 60..... 90.....done -------------------------------------------------Maintenance/Replacement Menu Status of the replaced unit. FRU Status ------------- -------FAN_A#0 Normal -------------------------------------------------The replacement of FAN_A#0 has completed, normally.
1. Configuring into a domain To configure CMU or IOU into the domain, enter the following command from the terminal that is connected to the XSCF: XSCF> addboard -c configure -d 0 00-0 The system administrator permission is required for executing this command. 2. Checking resources Log in to the XSCF and execute the showboard command to verify that the target CMU or IOU has been correctly configured. 4.2.3.2 Configuring a PCI card 1.
Note – The display may vary depending on the XCP version. For details of the showhardconf command and showstatus command, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide, the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF Reference Manual, or the man page.
XSCF> showhardconf SPARC Enterprise M9000; + Serial:PA30601004; Operator_Panel_Switch:Locked; + Power_Supply_System:Single-1Phase; Ex:Single-1Phase; SCF-ID:XSCF#0; + System_Power:On; System_Phase:Cabinet Power On; Domain#0 Domain_Status:Running; CMU#0 Status:Normal; Ver:0101h; Serial:PP0616B579 ; + FRU-Part-Number:CA06629-D001 A4 ; + Memory_Size:128 GB; + Type:A CPUM#0-CHIP#0 Status:Normal; Ver:0101h; Serial:PP091505ZY ; + FRU-Part-Number:CA06620-D021 A3 + Freq:2.
Output for the showhardconf command (Continued) + Core:4; Strand:2; CPUM#1-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0620P552 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 ; + Freq:2.880 GHz; Type:32; + Core:4; Strand:2; CPUM#2-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0631Q396 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 ; + Freq:2.880 GHz; Type:32; + Core:4; Strand:2; CPUM#3-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0629H443 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 ; + Freq:2.
Output for the showhardconf command (Continued) OPNL#0 Status:Normal; Ver:0101h; Serial:PP06058246 ; + FRU-Part-Number:CA06629-D061 A1 PSU#0 Status:Normal; Serial:; + FRU-Part-Number:CA01022-0690; + Power_Status:On; : FANBP_A#0 Status:Normal; Ver:0101h; Serial:PP0607D266 ; + FRU-Part-Number:CA21123-B54X 003AC FAN_A#0 Status:Normal; Serial:PA0605B287; + FRU-Part-Number:CA06501-D023 A2 /371-2222-00 : FAN_A#15 Status:Normal; Serial:PA0605B303; + FRU-Part-Number:CA06501-D023 A2 /371-2222-00 : FANBP_B#0 Status:N
4.3 Hot Replacement In hot replacement the Solaris OS does not need to be configured to allow the component to be replaced. The domain might need to be stopped and restarted in the case of CD-RW/DVD-RW drive unit or tape drive units. 4.3.1 ■ Section 4.3.1, “Disconnecting and Replacing a FRU” on page 4-12 ■ Section 4.3.2, “Confirming the Hardware” on page 4-15 Disconnecting and Replacing a FRU 1. From the XSCF Shell prompt, specify the domain, and execute the poweroff command.
The command is menu-driven. The example continues using a FAN unit. Select [1,2|c:cancel] :1 --------------------------------------------------Maintenance/Replacement Menu Please select a FAN to be replaced. No. FRU Status --- --------------- -----------------1. FAN_A#0 Normal 2. FAN_A#1 Normal 3. FAN_A#2 Normal 4. FAN_A#3 Normal --------------------------------------------------Select [1-4|b:back] :1 You are about to replace FAN_A#0.
Diagnostic tests for FAN_A#0 have started. [This operation may take up to 2 minute(s)] (progress scale reported in seconds) 0..... 30..... 60..... 90.....done --------------------------------------------------Maintenance/Replacement Menu Status of the replaced unit. FRU Status -------------------FAN_A#0 Normal --------------------------------------------------The replacement of FAN_A#0 has completed, normally.
4.3.2 Confirming the Hardware 1.
Output for the showhardconf command (Continued) + Memory_Size:32 GB; + Type:B CPUM#0-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0608J517 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 + Freq:2.880 GHz; Type:32; + Core:4; Strand:2; CPUM#1-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0620P552 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 + Freq:2.880 GHz; Type:32; + Core:4; Strand:2; CPUM#2-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0631Q396 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 + Freq:2.
Output for the showhardconf command (Continued) XBU_B#0 Status:Normal; Ver:0201h; Serial:PP0641X324 ; + FRU-Part-Number:CA06620-D301 A6 : CLKU_B#0 Status:Normal; Ver:0101h; Serial:PP0542M679 ; + FRU-Part-Number:CA06629-D042 A1 : OPNL#0 Status:Normal; Ver:0101h; Serial:PP06058246 ; + FRU-Part-Number:CA06629-D061 A1 PSU#0 Status:Normal; Serial:; + FRU-Part-Number:CA01022-0690; + Power_Status:On; : FANBP_A#0 Status:Normal; Ver:0101h; Serial:PP0607D266 ; + FRU-Part-Number:CA21123-B54X 003AC FAN_A#0 Status:Norma
4.4 Cold Replacement In cold replacement all business operations are stopped. Cold replacement is the act of powering off the server and disconnecting input power. This is normally required for safety when the inside of the server is accessed. The work of cold replacement consists of the following three steps: 4.4.1 ■ Section 4.4.1, “Powering the Server Off” on page 4-18 ■ Section 4.4.2, “Powering the Server On” on page 4-19 ■ Section 4.4.
2. Back up the system files and data as necessary. 3. Log into the XSCF Shell and type the poweroff command XSCF> poweroff -a The following activities occur when the poweroff command is used: ■ The Solaris OS shuts down cleanly. ■ The server powers off to Standby mode (the XSCF Unit and one fan will still have power). For details of the command, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. 4. Switch off all main line switches of the AC section.
4.4.2.2 From the Maintenance Terminal Follow the power-on procedure below using the XSCF maintenance functions. 1. Make sure that the server has enough power supply units to run the desired configuration. 2. Switch on all main line switches of the AC section. 3. Make sure the XSCF Unit Ready LED is lit. 4. Log into the XSCF Shell and type the poweron command. XSCF> poweron Refer to the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide for details. 5.
XSCF> showhardconf SPARC Enterprise M9000; + Serial:PA30601004; Operator_Panel_Switch:Locked; + Power_Supply_System:Single-1Phase; Ex:Single-1Phase; SCF-ID:XSCF#0; + System_Power:On; System_Phase:Cabinet Power On; Domain#0 Domain_Status:Running; CMU#0 Status:Normal; Ver:0101h; Serial:PP0616B579 ; + FRU-Part-Number:CA06629-D001 A4 ; + Memory_Size:128 GB; + Type:A CPUM#0-CHIP#0 Status:Normal; Ver:0101h; Serial:PP091505ZY ; + FRU-Part-Number:CA06620-D021 A3 + Freq:2.
Output for the showhardconf command (Continued) CPUM#0-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0608J517 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 + Freq:2.880 GHz; Type:32; + Core:4; Strand:2; CPUM#1-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0620P552 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 + Freq:2.880 GHz; Type:32; + Core:4; Strand:2; CPUM#2-CHIP#0 Status:Normal; Ver:0901h; Serial:PP0631Q396 ; + FRU-Part-Number:CA06620-D051 A0 /371-4616-01 + Freq:2.
Output for the showhardconf command (Continued) XBU_B#0 Status:Normal; Ver:0201h; Serial:PP0641X324 ; + FRU-Part-Number:CA06620-D301 A6 : CLKU_B#0 Status:Normal; Ver:0101h; Serial:PP0542M679 ; + FRU-Part-Number:CA06629-D042 A1 : OPNL#0 Status:Normal; Ver:0101h; Serial:PP06058246 ; + FRU-Part-Number:CA06629-D061 A1 PSU#0 Status:Normal; Serial:; + FRU-Part-Number:CA01022-0690; + Power_Status:On; : FANBP_A#0 Status:Normal; Ver:0101h; Serial:PP0607D266 ; + FRU-Part-Number:CA21123-B54X 003AC FAN_A#0 Status:Norma
4.5 Power-On/Off of Main Line Switch 4.5.1 Types of Power Supply TABLE 4-1 shows how the components of the server vary depending on the type of power supply. For details on the electrical specifications, see the SPARC Enterprise M8000/M9000 Servers Site Planning Guide.
FIGURE 4-1 Internal Connection of Power Supply (Single-Phase Power Feed) Distribution panel Main line switch AC ACS (ON/OFF) (ON/OFF) Power supply unit PSU PSU PSU FIGURE 4-2 Internal Connection of Power Supply (Three-Phase Power Feed) Distribution panel Main line switch Switch Power supply unit PSU AC ACS ACS_D (ON/OFF) (ON/OFF) (ON/OFF) ACS_D PSU ACS_D PSU ACS_D PSU To use three-phase power feed, you need the three-phase power feed option and the power cabinet for mounting the
4.5.1.2 Power System There are two types of power system: single power and dual power. FIGURE 4-3 shows the conceptual diagram of power system. FIGURE 4-3 Power System (Single Power / Dual Power) Dual power feed Single power feed Cabinet Cabinet PSU PSU PSU ACS ACS ACS AC AC AC Distribution panel Distribution panel Distribution panel The dual power feed option receives power from two external AC power sources that are independent of each other, and it duplicates the input power system.
4.5.2.1 Power-On 1. Turn on all the main line switches of the expansion cabinet. If an optional power cabinet is connected, also turn on all the main line switches of the power cabinet. 2. Turn on all the main line switches of the base cabinet. If an optional power cabinet is connected, also turn on all the main line switches of the power cabinet. 4.5.2.2 Power-Off 1. Turn off all the main line switches of the base cabinet.
4.5.3.
4.5.3.
4.5.3.
4.5.3.
4.5.3.
4.5.3.
4.6 Emergency Switch-Off In emergency situations (such as smoke or flames are coming from the server), you need to immediately stop the use and turn off the power supply. Caution – If smoke or flames are coming from the server, switch off all main line switches immediately, or use the customer's distribution panel to immediately stop the power supply. Regardless of the job at hand, remember to give top priority to fire prevention. 4.
FIGURE 4-10 4.7.1.1 Example of the External I/O Expansion Units Mounted in the Rack Space for the SPARC Enterprise M8000 Server (front) Precautions For Cable Routing Note – In principle, run LAN cables to the external I/O expansion units along the route on the right side of the rack (right route). If the total number of LAN cables exceeds 72, place only 72 of the LAN cables along the right route, and place the remaining LAN cables along the route on the left side of the rack (left route).
Note – The SPARC Enterprise M8000 server comes with 22 short hook-and-loop fasteners and 4 long hook-and-loop fasteners for cables. Use the type of hook-and-loop fastener appropriate to the situation. Note – To do maintenance on the IOU or the FAN unit of the SPARC Enterprise M8000 server, push aside the cables. Caution – Do not place any cable over a PCI cassette at the front of an IOU in the SPARC Enterprise M8000 server. (Adjust the positioning to divide and bundle the cables.
4.7.1.2 When Three External I/O Expansion Units Mounted This section provides an example of the cable routing procedure applied when three external I/O expansion units are mounted in the rack space for the SPARC Enterprise M8000 server. Cable Routing Procedure The cable routing procedure explained below is an example of the recommended procedure.
2. Run the remaining cables along the left route.
3. Bind the cables connected to each external I/O expansion unit with a short hook-and-loop fastener. (See (1) in FIGURE 4-12.) Caution – Keep optical fiber cables as far apart from other cables as possible to prevent the optical fiber cables from being covered by the other cables. 4. Bind the bundles of cables with a long hook-and-loop fastener at location (2) shown in FIGURE 4-12 while taking care not to touch the main line switch of the ACS. 5.
FIGURE 4-12 4-40 Magnified View of Part A (M8000, rear view) SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
4.7.1.3 For Server Using Three-phase Power Feed Note – The server using three-phase power feed does not have any cable holder. In this case, fasten the cables by running the hook-and-loop fastener through the slit between the bracket at the cable holder mounting location and a column in the cabinet. (FIGURE 4-13 shows the fastening location on the left route.
10. Divide the remaining cables into three routes (about 12 cables per route), bind the cables on each route with a short hook-and-loop fastener, and allow them to hang from the fastener. (See (8) in FIGURE 4-14.) Caution – Adjust the positioning of the routes appropriately so that the cables do not cover a PCI cassette at the front of an IOU.
11. Fasten the cables running along the left route to the cable holder with a short hook-and-loop fastener. (See (9) in FIGURE 4-15.) 12. Bind each remaining bundle of hanging cables with a short hook-and-loop fastener. (See (10) in FIGURE 4-15.) 13. Also bind each of these bundles of cables with a short hook-and-loop fastener at a location near the floor panel. (See (11) in FIGURE 4-15.) Binding the cables near the floor panel facilitates cable handling through floor panel openings.
FIGURE 4-16 4-44 Completed Cable Routing with Three External I/O Expansion Units Mounted (M8000, rear view) SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
4.7.1.4 When One External I/O Expansion Unit Mounted This section provides an example of the cable routing procedure applied when one external I/O expansion unit is mounted in the rack space for the SPARC Enterprise M8000 server. Cable Routing Procedure The cable routing procedure explained below is an example of the recommended procedure.
FIGURE 4-17 Cable Routing When One External I/O Expansion Unit Mounted (M8000, rear view) Part A Ventilation slits Part B 4-46 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
2. Bind the cables from the external I/O expansion unit with a short hook-and-loop fastener while taking care not to touch the main line switch of the ACS. (See (1) in FIGURE 4-18.) 3. Fasten the cables with a short hook-and-loop fastener to the cable holder at location (2) shown in FIGURE 4-18. FIGURE 4-18 Magnified View of Part A (M8000, rear view) 4. Fasten the cables with a short hook-and-loop fastener to the cable holder at location (3) shown in FIGURE 4-19. 5.
FIGURE 4-19 4-48 Magnified View of Part B (M8000, rear view) SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 4-20 Completed Cable Routing with One External I/O Expansion Unit Mounted (M8000, rear view) Chapter 4 FRU Removal Preparation 4-49
4-50 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
PA RT II Maintenance Part II explains how to remove and replace FRUs. The chapters are organized according to the FRU mounting locations. Chapter 6 starts with an overview of FRUs, followed by explanations of the FRU removal and replacement procedures.
CHAPTER 5 Internal Components Access This chapter explains how to access each unit of the server, in the following sections: ■ ■ ■ ■ 5.1 Section 5.1, Section 5.2, page 5-2 Section 5.3, Section 5.4, “How to Open and Close Doors” on page 5-1 “Corresponding Components and Doors That Can Be Accessed” on “How to Remove a Door” on page 5-3 “How to Remove a Side Panel” on page 5-4 How to Open and Close Doors The high-end server is equipped with doors at the front and rear.
2. Push the projection down below the handle. The part below the handle projects outward. 3. Turn the handle counterclockwise and pull it out to open the door. To close the door, reverse the above steps for opening it. FIGURE 5-1 5.2 How to Open a Door (Front Door as an Example) Corresponding Components and Doors That Can Be Accessed TABLE 5-1 indicates which door should be opened to access a certain component. Components are represented by their abbreviated names.
TABLE 5-1 Model Relationship Between Components and Doors That Can Be Accessed Component (abbreviation) maintained through front door Component (abbreviation) maintained through rear door M8000 CMU, XSCFU, TAPEU, DVDU, DDC_A, PSU, IOU, ACS_A, and FAN_B (2-FAN #2 to #7) FAN_A (3-FAN), FAN_B (2-FAN #0 and #1), OPNL, SNSU, and RDPF M9000 CMU, IOU (odd ID), and FAN_A (3-FAN #4 PSU, IOU (even ID), XSCFU_B, XSCFU_C to #15) (expansion cabinet), CLKU_B, XBU_B, TAPEU, DVDU, PSU, ACS_B, FAN_A (3-FAN #0 to #3),
5. Keep the door in a safe place. To reinstall the door, reverse the above steps for removing it. FIGURE 5-2 How to Remove a Door (Front Door) Operator panel Cable holder Sensor unit 5.4 How to Remove a Side Panel To install an optional power supply cabinet or to connect an expansion cabinet to the M9000 base cabinet, the side panel of the cabinet must be removed. The side panel hangs from a bracket on the top of the cabinet, and the lower part of the side panel is secured with a screw or screws.
The procedure for removing the side panel is follows. 1. Using a Phillips screwdriver, completely loosen the one or two screws at the lower part of the side panel. 2. Lift the side panel, and remove it from the cabinet. 3. Keep the side panel in a safe place. To reinstall the side panel, reverse the above steps for removing it.
