ATCA-2210 Switch and Control Reference · NTS API ... This guide describes the ATCA-2210 switch and...

www.radisys.com

PROMENTUM™SWITCH AND CONTROL MODULE REFERENCE

ATCA-2210WITH SOFTWARE VERSION 1.8

007-02376-0001 • September 2007

2

Table 1. Release history

Release Date Description-0000 December 2006 First release.-0001 September 2007 Editorial changes and new features.

Copyright © 2006-2007 by RadiSys Corporation. All rights reserved. RadiSys is a registered trademark and Promentum is a trademark of RadiSys Corporation. AdvancedTCA, ATCA, and PIGMG are registered trademarks of PCI Industrial Computer Manufacturers Group. Linux is a registered trademark of Linus Torvalds. All other trademarks, registered trademarks, service marks, and trade names are the property of their respective owners.

3

TABLE OF CONTENTS

Preface ................................................................................................................................................................ 7About this manual....................................................................................................................................................................... 7Notational conventions .............................................................................................................................................................. 7Electrostatic discharge ................................................................................................................................................................ 8Where to get more product information ............................................................................................................................... 8

Chapter 1: Features Overview ....................................................................................................................... 11Local management processor .................................................................................................................................................11Ethernet switches........................................................................................................................................................................11Network timing subsystem...................................................................................................................................................... 12Intelligent Platform Management Controller....................................................................................................................... 12Installed software....................................................................................................................................................................... 12External connectivity ................................................................................................................................................................. 12

Chapter 2: Theory of Operation .................................................................................................................... 13Possible applications of the SCM........................................................................................................................................... 13Overview of functionality......................................................................................................................................................... 13

Chapter 3: Subsystems ................................................................................................................................... 19Local management processor ................................................................................................................................................ 20Clock synthesizer subsystem .................................................................................................................................................. 23Base Ethernet switch subsystem............................................................................................................................................ 24Fabric Ethernet switch subsystem.......................................................................................................................................... 28LED controller interface subsystem....................................................................................................................................... 31Network timing subsystem...................................................................................................................................................... 36Rear transition module............................................................................................................................................................. 38Hardware management subsystem ...................................................................................................................................... 42Power supplies........................................................................................................................................................................... 44

Chapter 4: Physical Interfaces........................................................................................................................ 47Mechanical specification .......................................................................................................................................................... 47Front panel interfaces............................................................................................................................................................... 48

4

Table of Contents

Backplane interfaces ................................................................................................................................................................. 52Internal interfaces......................................................................................................................................................................58

Chapter 5: Initial Verification and Configuration ........................................................................................ 61Configuration checklist ............................................................................................................................................................. 61Connecting to the SCM............................................................................................................................................................ 62Logging in to the SCM ............................................................................................................................................................. 62Adding user accounts and setting passwords..................................................................................................................... 63Rebooting the SCM................................................................................................................................................................... 63Configuring IP addresses ......................................................................................................................................................... 63Per-VLAN interfaces ..................................................................................................................................................................66Configuring SNMP trap destinations and security access ................................................................................................ 67Enabling the SNMP agent .......................................................................................................................................................68Enabling the DHCP server.......................................................................................................................................................69Enabling and disabling the Telnet server and TFTP server .............................................................................................. 70Enabling and disabling syslog and other services ............................................................................................................. 71Copying a configuration from existing CLI configuration files ........................................................................................ 72Copying CLI configuration files to a TFTP server ............................................................................................................... 72Copying any file to the SCM ................................................................................................................................................... 73

Chapter 6: Software Features ........................................................................................................................ 75Shelf Manager ............................................................................................................................................................................ 76Ethernet switching ..................................................................................................................................................................... 76IP routing..................................................................................................................................................................................... 76Linux and boot loader .............................................................................................................................................................. 77Flash memory usage ................................................................................................................................................................ 79RAM and file system usage..................................................................................................................................................... 81IPv4 interface and Ethernet switch port identifiers ............................................................................................................84Pre-defined IP addressing scheme........................................................................................................................................86DHCP server ...............................................................................................................................................................................90Telnet and SSH...........................................................................................................................................................................90Network timing subsystem (NTS) software.........................................................................................................................90NTP server................................................................................................................................................................................... 93Module state management ..................................................................................................................................................... 93Flash and file system management....................................................................................................................................... 93Software image management ................................................................................................................................................ 93Firmware and software upgrades ..........................................................................................................................................94

Table of Contents

5

Logging service ..........................................................................................................................................................................94E-Key authorization and notification .....................................................................................................................................94

Chapter 7: SNMP Agent Support .................................................................................................................. 95Base or Fabric switch selection............................................................................................................................................... 95ifIndex values for IP interfaces................................................................................................................................................ 95MIB module support ................................................................................................................................................................96Notification support ..................................................................................................................................................................98MIB modules supplied.............................................................................................................................................................98MIB browser utility applications.............................................................................................................................................99

Chapter 8: API Overview ...............................................................................................................................101Where to find related files and additional information ...................................................................................................101Ethernet switching API.............................................................................................................................................................101NTS API...................................................................................................................................................................................... 102Module state management API............................................................................................................................................ 102Flash and file system management API ............................................................................................................................. 103Software image management API ....................................................................................................................................... 103Management plane framework (MPF) API ....................................................................................................................... 104

Chapter 9: Using the Command Line Interface......................................................................................... 105Accessing the master CLI....................................................................................................................................................... 105Displaying available commands and options ................................................................................................................... 105Master CLI command modes ............................................................................................................................................... 106“No” commands...................................................................................................................................................................... 106“Show” commands ................................................................................................................................................................. 107Command line completion ................................................................................................................................................... 107Editing and history keys ......................................................................................................................................................... 107Saving configurations persistently ....................................................................................................................................... 107Master CLI commands ........................................................................................................................................................... 108

Chapter 10: Blade Management Commands .............................................................................................111Accessing blade-management command modes.............................................................................................................111SNMP commands ....................................................................................................................................................................111Network timing subsystem (NTS) commands ..................................................................................................................112

Chapter 11: Maintenance ............................................................................................................................. 125Installing the SCM ................................................................................................................................................................... 125

6

Table of Contents

Removing the SCM ................................................................................................................................................................ 125Troubleshooting....................................................................................................................................................................... 126

Appendix A: Specifications ............................................................................................................................131Environmental .......................................................................................................................................................................... 131Safety.......................................................................................................................................................................................... 132Mechanical dimensions ......................................................................................................................................................... 132Electromagnetic compatibility (EMC) ................................................................................................................................. 133Power ........................................................................................................................................................................................ 134Reliability ................................................................................................................................................................................... 134

Appendix B: IPMI commands and Managed Sensors .............................................................................. 135IPMI commands ...................................................................................................................................................................... 135Managed sensors .................................................................................................................................................................... 137Board health and event generation .................................................................................................................................... 140Sensor alarms troubleshooting ............................................................................................................................................ 143

7

PREFACE

About this manualThis manual describes the Promentum™ ATCA-2210, a switch and control module (SCM), which is fully compliant with AdvancedTCA® (ATCA®). The SCM is designed to be incorporated into High Availability (HA) systems such as the Promentum platforms SYS-6006 and SYS-6010. Use this manual as a reference for the operation and the maintenance of the ATCA-2210 SCM. The material presented here is not introductory; it is assumed that you are already familiar with the intended use of the ATCA-2210 SCM in your organization’s ATCA platform. For instructions on initial setup of the ATCA-2210 SCM, refer to the ATCA-2210 Switch and Control Module Installation Guide. The simplified names “SCM” and “module” will be used in place of “ATCA-2210 SCM” for the rest of the manual.

Notational conventionsThis manual uses the following conventions

All numbers are decimal unless otherwise stated.

BoldText A keyword.

ItalicText File, function, and utility names.

MonoText Screen text and syntax strings.

BoldMonoText A command to enter.

ItalicMonoText Variable parameters.

Brackets [ ] Command options.

Curly braces { } A grouped list of parameters.

Vertical line | An “OR” in the syntax. Indicates a choice of parameters.

Preface

8

Electrostatic discharge WARNING! This product contains static-sensitive components and should be handled with care. Failure to employ adequate anti-static measures can cause irreparable damage to components.

Electrostatic discharge (ESD) damage can result in partial or complete device failure, performance degradation, or reduced operating life. To avoid ESD damage, the following precautions are strongly recommended.

Keep each module/PCB in its ESD shielding bag until you are ready to install it.

Before touching a module, attach an ESD wrist strap to your wrist and connect its other end to a known ground.

Handle the module only in an area that has its working surfaces, floor coverings, and chairs connected to a known ground.

Hold modules only by their edges and mounting hardware. Avoid touching PCB components and connector pins.

For further information on ESD, visit www.esda.org.

Where to get more product information Please visit the RadiSys Web site at www.radisys.com for product information and other resources. Downloads (manuals, release notes, software, etc.) are available via the Technical Support Library product links at www.radisys.com/support or the product pages at www.radisys.com/products.

About related RadiSys ProductsThe SCM is part of the Promentum SYS-6006 and SYS-6010 platforms. For information on the Promentum product family and other RadiSys products, see the RadiSys Web site at www.radisys.com.

Standards informationFor information about the PCI Industrial Computer Manufacturers Group (PICMG®) and the AdvancedTCA standard, consult the PICMG Web site at this URL:http://www.picmg.org

http://www.esda.org

http://www.radisys.com

http://www.radisys.com/support

http://www.radisys.com/products


http://www.picmg.org

Where to get more product information

9

Related DocumentsCAN/CSA 22.2 #60950-1-03 Safety for Information Technology Equipment, CSA

EN 60950-1:2002 Safety for Information Technology Equipment, CENELEC

GR-63-CORE NEBS Requirements Physical Protection, Issue 2, April 2002, Telcordia

GR-78-CORE Generic Requirements for the Physical Design and Manufacture of Telecommunications Products and Equipment, Issue 1, September 1997, Telcordia

GR-1089-CORE Electromagnetic Compatibility and Electrical Safety – Generic Criteria for Network Telecommunications Equipment, Issue 3, October 2002, Telcordia

IEC 60950-1 Safety for Information Technology Equipment, IEC

INF-8077i 10 Gigabit Small Form Factor Pluggable Module (XFP), Revision 4.5, XFP Multi Source Agreement (MSA) Group

PICMG 3.0 R2.0 Advanced Telecommunications Computing Architecture (AdvancedTCA), PICMG

PICMG 3.1 R1.0 Ethernet/Fibre Channel over PICMG 3.0, PICMG

PICMG COM.0 R1.0 COM Express Module Base Specification, PICMG

Small Form-factor Pluggable (SFP) Transceiver MultiSource Agreement (MSA), Cooperation Agreement for Small Form-Factor Pluggable Transceivers, September 14, 2000, Agilent Technologies, Blaze Network Products, E2O Communications, Inc., ExceLight Communications, Finisar Corporation, Fujikura Technology America Corp., Hitachi Cable, Infineon Technologies Corp., IBM Corp., Lucent Technologies, Molex, Inc., Optical Communication Products, Inc., Picolight, Inc., Stratos Lightwave, Tyco Electronics.

UL 60950-1 Safety for Information Technology Equipment

Preface

10

11

1FEATURES OVERVIEW

This guide describes the ATCA-2210 switch and control module (SCM) and serves as a hardware and software reference for its electrical, mechanical, and environmental aspects. The SCM is fully compliant with the ATCA standard and is designed to be incorporated into high availability (HA) systems, which meet 99.999% availability. The module provides highly integrated, centralized common equipment functions—packet switching, shelf management, network timing, and provides a site for a daughter card processor, which can be used for system management. The SCM has a comprehensive hardware management subsystem, managed by the IPMC, which interfaces to the Shelf Manager and functions as a fully managed building block.The HA infrastructure uses two SCMs that each contain a local management processor (LMP), a 1 Gigabit Base Ethernet switch, a 10 Gigabit Fabric Ethernet switch, a Shelf Manager, a network timing subsystem (NTS), and an Intelligent Platform Management Controller (IPMC). Duplication of the SCM and its supporting components allows for redundancy–a means of providing continuing functionality even if a failure occurs on one of the SCMs.This chapter provides a features overview of the SCM and its components. Detailed information on the theory and the operation of the SCM is covered in the chapters that follow.

Local management processorThe SCM incorporates a powerful on-board PowerQUICC III processor for its local management functions. The local management processor (LMP) is used to manage the Ethernet switches, manage the network timing subsystem (NTS), and to provide access to the hardware management subsystem. The LMP may also function as the Shelf Manager, depending on the software configuration you implement through your application software. See page 20 for an overview of the LMP’s features, and subset of features, as they apply to the implementation of the SCM.

Ethernet switchesThe SCM is has two Ethernet switches:

1 Gigabit Base Ethernet switch — The Base Ethernet consists of a single Broadcom BCM56304 Ethernet switch. See page 24 for more information on the Base Ethernet switch subsystem.

1 Gbps or 10 Gigabit Fabric Ethernet switch — The Fabric Ethernet interface uses a single Broadcom BCM56800 Ethernet switch, which has twenty ports that can be configured for 1Gbps or 10 Gb operation. See page 28 for more information on the Fabric Ethernet switch subsystem.

1 Features Overview

12

Network timing subsystem The network timing subsystem (NTS) consists of a circuit block and software. The circuit block provides a centralized timing source for the shelf based on the AdvancedTCA synchronization clock Interface specification. The NTS selects between several internal and external timing references, and provides a phase-locked, highly stable system clock to the platform modules managed by the SCM’s Shelf Manager and the NTS software. See page 36 for information on the circuit block and page 90 for details about the software aspects of the NTS.

Intelligent Platform Management ControllerThe SCM uses a Renesas H8S/2166 micro-controller for the Intelligent platform management controller (IPMC). This device manages commands and data as part of the hardware management subsystem, which includes the IPMCs and the sensors from installed modules and field replaceable units (FRUs) and their communication with the Shelf Manager. See page 42 for more information on the hardware management system and the IPMC.

Installed softwareThe SCM’s Linux-based software includes many features and several management interfaces, including Ethernet switching protocol support, a PICMG-compliant Shelf Manager, a command line interface (CLI), application programming interfaces (APIs), and an SNMP agent. For a complete overview, see Software Features on page 75.

External connectivityThe SCM provides connectivity for communicating with external equipment through three interfaces:

Front panel interfaces — Front panel interfaces are accessed through each SCM’s faceplate. The connection types are labeled on the faceplate. See page 48 for an illustration of the front panel and a description of all its components, including LEDs and port pinouts.

Backplane interfaces — The backplane is the primary electrical interconnection between the modules in a platform. The Zone 2 connectors of the backplane provide connectivity for the Base, the Fabric, the update channel interfaces, and synchronization clocks. See page 52 for details on the backplane contact assignments and connectors.

Rear transition module interfaces — When installed in a RadiSys shelf, each SCM has an associated rear transition module (RTM) site, which can provide external access to SCM’s signals and connections. See page 38 for descriptions of the RTM’s electrical connections and signal interfaces.

13

2THEORY OF OPERATION

Possible applications of the SCM The SCM is ideal for next generation network elements in third generation (3G) wireless and wire-line infrastructures because it provides highly integrated, centralized common equipment functions—switching, shelf management, and network timing.

Possible applications include the following:

Overview of functionalityThe SCM supports inter-node connectivity by providing Base and Fabric Ethernet switching and system clock synchronization. The SCM can act as a Shelf Manager and also includes a site for a COM Express processing module. The functions on the SCM are managed by the local management processor (LMP) block, based on a PowerQUICC III processor. The LMP has four Ethernet ports, two PCI buses, two serial ports, a local Inter-integrated circuit (I2C) bus, a SPI port, DRAM and flash memory. A 10/100 Ethernet channel from the PowerQUICC III is routed to the front panel to provide an interface for switch management. The memory bus provides interfaces to the DRAM and flash memory. The PCI busses provide the interfaces to the Base Ethernet and the Fabric Ethernet switches. Figure 1 on page 14 shows a block diagram of the SCM.

Carrier grade billing and network serversIPTV switching infrastructureMedia gateway (MGW)Push-to-connect/Push-to-talk (POC/PTT) server

Radio network controller (RNC)SoftswitchesSession border controllers (SBC)Web servers

2 Theory of Operation

14

Figure 1. SCM system block diagram

The SCM as a Shelf ManagerThe SCM is configured to act as the Shelf Manager by default. The shelf management duties are divided between the LMP and the Intelligent Platform Management Controller (IPMC). The IPMC on the SCM is equivalent to a shelf management controller (ShMC). The Shelf Manager is responsible for monitoring and controlling the behavior of the components within the shelf and the ShMC is the device that communicates with the IPMCs on the field replaceable units (FRUs). The high-level aspects of shelf management are handled by the Shelf Manager software, which is described in detail in the Shelf Management Software Reference.

E1/DS1framer

RJ45

RJ45

RJ45

SFP

SFP

SFP

SFP

Zone3

NTS

Zone2

Zone1

ShMCIPMC

SFP

SFP

SFP

SFP

BaseEthernetswitch

XFP

FabricEthernetswitch

XFP

XFP

RJ45

RJ45

USB

RJ45

COM Express * Harddrive

12V

USB

1Gb Eth

FE

10Gb

10Gb

48V

1Gb

100Mb

Clk Clk

Clk

1Gb-Tx

Local powerconversion

Serial

Updchan

1Gb

PCI

Syncchan

Nodeslots

FE

0-4

1Gb

SFP

RTC

FE

XFP

XFP

10Gb Nodeslots

Sync

RS-232

10Gb

RJ45RS-232

1Gb

1Gb

1Gb

-48V powerconversion

3-4

12-14

12-14

1

1

2

2-8

ShMgr

SAS/SATA*

1Gb

10Gb

2

SFP

SFP

1Gb

IPMB

1-4

* Reserved for future implementation.

RTM (Illustration)

RJ45

SFP*

SCM

PCI

Blue = Base EthernetRed = Fabric Ethernet

PowerQUICC III local

management processor

switchFE

2Overview of functionality

15

Intelligent Platform ManagementIntelligent Platform Management (IPM) is a subset of shelf management. IPM handles aspects of shelf management that involve communication between the Shelf Manager and the IPMCs on the FRUs. The Shelf Manager and all the IPMCs communicate through the Intelligent Platform Management Bus (IPMB). This communication uses the Intelligent Platform Management Interface (IPMI) protocol. Figure 2 provides an overview of the IPMB architecture on RadiSys 14-slot shelf.

Figure 2. Example IPMB architecture on a RadiSys 14-slot shelf

I2C

Note:IPMC — Intelligent platform managment controllerLMP — Local management processorPEM — Power entry moduleSCM — Switch and control moduleShMC – Shelf management controllerSPM — Shelf peripheral module

IPMC IPMC

IPMC

IPMC IPMC

IPMC IPMC

IPMC

IPMCIPMC

IPMC

ShMCShMC

ModuleModuleSCMSCMModuleModule

Fan

FanFan

Fan

Shelf display or alarm panel

Sensors

SPM

(Modules insome slotsnot shown)

(Modules insome slotsnot shown)

PEMPEM

LMP LMP


16

The Shelf Manager retrieves FRU information from the IPMCs and stores this information. Possible types of information that can be retrieved from a FRU include information on its capabilities, sensor readings, and hot-swap state. To provide redundancy, the active Shelf Manager also sends the FRU information to the inactive Shelf Manager, which also stores the FRU information. To implement full redundancy, the IPMB is set up as two I2C buses. This dual-bus architecture prevents shelf management from being interrupted by the failure of either bus. Each ShMC and each IPMC connects to both I2C buses, so if one bus becomes unavailable, communication takes place on the other bus. To prevent a failed ShMC or an IPMC from significantly disrupting either or both buses, each ShMC and IPMC has associated circuitry that isolates the ShMC or the IPMC from both buses if its associated watchdog timer fails to be triggered periodically.

Base Ethernet interfaceWhen the SCMs are installed on a platform, like the RadiSys SYS-6010, they provide a Base interface that supports the 10/100/1000Base-T Ethernet standard on the backplane. Each SCM also provides 10/100/1000Base-T Ethernet uplinks on its front panel. The topology of the Base interface forms a dual star. Figure 3 illustrates this topology, showing the relationship between the node and the hub modules.

Figure 3. Dual star topology

Figure 4 on page 17 illustrates the dual star topology in terms of the backplane Base connections between modules on a RadiSys 14-slot shelf.

Hub(SCM)

Hub(SCM)

Node Node Node

Node Node Node

2Overview of functionality

17

Figure 4. Backplane Base connections to modules on a RadiSys 14-slot shelf

Each Ethernet backplane link consists of four differential signal pairs. Each SCM is installed in a hub slot and acts as a switch that connects to every node (non-SCM) module to form the star topology. This redundant topology provides an alternate path if one Ethernet path fails, or if one SCM fails. The SCMs connect to each other through a direct Ethernet connection (Base channel 2), which the high-level shelf-management software uses. Traffic through the SCM-to-SCM link is controlled by the Base Ethernet switch, which can also pass user traffic.For information on the software managing the Base Ethernet interface, see The Ethernet Switching Software Reference.

SCMSCMModule

Channel

2 1

Slot 1 Slot 2

(Modulesnot shown

in some slots)

Slot 7 Slot 8

Hub slotsNode slots Node slots

(Modulesnot shown

in some slots)

Module

Channel

2 1

Module

Channel

2 1

Slot 13 Slot 14

UplinksUplinks

Ethernet interface

Module

Channel

2 1


18

Fabric Ethernet interfaceThe SCM can support either 1 Gbps or 10 Gb Ethernet links for the Fabric interconnect, depending on the node module connections. The full Fabric interface channel is implemented using a 10 Gb XAUI configuration as described in the PICMG 3.1 specification. The Fabric Ethernet interface also implements a dual star topology similar to the topology of the Base Ethernet interface. However, there is no direct switch-to-switch link. For information on the software managing the Fabric Ethernet interface, see The Ethernet Switching Software Reference.

RedundancyTypically, two SCMs are installed on a platform for redundancy purposes. Redundancy provides a means of continuing functionality even if a failure occurs on one of the SCMs:

Each SCM contains switches for both the Base Ethernet and the Fabric Ethernet. If either SCM fails or is hot-swapped, the other SCM switches all the traffic for both the Base and the Fabric interface. The SCMs operate on an active/active basis, which means the switches in both SCMs are always operating.

Both SCMs run the same switch configuration and management software. If one SCM is removed or fails, the copy in the other SCM continues to control the switches in the remaining active SCM.

Each SCM has a Shelf Manager. If one SCM is removed or fails, the Shelf Manager in the other SCM operates as the active Shelf Manager. The dual Shelf Managers operate on an active/standby basis, which means only one Shelf Manager is active at any time.

SCMs act and operate independently from each other and, in most cases, need to be configured separately. However, some shelf management settings do not require a separate configuration. These shelf-specific configuration changes are only performed on the SCM containing the active Shelf Manager. See the Shelf Management Software Reference for information on which shelf management settings require a separate configuration on each SCM and which do not.

Note: The SCMs must be of the same model. For example, an ATCA-2210 SCM is not compatible with an ATCA-2100 SCM.

19

3SUBSYSTEMS

The SCM consists of a series of subsystems that work together to provide the network elements required for third generation wireless and wire-line infrastructures. This chapter covers the details of the subsystems that make up the SCM.

Figure 5. Detailed block diagram of SCM

BITS/SSU clk

MagI/F

RJ -45

RJ -45

SFP

SFP

Zone 3conn

Local powerponversion

NTM

Zone 2conn

Zone 1conn

DC/DCpower

converter

IPMCSFP

SFP

SFP

SFP24-port Gb

+ 2 port10Gb Eth

switch

16-portmag

16-portGb EthPHY

ID

XFP

20-port10Gb EthSwitch

XFP

XFP

RJ -45mag

RJ -45mag

USB

RJ45

232Buf

DDR1SODIMM

BufDual 1 Gb

MACSerDes

Tyco PIM-200power cond

SAS/SATAI2C

SPI

USB

1Gb Eth

10/100-Tx

PCI-E

48V12V

IPMB

12V

ClkInClkInRefClkRefClk

Clk

1Gb-Tx

1Gb SGMII


Serial

14

Updchan 3,4

2

2

3

Updchan 0

Sync I/F

14

1

ShMCBase Ch1

Nodeslots

F-LinkBase Ch2

210/100-Tx

Dual-bootflash

Ser-

4

Buf

XAUI toXFI

XAUI toXFI

XAUI toXFI

10/10PHY

10/10PHY

2

2

2

X

Switch

10Gb XAUI

Buf ID & tempBufI2C

EE -PROMTemp

sensors

EE -PROM

SFP

SFP

Switch12V

FEswitch

COM-EType Det

SPI

MIIMII

232Buf

SIO

LPC

I2C

Flash sel

2

Serial

XFP

XFP XAUI toXFI

XAUI toXFI

10GbXAUI

14

Nodeslots

Sync

1Gb SD

TBI

PCI-32

10 Gb XAUI

1

SCM

4

Fuses

1Gb-Tx Updchan 1,2

PCI-E

3.3VSwitch

Mag

X

SPI

2

2X10 Gb XAUI

1

2

1

1Gb SD

2

SFP

SFP

SFP

1Gb SD

RJ45

43

BITS/SSU clk

BITS/SSU clk

MagI/F

BITS/SSU clk

MagI/F

MagI/F

RJ -45

RJ -45

PowerQUICC IIIlocal

management processor

Des

RTM (Illustration)

SFP*

COMExpress*

SAS/SATAhard drive*


Blue = Base EthernetRed = Fabric Ethernet

10/100-Tx

Clock/calendar

1Gb

10Gb XAUI

3 Subsystems

20

Local management processor The local management processor (LMP) is used to manage the Ethernet switches, to manage the network timing subsystem (NTS), and to provide access to the hardware management subsystem. The LMP provides Ethernet ports, serial communications, an inter-integrated circuit (I2C) controller, a serial peripheral interface (SPI), memory, and peripheral component interconnect (PCI) bus interfaces. A 1 Gigabit Ethernet port provides a high-speed interface to the Base switch. A 10/100 Mb Ethernet channel from the LMP is routed to the front panel and provides an interface for switch management. A 10/100 Mb Ethernet channel is routed across the backplane to the redundant SCM. The I2C controller and the SPI buses control on-board peripheral devices. The memory buses provide interfaces to the local synchronous dynamic random access memory (SDRAM), and the boot flash memories. The PCI busses connect to the management ports of the Base and the Fabric Ethernet switches.The LMP includes the following components:

The block diagram in Figure 6 shows the relationship between the LMP and the SCM.

