New products: M5000-GR20, M5000-CL09-PI01, M5000-CL10-PI01, M5000-CL11-12-PI01
Withdrawn products:
Some minor corrections: Chapter
SystemArchitect PRIMEPOWER (PW) Hotline
For further questions related to configurations for SE M5000 Systems with PC-ARCHITECT or SYSTEM-ARCHITECT
do not hesitate to get in contact with SystemArchitect PRIMEPOWER (PW) Hotline:
mailto:SA-PW.Hotline@fujitsu-siemens.com
Please add the respective configuration file (SAR-FILE)!
Contents
Recommended 19"- rack PRIMECENTER 2 with 1100 mm depth 3
Overview M4000 and M5000 4
System Architecture 5
Product Overview SPARC® Enterprise M5000 6
Features 6
M5000 Server Product ID List 9
M5000 COD (Capacity on Demand) Server Product ID List 10
M5000 COD (Capacity on Demand) CPU Right to Use (License) Product ID List 10
M5000 Internal Storage Device Product ID List 10
M5000 Filter Product ID List 10
M5000 Power Cords for Rack internal use Product ID List 11
Rack integration order numbers 11
Power Cords for PSU connection directly to Rack external Jacks: 11
SolarisTM Operating System (OS) and other Software Product ID List 12
M5000 PCI Card Product ID List 13
Ethernet Controller 13
Serial Interface RS232 Controller 13
Cryptographic Host Bus Adapters 13
Fibre Channel Host Bus Adapters 14
SCSI Host Bus Adapters 14
Notes on Memory ordering: 26
Memory Exception rules requires a “Request for Project Release”: 26
Memory Mirroring 26
Mounting additional Memory with different sizes 27
Memory Interleaving 29
I/O Board 30
M5000 second I/O Board Product ID List 30
I/O-BOX 31
M5000 external I/O-Box Product ID List 33
External storage units for M5000 34
External storage units for HDDs or MTC drives (differential SCSI) 34
DN413 disk node for high performance HDD 34
Hard Disk drives for DN413* 35
Disk subsystems, cable connection diagrams with cable list 35
Connecting Disk node DN413 to SCSI HBAs within single node configurations 35
BG57 for accessible devices 36
Order number for BG57 36
Drives for BG57 36
BG57 for accessible devices, cable connection diagrams with list of cables 36
Connecting BG57 to controller with LVD I/F 36
Service clearance 37
M5000 Connectivity 37
Further information and contacts 38
Recommended 19"- rack PRIMECENTER 2 with 1100 mm depth
24 HU height 38 HU height 46 HU height
+ 2 x 2 HU vertical + 3 x 2 HU vertical + 4 x 2 HU vertical
S26361-K826-V212 S26361-K826-V213 S26361-K826-V214
SPARC® Enterprise M5000 servers can be mounted in PCR2 Rack versions
with 1100 mm depth.
If the mounting kits do NOT fit to the required rack, please order additional kits
separately. For more details refer to the Product Facts: PrimeCenter RackDesign Guide and Configuration PrimeCenter Rack.
To integrate a server or I/O-Box into a Rack by Factory additional Order
numbers are mandatory (see chapter “Rack integration order numbers”)!
Safety Attention:
To mount an M5000 server into a rack you have to take notice for the rules
which are described in the “Installation Guide” for Form factor models.
M5000 System Base Unit is only allowed to be mounted up to Rack height
unit 27 (lowest system height unit)
Up to 3 x M5000 serves in a common Rack are allowed!
Primecenter_24 Primecenter_38 Primecenter_46
Installation of PRIMEPOWER servers and SPARC® Enterprise Servers in other 19" racks than PCR2 is supported only, if „the
PRIMEPOWER - Rack Installation Guidelines for PRIMEPOWER systems“ are fulfilled.
For more information see the FSC extranet at:
Removal media DVD (Standard); DAT (Option)
On-board GbE port 1000Base-T 2 4 (2 on each IOU)
On-board Serial ----External IO exp. Yes
Max. Number of IO-Boxes 2 Boxes (each max 12 PCI slots)
Max. Number of PCI slots
(internal plus I/O-Boxes
Redundant parts HDD*, PSU, FAN, PCI-Cards *, DIMMs**
Hot-replacable parts (Hot swap) HDD, PSU, FAN, PCI-cards, I/O-Boxes, DAT
Dynamic degradation CPU chips, memory, power supply units, fans, HDDs,
RAS Management XSCF
RAS features All major components can be automatically deconfigured after rebooting,
XSCF - Interfaces 2 LAN interface 100Base-TX (RJ45)
4 x 2.15GHz SPARC®64VI
(2 per module)
5 MB
2 CPU Modules
(each 2 CPU chips on CPU Module)
4 Memory Boards
(each 8 slots DDR-II SDRAM)
(8 DIMMs per unit)
1
1 PCI-X (133MHz)
mounted to 1 IOU,
occupies 4 internal PCI-E slots
25 50
(*Software support & consideration on configuration are required;
**Memory Mirror function)
PCI cards (only in multi-path configuration)
Deconfiguration minimizes system impact in case of an uncorrectable error
Finer granularity to prevent propagation of errors
Buses between ASICs are protected by ECC
Memory mirroring (optional)
1 serial port RS232 (RJ45)
1 RCI interface, Fujitsu/FSC proprietary interface
1 USB port (for service only)
SPARC Enterprise M5000
Basic Config up to 8x CPU (up to 16 core) / 256 GB MEM / 4x Hard disk /
2x PCI-X and 8x PCI- expr ess slot
Up to 8 CPU, 4 Module (SPARC64 VI). Up to 16 core (2 Threads
per core) 2.15GHz / 5MB SLC for 1 CPU (2 cores)
CPU order#Type
SEL[XY]1A1U2 CPU, 2.15GHz on a Processor Module
CPU ModuleCPU ModuleCPU ModuleCPU Module
24U, 38U, 46U
CPUCPUCPUCPU
CPU ModuleCPU ModuleCPU ModuleCPU Module
CPU
Main Memory options: up to 256 GB within 4GB DIMMs
MEM order#Type
SEL[XY]2A1U8 x 1 GB DIMM
SEL[XY]2B1U8 x 2 GB DIMM
SEL[XY]2C1U8 x 4 GB DIMM
MEMMEMMEMMEMMEMMEMMEMMEM
MEM MEM MEM MEM MEM MEM MEM MEM
MEMMEMMEMMEMMEMMEMMEMMEM
MEM MEM MEM MEM MEM MEM MEM MEM
MEMMEMMEMMEMMEMMEMMEMMEM
MEM MEM MEM MEM MEM MEM MEM MEM
MEMMEMMEMMEMMEMMEMMEMMEM
MEM MEM MEM MEM MEM MEM MEM MEM
MEMMEMMEMMEMMEMMEMMEMMEM
MEM MEM MEM MEM MEM MEM MEM MEM
MEMMEMMEMMEMMEMMEMMEMMEM
MEM MEM MEM MEM MEM MEM MEM MEM
MEMMEMMEMMEMMEMMEMMEMMEM
MEM MEM MEM MEM MEM MEM MEM MEM
MEMMEMMEMMEMMEMMEMMEMMEM
MEM MEM MEM MEM MEM MEM MEM MEM
Internal I/O Board (standard) :
1 internal DVD (standard)
1 internal DAT (optional)
2 internal disk bays (standard),
Addition al drives need to order the following HDD’ s.