5-6 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 6 Replacement of CPU/Memory Board Unit (CMU), CPU, and DIMM This chapter explains the procedures for replacing the CPU/Memory board unit (CMU) and the CPUs and DIMMs mounted in the CMU. It covers the following: ■ ■ ■ ■ Section 6.1, Section 6.3, Section 6.4, Section 6.
For the version matching of firmware, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. FIGURE 6-1, FIGURE 6-2, and FIGURE 6-3 show the mounting locations of CMUs in the high-end server cabinets.
FIGURE 6-2 CMU Mounting Locations (Rear of the M9000 Base Cabinet) CMU#7 CMU#6 CMU#5 CMU#4 CMU#3 CMU#2 CMU#1 CMU#0 Chapter 6 Replacement of CPU/Memory Board Unit (CMU), CPU, and DIMM 6-3
CMU Mounting Locations (Rear of M9000 with the Expansion Cabinet) FIGURE 6-3 CMU#15 CMU#14 CMU#13 CMU#12 CMU#11 CMU#10 CMU#9 CMU#8 CMU#7 CMU#6 CMU#5 CMU#4 CMU#3 CMU#2 CMU#1 CMU#0 TABLE 6-1 lists the abbreviated names and component numbers of the CMUs.
6.2 CPU Upgrade This section describes the procedure to mount the SPARC64 VII processor on the SPARC Enterprise M8000/M9000 server. Caution – The SPARC64 VII processors are supported by the XCP firmware (beginning with XCP 1070) and Solaris software. For more information about these minimum software and firmware requirements, see the latest version of the Product Notes (no earlier than the XCP 1070 edition) for your server.
4. Collect an XSCF snapshot to archive system status prior to upgrade. This will be help in case any problem occurred in this procedure. XSCF> snapshot -t user@host:directory 5. Update the XCP version. For the firmware updating procedures, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. 6. After updating the XCP, reset the XSCF. XSCF> rebootxscf 7. After resetting the XSCF, log in to the XSCF again. 8. Mount the CPU module (CPUM) on the CMU for add-on.
13. Sets the following for the CMU. ■ Set up XSB. ■ Set up the domain ■ Set up the CPU operational mode of the domain. For each setting, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. 14. Use the setdomainmode(8) command, disable the autoboot function of the domain. Refer to the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide and the setdomainmode(8) man page for more information. 15. Power on the domains. XSCF> poweron -d domain_id 16.
6.2.2 SPARC64 VII Processors Added to an Existing Domain Adding SPARC64 VII Processors to an existing domain is a two step process. First you must prepare the system (see Section 6.2.2.1, “Preparing to Add SPARC64 VII Processors to an Existing Domain” on page 6-8) and then you must install the Processors using the instructions that correspond to your installation scenario. 6.2.2.1 Preparing to Add SPARC64 VII Processors to an Existing Domain 1.
8. Update the XCP version. For the firmware updating procedures, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. 9. After updating the XCP, reset the XSCF. XSCF> rebootxscf 10. After resetting the XSCF, log in to the XSCF again. 11. Power on the target domain, and apply OpenBoot PROM firmware. XSCF> poweron -d domain_id It stops at ok prompt. It is not necessary to start Solaris OS. 12. Check the updated OpenBoot PROM version. The OpenBoot PROM version of XCP 1082 is 02.
■ 6.2.2.2 If you are upgrading an existing SPARC64 VI CPU module in an existing domain to SPARC64 VII processors, continue with Section 6.2.2.3, “Upgrading a SPARC64 VI CPU Module to SPARC64 VII on an Existing Domain” on page 6-11. Adding a SPARC64 VII CPU Module to a Domain Configured With SPARC64 VI This procedure must be preceded by Section 6.2.2.1, “Preparing to Add SPARC64 VII Processors to an Existing Domain” on page 6-8. If you have not completed that procedure, do so before continuing. 1.
8. Confirm that the target domain has been correctly started. XSCF> showlogs power 9. Confirm no abnormality occurred by using showlogs error -v and showstatus(8) commands. XSCF> showlogs error -v XSCF> showstatus 6.2.2.3 Upgrading a SPARC64 VI CPU Module to SPARC64 VII on an Existing Domain This procedure must be proceeded by Section 6.2.2.1, “Preparing to Add SPARC64 VII Processors to an Existing Domain” on page 6-8. If you have not completed that procedure please do so before continuing. 1.
8. Confirm that the target domain has been correctly started. XSCF> showlogs power 9. Confirm no abnormality occurred by using showlogs error -v and showstatus(8) commands. XSCF> showlogs error -v XSCF> showstatus 6.3 Active Replacement and Hot Replacement The active replacement procedure and hot replacement procedure are almost the same. This section explains both procedures together.
b. If it is logically divided, execute the showdcl command to check whether the CMU to be replaced and another CMU comprise a domain. c. Execute the showboards command to check the use states of the devices and resources in the CMU. For details on how to log in to XSCF and commands, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. For details of commands, see also the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF Reference Manual, or the man page. 3.
4. Use the maintenance menu to select the CMU to be replaced. (The maintenance menu is used up to Step 20.) The fieldeng permission is required for executing this command. a. From the console that is connected to XSCF, log in to XSCF, and execute the replacefru command. As shown in FIGURE 6-4, the initial page of the maintenance menu is displayed.
6. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat.
FIGURE 6-5 Removing the CMU (Front of the M8000) 3 1 2 1 6-16 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 6-6 Removing the CMU (Rear of the M9000 Base Cabinet) 3 1 2 1 8. Place the removed CMU with the cover and its label facing upward on an antistatic mat. 9. Remove the cover of the CMU. a. Pull the two latches on the connector side of the CMU to unlock the latches. b. Slide the cover of the CMU toward the connector side to release it from the guide pins (two pins on each side). c. Raise the cover of the CMU to remove it. 10. Remove the DIMM duct cover.
11. Replace parts. In CMU replacement: Remove all the CPUs and DIMMs mounted in the removed CMU, and remount them in the replacement CMU. Caution – Be sure to remount the CPUs and DIMMs in the same locations in the replacement CMU as their mounting locations in the removed CMU. For CPU replacement: Replace each swap target CPU mounted in the removed CMU with a replacement CPU. For the replacement procedure, see Section 6.5, “CPU and DIMM Replacement” on page 6-29.
For DIMM replacement: Replace each swap target DIMM mounted in the removed CMU with a replacement DIMM. For the replacement procedure, see Section 6.5, “CPU and DIMM Replacement” on page 6-29. FIGURE 6-7 CPU and DIMM Locations Duct cover DIMM (x32) CPU (x4) 12. Place the DIMM duct cover back at its original location. Insert the L-shaped projections on the CMU frame into the two positioning slits of the DIMM duct cover. 13.
14. Place the replacement CMU on the antistatic mat, and touch the metallic chassis for five or more seconds with your bare hand wearing the antistatic wrist strap, to remove static electricity. Caution – To mount a new CMU, ensure that the board is placed on a grounded antistatic mat. Then, touch it with your bare hand while wearing a properly connected antistatic wrist strap. This will remove any static electricity before installing it in the server. Take care not to damage the connector on the CMU edge.
15. Touch each of the designated points on the guide blocks for 5 or more seconds with your bare hand wearing the antistatic wrist strap. FIGURE 6-9 Guide Block (CMU) Touch for 5 or more seconds. 16. Mount the replacement CMU by following the removal instructions in Step 7 in reverse order. Align the CMU with the slot guides, insert it carefully, and secure it firmly.
b. Exit the maintenance menu. 18. Configure the CMU into the system. ■ Active replacement: Use the addboard command to dynamically configure the relevant CMU into a domain. Perform this operation for all the domains that have been recognized in Step 2 as including the CMU to be replaced. For details, see Section 4.2.2, “Disconnecting and Replacing a FRU” on page 4-5, and the SPARC Enterprise M4000/M5000/M8000/M9000 Servers Dynamic Reconfiguration (DR) User’s Guide.
6.4 Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching CMU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2.
5. Switch off all main line switches for the AC section (ACS) in the power supply system. For details, see Section 4.5.3, “Main Line Switch Locations” on page 4-27. Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units.
e. While grasping the front and upper handles of the CMU, pull it out slowly from the slot. Note – To upgrade the CMU, remove the dummy (filler) unit before mounting the CMU to be added. Since the same levers are used on the dummy (filler) unit, please operate the eject/lock levers in the same manner. Caution – A CMU with mounted components in the maximum configuration weighs about 22 kg. 9. Place the removed CMU with the cover facing upward on an antistatic mat. 10. Remove the cover of the CMU. a.
13. Place the DIMM duct cover back at its original location. Insert the L-shaped projections on the CMU frame into the two positioning slits of the DIMM duct cover. 14. Follow the removal instructions in Step 10 in reverse order to attach the cover of the CMU, and press the latch pins to lock the latches. 15. Place the replacement CMU on the antistatic mat, and touch the metallic chassis for five or more seconds with your bare hand wearing the antistatic wrist strap, to remove static electricity.
16. Touch each of the designated points on the guide blocks for 5 or more seconds with your bare hand wearing the antistatic wrist strap. FIGURE 6-11 Guide Block (CMU) Touch for 5 or more seconds. 17. Mount the replacement CMU by following the removal instructions in Step 8 in reverse order. Align the CMU with the slot guides, insert it carefully, and secure it firmly.
20. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From the console that is connected to XSCF, execute the poweron -d command of XSCF for the domain of the swap target new CMU. Power to all the domains is turned on. Firmware version synchronization of the target domain is automatically performed. For details of the power-on operation, see Section 4.4.
6.5 CPU and DIMM Replacement When replacing a CPU module or DIMM, replace it with one of the same type. For details of the mounting locations indicated in messages, see FIGURE 6-12.
6.5.1 Replacing a CPU Module To replace the CPU module, use the CPU Module insertion/extraction tool. Side A is for insertion, and side B is for extraction. The CPU Module insertion/expansion tool is stored inside the server. In the SPARC Enterprise M8000 server, the tool can be found on the left-side wall when the front door of the server is open. In the SPARC Enterprise M9000 server base cabinet, the tool can be found on the upper part of the unit when the rear door is open.
1. Remove the CPU Module. a. Insert the tips on the extraction side of the CPU Module insertion/extraction tool (B) under the CPU Module guide pins. Insert the shorter guide pin into the groove of the guide lock of the CMU.
FIGURE 6-15 CMU Guide Lock (Enlarged) Caution – When using the CPU Module insertion/extraction tool, make sure that it is set in position and used in the manner shown in the figure. If the CPU Module insertion/extraction tool is set and used in a different direction from the figure, serious damage may result. b. Tilt the CPU Module insertion/extraction tool to the DIMM side, and pull up the CPU Module to disengage the CMU connector.
FIGURE 6-16 Lever Up the CPU Module Caution – When tilting the CPU Module insertion/extraction tool toward the DIMM, do not tilt it to such an extent that the CPU Module is raised beyond its limit height. Caution – Be sure not to remove the tool before removing the CPU Module. Otherwise, the CMU may be damaged. c. Slide the CPU Module upward at an inclined angle to remove it from the guide lock.
FIGURE 6-17 Remove the CPU Module d. Remove the CPU Module insertion/extraction tool.
2. Attach the replacement CPU Module. a. Slide and insert the guide pin on the DDC side of the CPU Module along the groove of the CMU guide lock.
FIGURE 6-19 6-36 CMU Guide Lock (Enlarged) SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
b. Fit the guide pin on the heat sink side of the CPU Module into the groove of the CMU guide lock. FIGURE 6-20 Set CPU Module on the Guide Lock (2) Note – Each CPU Module has two guide pins and two CMU guide locks on the DDC side. Perform the work in such a way that both guide pins and both guide locks are mounted correctly. c. Move the CPU Module back and forth to confirm that the pins are in the grooves. d.
FIGURE 6-21 6-38 Set CPU Module Insertion/Extraction Tool SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 6-22 CMU Guide Lock (Enlarged) Caution – When using the CPU Module insertion/extraction tool, make sure that it is set in position and used in the manner shown in the figure. If the CPU Module insertion/extraction tool is set and used in a different direction from the figure, serious damage may result. e. Raise the CPU Module insertion/extraction tool on the CPU Module side until the tool is standing upright, perpendicular to the ground.
FIGURE 6-23 Mounting the CPU Module Caution – Each CPU Module has two CMU guide locks. Perform the work in such a way that both guide locks are correctly set. Raise the insertion/extraction tool until it touches the stopper and is perpendicular to the ground. The connector is securely engaged when the insertion/extraction tool is raised vertically and it comes into contact with the stopper. Do not forcibly raise the insertion/extraction tool further. f.
The CMU has 16 DIMMs as standard. The number of DIMMs that can be mounted on the CMU is 16 or 32, in sets of 16 DIMMs. Therefore, 16 DIMMs can be added to the standard product. Caution – To replace or install a DIMM, do not fail to confirm the DIMM information and comply with the conditions to mount the memory. 6.5.2.1 Confirmation of DIMM Information Confirm the DIMM information (size/rank) in the following way. ■ Execute the showhardconf(8) command on XSCFU.
FIGURE 6-24 shows an example of how to read the DIMM information displayed in the Type field. FIGURE 6-24 Example of Reading DIMM Information MEM#00A Status:Normal; Memory slot No. + Code:2cffffffffffffff0836HTF25672Y-53EB1 0100-d409da25; + Type:2B; Size:2 GB; DIMM size 1: 1 GB 2: 2 GB 4: 4 GB ... DIMM rank A: 1 rank B: 2 rank 6.5.2.2 DIMM Mounting Conditions ■ ■ ■ ■ ■ Memory modules are added in sets of 16 DIMMs. The DIMMs mounted in the MEM#xxA slot are defined as Group A.
FIGURE 6-25 Memory Mounting Conditions DIMM MEM#30A CPUM#3 MEM#30B MEM#31A MEM#31B MEM#21B CPUM#2 MEM#21A MEM#20B MEM#20A MEM#10A CPUM#1 MEM#10B MEM#11A MEM#11B MEM#01B CPUM#0 MEM#01A MEM#00B MEM#00A Memory modules mounted as standard MEM#32A MEM#32B MEM#33A MEM#33B MEM#23B MEM#23A MEM#22B MEM#22A MEM#12A MEM#12B MEM#13A MEM#13B MEM#03B MEM#03A MEM#02B MEM#02A Added memory modules Conceptual diagram of CPU and memory module locations Chapter 6 Mounting locations of CPUs and memory modules Replacemen
6.5.3 DIMM Replacement 1. Open outward the tabs of the connector in which the DIMM to be replaced is mounted, and pull the DIMM out vertically. FIGURE 6-26 Removing the DIMM Note – While grasping both ends of the DIMM, pull it out vertically.
2. Perform the above operation in reverse order to mount the DIMM. Note – To mount DIMM, position the notch of DIMM to the corresponding part of the connector.
6-46 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 7 I/O Unit (IOU) Replacement This chapter explains the replacement procedures for the I/O unit (IOU). It covers the following: ■ ■ ■ Section 7.1, “Overview of the IOU” on page 7-2 Section 7.2, “Active Replacement and Hot Replacement” on page 7-8 Section 7.3, “Cold Replacement” on page 7-17 There are three methods for replacing an IOU: active replacement, hot replacement and cold replacement. For the definition of each type of method, see Section 4.
Caution – Adjust the positioning of the routes appropriately so that the cables do not cover a PCI cassette at the front of an IOU in the SPARC Enterprise M8000 server. 7.1 Overview of the IOU This section provides an overview of IOUs and describes their mounting locations. Each IOU consists of one I/O board (IOB), eight PCI slots, and up to four hard disk drives (HDD). A PCI card or link card is mounted in a PCI slot by using a PCI cassette.
FIGURE 7-1, FIGURE 7-2, FIGURE 7-3, FIGURE 7-4, and FIGURE 7-5 indicate the mounting locations of IOUs.
FIGURE 7-2 IOU mounting Locations (Front of the M9000 Base Cabinet) IOU#0 IOU#2 IOU#4 IOU#6 7-4 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 7-3 IOU Mounting Locations (Rear of the M9000 Base Cabinet) IOU#7 IOU#5 IOU#3 IOU#1 Chapter 7 I/O Unit (IOU) Replacement 7-5
FIGURE 7-4 IOU Mounting Locations (Front of M9000 with the Expansion Cabinet) IOU#0 IOU#2 IOU#4 IOU#6 IOU#8 IOU#10 IOU#11 IOU#12 7-6 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 7-5 IOU Mounting Locations (Rear of M9000 with the Expansion Cabinet) IOU#15 IOU#13 IOU#11 IOU#9 IOU#7 IOU#5 IOU#3 IOU#1 TABLE 7-1 lists the abbreviated names and component numbers of the IOUs.