A PowerQUICC III processor chip with 833 MHz core frequencyA 200-pin SODIMM memory connector for connecting to a memory module with a capacity of 1Gb DDR SDRAMA redundant pair of 64 MB Spansion flash memory components (128MB total) for PowerQUICC III configuration, Linux operating system boot image, and file systemClock generation circuitryPower good generation circuitryTwo 10/100 Fast Ethernet interfaces to a front panel connector and the update channel

A serial interface to a front panel connectorA serial bus interface to the Intelligent Platform Management controller (IPMC)An I2C bus interface to the serial EEPROMs, the clock/calendar, and the rear transition module (RTM)A SPI bus to the NTS, the Fast Ethernet switch, and the RTMTwo peripheral component interconnect (PCI) bus configuration interfaces to the Base and the Fabric Ethernet switchesA 1 Gb Ethernet port to the Base switch

3Local management processor

21

Figure 6. Local management processor with SCM

PowerQUICC III processorThe LMP is based on a PowerQUICC III processor. The PowerQUICC III LMP is packaged in a 783-pin ball grid array and is highly integrated with an embedded e500 core; integrated instruction and data caches, a system interface unit, and an integrated reduced instruction set computer-based (RISC) communications processor.

The PowerQUICC III LMP has the following characteristics:

JTAG/COP

Serial port to IPMC

UART1

UART0

DRAM bus

10/100Base-TEthernet

Localmanagement

processor

I2C bus

Front panelconsole port

COP debugheader

10/100Base-TEthernet

Update channel link

Front panelmgmt port

512MB DDR1SDRAM

64MB FlashROM 0

64MB FlashROM 1

Local bus

PCI bus 0

Fabric interfaceswitch

PCI bus 1

SPI bus

Clock/calendar

UserEEPROM

MACaddressEEPROM

RTM

NetworkTimingmodule

RTM

Base interfaceswitch

FastEthernetswitch

MDIO

Base interfaceswitch

10/100/1000 MBEthernet

10/100/1000 MBEthernet

ProcessorCore Freq. (MHz)

PCI Bus Freq. (MHz)

SYSCLK Freq. (MHz) VDD/Core

I Cache(Kbytes)

DCache(Kbytes)

Processor Version

MPC8541 833 66 66 3.3V/1.2V 32 32 0x0080

3 Subsystems

22

The PowerQUICC III processor chip incorporates the following elements:

CPU with e5000 core that implements Book E 32-bit architecture256KB on-chip memoryDDR memory controllerProgrammable interrupt controller (PIC) compliant with OpenPIC architectureLocal Bus Controller (LBC), which connects to the two 64MB flash components Two PCI bus controllers used to configure the Base and the Fabric switchesI2C controller providing access to the SDRAM module, two serial EEPROMs, the clock/calendar, and devices on the RTMTwo UART interfaces for:

Serial console port routed through front panel Serial port providing link between the PowerQUICC III and the on-board IPMC

Two 10/100Base-T Ethernet interfaces (Fast Ethernet)

One interface is used for management and maintenance purposes One interface is used for connecting to the redundant SCM through the update channel

One 10/100/1000 interface configured as a ten-bit interface (TBI), which is linked to a serializer/deserializer (SerDes). The interface is connected to the Base Ethernet switch as an alternate port to the PCI bus.

Serial management data input/output (MDIO) port for managing the physical layer devices (PHYs) on the 10/100Base-T interfacesSPI controller providing a four-wire serial bus for accessing the NTS, the 10/100Base-T Ethernet switch, and devices on the RTMEEPROM used to store media access control (MAC) addresses for Ethernet portsEEPROM used for storing user-defined informationClock/calendar to keep a running time and date. Includes lithium battery to maintain timekeeping for a specific minimum period of time when power is not available to the SCM.General purpose input/output (I/O) signals to and from various functions on the SCM for control and statusSynchronous dynamic random access memory (SDRAM) interface for the small outline dual in-line memory module (SODIMM).Two Spansion flash memory devices used as a redundant set of 64MB flash stores.Reads I/O bus information on the SCM version and revision historyIntegrated reset logic

3Clock synthesizer subsystem

23

Communications controller assignmentsThe PowerQUICC III communications functions as they pertain to the SCM are mapped out in the following table.

Clock synthesizer subsystemThe SCM has a clock synthesizer subsystem, which includes two clocks used by the PowerQUICC III. The other clocks are buffered and distributed to several Ethernet chips. These include Ethernet switch, PHY reference, XAUI-to-XFI interface, and the serial communications clocks. All system clocks are series-terminated at the clock source.

Table 1. Communication controller assignments

PowerQUICC III communications function SCM functionPCI0 Base interface switchPCI1 Fabric interface switchTSEC1 Gb link to Base interface switchUART0 Faceplate RS232 serial console portUART1 IPMC serial interface I2C SMBus to SDRAM, EEPROM, clock/calendarCPM:FCC0 10/1000Base-T Ethernet port (front panel) CPM:FCC1 10/100Base-T Ethernet port (update channel) CPM:SPI NTS, Fast Ethernet switch, RTM interfaceCPM:I2C Not used

3 Subsystems

24

Base Ethernet switch subsystem One of the primary functions of the SCM is to provide the Ethernet switch function for ATCA platforms. This is implemented using a Broadcom 28-port Gigabit switch. Twenty-four of the ports are 1 Gb and four of the ports are 10 Gb. The key features of the 28-port Gigabit Ethernet switch include the following:

Figure 7 on page 25 is a block diagram showing how the Base Ethernet switch communicates with the rest of the SCM.

Twenty-four 10/100/1000Mb ports Four 10 Gb portsEthernet switch/router with integrated MACs, packet buffer engine and switching engineSupports line-rate Layer 2 switching and Layer 2 multicast for all packet sizes and conditions.Line-rate switching for all packet sizes and conditionsEight layer Quality of Service (QoS) on a per-port basisSupports multiple (256) Spanning Trees (IEEE 802.1D)Supports Flow Control (IEEE 802.3x)VLAN Support based on MAC, port, subnet, and protocol (IEEE 802.1Q)

Supports link aggregation (up to 8 ports) (IEEE 802.3ad)IPSec engine supports DES, 3DES, HMAC-SHA-1, and HMAC-MD5, and is compatible with IPSec, SSL, TLS, 802.1x, and 802.11i.Head-of-line blocking preventionPer-port packet rate (storm) controlSupports port mirroring PCI interface to PowerQUICC III host processorSGMII interface support to external 1000 Mb PHYsIntegrated SerDesThermally enhanced 1156-pin FCBGA package

3Base Ethernet switch subsystem

25

Figure 7. Base Ethernet switch

ConfigurationThe switch has built in media access controllers (MACs) for each port that interface to the external physical layer devices (PHYs). The 1 Gigabit Ethernet ports can each be configured to operate in:

Serial Gigabit Media Independent Interface (SGMII) mode, which has Gigabit Ethernet ports going to the backplane, faceplate RJ-45 ports, and RTM.

Serializer/deserializer (SerDes) modes, which support Gigabit Ethernet ports going to optical SFP modules, 10/100/1000 Mbit ports going to copper SFP modules, or internal links.

The 10 Gigabit Ethernet ports are configured to operate in 10 Gigabit attachment unit interface (XAUI) mode, with ports routed to Zone 3 and the faceplate XFP ports.

Port 0Port 1Port 2Port 3

Port 7

Port 4Port 5Port 6

Frontpanel

Port 20

Port 23

Port 21

Port 25

Port 24

BCh3

BCh7

BCh4

BCh5

BCh8

BCh9

BCh6

BCh10BCh11

BCh14

BCh12

BCh13

ShMC1

BCh16BCh15

Backplane

Base Ethernetswitch

Quad GbEPHY

Dual GbEmagnetics

PowerQUICC III

PCI Bus 1

Port 8Port 9

Port 10Port 11

Port 15

Port 12Port 13Port 14

Port 19

Port 16

Port 17

Port 18

Port 22

Dual GbEmagnetics

Quad GbEPHY

Quad GbEPHY

Quad GbEPHY

GPIO Reset

PCI Bus

TSEC1

TBI GbSerDes

Dual GbEmagneticsDual GbEmagnetics



COM Express *

PCI-Express

GbE MACSerDes

XAUI toXFI

XFP

Zone 3conn

FE switchw/ mags

ShMC2

BCh2

RTM

Frontpanel

Quad GbESFP

BCh1

Port 27

Port 26* Reserved for future implementation

3 Subsystems

26

Port mapping

Table 2. Base Ethernet port mapping

Port # PHY/PortMDIO interface address Port destination

IPMI hardware address Logical slot

0 0 / 0 0x00 FE Switch (Shelf Manager) N/A N/A1 0 / 1 0x01 Base channel 2 (link to redundant

SCM)0x41 / 42 1 / 2

2 0 / 2 0x02 Base channel 3 0x43 33 0 / 3 0x03 Base channel 4 0x44 44 1 / 2 0x06 Base channel 5 0x45 55 1 / 3 0x07 Base channel 6 0x46 66 1 / 0 0x04 Base channel 7 0x47 77 1 / 1 0x05 Base channel 8 0x48 88 2 / 2 0x0A Base channel 9 0x49 99 2 / 3 0x0B Base channel 10 0x4A 1010 2 / 0 0x08 Base channel 11 0x4B 1111 2 / 1 0x09 Base channel 12 0x4C 1212 3 / 2 0x0E Base channel 13 0x4D 1313 3 / 3 0x0F Base channel 14 0x4E 1414 SerDes/SGMII N/A RTM or Base channel 15 N/A N/A15 SerDes/SGMII N/A RTM or Base channel 16 N/A N/A16* SerDes N/A COM Express N/A N/A17 SerDes/SGMII N/A RTM N/A N/A18 SFP N/A Front panel 1/6: 1G optical or

10/100 Mbit copper N/A N/A

19 SFP N/A Front panel 1/7: 1G optical or 10/100 Mbit copper

N/A N/A


N/A N/A


N/A N/A

22 SerDes/SGMII N/A RTM N/A N/A23 SerDes N/A PowerQUICC III N/A N/AX0 XAUI IF 0x10 Front panel 1/5: XFP N/A N/AX1 XAUI IF TBD RTM XAUI N/A N/AX2 XAUI IF N/A Not used N/A N/AX3 XAUI IF TBD RTM XAUI N/A N/A* Reserved for future implementation.

3Base Ethernet switch subsystem

27

10/100/1000Base-T Gigabit Ethernet portsSixteen Gigabit Ethernet PHYs provide 10/100/1000Base-T connectivity to the backplane and the Fast Ethernet (FE) switch. All of the 1000Base-T ports use a Gigabit PHY channel along with Gigabit Ethernet transformers to couple to the backplane.

12 or 14 ports to backplane node slots, depending on the board configuration.

One port to an FE switch, allowing two FE connections to an external Shelf Manager (if present) through Base channel 1.

One port to the Base interface of a redundant SCM (if present) through Base channel 2.

SerDes Gigabit Ethernet portsSix ports are configured as serializer/deserializer (SerDes) ports for a 14-slot (default) backplane. Two of the SerDes ports go to the on-board LMP and the COM Express site. Four of the SerDes ports are directed to SFP sockets on the SCM front panel.

10 Gigabit Ethernet portsThree 10Gb Base Ethernet ports provide external access on the front and rear faceplates.

1 XFP port on the front faceplate, XFI signaling converted from XAUI

2 XAUI signaling to RTM

The XFP 10Gb connector uses XFI signaling, composed of single transmit and receive 10Gb differential pairs. A XAUI-to-XFI interface chip converts the signaling from the XAUI signals (connected to the switch) to XFI signals (connected to the XFP port) at the front panel.

Fast Ethernet switchA 5-port 10/100Base-T Ethernet switch connects one Base interface port to one or two off-board Shelf Managers through Base channel 1.The Fast Ethernet (FE) switch is managed by the PowerQUICC III using the SPI port.

Table 3. Fast Ethernet switch port mapping

Port # Port destination Port states1 Base interface Port 1 should always be enabled.2 Base channel 1a Ports 2 and 3 may be enabled or disabled (together), depending on SCM

configuration for Base connectivity to an off-board Shelf Manager.3 Base channel 1b 4 Not used Ports 4 and 5 are not used, and should be disabled.5 Not used

3 Subsystems

28

Port status LEDsThis table describes the LEDs specific to the Base Ethernet.

Fabric Ethernet switch subsystem The SCM provides a 10 Gb Fabric Ethernet connection to 12 or 14 node slots (depending on the board configuration) based on AdvancedTCA 3.1. Four ports of the Fabric interface are also available on the front panel and three ports are routed to the RTM through the Zone 3 area.The Fabric interface is implemented using a Broadcom 20-port 10 Gb Ethernet switch. The backplane ports use XAUI, made up of four lanes (Tx and Rx differential pairs) operating at 3.125 Gb. The faceplate uses XFI on an XFP connector, converted from the XAUI signaling on the switch ports by a XAUI-to-XFI transceiver.

The Fabric Interface connects:

Note that there is no direct Base-to-Fabric connection on an SCM and no Fabric-to-Fabric connection between the two SCMs.Figure 8 is a block diagram showing how the Fabric Ethernet switch communicates with the rest of the SCM.

Table 4. Base Ethernet port status LEDS

Port LED and possible states1 Gb optical or 10/100 Mbit copper SFP portsEach Base Ethernet SFP port on the front panel or on the RTM has two LEDs

1 green link status LEDGreen – link establishedBlinking green – link activityOff – link fail or port is disabled

1 green port status LED Green – port is enabled Off – port is disabled

10 Gb XFP portsEach Base Ethernet XFP port on the front panel or on the RTM has two associated LEDs.

1 green link/activity status LEDGreen – link establishedBlinking green – link activityOff – link fail or port is disabled


Node slots as 10Gb XAUIFront panel XFP connector ports as 10Gb XFI1 or 3 ports (depending on the board configuration) to the RTM as 10Gb XAUI or 1Gb SerDes

One port to the COM Express site as 1 Gb SerDes (reserved for future implementation)The LMP on PCI Bus 2

3Fabric Ethernet switch subsystem

29

Figure 8. Fabric Ethernet switch


Port 7

Port 4Port 5Port 6

Port 18Port 16

Port 17

FCh2

FCh6

FCh3FCh4

FCh7FCh8

FCh5

FCh9FCh10

FCh13

FCh11FCh12

FCh15FCh14

Backplane

PowerQUICC III

PCI bus 2

Port 8Port 9


Port 19

GPIO Reset

PCI bus

TSEC2

COM Express

PCI-Express

GbE MACSerDes

XAUI toXFI

XFP

Frontpanel

XAUI toXFI

XFP

Port 15 XAUI toXFI

XFP

Port 14 XAUI toXFI

XFP

RTM

Fabric Ethernet

3 Subsystems

30

Port mapping

ConfigurationThe Fabric interface on the SCM is a managed 10Gb Ethernet switch with the following features:

Switch configuration via the PowerQUICC III PCI bus

Transmit (Tx) port disable based on E-Key port state

The Fabric Ethernet configuration port is connected to the PowerQUICC III through PCI bus 2, a 32-bits wide bus operating at 66MHz.

Table 5. Fabric Ethernet port mapping

Port # PHY/PortMDIO interface address

Port destination IPMI hardware address Logical slot

0 XAUI N/A Fabric channel 2 0x43 31 XAUI N/A Fabric channel 3 0x44 42 XAUI N/A Fabric channel 4 0x45 53 XAUI N/A Fabric channel 5 0x46 64 XAUI N/A Fabric channel 6 0x47 75 XAUI N/A Fabric channel 7 0x48 86 XAUI N/A Fabric channel 8 0x49 97 XAUI N/A Fabric channel 9 0x4A 108 XAUI N/A Fabric channel 10 0x4B 119 XAUI N/A Fabric channel 11 0x4C 1210 XAUI N/A Fabric channel 12 0x4D 1311 XAUI N/A Fabric channel 13 0x4E 1412 XAUI N/A RTM or Fabric channel 14 N/A N/A13 XAUI N/A RTM or Fabric channel 15 N/A N/A14 XAUI 0x10 Front panel XFP 1/1 N/A N/A15 XAUI 0x11 Front panel XFP 1/2 N/A N/A16 XAUI 0x12 Front panel XFP 1/3 N/A N/A17 XAUI 0x13 Front panel XFP 1/4 N/A N/A18* SerDes N/A COM Express N/A N/A19 XAUI N/A RTM N/A N/A* Reserved for future implementation.

3LED controller interface subsystem

31

10 Gigabit Ethernet portsFor the standard SCM configuration, twenty 10Gb Fabric Interface ports are available on the Fabric Ethernet. One is configured for 1Gb operation, connecting to the COM Express site, which is reserved for future implementation. Twelve connect directly to backplane node slots, four provide faceplate XFP connections, and one or three are routed to the RTM.

12 or 14 (depending on the board configuration) direct XAUI node-slot connections, which can be configured as 10 Gb XAUI or 1 Gb SerDes.4 XFP ports on the front faceplate. Each XFP 10Gb connector uses XFI signaling, composed of single transmit and receive 10Gb differential pairs. A XAUI to XFI interface chip converts the signaling from the XAUI signals at each switch port to the XFI signals on the XFP connector.1 or 3 ports (depending on the board configuration) routed to the RTM, which can be configured as 10 Gb XAUI or 1 Gb SerDes.

XFP port LEDsThis table describes the LEDs specific to the Fabric Ethernet. The LEDs are controlled by the Fabric Ethernet switch through a serial LED bus.

LED controller interface subsystem Ethernet port status LED indicators are controlled from the Base Ethernet and the Fabric Ethernet switches through serial LED data streams consisting of clock and data signals. One stream comes from the Fabric Ethernet switch and one comes from the Base Ethernet switch. Each bit position in the bit-stream indicates the on/off value of one port status LED (see Table 7 on page 32 and Table 8 on page 33 for LED controller assignments). The streams are routed to a complex programmable logic device (CPLD), which acts as a hardware LED controller-driver. The stream is then routed to the Zone 3 connector to support a LED controller on an optional rear I/O module (e.g. RTM).

A three-wire serial peripheral interface port (SPI) is available on the LED control CPLD. The SPI supports read operations on the CPLD Revision ID register. Write operations are not supported.

Table 6. XFP port LEDs

Port LED and possible statesXFP port LEDsEach of the Fabric Ethernet XFP ports on the front panel or the RTM has two LEDs.

1 green link/activity status LEDGreen – link establishedBlinking green – transmit/receive activityOff – no link or port disabled


3 Subsystems

32

LED controller Base assignments

Table 7. LED controller Base assignments

Eth switchBit stream position

Faceplate port label Switch port Indicator label Position color

Base 1 1/6 GE18 Link GreenBase 2 1/6 GE18 Link AmberBase 3 1/6 GE18 Port GreenBase 4 1/6 GE18 Port AmberBase 5 1/7 GE19 Link GreenBase 6 1/7 GE19 Link AmberBase 7 1/7 GE19 Port GreenBase 8 1/7 GE19 Port AmberBase 9 1/8 GE20 Link GreenBase 10 1/8 GE20 Link AmberBase 11 1/8 GE20 Port GreenBase 12 1/8 GE20 Port AmberBase 13 1/9 GE21 Link GreenBase 14 1/9 GE21 Link AmberBase 15 1/9 GE21 Port GreenBase 16 1/9 GE21 Port AmberBase 17 1/5 XG0 Link GreenBase 18 1/5 XG0 Link AmberBase 19 1/5 XG0 Port GreenBase 20 1/5 XG0 Port Amber


33

LED controller Fabric assignments

Table 8. LED controller Fabric assignments

Eth switchBit stream position

Faceplate port label Switch port Indicator label Position color

Fabric 1 1/1 XG15 Link GreenFabric 2 1/1 XG15 Link AmberFabric 3 1/1 XG15 Port GreenFabric 4 1/1 XG15 Port AmberFabric 5 1/2 XG16 Link GreenFabric 6 1/2 XG16 Link AmberFabric 7 1/2 XG16 Port GreenFabric 8 1/2 XG16 Port AmberFabric 9 1/3 XG17 Link GreenFabric 10 1/3 XG17 Link AmberFabric 11 1/3 XG17 Port GreenFabric 12 1/3 XG17 Port AmberFabric 13 1/4 XG18 Link GreenFabric 14 1/4 XG18 Link AmberFabric 15 1/4 XG18 Port GreenFabric 16 1/4 XG18 Port Amber

3 Subsystems

34

Front panel LED positionsFigure 9. LED positions on SCM front panel

1/1 – Link status1/1 – Port status






1/7 – Link status1/7 – Port status1/8 – Link status1/8 – Port status


Base 1 GE, SFP qty 4

Base 10 Gb, XFP

Fabric 10 GE, XFP qty 4

LMP Fabric Ether-net(LMP ETH), RJ-45

LMP serial port(LMP SER), RJ45


35

Front panel LED descriptions and states

Table 9. Front panel LEDs

Description StatesOne bi-color Out of Service (OOS)“LED1” controlled by GPO signals on the IPMC.

Amber = OOS Off = normal operation

One green Power Good“LED2” controlled by GPO signals on the IPMC.

Green = power is good

One amber Module State“LED3” controlled by GPO signals on the IPMC.

Amber flashing = bootingAmber on = functioningOff = inactive

One blue Hot Swap ready“Blue LED” controlled by an IPMC output signal.

Blue = ready for hot swap

Two green COM Express and LMP Ethernet console port linksLED part of the RJ-45 connector - indicates PHY port activity. Note: The COM Express connections are reserved for future implementation.

Green = linkOff = no link

Two amber COM Express and LMP Ethernet console port activity

LED part of the RJ-45 connector - indicates PHY port activity. Note: The COM Express connections are reserved for future implementation.

Amber blink = activityOff = no link or port disabled

Four green Fabric Ethernet XFP port status LEDs controlled by the LMP through the 10 G PHY LED signals.

Green = link is goodOff = no link or port disabled

Four green Fabric Ethernet XFP link/activity LEDs controlled by the LMP through the 10 G PHY LED signals.

Green = enabledGreen blink = activityOff = no link or port disabled

One green Base Ethernet XFP port status LEDs controlled by the LMP through the 10 G PHY LED signals.

Green = link is goodOff = no link or port disabled

One green Base Ethernet XFP link/activity LEDs controlled by the LMP through the 10 G PHY LED signals.

Green = enabledGreen blink = activityOff = no link or port disabled

One green Base Ethernet SFP link/activity LEDs part of the RJ-45 connector – indicates GE PHY port status, driven through the PHY LED signals.

Green = link is goodGreen blink = activityOff = no link or port disabled

Four green Base Ethernet SFP port statusLEDs part of the RJ-45 connector – indicates GE PHY port status, driven through the PHY LED signals.

Green = enabled Off = port disabled

3 Subsystems

36

Network timing subsystemThe NTS is the centralized system timing function provided by the SCM. The timing function is based on the backplane synchronization clock interface described in the AdvancedTCA Base Specification.

The list below and Figure 10 provide an overview of the operation of the NTS.

Provides a Synchronous Equipment Timing Source (SETS) for all shelf node modulesAccepts system timing from up to four (redundant) Building Integrated Timing Source (BITS) frames, and generates up to four framed synchronization signals for shelf-to-shelf chainingAccepts and generates Synchronization Supply Unit (SSU) frames per ITU-T Recommendation G.812Accepts timing references from backplane node module synchronization clocks (CLK3-A and CLK3-B)

Generates 8 KHz and 19.44 MHz synchronization clocks on backplane (CLK1-A/B and CLK2-A/B, respectively) Generates 8 KHz, 2.048 MHz, 1.544 Mhz, and 19.44 Mhz synchronization clocks for the RTMCan provide automatic or manual hitless input switchover on loss of input, if configured using input priority levelsProvides automatic hitless protection of the system clock when operating with a redundant SCM

3Network timing subsystem

37

Figure 10. Network timing subsystem

The NTS can select reference timing sources from the synchronization bus, the RTM, or the update channel and be configured with priority levels for each input. By monitoring each input, the NTS can automatically switch between valid sources based on those priorities. Information on configuring the different timing sources is covered in Network timing subsystem (NTS) software on page 90.

CLK3 -A

2 KHz IN

6.48 MHz IN

2 KHz OUT UPDCH3Tx

BITS/SSU-4

TO2

PowerQUICC III

SPI

GPIO

SPI

12.8MHzTCXO

12.8MHzOCXO

I9

I6

I5

DS26502Framer /Deframer


REF-4

UPDCH4Tx

UPDCH3Rx

UPDCH4Rx

CLK3 -B

19.44 MHz

8 KHz

CLK1 -A

CLK1 -B

CLK2 -A

CLK2 -B

I3

I4

SYNC2K

I11

TO4

TO5

TO11

TO1

REFCLK

CSB

I10

12.8MHzXO

19.44 MHzTO3



Networktiming

subsystem

RTM (Illustration)

SCM

6.48 MHZ OUT

BITS/SSU-3

REF-3

BITS/SSU-2

REF-2

REF-1

BITS/SSU-1

EXT CLK OUT

EXT CLK IN 1

EXT CLK IN 2

EXT CLK IN 3

EXT CLK IN 4

3 Subsystems

38

Synchronization clock and update channel buffer enables The PowerQUICC III can enable or disable the clock buffers to the synchronization clock and update channel outputs. Permission to drive these signals comes from the Shelf Manager through the IPMC, as part of E-Keying.

Rear transition module When an SCM is installed on a RadiSys shelf, an associated rear transition module (RTM) can be connected to provide external access to some of the SCM’s signals and electrical connections. The RTM is installed in the rear of the platform and connects to the SCM using the Zone 3 connectors. A shelf peripheral module (SPM), such as the ATCA-5010, is used as the RTM when an SCM is installed in a 14-slot RadiSys shelf like the ATCA-6000. The SPM provides back-of-shelf access, including ATCA shelf management interfaces to various shelf peripherals and sensors. SPMs are optional when SCMs are installed in RadiSys 6-slot shelves, such as the ATCA-6006, or in non-RadiSys shelves.

Electrical connections between the SCM and the associated RTM include:

Two or four 1 Gb SGMII/SerDes ports from the Base interface Two 10 Gb XAUI port from Base interfaceOne or three 10Gb XAUI ports from Fabric switch Four external reference clock inputsThree external system clock outputsI2C interface to IPMC for ID, temperature, temperature sensors, and control management

I2C interface to the LMP for IDSPI to the LMPPCI Express interface from COM Express site (the site is reserved for future implementation)Switched 12 V from payload supplyCurrent limited 3.3 V from IPMC supplyRTM hot-swap switch closure signalHot-swap RTM LED output from the IPMC

3Rear transition module

39

Figure 11 shows a block diagram example of an SPM designed for a 14-slot RadiSys shelf.

Figure 11. SPM block diagram

BITS/SSU clk

MagI/F

RJ -45

RJ -45

SFP

SFP

Zone 3conn


ID

ClkInRefClk

Buf

ID & tempBuf

SFP

SFP

SPI

1Gb SD

SFP

SFP

SFP

1Gb SD

BITS/SSU clk

BITS/SSU clk

MagI/F

BITS/SSU clk

MagI/F

MagI/F

RJ -45

RJ -45

SPM (example)

SFP*


3 Subsystems

40

RTM hot-swap support buffersIn order to protect the RTM interface signals from disruption during insertion or removal of the RTM, buffers have been added to many of the RTM signals. These buffers are controlled through RTM_BUF_ENA~ on GPIO PC12 of the LMP. The buffer will be disabled whenever the RTM is not installed, or is being changed.

