HP Hewlett Packard INTEGRITY RX3600 User Manual

HP Integrity rx3600  
User Service Guide  
HP Part Number: AB463-9003C  
Published: November 2007  
Edition: Third edition  
Table of Contents  
Table of Contents  
3
4
Table of Contents  
Table of Contents  
5
6
Table of Contents  
Table of Contents  
7
8
Table of Contents  
Table of Contents  
9
10  
Table of Contents  
Table of Contents  
11  
12  
Table of Contents  
14  
List of Figures  
15  
16  
List of Figures  
List of Tables  
17  
18  
List of Tables  
List of Examples  
19  
20  
About This Document  
This document describes how to troubleshoot and diagnose server problems, and remove and  
replace server components for the HP Integrity rx3600, Regulatory Model Number: RSVLA-0404.  
The document printing date and part number indicate the documents current edition. The  
printing date will change when a new edition is printed. Minor changes may be made at reprint  
without changing the printing date. The document part number will change when extensive  
changes are made.  
Document updates may be issued between editions to correct errors or document product changes.  
To ensure that you receive the updated or new editions, you should subscribe to the appropriate  
product support service. See your HP sales representative for details.  
The latest version of this document can be found online at http://www.docs.hp.com.  
Intended Audience  
This document is intended to provide technical product and support information for authorized  
service providers, customer system administrators, and HP support personnel.  
New and Changed Information in This Edition  
This document has been updated to include new processor information.  
Publishing History  
Table 1 lists the publishing history details for this document.  
Table 1 Publishing History Details  
Document Manufacturing Part Number  
AB463-9003A  
Publication Date  
September 2006  
February 2007  
November 2007  
AB463-9003B  
AB463-9003C  
Document Organization  
The HP Integrity rx3600 User Service Guide is divided into several chapters, each chapter contains  
information about servicing the HP Integrity rx3600. The appendixes contain supplemental  
information.  
Use this chapter to develop a high-level understanding of the major  
server subsystems.  
Use this chapter to learn about the system specifications.  
Use this chapter to learn how to install the system.  
Use this chapter to learn about booting and shutting down the  
operating system.  
Use this chapter to help troubleshoot and diagnose server problems.  
Use this chapter to learn how to remove and replace server  
components.  
Use this appendix for parts information.  
Use this appendix for instructions on how to configure core I/O  
cards.  
Use this appendix to learn about utilities such as EFI and the iLO  
2 MP.  
Intended Audience  
21  
           
Typographic Conventions  
This document uses the following conventions:  
audit(5)  
An HP-UX manpage. In this example, audit is the name and 5 is the section  
in the HP-UX Reference. On the Web and on the Instant Information CD, it  
may be a hot link to the manpage itself. From the HP-UX command line,  
you can enter “man audit” or “man 5 audit” to view the manpage.  
See man(1).  
Book Title  
The title of a book. On the Web and on the Instant Information CD, it may  
be a hot link to the book itself.  
KeyCap  
The name of a keyboard key. Note that Return and Enter both refer to the  
same key.  
Emphasis  
Text that is emphasized.  
Bold  
Text that is strongly emphasized.  
Text displayed by the computer.  
Commands and other text that you type.  
A command name or qualified command phrase.  
The name of a variable that you can replace in a command or function or  
information in a display that represents several possible values.  
ComputerOut  
UserInput  
Command  
Variable  
[ ]  
{ }  
The contents are optional in formats and command descriptions. If the  
contents are a list separated by a pipe (|), you must select one of the items.  
The contents are required in formats and command descriptions. If the  
contents are a list separated by a pipe (|), you must select one of the items.  
...  
|
The preceding element can be repeated an arbitrary number of times.  
Separates items in a list of choices.  
Related Documents  
The HP Server Documentation CD has been provided with the server. It contains a complete  
documentation set for the server, including localized versions of key documents.  
Other documents in this collection include:  
HP Integrity rx3600 Installation Guide  
HP Integrity rx3600 Site Preparation Guide  
HP Integrity rx3600 Safety and Regulatory Information  
HP Integrity rx3600 & rx6600 Console Quick Start  
HP Integrity iLO 2 MP Operations Guide  
Generalized Site Preparation Guide  
The latest versions of these documents, and any updates to these documents are posted under  
the Enterprise Servers, Workstations and Systems Hardware collection under HP Integrity rx3600 at  
Warranty Information  
The latest versions of the BCS Global Limited Warranty and Technical Support document is posted  
in the Enterprise Servers, Workstations and Systems Hardware collection under each server to which  
Related Information  
You can find other information on HP server hardware management, Microsoft® Windows®,  
and diagnostic support tools at the following Web sites.  
22  
       