7.2 Active Replacement and Hot Replacement The active replacement procedure and hot replacement procedure are almost the same. This section explains both procedures together. Each step that is different between these procedures is distinguished by "Active replacement" or "Hot replacement" as a header. The main difference between active replacement and hot replacement is as follows: ■ Active replacement: The CMU of an IOU is disconnected from or configured into a domain by dynamic reconfiguration (DR).
3. Prepare the IOU that is to be replaced. ■ Active replacement: From the terminal connected to the XSCF, execute the deleteboard command to dynamically disconnect the target CMU from all the domains that have been recognized in Step 2 as including the target CMU. (The IOU status can be checked by executing the showboards command.) To execute this command, you need the privilege of platform administrator (platadm) or domain administrator (domainadm). For details, see Section 4.2.
Note – DDC_A is displayed only for the M8000. b. From the replacement parts list in the maintenance menu, select 1, "CMU/IOU." c. From here, perform the operation according to the guidance displayed in the maintenance menu. Check the states of LEDs (POWER LED: off; CHECK LED: blinking) on the IOU to be replaced, and begin the actual replacement work when a message instructing the user to replace the target IOU is displayed. Leave the maintenance menu as is until Step 13. 5.
Caution – An IOU with all options mounted weighs about 21 kg.
FIGURE 7-8 Removing the IOU (Front of M9000) 3 1 2 1 7-12 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 7-9 Removing the IOU (Rear of M9000) 3 1 2 1 8. Place the removed IOU on an antistatic mat. 9. Dismount the HDD (or filler panel) and the PCI cards on the detached IOU, and remount them all on the replacing IOU. As for the PCI cassette, which serves as the housing for PCI card, use the one which is mounted on the replacing IOU. For the remounting procedures, see Chapter 12 and Chapter 13.
FIGURE 7-10 Locations of PCI Slots and HDDs HDD#0 HDD#1 HDD#2 HDD#3 PCI#4 PCI#5 PCI#6 PCI#7 PCI#0 PCI#1 PCI#2 PCI#3 10. Place the replacement IOU on the antistatic mat, and touch the metallic chassis for five or more seconds with your bare hand wearing the antistatic wrist strap, to remove static electricity. Caution – To mount a new CMU, ensure that the board is placed on a grounded antistatic mat. Then, touch it with your bare hand while wearing a properly connected antistatic wrist strap.
FIGURE 7-11 Metallic Chassis (IOU) Touch for 5 or more seconds. 11. Touch each of the designated points on the guide blocks for 5 or more seconds with your bare hand wearing the antistatic wrist strap. FIGURE 7-12 Guide Block (IOU) Touch 5 or more seconds.
12. Mount the replacement IOU by following the removal instructions in Step 7 in reverse order. Align the IOU with the slot guides, insert it carefully, and secure it firmly. Push the eject/lock levers as far as they will go, and then push the knob screws on the upper and lower eject/lock levers to lock them in position. When the knob screws remain in position, the IOU is securely mounted. Note – Mount the dummy (filler) unit in the same manner as IOU.
■ Hot replacement: From the console that is connected to XSCF, execute the poweron -d command of XSCF to turn on power to the domain. When the power to the domain is turned on, the power-on self test (POST) is performed. When ok prompt displayed on the console of relevant domain, POST is completed. 15. Confirm that the replacement components are normal by using the showhardconf or showstatus command of XSCF. Unless "*" is displayed for a replacement IOU, CPU, or DIMM, it is operating normally.
3. Stop (turn off power) all domains together. Take either of the following two actions: ■ Press and hold the POWER switch on the operator panel for at least four seconds. ■ From a console that is connected to XSCF, log in to XSCF, and execute the poweroff -a command. The OS shutdown processing is executed for all the domains, and then the power-off processing is executed. 4. Confirm that the power-off processing is completed by confirming that the POWER LED (green) on the operator panel is off. 5.
8. Remove the IOU to be replaced. (See FIGURE 7-7 - FIGURE 7-9) a. Using a Phillips screwdriver, turn the knob screws securing the upper and lower eject/lock levers 90 to 180 degrees to unlock the levers. (Turn the knob screw while lightly pulling the lever by one hand.) b. Unlock the eject/lock levers to pull the IOU off the backplane. c. While grasping the handle at the front of the IOU, pull the IOU out partially until the [NEAR TO END] label comes into view. d.
11. Place the replacement IOU on the antistatic mat, and touch the metallic chassis for five or more seconds with your bare hand wearing the antistatic wrist strap, to remove static electricity. Caution – To mount a new CMU, ensure that the board is placed on a grounded antistatic mat. Then, touch it with your bare hand while wearing a properly connected antistatic wrist strap. This will remove any static electricity before installing it in the server. Take care not to damage the connector on the FRU edge.
12. Touch each of the designated points on the guide blocks for 5 or more seconds with your bare hand wearing the antistatic wrist strap. FIGURE 7-14 Guide Block (IOU) Touch 5 or more seconds. 13. Mount the replacement IOU by following the removal instructions in Step 8 in reverse order. Align the IOU with the slot guides, insert it carefully, and secure it firmly.
14. Switch on all main line switches for the AC section (ACS) in the power supply system. 15. Confirm that the XSCF STANDBY LED (green) on the operator panel remains lit. If it is blinking, wait until it remains lit. 16. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel.
CHAPTER 8 FAN Unit Replacement This chapter explains the replacement procedures for the FAN unit (FANU). It covers the following: ■ ■ ■ Section 8.1, “Overview of the FAN Unit” on page 8-2 Section 8.2, “Active Replacement and Hot Replacement” on page 8-8 Section 8.3, “Cold Replacement” on page 8-14 There are three methods for replacing a FAN unit: active replacement , hot replacement, and cold replacement. For the definition of each type of method, see Section 4.
8.1 Overview of the FAN Unit This section provides an overview of FAN unit and describes their mounting locations. Each FAN unit consists of two or three cooling fans, and two or more FAN unit are used for cooling components in a cabinet. There are two types of FAN unit: FAN_A (3-FAN unit), which is used commonly with the high-end server, and FAN_B (2-FAN unit), which is used only with the M8000. FAN units have an N+1 redundancy configuration, and active maintenance can be performed on them.
FIGURE 8-2 FAN Mounting Locations (Rear of the M8000) FAN_B#2-#7 #7 #5 #3 #6 #4 #2 Chapter 8 FAN Unit Replacement 8-3
FIGURE 8-3 FAN Mounting Locations (Front of the M9000 Base Cabinet) FAN_A#0-#3 8-4 #1 #3 #0 #2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 8-4 FAN Mounting Locations (Rear of the M9000 Base Cabinet) FAN_A#4-#15 #15 #12 #9 #6 #14 #11 #8 #5 #13 #10 #7 #4 Chapter 8 FAN Unit Replacement 8-5
FIGURE 8-5 FAN Mounting Locations (Front of M9000 with the Expansion Cabinet) FAN_A#0-#3 #1 #3 #0 #2 FAN_A#20-#23 #21 #23 #20 #22 base cabinet Expansion cabinet 8-6 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 8-6 FAN Mounting Locations (Rear of M9000 with the Expansion Cabinet) FAN_A#4-#15 #15 #12 #9 #6 #14 #11 #8 #5 #13 #10 #7 #4 FAN_A#24-#35 #35 #32 #29 #26 #34 #31 #28 #25 #33 #30 #27 #24 base cabinet Expansion cabinet TABLE 8-1 lists the abbreviated names and component numbers of the FAN unit.
8.2 Active Replacement and Hot Replacement Caution – Each cabinet contains multiple FAN units so that the cooling specifications are satisfied even when one individual fan inside the FAN units fails. When replacing multiple FAN units, replace one at a time. The active replacement procedure and hot replacement procedure for the FAN units are the same. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5.
FIGURE 8-7 Initial Page of the Maintenance Menu --------------------------------------------------Maintenance/Replacement Menu Please select a type of FRU to be replaced. 1. CMU/IOU(CPU Memory Board Unit/IO Unit) 2. FAN(Fan Unit) 3. PSU(Power Supply Unit) 4. XSCFU(Extended System Control Facility Unit) 5. DDC_A(DDC for BP_A) --------------------------------------------------Select [1-5|c:cancel] : Note – DDC_A is displayed only for the M8000. b.
6. While grasping the handle of the FAN unit to be replaced, pull it out slowly from the slot.
FIGURE 8-9 Removing the FAN Unit (Rear of M8000) 1 2 Chapter 8 FAN Unit Replacement 8-11
FIGURE 8-10 Removing the FAN Unit (Front of M9000) 1 2 8-12 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 8-11 Removing the FAN Unit (Rear of M9000) 1 2 7. Place the removed FAN unit on an antistatic mat. 8. Mount the replacement FAN unit by following the removal instructions in Step 5 and Step 6 in reverse order. Align the FAN unit with the slot guides, insert it carefully, and secure it firmly. Caution – Each cabinet contains multiple FAN unit so that the cooling specifications are satisfied even if one FAN unit fails. When replacing a FAN unit, however, replace it as quickly as possible.
9. Go back to the terminal that is connected to XSCF, and perform the input operation according to the instructions of the maintenance menu that is displayed after the replacement. a. Perform operation according to instructions of messages. When the display of the maintenance menu returns to the initial page, the operation is completed. b. Exit the maintenance menu. 10. Confirm that the replacement components are normal by using the showhardconf or showstatus command of XSCF.
3. Stop (turn off power) all domains together. Take either of the following two actions: ■ Press and hold the POWER switch on the operator panel for at least four seconds. ■ From a console that is connected to XSCF, log in to XSCF, and execute the poweroff -a command. The OS shutdown processing is executed for all the domains, and then the power-off processing is executed. 4. Confirm that the power-off processing is completed by confirming that the POWER LED (green) on the operator panel is off. 5.
9. While grasping the handle of the FAN unit to be replaced, pull it out slowly from the slot. 10. Place the removed FAN unit on an antistatic mat. 11. Mount the replacement FAN unit by following the removal instructions in Step 8 and Step 9 in reverse order. Align the FAN unit with the slot guides, insert it carefully, and secure it firmly. Caution – Do not forcibly push the FAN unit when inserting it, even if it is not moving smoothly.
16. Restore the original mode switch setting on the operator panel. Note – When the maintenance work is completed, be sure to confirm the mode switch setting and return the switch to Locked. 17. Close the door on the server, and lock it.
8-18 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 9 Power Supply Unit (PSU) Replacement This chapter explains the replacement procedures for the power supply unit (PSU). It covers the following: ■ ■ ■ Section 9.1, “Overview of the PSU” on page 9-1 Section 9.2, “Active Replacement and Hot Replacement” on page 9-8 Section 9.3, “Cold Replacement” on page 9-12 There are three methods for replacing a PSU: active replacement , hot replacement, and cold replacement. For the definition of each type of method, see Section 4.
For details of the types and locations of power supply, see Section 4.5, “Power-On/Off of Main Line Switch” on page 4-24.
FIGURE 9-2 PSU Mounting Locations (Rear of M8000 + Power Cabinet) Power cabinet Chapter 9 Power Supply Unit (PSU) Replacement 9-3
FIGURE 9-3 PSU Mounting Locations (Front of the M9000 Base Cabinet + Power Cabinet) Power cabinet PSU#0-#15 Power supply unit PSU#20-#34 #2 #5 #8 #11 #14 #1 #4 #7 #10 #13 #0 #3 #6 #9 #34 #33 #32 #31 #30 #29 #28 #27 #26 #25 #24 #23 #22 #21 #20 9-4 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010 #12
FIGURE 9-4 PSU Mounting Locations (Rear of the M9000 Base Cabinet + Power Cabinet) Power cabinet Chapter 9 Power Supply Unit (PSU) Replacement 9-5
FIGURE 9-5 PSU Mounting Locations (Front of M9000 with an Expansion Cabinet + Power Cabinet) PSU#20-#34 #22 #25 #28 #31 #34 #21 #24 #27 #30 #33 #20 #23 #26 #29 #32 Power cabinet Power cabinet PSU#0-#14 #2 #5 #8 #11 #14 #1 #4 #7 #10 #13 #0 #3 #6 #9 #12 PSU#40-#54 PSU#60-#74 #54 #74 #53 #73 #52 #72 #51 #71 #50 #70 #49 #69 #48 #68 #47 #46 9-6 #67 Base cabinet Expansion cabinet #66 #45 #65 #44 #64 #43 #63 #42 #62 #41 #61 #40 #60 SPARC Enterprise M8000/M9000 Serve
FIGURE 9-6 PSU Mounting Locations (Rear of M9000 with an Expansion Cabinet + Power Cabinet) Power cabinet Power cabinet Base cabinet Expansion cabinet TABLE 9-1 lists the abbreviated names and component numbers of the PSUs.
9.2 Active Replacement and Hot Replacement Caution – Each cabinet contains multiple PSUs so that the power output specifications are satisfied even if one PSU fails. However, avoid continuing operation for a long period with a faulty PSU left in the cabinet. The active replacement procedure and hot replacement procedure for the PSUs are the same. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5.
FIGURE 9-7 Initial Page of the Maintenance Menu --------------------------------------------------Maintenance/Replacement Menu Please select a type of FRU to be replaced. 1. CMU/IOU(CPU Memory Board Unit/IO Unit) 2. FAN(Fan Unit) 3. PSU(Power Supply Unit) 4. XSCFU(Extended System Control Facility Unit) 5. DDC_A(DDC for BP_A) --------------------------------------------------Select [1-5|c:cancel] : Note – DDC_A is displayed only for the M8000. b.
4. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat.
FIGURE 9-9 Removing the PSU (Front of the M9000 Base Cabinet) 3 1 2 7. Place the removed PSU on an antistatic mat. 8. Mount the replacement PSU by following the removal instructions in Step 5 and Step 6 in reverse order. Align the PSU with the slot guides, insert it carefully, and secure it firmly. Caution – Each cabinet contains multiple PSUs so that the power output specifications are satisfied even if one PSU fails. However, please refrain from the long time operation as one PSU remains failed.
a. Perform operation according to instructions of messages. When the display of the maintenance menu returns to the initial page, the operation is completed. b. Exit the maintenance menu. 10. Confirm that the replacement components are normal by using the showhardconf or showstatus command of XSCF. Unless "*" is displayed for the replacement PSU, it is operating normally.
■ Press and hold the POWER switch on the operator panel for at least four seconds. ■ From a console that is connected to XSCF, log in to XSCF, and execute the poweroff -a command. The OS shutdown processing is executed for all the domains, and then the power-off processing is executed. 4. Confirm that the power-off processing is completed by confirming that the POWER LED (green) on the operator panel is off. 5. Switch off all main line switches for the AC section (ACS) in the power supply system.
7. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat.
15. Confirm that the replacement components are normal by using the showhardconf or showstatus command of XSCF. Unless "*" is displayed for the replacement PSU, it is operating normally. For details of the showhardconf command and showstatus command, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide, the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF Reference Manual, or the man page. Note – The showstatus command displays information on degraded components. 16.
9-16 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 10 Operator Panel Replacement This chapter explains the replacement procedures for the operator panel (OPNL). It covers the following: ■ ■ Section 10.1, “Overview of the Operator Panel” on page 10-1 Section 10.2, “Cold Replacement” on page 10-4 Cold replacement is the only swapping method that can be used for the OPNL. For the definition of each type of swapping method, see Section 4.1, “Types of Replacement Procedures” on page 4-2.
FIGURE 10-1 OPNL Mounting Location (Front of the M8000) OPNL 10-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 10-2 OPNL Mounting Location (Front of the M9000 Base Cabinet) OPNL Chapter 10 Operator Panel Replacement 10-3
FIGURE 10-3 OPNL Mounting Location (Front of M9000 with the Expansion Cabinet) OPNL Base cabinet Expansion cabinet TABLE 10-1 lists the abbreviated name of the OPNL. TABLE 10-1 10.2 Abbreviated Name of the Operator Panel Model name Abbreviated name M8000/M9000 Operator Panel OPNL Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server.
Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Change the mode switch setting on the OPNL from the Locked position to the Service position. Note – When the maintenance work is completed, be sure to confirm the mode switch setting and return the switch to Locked. 3. Stop (turn off power) all domains together.
7. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat.
11. Place the removed OPNL on an antistatic mat. 12. Mount the replacement OPNL by following the removal instructions in Step 8 to Step 10 in reverse order. Caution – Do not forcibly push the cable in when inserting it. If there is problem with the connector pin, serious damage may result. 13. Insert the OPNL key, and then set the key to the Service position. 14. Switch on all main line switches for the AC section (ACS) in the power supply system. 15.