RTM interface signalsSpace has been allocated in the Zone 3 connectors for signals passing between the SCM and the RTM, including data path, configuration, and hot-swap signals. This represents a superset of the signals that may be used for various combinations of the SCM and the RTM.

Table 10. RTM interface signals

Signal nameDirection from SCM to RTM

Pins Description

+12V_RTM Out 6 Power source for RTM circuits, 3 Amp limit. This is current limited, and can be turned on and off by the IPMC.

+3.3V_IPMC_LTD Out 1 Power source for RTM hot-swap control circuits, 0.5 Amp limit. This is current limited, and can be turned on and off by the IPMC.

RTM_EJECT_IN~ In 1 Signal is grounded on RTM when ejector handle is closed (RTM fully seated). Pull-up to +3.3V_IPMI on SCM.

RTM_Present~ In 1 Indicates RTM is connected. This pin should be grounded on the RTM. Pulled up to +3.3V_IPMI on SCM.

RTM_HOTSWAP~ Out 1 Drive signal for RTM hot-swap indicator LED. 3.3V Signal goes low to turn on LED. Pull up to +3.3V_IPMC on RTM.

RTM_ENA~ Out 1 Low-true Enable signal for RTM on-board controller. Low enables, high disables (resets). Pull up to +3.3V_IPMC on RTM.

RTM_INTR~ In 1 Low-true interrupt signal from the RTM to the IPMC. Pulled up to +3.3V_IPMI on SCM.IPMC_I2C_CLK I/O 1 IPMC control bus clock. Pulled up to +3.3V_IPMI on SCM.IPMC_I2C_DAT I/O 1 IPMC control bus data. Pulled up to +3.3V_IPMI on SCM.PQ_12C_CLK I/O 1 PowerQUICC III control bus clock. Pulled up to Payload +3.3V on SCM.PQ_12C_DAT I/O 1 PowerQUICC III control bus data. Pulled up to Payload +3.3V on SCMSPI_RTM_SEL~ Out 1 SPI target select, low-true. RTM SPI bus slave devices will use this select along with

address bits to select a certain device. 3.3V signal from LMP goes low to select target device. Pull up to Payload +3.3V on RTM.

SPI_SLVA[0..2] Out 3 SPI bus target device address. The MSB indicates SMC3 (0) or RTM3 (1) as the target board. 3.3V signal from LMP.

BLED_CLK Out 1 Base switch LED control stream clock signal. 3.3 V signal from LMP controls RTM faceplate LEDs.

BLED_DAT Out 1 Base switch LED control stream data signal. 3.3 V signal from LMP controls RTM faceplate LEDs.

FLED_CLK Out 1 Fabric switch LED control stream clock signal. 3.3 V signal from LMP controls RTM faceplate LEDs.

3Rear transition module

41

FLED_DAT Out 1 Fabric switch LED control stream data signal. 3.3 V signal from LMP controls RTM faceplate LEDs.

RTM_RST~ Out 1 Reset output to RTM peripherals (active low). Pull up to +3.3V_IPMC on RTM.SPI_CLK Out 1 SPI data transfer clock. 3.3 V signal from LMP.SPI_SDO Out 1 SPI data out (from LMP to target). 3.3V signal from LMP.SPI_SDI In 1 SPI data in (3.3 V signal from RTM target to LMP).BASE_MDC I/O 1 Base 1Gb interface MDIO clock. 3.3 V signal from Base switch.BASE_MDIO Out 1 Base 1Gb interface MDIO data. 3.3 V bidirectional signal from Base switch.BASE_XMDC I/O 1 Base 10 Gb interface MDIO clock. 2.5 V signal from Base switch.BASE_XMDIO Out 1 Base 10 Gb interface MDIO clock. 2.5 V bidirectional signal from Base switch.FABRIC_XMDC Out 1 Fabric 10 Gb interface MDIO clock. 2.5 V signal from Base switch.FABRIC_XMDIO I/O 1 Fabric 10 Gb interface MDIO data. 2.5 V bidirectional signal from Base switch.EXTCLK1_IN+/- In 2 Recovered clock 1 in, LVDS differentialEXTCLK2_IN+/- In 2 Recovered clock 2 in, LVDS differentialEXTCLK3_IN+/- In 2 Recovered clock 3 in, LVDS differentialEXTCLK4_IN+/- In 2 Recovered clock 4 in, LVDS differentialEXTCLK_OUT+/- Out 2 Clock out, LVDS differentialRTM_LMP_INTR5~ In 1 Interrupt (low-true) from the RTM to the LMP. Pulled up on SCM.RTM_LMP_INTR6~ In 1 Interrupt (low-true) from the RTM to the LMP. Pulled up on SCM.NTS_8K_OUT+/- Out 2 8KHz external clock out, LVDS differentialNTS_19M_OUT+/- Out 2 19.44MHz external clock out, LVDS differentialPCI_E_TX+/- Out 2 One PCI Express Tx lane, differentialPCI_E_RX+/- In 2 One PCI Express Rx lane, differentialBASE_GE_TX[0..1]+/- Out 164 Base interface 1Gb Tx port, differential. Base interface SerDes signals. The number of

Base ports available depends on SCM component loading (stuffing).BASE_GE_RX[0..1]+/- In 16 Base interface 1Gb Rx port, differential. Base interface SerDes signals. The number of

Base ports available depends on SCM component loading (stuffing).BASE_10G_TX[0..3]+/- Out 16 Base interface 10 Gb Tx port, differential. Base interface XAUI signals. The number of

Base ports available depends on SCM component loading (stuffing).BASE_10G_RX[0..3]+/- In 16 Base interface 10 Gb Rx port, differential. Base interface XAUI signals. The number of

Base ports available depends on SCM component loading (stuffing).FABRIC_10G_TX[0..3]+/- Out 32 Fabric interface 10 Gb Tx port, differential. Fabric interface XAUI signals. The number of

Fabric ports available depends on SCM component loading (stuffing).FABRIC_10G_RX[0..3]+/- In 32 Fabric interface 10 Gb Rx port, differential. Fabric interface XAUI signals. The number of

Fabric ports available depends on SCM component loading (stuffing).

Table 10. RTM interface signals(continued)

Signal nameDirection from SCM to RTM

Pins Description

3 Subsystems

42

Hardware management subsystem The function and operation of the SCM hardware management subsystem is controlled by the Intelligent Platform Management Controller (IPMC). The hardware management subsystem is the collection of IPMCs and sensors on the installed modules and field replaceable units (FRUs), and the communication between these devices and the Shelf Manager. The SCM’s IPMC manages the commands and the data portion of this subsystem.

The IPMC has the following features:

Figure 12 on page 43 shows how the IPMC communicates with other subsystems like the LMP and the RTM.

Dual I2C interface to backplane intelligent platform management bus (IPMB) with programmable pull-upsSerial interfaces to LMP and COM Express siteI2C interfaces to I/O expander, field-programmable gate array (FPGA), temperature sensors, voltage sensors, and RTMSystem event log (SEL) capabilities

Remote executable flash and micro-controller software update supportElectronic keying (E-Keying) support for Base interface, Fabric interface, synchronization clocks, and update channelsRemote micro-controller software update support

3Hardware management subsystem

43

Figure 12. IPMC control interface to payload and RTM

IPMC

48 V DC/DC

+12V

LMP

FlipFlopGPIO5

Powerswitch

Powerswitch

-48 V

COM Express*

+12V +3.3V

+3.3V_CE

+12V_CEPowerswitch

1 GbEEna

CE_GBE1

Update channel

COM0_TX

COM0_RX

IRQ 0

SRESET#

BUF_MRESET#

PORESET#FLSH_SEL

PL-RST#0

CE_UC_ENA

CECOM1_RX

CECOM1_TX

COME:RESET#

COM_E_TYPE0

COM_E_TYPE1

COM_E_TYPE2

COM_E_PRST~

COME:PWR_FAIL~

COME:PWR_ENA~

IPMC:I2C_SDA

IPMC:I2C_SCL

RTM_ENA~

RTM:HOTSWAP

RTM_RST~

RTM_EJECT_IN~RTM_PRESENT~

RTM_INTR~

RTM:+3.3V_FAIL~

RTM_+3.3V_ENA~

RTM:+12V_FAIL~

RTM_+12V_ENA~

IPMC:MSG_FLAG_STBIPMC:MSG_FLAG

IPMC:MSG_FLAG_CLR

+3.3V_IPMC

PAYLOAD_PWR#

+12V_RTM

+3.3V_RTM

RTM

+12V +3.3V_IPMI

3 Subsystems

44

Non-volatile random access memoryThe non-volatile random access memory (NVRAM) in the IPMC on the SCM is a 64 KB device. The NVRAM stores the following types of information about the field replaceable units (FRUs):

The SCM acting as an active Shelf Manager, retrieves this information from the IPMC and stores the information. To provide redundancy, the active Shelf Manager also sends the FRU information to the inactive Shelf Manager, which also stores the information. This information is recorded in the system event log (SEL) to monitor system events.

IPMC watchdog timerThe IPMC entity includes a hardware watchdog timer. Once the watchdog is activated, the IPMC must strobe it in order to keep it from timing out. If a firmware or a hardware problem on the IPMC causes it to stop strobing the watchdog timer, the IPMC is automatically isolated from the IPMB. This isolation of the IPMC keeps the IPMI buses functional for the remaining IPMC devices. A two-pin header can be used to disable the watchdog timer during debug.

Power suppliesInformation on the SCM’s power consumption and emissions is provided in Specifications on page 131.

Power input module (PIM)The SCM uses a power input module (PIM), which provides input conditioning of the -48V backplane power feeds. This module is between the input protection fuses and the main 48V to 12V DC/DC converter. Features of the PIM include:

Serial numberPart numberManufacturerDate and time of manufacture

Product nameFRU capabilitiesPoint to point connectivity records for modules that plug into the backplane, which is used for E-Keying.

Inrush current limit protectionIntegrated conducted noise filterIsolated auxiliary 3.3V power for IPMI circuitORing field effect transistors (FETs) on A and B -48V feeds

A/B feed loss alarmHot-swap controlGlitch hold-up circuit based on external capacitor and 72V charging circuitInput under voltage and over voltage protectionOver current and thermal protection

3Power supplies

45

-48V to 12V primary voltage conversion A DC/DC converter is used to generate 12V for the SCM. This power supply provides up to 18 Amps of current to power the SCM payload and RTM. The RTM power buses are switched from this supply through integrated FET switches, controlled by the IPMC. Output voltage tolerance: ±5%. The DC/DC is switched on and off through an optically isolated control signal from the IPMC.

Power converters The following non-isolated supplies are generated from the 12V intermediate supply by switching regulators:

5.0 V DC for components such as the hard drive and XFP modules. 3.3 V DC for components such as the hard drive, the XFP modules, and the NTS block. 2.5 V DC for use components such as the PowerQUICC III, the DRAM, the 1Gb SerDes, and the FE switch. 1.8 V DC for components such as the XFP modules, the XAUI-to-XFI interfaces, the PCI Express 1Gb controllers.

1.25 V DDR for the DDR DRAM module. 1.25 V DC for the Base and the Fabric Ethernet switches. 1.2 V DC for the PowerQUICC III, Gigabit Ethernet Quad PHYs, and the XAUI-XFI interfaces. 1.1 V DC for use by the PCI Express 1Gb controllers. 1.0 V DC for the 10 Gb Ethernet switch.

3 Subsystems

46

47

4PHYSICAL INTERFACES

Mechanical specificationThe SCM dimensions conform to the PICMG AdvancedTCA 3.0 R2.0 Specification.

Table 11. Base board layout

PowerQUICC III

Fabric Ethernet switch

Base Ethernet switch

Power input module

COM Express site*SAS/SATA site*

* Reserved for future implementation

Consoleports

Base andFabric XFPsockets

Base SFP sockets

DC/DCConverter

DRAMSODIMM

Clock/calendar

4 Physical Interfaces

48

Thermal designHeat sinks are used in the SCM’s thermal design. Extruded and crosscut aluminum heat sinks with thermal interface pads are used to cool the Base Ethernet switch, the Fabric Ethernet switch, the Base Ethernet PHYs, and the PowerQUICC III processor. The heat sinks for the XFPs are a pin fin configuration and are compliant with the specifications defined in the XFP MSA.

Front panel interfacesEach SCM includes a sheet metal front panel that will serve as an EMI/RFI barrier and complies with PICMG 3.0 Revision 2.0. The front panel provides access to the LMP, the Base Ethernet switch, and the Fabric Ethernet switch.

The front panel includes: RJ-45 connector for the RS-232 serial port (labeled LMP SER) from the LMP.RJ-45 connector (with integrated LEDs) for the LMP 10/100Base-T copper port (labeled LMP ETH). Two LEDs on each connector indicate link status and activity. Four 10 Gb XFP sockets with 10Gb XFI-to-XAUI interfaces to the Fabric Ethernet switch. Two LEDs next to connector indicate link, activity, and port status. XFP module loading is optional. One XFP socket with a 10Gb XFI-to-XAUI interface to the Base Ethernet switch. Two LEDs next to connector indicate link, activity, and port status. XFP module loading is optional. Four SFP sockets with integrated status LEDs. Two LEDs on each connector indicate link, activity, and port status. The sockets have 1Gb SerDes interfaces to the Base Ethernet switch when optical SFPs are loaded or 10/100 Mbit SerDes interfaces to the Base Ethernet switch when copper SFPs are loaded. SFP loading is optional.Three connection types on the front panel are reserved for future implementation of the COM Express: one USB port, one RJ-45 connector for a RS-232 serial port, and one RJ-45 connector for a 10/100/1000 Base-T copper interface.Reset button (recessed to prevent accidental actuation)Blue Hot-swap LED (ATCA “Blue LED”)Red or amber Out Of Service LED (ATCA “LED 1”)Green Power Good LED (ATCA “LED 2”)Amber SCM Status LED (ATCA “LED 3”)Five green/amber 10Gb Ethernet port Link status LEDsFive green/amber 10Gb Ethernet port Activity status LEDsHot-swap ejector latches

See Front panel interfaces on page 49 for the locations of the front panel interfaces. For descriptions of the front panel LEDs and their various states, see LED controller interface subsystem on page 31.

4Front panel interfaces

49

Figure 13. Front panel interfaces

AdvancedTCA R

ATCA - 2210

BASE

FABRIC

CO

M-E

US

BC

OM

-ES

ER

CO

M-E

ETH

XFP 1/1

XFP 1/2

XFP 1/3

LINK

POR

TLIN

KPO

RT

LINK

POR

TLIN

KPO

RT

SFP

1/9S

FP 1/8

SFP

1/7S

FP 1/6

RESET

XFP 1/5

XFP 1/4

LMP

ETH

LMP

SE

R

OO

SPW

RA

CT

PORT

H/S

LINK

LINK

LINK

LINK

LINK

PORT

PORT

PORT

PORT

LMP serial port

LMP Ethernet port

Fabric Ethernet XFPs

Base Ethernet SFPs

Base Ethernet XFP

Reset


50

Push Buttons/SwitchesThe SCM includes a reset button and a hot-swap switch. The recessed reset push button only resets the payload section of the SCM when pressed. The hot-swap switch closes when the ejector handle is fully latched.

LMP serial console portAn RJ-45 connector provides serial access to the console port of the LMP. This may be used as a console port for setup and management of applications running on the PowerQUICC III processor chip.

Note: Use the LMP serial console port for temporary connections only. Use a shielded cable with the port.

LMP Ethernet maintenance portAn RJ-45 connector is used to provide a maintenance port for the SCM. This Ethernet port interfaces to the PowerQUICC III processor chip. This connector includes the isolation magnetics and LEDs for Link (green) and Activity (amber).

*Unused pairs are terminated.Note: Use the LMP Ethernet port for connections to indoor cables only. Use a shielded cable with the port for applications that need to fully comply with NEBS GR 1089 requirements.

Table 12. LMP serial console port connector pinouts

Pin Description Pin Description1 LMP_RTS 5 GND2 NC 6 LMP_RXD3 LMP_TXD 7 NC4 GND 8 LMP_CTS

Table 13. LMP maintenance port connector pinouts

Pin Description Pin Description1 TX+ 5 Unused (DC-)*2 TX- 6 RX-3 RX+ 7 Unused (DD+)*4 Unused (DC+)* 8 Unused (DD-)*

4Front panel interfaces

51

SFP Base Ethernet interface portsThe SFP Base Ethernet connectors are used to provide front panel access to the Base Ethernet switch. Each connector is inside a metal cage, which is directly coupled to a port of the Base Ethernet switch. You can install or remove SFP transceivers while the SCM is running. The status of SFPs can be detected and shown. An SFP port will not work if:

no module is installedthe port is administratively disabled

See the ATCA-2210 Support page of RadiSys Web site at www.radisys.com for a list of SFP configurations and products that have been tested and confirmed by RadiSys to work with the SCM. The SFP Base Ethernet connector pinouts are listed in the following table.

XFP Base Ethernet and Fabric Ethernet portsThe XFP connectors are used to provide front panel access to the 10Gb ports of the Base and the Fabric Ethernet switches. Each connector is coupled to a switch port through a XAUI-to-XFI interface. The connector is inside a metal cage, which is connected to the SCM. You can install or remove XFP transceivers while the SCM is running, but they are not autodetected. An XFP port will not work if:

no module is installedthe port is administratively disabled

See the ATCA-2210 Support page of RadiSys Web site at www.radisys.com for a list of XFP configurations and products that have been tested and confirmed by RadiSys to work with the SCM.

Table 14. SFP Base Ethernet connector pinouts

Pin Description Pin Description1 GND 11 GND2 TXFAULT 12 RD-3 TXDIS 13 RD+4 MODDEF2 14 GND5 MODDEF1 15 +3,3V6 MODDEF0 16 +3.3V7 RATE SEL 17 GND8 LOS 18 TD+9 GND 19 TD-10 GND 20 GND11 GND 20 GND




52

The XFP Base Ethernet & Fabric Ethernet connector pinouts are listed in the following table.

Backplane interfacesThe following SCM interfaces are available through the backplane:

Update channelsUpdate channels are the backplane connections that exist between a pair of SCMs operating on a redundant basis. Application software can use update channels for redundancy interlock. If you configure a pair of SCMs to use the update channel for redundancy support, you must install the SCMs into hub slots linked by an update channel. The update channel for SCMs exists between physical slots 1 and 2 in RadiSys 6-slot shelf and 7 and 8 in a RadiSys 14-slot shelf.For more information on update channels and the slots that link redundant modules, see the appropriate Platform Hardware Reference manual (SYS-6000/6010 or SYS-6006) for the platform you are using.

Table 15. XFP Base Ethernet & Fabric Ethernet connector pinouts

Pin Description Pin Description Pin Description1 GND 11 SDA 21 P_DOWN2 NC 12 MOD_ABS 22 +1.8V3 MOD_DESEL 13 MOD_NR 23 GND4 INTERRUPT 14 RX_LOS 24 REFCLK+5 TX_DIS 15 GND 25 REFCLK-6 +5.0V 16 GND 26 GND7 GND 17 RD- 27 GND8 +3.3V 18 RD+ 28 TD-9 +3.3V 19 GND 29 TD+10 SCL 20 +1.8V 30 GND

Twelve or fourteen 10/100/1000Base-T Base interface node portsOne inter-LMP link to a redundant SCM through the update channelOne Base Ethernet switch-to-switch link port through Base channel 2Dual 10/100Base-T Base interface port to one or two external Shelf Managers through Base channel 1Twelve or fourteen 10 Gb XAUI Fabric interface node ports

Six synchronization channel clock input/outputs to the NTSMaster/slave link to NTS module clock through the update channelTwo -48V power railsIPMB interface (two I2C ports) with programmable pull-upsEight hardware address lines, seven address/one parity

4Backplane interfaces

53

Power connector signals (P10)The Zone 1 contact assignments for P10 are included in the following table.

Table 16. Zone 1 contact assignments P10

Contact number Designation Description Mating sequence

1 Reserved Reserved N/A2 Reserved Reserved N/A3 Reserved Reserved N/A4 Reserved Reserved N/A5 HA0 Hardware address bit 0 Third6 HA1 Hardware address bit 1 Third7 HA2 Hardware address bit 2 Third8 HA3 Hardware address bit 3 Third9 HA4 Hardware address bit 4 Third10 HA5 Hardware address bit 5 Third11 HA6 Hardware address bit 6 Third12 HA7/P Hardware address bit 7 (odd parity bit) Third 13 SCL_A IPMB clock, port A Third14 SDA_A IPMB data, port A Third15 SCL_B IPMB clock, port B Third16 SDA_B IPMB data, port B Third17 MT1_TIP Metallic test # 1 tip Third18 MT2_TIP Metallic test # 2 tip Third19 -RING_A Ringing generator A negative ringing Third20 -RING_A Ringing generator B negative ringing First21 MT1_RING Metallic test #1ringing Third22 MT2_RING Metallic test #2 ringing Third23 RRTN_A Ringing generator A return Third24 RRTN_B Ringing generator B return Third25 SHELF_GND Shelf ground and safety ground First26 LOGIC_GND Logic ground (reference and return for front

module-to-front module logic signals)Fourth

27 ENABLE_B Enable B (short pin for power sequencing, tied to VRTN_B on backplane)

First

28 VRTN_A Voltage return A First29 VRTN_B Voltage return B First30 EARLY_A -48 volt early (pre-charge) A First31 EARLY_B -48 volt early (pre-charge) B First32 ENABLE_A Enable A (short pin for power sequencing, tied to VRTN_A

on backplane)Fourth


54

Backplane connector J20 signals


33 -48V_A -48 volt A, uses ENABLE_A to enable converters Second34 -48V_B -48 volt B, uses ENABLE_B to enable converters Third

Unused pin on SCM

Table 17. Backplane connector J20 signals

Row #Interface designation ab cd ef gh

1 Clks CLK1A+ CLK1A- CLK1B+ CLK1B- CLK2A+ CLK2A- CLK2B+ CLK2B-2 Update channel and

clksTx4(UP)+ Tx4(UP)- Rx4(UP)+ Rx4(UP)- CLK3A+ CLK3A- CLK3B+ CLK3B-

3 Tx2(UP)+ Tx2(UP)- Rx2(UP)+ Rx2(UP)- Tx3(UP)+ Tx3(UP)- Rx3(UP)+ Rx3(UP)-4 Tx0(UP)+ Tx0(UP)- Rx0(UP)+ Rx0(UP)- Tx1(UP)+ Tx1(UP)- Rx1(UP)+ Rx1(UP)-5 Fabric channel 15 Tx2[15]+ Tx2[15]- Rx2[15]+ Rx2[15]- Tx3[15]+ Tx3[15]- Rx3[15]+ Rx3[15]-6 Tx0[15]+ Tx0[15]- Rx0[15]+ Rx0[15]- Tx1[15]+ Tx1[15]- Rx1[15]+ Rx1[15]-7 Fabric channel 14 Tx2[14]+ Tx2[14]- Rx2[14]+ Rx2[14]- Tx3[14]+ Tx3[14]- Rx3[14]+ Rx3[14]-8 Tx0[14]+ Tx0[14]- Rx0[14]+ Rx0[14]- Tx1[14]+ Tx1[14]- Rx1[14]+ Rx1[14]-9 Fabric channel 13 Tx2[13]+ Tx2[13]- Rx2[13]+ Rx2[13]- Tx3[13]+ Tx3[13]- Rx3[13]+ Rx3[13]-10 Tx0[13]+ Tx0[13]- Rx0[13]+ Rx0[13]- Tx1[13]+ Tx1[13]- Rx1[13]+ Rx1[13]-Note: Each differential pair has an individual L-shaped ground contact (not shown).


Row # Interface designation ab cd ef gh

1 Fabric channel 12 Tx2[12]+ Tx2[12]- Rx2[12]+ Rx2[12]- Tx3[12]+ Tx3[12]- Rx3[12]+ Rx3[12]-2 Tx0[12]+ Tx0[12]- Rx0[12]+ Rx0[12]- Tx1[12]+ Tx1[12]- Rx1[12]+ Rx1[12]-3 Fabric channel 11 Tx2[11]+ Tx2[11]- Rx2[11]+ Rx2[11]- Tx3[11]+ Tx3[11]- Rx3[11]+ Rx3[11]-4 Tx0[11]+ Tx0[11]- Rx0[11]+ Rx0[11]- Tx1[11]+ Tx1[11]- Rx1[11]+ Rx1[11]-5 Fabric channel 10 Tx2[10]+ Tx2[10]- Rx2[10]+ Rx2[10]- Tx3[10]+ Tx3[10]- Rx3[10]+ Rx3[10]-6 Tx0[10]+ Tx0[10]- Rx0[10]+ Rx0[10]- Tx1[10]+ Tx1[10]- Rx1[10]+ Rx1[10]-7 Fabric channel 9 Tx2[9]+ Tx2[9]- Rx2[9]+ Rx2[9]- Tx3[9]+ Tx3[9]- Rx3[9]+ Rx3[9]-8 Tx0[9]+ Tx0[9]- Rx0[9]+ Rx0[9]- Tx1[9]+ Tx1[9]- Rx1[9]+ Rx1[9]-9 Fabric channel 8 Tx2[8]+ Tx2[8]- Rx2[8]+ Rx2[8]- Tx3[8]+ Tx3[8]- Rx3[8]+ Rx3[8]-10 Tx0[8]+ Tx0[8]- Rx0[8]+ Rx0[8]- Tx1[8]+ Tx1[8]- Rx1[8]+ Rx1[8]-Note: Each differential pair has an individual L-shaped ground contact (not shown).

Table 16. Zone 1 contact assignments P10 (continued)

Contact number Designation Description Mating sequence


55




Row #Interface designation ab cd ef gh

1 Fabric channel 7 Tx2[7]+ Tx2[7]- Rx2[7]+ Rx2[7]- Tx3[7]+ Tx3[7]- Rx3[7]+ Rx3[7]-2 Tx0[7]+ Tx0[7]- Rx0[7]+ Rx0[7]- Tx1[7]+ Tx1[7]- Rx1[7]+ Rx1[7]-3 Fabric channel 8 Tx2[6]+ Tx2[6]- Rx2[6]+ Rx2[6]- Tx3[6]+ Tx3[6]- Rx3[6]+ Rx3[6]-4 Tx0[6]+ Tx0[6]- Rx0[6]+ Rx0[6]- Tx1[6]+ Tx1[6]- Rx1[6]+ Rx1[6]-5 Fabric channel 5 Tx2[5]+ Tx2[5]- Rx2[5]+ Rx2[5]- Tx3[5]+ Tx3[5]- Rx3[5]+ Rx3[5]-6 Tx0[5]+ Tx0[5]- Rx0[5]+ Rx0[5]- Tx1[5]+ Tx1[5]- Rx1[5]+ Rx1[5]-7 Fabric channel 4 Tx2[4]+ Tx2[4]- Rx2[4]+ Rx2[4]- Tx3[4]+ Tx3[4]- Rx3[4]+ Rx3[4]-8 Tx0[4]+ Tx0[4]- Rx0[4]+ Rx0[4]- Tx1[4]+ Tx1[4]- Rx1[4]+ Rx1[4]-9 Fabric channel 3 Tx2[3]+ Tx2[3]- Rx2[3]+ Rx2[3]- Tx3[3]+ Tx3[3]- Rx3[3]+ Rx3[3]-10 Tx0[3]+ Tx0[3]- Rx0[3]+ Rx0[3]- Tx1[3]+ Tx1[3]- Rx1[3]+ Rx1[3]-Note: Each differential pair has an individual L-shaped ground contact (not shown).