Web Site for HP Technical Documentation: http://docs.hp.com  
This is the main Web site  
for HP technical documentation. This site offers comprehensive information about HP products  
available for free.  
Server Hardware Information: http://docs.hp.com/hpux/hw/ This Web site is the systems  
hardware portion of the docs.hp.com site. It provides HP nPartition server hardware management  
details, including site preparation, installation, and more.  
Diagnostics and Event Monitoring: Hardware Support Tools:  
hardware support tools, including online and offline diagnostics and event monitoring tools.  
This site has manuals, tutorials, FAQs, and other reference material.  
This site contains complete information about HP  
Web Site for HP Technical Support: http://us-support2.external.hp.com  
The HP IT  
resource center Web site provides comprehensive support information for IT professionals on a  
wide variety of topics, including software, hardware, and networking.  
Server Support Information: http://www.hp.com/support  
Refer to this Web site for online  
access to technical support information, self-solve tools, online assistance, community forums  
of IT experts, broad multi-vendor knowledge base, monitoring and diagnostic tools. Support  
can be obtained from the US and Canada at : 1-800-593-6631  
Alerts and Notification: http://www.hp.com/go/e-updates  
HP's driver and support alerts/notifications program.  
Use this site to subscribe to  
HP Encourages Your Comments  
HP encourages your comments concerning this document. We are truly committed to providing  
documentation that meets your needs.  
Send comments to:  
Include document title, manufacturing part number, and any comments, errors found, or  
suggestions for improvement you have concerning this document. Also, include what we did  
right so we can incorporate it into other documents.  
HP Encourages Your Comments  
23  
 
24  
1 Overview  
The HP Integrity rx3600 server is a high performance, high availability server that provides the  
following features:  
Form factor - 4U rack- or pedestal-installed  
Internal peripherals - serial-attached SCSI (SAS) hard drives and a DVD or DVD+RW drive  
High availability components - N+1 hot-swappable fans, 1+1 hot-swappable power supplies,  
and hot-pluggable PCI-X or PCIe cards and SAS drives  
Processors - one or two dual-core Intel® Itanium® processors  
Memory - up to 96 GB of memory  
This chapter addresses the following topics:  
Server Subsystems  
I/O  
The server may contain either a 10-slot PCI/PCI-X IOBP or a 10-slot PCI/PCI-X/PCIe IOBP. The  
PCI/PCI-X/PCIe interfaces are classified under two major categories: public and private.  
Public interfaces Public interfaces are PCI slot connectors that are left available to the customer  
for the I/O-card adapters which they wish to install, based on their application needs, provided  
the adapter is supported in the slot. All public slots support HP server traditional OL* hot-plug  
operations.  
Private interfaces  
Private interfaces are interfaces dedicated to meeting the core I/O system  
requirements. There are two types of core-I/O: fast-core and slow-core, also known as the UCIO  
(Unified Core-I/O). The Private slots and UCIO are not hot-pluggable.  
Hot-plug  
PCI/PCI-X/PCIe hot-pluggable operations enable you to add or remove a  
PCI/PCI-X/PCIe card while the server is powered on. Each slot is physically separated with a  
divider that contains a card extraction mechanism and a mechanical retention latch (MRL). Slots  
3 - 10 on the 10 slot PCI/PCI-X IOBP and slots 5 - 10 on the 10 slot PCI/PCI-X/PCIe IOBP support  
HP-server traditional OL* "hot-plug" operations. PDHP slots rely on the facilities of DHPC (Dual  
Hot-plug Controller) FPGAs to enable OL* hot-plug functionality. See Table 1-1 (page 27) and  
Table 1-2 (page 28) for details.  
LEDs Attention LEDs, doorbell buttons, and power LEDs for each hot-pluggable PCI/PCI-X/PCIe  
slot are located on the PCI bulkhead at the rear of the server chassis.  
Server Subsystems  
25  
       