SEP[XY]X3A11U 1 x 73GB 2,5” SAS 10k rpm
Second (Internal) I/O Board (optional) :
2 internal disk bays (standard),
Addition al drives need to order the following HDD’ s.
SEP[XY]X3A11U 1 x 73GB 2,5” SAS 10k rpm
This pic t ur e is an ove r a l l vie w only .
Not all order numbers are i np l emented in t his picture.
To followthe rules of the configurationdiagram is mandatory!
oror
.
<-> means OR
.
.
.
.
.
.
.
.
.
LAN
Console
Solaris
PC
LAN HUB
oror
U320 SCSI-Contr. PCI-e
LVD 2 portSG(X)PCIE2SCSIU320Z
.
1 G-bit LAN Contr. PCI-e
2 port MM F (optic)
(X)7281A-2
.
.
Fibre Channel 4Gb
SG-(X)PCIE1FC-EM4
.
.
.
.
.
.
.
Fibre Channel 4Gb
SG-(X)PCIE2FC-EM4
.
.
.
.
Fibre Channel 4Gb
SG-(X)PCI1FC-EM4-Z
Fibre Channel 4Gb
SG-(X)PCI2FC-EM4-Z
Serial (RS232)
(X)2156A-2
1 port
2 port
1 port
.
2 port
.
8 port
1 G-bit LAN Contr. PCI-e
2 port UTP (copper)
(X)7280A-2
Crypto Accelerator 6000
PCI-e
(X)6000A
Fibre Channe l 4 G b
1 port
SG-(X)PCIE1FC-QF4
Fibre Channel 4Gb
2 port
SG-(X)PCIE2FC-QF4
Fibre Channel 4Gb
1 port
SG-(X)PCI1FC-QF4
Fibre Channel 4Gb
2 port
SG-(X)PCI2FC-QF4
D:GPDS2-DN413 SCSI –HDD -JBOD
D:GPDS2-DN413 SCSI – HDD - JBOD
UltraSCSII/O module
7 disk bay
7 disk bay
7 disk bay
7 disk bay
D:GPDS2-SV30Redundant PSU
D:GPDS 2-SV3 0 Redundant PSU
D:EPRCK-BG57 Peripheral Box 2x 2bay (1,6 inch hight) or
For changeable media drives!
2x 1bay 3,2 inchhight)
UltraSCSI I/O module
UltraSCSII/O module
UltraSCSI I/O module
1,6 inch hight1,6 inch hight1,6 inchhight1,6 inch hight
SPARC® ENTERPRISE M5000 is a midrange server developed using the expertise of Fujitsu and Sun Microsystems. It
has been designed to be the best choice in mission-critical computing solutions. High scalability comes from eight CPU
chips (16 cores) and is backed by exceptional reliability and excellent mea ns for asset investment protection. Based on
SMP (Symmetric Multi Processing) architecture it delivers high scalability without change to applications.
Features
1. SPARC
• Up to eight dual core CPU chips can be installed.
• Up to 256 GB of memory and 50 I/O slots (when four I/O Boxes are connected) are supported for excellent scalability.
• The maximum data transfer performance of 64 GB/s ensures excellent linear scalability up to 16 processor-cores in an
®
ENTERPRISE M5000 is a scalable enterprise server.
SMP configuration.
2. Dual-core SPARC
®
64 VI processors are used in this server.
• 90 nm copper wiring technology is used in the processor to provide higher performance at lo wer power consumption.
• The dual-core processors have large caches (primary cache: instructions 128 KB + data 128 KB per CPU core;
secondary cache: 5 MB per CPU chip) appropriate for enterprise class server systems.
• Multi-core and multi-threading technologies ensure high parallel-processing performance.
• Both the primary and secondary cache use ECC (Error Checking & Correction) for high reliability.
• Registers and ALU are all parity protected. This further enhances the high reliabilit y of the server.
3. Integrated high-reliability makes this server suitable for 24-hour operation.
•All address/data paths between key LSIs, such as CPU chips, memory controllers, I/O controllers and system
controllers, are ECC protected for higher reliability.
•Key components are redundant and hot replacement is supported.
4. Very flexible partitioning functions are provided.
• Multiple independent systems can be built on a single server called partitions or domains.
• Each partition can individually be configured.
• Up to four partitions are supported.
• The motherboard on which CPU- and memory-modules are mounted is sub-divided into eight parts, and the two I/O
board are sub-divided into four. A combination of sub-parts from motherboard and I/O board form a partition, in
maximum 4.
5. Partition independent processor based system control facility: XSCF (eXtended System Control Facility)
•XSCF detects abnormal conditions and controls the system without relying on system status, such as CPU abnormality
or system hang. (Mutual monitoring/abnormal CPU notification, fan rotation frequency monitoring/control, power
monitoring/control, humidity/temperature monitoring, etc., are provided)
•Monitoring/control of peripheral devices such as RCI-connected units (cluster nodes) and expansion cabinets is
available via RCI (Remote Cabinet Interface). Remote power-on/off is also available.
• The operational status of the server can also be monitored by XSCF.
• The embedded partition control facility renders the need for an external control console te rminal unnecessary.
6. Remote Customer Support
• Remote Customer Support will be handled via Teleservice.