Row # Interface designation ab cd ef gh

1 Fabric channel 2 Tx2[2]+ Tx2[2]- Rx2[2]+ Rx2[2]- Tx3[2]+ Tx3[2]- Rx3[2]+ Rx3[2]-2 Tx0[2]+ Tx0[2]- Rx0[2]+ Rx0[2]- Tx1[2]+ Tx1[2]- Rx1[2]+ Rx1[2]-3 Fabric channel 1 Tx2[1]+ Tx2[1]- Rx2[1]+ Rx2[1]- Tx3[1]+ Tx3[1]- Rx3[1]+ Rx3[1]-4 Tx0[1]+ Tx0[1]- Rx0[1]+ Rx0[1]- Tx1[1]+ Tx1[1]- Rx1[1]+ Rx1[1]-5 ShMC Port BI_SHMC_

Tx0+BI_SHMC_Tx0-

BI_SHMC_Rx0+

BI_SHMC_Rx0-

BI_SHMC_Tx1+

BI_SHMC_Tx1-

BI_SHMC_Rx1+

BI_SHMC_Rx1-

6 Base channel 2 BI_DA2+ BI_DA2- BI_DB2+ BI_DB2- BI_DC2+ BI_DC2- BI_DD2+ BI_DD2-7 Base channel 3 BI_DA3+ BI_DA3- BI_DB3+ BI_DB3- BI_DC3+ BI_DC3- BI_DD3+ BI_DD3-8 Base channel 4 BI_DA4+ BI_DA4- BI_DB4+ BI_DB4- BI_DC4+ BI_DC4- BI_DD4+ BI_DD4-9 Base channel 5 BI_DA5+ BI_DA5- BI_DB5+ BI_DB5- BI_DC5+ BI_DC5- BI_DD5+ BI_DD5-10 Base channel 6 BI_DA6+ BI_DA6- BI_DB6+ BI_DB6- BI_DC6+ BI_DC6- BI_DD6+ BI_DD6-Note: Each differential pair has an individual L-shaped ground contact (not shown).

Unused pin on SCM


56


Zone 3 connector J30 signals


Row #Interface

designation ab cd ef gh

1 Base channel 7 BI_DA7+ BI_DA7- BI_DB7+ BI_DB7- BI_DC7+ BI_DC7- BI_DD7+ BI_DD7-2 Base channel 8 BI_DA8+ BI_DA8- BI_DB8+ BI_DB8- BI_DC8+ BI_DC8- BI_DD8+ BI_DD8-3 Base channel 9 BI_DA9+ BI_DA9- BI_DB9+ BI_DB9- BI_DC9+ BI_DC9- BI_DD9+ BI_DD9-4 Base channel 10 BI_DA10+ BI_DA10- BI_DB10+ BI_DB10- BI_DC10+ BI_DC10- BI_DD10+ BI_DD10-5 Base channel 11 BI_DA11+ BI_DA11- BI_DB11+ BI_DB11- BI_DC11+ BI_DC11- BI_DD11+ BI_DD11-6 Base channel 12 BI_DA12+ BI_DA12- BI_DB12+ BI_DB12- BI_DC12+ BI_DC12- BI_DD12+ BI_DD12-7 Base channel 13 BI_DA13+ BI_DA13- BI_DB13+ BI_DB13- BI_DC13+ BI_DC13- BI_DD13+ BI_DD13-8 Base channel 14 BI_DA14+ BI_DA14- BI_DB14+ BI_DB14- BI_DC14+ BI_DC14- BI_DD14+ BI_DD14-9 Base channel 15 BI_DA15+ BI_DA15- BI_DB15+ BI_DB15- BI_DC15+ BI_DC15- BI_DD15+ BI_DD15-10 Base channel 16 BI_DA16+ BI_DA16- BI_DB16+ BI_DB16- BI_DC16+ BI_DC16- BI_DD16+ BI_DD16-Note: Each differential pair has an individual L-shaped ground contact (not shown).

Table 22. Zone 3 connector J30 signals

Row # ab cd ef gh1 +12V_RTM +12V_RTM +12V_RTM +3.3V_IPMC_

LTDRTM_EJECT RTM_

PRESENT~RTM_HOTSWAP~

RTM_ENA~

2 +12V_RTM +12V_RTM +12V_RTM RTM_INTR~ IPMC_I2C_CLK IPMC_I2C_DAT PQ_I2C_CLK PQ_I2C_DAT3 SPI_RTM_

SEL~SLVA2 SLVA1 SLVA0 BLED_CLK BLED_DAT FLED_CLK FLED_DAT

4 RTM_RST~ SPI_CLK SPI_SDO SPI_SDI BASE_MDC BASE_MDIO FAB_XMDC FAB_XMDIO5 EXTCLK_

OUT-EXTCLK_OUT+

RTM_LMP_IRQ5

RTM_LMP_IRQ6

EXTCLK2_IN+ EXTCLK2_IN- EXTCLK1_IN+ EXTCLK1_IN-

6 NTS_8K_OUT+

NTS_8K_OUT-

NTS_19M_OUT+

NTS_19M_OUT-

EXTCLK4_IN+ EXTCLK4_IN- EXTCLK3_IN+ EXTCLK3_IN-

7 FAB:XERTMxTX0+

FAB:XERTMxTX0-

FAB:XERTMxTX1+

FAB:XERTMxTX1-

FAB:XERTMxTX2+

FAB:XERTMxTX2-

FAB:XERTMxTX3+

FAB:XERTMxTX3-

8 FAB:XERTMxRX0+

FAB:XERTMxRX0-

FAB:XERTMxRX1+

FAB:XERTMxRX1-

FAB:XERTMxRX2+

FAB:XERTMxRX2-

FAB:XERTMxRX3+

FAB:XERTMxRX3-

9 BASE:GERTM6TX+

BASE:GERTM6TX-

BASE:GERTM6RX+

BASE:GERTM6RX-

BASE:GERTM7TX+

BASE:GERTM7TX-

BASE:GERTM7RX+

BASE:GERTM7RX-

10 BASE:GERTM4TX+

BASE:GERTM4TX-

BASE:GERTM4RX+

BASE:GERTM4RX-

BASE:GERTM5TX+

BASE:GERTM5TX-

BASE:GERTM5RX+

BASE:GERTM5RX-

Note: Each differential pair has an individual L-shaped ground contact (not shown).


57



Row # ab cd ef gh1 FAB:XE

RTM1TX0+FAB:XERTM1TX0-

FAB:XERTM1TX1+

FAB:XERTM1TX1-

FAB:XERTM1TX2+

FAB:XERTM1TX2-

FAB:XERTM1TX3+

FAB:XERTM1TX3-

2 FAB:XERTM1RX0+

FAB:XERTM1RX0-

FAB:XERTM1RX1+

FAB:XERTM1RX1-

FAB:XERTM1RX2+

FAB:XERTM1RX2-

FAB:XERTM1RX3+

FAB:XERTM1RX3-

3 FAB:XERTM3TX0+

FAB:XERTM3TX0-

FAB:XERTM3TX1+

FAB:XERTM3TX1-

FAB:XERTM3TX2+

FAB:XERTM3TX2-

FAB:XERTM3TX3+

FAB:XERTM3TX3-

4 FAB:XERTM3RX0+

FAB:XERTM3RX0-

FAB:XERTM3RX1+

FAB:XERTM3RX1-

FAB:XERTM3RX2+

FAB:XERTM3RX2-

FAB:XERTM3RX3+

FAB:XERTM3RX3-

5 COM_PCIE_TX1+

COM_PCIE_TX1-

COM_PCIE_RX1+

COM_PCIE_RX1-

NC NC BASE_XMDC BASE_XMDIO

6 NC NC NC NC NC NC NC NC7 FAB:XE

RTMzTX0+FAB:XERTMzTX0-

FAB:XERTMzTX1+

FAB:XERTMzTX1-

FAB:XERTMzTX2+

FAB:XERTMzTX2-

FAB:XERTMzTX3+

FAB:XERTMzTX3-

8 FAB:XERTMzRX0+

FAB:XERTMzRX0-

FAB:XERTMzRX1+

FAB:XERTMzRX1-

FAB:XERTMzRX2+

FAB:XERTMzRX2-

FAB:XERTMzRX3+

FAB:XERTMzRX3-

9 BASE:XERTM1TX0+

BASE:XERTM1TX0-

BASE:XERTM1TX1+

BASE:XERTM1TX1-

BASE:XERTM1TX2+

BASE:XERTM1TX2-

BASE:XERTM1TX3+

BASE:XERTM1TX3-

10 BASE:XERTM1RX0+

BASE:XERTM1RX0-

BASE:XERTM1RX1+

BASE:XERTM1RX1-

BASE:XERTM1RX2+

BASE:XERTM1RX2-

BASE:XERTM1RX3+

BASE:XERTM1RX3-

Note: Each differential pair has an individual L-shaped ground contact (not shown).NC = No connection


58


Internal interfaces

PowerQUICC III COP/JTAG headerThe common on-chip debug processor/Joint Test Action Group (COP/JTAG) port provides a full duplex serial interface for communications between the internal PowerQUICC III development support logic and an external development tool. The development port signals are available for use on a 16-pin surface mount COP/JTAG header located on the SCM. The following tables provide the connector and pin functions for the header.


Row # ab cd ef gh1 BASE:XE

RTM3TX0+BASE:XERTM3TX0-

BASE:XERTM3TX1+

BASE:XERTM3TX1-

BASE:XERTM3TX2+

BASE:XERTM3TX2-

BASE:XERTM3TX3+

BASE:XERTM3TX3-

2 BASE:XERTM3RX0+

BASE:XERTM3RX0-

BASE:XERTM3RX1+

BASE:XERTM3RX1-

BASE:XERTM3RX2+

BASE:XERTM3RX2-

BASE:XERTM3RX3+

BASE:XERTM3RX3-

3 BASE:GERTM18TX+

BASE:GERTM18TX-

BASE:GERTM18RX+

BASE:GERTM18RX-

BASE:GERTM19TX+

BASE:GERTM19TX-

BASE:GERTM19RX+

BASE:GERTM19RX-

4 BASE:GERTM20TX+

BASE:GERTM20TX-

BASE:GERTM20RX+

BASE:GERTM20RX-

BASE:GERTM21TX+

BASE:GERTM21TX-

BASE:GERTM21RX+

BASE:GERTM21RX-

5 NC NC NC NC NC NC NC NC6 NC NC NC NC NC NC NC NC7 NC NC NC NC NC NC NC NC8 NC NC NC NC NC NC NC NC9 NC NC NC NC NC NC NC NC10 NC NC NC NC NC NC NC NCNote: Each differential pair has an individual L-shaped ground contact (not shown).NC = No connection

Table 25. PowerQUICC III COP/JTAG connector J3

Pin # Description Pin # Description1 JTAG_TDO 2 GND3 JTAG_TDI 4 JTAG_TRST~5 QREQ~ 6 PULL-UP7 JTAG_TCK 8 CKSTP_IN~9 JTAG_TMS 10 N/C11 COP_SRESET~ 12 GND13 COPHRESET~ 14 N/C15 CKSTP_OUT~ 16 GND

1

3

5

7

9

11

13

15

2

4

6

8

10

12

14

16

4Internal interfaces

59

ShMC headersThe following headers on the SCM are embedded in the shelf management controller (ShMC) entity.

H3 Serial debug and programming cable header

H1 JTAG debug cable header

J33 Auxiliary function and I/O FPGA programming cable header

H5 Watchdog and power jumper post header

Note: When the SCM is installed on a shelf that includes a non-RadiSys Shelf Manager, the IPMI controller is an IPMC rather than a ShMC.

Table 26. PowerQUICC III COP/JTAG pin functions

Pin # Name Description1 TDO Development serial data out2 QACK Quiescent Acknowledge (GND)3 TDI Development serial data in (Pull-up)4 TRST Test (soft) reset (Pull-down)5 QREQ~ Quiescent request (Pull-up)6 +3.3V +3.3V supply through 1K)7 TCK Development serial clock (Pull-up)8 CKSTP_IN~ Check Stop Input (Pull-up)9 TMS Test mode select input (Pull-up)10, 14 N.C. No Connect11 SOFTREST~ Soft Reset on PowerQUICC III (Pull-up)12 GND Digital GND13 HREST# Hard Reset (Pull-up)14 Key Mechanical Key (No connect)15 CKSTP_OUT~ Check stop output (Pull-up)


60

PowerQUICC III port 80 interfaceThe PowerQUICC III Port 80 interface is used to connect to a RadiSys 84-6000-01 hex display assembly. The interface connects to the GPIO signals of the parallel bus of the PowerQUICC III.

Figure 14. Port 80 debug socket J4

Pin # Description Pin # Description1 +3.3V 2 PRT80D03 PRT80D1 4 PRT80D25 PRT80D3 PRT80D47 PRT80D5 PRT80D69 PRT80D7 MRESET# (reset)11 GND 12 LCS4# (chip select)13 GND 14 N/C

13

75

119

13

42

1086

1412

61

5INITIAL VERIFICATION AND CONFIGURATION

Configuration checklistThis checklist identifies tasks to be done soon after installing the SCM, and points to information that can help with each task. Some of these tasks are optional.Note: When an platform has more than one SCM installed, each SCM must be configured separately.

Task: Where to find information:

_____ Verifying Base interface operation SCM Installation Guide

_____ Verifying Fabric interface operation SCM Installation Guide

_____ Verifying Shelf Manager operation SCM Installation Guide

_____ Adding user accounts and setting passwords page 63

_____ Configuring IP addresses for SCM interfaces page 63

_____Configuring SNMP trap destinations and security access (optional)

page 67

_____ Enabling the SNMP agent (optional) page 68

_____ Enabling the DHCP server (optional) page 69

_____Enabling and disabling the Telnet server and TFTP server (optional)

page 70

_____Enabling and disabling syslog and other services (optional)

page 71

_____Copying a configuration from existing CLI configuration files (optional)

page 72

_____Copying CLI configuration files to a TFTP server (optional)

page 72

_____ Copying any file to the SCM (optional) page 73

5 Initial Verification and Configuration

62

Connecting to the SCM

Serial connection1. Connect the serial cable to the serial port of the SCM (labeled LMP SER). The appropriate

9-pin D-shell to RJ-45 cable was included in SCM package.

2. Connect the cable’s other end to the COM1 or COM2 serial port of an external computer (or to a dumb terminal).

3. Start a terminal-emulator application on the external computer. Specify 115200 baud, 8 data bits, no parity, one stop bit, and no hardware or software flow control. When the terminal emulator is set up and connected, you will receive a login prompt.

Telnet connection1. Connect an Ethernet cable to the SCM’s Ethernet maintenance port (labeled LMP ETH).

2. Connect the cable’s other end to the Ethernet port of a computer that is connected to your network.

3. Configure your computer to a subnet IP address (10.0.0.x, but not 10.0.0.1) and set the netmask to 255.255.255.0.

4. Telnet to IP address 10.0.0.1 (port 23). When the connection is made, you will receive a login prompt.

Logging in to the SCMFrom the ATCA-2210 login prompt, the user name you enter determines whether you begin in the command line interface (CLI) or in the Linux shell.

If you have not yet set passwords, no password is required for the root and admin accounts.

To access the... ...from the... ...enter:CLI ATCA-2210 login prompt admin and password

Linux shell mcli

Linux shell ATCA-2210 login prompt root and passwordCLI linux-shell

5Adding user accounts and setting passwords

63

Adding user accounts and setting passwordsNote: To connect to the SCM, use either Telnet or log in through the Shelf Manager. Establish passwords for the root and admin users during initial setup. You should also create user accounts for those who do not require administrative privileges. For details on account administration, see the documentation for the version of Linux that you are using. To set up another account that goes directly to the CLI, set the login script for the account to /usr/bin/mcli. This mimics how the admin user is set up in the /etc/passwd file.Once logged in, you are not prompted for a password when moving from the CLI to the Linux shell and back.

Rebooting the SCMTo reboot the SCM from the Linux shell, enter:

reboot

To physically initiate a reboot of the SCM, use a pen or narrow instrument (not a pencil) and firmly push the recessed reset button on the front panel. See Effects of rebooting the SCM on page 78.

Configuring IP addressesIP interfaces are configured by default to minimize your work in setting them up. Some of these interfaces are set up to get IP addresses from a DHCP server. Others are assigned unique static addresses by including variables for the logical shelf, physical slot, or logical slot numbers in the IP address. A few interfaces have static, literal IP addresses assigned to them that should not be changed.We recommend that you:

Keep the default IP addresses for certain interfaces. See Keeping default IP addresses for certain interfaces on page 64.

Obtain dynamic IP addresses for interfaces configured for DHCP by connecting the SCM to a network that has a DHCP server and running the DHCP client. See Requesting IP addresses through DHCP on page 64.

Add static, literal IP addresses as alternates to some of the literal addresses already assigned (as subinterfaces). See Adding static IP addresses for management ports on page 64.

For more information and other possibilities, see Pre-defined IP addressing scheme on page 86.


64

Keeping default IP addresses for certain interfacesDo not change the default IP addresses for these interfaces:

eth1 - This update channel is used to communicate with the redundant SCM to synchronize the Shelf Managers. Changing this address would reduce performance of Shelf Manager communications with its peer.

lo - This internal loopback interface is not visible to any network. Changing it may disrupt occasional communications within the SCM between applications and the local management processor (LMP).

For the default addresses, see Table 32 on page 86.

Requesting IP addresses through DHCPThe dtl0 and dtl1 interfaces are configured to request addresses through DHCP, with RadiSys-defined client IDs. To modify the client IDs, see DHCP client on page 88.To get addresses:1. Physically connect the SCM to a network that has a DHCP server.

2. Connect to the SCM through the serial port or with Telnet.

3. Ensure that the interfaces on the Base and Fabric switches that are expected to communicate with the DHCP server are administratively enabled. See the Ethernet Switching Software Reference for details about the CLI port numbering and the no shutdown command that enables ports.

4. Access the Linux shell.

5. Verify that the interfaces have received IP addresses by entering:ifconfig dtl0ifconfig dtl1

Look for the “inet addr” values. If they do not show, wait five minutes and look again.

When the SCM reboots, the addresses will be reused or new ones will be requested.

Adding static IP addresses for management portsThis procedure tells how to configure subinterfaces with alternate IP addresses for certain ports:

eth0:1—subinterface for the front panel Ethernet maintenance port

dtl0:2—subinterface for the Base interface connection to the local management processor (LMP)

dtl1:0—subinterface for the Fabric interface connection to the LMP

To configure alternate IP addresses for the ports:

1. Connect to the SCM’s serial port.

2. Access the CLI.

5Configuring IP addresses

65

3. Select the Base interface switch by entering:base-ethernet

4. Configure the front panel port (eth0) with a new IP address, subnet mask, and (optional) default gateway, using this syntax:serviceport ip <ipaddr> <netmask> [gateway]

This establishes the subinterface eth0:1 with the IP address you assigned.

5. Configure dtl0 with a new IP address, subnet mask, and (optional) default gateway, using this syntax: network parms <ipaddr> <netmask> [gateway]

This establishes the subinterface dtl0:2 with the IP address you assigned.

6. Exit the Base interface by entering:exit

7. Select the Fabric interface switch by entering:fabric-ethernet

8. Configure the Fabric interface connection to the LMP with a new IP address, subnet mask, and (optional) default gateway, using this syntax: network parms <ipaddr> <netmask> [gateway]

This establishes the subinterface dtl1:0 with the IP address you assigned.

9. Exit the switch portion of the CLI by typing:exit

10. When prompted to save changes, say yes (“y”) to save the changes to the SCM’s startup configuration.


66

Per-VLAN interfacesIn some use cases for the SCM, customers may need to route management traffic over more than one VLAN. Figure 15 shows an example of such a use case.

Figure 15. Use case for per-VLAN interfaces

As shown in Figure 15 on page 66, shelf management traffic can be carried over a VLAN that exists only inside the shelf, while higher level management traffic, such as SNMP, can be carried over a VLAN set up for an entire system. This system could consist of multiple shelves.

The SCM supports this use case with per-VLAN interfaces. Per-VLAN interfaces are created using the 802.1Q VLAN driver and configured with the vconfig application. The syntax for adding and removing per-VLAN interfaces is:

vconfig add [interface-name] [vlan_id]vconfig rem [interface-name] [vlan_id]

The following example shows a per-VLAN interface (dtl0.5) being created for a VLAN (VLAN 5). The IPv4 address 192.168.5.1/24 is assigned to the per-VLAN interface:

vconfig add dtl0 5 ifconfig dtl0.5 192.168.5.1 netmask 255.255.255.0

Node module

BaseEthernet

Switchingsoftware

Managementsoftware

ATCA shelf

ATCA-2210 SCM

External system manager

VLAN x used to carry SNMP traffic

VLAN y used to carry shelf management traffic

Node module: Refers to a front module that is not a switch and control module. For example, node modules can include carrier modules, compute modules, storage modules, I/O modules, and line cards.

LMP

5Configuring SNMP trap destinations and security access

67

When setting up your own per-VLAN interfaces, be sure to verify the following once the per-VLAN interface has been created and configured:

the appropriate Base or Fabric interfaces have been configured to participate in the VLAN the traffic being sent from the per-VLAN interface is reaching the desired destination or destinations

For more information on VLANs, see the Ethernet Switching Software Reference.

Configuring SNMP trap destinations and security accessTo configure the SNMP notification (trap and inform) destinations and community names:

1. Connect to the SCM.

2. Access the Linux prompt.

3. Make a backup copy of the configuration file:

cp /etc/snmp/snmpd.conf /etc/snmp/snmpd.conf.orig

4. Edit the configuration file /etc/snmp/snmpd.conf.

5. (Optional) Configure a trap destination by uncommenting a template line and filling in values as follows:

A. Locate the line that begins with “#trap2sink”.

B. Fill in the host IP address and community name in the correct positions as follows:trap2sink <ip_addr> <community_name> 162

C. Uncomment the line by removing the # character, and add more lines as necessary.

6. (Optional) Configure an inform destination by uncommenting a line and filling in values as follows:

A. Locate the line that begins with “#informsink”.B. Fill in the host IP address and community name in the correct positions as follows:

informsink <ip_addr> <community_name> 162

C. Uncomment the line and add more lines as necessary.

7. Configure community names by uncommenting lines and filling in values as follows:

A. Locate the lines that begin with “#com2sec”.The lines that specify the context name as “-Cn fi” configure access to the Fabric interface switch objects. The other lines configure access to all other objects, including Base interface switch objects. The items ending in “RO” or “RW” are the security names.

WARNING! Do not change the context name “fi”, or you will lose access to many Fabric interface switch objects.

B. Change the community names. This step is optional, but highly recommended.C. Uncomment these lines and add more lines as appropriate.


68

8. Save the file. The saved changes take place when you enable or restart the SNMP agent. Changes will also take effect if the agent reads its own configuration file. See Enabling the SNMP agent on page 68 or the CLI command service snmpd on page 111. Changes to the file are saved persistently, so they will be restored after the SCM is rebooted.

Note: If you have installed the Net-SNMP man pages onto a system, you can find more details on the snmpd.conf file options using the command:

man snmpd.conf

Enabling the SNMP agentThis procedure assumes that you already performed the steps in Configuring SNMP trap destinations and security access on page 67.To enable the SNMP agent:


2. Access the CLI.

3. Access blade-management mode by entering:blade-mgmt

4. Enable the master agent by entering:service snmpd

5. Exit blade-management mode by entering:exit

6. Make the change to the agent’s enable status persistent across reboots by entering:copy system:running-config nvram:startup-config

For more information, see the service snmpd command description on page 111 and the copy system:running-config description on page 108.

5Enabling the DHCP server

69

Enabling the DHCP serverThe SCM’s DHCP server is set up to work with a redundant SCM to provide primary and secondary DHCP servers. Configuration files are supplied in /etc as pri.dhcpd.conf.example, sec.dhcpd.conf.example, and common.dhcpd.conf.example, but they must be modified to fit your network.If you want to use the SCM’s DHCP server to provide IP addresses, configure and enable the DHCP server as follows:



3. Change to the directory containing the DHCP server configuration files:cd /etc

4. Copy the example configuration files or your replacement files to the required file names as follows:cp pri.dhcpd.conf.example pri.dhcpd.confcp sec.dhcpd.conf.example sec.dhcpd.confcp common.dhcpd.conf.example common.dhcpd.conf

5. Modify the interface configuration for the SCM interface that will serve addresses. The DHCP server configuration uses dlt0 by default (as specified in the /etc/default/dhcp file), but dtl0 uses a dynamic IP address by default.

6. Verify that the interface that will serve addresses is administratively enabled through the Base or Fabric Ethernet CLI.

7. Modify the configuration files as necessary for your network. Make the address and the peer address match the IP addresses that the two SCMs will use to communicate with each other in both the primary and secondary files. See page 86 for information about the eth1 update channel IP address, which can be used for this purpose.

8. If this SCM will host the secondary DHCP server, change the configuration as follows:

A. Edit the DHCP server initialization script /etc/rc.d/init.d/dhcpd.

B. Locate the FAILOVER_PRIMARY line that appears after the commented Read Me text. The line should look like this:FAILOVER_PRIMARY=1

The default value is 1, indicating the primary, so no changes are needed on the SCM hosting the primary DHCP server.

C. Change the FAILOVER_PRIMARY value to 0 to indicate the secondary. The line should look like this when you are done: FAILOVER_PRIMARY=0

D. Save the file.

9. Create a directory for the lease database file:mkdir -p /rsys/onboot.data/bindmount/var/state/dhcp


70

10. Make the directory persistent:

touch /rsys/onboot.data/bindmount/var/state/dhcp/.bindmount

The directory appears in the file system as /var/state/dhcp, and is not present until the SCM reboots.

11. Create a symbolic link from the /etc/rc.d/rc2.d directory to the startup script:

ln -s ../init.d/dhcpd /etc/rc.d/rc2.d/<link_name>

For example:

ln -s ../init.d/dhcpd /etc/rc.d/rc2.d/S53dhcpd

This causes the DHCP server to be started automatically at SCM bootup. The value of <link_name> determines the order of startup relative to other scripts in the rc2.d directory and should place this service after all other scripts upon which this service depends.

