Académique Documents
Professionnel Documents
Culture Documents
ATCA-2210
www.radisys.com
Release history Release -0000 Date December 2006 Description First release.
Copyright 2006 by RadiSys Corporation. All rights reserved. RadiSys is a registered trademark and Promentum is a trademark of RadiSys Corporation. MontaVista is a registered trademark of MontaVista Software Inc. Red Hat and Red Hat Linux are registered trademarks of Red Hat, Inc. 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.
TABLE OF CONTENTS
Preface ...........................................................................................................................................................
About this manual.................................................................................................................................................................. Notational conventions......................................................................................................................................................... Electrostatic discharge .......................................................................................................................................................... Where to get more product information..........................................................................................................................
7
7 7 7 8
Chapter 3: Subsystems.................................................................................................................................. 19
Local management processor............................................................................................................................................. Clock synthesizer subsystem ............................................................................................................................................... Base Ethernet switch subsystem......................................................................................................................................... Fabric Ethernet switch subsystem ...................................................................................................................................... LED controller interface subsystem ................................................................................................................................... Network timing subsystem .................................................................................................................................................. Rear transition module ......................................................................................................................................................... Hardware management subsystem ................................................................................................................................... Power supplies........................................................................................................................................................................ 20 23 24 28 31 35 37 41 43
Table of Contents
72 72 72 73 74 77 79 82 83 83 86 86 86 86 86 87
87
Table of Contents
Troubleshooting...................................................................................................................................................................... 120
PREFACE
Notational conventions
This manual uses the following conventions: BoldText
MonoText ItalicText
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.
Preface
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.
Standards information
For 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
Related Documents
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. INF-8077i 10 Gigabit Small Form Factor Pluggable Module (XFP), Revision 4.5, XFP Multi Source Agreement (MSA) Group UL 60950-1 Safety for Information Technology Equipment CAN/CSA 22.2 #60950-1-03 Safety for Information Technology Equipment, CSA EN 60950-1:2002 Safety for Information Technology Equipment, CENELEC IEC 60950-1 Safety for Information Technology Equipment, IEC
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
Preface
10
FEATURES 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 functionspacket switching, shelf management, network timing, and provides a site for a daughter card processor, which can be used for system management. It has 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 network timing subsystem (NTS), an Intelligent Platform Management Controller (IPMC), and a Shelf Manager. Duplication of the SCM and its supporting components allows for redundancya 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.
Ethernet switches
The SCM is has two Ethernet switches: 1 Gigabit Base Ethernet switch The Base Ethernet is comprised of a single Broadcom BCM56304 Ethernet switch. See page 24 for more information on the Base Ethernet switch subsystem. 10 Gigabit Fabric Ethernet switch The Fabric Ethernet interface uses a single Broadcom BCM56800 Ethernet switch, which has twenty ports configurable for 1Gb or 10 Gb operation. See page 28 for more information on the Fabric Ethernet switch subsystem.
11
Features Overview
Installed software
The SCMs 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 71.
External connectivity
The SCM provides connectivity for communicating with external equipment through three interfaces: Front panel interfaces Front panel interfaces are accessed through each SCMs faceplate. The connection types are labeled on the faceplate. See page 46 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 49 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 SCMs signals and connections. See page 37 for descriptions of the RTMs electrical connections and signal interfaces.
12
THEORY OF OPERATION
Overview of functionality
The SCM provides system interfaces through Base Ethernet, Fabric Ethernet, clock I/O, and two Intelligent Platform Management Buses (IPMB). An LMP with a PowerQUICCIII chip has two serial interfaces, a local Inter-integrated circuit (I2C) bus, a PCI bus, and a memory bus. A 10/100 Ethernet channel from the PowerQUICCIII is routed to the front panel to provide an interface for switch management. The memory bus provides interfaces to the DRAM and boot flash memory. The PCI bus provides the interface to the Base Ethernet and the Fabric Ethernet switches. Figure 1 shows a block diagram.
13
Theory of Operation
1Gb 100Mb
10Gb 12-14 FE 2 FE
XFP
Sh Mgr
1Gb 0-4
RTM (Illustration)
ShMC IPMC
SCM
* Reserved for future implementation.
14
Overview of functionality
Fan
IPMC
Fan
IPMC
Sensors
SPM IPMC
I2C
IPMC
IPMC
ShMC
ShMC
IPMC
IPMC
Module
Module
SCM
SCM
Module
Module
Note: IPMC Intelligent platform managment controller SCM Switch and control module ShMC Shelf management controller SPM Shelf peripheral module PEM Power entry module
PEM
IPMC
PEM
IPMC
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
15
Theory of Operation
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.