18. Restore the original mode switch setting on the operator panel. Note – When the maintenance work is completed, be sure to confirm the mode switch setting and return the switch to Locked. 19. Close the door on the server, and lock it.
CHAPTER 11 XSCF Unit Replacement This chapter explains the replacement procedures for an eXtended System Control Facility unit (XSCFU). It covers the following: ■ ■ ■ Section 11.1, “Overview of the XSCFU” on page 11-1 Section 11.2, “Active Replacement and Hot Replacement” on page 11-5 Section 11.3, “Cold Replacement” on page 11-12 There are three methods for replacing an XSCFU: active replacement, hot replacement, and cold replacement. For the definition of each type of method, see Section 4.
FIGURE 11-1 XSCFU Mounting Locations (Front of the M8000) XSCFU_B#0 XSCFU_B#1 11-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 11-2 XSCFU Mounting Locations (Front of the M9000 Base Cabinet) XSCFU_B#1 XSCFU_B#0 Chapter 11 XSCF Unit Replacement 11-3
FIGURE 11-3 XSCFU Mounting Locations (Front of M9000 with the Expansion Cabinet) XSCFU_C#1 XSCFU_C#0 XSCFU_B#1 XSCFU_B#0 TABLE 11-1 lists the abbreviated names and component numbers of the XSCFUs.
11.2 Active Replacement and Hot Replacement Note – To replace the XSCF Unit, you must consider the version of the XCP that is currently in the server. To determine the firmware version of the server, type: XSCF> version -c xcp For more information, refer to the version (8) man page or the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF Reference Manual. If the version of XCP in your server is XCP 1041 or earlier, see Appendix E.
3. Use the maintenance menu to select the XSCFU to be replaced. (The maintenance menu is used up to Step 10.) The fieldeng permission is required for executing this command. a. From the console that is connected to the Active XSCF, log in to the XSCF, and execute the replacefru command. As shown in FIGURE 11-4, the initial page of the maintenance menu is displayed.
5. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat.
FIGURE 11-5 Removing the XSCFU (Front of the M8000) 1 2 1 11-8 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 11-6 Removing the XSCFU (Front of the M9000 Base Cabinet) 1 2 1 8. Place the removed XSCFU on an antistatic mat. 9. Mount the replacement XSCFU by following the removal instructions in Step 6 and Step 7 in reverse order. Align the XSCFU with the slot guides, insert it carefully, and secure it firmly. Push the levers as far as they will go, and then push the knob screws on the eject/lock levers to lock them in position. When the knob screws remain in position, the XSCF unit is securely mounted.
Note – When your system is the SPARC Enterprise M9000 server with Expansion Cabinet, pay attention to the XSCF message or the error log information which pointing out the suspect unit. * When it is "XSCFU_C," replace XSCFU_C. * When it is "XSCFU_B," you must always once extract both XSCFU_B and XSCFU_C. For insertion, insert XSCFU_C first, and then XSCFU_B. Caution – Do not forcibly push the XSCFU in when inserting it.
XSCF> unlockmaintenance For details of the unlockmaintenance command, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide, the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF Reference Manual, or the man page. ii. Execute the showhardconf command. And, confirm that the currently active XSCF unit is the one, which you are going to replace with using replacefru command. iii.
For the version matching of firmware, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. 13. Close the door on the server, and lock it. 11.3 Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.
Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units. Caution – After stopping all the domains, be sure to switch off the main line switches in the power supply system to completely shut off the supply of power. 6.
11. Mount the replacement XSCFU by following the removal instructions in Step 8 and Step 9 in reverse order. Align the XSCFU with the slot guides, insert it carefully, and secure it firmly. Push the levers as far as they will go, and then push the knob screws on the eject/lock levers to lock them in position. When the knob screws remain in position, the XSCF unit is securely mounted. Caution – Do not forcibly push the XSCFU in when inserting it.
16. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.
11-16 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 12 Hard Disk Drive (HDD) Replacement This chapter explains the replacement procedures for the hard disk drive (HDD). It covers the following: ■ Section 12.1, “Overview of the HDD” on page 12-1 ■ Section 12.2, “Active Replacement” on page 12-3 ■ Section 12.3, “Cold Replacement” on page 12-5 There are two methods for replacing a HDD: active replacement and cold replacement. For the definition of each type of method, see Section 4.1, “Types of Replacement Procedures” on page 4-2. 12.
FIGURE 12-1 HDD Mounting Locations HDD#0 HDD#1 HDD#2 HDD#3 TABLE 12-1 lists the abbreviated name and numbers of HDDs in an IOU.
12.2 Active Replacement The procedure for active replacement is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Prepare for the replacement of the target HDD. From the OS, execute the cfgadm -a command to verify Ap_Id.
6. Remove the HDD unit to be replaced. a. Press the square push button on the HDD unit to be replaced. The eject/lock lever protrudes at an angle of about 45 degrees. b. Pull the eject/lock lever to an angle of about 90 degrees to release the lock. c. While grasping the eject/lock lever, pull out the HDD unit slowly from the slot. FIGURE 12-2 Removing the HDD 2 1 7. Place the removed HDD on an antistatic mat.
8. Mount the replacement HDD by following the removal instructions in Step 6 in reverse order. Align the HDD with the slot guides, insert it carefully, and secure it firmly. Caution – Do not forcibly push the HDD when inserting it, even if it is not moving smoothly. If the HDD is forcibly inserted despite the presence of any problem with a connector pin, serious damage may result. Do not insert an HDD into a slot while the eject/lock lever is closed.
3. Stop (turn off power) all domains together. Take either of the following two actions: ■ Press and hold down the POWER switch on the operator panel for at least four seconds. ■ From a console that is connected to XSCF, log in to XSCF, and execute the poweroff -a command. The OS shutdown processing is executed for all the domains, and then the power-off processing is executed. 4. Confirm that shutdown processing is completed by confirming that the READY LED (green) on the HDD is off. 5.
8. Remove the HDD unit to be replaced. (See FIGURE 12-2) a. Press the square push button on the HDD unit to be replaced. The eject/lock lever protrudes at an angle of about 45 degrees. b. Pull the eject/lock lever to an angle of about 90 degrees to release the lock. c. While grasping the eject/lock lever, pull out the HDD unit slowly from the slot. 9. Place the removed HDD on an antistatic mat. 10. Mount the replacement HDD by following the removal instructions in Step 8 in reverse order.
16. Close the door on the server, and lock it.
CHAPTER 13 PCI Slot Device Replacement This chapter explains the replacement procedures for devices that are mounted in the PCI slots of an IOU. It covers the following: ■ Section 13.1, “Overview of PCI Slot Devices” on page 13-1 ■ Section 13.2, “Active Replacement” on page 13-3 ■ Section 13.3, “Hot Replacement” on page 13-17 ■ Section 13.4, “Cold Replacement” on page 13-20 There are three methods for replacing PCI slot devices: active replacement, hot replacement and cold replacement.
■ Link card If PCI cards mounted in all the IOUs are not sufficient in number, this card serves as a card for connecting an external I/O expansion unit that is used for mounting additional PCI cards. The PHP function enables active maintenance of link cards. A filler panel may be mounted in a PCICS. FIGURE 13-1 indicates the PCI slot locations in an IOU cabinet. For information on the IOU mounting locations, see FIGURE 7-1 to FIGURE 7-5 in Chapter 7.
TABLE 13-1 shows abbreviated name and numbers of PCI slots in an IOU. TABLE 13-1 Abbreviated Name and Numbers of PCI Slots Slot name Abbreviated name and numbers PCI slot PCI#0 to 7 For the following devices, replacement procedures have the same steps except that LEDs to be checked, eject levers, and names used for status display are different.
2. From the OS, execute the cfgadm -a command to verify the Ap_Id of the PCI card. Example: If the PCI card in PCI slot 3 of IOU#0 is replaced # cfgadm -a Ap_Id iou#0_pci#3 Type Receptacle Occupant Condition : pci_pci/hp connected configured ok The AP_Id (iou#0_pci#3) verified here is referred to as PCI_ApId, and the subsequent explanations of cfgadm operations for the PCI card use the term PCI_ApId. 3.
8. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat.
FIGURE 13-2 Removing the PCI Cassette 2 1 10. Remove the PCI card from the PCI cassette. (Perform this work on the conductive mat.) a. While grasping the lock bracket, turn it in the direction of the arrow and push the lever a little. Then, push in the lever completely. (See FIGURE 13-3) Once you start moving the lever, you need not continue holding the lock bracket. b. Loosen the two screws on the back of the PCI cassette, slide out the latch, and temporarily secure it with screws.
FIGURE 13-3 Pushing In the Lever Lever Lock bracket Chapter 13 PCI Slot Device Replacement 13-7
FIGURE 13-4 13-8 Loosening Screws of Latches SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 13-5 Removing the PCI Card Chapter 13 PCI Slot Device Replacement 13-9
11. Mount the replacement PCI card on the PCI cassette. (Perform this work on the conductive mat.) FIGURE 13-6 shows card positioning post and card positioning tab. FIGURE 13-6 Card positioning post and card positioning tab 1 2 TABLE 13-2 Card positioning post and card positioning tab Item Description 1 Card positioning post 2 Card positioning tab The post (item 1 in FIGURE 13-6) is seated in the notch of the bracket of the PCI card.
Note – When the PCI cassette is removed from the IOU, the tab raises the front of the card from the card connector. a. Move the PCI card in the direction of the arrow, and mount it on the PCI cassette by inserting the part projecting out at the bottom of the bracket of the PCI card into the hole at the bottom of the front panel of the PCI cassette and inserting the pin of the PCI cassette into the oval hole at the top of the bracket of the PCI card. FIGURE 13-7 Mounting the PCI Card b.
c. Align the PCI card with the center part of the lower latch of the cassette. Then, while pressing the PCI card against the center part of each of the upper latches of the cassette, tighten the fixing screws in the sequence shown in FIGURE 13-8 to fix the card in position.. Note – To ensure that the PCI card is properly mounted, you need to confirm that the PCI card base is properly centered at the latches and securely fastened in position.
Caution – If the card is short, move the screw at the bottom of the cassette as required. In such cases, place the latches as close as possible to the far end of the card. (FIGURE 13-9 reference) To shift the location of the bottom screw for the PCI card, first secure the card at the lower latch (part a in FIGURE 13-8). Then, tighten the fixing screws in the sequence of 1-2 shown in FIGURE 13-8 to fix the card in position. FIGURE 13-9 Latch Positions at the Bottom of the Cassette 12.
e. Using a Phillips screwdriver, tighten the green screw. Be sure to stop tightening the screw when the bottom of the screw head is flush with the projecting part of the handle. Otherwise, over tightening the screw may cause the projecting part of the handle to break. (See FIGURE 13-10.) FIGURE 13-10 Tightening the Screw Screw head Projecting part of the handle Caution – Align the PCI card with the slot guides, insert it carefully, and secure it firmly.
■ If the lever of a PCI cassette cannot be inserted: a. When the lever of the PCI cassette is pushed in but it stops halfway without reaching the far end, do not forcibly push it in any further. Instead, pull it out once and then insert it again.
b. If the lever stops halfway even when the PCI cassette is inserted again, pull the PCI cassette out from the IOU, and then check whether the PCI card is set in the normal position (see a, b, and c of Step 11). Then, reinsert the PCI cassette. Caution – Do not forcibly push the lever of the PCI cassette when the lever is pushed but stops halfway. In this case, restore the lever to its original position, and then reinsert it.
13.3 Hot Replacement The procedure for hot replacement is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Turn off power to the relevant domain. From a console that is connected to XSCF, log in to XSCF, and execute the poweroff -d command.
6. Remove the PCI cassette of the PCI card to be replaced. (See FIGURE 13-2) Caution – To disconnect the link cable, which is a cable connected between the link card mounted on the server and external I/O expansion unit, pull the latch lock of the connector. If you forcibly pull the connector or the cable, the cable may be damaged. Caution – When removing cables such as LAN cable, if your finger can't reach the latch lock of the connector, press the latch with a flathead screwdriver to remove the cable.
c. Set the screws in the center part of each of the two latches, and then tighten the screws while pressing them against the PCI card. Note – When affixing the PCI card after aligning its position with the card-positioning tab, do so while pulling the lever about 2 cm from the frame. Caution – If the card is short, move the screw at the bottom of the cassette as required. In such cases, place the latches as close as possible to the far end of the card. (See FIGURE 13-9) 9. Mount the PCI cassette. a.
b. If the lever stops halfway even when the PCI cassette is inserted again, pull the PCI cassette out from the IOU, and then check whether the PCI card is set in the normal position (see a, b and c of Step 8). Then, reinsert the PCI cassette. Caution – Do not forcibly push the lever of the PCI cassette when the lever is pushed but stops halfway. In this case, restore the lever to its original position, and then reinsert it.
2. Change the mode switch setting on the OPNL from the Locked position to the Service position. Note – When the maintenance work is completed, be sure to confirm the mode switch setting and return the switch to Locked. 3. Stop (turn off power) all domains together. Take either of the following two actions: ■ Press and hold the POWER switch on the operator panel for at least four seconds. ■ From a console that is connected to XSCF, log in to XSCF, and execute the poweroff -a command.
7. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat.
10. Mount the replacement PCI card on the PCI cassette. (Perform the work on the conductive mat.) a. Move the PCI card in the direction of the arrow, and mount it on the PCI cassette by inserting the part projecting out at the bottom of the bracket of the PCI card into the hole at the bottom of the front panel of the PCI cassette and inserting the pin of the PCI cassette into the oval hole at the top of the bracket of the PCI card. b.
14. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.
CHAPTER 14 CD-RW/DVD-RW Drive Unit Replacement This chapter explains the replacement procedures for a CD-RW/DVD-RW drive unit (DVDU). It covers the following: ■ ■ ■ ■ Section 14.1, Section 14.2, Section 14.3, Section 14.4, “Overview of a CD-RW/DVD-RW Drive Unit” on page 14-1 “Active Replacement” on page 14-5 “Hot Replacement” on page 14-9 “Cold Replacement” on page 14-11 There are three methods for replacing a CD-RW/DVD-RW drive unit: active replacement, hot replacement, and cold replacement.
FIGURE 14-1 CD-RW/DVD-RW Drive Unit Mounting Location (Front of the M8000) DVDU 14-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 14-2 CD-RW/DVD-RW Drive Unit Mounting Location (Front of the M9000 Base Cabinet) DVDU Chapter 14 CD-RW/DVD-RW Drive Unit Replacement 14-3
FIGURE 14-3 CD-RW/DVD-RW Drive Unit Mounting Location (Front of M9000 with the Expansion Cabinet) DVDU DVDU Base cabinet Expansion cabinet TABLE 14-1 shows the abbreviated name of the CD-RW/DVD-RW drive unit.
14.2 Active Replacement The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Stop all connection to the target CD-RW/DVD-RW drive unit. 3. From the OS, execute the cfgadm -a command to verify the Ap_Id of CD-RW/DVD-RW drive unit.
7. From the OS, execute the cfgadm -c unconfigure TAPE_Ap_Id command to unconfigure the tape drive unit from the domain. # cfgadm -c unconfigure c0::rmt/0 8. From the OS, execute the cfgadm -a command to verify that the status of CD-RW/DVD-RW drive unit and tape drive unit has changed from "configured" to "unconfigured." # cfgadm -a Ap_Id c0::dsk/c0t4d0 c0::rmt/0 Type Receptacle Occupant Condition : CD-ROM connected unconfigured unknown tape connected unconfigured unknown : 9.
13. Remove the CD-RW/DVD-RW drive unit to be replaced. a. Using a Phillips screwdriver, loosen the two knob screws securing the CD-RW/DVD-RW drive unit to be replaced. b. While grasping the knob screws, pull out the CD-RW/DVD-RW drive unit slowly from the slot.
FIGURE 14-5 Removing the CD-RW/DVD-RW Drive Unit (Front of the M9000) 1 2 14. Place the removed CD-RW/DVD-RW drive unit on an antistatic mat. 15. Mount the replacement CD-RW/DVD-RW drive unit by following the removal instructions in Step 13 in reverse order. Align the CD-RW/DVD-RW drive unit with the slot guides, insert it carefully, and secure it firmly. 16. Attach the CD-RW/DVD-RW drive unit to the port. Execute the cfgdevice -c attach -p port no. command of XSCF. 17.