12. Enable the DHCP server immediately by running the startup script:

/etc/rc.d/init.d/dhcpd start

Enabling and disabling the Telnet server and TFTP serverIf you want to restrict access to the SCM, you can disable the Telnet server and the TFTP server, which are enabled by default. These services are controlled by the /etc/inetd.conf file.WARNING! Do not disable the Telnet server until you confirm access to the SCM through other means. For example, the SSH service is enabled by default, but a user account with a password must be set up before SSH can be used.

To enable or disable the Telnet server or TFTP server:



3. Change to the directory containing the DHCP server configuration files:

cd /etc

4. Edit the inetd.conf file as appropriate:

A. To disable the Telnet server, comment out the line beginning with telnet. You can comment out the line by inserting a pound character (#) at the beginning of the line.

B. To disable the TFTP server, comment out the line beginning with tftp.

C. To enable the Telnet server, uncomment the line beginning with telnet (by removing the pound character).

D. To enable the TFTP server, uncomment the line beginning with tftp.

This changes the enable states of the services the next time the SCM boots up.

5. Enable or disable the services immediately by restarting inetd:

/etc/init.d/inetd restart

5Enabling and disabling syslog and other services

71

Enabling and disabling syslog and other servicesIf you want the SCM to log messages to a file, enable the syslog service. You may want to enable other services—such as NTP—that are disabled by default, or to re-enable a service after it has been disabled.To enable a service:

1. Prepare a configuration file for the service with appropriate contents (for example, syslog.conf).

2. Copy the configuration file to the correct location (/etc) on the SCM.

3. Create a symbolic link from the /etc/rc.d/rc2.d directory to the service’s startup script:

ln -s ../init.d/<startup_script> /etc/rc.d/rc2.d/<link_name>

For example:

ln -s ../init.d/syslog /etc/rc.d/rc2.d/S12syslog

This causes the service to be started automatically at SCM bootup. The value of <link_name> determines the order of startup relative to other scripts in the rc2.d directory and should place this service after all other scripts upon which this service depends.

4. Enable the service immediately by running the startup script:

/etc/rc.d/init.d/<startup_script> start

For example:

/etc/rc.d/init.d/syslog start

To disable a service:

1. Change to the /etc/rc.d/rc2.d directory:

cd /etc/rc.d/rc2.d

2. Identify the symbolic link to the appropriate startup script.

For example, listing the details of the syslog link might show:

lrwxrwxrwx 1 root root 16 May 29 08:57 S12syslog -> ../init.d/syslog

3. Disable the service immediately by running the startup script with the stop option:

/etc/rc.d/init.d/<startup_script> stop

For example:

/etc/rc.d/init.d/syslog stop

4. Keep the service from restarting when the SCM reboots by removing the symbolic link:

rm <link_name>

For example:

rm S12syslog

The service can be re-enabled later by adding the link again as describe above.


72

Copying a configuration from existing CLI configuration filesYou can change the configuration of the SCM using CLI configuration files created on another SCM or created on this SCM but stored remotely. The files should be placed on an accessible TFTP server (for example, in the /tftpboot directory). This procedure copies the files to the SCM and makes them the current and persistent configurations.To change to configuration settings created by CLI commands:


2. Access the CLI.

3. Retrieve the Base Ethernet switch configuration file and make it the running configuration:

base-ethernetcopy tftp://<IP address>/<file-1> nvram:startup-config

Executing the command returns you to the master CLI.

4. Retrieve the Fabric Ethernet switch configuration file and make it the running configuration:

fabric-ethernetcopy tftp://<IP address>/<file-2> nvram:startup-config

Executing the command returns you to the master CLI.

5. Retrieve the blade-management configuration file and make it the running configuration (from the master CLI):

copy tftp://<IP address>/<file-3> nvram:startup-configcopy nvram:startup-config system:running-config

The running configuration and the persistent startup configuration has changed to reflect the content of the configuration files.

Copying CLI configuration files to a TFTP serverYou can copy the CLI configuration files from the SCM to a TFTP server. This procedure also saves the current running configuration to be the SCM’s persistent startup configuration, if it wasn’t already.Most TFTP servers require that the target files being copied already exist on the target TFTP server. The target files must also have “global write” permissions. To avoid having to create the files in advance, start the TFTP server with the -c option. If you have the Linux man pages installed on a system, see the tftpd man page for details.To save the current configuration to be the startup configuration and to copy the configuration files to a TFTP server:


2. Access the CLI.

5Copying any file to the SCM

73

3. Save the Base Ethernet switch configuration file to persistent storage and copy it to the server:base-ethernetcopy system:running-config nvram:startup-configcopy nvram:startup-config tftp://<IP address>/<file-1>exit

4. Save the Fabric Ethernet switch configuration file and copy it to the server:fabric-ethernetcopy system:running-config nvram:startup-configcopy nvram:startup-config tftp://<IP address>/<file-2>exit

5. Save the blade-management configuration file and copy it to the server (from the master CLI):copy tftp://<IP address>/<file-3> nvram:startup-configcopy system:running-config nvram:startup-configcopy nvram:startup-config tftp://<IP address>/<file-3>

The configuration now resides on the TFTP server, and will be restored from persistent storage when the SCM is rebooted.

Copying any file to the SCMYou can copy any file to the SCM using TFTP. The files should be placed on an accessible TFTP server (for example, in the /tftpboot directory). To copy a file to the SCM:

1. Change to the directory where the files should be placed on the SCM.

2. Connect to the TFTP server:tftp <IP address of TFTP server>

3. (Optional) For binary files, specify binary mode:bin

4. Retrieve the file:get <myfile>

Repeat this step for each additional file.

5. Exit TFTP:quit


74

75

6SOFTWARE FEATURES

This chapter describes the major software features of the SCM. The tables that follow provide an overview of the management interfaces and features supported on the SCM and where to find additional on them.

Table 27. Management interfaces

Management Interface Described inCommand Line Interface (CLI) This manual, starting in Using the Command Line Interface on page 105. Details on

the some portions of the CLI appear in other manuals as identified on page 106. SNMP agent This manual in SNMP Agent Support on page 95. In addition, the Ethernet Switching

Software Reference describes controls for Layer 2 trap notifications.APIs (application programming interfaces) This manual in API Overview on page 101.HPI (Hardware Platform Interface) The Shelf Management Software Reference.IPMI (Intelligent Platform Management Interface) The Shelf Management Software Reference.

Table 28. Software features

Feature Description Procedures and Command Reference

Management Interfaces

PICMG-compliant Shelf Manager page 76 Shelf Management Software Reference

CLI, HPI, IPMI

Ethernet switching (VLANs, spanning tree, link aggregation, Class of Service queuing, storm control, and more)

page 76 and Ethernet Switching Software Reference

Ethernet Switching Software Reference

CLI, API, SNMP MIBs for certain aspects

IP routing page 76 and IP Routing Software Reference

IP Routing Software Reference

CLI

Linux and boot loader page 77 Linux and U-Boot documentation

standard Linux and U-Boot

RAM and file system usage (JFFS, persistent files) page 81 N/A standard LinuxIPv4 interface and Ethernet switch port identifiers page 84 N/A N/APre-defined IP addressing scheme (DHCP client, defaults) page 86 page 63 and page 86 standard LinuxDHCP server page 90 page 69 standard LinuxTelnet and SSH page 90 N/A standard LinuxNetwork timing subsystem (NTS) software (system clock) page 90 and page 102 page 112 CLI, APINTP server (time of day) page 93 N/A standard LinuxModule state management page 93 and page 102 API Reference APIFlash and file system management (moving files, persistency) page 93 and page 103 page 107 and API Reference CLI, APISoftware image management (versions) page 93 and page 103 API Reference API

6 Software Features

76

Shelf ManagerThe SCM provides a PICMG-compliant Shelf Manager and the IPMI infrastructure for managing the shelf and FRU data. The Shelf Manager keeps track of alarms and provides E-Key services to modules to authorize them to power up (if appropriate) when inserted into the shelf. The Shelf Manager also corrects many error conditions and controls overheating situations by increasing fan speeds. The platform-management CLI, the standard hardware platform interface (HPI), and an example HPI application provide interactive and programmatic management access to the shelf. For more information, see the Shelf Management Software Reference.

Ethernet switchingThe SCM provides Ethernet switching software that runs on both the Base and Fabric switches. The software contains support for many Layer 2 switching protocols such as virtual LANs, spanning tree, and Class of Service queuing. The SCM provides a CLI, an API, and some SNMP MIB support to manage the switch. For information about the Layer 2 protocols supported and the CLI that controls them, see the Ethernet Switching Software Reference. For information about the SNMP MIB support, see Chapter 7, SNMP Agent Support, on page 95 and the Ethernet Switching Software Reference. For information about the API, see Chapter 8, API Overview, on page 101.

IP routingThe SCM includes IPv4 routing features as a licensed software option. These unicast and multicast routing features are supported on the Fabric interface only. With the IP routing feature set enabled, the SCM supports:

OSPF (open shortest path first) version 2

RIP (router information protocol) version 1 and 2

DVMRP (distance vector multicast routing protocol)

VRRP (virtual router redundancy protocol)

The SCM provides CLI commands to manage the Layer 3 configuration. For more information on the IP routing features and the CLI, see the IP Routing Software Reference.

Firmware and software upgrades page 94 and page 103 Firmware and Software Upgrades

Linux utilities, API

Logging service (syslog errors and messages) page 94 page 71 and page 94 standard Linux, CLIE-Key authorization and notification page 94 N/A None

Table 28. Software features (continued)

Feature DescriptionProcedures and Command Reference

Management Interfaces

6Linux and boot loader

77

Linux and boot loaderThe installed SCM software includes the Wind River Platform for Network Equipment, Linux edition, which supplies the Linux 2.6 kernel along with user space applications and utilities. For information on Linux, see http://www.windriver.com or http://www.linux.org/docs.The SCM uses the RadiSys onboot utility to load its software into RAM so it can run in a way that resembles diskless Linux. For details, see RAM and file system usage on page 81.The software includes the Das U-Boot Universal Bootloader, version 1.1.4, which initializes the SCM hardware and the Linux operating system. When the software image is upgraded, U-Boot should also be upgraded to ensure it includes the most current changes. When U-Boot is reflashed, the image and the environment variables are reset to the factory defaults.U-Boot also provides an environment for debugging, changing the flash contents, and scripting. The primary and secondary boot devices each contain a copy of the U-Boot image and its environment variables. When U-Boot is powered on, the IPMC starts a watchdog timer called the corrupt flash detection (CFD) watchdog timer. If U-Boot does not issue a “Stop Watchdog timer” command to the IPMC, the CFD timer will expire, and U-Boot will switch to the alternate U-Boot flash bank and reset the LMP. If the CFD timer expires, the alternate bank will always be used.For information on diagnosing bootup issues with U-Boot see Diagnosing bootup issues by disabling the BMC watchdog timer on page 130. For more information on U-Boot, see http://u-boot.sourceforge.net.

Linux servicesAll standard services of this Linux version are included on the SCM. While RadiSys provides additional management interfaces and documentation for some services, the following are included without additional RadiSys support:

NFS clientFTP clientTFTP server and clientWatchdog daemonBootP

http://www.windriver.com

http://www.linux.org/docs

http://u-boot.sourceforge.net

6 Software Features

78

Effects of rebooting the SCMRebooting the SCM causes the following to occur:

Base and Fabric interface connections from node modules may change over to the redundant SCM and possibly lose traffic.

Spanning tree topology changes occur, possibly resulting in lost traffic.

If the rebooted SCM hosts the active Shelf Manager or master network timing subsystem (NTS), they will fail over. For more information, see the Shelf Management Software Reference and Network timing subsystem (NTS) software on page 90.

The SCM terminates all of its Telnet sessions.

Configuration settings revert to those saved as startup configurations (configuration changes not saved are lost). For more information, see Saving configurations persistently on page 107.

The file system is recreated, so files and directories revert to those saved persistently (changes not made persistent are lost). For more information, see File persistence on page 81.

Linux promptThe default Linux shell prompt can be very informative and follows this format:

<userName>@<hostName>@<LCAB>-<LCHAS>-<SLOT>:<pwd><symbol>

where

<userName> = the name of the user currently logged in<hostName> = the system name, ATCA-2210<LCAB> = the logical cabinet number stored in the Shelf FRU Information<LCHAS> = the logical chassis (shelf address) stored in the Shelf FRU Information<SLOT> = the physical slot number in the chassis<pwd> = the present working directory<symbol> = # for the root user or $ for any other user.

For example:

root@ATCA-2210@1-3-8:/etc/sysconfig#

If the Shelf FRU Information including the logical cabinet cannot be read at startup, it is considered an error condition, and the prompt uses this format:

<userName>@<hostName>@ DEBUGMODE :<pwd><symbol>

For information about debug mode, see the troubleshooting information in the Shelf Management Software Reference.

The prompt is defined in /etc/profile.

6Flash memory usage

79

Flash memory usageFigure 16 shows the layout of the flash memory devices used in the SCM. Each flash memory device consists of 64MB flash parts. The part of the flash memory selected by the IPMC as the boot device is located in the memory map at address FC000000h. Flash part 0 is the primary boot device and is located at address E0000000h. Flash part 1 is the secondary boot device and is located at address E4000000h.

Figure 16. Flash device memory layout

Unused/misc. registers

Flash part 0primary boot

E000 0000h

E800 0000h

F000 0000h

F800 0000h

FFFF FFFFh

64MB

64MB

128MB

64MB

64MB

Flash part 1secondary boot

Flash partselected by IPMCas the boot device

FC00 0000h

128MBUnused

Unused

E400 0000h

6 Software Features

80

Figure 17 shows the layout of the flash memory device used in the SCM. The JFFS is used to store configuration, log, and other non-volatile files. The JFFS image comes from the active part of the memory. The address of the JFFS image is FC000000.

Figure 17. Flash memory device map

ubootU-Boot

Kernel andfile system image

JFFS Image(Unused in 2nd Part)

JFFS image

512KB

48MB

15.5MB

Lower memory

6RAM and file system usage

81

Figure 18 shows the layout of memory technology device (MTD) partitions used in the SCM. These partitions are used to load items from the flash and also to program the flash memory.

Figure 18. Flash memory device partitions

RAM and file system usageThe SCM uses the RadiSys onboot utility to load its software into RAM so it can run in a way that resembles diskless Linux.The SCM runs its software from the RAM disk and uses a JFFS2 (journaling flash file system version 2) for persistent file storage. At SCM bootup, U-Boot copies the kernel and initial RAM disk (initrd) from flash memory into RAM. Then Linux converts the initial RAM disk into a normal RAM disk mounted as the root file system. During the boot process, certain files in the JFFS are linked (bind mounted) to the RAM disk, causing the files in the JFFS to appear in RAM. Certain other files are copied (overlayed) to RAM. The linking and copying of files to RAM occurs before the initialization scripts are run.

File persistenceBecause the SCM file system is recreated in RAM at each reboot, changes made to files must be saved persistently to the JFFS or be lost upon reboot. Changes to a RadiSys-defined set of configuration files can be preserved by deliberately saving them persistently through the CLI or API. (See Saving configurations persistently on page 107 and Flash and file system management on page 93.) Other files must be included in a persistency directory.

mtd3

JFFS2

mtd4

Primary kerneland file system

Primary U-Boot

Secondary U-Boot

mtd6

mtd7

mtd0

mtd1

E000 0000h

E400 0000h

E800 0000h

mtd2

Secondary kerneland file system

mtd5JFFS2

6 Software Features

82

The persistency directories include files that should replace existing files or be added to the RAM file system upon bootup. The types of persistency directories are:

Bind mount. The files are linked to the RAM disk, and any changes made to these files in RAM are persistent in the JFFS and relinked in RAM upon reboot. Use the bind mount directory for files that will change over time and must be persistent.

The bind mount directory is /rsys/onboot.data/bindmount. The contents of the directory are linked to the RAM file system at the root level (/). For example, /rsys/onboot.data/bindmount/sbin/make-client-id is linked to /sbin/make-client-id.

Overlay. The files are copied to the RAM disk, and any changes made to these files in RAM are lost upon reboot. Use the overlay directory for files such as executables that are normally not changed. Overlaying is typically used to replace existing files.

The overlay directory is /rsys/onboot.data/overlay. The directory’s contents are recreated at the root level (/) in RAM. For example, /rsys/onboot.data/overlay/usr/bin/myfile.txt replaces the /usr/bin/myfile.txt file in RAM.

The JFFS partition is limited to 15.5 MB. Bind mounting and overlaying should be used conservatively, because the partition must hold the factory-installed persistent files, your persistent files, configuration files, and any changes made to the persistent files in RAM. In most cases, individual files should be made persistent instead of whole directories.The factory default configuration includes a number of files and directories that are bind mounted and overlayed. You can view the current overlay and bind mount lists of persistent files—generated each time the SCM boots—in /etc/version.All of the /etc directory, which includes all configuration files generated by the CLI, is bind mounted for persistence. In some portions of the CLI, before you request that the configuration be saved or copied persistently, the configuration change exists only within the application. When you request that the current configuration be made persistent, the configuration file is written to the RAM disk, which is also saved to the JFFS because of the bind mount settings. See Saving configurations persistently on page 107.

Adding a persistent directoryThis procedure explains how to bind mount a directory to the RAM file system. This procedure assumes that the original directory exists in RAM on the SCM, but directories can be moved directly to the persistency directories from another system.To make a new or existing directory and its contents persistent:

1. Copy the directory and its contents into the bind mount tree:

cp -a <path><dir> /rsys/onboot.data/bindmount<path><dir>

For example, when <path>=/var/lib/ and <dir>=misc:

cp -a /var/lib/misc /rsys/onboot.data/bindmount/var/lib/misc

6RAM and file system usage

83

2. Indicate that the directory should be made persistent and not the individual files inside it by creating a .bindmount file in the directory:touch /rsys/onboot.data/bindmount<path><dir>/.bindmount

For example, when <path>=/var/lib/ and <dir>=misc:touch /rsys/onboot.data/bindmount/var/lib/misc/.bindmount

3. Cause this persistency configuration to take effect by rebooting the SCM:

reboot

After reboot, the directory and its contents immediately appear in RAM. After this point, any changes within a bind mounted directory are saved persistently. Tip: If you accidentally made an empty directory persistent and overwrote a directory in RAM that had valuable contents, you can retrieve the contents by unmounting the directory (with the umount command). Then you can repeat the above procedure. If you have rebooted the SCM, your contents are lost.

Adding a persistent fileThis procedure explains how to bind mount a file to the RAM file system. Overlaying a file is very similar, but instead uses the overlay directory (/rsys/onboot.data/overlay). This procedure assumes that the original file exists in RAM on the SCM, but files can be moved directly to the persistency directories from another system.To make a file persistent:

1. Copy the file into the bind mount tree:cp -a <path><file> /rsys/onboot.data/bindmount<path><file>

For example, when <path>=/sbin/ and <file>=make-client-id:cp -a /sbin/make-client-id /rsys/onboot.data/bindmount/sbin/make-client-id

2. Cause this persistency configuration to take effect by rebooting the SCM:reboot

After reboot, the file appears in RAM and any changes are saved persistently.

Notes: The procedure for overlaying a file is very similar, but the overlay directory (/rsys/onboot.data/overlay) is used instead. In addition, any changes to an overlayed file are not saved persistently after reboot.

Modifications made to the JFFS2 partition are currently not preserved across software upgrades. Track any changes made directly to the file system (bypassing the CLI), so they can be reapplied following an upgrade.

6 Software Features

84

IPv4 interface and Ethernet switch port identifiersThis section describes the IPv4 interfaces on the local management processor (LMP) and the Ethernet switch port interfaces on both the Base and Fabric switches. An SNMP ifIndex is listed for each interface, and a CLI slot/port number is listed for each Ethernet switch port interface.Table 29 lists the LMP IPv4 interfaces. Table 30 and Table 31 list the Ethernet switch port interfaces.

Table 29. LMP IPv4 interface identifiers and descriptions

LMP IPv4 Interface SNMP ifIndex Descriptioneth2 1 Ethernet connection to the Base switch port 2/2.lo 2 Software loopback interface.eth0 3 Ethernet connection to the SCM front panel maintenance port labeled LMP ETH.eth1 4 Update channel between the two slots designated for SCMs, used by the Shelf Manager to

synchronize with its peer. dtl0 5 Pseudo-Ethernet connection to the Base switch, implemented over PCI. dtl1 6 Pseudo-Ethernet connection to the Fabric switch, implemented over PCI.

Table 30. Base Ethernet switch port identifiers and descriptions

Base Switch Port SNMP ifIndex CLI Slot/Port DescriptionTo front slots 1000 + <slotnum>

For example, 1003 for slot 3.

0/<slotnum>For example, 0/3 for slot 3.

Backplane Base Ethernet connections to front slots, where <slotnum> is the physical chassis slot number.

To front panel ports 1021 1/5 10 Gbps Base Ethernet connection to front panel port 1/5.1022 1/6 1 Gbps Base Ethernet connection to front panel port 1/6.1023 1/7 1 Gbps Base Ethernet connection to front panel port 1/7.1024 1/8 1 Gbps Base Ethernet connection to front panel port 1/8.1025 1/9 1 Gbps Base Ethernet connection to front panel port 1/9.

To other SCM 1033 2/1 Backplane Base Ethernet connection to the other, redundant SCM (AdvancedTCA 3.0 Base channel 2).

To LMP 1034 2/2 Base Ethernet connection to the LMP eth2 interface.To non-RadiSys Shelf Manager

1036 2/4 Base Ethernet connection to a non-RadiSys Shelf Manager (AdvancedTCA 3.0 Base channel 1). This is used when the SCM is in a non-RadiSys shelf.

To RTM 1053 3/5 Base Ethernet connections to the RTM through the zone 3 connector.1054 3/61055 3/71056 3/8

To LMP 1065 N/A Base pseudo-Ethernet connection to the LMP dtl0 interface, implemented over PCI.

6IPv4 interface and Ethernet switch port identifiers

85

Link aggregation connections

1066 4/1 Base Ethernet link aggregation (LAG) logical port-channels. These exist only if created by users and are numbered in the order in which they were created.

1067 4/21068 4/31069 4/41070 4/51071 4/61072 4/71073 4/8

Table 31. Fabric interface Ethernet interface identifiers and descriptions

Fabric Switch Port SNMP ifIndex CLI Slot/Port DescriptionTo front slots 2000 + <slotnum>

For example, 2005 for slot 5.

0/<slotnum>For example, 0/5 for slot 5.

Backplane Fabric Ethernet connections to front slots, where <slotnum> is the physical chassis slot number.

To front panel ports 2017 1/1 10 Gbps Fabric Ethernet connection to front panel port 1/1.2018 1/2 10 Gbps Fabric Ethernet connection to front panel port 1/2.2019 1/3 10 Gbps Fabric Ethernet connection to front panel port 1/3.2020 1/4 10 Gbps Fabric Ethernet connection to front panel port 1/4.

To RTM 2049 3/1 Fabric Ethernet connections to the RTM.2050 3/22051 3/3

To LMP 2065 N/A Fabric pseudo-Ethernet connection to the LMP dtl1 interface, implemented over PCI.

Link aggregation connections

2066 4/1 Fabric Ethernet link aggregation logical port-channels. These exist only if created by users and are numbered in the order in which they were created.

2067 4/22068 4/32069 4/42070 4/52071 4/62072 4/72073 4/8

VLAN routing interfaces N/A 5/1 to 5/28 Fabric Ethernet VLAN router interfaces. These exist only for VLANs on which routing is enabled. They are numbered in the order in which the VLANs had routing enabled.

Table 30. Base Ethernet switch port identifiers and descriptions (continued)

Base Switch Port SNMP ifIndex CLI Slot/Port Description

6 Software Features

86

Pre-defined IP addressing schemeThe SCM supports a number of IP interfaces that are used as management interfaces to the module. The interfaces are configured by default to minimize your work in setting them up. This section describes the default IP addresses, the pre-defined address assignment scheme, and the options for customizing address assignment if necessary.For a basic overview and the initially recommended steps, see Configuring IP addresses on page 63.

Default IPv4 addressesTable 32 summarizes the IPv4 interfaces and their default IP addresses.

Table 32. Default IP addresses and address usage for IP interfaces

Interface Name

Default IPv4 Address/Subnet Mask

Description Interface and IP Address Usage

lo 127.0.0.1/8 Software loopback interface on the LMP.

IPv4 datagrams sent to this interface by the LMP are received immediately by the LMP. Do not change this address.

eth0 10.0.0.1/24 Ethernet maintenance port labeled LMP ETH on the SCM front panel.

This interface can be used as a maintenance port for the entire ATCA shelf. From an attached remote console, a user can update the SCM’s firmware and software, and also perform configuration and service of other modules that have IPv4 connectivity to the SCM.This private IPv4 address allows for easy connectivity from a service terminal connected to it, provided that the terminal is also on the 10.0.0.0/24 subnet.If IPv4 Masquerade is enabled on the SCM, this allows the terminal to easily communicate with other modules beyond the SCM’s switch fabric, since the SCM translates the private addresses on the 10.0.0.0/24 network to the public or private address assigned on its dtl0 interface.

eth0:0 10.1.<LCHAS>.<SLOT>/16(where <LCHAS> is the logical chassis and <SLOT> is the physical slot)

Subinterface. This subinterface allows users and validators to use a shared LAN segment to maintain concurrent IPv4 connectivity to SCMs via their eth0:0 subinterfaces. The subinterface also enables multiple SCMs to be networked together via their eth0 physical interfaces without reconfiguring the eth0 IP addresses.

eth0:1 None, interface not created Subinterface. This subinterface is configured through the Base interface CLI using the serviceport ip command. See page 64.

eth1 10.0.1.<LSLOT>/24(where <LSLOT> is the logical slot)

PICMG 3.0 update channel to the redundant SCM, connecting logical slots 1 and 2.

This interface is used by the Shelf Manager to synchronize with its peer on the redundant SCM. The interface could also be used for management.Changing this address reduces performance of Shelf Manager synchronizations, which then use a slower path.

dtl0 Dynamically assigned by DHCP PCI bus connection between the Base switch and the LMP.

This is a pseudo-Ethernet interface.

dtl0:0 10.2.<LCHAS>.<SLOT>/16 Subinterface. This subinterface can be used to configure the SCM or the shelf.

6Pre-defined IP addressing scheme

87

Static IP address assignment configuration file setupThe static IP address assignments are determined by configuration files.The /etc/sysconfig/network-devices directory contains the configuration files that determine the IP assignments at SCM bootup. Files called template.ifconfig.<interfaceName> are interpreted and used to create ifconfig.<interfaceName> files. The interpretation process replaces variables with appropriate values. Interfaces that receive literal IP addresses or that use DHCP have just an ifconfig.<interfaceName> file and no template file, because there is nothing to interpret.The IP address configuration files are saved persistently because the /etc directory is bind mounted as a factory default setting.

dtl0:1 <ShelfMgrIPAddr>/24,which is 192.168.16.17/24 by default

Used by the LMP to communicate with the Shelf Manager through the Base interface.