7. ServerView Enterprise Edition 2.0 is available for download in the FSC Software Pool.
•ServerView Enterprise Edition 2.0 provides consolidated management of all FSC servers including PRIMEPOWER,
SPARC
®
Enterprise, PRIMEQUEST and PRIMERGY. Server managing functions for all managed servers, such as
hardware configuration display and status monitoring, are provided for system administrators via a unified GUI and
standard operating processes.
a. The status monitoring of all servers is available from a single management client. The status of key components,
such as CPU/memory of each server, can be displayed in a unified tree view.
b. When server components fail, SVS EE displays each event graphically together with the relevant message for
prompt defective part location.
c. Server management can be performed with unified operability.
8. The DR (Dynamic Reconfiguration) functions are supported.
* Areas indicated in blue show components belonging to the base units; those indicat ed in red show options (purchased
separately, but see also base unit description).
10. Hardware for M5000
• Motherboard
The motherboard carries CPU modules and memory modules. Four CPU modules (eight CPU chips) an d eight
memory modules (64 DIMMs) can be mounted. The motherboard is part of the base unit.
st
I/O board (standard)
• 1
The I/O board has a connection to two Hard Disk Drive bays, one DVD Drive bay and one DAT Drive bay. It contains
two onboard LAN
nd
• 2
I/O board (optional)
nd
The 2
I/O board has a connection to two Hard Disk Drive bays. It contains two onboard LAN*2 ports, four PCI Express
slots and one PCI-X slot.
• System disk
Each partition requires at least one system disk for Solaris-installation. System disks can be realized in any of the
following ways. To improve system reliability it is recommended to duplicating (mirroring) the system disk.
A) Mount a Hard Disk Drive i n a disk bay inside the server.
B) Ultra 320 SCSI cards, expansion file units together with external HDDs can also be used as system disks.
C) When using SAN Boot, fibre channel cards, RAID devices and system disks must be prepared. Use a disk in a
RAID device as the system disk. When installing the OS on the RAID device, you need to prepare another
installation server, or use one partition as an installation server using method A) above.
D) LAN Boot from other SPARC Enterprise Server is also possible.
M5000 COD (Capacity on Demand) Server Product ID List
PART NUMBER LONG DESCRIPTION SHORT
DESCRIPTION
M5000family:
SEFRCDB1S FSC SPARC Enterprise M5000 COD (Capacity On Demand) server.
Includes 8x 2.15GHz CPUs (4 CPU boards with 2 x CPUs each, 5MB onchip L2 cache) and 64GB system memory (4 memory modules with 8 x
2GB DDR2 DIMMs), 4 x 73GB SAS hard disks, 1 DVD-ROM,
2x 1Gb Ethernet ports, 2 I/O trays with 8 PCI-E and 2 PCI-X slots total,
4 power supplies (includes redundancy and DPF).
As minimum to each COD System 2 RTU (Right To Use) licenses are
required (must be ordered together with the system)!
M5CoD 2.1GHz
8P64GB 4x73GB
2xI/O tray
M5000 COD (Capacity on Demand) CPU Right to Use (License) Product ID List
PART NUMBER LONG DESCRIPTION SHORT
DESCRIPTION
M5000 family:
SELX9RU1S FSC SPARC Enterprise Server COD 2.0 Right-To-Use (RTU) license for
one Capacity-on-Demand (COD) CPU, for use with SPARC Enterprise
M4000 and M5000 servers.
CoD RTU license for
1 CPU (FF model)
Note to COD:
COD Licenses are automatically arranged by the service processor (XSCF) (see also: Server Admin istration Guide).
Licenses can only manually be assigned to domains, but not to hardware components like CPU.
If licenses are manually assigned to domains, the licenses will be automatically arranged to the CPU’s of this domai n,
first CPU gets first license which is assigned for this domain, second CPU gets second license and so on.
If there is only one big domain, the arrangement is the same: First CPU gets first license which is assigned for this
domain, second CPU gets second license and so on.
Result for M5000: To use the second system-board in one big domain over the complete system a minimum of 5
licenses is necessary!
In addition the “Note to Memory ordering” is also permissible for COD server!
M5000 Internal Storage Device Product ID List
Marketing P/N Option Description
MR 73GB 10K RPM SAS Hard Disk Drive.
SEPX3A11U
SEPY3A11U
SEPX3C11U
SEPY3C11U
SELX9DT1U
SELY9DT1U
Without factory integration!
MR 73GB 10K RPM SAS Hard Disk Drive.
Factory integration only!
MR 146GB 10K RPM SAS Hard Disk Drive.
Without factory integration!
MR 146GB 10K RPM SAS Hard Disk Drive.
Factory integration only!
DAT 72 Tape Drive.
Without factory integration!
DAT 72 Tape Drive.
Factory integration only!
Maximum Number
Supported per System Comments
2nd I/O tray required
for HDD 3 & 4 on
Up to 4 per M5000
Up to 4 per M5000
Up to 4 per M5000
Up to 4 per M5000
Up to 1 per M4000
Up to 1 per M5000
Up to 1 per M4000
Up to 1 per M5000
M5000
nd
2
I/O tray required
for HDD 3 & 4 on
M5000
nd
2
I/O tray required
for HDD 3 & 4 on
M5000
2nd I/O tray required
for HDD 3 & 4 on
M5000
M5000 Filter Product ID List
Maximum Number
Marketing P/N Option Description
SEEX9FL1U M4000 air filter kit 1 per system
SEFX9FL1U M5000 air filter kit 1 per system
To integrate a server or I/O-Box into a Rack by Factory additional Order numbers are mandatory!
Integrate a M4000 or M5000 server into a Rack by Factory RACK-SER2
Integrate a I/O-Box into a Rack by Factory RACK-SER1
Power Cords for PSU connection directly to Rack external Jacks:
Must only be ordered for direct Power connections outside of the Rack to a country specific jack.
Overview Order
Numbers Power Cord:
SELX9P21U MR AC Power Cord (for Continental Europe) 2 4
Description:
It is recommended to order 1 Cord to each PSU:
ENTERPRISE M5000 has 2+2 redundant power supply units (four in total).
HDD#1HDD#3
HDD#0
Power
supply
unit #0
HDD#2
Power
supply
unit #1
DAT
Power
supply
unit #2
DAT
Power
supply
unit #3
If an uninterruptible Power Supply (UPS) is used it shall only
supply one of the two power feeds.
SolarisTM Operating System (OS) and other Software Product ID List
The SPARC
Order Numbers
Media Kit:
SOLZ9-100C9A7M Solaris 10 OS Media kit
U24529-C590A Solaris 10 CD Set Premium Pack
U24529-C592A Solaris 10 Accessories Kit SPARC® Enterprise Server Low End and Midrange
X6099A IPsec enabler CD for Crypto Accelerator 6000
®
Enterprise M5000 servers come with the Solaris 10 Operating System preinstalled on an HDD.