Wake on LAN  
computers through special network packets, is enabled in core PCI/PCI-X slots one and two.  
Ropes Ropes is an HP-proprietary, custom bus interface. It clocks data packets across  
Wake on LAN, a hardware feature that enables the remote power on of  
long-length nets using source synchronous clocking schemes. There are eighteen signals bundled  
per Ropes group, of which there are a total of 16 in the rx3600. Each group is capable of 0.5GB/s  
peak bandwidth. Table 1-1 (page 27) and Table 1-2 (page 28) display the association of the Ropes  
group(s) to the PCI slot / HBA to which they connect.  
Figure 1-1 (page 26) is a block diagram of the PCI/PCI-X I/O subsystem, and Figure 1-2 (page 27)  
is a block diagram of the PCI/PCI-X/PCIe I/O subsystem.  
Figure 1-1 I/O Subsystem Block Diagram  
Unified Core IO Board  
USB  
USBB  
USB  
32 bit  
PCI-33  
Single - Rope  
DMD  
ZX1 PCI Bridge  
I2C  
USB  
DMD  
Bridge  
BMC  
IPMB  
RMP3  
Video  
LPC  
VGA  
TPM  
MP LAN  
ROM  
UART  
RTC  
COM  
COM  
DVD  
SRAM  
Common Display Panel board  
Public PHP I/O Slots  
DHPC  
DHPC  
Single - Rope  
Single - Rope  
PCIx-66  
ZX1 PCI-X Bridge  
ZX1 PCI-X Bridge  
PCIx-66  
PCIx-133  
PCIx-133  
Dual - Rope  
Dual - Rope  
Quad - Rope  
Quad - Rope  
ZX1 PCI-X Bridge  
ZX1 PCI-X Bridge  
ZX2 PCI-X Bridge  
ZX2 PCI-X Bridge  
PCIx-266  
PCIx-266  
Ports From  
Core Disk Storage  
PORT  
PORT  
Chassis Rear  
Single - Rope  
PCIx-66  
ZX1 PCI-X Bridge  
Private Slots  
Ports From  
Chassis Front  
Core LAN  
LAN AN  
Common IO Board  
26  
Overview  
     