19. From the OS, execute the cfgadm -c configure TAPE_Ap_Id command to configure the tape drive unit to the domain. # cfgadm -c configure c0::rmt/0 20. From the OS, execute the cfgadm -a command to verify that the status of CD-RW/DVD-RW drive unit and the tape drive unit has changed from "unconfigured" to "configured." # cfgadm -a Ap_Id Type Receptacle Occupant : c0::dsk/c0t4d0 CD-ROM connected configured c0::rmt/0 tape connected configured : Condition unknown unknown 21.
5. Confirm that the target CD-RW/DVD-RW drive unit is unconfigured from the domain. From XSCF, execute cfgdevice -l to confirm that the relevant port is not displayed in the current connection. 6. Connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port. See Section 1.4, “Antistatic Precautions” on page 1-6. 7. Ensure that the metallic underside of the wrist strap is in direct contact with your skin.
13. Turn on the power to the relevant domain. From a console that is connected to XSCF, execute the poweron -d command of XSCF. For details of the power-on operation, see Section 4.4.2, “Powering the Server On” on page 4-19, and the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. 14. Start up domains as necessary. 15. Close the door on the server and lock it. 14.4 Cold Replacement The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5.
5. Switch off all main line switches for the AC section (ACS) in the power supply system. For details, see Section 4.5.3, “Main Line Switch Locations” on page 4-27. Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units.
10. Mount the replacement CD-RW/DVD-RW drive unit by following the removal instructions in Step 8 in reverse order. Caution – Do not forcibly push the CD-RW/DVD-RW drive unit in when inserting it. If the CD-RW/DVD-RW drive unit is forcibly inserted despite the presence of any obstruction in a slot or any problems with the connector pin, serious damage may result. 11. Switch on all main line switches for the AC section (ACS) in the power supply system. 12.
14-14 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 15 Tape Drive Unit Replacement This chapter explains the replacement procedure for the tape drive unit (TAPEU). It covers the following: ■ ■ ■ ■ Section 15.1, Section 15.2, Section 15.3, Section 15.4, “Overview of the Tape Drive Unit” on page 15-1 “Active Replacement” on page 15-5 “Hot Replacement” on page 15-9 “Cold Replacement” on page 15-11 There are three methods for replacing a tape drive unit: active replacement, hot replacement, and cold replacement.
FIGURE 15-1 Tape Drive Unit Mounting Locations (Front of the M8000) TAPEU 15-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 15-2 Tape Drive Unit Mounting Locations (Front of the M9000 Base Cabinet) TAPEU Chapter 15 Tape Drive Unit Replacement 15-3
FIGURE 15-3 Tape Drive Unit Mounting Locations (Front of M9000 with the Expansion Cabinet) TAPEU TAPEU Base cabinet Expansion cabinet TABLE 15-1 shows the abbreviated name of the tape drive unit.
15.2 Active Replacement The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Stop all connection to the target tape drive unit. 3. From the OS, execute the cfgadm -a command to verify the Ap_Id of tape drive unit.
7. From the OS, execute the cfgadm -c unconfigure DVD_Ap_Id command to unconfigure the CD-RW/DVD-RW drive unit from the domain. # cfgadm -c unconfigure c0::dsk/c0t4d0 8. From the OS, execute the cfgadm -a command to verify that the status of tape drive unit and CD-RW/DVD-RW drive unit has changed from "configured" to "unconfigured." # cfgadm -a Ap_Id c0::dsk/c0t4d0 c0::rmt/0 Type Receptacle Occupant : CD-ROM connected unconfigured tape connected unconfigured : Condition unknown unknown 9.
13. Remove the tape drive unit to be replaced. a. Using a Phillips screwdriver, loosen the two knob screws securing the tape drive unit to be replaced. b. While grasping the knob screws, pull out the tape drive unit slowly from the slot.
FIGURE 15-5 Removing the Tape Drive Unit (Front of the M9000) 1 2 14. Place the removed tape drive unit on an antistatic mat. 15. Mount the replacement tape drive unit by following the removal instructions in Step 13 in reverse order. Align the tape drive unit with the slot guides, insert it carefully, and secure it firmly. 16. Attach the tape drive unit to the port. Execute the cfgdevice -c attach -p port no. command of XSCF. 17. Confirm that the replaced tape drive unit is configured to the domain.
20. From the OS, execute the cfgadm -a command to verify that the status of tape drive unit and the CD-RW/DVD-RW drive unit has changed from "unconfigured" to "configured." # cfgadm -a Ap_Id Type Receptacle Occupant Condition : c0::dsk/c0t4d0 CD-ROM connected unconfigured unknown c0::rmt/0 tape connected unconfigured unknown : 21. Close the door on the server, and lock it. 15.3 Hot Replacement The procedure is as follows. 1. Unlock and open the appropriate door on the server. For details, see Chapter 5.
7. Connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port. See Section 1.4, “Antistatic Precautions” on page 1-6. 8. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6.
15.4 Cold Replacement The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Change the mode switch setting on the operator panel from Locked to Service. 3. Stop (turn off power) all domains together.
5. Switch off all main line switches for the AC section (ACS) in the power supply system. For details, see Section 4.5.3, “Main Line Switch Locations” on page 4-27. Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units.
10. Mount the replacement tape drive unit by following the removal instructions in Step 8 in reverse order. Caution – Do not forcibly push the tape drive unit in when inserting it. If the tape drive unit is forcibly inserted despite the presence of any obstruction in a slot or any problems with the connector pin, serious damage may result. 11. Switch on all main line switches for the AC section (ACS) in the power supply system. 12.
15-14 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 16 Clock Control Unit Replacement This chapter explains the replacement procedure for the clock control unit (CLKU). It covers the following: ■ ■ Section 16.1, “Overview of the CLKU” on page 16-1 Section 16.2, “Cold Replacement” on page 16-4 Cold replacement is the only swapping method that can be used for the CLKU. For the definition of the swapping method, see Section 4.1, “Types of Replacement Procedures” on page 4-2. 16.
FIGURE 16-1 CLKU Mounting Locations (Front of the M9000 Base Cabinet) CLKU_B#0 CLKU_B#1 16-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 16-2 CLKU Mounting Locations (Front of M9000 with the Expansion Cabinet) CLKU_B#0 CLKU_B#1 Base cabinet CLKU_B#2 CLKU_B#3 Expansion cabinet TABLE 16-1 lists the abbreviated names and component numbers of CLKUs.
16.2 Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2.
5. Switch off all main line switches for the AC section (ACS) in the power supply system. For details, see Section 4.5.3, “Main Line Switch Locations” on page 4-27. Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units.
8. Remove the CLKU to be replaced. (See FIGURE 16-3) a. If the M9000 expansion cabinet is installed, slide the cable support bracket at the front of the clock control unit up or down by one hole, and then use a flathead screwdriver to remove the cable connected to the clock control unit. (four cables from the base unit and two from the expansion unit). b. Using a Phillips screwdriver, turn the knob screws securing the left and right eject/lock levers 90 to 180 degrees to unlock the levers.
10. Mount the replacement CLKU by following the removal instructions in Step 8 in reverse order. Align the CLKU with the slot guides, insert it carefully, and secure it firmly. Push the eject/lock levers as far as they will go, and then push the knob screws on the right and left eject/lock levers to lock them in position. If the knob screws remain in position, the clock control unit is securely mounted. Note – Use a torque screwdriver to secure the clock cable connector with the torque of 0.2N*m (2.
14. Confirm that the replacement components are normal by using the showhardconf or showstatus command of XSCF. Unless "*" is displayed for the replacement CLKU, it is operating normally. For details of the showhardconf command and showstatus command, see the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide, the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF Reference Manual, or the man page. Note – The showstatus command displays information on degraded components. 15.
CHAPTER 17 Crossbar Unit Replacement This chapter explains the replacement procedure for crossbar units (XBUs). It covers the following: ■ ■ Section 17.1, “Overview of XBUs” on page 17-1 Section 17.2, “Cold Replacement” on page 17-4 As a replacement method for XBUs, only the cold replacement method can be used. For the definition of the replacement method, see Section 4.1, “Types of Replacement Procedures” on page 4-2. 17.
FIGURE 17-1 XBU Mounting Locations (Front of the M9000 Base Cabinet) XBU_B#0-#7 #7 #6 #5 #4 #3 #2 #1 #0 17-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 17-2 XBU Mounting Locations (Front of M9000 with the Expansion Cabinet) XBU_B#0-#7 XBU_B#8-#15 #7 #15 #6 #14 #5 #13 #4 #12 #3 #11 #2 #10 #1 #9 #0 #8 Base cabinet Expansion cabinet TABLE 17-1 lists the abbreviated names and numbers of the XBU.
17.2 Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2.
Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units. Caution – After stopping all the domains, be sure to switch off the main line switches in the power supply system to completely shut off the supply of power. 6. Connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port. See Section 1.4, “Antistatic Precautions” on page 1-6. 7.
FIGURE 17-3 Removing the XBU (Front of the M9000 Base Cabinet) 2 1 1 9. Place the removed crossbar unit on an antistatic mat, and then install the replacement crossbar unit. 10. While grasping the eject/lock levers, properly align the crossbar unit with the slot guide, and then gently insert it. Caution – If you are unable to obtain a torque screwdriver, finger tighten the clock cable connectors. Do not secure them with a regular screwdriver.
12. Push the eject/lock levers as far as they will go, and then push the knob screws on the right and left eject/lock levers to lock them in position. If the knob screws remain in position, the crossbar unit is securely mounted. 13. If a SPARC Enterprise M9000 server expansion cabinet is installed, connect cables for connecting the base cabinet to the expansion cabinet crossbar unit. a.
15. Switch on all main line switches for the AC section (ACS) in the power supply system. 16. Confirm that the XSCF STANDBY LED (green) on the operator panel remains lit. If it is blinking, wait until it remains lit. 17. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF.
CHAPTER 18 AC Section Replacement This chapter explains the replacement procedures for AC sections (ACSs). It covers the following: ■ ■ Section 18.1, “Overview of ACSs” on page 18-1 Section 18.2, “Cold Replacement” on page 18-5 As a replacement method for ACSs, only the cold replacement method can be used for the sensor units. For the definition of the replacement method, see Section 4.1, “Types of Replacement Procedures” on page 4-2.
An ACS distributes AC input power from external sources to power supply units (PSUs). If the dual power feed option is not installed, a main line switch is mounted in an ACS. An ACS can be duplicated, but active replacement cannot be performed even in such a case. FIGURE 18-1, FIGURE 18-2, and FIGURE 18-3 show the mounting locations of CMUs.
FIGURE 18-2 ACS Mounting Location (Front of the M9000 Base Cabinet) ACS_B Chapter 18 AC Section Replacement 18-3
FIGURE 18-3 ACS Mounting Locations (Front of M9000 with the Expansion Cabinet) ACS_B#0 ACS_B#1 Base cabinet Expansion cabinet TABLE 18-1 lists the abbreviated names and component numbers of the ACSs.
18.2 Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2.
5. Switch off all main line switches for the AC section (ACS) in the power supply system. For details, see Section 4.5.3, “Main Line Switch Locations” on page 4-27. Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units.
a. Loosen the cable support plate retaining screw, and remove the plate by sliding it to the left. b. Using a Phillips screwdriver, loosen the two thumb screws to remove the connector protecting plate. c. Using a Phillips screwdriver, loosen the connector retaining screws to disconnect five power cords. d. Move the eject lever up to the open position. e. Using a Phillips screwdriver, loosen the two thumb screws in the upper part of the ACS. Pull up the two levers to the open position. f.
FIGURE 18-4 1 Removing the AC Section (Rear of the M8000) 1 2 18-8 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 18-5 1 Removing the AC Section (Front of the M9000) 1 2 Chapter 18 AC Section Replacement 18-9
FIGURE 18-6 Removing the Three-phase AC Section (Power Cabinet) 9. Place the removed AC section on an antistatic mat. 10. Mount the replacement AC section by following the removal instructions in Step 8 in reverse order. Align the AC section with the slot guides, insert it carefully, press the lever to the closed position, and secure it firmly. Caution – Do not forcibly push the ACS when inserting it, even if it is not moving smoothly.
13. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.
18-12 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 19 DDC Replacement This chapter explains the replacement procedures for DDCs (DC-to-DC Converters). It covers the following: ■ ■ ■ Section 19.1, “Overview of the DDC” on page 19-1 Section 19.2, “Active Replacement and Hot Replacement” on page 19-3 Section 19.3, “Cold Replacement” on page 19-7 There are three methods for replacing DDCs: active replacement, hot replacement and cold replacement. For the definition of the replacement method, see Section 4.
FIGURE 19-1 Mounting Locations of the DDCs (Front of the M8000) DDC_A#0 DDC_A#1 TABLE 19-1 lists the abbreviated names and numbers of the DDCs.
19.2 Active Replacement and Hot Replacement The active replacement procedure and hot replacement procedure for the DDC are the same. The procedure is as follows. 1. Unlock and open the appropriate door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2.
c. From here, perform operations according to the guidance displayed in the maintenance menu. Check the states of LEDs (POWER LED: off; CHECK LED: blinking) on the DDC to be replaced, and begin the actual replacement work when a message instructing the user to replace the target DDC is displayed. Leave the maintenance menu as is until Step 7. 3. Connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port. See Section 1.4, “Antistatic Precautions” on page 1-6. 4.
5. Remove the DDC to be replaced. a. Using a Phillips screwdriver, turn the knob screw securing the bottom eject/lock lever 90 to 180 degrees to unlock the lever. (Turn the knob screw while lightly pulling the lever with one hand.) b. Unlock the eject/lock levers to pull the DDC off the backplane. c. While grasping the eject/lock levers, pull out the DDC slowly from the slot. FIGURE 19-3 Removing the DDC (Front of the M8000) 1 2 1 6. Place the removed DDC on an antistatic mat.
7. Mount the replacement DDC by following the removal instructions in Step 5 and Step 6 in reverse order. Align the DDC with the slot guides, insert it carefully, and secure it firmly. Push the bottom eject/lock lever as far as it will go, and then push the knob screw on the lever to lock it in position. If the knob screws remain in position, the DDC is securely mounted. Caution – Do not forcibly push the DDC when inserting it, even if it is not moving smoothly.
19.3 Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2.
5. Switch off all main line switches for the AC section (ACS) in the power supply system. For details, see Section 4.5.3, “Main Line Switch Locations” on page 4-27. Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units.
Caution – Do not forcibly push the DDC when inserting it, even if it is not moving smoothly. If the DDC is forcibly inserted despite the presence of any obstruction in a slot or any problem with a connector pin, serious damage may result. 11. Switch on all main line switches for the AC section (ACS) in the power supply system. 12. Confirm that the XSCF STANDBY LED (green) on the operator panel remains lit. If it is blinking, wait until it remains lit. 13. Start (turn on power) all domains together.
15. Restore the original mode switch setting on the operator panel. Note – When the maintenance work is completed, be sure to confirm the mode switch setting and return the switch to Locked. 16. Close the door on the server, and lock it.
CHAPTER 20 Backplane Replacement This chapter explains the replacement procedure for backplanes (BPs). It covers the following: ■ ■ Section 20.1, “Overview of the BP” on page 20-1 Section 20.2, “Cold Replacement” on page 20-1 As a replacement method for BPs, only the cold replacement method can be used. For the definition of the replacement method, see Section 4.1, “Types of Replacement Procedures” on page 4-2. 20.
In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. 20.2.1 M8000/M9000 BPs FIGURE 20-1, FIGURE 20-2, and FIGURE 20-3 show the mounting locations of BPs of M8000, BP for M9000 (base cabinet), and M9000 (with the expansion cabinet), respectively.
FIGURE 20-2 BP Mounting Locations (Front of the M9000 Base Cabinet) BP_B Chapter 20 Backplane Replacement 20-3
FIGURE 20-3 BP Mounting Locations (Front of M9000 with the Expansion Cabinet) BP_B#0 BP_B#1 Base cabinet Expansion cabinet TABLE 20-1 lists the abbreviated names and numbers of the BPs.
The cold replacement procedure is as follows: 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Change the mode switch setting on the operator panel from Locked to Service.
5. Switch off all main line switches for the AC section (ACS) in the power supply system. The XSCF STANDBY LED (green) on the operator panel goes out. For details, see Section 4.5.3, “Main Line Switch Locations” on page 4-27. Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units.
Perform the following operations from the rear of the unit (See FIGURE 20-4): a. Remove the grounding wire on one side at the top of the rear door so that the door can be opened at least 180 degrees. b. Remove all FAN units on the I/O unit. (See Chapter 8) Mount the fixing brackets of the FAN unit at their original locations, and secure them temporarily. c. Using a Phillips screwdriver, loosen the two thumb screws securing the duct on the fan shelf. While grasping the handle, remove the duct. d.