Description: Attention:
Solaris 10 (latest release) DVD Media Kit. Multilingual,
no hardcopy documentation, no license.
Note: Must be ordered with each system if Premium Pack is not ordered!
With additional SW-Components on the control DVD (ServerView Suite, etc.)
This includes:
SOLZ9-100C9A7M Solaris 10 (latest release) DVD Media Kit
U24529-C503A control DVD
Note: Must be ordered with each system if OS Media kit is not ordered!
and ESF CD (Partition DR, SCF interface etc.) U24529-C502A
Note: Can be ordered with Crypto Accelerator!
One of these two
positions must be
ordered for each
system!
Single 4Gb Fiber Channel Network
Adapter, RoHS 6 compliant
8 6
Yes
8 6
Yes
8 6
Yes
8 6
Yes
No
2 ----
No
No
2 ----
No
No
2 ----
No
SG-XPCI1FC-EM4-Z
PCI-X
SG-PCI1FC-EM4-Z
X-option
Fact.-
Integrat.
Single 4Gb Fiber Channel Network
Adapter, RoHS 6 compliant
2 ----
SCSI Host Bus Adapters
Overview Or der
Numbers options:
SG-XPCIE2SCSIU320Z
SG-PCIE2SCSIU320Z
PCI
Bus
type
PCI-e
for Factory
Integration
or X-option
X-option
Fact.-
Integrat.
Description Max
within a
System
Max in
an
I/O-boat
Supported
in I/O-Box
PCI-E Dual channel Ultra320 low
voltage differential SCSI host bus
adapter includes standard and low
6 2
profile brackets. RoHS 6 compliant
Note:
•SCSI Controller SG-(X)PCIE2SCSIU320Z using Solaris 10 Update 3 can only be placed in PCI-express slot #2, #3 or
#4 from both IOUs, not in PCI-express slot #1 nor in I/O-Box which is connected to this slot!
Using Solaris 10 Update 4 or higher versions this SCSI-Controller is also allowed in PCI-express Slot #1, see table
above.
For APL a personal computer (PC) or workstation is required for system control and maintenance work. This console must be
setup prior to installation of APL.
The XSCF is accessed using this console. The console ports of each partition are accessed via XSCF. XSCF and the console
need to be connected via LAN.
If you forget the XSCF-sign on password, you will need to connect to XSCF via RS-232C for recovery.
The customers console terminal should meet the following conditions.
• https (or http) and ssh (or telnet) available (for details, see the following specifications).
• The terminal needs to be placed within 15 m from the server.
• A maintenance account (account for field engineer privileges) for field engineers set up in advance. (For details on
account creation, see “eXtended System Control Facility User’s Guide”.)
Where the customer has no prepared terminal, or the field engineer cannot use the terminal prepare d b y the customer, approval
to use the field engineer’s own PC should be gained from the customer.
Field engineers can connect their own PC to the XSCF via XSCF-LAN or RS-232C. (XSCF-LAN is used as standard. However,
in special cases, if XSCF-LAN is not available or the XSCF-sign on password needs initialization, RS-232C is used.)
Console connection example (field engineer uses a terminal setup b y custom er)
Console connection example (field engineer uses a terminal setup by customer and the console path is duplicated)
Note: For the HUB in the pictures our LAN Switch D:GP70F-AL21 can be ordered.
Console connection example (Field engineer connects temporaril y a PC via LAN (shown as blue line) or RS-232C (shown as
dotted line) as required to provide maintenance)
Note: For the HUB in the picture our LAN Switch D:GP70F-AL21 can be ordered.
Console connection example (the console path is duplicated. Field engineer connects temporarily a PC via LAN (shown as
blue line) or RS-232C (shown as dotted line) as required to provide maintenance)
Note: For the HUB in the picture our LAN Switch D:GP70F-AL21 can be ordered.
UTP category 5 enhanced patch cable for Gbit Ethernet
(1000Base-T) to connect controllers with a hub or switch.
UTP category 5 enhanced patch cable for Gbit Ethernet
(1000Base-T) to connect controllers with a hub or switch.
UTP category 5 enhanced crossover cable for Gbit
Ethernet (1000Base-T) to make direct connection
between Ethernet controllers in a two node configuration
SPARC® ENTERPRISE M5000 supports partitioning in which a single server can be divided into multiple independent systems.
• Partitioning creates multiple independent systems which are called partitions or domains.
• The basic hardware resource required for partitioning is a Physical System Board (hereafter referred to as PSB).
A PSB consists of a motherboard with or without connection to an IOU.
•A PSB can logically be divided into four parts. In this case it is called a Quad-XSB.
A PSB without such a division is called a Uni-XSB.
• The physical unit configuration of each divided PSB part is called eXtended System Board (XSB).
• Partitions can be configured to form a domain using any combination of XSBs. XSBs divided with Uni-XSB and Quad-
XSB can be mixed together within one domain.
• The XSCF is used to configure a partition and specify the PSB division type.
• Up to 4 partitions are supported.
• A minimum of one system hard disk drive per partition is necessary as a boot device.
PSB Division Types:
1) Uni-XSB
The motherboard is divided into two physical system boards (PSB). Each PSB
XSB format. The entire motherboard cannot be used as a single PSB.
Uni-XSB is equivalent to the PSB definition for PRIMEPOWER.
Division Type: Uni-XSB
represents an own logical system in UNI-
The two PSBs are not divided and used as two logical system boards (SB).
CPU chips, memory, PCI slots, hard disk drives assigned to XSB are predetermined.
Mount as a minimum one CPU module and two MEM modules per operating XSB.
If only one partition is configured with both PSBs in Uni-XSB mode but no CPU- and MEM-modules are mounted on
PSB#1, then HDDs and PCI slots on the additional I/O board cannot be used (Areas shaded black in the figure below are
not equipped with CPU- and MEM-modules).
Because Quad-XSB allows to configure several partitions in units of individual CPU chips (two cores) it is the optimized way
to create small partitions.
Quad-XSB is equivalent to the XSB definition for PRIMEPOWER.