Figure 1-2 PCI/PCI-X/PCIe I/O Subsystem Block Diagram  
Unified Core IO Board  
USB  
USBB  
USB  
32 bit  
PCI-33  
Single - Rope  
DMD  
ZX1 PCI Bridge  
I2C  
USB  
DMD  
Bridge  
BMC  
IPMB  
SP  
Video  
LPC  
VGA  
TPM  
MP LAN  
ROM  
UART  
RTC  
COM  
COM  
SRAM  
DVD  
Common Display Panel board  
Public PHP I/O Slots  
DHPC  
Single - Rope  
Single - Rope  
PCIx-66  
ZX1 PCI-X Bridge  
ZX1 PCI-X Bridge  
PCIx-133  
PCIx-133  
X8 PCI-E  
Dual - Rope  
Quad - Rope  
Quad - Rope  
Dual - Rope  
ZX1 PCI-X Bridge  
ZX2 PCI-E Bridge  
ZX2 PCI-E Bridge  
ZX2 PCI-E Bridge  
X8 PCI-E  
X8 PCI-E  
Express  
Switch  
Ports From  
Core Disk Storage  
PORT  
PORT  
Chassis Rear  
Single - Rope  
PCIx-66  
ZX1 PCI-X Bridge  
Private Slots  
Ports From  
Chassis Front  
Core LAN  
LAN AN  
Common IO Board  
PCI/PCI-X IOBP  
On the 10 slot PCI/PCI-X IOBP there are a total 8 Public slots (PCI-X mode1/2), 2 Private Fast-core  
slots (PCI-X mode1), and 1 Private Slow-core UCIO slot (PCI 32-bit/33-MHz). The eight Public  
slots are further subdivided into three speed/bandwidth configurations:  
Four PDHP (Public Dual Hot-plug), which operate at 64-bit/66-MHz PCI-X  
Two PSHP-SDR (Public Single Hot-plug - Single Data Rate), operating at 64-bit/133-MHz  
PCI-X  
Two PSHP-DDR (Public Single Hot-plug - Double Data Rate), at 64-bit/266-MHz (133-MHz  
double clocked) PCI-X mode2  
Four 66-MHz PCI/PCI-X slots are shared in groups of two. Shared slots have many speed and  
mode change restrictions during hot-plug add or remove operations. For more information on  
PCI/PCI-X card configuration and restrictions, see “PCI/PCI-X/PCIe Configurations” (page 62).  
Four of the server PCI/PCI-X slots are not shared. Two of the nonshared slots are 133 MHz  
PCI/PCI-X, and two of the slots are 266-MHz PCI/PCI-X. Shared slots have many speed and  
mode change restrictions during hot-plug add or remove operations.  
Table 1-1 PCI/PCI-X I/O Rope Groups  
Slot #  
Rope Numbers PCI Bus Bits  
Speed  
Function  
Hot Swap /  
OL*  
1, 2  
3
8
1
2
64  
64  
66 MHz  
Core I/O (Private)  
N
Y
12, 13, 14, 15  
(Quad  
Bandwidth)  
266 MHz  
PCI-X 2.0, DDR (Public)  
4
5
4, 5, 6, 7 (Quad  
Bandwidth)  
3
4
64  
64  
266 MHz  
133 MHz  
PCI-X 2.0, DDR (Public)  
Y
Y
10, 11 (Dual  
Bandwidth)  
High-Speed PCI-X (Public)  
Server Subsystems  
27  
     
Table 1-1 PCI/PCI-X I/O Rope Groups (continued)  
Slot #  
Rope Numbers PCI Bus Bits  
Speed  
Function  
Hot Swap /  
OL*  
6
2, 3 (Dual  
Bandwidth)  
5
64  
133 MHz  
High-Speed PCI-X (Public)  
Y
7, 8  
9, 10  
-
9
1
0
6
7
0
64  
64  
32  
66 MHz  
66 MHz  
33 MHz  
General PCI-X (Public)  
General PCI-X (Public)  
UCIO (Private)  
Y
Y
N
PCI/PCI-X/PCIe IOBP  
On the 10 slot PCI/PCI-X/PCIe IOBP there are a total of eight Public slots (four PCI-X mode 1  
and four PCIe), two Private Fast-core slots (PCI/PCI-X mode 1 64-bit/66-MHz), and one Private  
Slow-core UCIO slot (PCI 32-bit/33-MHz). The eight Public slots are further divided into three  
speed/bandwidth configurations:  
Two PDHP, which operate at 64-bit/66-MHz PCI-X  
Two PSHP, operating at 64-bit/133-MHz PCI-X  
Four PCIe 8-lane (x8) 2.5 Gbps, two of which are switched  
The two 66 MHz PCI/PCI-X slots are shared. Shared slots have many speed and mode change  
restrictions during hot-plug add or remove operations.  
Table 1-2 PCI/PCI-X/PCIe I/O Rope Groups  
Slot #  
1, 2  
Rope Numbers  
Bits  
64  
Speed  
Function  
Hot Swap/OLR  
8
66 MHz  
2.5 Gbps  
Core I/O (Private)  
N
N
3, 4  
10, 11  
x8  
PCIe x8 (Public or Private depending upon  
Core I/O)  
5
12, 13, 14, 15  
x8  
x8  
64  
64  
64  
32  
2.5 Gbps  
2.5 Gbps  
133 MHz  
133 MHz  
66 MHz  
33 MHz  
PCIe x8 (Public)  
Y
Y
Y
Y
Y
N
6
4, 5, 6, 7  
PCIe x8 (Public)  
7
2, 3  
9
High-Speed PCI-X (Public)  
High-Speed PCI-X (Public)  
General PCI-X (Public)  
UCIO (Private)  
8
9, 10  
-
1
0
PCIe MPS Optimization  
For PCIe-based systems, each PCIe device has a configurable MPS (maximum payload size)  
parameter. Larger MPS values can enable the optimization to gain higher performance.MPS  
Optimization is supported on PCIe systems running HP-UX, Open VMS, and Linux. System  
firmware level greater than 02.03 performs an optimization during boot time to set the MPS  
value to the largest size supported by both a PCIe root port and the devices below it.  
The default server state is optimization disabled. When disabled system firmware sets MPS to  
the minimum value on each PCIe device.  
The info iocommand will display the current PCIe MPS optimization setting. See “info”  
To enable PCIe MPS optimization use the ioconfig mps_optimizecommand. See “ioconfig”  
For non-PCIe systems, ioconfigand info iowill not display the MPS optimization policy  
setting. The Set PCIe MPS Optimization boot manager menu also will not be displayed. Running  
28  
Overview  
           