This subinterface is dedicated to providing access to the Shelf Manager server. This subinterface exists only on the SCM with the active Shelf Manager, and is enabled and disabled by the Shelf Manager server. This setup allows the system manager to always reach the Shelf Manager, regardless of which SCM is active at a given time. The Shelf Manager IP address is stored in the Shelf FRU Device and can be configured via the HPI example application, hpiapp. Updating the Shelf Manager IP address immediately updates the address of dtl0:1, bringing the interface down and then up with the new address.

dtl0:2 None, interface not created Subinterface. This subinterface is configured through the Base interface CLI using the network parms command. See page 64.

dtl1 Dynamically assigned by DHCP PCI bus connection between the Fabric switch and the LMP.

This is a pseudo-Ethernet interface.

dtl1:0 None, interface not created Subinterface. This subinterface is configured through the Fabric interface CLI using the network parms command. See page 64.

Table 32. Default IP addresses and address usage for IP interfaces (continued)

Interface Name

Default IPv4 Address/Subnet Mask Description Interface and IP Address Usage

6 Software Features

88

Variables used to assign IP addressesTable 33 explains the use of specific variables to assign static IP addresses.

Note: The factory default Logical Shelf address is 255. If two shelves are connected which both have factory default settings, there will be multiple SCMs with the same eth0:0 IP address (for example, 10.1.255.7 for the slot 7 SCM and 10.1.255.8 for the slot 8 SCM). To avoid this, add a shelf to the network and commission it with a new shelf address before adding another shelf.

DHCP clientSome interfaces are set up to use the DHCP client (/sbin/dhclient) to request an IP address from a DHCP server on the network. The shelf address and the physical slot number are used to derive the DHCP client identifier string sent in DHCP client requests. The client ID is formatted as a colon-delimited hexadecimal string containing the logical shelf and physical slot. The ID always begins with a null character (for example, 00:34:2d:38 for logical shelf 4 and physical slot 8). To modify the format of the client ID, edit the /etc/sysconfig/network-devices/make-interfaces script. The client ID appears as the variable CLIENT_ID in /etc/template.dhclient.conf and is replaced by the actual value in the dhclient.conf file.

As the SCM boots up, the DHCP client works in the background requesting IP addresses. By default, the requests from the DHCP client are no more than 5 seconds apart for a total duration of 300 seconds or until an address is received. If no DHCP server has responded to the requests after 300 seconds, a timeout of 5 minutes occurs before the DHCP client starts the request sequence again.

Table 33. Use of variables in IP address templates

Variable How it is useful How the value is set and obtained<LCHAS> - logical chassis

The logical shelf ID, or shelf address, can identify the physical location of the shelf within a central office. The ID can also be used to generate unique IP addresses and prompts to help users distinguish between shelves.

RadiSys assigns the default value of 255 to the shelf.During the commissioning process, a shelf should be assigned a shelf address that is unique within a particular central office, or at least within the domain in which the SCM communicates. This information is stored in the Shelf FRU Device on the shelf, and is read during host processor initialization. If the shelf address does not conform to the RadiSys-defined format (described in the Shelf Management Software Reference), the SCM cannot read it, and the default value is used.

<SLOT> -physical slot in shelf

The physical slot location uniquely identifies this SCM’s location in the shelf.

The slot number is read during host processor initialization.

<LSLOT> -logical slot

The logical slot is an easy way for the SCM to distinguish itself from the other SCM.

The logical slot identifies the position of the SCM in the chassis in relation to the other SCM. The lower-numbered physical slot is logical slot 1. The higher numbered slot is logical slot 2.

6Pre-defined IP addressing scheme

89

You can configure the values for the maximum request interval, the timeout period, and the length of time before retrying the request sequence in the /etc/template.dhclient.conf file. If you have Linux man pages installed on a system, refer to the dhclient.conf man page for more information.To request addresses for the first time, see Requesting IP addresses through DHCP on page 64.

Customizing IP address assignments

Addresses to avoid changingWe recommend keeping the default addresses for:

loeth1

In addition, it may not be necessary to change the other statically configured IP addresses. For the default values, see Table 32 on page 86. For further explanation, see Keeping default IP addresses for certain interfaces on page 64. For a procedure, see Adding static IP addresses for management ports on page 64.

Options for customizationUsers who wish to further change the default configuration of the IPv4 interfaces have these choices:

Change the /etc/sysconfig/network-devices/template.ifconfig.<interface> template files to modify the variables used in assigning addresses.

Change the /etc/sysconfig/network-devices/ifconfig.<interface> files to change the static address assignments.

Change the external DHCP server configuration to change the dynamic address assignments based on the client identifier.

Change the run level startup script (/etc/rc.d/init.d/network) to modify the default behavior for configuring the static and dynamic address assignments.

Users who wish to change the DHCP client behavior have these choices:

Change the /etc/template.dhclient.conf files to change the protocol timers and options used by the DHCP client. See DHCP client on page 88.

Change the run level startup script (/etc/rc.d/init.d/network) to modify the default behavior for launching the DHCP client application.

Change the DHCP client application itself and the scripts that invoke it.

6 Software Features

90

DHCP serverThe SCM supports the standard Linux DHCP server, which can supply IP addresses for blades and other network equipment on the same network as the SCM. For more information on setting up the DHCP server, see Enabling the DHCP server on page 69.

Telnet and SSHThe SCM supports the standard Linux Telnet server and client to enable connections to and from the SCM. The Telnet server daemon is enabled by default, responding to requests on TCP port 23 and bringing you to the SCM login prompt. See Logging in to the SCM on page 62.The SCM also supports secure shell (SSH) client and server, along with the rest of the full OpenSSH distribution. SSH allows secure connections to and from the SCM, and requires a password. SSH protocol version 2.0 is supported. The SSH server is enabled by default, responding to requests on TCP port 22.To disable the Telnet server, see Enabling and disabling the Telnet server and TFTP server on page 70. To disable the SSH server, see Enabling and disabling syslog and other services on page 71.

Network timing subsystem (NTS) softwareThe SCM’s network timing subsystem provides centralized system timing based on the synchronization clock interface defined in the PICMG® 3.0 Revision 2.0 AdvancedTCA Base Specification.The SCM can receive timing reference signals from the network interfaces and external clock sources and can provide a network-synchronized system clock to the other interconnected blades in the shelf. Working together with capable node blades, the SCM can provide uninterrupted (hitless failover) network timing for the whole shelf.Primary reference sources (PRSs) provide the highest level of time and frequency accuracy for telecommunications networking. International standards require that all clocks in a synchronization network be referenced or traceable to a PRS. PRS timing can come from a variety of sources, such as a GPS, synchronous supply unit (SSU), or building integrated timing supply (BITS). These sources distribute precise time and frequency information to equipment in central offices, enabling voice, video, and data traffic to be transported reliably and efficiently throughout the network.

Clock bus signalsThe SCM requires that the backplane synchronization clock interface buses be used as follows:

CLK1A and CLK1B - 8 kHz system clock distributed by an SCM to node blades.

CLK2A and CLK2B - 19.44 MHz system clock distributed by an SCM to node blades.

CLK3A and CLK3B - 8 kHz, 1.544 MHz, 2.048 MHz, or 19.44 MHz network timing reference input from node blades.

6Network timing subsystem (NTS) software

91

The SCM allows flexibility in the usage of the clock buses, allowing a clock bus’s A and B halves to be driven by two separate sources. For example, to provide redundancy, the CLK1 and CLK2 pairs can be split with the A sides being driven by one SCM and the B sides by the other SCM. For details, see System clock redundancy on page 92.

Selectable timing sourcesThe SCM can select reference timing sources from:

Synchronization clock interface. CLK3A and CLK3B can each carry a timing signal from a node module.

RTM. The RTM may have external unframed sources or framed sources that it de-frames and provides to the SCM. The SCM receives these reference clock inputs at multiples of 8 kHz up to 19.44 MHz.

Update channel interface. The SCM receives high speed and frame alignment clock signals from the redundant SCM over dedicated links.

System clock outputsThe SCM can output network-synchronized system clocks to:

Synchronization clock interface. CLK1A, CLK1B, CLK2A, and CLK2B can each carry a system clock signal to the node blades.

RTM. The SCM drives system clock signals to the RTM that may be used by the RTM to generate framed or unframed shelf outputs. The SCM supplies the clocks at 8 kHz, 19.44 MHz, and either 1.544 MHz or 2.048 MHz.

Update channel interface. The SCM outputs high speed and frame alignment clock signals to the redundant SCM over dedicated links.

Source prioritiesThe SCM can be configured with source priority levels for each input clock. The SCM can automatically respond to a faulty clock source by selecting the next best source based on configured priorities. Alternatively, an application such as a system manager can explicitly specify the input to use as the reference clock source.

Hitless failover abilityThe SCMs and the node modules work together to provide hitless failover network timing for the ATCA shelf. The SCM provides a consistent system clock, even when its primary reference source becomes invalid or it transitions to a secondary source. Two SCMs act in parallel to provide redundant in-phase system clock sources for the node blades. In case of an interruption to an SCM's system clock output, the node blades must also maintain stable clocking by switching over to the system clock source on the redundant SCM. All of these safeguards together provide for hitless failover for the entire shelf.

6 Software Features

92

Transition to a secondary input timing sourceThe SCM monitors incoming timing sources and qualifies their signals against network timing standards. Timing sources are selected based on user-assigned priorities. The SCM locks to its primary reference and uses it to distribute precise, network-synchronized, 8 kHz and 19.44 MHz system clock signals to the node blades.If the SCM's primary reference clock fails or is out of capture range, the SCM selects a secondary source based on the user-configured priorities. The internal digital PLL (phase-locked loop) immediately enters digital holdover mode and then locks to the new timing source. Digital holdover mode maintains system clock stability while not locked to a timing source. The mode accomplishes this by using consistent timing from the SCM's 12.8 MHz local oscillator in conjunction with stored, averaged frequency data collected while locked to a qualified source.If all incoming timing references become disqualified, the PLL immediately enters digital holdover mode to maintain a stable system clock until it identifies and locks to a qualified source.

System clock redundancyBy default, the pair of redundant SCMs are configured to provide hitless failover, with one acting as the network timing master and one as a slave. The slave SCM locks to the master’s system clock output, based on the clock signals received through the update channel. Since the slave can output the same signals as the master, seamlessly redundant clocking is possible if you configure your system as follows:

The master SCM drives CLK1A and CLK2A, and the slave drives CLK1B and CLK2B.

Node blades prioritize the master SCM clocks as top priority, with the slave clocks as second priority.

With this configuration, the slave provides an immediate failover alternative without any clock cycle loss. For example, if the master SCM clock on CLK1A becomes invalid, the node module can switch to CLK1B and lock to the new source.The roles of the SCMs can be reversed, or they can be autonomous, where they do not exchange timing information.

Node module transition to redundant system clockHitless failover of timing relies on node blades that can seamlessly transition to a secondary system clock upon failure of the primary clock. If the node module’s primary reference clock (master SCM system clock) becomes disqualified, the node blades must remain active and maintain stable clocking while transitioning to the slave SCM system clock.

6NTP server

93

E-Key management of clock busesThe SCM can restrict entities from driving each clock bus by requiring E-Key authorization. With the SCM configured to require authorization, blades and applications must request permission to drive a clock bus. The Shelf Manager handles the details of enabling the bus by granting tokens to requesters. A system manager or custom application can control the granting of tokens through the NTS API. See NTS API on page 102 and network-clock-select request-bused-resource on page 117.

NTP serverThe SCM supports the Linux standard NTP server, which allows it to provide time of day services to other blades or systems. The NTP server is disabled by default. To enable it, see Enabling and disabling syslog and other services on page 71.

Module state managementThe SCM provides an API to display and set the status of the SCM’s rear transition module. For more information on the API, see Module state management API on page 102.Additionally, the SCM provides control over the hot-swap state of any shelf FRU through the platform-management CLI. See the Shelf Management Software Reference for details.

Flash and file system managementThe SCM provides CLI commands and an API to help you manage configuration files and other files on the SCM. These help you to:

Copy RadiSys-defined configuration files to new locations, on or off the SCM.

Retrieve configuration files from another system to the SCM.

Save the current configuration persistently.

Reset the configuration to the one in persistent storage (the startup configuration).

The RadiSys-defined configuration files are those operated on by the CLI. They cover all configuration performed in the CLI. For more details on performing these functions through CLI commands see Saving configurations persistently on page 107 and Master CLI commands on page 108.For more information about the API, see Flash and file system management API on page 103.

Software image managementThe SCM provides an API to report the software images residing on the SCM. For more information about the API, see Software image management API on page 103.

6 Software Features

94

Firmware and software upgradesThe SCM provides Linux utilities to help you upgrade the firmware and the software. An API can also handle certain aspects of software upgrades, and a software upgrade example application is provided.For details on the Linux utilities, see Firmware and Software Upgrades, which is included in the Promentum software distribution. For an introduction to the API, see Flash and file system management API on page 103.

Logging serviceThe SCM supports the Linux standard syslog service, which allows it to collect messages from various SCM software components and control the output. By default, the module-wide syslog service is disabled, but logging is enabled for several components. Messages from the components are ignored unless the module-wide service is enabled. The module’s syslog configuration also determines the destination for the messages. See Enabling and disabling syslog and other services on page 71.The Ethernet switching software is an example of a component that enables syslog generation by default. Generation of syslog messages is controlled separately on the Base and Fabric switches. To disable it or re-enable it, see the logging syslog command in the Ethernet Switching Software Reference.

E-Key authorization and notificationThe SCM’s Shelf Manager provides E-Key authorization for blades to enable and disable the backplane ports between blades and modules (such as AMCs). The E-Key notification handler communicates with the Shelf Manager and notifies interested applications when ports change E-Key states.For example, the Ethernet switching software needs to know the E-Key state of relevant ports. The Ethernet switching software registers itself with the E-Key notification handler through the internal E-Key API to receive notifications when its switch ports become enabled or disabled. Other software components may also register themselves.In addition, the Shelf Manager tracks current authorization for the synchronization clock interface using bused E-Key commands, giving permission to blades as appropriate to drive clock signals onto the clock buses. For more information, see Network timing subsystem (NTS) software on page 90.

95

7SNMP AGENT SUPPORT

The SCM can be remotely managed by an SNMP agent. The SNMP agent implementation is modular and extensible, dividing the functionality into a master agent and two subagents that represent the Base interface switch and the Fabric interface switch. The master agent receives requests from the SNMP manager and completes the requested action, coordinating with its subagents when necessary. Notifications generated by the master agent and the subagents are relayed by the master agent to registered destinations. The master agent communicates with the subagents using the AgentX protocol.Subagents are provided for Ethernet switch management (switchdrvr) and for SNMP notification logging (snmptrapd). The SNMP agent is configured and enabled as described in Configuring SNMP trap destinations and security access on page 67 and Enabling the SNMP agent on page 68. The SCM implements its SNMP agent by incorporating Net-SNMP open-source software, version 5.3.1. Usage details are covered in Net-SNMP man pages and on the http://www.net-snmp.com Web site. The SNMP agent currently supports SNMPv2c. SNMPv1 and SNMPv3 are currently implemented.

Base or Fabric switch selection The SNMP objects in the BRIDGE-MIB, P-BRIDGE-MIB, and Q-BRIDGE-MIB modules are instantiated separately for the Fabric interface and Base interface switches. For the Fabric interface switch, use community names that map to the "fi" context name to access these objects. To access all other objects, use community names that map to the null context name.Spanning tree notifications (newRoot and topologyChange) include the variable dot1dBaseBridgeAddress.0 to indicate the appropriate switch. The variable consists of the 6-byte MAC address used in the switch's bridge priority value. The dot1dBaseBridgeAddress.0 value for the switches is shown in the output of the Ethernet CLI show spanning-tree brief command for either the Base or the Fabric interfaces. The last six bytes of the Bridge Priority field determine the value.

ifIndex values for IP interfacesFor a list of ifIndex values that identify the SCM ports, see IPv4 interface and Ethernet switch port identifiers on page 84.

http://www.net-snmp.com

7 SNMP Agent Support

96

MIB module supportTable 34 lists the MIB modules supported by the SNMP agent. For each MIB module, the table also references the associated Request for Comments (RFC) document (if applicable), identifies whether the MIB module is supported by the master agent or one of the subagents, and provides the support details for the MIB module.

Exceptions

BRIDGE-MIB support details The following tables are not populated:

dot1dStaticTable

EtherLike-MIB support details dot3CollTable is not populated.

IEEE8023-LAG-MIB support detailsdot3adAggPortDebugTable is not populated.

IF-MIB support details ifTestTable is not populated.

Table 34. MIB module support

MIB RFC document Supported by Support detailsBRIDGE-MIB RFC 1493 switchdrvr subagent exceptions, see page 96EtherLike-MIB RFC 2665 switchdrvr subagent exceptions, see page 96IEEE8023-LAG-MIB n/a, IEEE Web site master agent and switchdrvr subagent exceptions, see page page 96IF-MIB RFC 2863 master agent and switchdrvr subagent exceptions, see page 96IP-MIB RFC 2011 master agent no major exceptionsP-BRIDGE-MIB RFC 2674 switchdrvr subagent exceptions, see page 97Q-BRIDGE-MIB RFC 2674 switchdrvr subagent exceptions, see page 97RFC1213-MIB RFC 1213 master agent exceptions, see page 97SNMPv2-MIB RFC 3418 master agent exceptions, see page 97SNMP-FRAMEWORK-MIB RFC 3411 master agent no major exceptionsSNMP-MPD-MIB RFC 3412 master agent no major exceptionsSNMP-TARGET-MIB RFC 3413 master agent no major exceptionsSNMP-NOTIFICATION-MIB RFC 3413 master agent no major exceptionsSNMP-USER-BASED-SM-MIB RFC 3414 master agent no major exceptionsSNMP-VIEW-BASED-ACM-MIB RFC 3415 master agent no major exceptionsTCP-MIB RFC 2012 master agent no major exceptionsUDP-MIB RFC 2013 master agent no major exceptions

7MIB module support

97

P-BRIDGE-MIB support details The following tables are not populated:

dot1dUserPriorityRegenTabledot1dPortOutboundAccessPriorityTable

Q-BRIDGE-MIB support detailsThe following tables are not populated:

dot1qForwardAllTabledot1qForwardUnregisteredTabledot1qPortVlanStatisticsTabledot1qPortVlanHCStatisticsTabledot1qLearningConstraintsTable

The following scalars are not populated:dot1qConstraintSetDefaultdot1qContraintTypeDefault

RFC1213-MIB support detailsThe following objects are not populated:

egp scalarsegpNeighTable

The following tables are supported as originally specified in RFC1213-MIB:atTableipRouteTable

The remaining scalars and tables are supported as re-specified in SNMPv2-MIB, IF-MIB, IP-MIB, TCP-MIB, and UDP-MIB.

SNMPv2-MIB support details warmStart notification is not supported.


98

Notification supportThe following table lists the notifications generated by the SNMP agent, the MIB modules in which they are specified, and the part of the SNMP agent that generates them.

MIB modules suppliedThe following MIB modules are supplied as text files installed on the SCM’s file system. The files include the MIB module name and end with .txt.

The text files are located in the /usr/share/snmp/mibs directory on the SCM’s file system. The directory also contains a README.txt file and a zipped archive of all the MIB files. The README.txt file explains how the MIB module files were created. The zipped archive of MIB files facilitates the transfer of files to other hosts.

Table 35. Notification support

Notifications MIBs Generated byauthenticationFailure SNMPv2-MIB master agentcoldStart SNMPv2-MIB master agentlinkDown IF-MIB master agent and switchdrvr subagentlinkUp IF-MIB master agent and switchdrvr subagentnewRoot BRIDGE-MIB switchdrvr subagentnsNotifyRestart NET-SNMP-AGENT-MIB master agentnsNotifyShutdown NET-SNMP-AGENT-MIB master agenttopologyChange BRIDGE-MIB switchdrvr subagent

BRIDGE-MIBEtherLike-MIBIANAifType-MIBIEEE8023-LAG-MIBIF-MIBIP-MIBNET-SNMP-AGENTNET-SNMP-MIBNET-SNMP-TCP-BRIDGE-MIB

Q-BRIDGE-MIBRADISYS-ATCA-MIBRADISYS-MIBRFC1155-SMIRFC1213-MIBRMON-MIBSNMP-FRAMEWORK-MIBSNMP-MPD-MIBSNMP-NOTIFICATION-MIBSNMP-TARGET-MIB

SNMP-USER-BASED-SM-MIBSNMP-VIEW-BASED-ACM-MIBSNMPv2-CONFSNMPv2-MIBSNMPv2-SMISNMPv2-TCSNMPv2-TMTCP-MIBUDP-MIB

7MIB browser utility applications

99

MIB browser utility applicationsThe SCM provides MIB browser utility applications from the Net-SNMP package. These applications can simulate an SNMP manager’s interaction with the master agent and are installed directly on the SCM in the /usr/bin directory. If you have the NET-SNMP 5.3.1 man page installed on a system, execute the man snmpcmd for more information.

The applications include:

snmptbulkgetsnmpbulkwalksnmpdeltasnmpget

snmpgetnextsnmpsetsnmpstatussnmptable

snmptestsnmptranslatesnmpwalk


100

101

8API OVERVIEW

This chapter provides an introduction to the C-language APIs provided for management of the SCM. The APIs initially cover a portion of the configuration and management features found in the CLI, with more features to be added incrementally. Currently, applications must run locally on the SCM’s LMP.

Where to find related files and additional informationTo support application development, the RadiSys Promentum software distribution includes the following resources:

API libraries and header files in RPM packagesAPI client example source codeAPI reference manuals

README files are also included to explain: the distribution contents the development host requirementsthe instructions for library installation the instructions for building provided examples

Ethernet switching APIThe Ethernet switching API allows you to configure and monitor statistics on the Base and Fabric switches. Each switch is treated as a forwarding element (FE). A function call returns the array of FE information. Each FE is identified by a handle, which can be used in subsequent calls to identify the FE (switch). The Base switch is assigned the product designator type of zero (0), while the Fabric switch has a non-zero type.

The API’s function name prefix, shared library, and header file name are listed below:

Reference information for the Ethernet switching APIs are located in the Ethernet Switching Management API Reference, which is provided as part of the software distribution.

Function name prefix: RSYS_Esw

Shared library: librsyseswapi.so.0

Header file to include: rsys_esw_api.h

8 API Overview

102

NTS APIThe network timing subsystem API allows you to configure the SCM’s clock generation and select the network timing sources. The API is also for retrieving the NTS status and for monitoring asynchronous events.

NTS events include:Loss of signal on any inputChange of reference source state between valid and invalidChange in operating state, such as locked or holdoverTemporary interruption to the currently selected reference source


Reference information for the NTS APIs are located in the Network Timing Subsystem API Reference, which is provided as part of the software distribution.

Module state management APIThe module state management API is useful for monitoring and controlling the state of the SCM’s RTM. This API can keep both the controlling SCM and the system manager informed about the RTM.Through the module state management API, an application can:

Retrieve the module’s operational stateChange the module’s administrative stateRetrieve the module’s current usage state


Reference information for the module state management APIs are located in the Blade Management API Reference, which is provided as part of the software distribution.

Function name prefix: RSYS_Nts

Shared library: librsysntsapi.so

Header file to include: nts/rsys_nts_api.h

Function name prefix: RSYS_Mm

Shared library: librsysmmsapi.so

Header file to include: rsys_mm_api_d0.5.h

8Flash and file system management API

103

Flash and file system management APIThe flash and file system management API is for loading and managing configurations on the SCM and its subsystems. This includes capabilities for:

Reporting the CLI-created configuration files stored on the SCM

Restoring the startup configuration

Copying a file to a local or remote location

Renaming, deleting, or listing details about a file


A provided example application copies a software image to the SCM and then upgrades the running software to that image by invoking the Linux software upgrade utility.Reference information for the flash and file system management APIs are located in the Blade Management API Reference, which is provided as part of the software distribution.

Software image management APIThe software image management API reports the software images and versions residing on the SCM.


Reference information for the software image management APIs are located in the Blade Management API Reference, which is provided as part of the software distribution.

Function name prefix: RSYS_Fm



Function name prefix: RSYS_Swm



8 API Overview

104

Management plane framework (MPF) APIThe RadiSys MPF API provides access to general purpose features that can be useful in application development. The API includes:

An event loop architectureMultiple socket connection support, signal, and timer eventsCommon error definition and loggingA syslog wrapperA runtime debug logging framework


Reference information for the management plane framework APIs are located in the Management Plane Framework API Reference, which is provided as part of the software distribution.

Function name prefix: RSYS_Mpf

Shared library: libmpf.so

Header file to include: mpf/rsys_mpf_api.h

105

9USING THE COMMAND LINE INTERFACE

The SCM’s command line interface (CLI) allows you to interactively configure and monitor a wide range of services. This chapter describes the CLI and provides basic information about using it.

Accessing the master CLITo access the master CLI:

Connect to the SCM as described on page 62.

Log in to the CLI as described on page 62.

Once logged in to the CLI, you should see the ATCA-2210 prompt.

Displaying available commands and optionsTo access online help while typing commands, enter a question mark (?). The CLI returns a list of the commands or options available in your current mode.To get a list of available commands, enter a question mark (?):

ATCA-2210# ? base-ethernet : Start base ethernet CLI fabric-ethernet: Start fabric ethernet CLI blade-mgmt : Blade management configuration and status linux-shell : Start the Linux shell show : Show configuration and status platform-mgmt : Platform management configuration and status copy : Copy information to and from the blade erase : Erase startup configuration exit : Logout from CLI session logout : Logout from CLI session help : Operation instructions

To get help on a specific command, type as much as you know of the command, followed by a question mark. Many commands require multiple arguments, each one shown as the next option when you type a question mark. After entering an option, type a question mark again to see if there are more options. The “<CR>” notation indicates that a carriage return is an option. This means that you have typed all required arguments, and you can press Enter to execute the command.

9 Using the Command Line Interface

106

Master CLI command modesThe top-level (master) CLI is divided into several command modes that give access to different subsets of commands. Some of the modes are documented in detail in this manual, while other modes are documented in manuals that describe those portions of the software in more detail.As indicated from the list of commands above, the master CLI has the following main command modes:

The master CLI also contains some commands that can be executed directly, without entering a mode. These are described in Master CLI commands on page 108.Many of these modes contain additional modes within them. The additional modes are described with the command reference information for each command mode.To exit any of the modes or the master CLI, enter:

exit

“No” commandsMany configuration commands have a “no” form that is used to disable, delete, or reset a configuration back to the factory defaults. For example, the following command administratively disables a port:

shutdown

The “no” form of this command administratively enables a port:no shutdown

Table 36. Main command modes

Mode Access Method Prompt Mode Description Described inBase Ethernet base-ethernet ATCA-2210-Base# Configure the Base Ethernet switch. Ethernet Switching

Software Reference and IP Routing Software Reference

Fabric Ethernet fabric-ethernet ATCA-2210-Fabric# Configure the Fabric Ethernet switch, including optional IP routing features.

Blade management

blade-mgmt ATCA-2210(blade-man)# Configure the services the module provides, including Linux services, SNMP, and network timing.