■ M9000 (See FIGURE 20-7): Perform the following operations from the rear of the unit: a. Pull all mounted CMUs off the BP. (See Chapter 6) b. Pull all mounted I/O units (IOUs) off the BP. (See Chapter 7) Perform the following operations from the front of the unit (See FIGURE 20-5): a. Remove all mounted crossbar units (XBUs). For a model with the expansion cabinet, remove all cables and cable support brackets as well. Disconnect the cables completely from both cabinets. (See Chapter 17) b.
FIGURE 20-4 Preparation for Accessing BP_A (Rear of the M8000) Chapter 20 Backplane Replacement 20-9
FIGURE 20-5 Preparation for Accessing BP_B (Front of M9000) 9. Remove the BP to be replaced. ■ M8000 (See FIGURE 20-6): a. Using a torque wrench (box driver size: M6), remove the four bolts that hold the part shaped like an inverted L at the top edge of the bus bar of the BP. b. While grasping the lock of the connector, remove every cable connected to the BP.
c. Arrange the cables so that the connectors do not get caught in anything while the BP is being removed. d. Using a Phillips screwdriver, loosen the eight thumb screws securing the BP. The BP does not fall because it is suspended from the two guide rails inside the upper part of the cabinet. e. While grasping the two handles of the BP, remove the BP by moving it carefully along the guide rails. ■ M9000 (See FIGURE 20-7): a. Using a torque wrench, remove the four hexagonal bolts on the bus bar of the BP.
FIGURE 20-6 20-12 Removing the BP (Rear of the M8000) SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 20-7 Removing the BP (Front of the M9000) 10. Place the removed BP on an antistatic mat. 11. Mount the replacement BP by following the removal instructions in Step 8 and Step 9 in reverse order. Caution – For tightening the bus bar, choose a torque depending on the bolt size. * For M8 bolts, use a torque of 8.24 N*m (84 kgf*cm). * For M6 bolts, use a torque of 3.73 N*m (38 kgf*cm). Caution – Remove the cover which is attached for shipment.
12. Switch on all main line switches for the AC section (ACS) in the power supply system. 13. Confirm that the XSCF STANDBY LED (green) on the operator panel remains lit. If it is blinking, wait until it remains lit. 14. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF.
20.2.2 PSU BP FIGURE 20-8, FIGURE 20-9, and FIGURE 20-10 show the mounting locations of PSUBPs of M8000, M9000 (base cabinet), and M9000 (with the expansion cabinet), respectively.
FIGURE 20-9 PSUBP Mounting Locations (Front of the M9000 Base Cabinet) PSUBP_A#0 PSUBP_A#1 PSUBP_A#2 20-16 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 20-10 PSUBP Mounting Locations (Front of M9000 with the Expansion Cabinet) PSUBP_A#3 PSUBP_A#0 PSUBP_A#4 PSUBP_A#1 PSUBP_A#5 PSUBP_A#2 Base cabinet Expansion cabinet TABLE 20-2 lists the abbreviated names and component numbers of the PSUBPs.
The cold replacement procedure is as follows: 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Change the mode switch setting on the operator panel from Locked to Service.
6. Connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port. See Section 1.4, “Antistatic Precautions” on page 1-6. 7. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6.
FIGURE 20-11 Removing the PSUBP (Front of the M8000) PSUBP_B PSUBP_A 20-20 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 20-12 Removing the PSUBP (Front of the M9000) 10. Place the removed BP on an antistatic mat. 11. Mount the replacement BP by following the removal instructions in Step 8 and Step 9 in reverse order. Align the holes by the side of the thumb screws of the PSUBP with the two guide pins in the cabinet. Note – To mount a FAN BP on a fan shelf that has a guide pin, align the FAN BP with the guide pin during mounting. Caution – Be sure not to hit connector parts of the BP against anything.
14. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.
20.2.3 FAN BP FIGURE 20-13, FIGURE 20-14, and FIGURE 20-15 show the mounting locations of FAN BPs of M8000, M9000 (base cabinet), and M9000 (with the expansion cabinet), respectively.
FIGURE 20-14 FANBP Mounting Locations (Rear of the M9000 Base Cabinet) FANBP_B FANBP_A 20-24 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 20-15 FANBP Mounting Locations (Rear of M9000 with the Expansion Cabinet) FANBP_A#1 FANBP_B#1 FANBP_A#0 FANBP_B#0 Expansion cabinet Base cabinet TABLE 20-3 lists the abbreviated names and component numbers of the FANBPs.
The cold replacement procedure is as follows: 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Change the mode switch setting on the operator panel from Locked to Service.
6. Connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port. See Section 1.4, “Antistatic Precautions” on page 1-6. 7. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6.
c. Remove all six FAN units on the left side. (See Chapter 8) After the FAN units are removed, mount the fan fixing brackets at their original locations, and secure them temporarily. d. Using a Phillips screwdriver, loosen the two thumb screws in the upper part of the fan shelf. e. While grasping the handle at the upper part of the fan shelf, pull out the shelf carefully. For FANBP_B, perform the following operations from the rear of the unit: a. Remove all six FAN units on the right side.
FIGURE 20-16 Removing the FANBP (Rear of the M8000) Chapter 20 Backplane Replacement 20-29
FIGURE 20-17 20-30 Removing the FANBP (Front of M8000) SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 20-18 Removing the FANBP (Rear of M9000) 10. Place the removed BP on an antistatic mat. 11. Mount the replacement BP by following the removal instructions in Step 8 and Step 9 in reverse order. Note – To mount a FAN BP on a fan shelf that has a guide pin, align the FAN BP with the guide pin during mounting. Caution – Be sure not to hit connector parts of the BP against anything. Otherwise, the connection pins may be seriously damaged. 12.
14. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.
CHAPTER 21 Sensor Unit Replacement This chapter explains the replacement procedure for the sensor unit (SNSU). It covers the following: ■ ■ Section 21.1, “Overview of the SNSU” on page 21-1 Section 21.2, “Cold Replacement” on page 21-4 As a replacement method for the SNSU, only the cold replacement method can be used. For the definition of the replacement method, see Section 4.1, “Types of Replacement Procedures” on page 4-2. 21.
FIGURE 21-1 Mounting Location of the SNSU (Front of the M8000) SNSU SNSU 21-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 21-2 SNSU Mounting Location (Front of the M9000 Base Cabinet) SNSU Chapter 21 Sensor Unit Replacement 21-3
FIGURE 21-3 SNSU Mounting Location (Front of M9000 with the Expansion Cabinet) SNSU Base cabinet Expansion cabinet TABLE 21-1 lists the abbreviated names of the SNSUs. TABLE 21-1 21.2 Abbreviated Names of the SNSU Part name Abbreviated name Sensor unit SNSU Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF.
The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Change the mode switch setting on the operator panel from Locked to Service.
6. Connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port. See Section 1.4, “Antistatic Precautions” on page 1-6. 7. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6.
8. Remove the SNSU to be replaced. (See FIGURE 21-4 or FIGURE 21-5) a. Using a Phillips screwdriver, loosen the two thumb screws in the SNSU. b. While grasping the thumb screws, pull out the SNSU carefully from the slot.
FIGURE 21-5 Removing the SNSU (Front of the M9000) 2 1 9. Place the removed SNSU on an antistatic mat. 10. Mount the replacement SNSU by following the removal instructions in Step 8 in reverse order. Caution – Do not forcibly push the SNSU in when inserting it. If the SNSU is forcibly inserted despite the presence of any obstruction in a slot or any problems with the connector pin, serious damage may result. 11. Switch on all main line switches for the AC section (ACS) in the power supply system. 12.
13. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.
21-10 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 22 Media Backplane Replacement This chapter explains the replacement procedure for media backplanes (MEDBPs). It covers the following: ■ ■ Section 22.1, “Overview of the MEDBP” on page 22-1 Section 22.2, “Cold Replacement” on page 22-5 As a replacement method for MEDBPs, only the cold replacement method can be used. For the definition of the replacement method, see Section 4.1, “Types of Replacement Procedures” on page 4-2. 22.
FIGURE 22-1 MEDBP Mounting Location (Rear of the M8000) MEDBP 22-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 22-2 MEDBP Mounting Location (Front of the M9000 Base Cabinet) MEDBP Chapter 22 Media Backplane Replacement 22-3
FIGURE 22-3 MEDBP Mounting Locations (Front of M9000 with the Expansion Cabinet) Base cabinet MEDBP#0 Expansion cabinet MEDBP#1 TABLE 22-1 lists the abbreviated names and numbers of the MEDBPs.
22.2 Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2.
5. Switch off all main line switches for the AC section (ACS) in the power supply system. For details, see Section 4.5.3, “Main Line Switch Locations” on page 4-27. Note – If the optional three-phase 200 VAC input power supply unit is installed, the main line switches are located on the three-phase ACS. Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units.
M9000 (See FIGURE 22-5): a. Remove the tape drive unit. (See Chapter 15) b. Remove the CD-RW/DVD-RW drive unit. (See Chapter 14) c. Remove I/O units #0, #2, and #4 (IOU#0, IOU#2, and IOU#4) on the right side of the CD-RW/DVD-RW drive unit. A filler panel may be mounted depending on the option configuration. (See Chapter 7) 9. Remove the MEDBP to be replaced. a. Using a Phillips screwdriver, loosen the two thumb screws securing the MEDBP. b. Slide the MEDBP to the right. c.
FIGURE 22-4 22-8 Removing the MEDBP (Front of the M8000) SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 22-5 Removing the MEDBP (Front of the M9000) 10. Place the removed MEDBP on an antistatic mat. 11. Mount the replacement MEDBP by following the removal instructions in Step 8 and Step 9 in reverse order. Caution – Do not forcibly push the MEDBP in when inserting it. If the MEDBP is forcibly inserted despite the presence of any obstruction in a slot or any problems with the connector pin, serious damage may result. 12.
■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.2, “Powering the Server On” on page 4-19, and the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. 15. Restore the original mode switch setting on the operator panel.
CHAPTER 23 Switch Backplane Replacement This chapter explains the replacement procedure for switch backplanes (SWBPs). It covers the following: ■ ■ Section 23.1, “Overview of SWBPs” on page 23-1 Section 23.2, “Cold Replacement” on page 23-5 As a replacement method for SWBPs, only cold replacement can be used. For the definition of the replacement method, see Section 4.1, “Types of Replacement Procedures” on page 4-2. 23.
FIGURE 23-1 SWBP Mounting Location (Rear of the M8000) SWBP 23-2 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
FIGURE 23-2 SWBP Mounting Location (Front of the M9000 Base Cabinet) SWBP Chapter 23 Switch Backplane Replacement 23-3
FIGURE 23-3 SWBP Mounting Locations (Front of M9000 with the Expansion Cabinet) Base cabinet SWBP#0 Expansion cabinet SWBP#1 TABLE 23-1 lists the abbreviated names and numbers of the SWBPs.
23.2 Cold Replacement In cold replacement, none of the replacement work is performed with the maintenance menu from a terminal connected to XSCF. The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2.
Caution – If the dual power feed option is installed, switch off the main line switches for both power supply units. Caution – After stopping all the domains, be sure to switch off the main line switches in the power supply system to completely shut off the supply of power. 6. Connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port. See Section 1.4, “Antistatic Precautions” on page 1-6. 7.
FIGURE 23-4 Removing the SWBP (Rear of the M8000) Chapter 23 Switch Backplane Replacement 23-7
FIGURE 23-5 Removing the SWBP (Front of the M9000) 10. Place the removed SWBP on an antistatic mat. 11. Mount the replacement SWBP by following the removal instructions in Step 8 and Step 9 in reverse order. Caution – Do not forcibly push the SWBP in when inserting it. If the SWBP is forcibly inserted despite the presence of any obstruction in a slot or any problems with the connector pin, serious damage may result.
14. Start (turn on power) all domains together. Take either of the following two actions: ■ Press and hold down (for less than four seconds) the POWER switch on the operator panel. ■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.
23-10 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 24 Addition and Deletion of a RDPF Option, Power Cabinet, and M9000 Expansion Cabinet This chapter explains the procedures for adding and deleting a rack-mountable dual power feed option (RDPF), a power cabinet, and an M9000 expansion cabinet, all of which are server options. In the chapter, deletion of such an option means that the work of removing it involves moving the server. These options are not FRUs.
24.1 Addition and Deletion of Rack-mountable Dual Power Feed This section explains the procedure for replacing a rack-mountable dual power feed unit (RDPF). The sections below explain the following: ■ ■ Section 24.1.1, “Overview of RDPF” on page 24-2 Section 24.1.2, “Addition and Deletion” on page 24-3 Turn off the server power before adding or deleting a rack-mountable dual power feed option. 24.1.
FIGURE 24-1 RDPF Mounting Location (M8000) RDPF Front Rear TABLE 24-1 lists the abbreviated name of rack-mountable dual power feed of M8000. TABLE 24-1 Abbreviated Name of Rack-mountable Dual Power Feed Component name Abbreviated name Rack-mountable dual power feed RDPF 24.1.2 Addition and Deletion 24.1.2.1 Addition Note – When you order the rack-mountable dual power feed, it is attached with the rack-mountable kit (bus bar, fixing bracket, screws etc.
The procedure is as follows. 1. Unlock and open a door of the server. For details, see Chapter 5. Caution – Before touching FRU or the dummy (filler) unit, see Section 1.4, “Antistatic Precautions” on page 1-6 and be sure to wear the antistatic wrist strap. Failure to do so might result in serious damage to operating domains. 2. Change the mode switch setting on the operator panel from Locked to Service.
7. Ensure that the metallic underside of the wrist strap is in direct contact with your skin. The wrist strap should be snug around the wrist so that it does not rotate. See Section 1.4, “Antistatic Precautions” on page 1-6. Caution – Before handling FRUs, be sure to connect the clip of the antistatic wrist strap and of the antistatic mat to the server grounding port, and attach the band of the wrist strap to one of your wrists. Place the FRUs on the grounded antistatic mat.
FIGURE 24-2 Mounting Locations of cage Nuts 14 14 2 4 f. Mount the dual power feed unit at the front of the equipment rack of the M8000.
g. Mount the dual power feed unit (with four screws) on the two poles at the front of the equipment rack of the M8000. Note – When securing the rack-mountable dual power feed unit on the equipment rack, keep it positioned close to the PSUs of the M8000. FIGURE 24-3 Mounting the Screws of Rack-mountable Dual Power Feed (Rear of M8000) screws (three on both sides) h.
FIGURE 24-4 Mounting the Cable (Rear of M8000) connector panel bus bar conductive-springattached bracket m. Mount the bracket with a conductive spring (with two screws) at the bottom rear of the dual power feed unit. n. Mount the connector unit that was removed in Step h at its original location.
FIGURE 24-5 Mounting the Connector Unit (Rear of M8000) o. Mount the two brackets (with two screws each) for securing the ACS on the side panel at the rear of the equipment rack of the M8000. For their mounting locations, use the second and third holes from the bottom, and secure them. p. From the rear of the server, mount the ACS mounting frame, and secure it at the rear of the server (with three screws each on both sides).
FIGURE 24-6 Mounting the Frame for ACS Mount (Rear of M8000) r. Using two thumb screws, mount the AC section (ACS) on the frame for mounting the ACS. For details, see Chapter 18.
FIGURE 24-7 Removing the ACS (Rear of M8000) s. Mount nine power supply units (PSUs) at the front of the dual power feed unit. For details, see Chapter 9.
FIGURE 24-8 Mounting the PSU (Front of the M8000) screws 10. Connect power cords to the added rack-mountable dual power feed unit. 11. Switch on all main line switches for AC section (ACS) of both power supply systems. 12. Confirm that the XSCF STANDBY LED (green) on the operator panel remains lit. If it is blinking, wait until it remains lit. 13. Start (turn on power) all domains together.
■ From a console that is connected to XSCF, execute the poweron -a command of XSCF. The power to all the domains is turned on, and then the power-on self test (POST) is executed. When ok prompt displayed on the console of relevant domain, POST is completed. For details of the power-on operation, see Section 4.4.2, “Powering the Server On” on page 4-19, and the SPARC Enterprise M3000/M4000/M5000/M8000/M9000 Servers XSCF User’s Guide. 14.
■ Adding or deleting a power cabinet for the SPARC Enterprise M9000 server base cabinet with single-phase or three-phase dual power feed configured ■ Adding or deleting a power cabinet for the SPARC Enterprise M9000 server expansion cabinet with single-phase or three-phase dual power feed configured FIGURE 24-9, FIGURE 24-10, and FIGURE 24-11 show installation diagrams for each case.