Division Type: Quad-XSB
PSB (consists of 1/2 motherboard and an I/O board) is divided into four partitions with 1 CPU chip and a set of memory (8
DIMMs) each. PSBs can be divided into XSBs (XSB#0-0, XSB#0-1, XSB#1-0, XSB#1-1) with connection to an I/O board
and XSBs (XSB#0-2, XSB#0-3, XSB#1-2, XSB#1-3) without connection to an I/O board. CPU chips, memory, PCI slots and
disks are permanently assigned to the XSBs as shown in the diagram above.
XSBs without connection to an I/O cannot be configured as an autonomous partition (refer to “8-4 Notes on
Configurations” for details).
PSBs divided into Quad-XSBs can be mixed together with Uni-XSB to configure a common partition. There is no difference
in settings and conditions for partitions configured with Uni-XSB only or Quad-XSB only. This option is useful if only a
partial memory mirroring within a certain partition is required but Uni-XSB-type XSB is too large to do so.
PAR#0 in the following example is a partition configured by using both Uni-XSB and Quad- XSB.
SPARC® ENTERPRISE M5000 servers use SPARC®64 VI processors, which are multi-core/multi-threading CPUs.
Each module contains two CPU chips (4 cores). Up to four CPU modules can be installed in SPARC
M5000.
Appearance
®
ENTERPRISE
CPU module for M4000/M5000 CPU module without cover viewed from the reverse side
Features
90 nm copper wiring CMOS technology is used for these processors to provide high performance while achieving low
power consumption.
Two CPU cores are built on each CPU chip using CMP (Chip Multi-Processing) technology.
CMT (Chip Multi-Threading) technology provides multi-thread processing that virtualizes each CPU core as two
threads.
The use of register renaming, out-of-order execution and non-blocking cache memory ensures high parallel-processing
performance.
These modules offer mainframe levels of reliability inherited from SPARC
®
64 V.
• • Data in both primary and secondary caches are protected by ECC (Error Checking and Correction).
• • Registers and ALU are protected by parity.
• • If an error occurs, the hardware re-executes the instruction that triggered the error.
(Not all errors are covered by this feature.)
• • If recoverable errors occur frequently in cache, hardware-set degrade mode is automatically invoked.
Type SPARC®64 VI
Frequency 2.15 GHz
Primary cache Instructions: 128 KB
per core
data: 128 KB
Secondary cache Both the instruction and data: 5 MB per CPU (2 cores)
Expansion unit 1 CPU module
(2 CPU chips, 4 cores)
Target servers SPARC® ENTERPRISE M4000
SPARC
®
ENTERPRISE M5000
M5000 CPU Product ID List
Marketing P/N Option Description
Dual-processor Modules
MR 2x 2.1GHz processor module.
SELX1A1U
SELY1A1U
Purchase “Factory-installed modules” if you need the CPU modules installed and shipped inside the server.
“Field expansion modules” are always shipped uninstalled even if they are purch ased together with the server.
SPARC® ENTERPRISE M5000 memory module is an expansion memory device consisting of memory DIMMs and a
memory controller equipped on a board that can be plugged to the motherboard. There are 8 GB, 16 GB, and 32 GB
additional memory module types with eight DIMMs each which can be added as a set. Mounting of additiona l memory
modules must be performed under specific conditions. For details on how to add memory refer to “Conditions for
Mounting Additional Memory”.
Appearance
Memory Module Memory module without a cover
Features
(1) 1 Gbit / 2 Gbit / 4 Gbit DDR2 SDRAM (Double Data Rate Synchronous Dynamic Random Access Memory) is used to
provide high-speed access.
(2) High-density DIMMs (Dual Inline Memory Module) are used to save space on the motherboard.
(3) ECC (Error Checking and Correction) functions provide higher system reliability.
(4) If a single memory chip on a DIMM fails completely, it is handled as a single-bit error. This technique drastically reduces
the number of system shutdowns caused by memory failures.
(5) Memory mirroring function enables continued operation even when an uncorrectable err or (multi-bit error) occurs.
(6) When memory failure is detected at initial diagnosis or during OS operation, Block Degradation or Area D egradation
occurs. Because the area to be degraded is designed to be a controllable minimum area, other areas remain intact.
Thus, operation can continue/resume using these areas. This minimizes the effects of failures.
(7) MEM controller patrols memory during normal operation independently from the CPU and OS. This hardware can detect
an error and degrade the area before either the OS or an application is affected. This procedure doesn’t have any effect
on other operations.
(Memory module: 8 DIMMs are always mounted)
M5000 Memory Product ID List
Maximum Number
Marketing P/N Option Description
8GB MR with server memory board and
8x 1GB DIMMs.
SELX2A1U
SELY2A1U
SELX2B1U
SELY2B1U
SELX2C1U
SELY2C1U
Purchase “Factory integration only” modules if you need the memory modules installed and shipped inside the server.
“Without factory integration” means modules are always shipped uninstal led even if they are ordered with the server.
Without factory integration!
8GB MR with server memory board and
8x 1GB DIMMs .
Factory integration only!
16GB MR with server memory board and
8x 2GB DIMMs.
Without factory integration!
16GB MR with server memory board and
8x 2GB DIMMs.
Factory integration only!
32GB MR with server memory board and
8x 4GB DIMMs.
Without factory integration!
32GB MR with server memory board and
8x 4GB DIMMs.
Factory integration only!
• Two sets of memory modules with the same size (16 DIMMs) must be mounted on a physical system board as a
minimum. 8 DIMMS are always handled as one set.
• When adding memory, two sets of memory modules with the same size must be added at the same time. Memory can
only be added in increments of 2x eight DIMMs. Only 2, 4, 6 or 8 Memory Modules are allowed.
• For mixing DIMMs with different sizes special conditions have to be notic ed, see chapter ”
Memory with different sizes
• If Memory Modules are installed on the second physical System Board a CPU-module on this PSB is mandatory and vice
versa.
• If a second internal I/O-tray is installed, at least one CPU-module and two Memory modules on the second physical
System Board are mandatory (see also Memory Exception rules).
• Min of 1 memory module per processor in Quad XSB Mode
Hint: When ordering 4Gb DIMMs the customer has to specify how FSC shall deal with the memory which comes with the base
configuration. It can either be delivered in a separate package or placed in MEM slots which remain open.
”.
Mounting additional
Memory Exception rules requires a “Request for Project Release”:
• To keep the initial costs low it is allowed as an exception to run a physical system board with only one MEM-module.
This is allowed for PSB#0 or PSB#1 or both of them. This is meaningful in cases where a basic server configuration with
only two MEM modules is ordered together with an additional 2
move a CPU module and a MEM module from the first PSB to the second PSB.