This manual on page 111

Linux shell linux-shell Linux prompt (see page 78) Access the Linux shell, which is not part of the CLI. When you exit the shell, you return to the CLI.

Standard Linux documentation

Platform management

platform-mgmt platform-mgmt# Access shelf management, alarm, and FRU information.

Shelf Management Software Reference

9“Show” commands

107

“Show” commandsA “show” command displays information about resources. You can use the applicable “show” command to show information for individual resources. The “show” commands are executed outside of configuration modes, usually directly within a main command mode such as Base Ethernet.

Command line completionPressing the tab key completes a partially typed command keyword. For example, typing a partial command, con, and pressing the tab key completes the keyword configure. Typing a question mark (?) lists valid entries after a command keyword. For example, typing the keyword configure and then a question mark brings up a list of valid options that follow configure.

Editing and history keysThe editing key sequences are similar to those used by EMACS text editors. Any character typed is inserted into the command line at the current cursor position, and all characters to the right of the typed character shift to the right. The history key sequence is similar to that used by the UNIX C shell.

You can obtain a list of key sequences by entering help at the main CLI prompt.

Saving configurations persistentlySome CLI configuration changes are kept within the application until you save them to persistent storage. If you do not save, your unsaved changes remain in effect until the module is rebooted, which restores the startup configuration.These methods for saving CLI configurations are currently required for the various CLI command modes:

Table 37. Methods for saving CLI configurations

Mode Persistent Save Method See alsoBase Ethernet Two alternatives:

Save when exiting these modes, as prompted.From within these modes, use this command: copy system:running-config nvram:startup-config

Ethernet Switching Software ReferenceFabric Ethernet

Blade management

From the master CLI, use this command:copy system:running-config nvram:startup-config

page 108

Platform management

Not necessary. The configuration is saved persistently when you execute the commands. N/A


108

Master CLI commands

show versionShows the software versions of the master CLI and various CLI components.Syntax: show version

show running-configShows the current configuration, displayed as a series of commands necessary to reproduce the current configuration. This command currently shows only the blade-management mode configuration. Syntax: show running-config

copy tftpCopies the specified file from a TFTP server and makes it the startup configuration in flash memory or the running configuration. This command currently replaces only the blade-management mode configuration.Syntax: copy tftp:<url> { nvram:startup-config | system:running-config }

Options:nvram:startup-config

Loads the configuration file so it becomes the persistent startup configuration.

system:running-configLoads the configuration file so it becomes the running configuration.

copy system:running-configSaves the current running configuration to persistent storage on the SCM or to another system. If saved to persistent storage, the saved configuration is loaded again when the module is rebooted. This command currently saves only the blade-management mode configuration. Syntax: copy system:running-config {tftp:<URL> | nvram:startup-config}

Options:tftp:<URL>

Saves the configuration to another system.

nvram:startup-configSaves the configuration to the SCM’s non-volatile memory. The saved configuration is reloaded when the SCM is rebooted.

9copy nvram:startup-config

109

copy nvram:startup-configCopies the persistently saved startup configuration to the specified URL, or reloads it so it becomes the running configuration. This command currently copies only the blade-management mode configuration.Syntax: copy nvram:startup-config {tftp:<URL> | system:running-config}

Options:tftp:<URL>

Saves the startup configuration to another system.

system:running-configReloads the configuration to RAM, making it the current configuration.

erase nvram:startup-configErases the saved CLI configuration files from persistent storage. This returns aspects of the module that are controlled through the CLI to the factory defaults.Syntax: erase nvram:startup-config

This command currently only affects the blade-management mode configuration.

exitLeaves the current CLI command mode, and goes to the next higher level. If you are already in the master CLI (at the top level), this is the same as the logout command, which exits the CLI entirely and logs out from the module.Syntax: exit

logoutExits the CLI and logs out from the module. This command works only from the master CLI (top level).Syntax: logout

helpShows the editing and history keys. This is similar to what is shown on page 107.Syntax: help


110

111

10BLADE MANAGEMENT COMMANDS

This chapter provides reference material for the commands within the blade-management CLI mode. These commands help you enable, disable, configure, and show status of various services that run on the SCM.

Accessing blade-management command modesTo enter blade-management commands:

1. Log in to the SCM CLI.

2. Access blade-management mode by entering:

blade-mgmt

When the command reference pages below do not indicate a mode, use the top-level blade-management mode.

ConfigTo access configuration mode from blade-management mode, enter:

config

The prompt changes each time you change modes. The command reference pages below indicate when to use the configuration mode.

SNMP commands

service snmpdEnables the SNMP master agent. The restart option causes the agent to reset itself by stopping, deleting any log files, and restarting with the current configuration files. The reconfigure option forces the agent to re-read the configuration files and change its operation based on their contents, without fully re-initializing.Syntax: service snmpd [ restart | reconfigure ]

10 Blade Management Commands

112

no service snmpdDisables the SNMP master agent.Syntax: no service snmpd

Note: To check the SNMP agent’s status, use the show running-config command from the master CLI and check the output for the service snmpd start command.

See also: Enabling the SNMP agent on page 68 and SNMP Agent Support on page 95.

Network timing subsystem (NTS) commandsUse these commands to configure the NTS behavior and to display the current settings. For a general description of NTS, see Network timing subsystem (NTS) software on page 90.

Master/slave interactions for NTS configurationWhen using two SCMs as an NTS master/slave pair, execute the configuration commands on the SCM acting as the master. The configuration is automatically copied to the running configuration of the partner SCM, with adjustments to the configured role of the SCM. For SCMs configured to use NTS autonomously, no configuration information is shared between them.

To set and display the master/slave roles, see the commands network-clock-select role and show network-clocks brief.Note: Save the configuration persistently by executing the copy system:running-config command on each SCM.

10Network timing subsystem (NTS) commands

113

Input clock definitionsThe following are the network timing reference sources the SCM can accept as input.

* xlink1 and xlink2 here could be called xlink1(rx) and xlink2(rx) to indicate they are received clocks.The same names are used for transmitted clocks, but the paths are electrically distinct. Both SCMstransmit and receive a pair of xlink clocks at all times.

Table 38. Input clock definitions

Clock Description Frequency Inputclk3a The “a” half of the CLK3 bus, which is a backplane bus that

is part of the PICMG-defined synchronization clock interface. A node module can input a signal on CLK3.

8 kHz, 1.544 MHz (T1 and J1 modes only), 2.048 MHz (E1 mode only), or 19.44 MHz

clk3b The “b” half of the CLK3 bus. 8 kHz, 1.544 MHz (T1 and J1 modes only), 2.048 MHz (E1 mode only), or 19.44 MHz

extclk1 The first input from the rear transition module (RTM). 8 kHzextclk2 The second input from the rear transition module (RTM). 8 kHzextclk3 The third input from the rear transition module (RTM). 8 kHzextclk4 The fourth input from the rear transition module (RTM). 8 kHzxlink1* The first update channel clock received from the partner

SCM. Update channel signals between partner SCMs maintain synchronization by providing phase alignment and allows redundancy by providing quick detection of peer loss.

6.48 MHz

xlink2* The second update channel clock received from the partner SCM.

2 kHz


114

Output clock definitionsThe following are the network clock outputs the SCM can provide.

**xlink1 and xlink2 here could be called xlink1(tx) and xlink2(tx) to indicate they are transmitted clocks. The same names are used for received clocks, but the paths are electrically distinct. Both SCMs transmit and receive a pair of xlink clocks at all times.

Table 39. Output clock definitions

Clock Description Frequency outputclk1a The “a” half of the CLK1 bus, which is a backplane bus that is part of the

PICMG-defined synchronization clock interface.8 kHz

clk1b The “b” half of the CLK1 bus. 8 kHzclk2a The “a” half of the CLK2 bus, which is a backplane bus that is part of the

PICMG-defined synchronization clock interface.19.44 MHz

clk2b The “b” half of the CLK2 bus. 19.44 MHzsysclk1 The first output to the rear transition module (RTM). The outputs to the

RTM are always being driven.2.048 MHz in E1 mode or 1.544 MHz in T1 or J1 mode

sysclk2 The second output to the RTM. 8 kHzsysclk3 The third output to the RTM. 19.44 MHzxlink1** The first update channel clock transmitted to the partner SCM. Update

channel signals between partner SCMs maintain synchronization by providing phase alignment and allows redundancy by providing quick detection of peer loss. The outputs to the update channel are always being driven.

6.48 MHz

xlink2** The second update channel clock to the partner SCM. 2 kHz

10network-clock-select

115

network-clock-selectSets the priority of a clock input, from 2 to 7, with 2 indicating the highest priority. The SCM uses priorities to select a reference clock when the mode is configured to automatic (see network-clock-select mode). Each clock must have a unique priority. If clocks are set to the same priority, the last clock configured retains that priority, while the others are set to zero (ignored). Priority 1 is reserved for the update channel clock (xlink1) from the SCM in the NTS master role, used as the input by the slave SCM.Syntax: network-clock-select <priority> { clk3a | clk3b | extclk1 |

extclk2 | extclk3 | extclk4 }

Mode: Config

Defaults: Source: Priority:

xlink1 0 (ignored) if master, 1 if slave

clk3a 2

clk3b 3

extclk1 4

extclk2 5

extclk3 6

extclk4 7

See also: show network-clocks priority, show network-clocks source priority, Input clock definitions

network-clock-select modeSelects the reference clock source. The automatic option selects the highest priority, validated clock, while the force mode with a clock specified selects a particular clock source. If a clock source specified in force mode is not valid, the SCM goes into digital holdover mode in the absence of a valid clock source.Syntax: network-clock-select mode [ automatic | force { clk3a | clk3b |

extclk1 | extclk2 | extclk3 | extclk4 } ]

Mode: Config

Default: automatic

See also: show network-clocks brief

no network-clock-select modeSets the reference clock source to automatic, which selects the highest priority clock.Syntax: no network-clock-select mode

Mode: Config


116

network-clock-select roleConfigures an SCM to act as the master, the slave, or to be autonomous in relationship to the other SCM in the shelf. A master SCM provides the system clocks. A slave SCM stays synchronized with the master and outputs the same system clocks, if so configured, provides a seamless failover source. In the autonomous role, SCMs do not communicate their timing or configuration information to each other and may be under manual control of a management application such as a system manager.Syntax: network-clock-select role [ master | slave | autonomous ]

Mode: Config

Default: the SCM in lower-numbered slot is the master and the other SCM is the slave

See also: show network-clocks brief

no network-clock-select roleSets the SCM roles to the default, making the SCM in the lower-numbered slot the master and the other the slave.Syntax: no network-clock-select role

Mode: Config

network-clock-select revertiveSets the reference clock to revert to the original source if the signal is lost. When this occurs, the SCM fails over to another source and the signal is restored. When the revertive mode is enabled and a newly validated or revalidated source has a higher priority than the current reference source, the SCM transitions to the higher priority source. When the revertive mode is disabled, a source transition does not occur and the number of source transitions is minimized. Syntax: network-clock-select revertive

Mode: Config

Default: Does not revert, stays with current source

See also: show network-clocks configuration

no network-clock-select revertiveSets the reference clock to stay with the current source, even if the original source is restored after a failover.Syntax: no network-clock-select revertive

Mode: Config

10network-clock-select manage-bused-resource

117

network-clock-select manage-bused-resourceEnables E-Key management of the clock bus resources (CLK1 and CLK2). With E-Key management enabled, the Shelf Manager responds to CLI or NTS API requests to drive a particular clock bus by granting or denying a token. Unless this SCM or its partner owns a token, the SCM cannot enable output on a bus. With management enabled, the Shelf Manager tracks token ownership, and the running configuration reflects the ownership, allowing it to be saved persistently to the startup configuration. If the NTS master fails over to the slave, the SCM notifies the Shelf Manager that a new token owner is driving the bus, and the configuration value is automatically changed.With management disabled, the SCM can drive output to a clock bus without requesting a token.Syntax: network-clock-select manage-bused-resource

Mode: Config

Default: Disabled

See also: network-clock-select request-bused-resource, show network-clocks bused-resource

no network-clock-select manage-bused-resourceDisables E-Key management of the clock bus resources.Syntax: no network-clock-select manage-bused-resource

Mode: Config

network-clock-select request-bused-resourceRequests permission to drive a clock bus pair. The Shelf Manager provides a token to the entity authorized to drive the clock. A result message indicates whether the token was granted. If E-Key management of clock bus resources is enabled, token ownership is enforced when allowing output to be enabled, and ownership is reflected in the running configuration, allowing it to be saved persistently to the startup configuration.Even though one SCM requests permission to drive both halves of the bus, the SCMs can each drive half. Output can be enabled if the SCM or its partner SCM owns the token. See the example in network-clock-select output. Syntax: network-clock-select request-bused-resource {clk1 | clk2 | clk3}

Mode: Config

See also: network-clock-select manage-bused-resource, show network-clocks bused-resource

no network-clock-select request-bused-resourceReleases a token previously granted to the SCM to allow it to drive a clock bus pair. This frees the bus resource for use by another entity.Syntax: no network-clock-select request-bused-resource { clk1 | clk2 |

clk3 }

Mode: Config


118

network-clock-select outputEnables output of the PICMG-defined system clock signal on the selected clock bus. Selecting a or b enables output on only one of the bus pairs for the selected bus resource. Otherwise, both pairs are driven. When management of bused resources is enabled, output can be enabled only for buses for which this SCM or its partner owns the token. This command does not affect output to the RTM or to the update channel between SCMs. Output to the RTM and to the update channel is always enabled. Syntax: network-clock-select output { clk1 | clk2 } [ a | b ]

Mode: Config

Default: No output

Example: To configure the partner SCMs to provide redundancy as described in System clock redundancy on page 92, use this setup:

On the master SCM:network-clock-select output clk1 anetwork-clock-select output clk2 a

On the slave SCM:network-clock-select output clk1 bnetwork-clock-select output clk2 b

Either SCM can request the bus tokens and share them with the other SCM.

See also: show network-clocks, network-clock-select manage-bused-resource, network-clock-select request-bused-resource

no network-clock-select outputDisables output of the system clock signal on the selected clock bus. Syntax: no network-clock-select output { clk1 | clk2 } [ a | b ]

Mode: Config

network-clock-select framerSets the expected framing format of the signal received from the RTM framer clock. The format also affects the CLK3 input frequency and the sysclk1 output frequency. With the format set to e1, 2.048 MHz is the expected RTM and CLK3 input frequency and sysclk1 output frequency. With the format set to t1 or j1, 1.544 MHz is the expected frequency.Syntax: network-clock-select framer [ e1 | t1 | j1 ]

Mode: Config

Default: e1

See also: show network-clocks configuration

10network-clock-select frequency

119

no network-clock-select framerResets the framing format expected from the framer clock to the default. Syntax: no network-clock-select framer

Mode: Config

network-clock-select frequencySpecifies the source frequency for the specified CLK3 input clock. The 2048-kHz option is available only when the framing format is set to e1. The 1544-kHz option is available only when the framing format is set to t1 or j1.Syntax: network-clock-select frequency { clk3a | clk3b } [ 8-kHz |

1544-kHz | 2048-kHz | 19.44-MHz ]

Mode: Config

Default: 19.44-MHz

See also: network-clock-select framer, show network-clocks source status

no network-clock-select frequencySets the selected input clock back to the default frequency.Syntax: no network-clock-select frequency { clk3a | clk3b }

Mode: Config

show network-clocksShows the frequency and enable status of each clock that the SCM may output.Syntax: show network-clocks

Output: Network Clock OutputThe output clock name. The output clocks are described in Output clock definitions on page 114.

Output FrequencyThe output frequency of each network clock.

Status EnabledWhether the SCM is driving each clock output.

show network-clocks briefShows a summary of the current SCM’s NTS role and operating status.Syntax: show network-clocks brief

Output: Mode Whether NTS is using the defined priorities to automatically select the highest priority, valid input clock or whether a specific input selection was forced by a user.


120

Role The current NTS role of the SCM. The possible values are master, slave, and autonomous. When configured as a master or slave, the value may not be what was configured, because an automatic failover may have changed the role.

Selected The source clock currently being used as a network timing reference. The sources are described in Input clock definitions on page 113.

Operating ModeThe mode of the PLL. The possible values are:

locked - The PLL is locked to a valid source.

holdover - The PLL is operating based on stored phase information from a previous source to maintain clock stability until it locks to another source.

freerun - The PLL phase is independent of any source.

pre-locked - After being in free-run mode, the SCM has selected a reference source and the PLL is attempting to lock to it. If the source is of good quality, the PLL enters locked state within 100 seconds. Otherwise, it reverts to free-run mode and the SCM selects another reference source.

pre-locked2 - After being in holdover mode, the SCM has selected a reference source and the PLL is attempting to lock to it. The PLL also enters this mode if the SCM is in revertive mode and is transitioning to a higher priority source that has been restored.

phase-lost - The PLL has lost phase lock and is attempting to lock to the input reference clock, if it exists. If the anomaly causing the loss of lock is serious, the SCM disqualifies the reference clock. The clock is automatically disqualified and generates an alarm if the PLL remains in phase-lost mode for more than 100 seconds. When the reference clock is disqualifies, the PLL proceeds to pre-locked2 mode if a valid reference clock is available. Otherwise, it goes into holdover mode.

Operating StateThe current alarms for the PLL. The possible values are:

t0-soft-alarm - The PLL is tracking its input reference clock beyond the ±11.43 ppm drift limit that generates a soft alarm.

sync2k-alarm - The xlink2 input from the redundant SCM has a problem.

RTM PresentWhether a rear transition module has registered its presence with the NTS API.

Partner VisibleWhether this SCM’s master/slave partner is visible.

10show network-clocks configuration

121

show network-clocks configurationSummarizes the NTS configuration settings for the current SCM. Syntax: show network-clocks configuration

Output: Mode Whether NTS is using the defined priorities to automatically select the input clock. “Off” indicates the forced selection of a user-specified input.

Role The current NTS role of the SCM. The possible values are master, slave, and autonomous. The value may or may not be what was configured, because an automatic failover may have changed the role.

RevertiveWhether the selected clock source reverts to the original if the original source signal is lost and subsequently restored.

Output The output clocks the SCM is currently driving. For source clock descriptions, see Input clock definitions on page 113.

Framer The framing format expected on RTM or CLK3 source clocks.

Sa Bit The bit used to receive and transmit the synchronization status message (SSM).

SSM The synchronization status message to transmit.

Manage EKey TokenWhether the ability to drive the clock buses (CLK1 - CLK3) is restricted by ownership of E-Key tokens.

Bused ResourceThe clock bus E-Key tokens this SCM currently owns, allowing it (and its partner SCM) to drive clock signals on those buses.

show network-clocks priorityShows the configured priorities of the source network clocks. The highest priority is 1, with 0 (zero) being ignored. If the module is in a forced mode, the priorities are ignored (see network-clock-select mode).Syntax: show network-clocks priority

See also: network-clock-select

Output: For a description of the source clocks, see Input clock definitions on page 113.

show network-clocks source priorityShows the actual priorities of the source network clocks, including the source currently used. The priorities shown here are based on configured priorities and clock validity status. The priorities may vary from the configured priorities because some inputs may not be valid. If the module is in a forced mode, the priorities are ignored (see network-clock-select mode).Syntax: show network-clocks source priority


122

Output: SelectedThe source clock currently in use.

1st PriorityThe source with the highest priority. This may match the selected clock.

2nd PriorityThe source with the 2nd priority.

3rd PriorityThe source with the 3rd priority.

See also: For source clock descriptions, see Input clock definitions on page 113.

show network-clocks source statusShows the current status of the source network clock signals. Syntax: show network-clocks source priority

Output: (Actual priority and usage)Briefly indicates the actual priorities and usage of the source clocks, also shown by show network-clocks source priority. The possible characters next to the corresponding clock are:

* - The validated input clock to which the PLL is currently locked. This matches the first priority validated input clock (denoted by 1) when configured with revertive mode enabled. With revertive mode disabled, they may or may not match.

1 - The first priority validated input clock.

2 - The second priority validated input clock.

3 - The third priority validated input clock.

Input The input source, described in Input clock definitions on page 113.

Local The validity status of the input from the perspective of this SCM.

Partner The validity status of the input from the perspective of the partner SCM in the protection pair. In the case of the xlink1 and xlink2 clocks, the partner SCM reports on the validity of clocks it receives from the local SCM.

SSM The synchronization status message (SSM) value received on the input.

FrequencyThe frequency of the input clock signal.

See also: show network-clocks source priority

10show network-clocks source alarms

123

show network-clocks source alarmsShows the current alarms for each source network clock. Not every alarm invalidates a source clock. An alarm is cleared when the signal improves.Syntax: show network-clocks source alarms

Output: For each source clock (described in Input clock definitions on page 113), zero or more of the following alarms are displayed.

soft out-of-bandThe source clock phase has drifted by more than ±11.43 ppm. This alarm does not invalidate the source clock.

hard out-of-bandThe source clock phase has drifted by more than ±15.24 ppm. This alarm invalidates the source clock.

no-activityThe source is within the hard out-of-band limits, but shows no activity.

phase-lockThe PLL could not achieve phase lock on this selected source within 100 seconds. The SCM checks for phase lock only on a source currently selected and in use. Once the alarm has been raised and the SCM selects another source, the SCM is no longer checking for phase lock on this source. The alarm is cleared after 128 seconds.

peer-lossThe xlink1 or xlink2 update channel from the partner SCM no longer has a valid signal. If the slave SCM loses the master’s xlink1 signal when it was previously valid, it causes a failover and becomes the master.

loss of frameThe SCM received a loss of frame condition for this source.

loss of signalThe SCM received a loss of signal condition for this source.

alarm indicationThe SCM received an alarm indication signal condition (AIS) for this source.

yellow alarmThe SCM received a yellow alarm condition for this source.


124

show network-clocks bused-resourceShows the E-Key token ownership of the clock buses and summarizes the last communications between this SCM and the Shelf Manager in regard to the tokens.Syntax: show network-clocks bused-resource

Output: Bused ResourceA backplane clock bus that is part of the PICMG-defined synchronization clock interface.

Token OwnerThe current entity authorized to drive signals on this clock bus, according to the Shelf Manager. The possible values are:

Local - This SCM holds the E-Key token.

Partner - The other SCM in this master/slave pair holds the token.

Unknown - An unknown entity or no one holds the token.

Board2ShelfThe last communication this SCM sent to the Shelf Manager regarding this token.

Shelf ReplyThe last reply from the Shelf Manager to this SCM regarding this token, as a standard IPMI return code.

Shelf2BoardThe last communication this SCM received from the Shelf Manager.

Board ReplyThe last reply from this SCM to the Shelf Manager regarding this token.

125

11MAINTENANCE

The SCM and the shelf are shipped with the jumper, connector, and pin configurations already set for the Promentum platform.Using two SCMs in the platform creates a redundancy that provides continued availability, even if a failure occurs in one of the modules. The Promentum platform includes interface management software from which the SCM can configure other platform supported modules and connect to external network systems.The interface management software is accessed during an initial installation through the serial port cable included with the SCM. The front panel 10/100/Base-T Ethernet connection can be used after the initial installation verification of interface operations. For additional information on the Ethernet connection and CLI commands, see the Ethernet Software Reference manual.

WARNING! This product contains static-sensitive components and should be handled with care. Failure to employ adequate anti-static measures can cause irreparable damage to components. Handle this product only when wearing a grounded wrist strap at a grounded work area. Additional precautions and ESD information are provided at www.radisys.com/esd.

Installing the SCM For information on installing or replacing an SCM, please see the ATCA-2210 Switch and Control Module Installation Guide.

Removing the SCM To remove the SCM from the shelf:

1. Disconnect all cables from the SCM front panel.

2. Loosen the SCM screws from the shelf.

3. Release the ejector latch in the position noted and then stop:

If the SCM is oriented vertically, release the lower ejector latch.

If the SCM is oriented horizontally, release the right ejector latch.

4. The H/S blue light starts flashing. When the H/S blue light turns solid blue you are ready to remove the SCM.

Note: When only one SCM is present in the shelf and its shelf manager function is enabled, the H/S blue light will never turn solid blue.

http://www.radisys.com/esd

11 Maintenance

126

5. Release the ejector latch in the position noted:

If the SCM is oriented vertically, release the top ejector latch.

If the SCM is oriented horizontally, release the left ejector latch.

6. Simultaneously pull both extractor latches to release the module from the slot.

7. Pull the SCM out of the slot.

8. Place the SCM on a flat, static-free surface.

TroubleshootingWhen you encounter a situation in which the Promentum platform does not perform as you expect, look for symptoms that might clarify the cause. Performing the following actions can aid you in diagnosing symptoms:

Check the state of the LEDs on all the modules in the platform, especially the power entry modules (PEMs) and rear transition modules (RTMs).

Check the shelf-management events logged in the system event log (SEL), which is accessible through the system manager. The Shelf Management Software Reference provides details on how to use the SEL.

Information about the Base Ethernet configuration, which can be generated using the show commands in the base-interface management CLI.

Information about the Fabric Ethernet configuration, which can be generated using show commands in the fabric-interface management CLI.

Verify the IP address and the subnet mask assignment. See Chapter 5, Initial Verification and Configuration, on page 61 for details on verifying these assignments.

Verify the Fabric Ethernet interface is communicating properly. See the Ethernet Switching Software Reference for information on diagnosing communication and connection issues.

Check temperatures on the module. The hpiapp application can be used to query the Alarm Table, which provides a way to query the actual temperature readings of devices. See the Shelf Management Software Reference for information on how to use the hpiapp application and how to perform these queries.

Verify SCMs are of the same type and are using the same software version, when installing two on the same platform for redundancy purposes. For example, you cannot use an ATCA-2210 SCM with an ATCA-2100 SCM.

11Troubleshooting

127

Symptoms and recommended actionsLook in this table for a symptom that applies to your situation, then follow the recommended action (or actions) for that symptom. When an action reveals the cause of the problem, resolve the problem as indicated.

Table 40. Troubleshooting actions based on symptoms

Symptom RecommendationThe power LED on the SCM or another module is not lit.

Verify that the module is fully inserted. Inspect the module connector pins for damage. If connector pins show no sign of damage, you can try the following:

carefully insert the module into a different slotcarefully insert a different module in the original slot

WARNING! Do not force the modules into the slots:To avoid damage to connectors, make sure the rear slot is either empty or contains a module that is compatible with the module being inserted. If the module does not slide easily, make sure you are inserting it into the correct slot and that it is aligned properly.

Check to see if the module’s Shelf Manager is disabled and whether the module is waiting for an external Shelf Manager to enable it.

Communication cannot be established between Ethernet nodes within the platform.

Connect to the SCM through the serial (LMP SER) or the Ethernet maintenance (LMP ETH) ports. Use base-interface management CLI commands to diagnose the cause. In particular, verify that an IP address has been assigned to the node, an appropriate subnet mask is assigned to the node, the node is within a reachable VLAN, and routing instructions do not prevent the communication. If switch control CLI commands fail to work with one of the nodes, check the system event log to verify that the node's Ethernet backplane interface was activated.