FIGURE 24-10 M9000 Base Cabinet + Power Cabinet Power cabinet Chapter 24 Addition and Deletion of a RDPF Option, Power Cabinet, and M9000 Expansion Cabinet 24-15
FIGURE 24-11 M9000 Expansion Cabinet + Power Cabinet Power cabinet Power cabinet 24.3 Addition and Deletion of M9000 Expansion Cabinet The procedure listed below can be used. For details, see the power cabinet installation procedure explained in the SPARC Enterprise M8000/M9000 Servers Installation Guide. ■ Adding or deleting an M9000 expansion cabinet for the M9000 base cabinet FIGURE 24-12 shows an installation diagram of the cabinets.
FIGURE 24-12 M9000 Base Cabinet + M9000 Expansion Cabinet M9000 expansion cabinet Chapter 24 Addition and Deletion of a RDPF Option, Power Cabinet, and M9000 Expansion Cabinet 24-17
24-18 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
CHAPTER 25 Addition and Deletion of CMU, DIMM, IOU, HDD, PCI Cards and TAPEU This chapter explains the procedures for installing new FRUs and removing existing FRUs. These procedures are called addition and deletion. The following topics are covered in this chapter. ■ ■ Section 25.1, “Addition” on page 25-2 Section 25.2, “Deletion” on page 25-4 All FRUs can be replaced, as shown in the chapters preceding , but addition and deletion are limited to the following FRUs.
Some FRUs can be actively added or actively deleted: ■ ■ ■ ■ Active addition or removal of a CMU or IOU is possible using the DR function. Maintenance or addition of a DIMM module involves removing or installing a CMU. Active addition or removal of PCI cards (including link cards) is possible using the PHP function. Active addition of TAPEUs is possible Caution – For deletion, a filler panel (TABLE 25-1) must be installed in the empty slot. Without filler panels, the system might overheat.
FIGURE 25-1 Initial Page of the Maintenance Menu --------------------------------------------------Maintenance/Addition Menu Please select a type of FRU to be added. 1. CMU/IOU (CPU Memory Board Unit/IO Unit) 2. FAN (Fan Unit) 3. PSU (Power Supply Unit) --------------------------------------------------Select [1-3|c:cancel]: Note – DDC_A is displayed only for the M8000. Perform the operation according to the instructions displayed in the addition menu.
4. Add the FRUs. Some steps for adding a FRU are the same as for replacing a FRU. For more details about the addition procedure, refer to the appropriate chapter on FRU replacement. Note – In addition of an IOU, mount the CMU and IOU in the proper location. Otherwise, the added IOU may not operate after power-on. 5. Switch on all main line switches for the AC section (ACS) in the power supply system. 6. Confirm that the XSCF STANDBY LED (green) on the operator panel remains lit.
25.2.1 Active Deletion First, select a unit to be deleted as described below. 1. Log in to the XSCF. 2. Execute the deletion command (deletefru). XSCF> deletefru 3. When the following deletion menu is displayed, select the type of FRU to be deleted. FIGURE 25-2 Initial Page of the Maintenance Menu --------------------------------------------------Maintenance/Removal (Deletion) Menu Please select a type of FRU to be removed (deleted). 1.
3. Switch off all main line switches for the AC section (ACS) in the power supply system. 4. Delete the FRUs. The procedures are the same for removing of the replacement for the appropriate FRU. Some steps for deleting a FRU are the same as for replacing a FRU. For more details about the deletion procedure, refer to the appropriate chapter on FRU replacement. Caution – For deletion, a filler panel (TABLE 25-1) must be installed in the empty slot. Without filler panels, the system might overheat. 5.
APPENDIX A System Configuration This appendix describes the installation conditions and configuration of the SPARC Enterprise server. A.1 Installation Conditions The basic component of a domain configuration is the system board. The system board consists of only a single CPU/Memory Board Unit (CMU), or it consists of one CMU and one I/O Unit (IOU). The system board mounting conditions are as follows: ■ CPU Module (CPUM) Mount either two sets of two CPUMs, or one set of four identical CPUMs on the CMU.
Be sure to use the same type of CPUM in each pair (CPUM#0 and CPUM#1, or CPUM#2 and CPUM#3). ■ Memory Module (DIMM) Mount DIMMs in sets of 16 DIMMs on the CMU. Mount memory modules of the same size and rank in each group. ■ I/O Unit (IOU) When mounting an IOU, also mount the CMU whose component number represents the same mounting location as the IOU. Example: Be sure to mount IOU#1 together with CMU#1. A.2 System Configuration TABLE A-1 to TABLE A-3 outline the system configuration.
TABLE A-1 System Configuration for SPARC Enterprise M8000 Server Unit Number Cabinet 1 CMU 1 to 4 CPU module Comments 2 or 4 • In case of SPARC64 VI processor 2.28 GHz, L2 cache 5M bytes 2.4 GHz, L2 cache 6M bytes Two CPU cores per CPU chip Two threads per CPU core (four threads per CPU module) Maximum of 64 threads per system • In case of SPARC64 VII processor 2.52 GHz, L2 cache 6M bytes 2.
TABLE A-1 System Configuration for SPARC Enterprise M8000 Server (Continued) Unit Number HDD 0 to 4 Comments HDD #0, 1: Slot#0 must have an IOU onboard device card. HDD #2, 3: Slot#4 must have an IOU onboard device card.
TABLE A-2 System Configuration for SPARC Enterprise M9000 Server Unit Number Base cabinet 1 CMU 1 to 8 CPU module Comments 2 or 4 • In case of SPARC64 VI processor 2.28 GHz, L2 cache 5M bytes 2.4 GHz, L2 cache 6M bytes Two CPU cores per CPU chip Two threads per CPU core (four threads per CPU module) Maximum of 128 threads per system • In case of SPARC64 VII processor 2.52 GHz, L2 cache 6M bytes 2.
TABLE A-2 System Configuration for SPARC Enterprise M9000 Server (Continued) Unit Number Comments HDD #2, 3: Slot#4 must have an IOU onboard device card.
A maximum of 16 CMUs (64 CPU chips / 128 cores) and 16 IOUs can be mounted in this model. A total of 128 core expansion is possible. The primary units such as the CMU, IOU, power supply unit, and FAN unit are common among the SPARC Enterprise servers. This is a floor-stand type.
TABLE A-3 System Configuration for M9000 with Expansion Cabinet (Continued) Unit Number Comments Slot#0, 2, 4, 6: PCI-EX short cards or IOU onboard device cards Slot#1, 3, 5, 7: PCI-EX short cards or external I/O expansion unit connection cards HDD 0 to 4 HDD #0, 1: Slot#0 must have an IOU onboard device card. HDD #2, 3: Slot#4 must have an IOU onboard device card.
TABLE A-3 System Configuration for M9000 with Expansion Cabinet (Continued) Unit Number Comments Single-phase and dual power feed option Power cabinet (three-phase power feed option: includes 2 units of AC section) * Even if the number of mounted CMUs or IOUs is small, the server is shipped with a full set of FAN units and power supply units installed.
A-10 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
APPENDIX B Components This Appendix describes each component making up SPARC Enterprise. For the most up-to-date information, contact a certified service engineer. TABLE B-1 gives the name and abbreviation of each component, and whether active replacement and hot replacement can be applied to each, together with the chapter and the section where the replacement procedure is explained. A circle symbol ( ) means a component is replaceable. A dash (—) means a component is not replaceable.
TABLE B-1 FRU Components of High-end Server Models (Continued) Name: Abbreviation Active replacement Hot replacement Replacement procedure XSCF unit (M8000/M9000 Servers for base XSCFU_B cabinet) B.11 Chapter 11 XSCF Unit (For M9000 Server Expansion cabinet) B.11 XSCFU_C Chapter 11 CD-RW/DVD-RW drive unit: See B.12 DVDU Chapter 14 Tape drive unit: See B.13 TAPEU Chapter 15 Operator Panel: See B.14 OPNL — — Chapter 10 Sensor Unit: See B.
TABLE B-1 FRU Components of High-end Server Models (Continued) Name: Abbreviation Active replacement Hot replacement Replacement procedure FAN Backplane (For M9000 Server) B.24 FANBP_A — — Section 20.2.3 FAN Backplane (For M9000 Server) B.24 FANBP_B — — Section 20.2.3 FAN Backplane (For M8000 Server) B.24 FANBP_C — — Section 20.2.3 Media Backplane: See B.25 MEDBP — — Chapter 22 Switch Backplane: See B.26 SWBP — — Chapter 23 Rack-mountable dual power feed option: See B.
B.1 CPU Memory Board Unit The CPU memory board unit (CMU) consists of a CPU memory board (CMB) with a DC-DC converter mounted, a memory, and a CPU module (CPUM). It can be combined with an I/O unit to configure a system (domain). FIGURE B-1 shows the CMU. FIGURE B-1 CPU Memory Board Unit (CMU) The CMU has the following characteristics. B-4 ■ Active maintenance of CMUs is possible using the DR function.
B.2 ■ For the bus connecting SC and SC as well as SC and crossbar unit, a high-speed link is used. ■ For the bus connecting SC and MAC as well as SC and CPU, a high-speed link is used. ■ For the memory, DIMM is used. ■ The CMU has a label bearing the printed component number and manufacture's serial number. CPU Module CPU module consists of CPU and DDC. Up to four CPU modules can be mounted on a CMU. There are four types of CPU modules. FIGURE B-2 shows the CPU module.
B-6 ■ A dedicated LSI for CPU is mounted. ■ One CPU chip is mounted on a CPU module. In case of SPARC 64 VI processor, one CPU chip has two CPU cores, each of which has two threads. This means four threads per CPU module. In case of SPARC 64 VII processor, one CPU chip has four CPU cores, each of which has two threads. This means eight threads per CPU module. A thread is a component of a CPU that can execute threads. When viewed from Solaris OS, one thread is seen as one CPU.
B.3 Memory The type of memory used is DIMM. FIGURE B-3 shows the memory. FIGURE B-3 Memory The memory has the following characteristics. ■ Active maintenance of memory is possible using the DR function. ■ Maintenance or installation of a DIMM module involves removing or installing a CMU.
B.4 I/O Unit An I/O Unit (IOU) consists of an I/O board (IOB) with two types of control LSIs and a DDC mounted, an HDD, and a PCI cassette. An IOU can be combined with a CMU to configure a system (domain). FIGURE B-4 shows the IOU. FIGURE B-4 I/O Unit (IOU) The IOU has the following characteristics. B-8 ■ Active maintenance of IOUs is possible using the DR function. ■ Active replacement of a failed IOU during system operation is possible using the DR function.
■ An IOU can be added or removed using the DR function during system operation. ■ Active maintenance of PCI slots is possible using the PHP function. ■ The IOU has eight PCI slots with PCI cassettes mounted. PCI cards can be mounted on PCI cassettes. Up to four HDDs (2.5 inch with SAS interface) can be mounted on an IOU. ■ Up to eight PCI-Express short cards can be mounted.
B.5 Hard Disk Drive The hard disk drive (HDD) is mounted on an IOU. Up to four HDDs can be mounted on an IOU. FIGURE B-5 shows the HDD. FIGURE B-5 Hard Disk Drive (HDD) The HDD has the following characteristics. B.6 ■ Active replacement of a failed HDD during system operation is possible. ■ Active installation of HDDs during system operation is possible. ■ It allows mounting of a 2.5 inch HDD with an SAS interface.
FIGURE B-6 shows the PCICS. FIGURE B-6 PCI Express Short card cassette (PCICS) The PCI cassette has the following characteristics. ■ Active maintenance of PCICSs is possible using the PHP function. ■ Active installation or removal of PCI cards (including link cards) is possible using the PHP function. ■ A PCICS can have a short card with the PCI-Express interface mounted. ■ A PCICS has a label bearing the printed component number and manufacturer's serial number.
B.7 IOU Onboard Device Card The IOU onboard device card (IOUA) is an option for mounting CD-RW/DVD-RW drive units or tape drive units in SPARC Enterprise server cabinets. Connect to the CD-RW/DVD-RW drive unit or tape drive unit is controlled by IOUA. The IOUA to control these devices must be mounted on PCI slot 0 or 4 of the IOU. FIGURE B-7 shows the IOUA. FIGURE B-7 IOU Onboard Device Card (IOUA) The IOUA has the following characteristics. B-12 ■ Hot replacement of an IOUA is possible.
B.8 Link Card (External I/O Expansion Unit Connection Card) The external I/O expansion unit connection card includes two link cards. One link card is mounted on the host server. The other link card is mounted on the I/O boat. The cards are supplied with a cable for connection. FIGURE B-8 shows the PCIB-EX. FIGURE B-8 External I/O Expansion Unit Connection Card (PCIB-EX) The PCIB-EX has the following characteristics. ■ Active maintenance of a PCIB_EX is possible using the PHP function.
B.9 Crossbar Unit A crossbar unit (XBU: formal abbreviation is XBU_B) is mounted only in the SPARC Enterprise M9000 server. The XBU consists of a connector and a crossbar switch (XB). The connector is used to physically connect a CMU and IOU, and the XB switches logical connections and controls data transfer between the CMU and IOU. Up to eight XBUs can be mounted in a cabinet. FIGURE B-9 shows the XBU_B. FIGURE B-9 Crossbar Unit (XBU_B) The XBU has the following characteristics.
■ The XBU has a normal mode and a degraded mode. In normal mode, all eight sets operate; and in degraded mode in which a failure has occurred, the four sets containing the failed set are disconnected, and operation continues with the remaining four normal sets. If an XB fails, the system is shut down to disconnect the four sets including the failed XB, and it is then automatically recovered. ■ A dedicated LSI chip is used for the crossbar switch (XB).
B.10 Clock Control Unit A clock control unit (CLKU: formal abbreviation is CLKU_B) is mounted only in the SPARC Enterprise M9000 server, and it supplies clock signals to a CMU and crossbar unit (XBU). Furthermore, in the SPARC Enterprise M9000 server with the expansion cabinet, both CLKUs in the base cabinet and expansion cabinet are connected to enable each to supply clock signals to the other cabinet. In the M8000, two LSI chips for clock signals instead of the CLKU are mounted on the backplane.
B.11 XSCF Unit XSCF unit (XSCFU: eXtended System Control Facility unit) controls the system. Two types of XSCFU are available: the XSCFU_B for the high-end servers and XSCFU_C for the expansion cabinet of the M9000. XSCFU_B has a connector for cable connection to XSCFU_C provided for expansion cabinet use. XSCFU_C has no functions for controlling the system, but it provides an interface for passing the status of each unit in the expansion cabinet to XSCFU_B.
FIGURE B-12 eXtended System Control Facility Unit (XSCFU_C) XSCFU has the following characteristics. B-18 ■ The XSCFU has a duplicated configuration as the standard specification, allowing hot replacement of a failed XSCFU during system operation. ■ XSCFU is the central control unit of the system and has the following functions: ■ It controls each unit. ■ It works as the console of each domain by using the TTY communication protocol.
B.12 CD-RW/DVD-RW Drive Unit The CD-RW/DVD-RW drive unit (DVDU) can be mounted on each cabinet. One for the M8000/M9000 with the base cabinet, and up to two for the M9000 with an expansion cabinet can be mounted. The CD-RW/DVD-RW drive unit is controlled by the optional IOU onboard device card (IOUA). The CD-RW/DVD-RW drive unit is shown in FIGURE B-13. FIGURE B-13 CD-RW/DVD-RW Drive Unit (DVDU) The CD-RW/DVD-RW drive unit has the following characteristics.
B.13 Tape Drive Unit Tape drive unit (TAPEU) is optional and can be mounted on each cabinet. One for theM8000 /M9000 with the base cabinet, and up to two for the M9000 with an expansion cabinet can be mounted. TAPEU is controlled by the optional IOU onboard device card (IOUA). The TAPEU is shown in FIGURE B-14. FIGURE B-14 Tape Drive Unit (TAPEU) Tape drive unit has the following characteristics. B-20 ■ Active replacement or hot replacement of tape drive units is possible.
■ B.14 The tape drive unit has a label bearing the printed component number and manufacture's serial number. Operator Panel The operator panel (OPNL) allows the operation of the high-end server and the display of condition. The OPNL is mounted on the front door. With the M9000, the OPNL is mounted only on the base cabinet. The front and rear views of the OPNL are shown in FIGURE B-15 and FIGURE B-16 respectively.
FIGURE B-16 Rear View of the Operator Panel (OPNL) OPNL has the following characteristics. B-22 ■ Only cold replacement is available for the OPNL. ■ The following switches are mounted on the OPNL to enable operation of the server. ■ POWER switch to turn on the power. ■ Mode switch to control the POWER switch ■ The LED on the front of the OPNL displays the condition of the server. The LED of the rear of the OPNL displays the status of OPNL itself.