Node
Node
Node
Hub (SCM)
Hub (SCM)
Node
Node
Node
16
Overview of functionality
Figure 4 illustrates the dual star topology in terms of the backplane connections between modules on a RadiSys 14-slot shelf. Figure 4. Backplane connections through modules on a RadiSys 14-slot shelf
Ethernet interface
2 1 Channel
2 1 Channel
Channel
Channel
Module
Module
SCM
SCM
Module
Module
Node slots
Hub slots
Node slots
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.
17
Theory of Operation
For information on the software managing the Fabric Ethernet interface, see The Ethernet Switching Software Reference.
Redundancy
Typically, 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.
18
SUBSYSTEMS
The SCM is comprised 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
RJ45 RJ45 232 Buf 232 Buf LPC USB USB COM Express* 12V Switch 2 1 RJ -45 Mag RJ -45 Mag 1Gb Eth 2 4 2 10/100-Tx 2 10/10 10/10 PHY PHY MII Serial XFP XAUI to XFI XAUI to XFI XAUI to XFI XAUI to XFI
Dual -Boot Flash
SIO
SAS/SATA PCI-E
Dual 1 Gb MAC SERDES
ID ID & Temp
SFP*
EE PROM
I2C
SFP
SPI 4 3 NTM
2 ClkIn RefClk
SFP
MII
SFP
Upd Chan 1,2 Upd Chan 3,4 Sync I/F Upd Chan 0
SFP
DDR1 SODIMM
XFP
SFP
XFP
Serial
RJ -45
XFP
RJ -45
RJ -45
XFP
XAUI to XFI
SER
DES
RJ -45
16-port Mag
14 1Gb-Tx 1 EE PROM
RTM (Illustration)
Temp Sensors
I2C
SFP Flsh Sel IPMC IPMB 2 Local Power Conversion 12V DC/DC Power Converter 48V Tyco PIM-200 Power Cond Fuses
Zone 2 Conn
Zone 1 Conn
SCM3
* Reserved for future implementation. Blue = Base Ethernet Red = Fabric Ethernet
19
The block diagram in Figure 6 shows the relationship between the LMP and the SCM.
20
PCI bus 1
10/100/1000 MB Ethernet
MAC address
User
EEPROM
Clock/ calendar
RTM
Network Timing module Fast Ethernet switch
RTM
JTAG/COP
PowerQUICCIII processor
The PowerQUICCIII is the processor chip used to run the LMP. The PowerQUICCIII 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 PowerQUICCIII used in the LMP has the following characteristics: Processor MPC8541 Core Freq. (MHz) 833 PCI Bus SYSCLK VDD/Core Freq. (MHz) Freq. (MHz) 66 66 3.3V/1.2V I Cache (Kbytes) 32 DCache (Kbytes) 32 Processor Version 0x0080
21
3
The PowerQUICCIII processor chip incorporates the following elements: CPU with e5000 core that implements Book E 32-bit architecture 256KB on-chip memory DDR memory controller Programmable interrupt controller (PIC) compliant with OpenPIC architecture Local Bus Controller (LBC), which connects to the two 64MB flash components Two PCI bus controllers used to configure the Base and the Fabric switches I2C controller providing access to the SDRAM module, two serial EEPROMs, the clock/calendar, and devices on the RTM Two UART interfaces for: Serial console port routed through front panel Serial port providing link between the PowerQUICCIII 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 interfaces SPI controller providing a four-wire serial bus for accessing the NTS, the 10/100Base-T Ethernet switch, and devices on the RTM EEPROM used to store media access control (MAC) addresses for Ethernet ports EEPROM used for storing user-defined information Clock/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 status Synchronous 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 history Integrated reset logic
22
23
Figure 7 is a block diagram showing how the Base Ethernet switch communicates with the rest of the SCM.
24
0 1 2 3 4 5 6 7
Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics Dual GbE Magnetics FE Switch w/ Mags
BCh2 BCh3 BCh4 BCh5 BCh6 BCh7 BCh8 BCh9 BCh10 BCh11 BCh12 BCh13 BCh14 BCh15 BCh16 BCh1 ShMC1 ShMC2 Backplane
TBI Gb SERDES
COM Express *
PCI -Express
Zone 3 Conn
RTM
Port 16
18 19 20 21 XAUI to XFI
Front Panel
Port 24 Port 26
XFP
Front Panel
Configuration
The 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 have Gigabit Ethernet ports going to optical 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.