• PSBs with only one MEM module can only be configured in Uni-XSB mode , but not in Quad-XSB mode.
• If you run a PSB with only one MEM module you have to be a ware of the risk in case:
- Of uncorrectable memory errors. In such a case the partition will panic and will not be bootable again,
until the memory fault is solved.
- If reconfigured to quad-XSB the CPU module, which is placed in a XSB without memory, will be
disabled.
nd
I/O-tray. To get the 2nd I/O tray working you have to
Memory Mirroring
SPARC® ENTERPRISE M5000 models provide hardware based memory mirroring to avert system downs even when multi-bit
errors are detected.
Features:
• Memory mirroring is available in both Uni-XSB and Qua d-XSB configurations. In Quad-XSB configuration, however,
mirroring should be configured in units of four XSBs (the same size as with Uni-XSB configuration). Mirroring cannot be
configured using a single XSB.
• Mixed configurations of mirrored and unmirrored SB/XSB i n one p artition are supported (on M5000).
• Memory mirroring configuration is performed from the XSCF. For more information, refer to “eXtended System Control
Facility User’s Guide”.
• When mirroring is configured, the available memory capacit y on each SB/XSB o ne-half of the total size.
• Mirroring is done within a group using combinations as shown below. Thus, the memory size in Groups A and B can be
To mix different DIMM types on a certain SB see chapter “Mounting additional Memory with different sizes”.
Condition 2:
Two sets of memory modules with the same size (16 DIMMs) must be mounted. When adding memory, two sets of
memory modules with the same size must be added at the same time.
Condition 3:
When mounting DIMMs to a specific group, the memory size must be identical within this group. DIMMs of different
memory sizes cannot be mounted within a group.
Condition 4:
Memory size in Group A must be greater than or equal to that in Group B.
These rules are true for mirrored and not mirrored memory configurations.
When mounting additional memory with different memory sizes transfer already-mounted DIMMs as necessary and
follow the above conditions.
Memory interleaving is a technique that divides main memory into several blocks (WAYs) and accesses each WAY
independently and in parallel to enable effective access to the contiguous ar ea.
The size of one WAY is 64 bytes, which is the line size of cache memory.
SPARC
WAYs in each configuration is as follows.
Uni-XSB configuration (16 DIMMs of memory are mounted in Groups A and B)
®
ENTERPRISE M4000 / M5000 perform memory interleaving in Group units as standard. The number of
Uni-XSB configuration (8 DIMMs of memory are mounted in Groups A and B)
Maximum number of HDDs Connection to 2 HDDs
Internal optical drive
Internal tape device
*2
*2
Connection to 1x DVD drive
(DVD-ROM: 8X, CD-ROM: 24X)
Connection to 1x DAT72
M5000 second I/O Board Product ID List
Maximum Number
Marketing P/N Option Description
MR I/O Tray (4 PCIe and 1 PCI-X slots) .
SEFX61U
Without factory integration!
Supported per System
Up to 1 per M5000
Comments
MR I/O Tray (4 PCIe and 1 PCI-X slots) .
SEFY61U
Factory integration only!
Up to 1 per M5000
Note for chapter I/O Board
*1: Notes on LAN connection
LAN ports on the I/O Board and on LAN PCI Express cards belonging to the same partition have the same MAC
address by default. Therefore, do not connect them to the same subnet before you have changed the d efault MAC
address. You can assign unique MAC addresses to respective LAN ports by changing the OBP environ ment variable
“local-mac-address?” to “true”.
*2: DVD-ROM drive and the internal DAT drive cannot be connected to the expansion (second) I/O board of the SPARC
The External I/O Expansion Unit (hereafter referred to as I/O Box) is a rack-mountable expansion unit used to expand the
server’s I/O connectivity with up to 12 PCI Express or PCI-X slots. Each I/O-box can contain two External I/O kits (hereafter
referred to as I/O boats) (giving 12 slots in total). Each I/O boat provides either six PCI Express or six PCI-X slots.
To connect to the I/O boat insert a Link Card into a PCI Express slot on the server and connect them with a cable kit.
Up to four I/O Boxes can be connected resulting in 48 external PCI slots for a M5000
Appearance (provisional):
Features
• Each I/O Box comes with one I/O boat for six PCI Express or six PCI-X slots as standard. By adding a second PCI
Express or PCI-X I/O boat the I/O Box can provide a maximum of 12 slots.
• There are two kinds of expansion I/O boats; PCI Express and PCI-X. Mixed I/O boat configuration is supported.
• Both PCI Express and PCI-X slot types support PCI HotPlug for I/O cards.
• Hot replacement and hot system expansion for each I/O boat is also supported.
• Each I/O Box has 1+1 redundant power supply units and supports single or dual power feed.
• Hot replacement for each power supply unit is supported.
• The I/O Box can be mounted in a 19-inch rack (4U).
Specifications
Items Specifications Remarks
In base
number
Redundant components Power supply units, power system
Hot replaceable
components
Target servers M4000/ M5000
IO boat PCI Slot Specifications
I/O boat for PCI Express
PCI Bus Slot No. No. of Lanes Card Size
LINK0 for exclusive use with a Link Card
PCI Express
I/O boat for PCI-X
PCI Bus Slot No. Bus Width Clock Rate Input Voltage Card Size
PCI-X
system
Maximum
number
PCIE1 8 L
PCIE2 8 L
PCIE3 8 L
PCIE4 8 L
PCIE5 8 L
PCIE6 8 L
LINK0
PCIX1 64/32 bit 133 MHz 3.3 V/Universal L
PCIX2 64/32 bit 133 MHz 3.3 V/Universal L
PCIX3 64/32 bit 133 MHz 3.3 V/Universal L
PCIX4 64/32 bit 133 MHz 3.3 V/Universal L
PCIX5 64/32 bit 133 MHz 3.3 V/Universal L
PCIX6 64/32 bit 133 MHz 3.3 V/Universal L
PCI Bus: indic ates the types of I/O boat.
LINK0: indicates slots where Link Cards, separately purchased PCI cards for connecting the server and the I/O Box,
are mounted.
No. of Lanes: indicates the width of PCI Express slot lane and shows eight lanes are supported.
Bus Width: shows PCI slots support a 64-bit width, and any PCI card with a 64-bit or 32-bit width can be used.
Clock Rate: indicates opera t ing frequencies supported by each PCI slot. When a 133 MHz operable PCI card is installed,
the clock rate becomes 133 MHz.