B.15 Sensor Unit The sensor unit (SNSU) monitors the intake-air temperature and humidity. In the M9000, this unit is mounted only on a base cabinet. When the output value of the sensor exceeds the reference value, XSCF issues an alarm to the system. FIGURE B-17 shows the SNSU. FIGURE B-17 Sensor Unit (SNSU) SNSU has the following characteristics. ■ Only cold replacement is available for the SNSU. ■ Only the temperature sensor and the humidity sensor are mounted.
B.16 Power Supply Unit The power supply unit (PSU) receives AC 200 to 240 V, and supplies each unit of the SPARC Enterprise with DC48 V and 12 VL. When all CMUs and IOUs are mounted without the dual power feed option, nine units for the M8000, 15 units for the M9000 (base cabinet only) and 30 units for the M9000 (with expansion cabinet) will be mounted. The number of mounted units also vary with the system configuration. The PSU is shown in FIGURE B-18.
B.17 AC Section AC section (ACS) is positioned at the input side of the power supply unit (PSU), to protect the system from electromagnetic interference (EMI). There are two types of ACS, ACS for M8000 and ACS for M9000. Both single-phase AC and three-phase AC types are available for both types. ACS consists of a terminal strip for power input, a main line switch, and so on.
■ The ACS_B is a single-phase power feed unit for the SPARC Enterprise M9000 server and the SPARC Enterprise M9000 server single-phase dual power feed cabinet (option). This unit uses five power cords to receive power and distribute AC power to 15 PSUs. FIGURE B-20 ■ The ACS_C is a single-phase power feed unit for the SPARC Enterprise M8000 server single-phase dual power feed cabinet (option). This unit uses three power cords to receive power and distribute AC power to nine PSUs.
FIGURE B-22 ■ The ACSTPH_B is a unit that receives three-phase (star) power for the SPARC Enterprise M8000 server three-phase dual power feed cabinet and switches between three-phase and single-phase. This unit uses two power cables to receive power, switch from three-phase to single-phase current internally, and distribute AC power to 18 PSUs.
FIGURE B-24 ■ The ACSTPH_D is a unit that receives three-phase (star) power for the SPARC Enterprise M9000 server three-phase dual power feed cabinet and switches between three-phase and single-phase. This unit uses two power cables to receive power, switch from three-phase to single-phase current internally, and distribute AC power to 30 PSUs.
■ The ACS_D is an internal current power feed unit for the SPARC Enterprise M8000 server/SPARC Enterprise M9000 server three-phase dual power feed cabinet. This unit supplies single-phase current to the server. It does not use external cables. FIGURE B-26 AC Section (ACS_D) ACS has the following characteristics. ■ Only cold replacement is available for the ACS.
B.18 FAN Unit The FAN unit consists of two to three cabinet cooling FAN units. Two types of FAN unit are available: the FAN_A type, which is the common FAN unit for SPARC Enterprise, and the FAN_B type, which is the a dedicated FAN unit for the M8000. FIGURE B-27 and FIGURE B-28 show the FAN unit.
FIGURE B-28 FAN Unit (FAN_B) The FAN unit has the following characteristics. ■ Active maintenance of a failed FAN unit during system operation is possible. ■ The FAN units are controlled by XSCF. ■ Even if a FAN unit fails during system operation, the redundant FAN unit configuration ensures that system operation can continue. ■ The FAN unit has a label bearing the printed component number and manufacturer's serial number.
B.19 Power Cabinet A dual power feeder (DPF) and a three-phase AC power feeder are mounted on the power cabinet. All of these are optional. The dual power feeder receives its power supply from two independent external power sources, ensuring that, even if a failure occurs in one source, power is supplied to the system from the other source that in normal state. It supports both single-phase and three-phase AC. The three-phase AC power feeder is required when the input power is three-phase AC.
The power cabinet is shown in FIGURE B-29. FIGURE B-29 Power Cabinet The power cabinet has the following characteristics. ■ Only cold replacement is available for the power cabinet. ■ It provides the mounting locations for all components of the dual power feeder (DPF) and the three-phase AC power feeder. There are no exclusive conditions for mounting. ■ The power cabinet has a label bearing the printed component number and manufacture's serial number.
B.20 Rack-mountable Dual Power Feed Rack-mountable Dual Power Feed (RDPF) is a dedicated option for M8000. The Rack-mountable Dual Power Feed receives its power supply from two independent external power sources, ensuring that, even if a failure occurs in one source, power is supplied to the system from the other source that in normal state. It is applicable only for single-phase AC input. For three-phase AC power input, a power cabinet is required.
B.21 Backplane The backplane (BP) is a printed circuit board (PCB) that has connectors and PC board used to interconnect the cabinet-mounted units. There are two types of BP for SPARC Enterprise available: namely, the BP_A for the M8000 and the BP_B for the M9000. The BP_A makes connections among CMU, IOU, XSCFU and DDC_A (See next section). The BP_B makes connections among CMU, IOU, XSCFU, CLKU and XBU. The BP is shown in FIGURE B-31 and FIGURE B-32.
FIGURE B-32 Backplane (BP_B) The BP has the following characteristics. B-36 ■ Only cold replacement is available for the BP. ■ Since there are no crossbar units for the M8000, such electronic devices as LSI for crossbar switches (XB) are mounted on BP_A. ■ BP_B, which consists of connectors and mechanical components only, has no electronic devices. ■ The BP for SPARC Enterprise has a label bearing the printed component number and manufacture's serial number.
B.22 DDC DC-DC converter (DDC_A) supplies DC power to electronic components mounted on the backplane (BP_A) for M8000. DDC_A receives 48 VDC and 12 VL and outputs 1.0 VDC, 1.8 VDC, and 3.3 VL. The DDC_A is shown in FIGURE B-33. FIGURE B-33 DC-DC Converter (DDC_A) DDC_A has the following characteristics. ■ Because DDC_As are normally operating in redundancy, active maintenance or hot maintenance of a failed DDC_A during system operation is possible.
B.23 PSU Backplane PSU backplane (PSUBP) is a PCB that has a connector for interconnecting multiple power supply units (PSU) concurrently operated. There are two types of PSUBP, namely the PSUBP_A commonly used for all SPARC Enterprise and the PSUBP_B for the M8000 only. The PSUBP is shown in FIGURE B-34 and FIGURE B-35. FIGURE B-34 PSU Backplane (PSUBP_A) FIGURE B-35 PSU Backplane (PSUBP_B) The PSUBP has the following characteristics.
B.24 FAN Backplane The FAN backplane (FANBP) is a PCB that has a connector for connecting multiple FAN units (FAN_A or FAN_B). Three types of FANBP are available: the FANBP_C for the M8000, and FANBP_A and FANBP_B for the M9000. FIGURE B-36, FIGURE B-37, and FIGURE B-38 show the FANBP.
FIGURE B-37 FAN Backplane (FANBP_B) FIGURE B-38 FAN Backplane (FANBP_C) FANBP has the following characteristics.
■ Because the FANBP is a built-in unit in the cabinet, only cold replacement is available for the FANBP. ■ FANBP connects FAN unit to the cabinet. ■ Up to 10 FAN units can be mounted on FANBP_A, and up to six FAN units can be mounted on FANBP_B and FANBP_C each. ■ The FANBP has circuits mounted for monitoring the fan rotation speed, switching between high/low speed, and controlling the CHECK LED. It is controlled by the XSCF.
B.25 Media Backplane The Media backplane (MEDBP) is a PCB that has connectors for mounting the CDRW/DVD-RW drive unit motherboard and the tape drive unit motherboard. On the CD-RW/DVD-RW drive unit motherboard and tape drive unit motherboard, the CD-RW/DVD-RW drive unit and the tape drive unit can be mounted, respectively. One MEDBP unit can be mounted on the M8000 and M9000 (base cabinet), and two units can be mounted on the M9000 (with expansion cabinet). The MEDBP is shown in FIGURE B-39.
B.26 Switch Backplane Switch backplane (SWBP) is a PCB that has connectors to connect Media BP to multiple IOUs. The SWBP is a unit that switches between IOUs to set which IOU is to control the CD-RW/DVD-RW drive unit and the tape drive unit (TAPEU). One SWBP unit can be mounted on the M8000 and M9000 (base cabinet), and two units can be mounted on the M9000 (with expansion cabinet). FIGURE B-40 shows the SWBP. FIGURE B-40 Switch Backplane (SWBP) SWBP has the following characteristics.
■ B-44 The SWBP has a label bearing the printed component number and manufacture's serial number.
CHAPTER C External Interface Specifications This appendix describes the specifications of the connectors provided on the SPARC Enterprise server unit: ■ ■ ■ ■ C.
C.2 UPC Port TABLE C-2 UPC Port Setting 5 4 3 2 1 9 8 7 6 C.
C.
C-4 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
APPENDIX D UPS Controller This appendix explains the UPS controller (UPC) that controls the uninterruptible power supply (UPS) unit. D.1 Overview The UPS is a unit to ensure a stable supply of power to the system even in the event of a power supply failure or a massive blackout.
D.3 Configuration of Signal Lines This section provides the signal definitions. FIGURE D-1 shows the configuration of signal lines when the UPS is connected. FIGURE D-1 Connection between the Server and UPS UPS TABLE D-1 7 *BTL 6 *BPS/*UALM 9 *ACOFF 5 SG UPS cable Server Signal Line Definitions Signal name Definitions Pin number Remarks *BPS/*UALM Indicates a UPS error.
TABLE D-1 Signal Line Definitions (Continued) Signal name Definitions Pin number Remarks *ACOFF Indicates that a power failure has occurred in the commercial power supply connected to the UPS. 9 Normal state: Off Abnormal state: On (Note 2) SG Signal ground 5 ER (Equipment Ready) Indicates that the server is operating 1 Do not connect anything to this pin. On: Indicates that the contact is closed. Off: Indicates that the contact is open.
D.4 Power Supply Conditions TABLE D-2 and TABLE D-3 show the power supply conditions of the UPC interface. D.4.1 Input Circuit TABLE D-2 Input Power Supply Conditions Signal name Input conditions *BPS/*UALM No voltage relay contact *BTL Contact rating DC 12 V, 10 mA or greater (0.5 A maximum) *ACOFF Use of a gold-plated contact or reed relay is recommended. Remarks: Limit the signal line chattering period to 1 ms. D.4.
D.5 UPS Cable The UPS cable specifications are as follows: ■ Connector shape D-SUB9 pin male (receptacle side: female) DEU-9PF-F0 ■ Terminal array FIGURE D-2 shows the pin signals of the UPC port and UPS cable. Do not use any pin that is not defined (Pin 2, 3, 4, and 8 in FIGURE D-2).
D.6 Connections This section explains the connection between the UPC port and UPS. D-6 ■ UPC#0 connects to UPS#0, and UPC#1 connects to UPS#1. ■ Single power feed uses only UPC#0. ■ Dual power feed uses UPC#0 and UPC#1.
D.7 UPC Port This chapter describes the location of the UPC port and the UPS connections. ■ UPC#0 connects with UPS#0. UPC#1 connects with UPS#1. ■ The single power feed uses UPC#0 only. ■ The dual power feed option uses UPC#0 and UPC#1. FIGURE D-3 show the location of the UPC port in the main unit.
D-8 SPARC Enterprise M8000/M9000 Servers Service Manual • June 2010
APPENDIX E XSCF Unit Replacement When XCP 1040 or 1041 Is in the Server The server must be updated to XCP 1050 or later. Note – XCP 1050 or later cannot be downgraded to XCP 1040 or XCP 1041. 1. Turn off all domains. 2. Turn the Mode switch to Service. 3. Turn off all main line switches. 4. Replace XSCFU#0 with spare material, and reconnect cables to XSCFU#0. 5. Extract XSCFU#1 6. Turn on all main line switches.
12. Wait until XSCF firmware reaches the ready state. This can be confirmed when the READY LEDs of XSCFU_B#0 and XSCFU_B#1 remain lit. 13. Log in to XSCFU#0 using a serial connection or LAN connection. 14. Confirm no abnormality occurred by using showlogs error -v and showstatus commands. XSCF> showlogs error -v XSCF> showstatus If you encounter any hardware abnormality of the XSCF, contact service engineer. 15. Confirm and update the imported XCP image again.
19. Confirm switching over between XSCFs works properly. XSCF> switchscf -t Standby The XSCF unit switch between the Active and Standby states. Continue? [y|n] :y a. When the READY LED on XSCFU_B#1 remains lit, log in to XSCFU#0 using a serial connection or LAN connection. b. Confirm switching over between XSCFs using the following commands: XSCF> showhardconf Confirm XSCFU#1 is now the standby, and that XSCFU#0 has become the active.
23. Turn the key switch on the operator panel from Service to Lock.
CHAPTER Abbreviations A E ACS AC Section ACSTPH ACS Three-Phase B BP Backplane BUI Browser User Interface C EMI Electromagnetic Interference F FAN FAN Unit FRU Field-Replaceable Unit H HDD Hard Disk Drive CMB CPU Memory Board CMU CPU/Memory Board Unit I CLI Command-Line Interface IOB I/O Board CLKU Clock Control Unit IOU I/O Unit CPUM CPU Module IOUA IOU Onboard Device Card_A M D DAT Digital Audio Tape MAC Memory Access Controller DDC DC to DC Converter MEDBP Me
P PCICS PCI Cassette PCI-ES PCI-Express Short PFC Power Factor Correction PHP PCI Hot Plug POST Power-On Self-Test PSU Power Supply Unit R RCI Remote Cabinet Interface RDPF Rack-mountable Dual Power Feed S SAS Serial Attached SCSI SATA Serial ATA (Advanced Technology Attachment) SC System Controller SNSU Sensor Unit SWBP Switch Backplane T TAPEU Tape Drive Unit U UPS Uninterruptible Power Supply X XBU Crossbar Unit XSCF eXtended System Control Facility XSCFU eXtended Syste
Index A AC section, B-25 AC section replacement, 18-1 ACS cold replacement, 18-5 active replacement, 4-2, 4-3 CD-RW/DVD-RW, 14-5 CMU, 6-12 DDC, 19-3 DVDU, 14-5 FAN, 8-8 HDD, 12-3 IOU, 7-8 PCI, 13-17 PSU, 9-8 TAPEU, 15-5 XSCFU, 11-5 addition and deletion, 24-1 addition and deletion of CMU, IOU, Fan, PSU, and PCI cards, 25-1 addition and deletion of M9000 expansion cabinet, 24-16 addition and deletion of power cabinet, 24-13 addition and deletion of rack mountable dual power feed, 24-2 air filter (I/O unit),
IOU, 7-17 MEDBP, 22-5 OPNL, 10-4 PCI, 13-20 PSU, 9-12 PSU BP, 20-1 servers BPs, 20-1 SNSU, 21-4 SWBP, 23-5 TAPEU, 15-9, 15-11 XBU, 17-4 XSCFU, 11-12 commands for checking hardware configuration and status, 2-24 commands for checking the messaging output, 2-30 commands for checking the monitoring output, 2-29 component, B-1 components to access and doors to open, 5-2 Conditions of memory mount, 6-40 confirming the hardware, 4-7, 4-15, 4-20 corresponding components and doors that can be accessed, 5-2 CPU and
I I/O box downlink card, B-13 I/O unit, B-8 I/O unit (IOU) replacement, 7-1 ID label, 2-11 incorporating CMU/IOU, 4-6 FRU into a domain, 4-6 internal components access, 5-1 iostat command, 2-38 IOU, B-12 active replacement, 7-8 cold replacement, 7-17 hot replacement, 7-8 K key, 2-1, 2-16 L labels about handling, 2-15 LED, 2-18 LED display, 2-31 LED error display, 2-30 M M9000 expansion cabinet, 24-1 MEDBP cold replacement, 22-5 media backplane, B-42 media backplane replacement, 22-1 memory, B-7 messaging
PSU active replacement, 9-8 cold replacement, 9-12 hot replacement, 9-8 PSU backplane, B-38 PSU BP cold replacement, 20-1 TAPEU active replacement, 15-5 cold replacement, 15-9, 15-11 traditional Solaris troubleshooting commands, 2-37 troubleshooting, 2-1 troubleshooting commands, 2-34 types of Replacement Procedures, 4-2 R U rack-mountable dual power feed, B-34 rack-mountable dual power feed option, 24-1 rating label, 2-11 replacement of CPU/memory board unit (CMU), CPU, and DIMM, 6-1 replacing a CPU mo