The SCM is in debug mode according to the login banner.

Restore valid Shelf FRU Information using the frurw utility as described in the Shelf Management Software Reference.

A module does not work correctly. Check the system event log for significant events related to the module. In particular, verify that the module worked correctly when it was installed, and look for any events since then that would account for why the module stopped working correctly. If the log information does not reveal useful symptoms, power down the module, slide it out, perhaps move it to a different slot, slide it back in, and power it back up. If appropriate, install a different module of the same kind to help determine whether the module might be defective.

The SCM’s embedded software is corrupted or it fails to boot properly with newly upgraded software.

See Loading software through U-Boot on page 128 or Loading software from the secondary flash device on page 130.

Intermittently, the SCM and other installed modules experience random data errors.

Verify that the platform’s frame-ground connection is properly connected to a high-quality earth-ground connection. Check for electrical noise at the backplane power connections and at the power entry module power inputs. Consider the possibility of a malfunctioning module causing electrical noise on backplane connections.

11 Maintenance

128

Loading software through U-BootIf the SCM’s software kernel becomes corrupted, you can load a software image using U-Boot if U-Boot is functioning. This procedure requires:

A serial connection to the SCM’s front panel serial port.

A valid SCM software image (unzipped, with a file name ending in .bin) placed on an accessible TFTP server (for example in the /tftpboot directory).

A network connection to the SCM front panel Ethernet maintenance port (LMP ETH).

To load software using U-Boot:

1. Establish a serial connection to the SCM. See Serial connection on page 62.

You may not see a prompt at this point, because the SCM has stopped its attempts to boot.

Tip: After the next step, you must enter a key combination within two seconds to break into U-Boot. Read the next steps carefully before performing them. If you do not enter the keystrokes in time, you can perform the steps again until you are successful.

2. Reboot the SCM by pressing the recessed reset button on the SCM’s front panel.

3. At the prompt to enter U-Boot, enter these key combinations within 2 seconds:

Ctrl-x Ctrl-w

Tip: If your terminal does not show any activity after pushing the reset button, the terminal may not be configured correctly. Check your terminal emulator settings again. If continued reboots show no activity, U-Boot is not working properly. Read Loading software from the secondary flash device on page 130.

4. Specify the IP address of the TFTP server by entering:

setenv serverip <server_ip_addr>

5. Configure the SCM Ethernet maintenance port to a subnet IP address accessible to the TFTP server by entering:

setenv ipaddr <SCM_ip_addr>

The SCM or another module overheats

Verify no cover plates are installed. Empty slots must have air management panels rather than cover plates installed to properly maintain airflow and emissions. Use shelf management software to check temperatures at the air intake, on the module, and at the platform’s air exhaust. Use the information to determine whether the overheating may be caused by warm facility air, a module failure, or a failed fan module. Try moving the module to a different slot to see if that resolves the overheating. Verify there is at least two inches of clearance between the side of the shelf and the side of the rack cabinet. Check the air filter for obstructions and dirt.

Table 40. Troubleshooting actions based on symptoms (continued)

Symptom Recommendation

11Troubleshooting

129

6. Transfer the software image to the SCM’s RAM by entering:

tftp 1000000 <image_name>

For example:

tftp 1000000 /ATCA-2210/release1.8.x_scm3-wr.bin

for the file /tftpboot/ATCA-2210/release1.8.x_scm3-wr.bin on the TFTP server.

Note: The x in 1.8.x_scm3-wr.bin is a variable. Enter the correct value for the release you are transferring.

WARNING! U-Boot does not verify the software image you supply, so be sure you selected a valid image before proceeding.

You are about to overwrite memory addresses, so type the next commands carefully.

7. Load the image from RAM to the primary flash device by entering:

protect off 2:0-507erase 2:0-507cp 1000000 fc000000 fe0000

This loads everything in the software image except U-Boot. Preserving U-Boot allows you to try the procedure again if the image is corrupted or if the reload doesn’t work.

8. Make the new software take effect by rebooting the SCM:

reset

If the software doesn’t boot successfully, try the procedure again, or read Loading software from the secondary flash device.

Rebooting the LMPTo issue a cold reset request on the LMP software, you can use the PICMG FRU control command “cold reset.” The LMP software on the SCM has no warm reset control.

Rebooting the flash devices To reboot to the primary flash device:

1. Connect to the SCM and access the Linux prompt.

2. Enter the commands below: reboot -0

11 Maintenance

130

To reboot to the secondary flash device:


2. Enter the commands below: reboot -1

Loading software from the secondary flash deviceIf the software located in the SCM’s primary flash device becomes corrupted and the secondary flash device is functioning, boot the SCM to the secondary flash device. Once the SCM has booted, you can load a new software image to the primary flash device from the secondary.


2. To reboot to the secondary flash, enter the command below:reboot -1

3. Once the reboot process has completed, reflash the primary image from the secondary flash device. A. To reflash the full flash image from a local file enter:

reflash -a -i <image>

The <image> parameter represents the path to and the name of the full flash image file.

For example: reflash -a -i flashimage.bin

B. To reflash the full image from a remote file via tftp, enter: reflash -a -i <image> -tftp <from_ip>

C. For information on the other reflash parameters and options, you can enter: reflash --help

If the procedure above still does not resolve the situation, please contact RadiSys Technical Support.

Diagnosing bootup issues by disabling the BMC watchdog timer If you encounter bootup issues on the SCM, you can disable the BMC watchdog timer and check the system event log (SEL) to aid you in your diagnosis. See the Shelf Management Software Reference for information on how to check the SEL. In cases where you are unable to access the CLI, you can persistently disable the watchdog through U-Boot using the following syntax:

setenv nobmcwatchdog 1saveenv

131

ASPECIFICATIONS

EnvironmentalRadiSys does not provide environmental certification testing, because any meaningful emissions agency certification must include the entire system. Thus, the SCM is designed and tested to pass the environmental specifications noted below, but is not certified. Please note the operating environment must provide sufficient airflow across the SCM to keep it within its temperature specification.

Table 41. Environmental

Characteristic State ValueTemperature (ambient)

Operating +5° C to +45° C30° C/hr rate of change

Short term operating 1 –5° C to +55° C30° C/hr rate of change

Storage –40° C to +70 ° C

Rate of change:23° C to -40° C at 30° C/hr-40° C to 23° C at 13° C/min+23o C to +70o C at 30° C/hr+70° C to +23° C at 10° C/min

Relative humidity Operating 5% to 85% RH non-condensingShort term operating1 5% to 90% RH non-condensing at +30° CStorage 5% to 90% RH non-condensing at +40o CShort term storage1 5% to 95% RH non-condensing at +40o C

Altitude Operating2 Up to 1800 meters (5,905 feet), +55° C> 1800 meters up to 4000 meters (13,123 feet), derated linearly to +45° C

Shock (drop) UnpackedFree fall, corners and edges

0 to < 10 kg = 100 mm drop 10 to < 25 kg = 75 mm drop

Packaged (Unpalletized)Free fall, corners and edges

0 to < 10 kg = 750 mm drop10 to < 25 kg = 600 mm drop

Palletized 300 mm free fall dropVibration Operating 0.1g, 5 to 100 Hz and back, 0.1 octave/min sine sweep 2

Transportation (packaged) 0.5g, 5 to 50 Hz and back, 0.1 octave/min sine sweep 2

3.0g, 50 to 500 Hz and back, 0.25 octave/min sine sweep 2

A Specifications

132

SafetyThe safety specifications are measured under laboratory ambient temperature and humidity (approximately 25°C and humidity between 30% and 50%). Testing was performed in partnership with a Nationally Recognized Testing Laboratory (NRTL) accredited to provide the required certifications.

Mechanical dimensions

Seismic Operating Per Zone 4 test method, GR-63-CORE1 “Short term” defined as a maximum of 96 consecutive hours.2 For a maximum of eight consecutive hours.

Table 41. Environmental (continued)

Characteristic State Value

Table 42. Safety

Characteristic Certification Standard and Test CriteriaUS Accessory Listing UL 60950-1 “Safety for Information Technology Equipment”Canada Approval CSA 22.2 #60950-1 “Safety for Information Technology

Equipment”EU Conformance with the Low

Voltage DirectiveEN 60950-1 “Safety for Information Technology Equipment”

Other CB Report IEC 60950-1 “Safety for Information Technology Equipment”

Table 43. Mechanical dimensions

Characteristic ValuePCB board Dimensions 322.25mm x 280.0mm +0, -0.3mm (12.687” x 11.023” +0.0, -0.012”)

Board thickness 2.0mm ±0.2mm (0.079” ± 0.007”)

AElectromagnetic compatibility (EMC)

133

Electromagnetic compatibility (EMC)The SCM has been tested and found to comply with the requirements detailed in the following standards when installed in a representative shelf:

ICES-003 Class A;FCC Part 15 Class A;EN 55022:1998 Class A;EN 300 386 V1.3.3EN 55024:1998 + A1 + A2

WARNING! This is a Class A product. In a domestic environment this product may cause radio interference in which case the user may be required to take adequate measures.

This equipment has been tested and found to comply with the limits for a Class A digital device, pursuant to Part 15 of the FCC Rules. These limits are designed to provide reasonable protection against harmful interference when the equipment is operated in a commercial environment. This equipment generates, uses, and can radiate radio frequency energy, and if not installed and used in accordance with the instruction manual, may cause harmful interference to radio communications. Operation of this equipment in a residential area is likely to cause harmful interference in which case the user will be required to correct the interference at his own expense.

Modifications made to the product, unless expressly approved by RadiSys Corporation, could void the user’s authority to operate the equipment.

The interface ports are designed for use as intra-building interfaces only (Type 2 or Type 4 ports as described in GR-1089-CORE, Issue 4) and require isolation from the exposed Outside Plant (OSP) cabling. The addition of Primary Protectors is not sufficient protection in order to connect these interfaces metallically to OSP wiring.

A Specifications

134

Power

Reliability The reported failure rates do not represent catastrophic failure. Catastrophic failure rates will vary based on application environment and features critical to the intended function.Failure rate (Fit): 8849.558 failures in 109 @ 55 C°Mean time between failures (MTBF): 113,000 hours @ 35 C°

Table 44. Power emissions

Characteristic State Standard and CriteriaNominal operating voltage -48V ETS 300 132-2 Static voltage levels

Table 45. Power consumption summary

Supply Amp WattTotal converted from 12V 86.9DC/DC Conversion Loss (85%) 15.3COM Express 12V 4.47 26.0RTM 12V 2.80 33.6Total 12V brick power output 161.8Brick Conversion Loss (93%) 12.2IPMC and RTM 3.3V & 2.5V 3.0Total PIM output 177.0PIM Loss (97%) 5.5Total -48V Power Required — 182.5

135

BIPMI COMMANDS AND MANAGED SENSORS

IPMI commandsThe following IPMI commands are supported by the SCM.

Table 46. Supported IPMI commands

Message commands NetFn CommandGet device ID App (06h) 01hCold reset App (06h) 02hWarm reset App (06h) 03hBroadcast “get device ID” App (06h) 01hReset watchdog timer App (06h) 22hSet watchdog timer App (06h) 24hGet watchdog timer App (06h) 25hGet channel info App (06h) 42hSet event receiver S/E (04h) 00hGet event receiver S/E (04h) 01hPlatform event (a.k.a. “event message”) S/E (04h) 02hGet device SDR info S/E (04h) 20hGet device SDR S/E (04h) 21hReserved device SDR repository S/E (04h) 22hSet sensor hysteresis S/E (04h) 24hGet sensor hysteresis S/E (04h) 25hSet sensor threshold S/E (04h) 26hGet sensor threshold S/E (04h) 27hSet sensor event enable S/E (04h) 28hGet sensor event enable S/E (04h) 29hGet sensor reading S/E (04h) 2DhGet FRU inventory area info Storage (0Ah) 10hRead FRU data Storage (0Ah) 11hWrite FRU data Storage (0Ah)‘ 12hGet SDR Storage (0Ah) 20hReserve SDR repository Storage (0Ah) 22hGet SDR Storage (0Ah) 23hPartial add SDR Storage (0Ah) 25hGet SEL info Storage (0Ah) 40hGet SEL entry Storage (0Ah) 43h

B IPMI commands and Managed Sensors

136

Add SEL entry Storage (0Ah) 44hClear SEL Storage (0Ah) 47hGet SEL time Storage (0Ah) 48hSet SEL time Storage (0Ah) 49hGet PICMG properties PICMG (2Ch) 00hGet address info PICMG (2Ch) 01hFRU control PICMG (2Ch) 04hGet FRU LED properties PICMG (2Ch) 05hGet LED color capabilities PICMG (2Ch) 06hSet FRU LED state PICMG (2Ch) 07hGet FRU LED state PICMG (2Ch) 08hSet IPMB state PICMG (2Ch) 09hSet FRU activation policy PICMG (2Ch) 0AhGet FRU activation policy PICMG (2Ch) 0BhSet FRU activation PICMG (2Ch) OChGet device locator record ID PICMG (2Ch) ODhSet port state PICMG (2Ch) 0EhGet port state PICMG (2Ch) 0FhCompute power properties PICMG (2Ch) 10hSet power level PICMG (2Ch) 11hGet power level PICMG (2Ch) 12hGet shelf address info (ShMC) PICMG (2Ch) 02hSet shelf address info (ShMC) PICMG (2Ch) 03h

Table 46. Supported IPMI commands (continued)

Message commands NetFn Command

BManaged sensors

137

Managed sensorsThe SCM supports the sensors described in Table 47 with sensor data record (SDR) support. The SCM implements a Type 0x12 record for the management controller and Type 0x01 records for the sensor descriptions. Type 0x12 is a management controller record and Type 0x01 is full sensor record.Sensors listed with a reading type of “threshold” include a range of possible thresholds readings:

LNR — Lower non-recoverable thresholds generate a critical alarm on the low side. LC — Lower critical thresholds generate a major alarm on the low side.LNC —Lower non-critical thresholds generate a minor alarm on the low side.UNC — Upper non-critical thresholds generate a minor alarm on the high side.UC — Upper critcal thresholds generate a major alarm on the high side.UR— Upper non-recoverable thresholds generate a critical alarm on the high side.

The following table provides a summary of sensors that are managed by the SCM.

Table 47. Managed sensors

Sensor #

Name Type Reading type Normal reading

Notes

0 ATCA Hotswap ATCA hot swap Sensor-specific 0x00-0x07 Return M0 to M7 ATCA hot-swap states for FRU 0 1 COME Hotswap

(COM Express hot swap)ATCA hot swap Sensor-specific 0x00-0x07 Return M0 to M7 ATCA hot-swap states for FRU 1

2 RTM Hotswap ATCA hot swap Sensor-specific 0x00-0x07 Return M0 to M7 ATCA hot-swap states for FRU 23 ATCA IPMB-0 ATCA IPMB type Sensor-specific 0x00-0xFF Bit [7] = IMPB B Override state

Bit [6:4] = IPMB B Local stateBit [3] = IPMB A Override stateBit [2:0} = IPMB A Local status

4 Ejector Closed Slot or connector Digital 1 0 1Ejector latch is open Ejector latch is closed

5 -48V Absent A Power supply Digital 0 0 1Power supply A

Presence detectedPower supply AFailure detected

6 -48V Absent B Power supply Digital 0 0 1Power supply B

Presence detectedPower supply BFailure detected

7 -48V Fuse Fault Power supply Digital 0 0 1Both A & B fuses OKPresence detected

Either A or B fuse blownFailure detected

8 SEL Overflow Event logging disabled

Digital 0 0 1No SEL overflow SEL overflow detected


138

9 IPMC WDT RESET(IPMC watchdog timer reset)

Slot or connector Digital 0 0 1IPMC watchdog not

expiredIPMC watchdog expired

10 HW REV(Hardware version)

Slot or connector Discrete 0x00 Hardware version number. Firmware acts differently for each hardware version.

11 +5V Voltage Threshold 5.00 This sensor measures voltage in Volts.Default thresholds

LNR LC LNC UNC UC UNR4.24 4.49 4.75 5.23 5.49 5.75

12 IPMC +3.3V Voltage Threshold 3.30 This sensor measures voltage in VoltsDefault thresholds


13 +3.3V Voltage Threshold 3.30 This sensor measures voltage in VoltsDefault thresholds








17 +12V Voltage Threshold 12.0 This sensor measures voltage in VoltsDefault thresholds




Table 47. Managed sensors (continued)

Sensor # Name Type Reading type

Normal reading Notes

BManaged sensors

139

19 Inlet Temp Temp Threshold 25 This sensor measures temperature in °CDefault thresholds

LNR LC LNC UNC UC UNR-10.0 -5.0 0.0 63 68 823

20 XFP Temp Temp Threshold 25 This sensor measures temperature in °CDefault thresholds

LNR LC LNC UNC UC UNR-10.0 -5.0 0.0 72 77 87

21 Fabric Temp Temp Threshold 25 This sensor measure temperature in °CDefault thresholds

LNR LC LNC UNC UC UNR-10 -5 0 65 70 85

22 PIM Alarm(Power input module alarm)

Power supply Digital 0 or 1 0 1PIM alarm not asserted PIM alarm asserted

23 PB Reset(Push button reset)

Slot or connector Digital 0 or 1 0 1Asserted Not asserted

24 IPMB 0 PU Ch A Slot or connector Discrete N/A 0 1 2No Pull-Up Half Pull-up Full Pull-up

25 IPMB 0 PU Ch B Slot or connector Discrete N/A 0 1 2No Pull-Up Half Pull-up Full Pull-up

26 ComE Present(COM Express present)

Module or board Digital 0 or 1 0 1COM Express present COM Express not

present27 IPMC Watchdog Slot or connector Discrete N/A Generates an event when the IPMC is reset 29 RTM Present Slot or connector Digital 0 or 1 0 1

RTM is present RTM is not presentFRU 1 COM Express sensors — only present if COM Express is present

160 ComE Pwr Fail Slot or connector Digital 0 0 1RTM power ok RTM power fail

161 ComE Type Slot or connector Discrete N/A 2-9Module type

162 ComE Eject Slot or connector Digital N/A 0 1COM Express not

presentCOM Express present





140

Board health and event generationThe managed sensors have settings that determine the health and the event generation for the SCM. The setting descriptions are as follows:

SCM health check: Sensor must be in a nominal state for the SCM to transfer from a state where no services are enabled on the module to a state where the module is going through hotswap.

Alarm level: The alarm level reported by the sensor when an event occurs. See Sensor alarms troubleshooting on page 143 for information on troubleshooting any alarm levels.

Local scanning disabled: Sensors events are not logged in the SCM’s IPMC log.

Remote scanning disabled: The status of a sensor is not updated or visible to remote modules. For example, a shelf manager would not be able to check the status of this sensor.

Remote event generation disabled: Sensors events are not generated by the SCM to the rest of the modules and components installed on the shelf.

Table 48 on page 141 shows the health and the event generation settings for each managed sensor.

FRU 2 RTM sensors — only present if RTM is present 163 RTM Temp Temp Threshold 25 This sensor measures temperature in °C

Default thresholdsLNR LC LNC UNC UC UNR-10 -5 0 65 70 80

164 RTM 12C Ready Slot or connector Digital 1 0 1RTM I2C is not ready RTM I2C is ready

165 RTM Pwr Fail Slot or connector Digital 0 0 1RTM power ok RTM power fail

166 RTM Eject Slot or connector Digital 1 0 1RTM latch is open RTM latch is closed




BBoard health and event generation

141

Table 48. Board health and event generation

Sensor #

NameSCM healthy check

Alarm level (if applicable)

Local scanning disabled

Local event generation disabled

Remote scanning disabled

Remote event generation disabled

0 ATCA Hotswap N/A No No No No1 COME Hotswap

(COM Express hot swap)N/A No No No No

2 RTM Hotswap N/A No No No No3 ATCA IPMB-0 N/A No No No No4 Ejector Closed N/A No No No No5 -48V Absent A N/A No No No No6 -48 V Absent B N/A No No No No7 -48 V Fuse Fault N/A No No No No8 SEL Overflow N/A No No No No9 IPMC WDT RESET

(IPMC watchdog timer reset)

N/A No No No No


Minor, Major, Critical voltage high

No No No No

Minor, Major, Critical voltage low

11 +5V Minor, Major, Critical voltage high

No Varies* No Varies*


12 IPMC +3.3V Minor, Major, Critical voltage high

No No No No

Minor Major, Critical voltage low

13 +3.3V Minor, Major, Critical voltage high










142




17 +12V Minor, Major, Critical voltage high






19 Inlet Temp Minor, Major, Critical temperature high

No No No No

Minor, Major, Critical temperature low

20 XFP Temp Minor, Major, Critical temperature high

No No No No


21 Fabric Temp Minor, Major, Critical temperature high

No No No No


22 PIM Alarm N/A No Varies* No Varies*23 PB Reset

(Push button reset)N/A No Varies* No Varies*

24 IPMB 0 PU Ch A N/A No Varies* No Varies25 IPMB 0 PU Ch B N/A No No No No26 ComE Present

(COM Express present)N/A No No No No

27 IPMC Watchdog N/A No No No No29 RTM Present N/A No No No Varies**

FRU 1 COM Express sensors — only present if COM Express is present160 ComE Pwr Fail N/A No Varies** No Varies**161 ComE Type N/A No Varies** No Varies**162 ComE Eject N/A No Varies** No Varies**

Table 48. Board health and event generation (continued)

Sensor # Name

SCM healthy check






BSensor alarms troubleshooting

143

Sensor alarms troubleshooting

FRU 2 RTM sensors — only present if RTM is present 163 RTM Temp Minor, Major, Critical

temperature highNo Varies*** No Varies***


164 RTM I2C Ready N/A No Varies*** No Varies***165 RTM Power Fail N/A No Varies*** No Varies***166 RTM Eject N/A No Varies*** No Varies**** Event generation is enabled when the payload power is turned on.** Event generation is enabled is a COM Express module is detected. *** Event generation is enabled if an RTM is detected.

Table 48. Board health and event generation (continued)

Sensor # Name

SCM healthy check






Table 49. Sensor alarms troubleshooting

Sensor # Name Alarm level

(if applicable) Possible fault or error Action to take

0 ATCA Hotswap N/A M7 state Remove SCM and reinsert to see if problem persists. If problem still persists, then remove SCM and contact customer support.

1 COME Hotswap(COM Express hot swap)

N/A M7 state Remove SCM and reinsert to see if problem persists. If problem still persists, then remove SCM and contact customer support.

2 RTM Hotswap N/A M7 state Remove SCM and reinsert to see if problem persists. If problem still persists, then remove SCM and contact customer support.

3 ATCA IPMB-0 N/A I2C data or clock lines problem

Investigate bus communication error. Possibly remove SCM and reinsert.

4 Ejector Closed N/A None No action required; information only.5 -48V Absent A N/A Voltage from -48V power

source ‘A’ is absent.Investigate power loss

6 -48 V Absent B N/A Voltage from -48V power source ‘B’ is absent.

Investigate power loss

7 -48 V Fuse Fault N/A A -48V fuse is not functioning.

Check the -48V fuses, and replace as necessary.

8 SEL Overflow N/A SEL full Read events, and clear9 IPMC WDT RESET

(IPMC watchdog timer reset)

N/A Missing strobe from the IPMC to the IPMC FPGA

Monitor to see if the alarm worsens or is persistent for more than 1 day.


144


N/A None No action required; information only.

11 +5V Minor voltage high Voltage over 5.23V threshold; SCM will not go into M4 state.

Monitor to see if the alarm worsens.

Major voltage high Voltage over 5.49V threshold; SCM will not go into M4 state.

Monitor to see if the alarm is persistent. If persistent, remove SCM and contact customer support.

Critical voltage high Voltage over 5.75V threshold; SCM will not go into M4 state.

Remove SCM and contact customer support.

Minor voltage low Voltage under 4.75V threshold; SCM will not go into M4 state.

Monitor to see if the alarm worsens

Major voltage low Voltage under 4.49V threshold; SCM will not go into M4 state.


Critical voltage low Voltage under 4.24V threshold; SCM will not go into M4 state.


12 IPMC +3.3V Minor voltage high Voltage over 3.48V threshold; SCM will not go into M4 state.
















145

13 +3.3V Minor voltage high Voltage over 3.48V threshold; SCM will not go into M4 state.
























15 +1.25 V Minor voltage high Voltage over 1.30V threshold; SCM will not go into M4 state.
















146

16 +1.8 V Minor voltage high Voltage over 1.88V threshold; SCM will not go into M4 state.












17 +12V Minor voltage high Voltage over 12.54V threshold; SCM will not go into M4 state.



Monitor to see if the alarm is persistent.

























147

19 Inlet Temp Minor temperature high Temperature over 63°C threshold.


Major temperature high Temperature over 68°C threshold.

Monitor to see if the alarm worsens or is persistent for more than 1 hour. If persistent, remove and/or hot swap SCM. If problem persists, remove SCM and contact customer support.

Critical temperature high Temperature over 82°C threshold; SCM will not stay in M4 state.

Remove and/or hot swap SCM. If problem persists, remove SCM and contact customer support.

Minor temperature low Temperature under 0°C threshold.


Major temperature low Temperature under -5°C threshold.


Critical temperature low Temperature under -10°C threshold; SCM will not stay in M4 state.


20 XFP Temp Minor temperature high Temperature over 72°C threshold.
















148

21 Fabric Temp Minor temperature high Temperature over 55°C threshold.












22 PIM alarm (Power input module)

N/A PIM Alarm Investigate error

22 PB reset N/A PB Reset Investigate error23 IPMB 0 PU Ch A N/A None Investigate bus communication error. Possibly

remove SCM and reinsert.24 IPMB 0 PU Ch B N/A None Investigate bus communication error. Possibly

remove SCM and reinsert.25 ComE present

(COM Express present)N/A None No action required; information only.

26 IPMC Watchdog N/A None Monitor to see if the alarm worsens or is persistent for more than 1 day.

29 RTM Present N/A None No action required; information only.FRU 1 COM Express sensors — only present if COM Express is present

160 COM Express Pwr Fail N/A ComE power failure Remove SCM and investigate power rail error.161 COM Express power fail N/A None No action required; information only.162 COM Express type N/A None No action required; information only.163 COM Express Eject N/A None No action required; information only.





149

FRU 2 RTM sensors — only present if RTM is present 164 RTM temp Minor temperature high Temperature over 65°C

threshold.Monitor to see if the alarm worsens.



Critical temperature Temperature over 80°C threshold; SCM will not stay in M4 state.








165 RTM I2C Ready N/A RTM I2C is not ready No action required; information only.166 RTM Power Fail N/A RTM Power failure Remove SCM and investigate power rail error.167 RTM Eject N/A None No action required; information only.





150

ATCA-2210 Switch and Control Reference · NTS API ... This guide describes the ATCA-2210 switch and...

Documents

Transcript of ATCA-2210 Switch and Control Reference · NTS API ... This guide describes the ATCA-2210 switch and...