the ioconfig mps_optimize [on|off]command from a non-PCIe system, the following  
output will be displayed:  
-------------  
Shell> ioconfig mps_optimize  
ioconfig: PCIe MPS optimization is not supported.  
Shell> ioconfig mps_optimize on  
ioconfig: PCIe MPS optimization is not supported.  
Exit status code: Unsupported  
Shell>  
-----------------  
To restore MPS to the default values use the default clearcommand. See “default” (page 327).  
Processor  
The server processor subsystem accommodates one or two dual-core Itanium® processor modules.  
The processor subsystem consists of the following elements:  
zx2 CEC front side bus, memory, and I/O controller  
System clock generation and distribution  
Circuitry for manageability and fault detection  
The zx2 CEC and the processor modules are located on the processor board assembly. Each  
processor connects to the processor board through a zero insertion force (ZIF) socket. The processor  
board is mounted on a removable carrier tray that is attached to the processor board access door.  
Access the assembly through the front of the server after the memory carrier is removed.  
Memory  
The server DIMMS are seated on memory boards that are enclosed in an extractable memory  
carrier assembly. The memory boards plug directly into sockets on the processor board when  
the memory carrier assembly is fully seated.  
Table 1-3 lists the two types of supported memory carriers and the memory configurations of  
each carrier.  
Table 1-3 Supported Memory Configurations  
Memory Carrier Type  
Memory Boards Installed  
Minimum Memory  
Configuration  
Maximum Memory  
Configuration  
8-DIMM memory carrier  
(standard)  
Two 4-DIMM memory  
boards  
2-GB (one pair: two 1-GB  
DIMMs)  
32-GB (eight 4-GB DIMMs)  
24-DIMM memory carrier Two 12-DIMM memory  
(optional, high-capacity) boards  
2-GB (one quad: four  
512-MB DIMMs)  
96-GB (24x4-GB DIMMs)  
The server supports the following DIMM sizes:  
512-MB  
1-GB  
2-GB  
4-GB  
Figure 1-3 is a block diagram of the 8-DIMM memory carrier that shows data, addresses, and  
controls that flow through the CEC and to and from the processors.  
Server Subsystems  
29  
           
Figure 1-3 8-DIMM Memory Carrier Block Diagram  
Figure 1-4 is a block diagram of the 24-DIMM memory carrier that shows data, addresses, and  
controls that flow from the processors through the CEC and memory mux to the DIMMs.  
30  
Overview  
   
Figure 1-4 24-DIMM Memory Carrier Block Diagram  
Cooling  
There are three system fans assemblies that cool the server. The fans are redundant, hot-swappable,  
and interchangeable. You service the fans through the top of the chassis. The fans are N+1,  
meaning that the server has three fans, but can operate for a very limited time with two fans  
running. If the time threshold is reached, the server automatically shuts down to prevent an  
overtemperature condition.  
Server Subsystems  
31  
         
Figure 1-5 Cooling Fans  
Power  
The server power subsystem is designed to provide high availability with 1+1 power supply  
redundancy. Figure 1-6 is a block diagram of the power subsystem, including voltage labels for  
each main server subsystem that requires power.  
32  
Overview  
     