Input Voltage: indicates the voltage levels of PCI cards that can operate in the slot. 3.3 V PCI cards and universal PCI
cards (PCI cards that can work at either 3.3 V or 5 V) can be mounted in any slot. PCI cards that only work
in 5 V slots cannot be used.
Card Size: S indicates short length PCI cards and L indicates long length PCI cards can be installed.
Connection Types (example)
Connection between server and I/O Box is realized by Fibre or Copper Link Card Kit (Link Card) and a dedicated cable
set (Link cable). Insert the Link Card in a PCI Express slot on the server.
Cascade connection of I/O boats is not supported.
Expansion
M5000: Up to 4 external I/O Boxes with 2 boats each are connectable (second internal IOU must be installed for third
and fourth external I/O-Box). Each I/O boat is connected directly to the server and occupies one PCI-express slot in
internal IOU.
PCI-Express External I/O Expansion Unit, w. 2x PSUs, w 1 PCI-E Boat,
ROHS-6
If “Factory integration” will be accomplished or not depends on
SENX8BE1S
SENX8BX1S
SENY8TE1U
SENX8TE1U
SENY8TX1U
SENX8TX1U
SELY8LK1U
SELX8LK1U
SENY8LK2U
SENX8LK2U
SENY8LK3U
SENX8LK3U
SENX8FC1U
SENX8FC2U
Notes:
- To connect to a server each I/O boat requires its own Link Card Kit. The Link Card Kits are not included in the basic I/O
Box configuration and must be purchased separately.
- It is not supported to connect the two IO boats within one I/O-Box to different servers. If two I/O boats are equipped they
have to be connected to the same server.
order sequencing!
PCI-X External I/O Expansion Unit, w. 2x PSUs, w 1 PCI-X Boat, ROHS-6
Without factory integration!
If “Factory integration”
order sequencing!
Current Status: For order of PCI-X I/O-Box a Request for Special
Release is necessary!
PCI-Express IO-Boat for External I/O Expansion Unit, ROHS-6
Factory integration only!
PCI-Express IO-Boat for External I/O Expansion Unit, ROHS-6
Without factory integration!
PCI-X IO-Boat for External I/O Expansion Unit, ROHS-6
Factory integration only!
Current Status: For order of PCI-X I/O-Boat a Request for Special
Release is necessary!
PCI-X IO-Boat for External I/O Expansion Unit, ROHS-6
Without factory integration!
Current Status: For order of PCI-X I/O-Boat a Request for Special
Release is necessary!
1x copper Link card Kit (2 boards + 2 x 4M copper cables) for External I/O
Expansion Unit.
Factory integration only!
1x copper Link card Kit (2 boards + 2 x 4M copper cables) for External I/O
Expansion Unit.
Without factory integration!
1x Optical Link card Kit (2 boards + 2 10M optical cables) for External I/O
Expansion Unit.
Factory integration only!
1x Optical Link card Kit (2 boards + 2 10M optical cables) for External I/O
Expansion Unit.
Without factory integration!
1x Optical Link card Kit (2 boards + 2 x 25M optical cables) for External
I/O Expansion Unit.
Factory integration only!
1x Optical Link card Kit (2 boards + 2 x 25M optical cables) for External
I/O Expansion Unit.
Without factory integration!
1x Optical Link Cable Set (10m optical cable) for External I/O Expansion
Unit.
Without factory integration!
1x Optical Link Cable Set (25m optical cable) for External I/O Expansion
Unit.
Without factory integration!
will be accomplished or not depends on
FSC Ext I/O Exp
Unit, 1 PCI-E
FSC Ext I/O Exp
Unit, 1 PCI-X
currently without
Controller support!
PCI-E IO-Boat for
I/O Exp Unit
PCI-E IO-Boat for
I/O Exp Unit
PCI-X IO-Boat for
I/O Exp Unit
External storage units for HDDs or MTC drives (differential SCSI)
DN413 disk node for high performance HDD
Overall view disk nodes (new generation)
Disk node DN413
Drive interface U320 SCSI Wide (LVD)
Node interface U320 SCSI Wide (LVD)
Capacity per disk 73 GByte, 146 GByte, 300 GByte
Maximum capacity / disk node 2x (7x300) = 4.2 TByte
Cabinet type 19" rack-mount (3 HU)
Redundant components Fan (std), PSU/DPF (option)
Controller to connect (HBA) SG-(X)PCIE2SCSIU320Z
Throughput 2x 320 MByte/s
Maximum cable length 10 m
Max. number of DN413 per dual
port HBA
RAID level 0, 1 by SW
Cluster NO
Disk nodes DN413 are for single server nodes. For the order numbers see table below.
Order number for disk node DN4x
Disk Subsystem for single node configuration
2 SCSI strings with 7 HDD bays each for disks with SCSI U320 LVD I/F
D:GPDS2-DN413 Disk Subsystem
D:GPDS2-SV30 Power supply unit
NOTES on electrical connection and installation of the box
The " Disk Node " (DN413) can be installed in the 19-inch rack PRIMECENTER.
The plug (IEC320) of the AC power cord, which is shipped with GPDS2-DN413, matches with the outlets of the multiple
socket outlets in the 19" rack.
The GPDS2-DN413 has two AC power cords if a redundant PSU is installed. Note that the boxes do support dual AC power
input/feed when a redundant PSU is installed.
To support dual AC power input/feed the AC power cords (of standard PSU and redundant PSU) must be plugged into
independent power grids (circuits) within the building's wiring.
When connecting the cables from standard PSU and redundant PSU to the same power grid (circuit) within the build ing's
wiring, you will get redundant power supply functionality instead of dual A C power supply functionality.
As long as standard PSU and redundant PSU are working accurately, they share the output load. The outputs of all PSUs
are connected in parallel.
NOTES on ordering
The (basic) disk node for 19" rack (GPDS2- DN413) does NOT include HDDs. HDDs must be ordered as separate items.
To configure DN413 with redundant power supply or support dual AC power input/feed, a redundant PSU (GPDS2-SV30)
must be ordered.
The maximum Disks are 7 per SCSI string and 14 per Disk Node Box GPDS2-DN413.
Standard cable carrier and support brackets for 19” racks PCR2 (BG5x)
All basic boxes for DN413 include support brackets.
1 cable carrier (ZB527) for mounting within 19“ rack BG5x belongs to all DN413
The diagnostic is done via SCSI in band SAF-TE (SCSI Access Fault-Tolerant Enclosure).