Figure 1-6 Power Subsystem Block Diagram  
System POL Converters  
PDH  
I/O BP  
Fan contlr  
12/5/06  
POL  
V2P 5  
POL  
V1P 2  
V12  
V12  
Processor BP  
Unified Core  
V3P3 is called  
V3P3_PCI  
POL  
Summit  
Loop  
V12  
Fan contlr  
POL  
Summit  
Loop  
CPU PODS  
V12  
V1P 5  
POL  
V12  
V12  
(2)  
V1P 0  
POL  
Summit  
Loop  
V1P 2_FPGA  
embedded  
V12  
V1P 2  
Summit  
POL  
Loop  
Tower of Power  
V5P 0  
POL  
V5P0  
V3P3  
V12  
Mem Ext (2)  
V2P 5_FPGA  
embedded  
V1P 5  
POL  
V3P 3  
POL  
V12  
V12  
V12  
V3P3_STBY  
V12  
Summit  
Loop  
12/24 DIMM  
V1P 8_SYS _PLL  
embedded  
V1P 2  
V12N  
POL  
POL  
Summit  
Loop  
V12  
V1P 8  
POL  
V1P 8_RMP 3  
embedded  
Summit  
Loop  
V1P 8  
POL  
Summit  
Loop  
V3P3_STBY  
V12  
Mid-Plane  
Interconnect  
CCA  
Embedded design  
V3P3_STBY  
V2P 5  
FPD CCA  
V3P3  
V12  
POL  
V3P3  
V3P3  
V3P3_STBY  
V0P 9  
embedded  
V12  
Summit  
Loop  
V5P 0  
V3P3_STBY  
SPocOha, JLim  
V3P3_STBY  
V2P5  
SAS CCA  
12/24 DIMM  
CIOBP – Target CCA  
V12  
V3P3  
V3P3_STBY  
V5P0  
V12  
BPS  
(2)  
Possibly share Source and Load  
Processor Side  
I/O Side  
Two hot-swappable ac/dc power supplies generate main system power, and a standby power  
voltage. One active power supply is sufficient to operate the system at maximum load. Each  
power supply receives ac power through the integrated ac inlet. The system can operate at 100-240  
VAC and achieve 1+1 redundancy. The power supplies are power factor corrected and the  
maximum dc power output of the power system is 1095 watts. Service the hot-swappable power  
supplies are serviced by sliding them out the rear of the chassis.  
Applying system power in normal customer usage, the rx3600 runs on 100 to 240 V. Standby  
power will be supplied on either; hence the BMC will power up when the power supplies are  
plugged in. The BPS0_AC_OK and BPS1_AC_OK signals indicate whether the ac voltage to the  
power supplies is within the required range. If neither BPS0_AC_OK nor BPS1_AC_OK is asserted,  
then the BMC should log an event and prevent the system from turning on.  
Power Button  
The power button on the rx3600 is a momentary contact push button. The BMC  
polls the front panel power button at a rate of at least 2 Hz. The power button is an input to the  
System Power State Management. If the system is off, a single button press will turn on the  
system. If the system has booted to an OS, and a short button press is detected, a  
graceful-shutdown request will be sent to the system by pulsing ACPI_PWR_BTN_L; when the  
ACPI bits are set to note the O/S has shut down, the BMC will perform a hard power down. If  
the system has not booted to an OS, or if a long (5 second) button press is detected, the system  
will do an immediate hard power off.  
System Power State Management  
The system power may be controlled from the power button,  
an IPMI Chassis command, Wake-On-LAN, loss or gain of ac.  
Power On Sequence:  
1. Update the cache of DIMM SPD information.  
2. Ensure that the memory board is detected and that the cpu board has a processor in socket  
0. If these FRUs are not detected the BMC logs an event against the Missing Device sensor  
(sensor 0x15).  
3. Check for a BPS0_AC_OK or a BPS1_AC_OK signal. If neither is asserted, then the ac supply  
has a problem.  
4. If any FRUs are missing or both ac supplies are not valid, then return to power off state.  
Server Subsystems  
33