Daisy chaining of disk nodes/ packages DN413) or daisy chaining of the 2 strings of one disk no de is not supported.
The disk node DN413 comes without SCSI cable. 1 SCSI cable has to be ordered for each connection from a SCSI string (=7
HDD bays) to a SCSI controller within a node as separate items. Order numbers see below chapter drawing.
You may connect the 2 strings of a disk node DN413 to different compute nodes. The DN413 is switched ON/OFF by
"terminator power“ from the SCSI controller, i.e. ON with ON of the first compute node and OFF with OFF of the last compute
node. If the temperature inside the DN413 reaches an abnormal level (because of a fatale error), all connected compute
nodes shut down and switch off the disk node.
2 SCSI connection module single port U320
rack-mountable box, needs 3U in a 19“-rack
1 cable carrier (ZB527) and 1 support bracket (ZB5211) for 19“-rack
BG5x are included
Power supply unit, providing redundant PSU configuration for
D:GPDS2-DN413.
The redundant PSU adds dual AC power input/feed to GPDS2- DN413
D:GPDS2-FP81Hard disk 73GB 15kRpm
D:GPDS2-FP84 Hard disk U320 10k 146 GB 146 GByte 10,000rpm, Ultra SCSI LVD320
D:GPDS2-FP85 Hard disk U320 10k 300 GB 300 GByte 10,000rpm, Ultra SCSI LVD320
D:GPDS2-HDP3 HD bundle with 4 x 73 GB 15Krpm292 GByte of disk space, 4 drives 73 GByte each
D:GPDS2-HDP4 HD packet with 4 x 146 GB U320 10Krpm
D:GPDS2-FP86 Hard disk 73GB 15kRpm
D:GPDS2-HDP5 HD bundle with 4 x 73 GB 15Krpm
NOTES on disks for DN413:
These HDDs can be mounted in „DN413 boxes“ only, but NOT in system cabinets.
Disk subsystems, cable connection diagrams with cable list
Connecting Disk node DN413 to SCSI HBAs within single node configurations
(up to 320 MByte/s)
DN413 is not usable for shared disks in cluster configurations.
73 GByte 15,000rpm, Ultra SCSI LVD320
Fast drives with short access time
584 GByte of disk space, 4 drives 146 GByte each
llowed in DN413.
73 GByte 15,000rpm, Ultra SCSI LVD320
Fast drives with short access time
292 GByte of disk space, 4 drives 73 GByte and 15,000rpm
each
D:GPDS2-DN413
D:GPDS2 -DN413
Disk Subsystem Dual Port
Disk Subsystem Dual Port
UltraSCSI
UltraSCSI
I/O module
I/O module
7 disk bay
7 disk bay
UltraSCSI
UltraSCSI
I/O module
I/O module
7 disk bay
7 disk bay
D:GPDS2-SV30
D:GPDS2 - SV30
The second UltraSCSI I/O module in this drawing can either be connected to the second port of a DualPort HBA or to another
HBA which may or may not belong to another partition.
Cables for connecting disk nodes -DN413 to U320 SCSI HBA (up to 320 MByte/s)
Rack-mountable box, 3 HU, 4 bays (1.6” high)
2 SCSI strings (2 bays each) as standard
D:EPRCK-BG57
D:EPRCK-AN60 Upgrade kit PW BG57 to 4channel U160
D:EPRCK-AN62 Change kit PW BG57 to 1channel U160 For connecting the 4 bays of one BG57 to 1 SCSI string
D:EPBOX-AN55 Adoption set for 8bit SE devices
19" Rack Box f. changeable Peripheral
2channel U160
4 SCSI strings (1 bay each) with optional EPRCK-AN60
1 SCSI strings (4 bays) with optional EPRCK-AN62
Each SCSI string may be connected to a different server.
Note: NO redundant PSU available.
For configurations with 4 SCSI strings with 1 bay each or 3
strings with 2x1 plus 1x2 bays.
1 adapter AN55 required for each 8-bit device within a
BG55/BG57 box
Drives for BG57
36 GByte (native) to 72 GByte (compressed)
16bit
D:EPBOX-MC57 MTC DAT72, 36-72GB, box
D:EPBOX-MC77 MTC 8mm, 80-160GB
D:EPBOX-MC88 MTC LTO 2 HH, 200/400GB
Install in EPRCK
LVD
Do NOT mix with 8bit drive within 1 box
Current Control²CD required
80 GByte (native) to 160 GByte (compressed)
16bit
Install in EPRCK
LVD
Do NOT mix with 8bit drive within 1 box
Current Control²CD required
200 GByte (native) to 400 GByte (compressed)
16bit
Install in EPRCK
LVD
Do NOT mix with 8bit drive within 1 box
-BG57 rack mountable box only
-BG57 rack mountable box only
-BG57 rack mountable box only
NOTES on BG57 MTC drives:
T hese devices can be mounted in BG57 boxes only, NOT in system cabinets.
BG57 for accessible devices, cable connection diagrams with list of cables
Cables for connecting BG57 boxes to SCSI controller (up to 80MByte/s, device dependent)
D:KB235-M3
D:KB235-M5
D
D:KB235-M8
D:KB243-M10 SCSI Connection Cable, 10m
Cables to configure the BG57 box
EPRCK-AN60 Upgrade kit PW BG57 to 4 channel U160
α
EPRCK-AN62 Change kit PW BG57 to 1 channel U160
β
SCSI cable 3m (68pin UHD and 68pin
mplimite)
SCSI cable 5m (68pin UHD and 68pin
mplimite)
SCSI cable 8m (68pin UHD and 68pin
mplimite)
These cables can be installed if the boxes contain drives
with
LVD I/F only.
The cable connects 1 or 2 bays in a BG57 to 1 input
connector of the box.
2 cables are included in the kit
1 kit (=2 cables) are included in BG57
This kit is needed when all 4 bays must be connected in
daisy chain to 1 input connector of the box. The
standard cables of the box cannot be used in this case.
Information about UPS (Uninterruptable Power Supplys):
UPSes are only available up to 20kVA for M8000 (Masterguard, rack mounted, single phase). UPSes with more power (or
3phase) are not available yet (at FSC), but the connection/installation is supported by partners like MASTERGUARD.
20kVA APC planned.
All rights, including rights created by patent grant or registration of a utility model or design, are
reserved. Delivery subject to availability; right of technical modifications reserved.
All hardware and software names used are trade names or trademarks of their respective
manufacturers.