25
3
Port mapping
Table 2. Base Ethernet port mapping MDIO interface Port destination address 0 0/0 0x00 FE Switch (Shelf Manager) 1 0/1 0x01 Base channel 2 (link to redundant SCM) 2 0/2 0x02 Base channel 3 3 0/3 0x03 Base channel 4 4 1/2 0x06 Base channel 5 5 1/3 0x07 Base channel 6 6 1/0 0x04 Base channel 7 7 1/1 0x05 Base channel 8 8 2/2 0x0A Base channel 9 9 2/3 0x0B Base channel 10 10 2/0 0x08 Base channel 11 11 2/1 0x09 Base channel 12 12 3/2 0x0E Base channel 13 13 3/3 0x0F Base channel 14 14 SERDES/SGMII N/A RTM 15 SERDES/SGMII N/A RTM 16* SERDES N/A COM Express 17 SERDES/SGMII N/A RTM 18 SFP N/A Front panel 1G 1/6 19 SFP N/A Front panel 1G 1/7 20 SFP N/A Front panel 1G 1/8 21 SFP N/A Front panel 1G 1/9 22 SERDES/SGMII N/A RTM 23 SERDES N/A PowerQUICCIII X0 XAUI IF 0x10 Front panel XFP 1/5 X1 XAUI IF TBD RTM XAUI X2 XAUI IF N/A Not used X3 XAUI IF TBD RTM XAUI * Reserved for future implementation. Port # PHY/Port IPMI hardware Logical slot address N/A N/A 0x41 / 42 1/2 0x43 0x44 0x45 0x46 0x47 0x48 0x49 0x4A 0x4B 0x4C 0x4D 0x4E N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 3 4 5 6 7 8 9 10 11 12 13 14 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
26
27
3
Port status LEDs
This table describes the LEDs specific to the Base Ethernet. Table 4. Base Ethernet port status LEDS Port
1 Gb SFP ports
Each Base Ethernet SFP port on the front panel or on the RTM has two LEDs
10 Gb XFP ports
Each Base Ethernet XFP port on the front panel or on the RTM has two associated LEDs.
LED and possible states 1 green link status LED Green link established Blinking green link activity Off link fail or port is disabled 1 green port status LED Green port is enabled Off port is disabled 1 green link/activity status LED Green link established Blinking green link activity Off link fail or port is disabled 1 green port status LED Green port is enabled Off port is disabled
Note that there is no direct Base-to-Fabric connection on an SCM and no Fabric-to-Fabric connection between the two SCM blades. Figure 8 is a block diagram showing how the Fabric Ethernet switch communicates with the rest of the SCM.
28
GPIO TSEC2
Port 0 Port 1 Port 2 Port 3 Port 4 Port 5 Port 6 Port 7 Port 8 Port 9 Port 10 Port 11 Port 12 Port 13
FCh2 FCh3 FCh4 FCh5 FCh6 FCh7 FCh8 FCh9 FCh10 FCh11 FCh12 FCh13 FCh14 FCh15
Backplane
COM Express
Port 19
RTM
PCI-Express
Port 14
XFP
Front panel
Port 15
XFP
XFP
XFP
29
3
Port mapping
Table 5. Fabric Ethernet port mapping MDIO Port # PHY/Port interface address 0 XAUI N/A 1 XAUI N/A 2 XAUI N/A 3 XAUI N/A 4 XAUI N/A 5 XAUI N/A 6 XAUI N/A 7 XAUI N/A 8 XAUI N/A 9 XAUI N/A 10 XAUI N/A 11 XAUI N/A 12 XAUI N/A 13 XAUI N/A 14 XAUI 0x10 15 XAUI 0x11 16 XAUI 0x12 17 XAUI 0x13 18* SERDES N/A 19 XAUI N/A * Reserved for future implementation. Port destination Fabric Channel 2 Fabric Channel 3 Fabric Channel 4 Fabric Channel 5 Fabric Channel 6 Fabric Channel 7 Fabric Channel 8 Fabric Channel 9 Fabric Channel 10 Fabric Channel 11 Fabric Channel 12 Fabric Channel 13 RTM RTM Front Panel XFP 1/1 Front Panel XFP 1/2 Front Panel XFP 1/3 Front Panel XFP 1/4 COM Express RTM IPMI hardware address 0x43 0x44 0x45 0x46 0x47 0x48 0x49 0x4A 0x4B 0x4C 0x4D 0x4E N/A N/A N/A N/A N/A N/A N/A N/A Logical slot 3 4 5 6 7 8 9 10 11 12 13 14 N/A N/A N/A N/A N/A N/A N/A N/A
Configuration
The Fabric interface on the SCM is a managed 10Gb Ethernet switch with the following features: Switch configuration via the PowerQUICCIII PCI bus Transmit (Tx) port disable based on E-Key port state The Fabric Ethernet configuration port is connected to the PowerQUICCIII through PCI bus 2, a 32-bits wide bus operating at 66MHz.
30
31
3
Front panel LED positions
Figure 9. LED positions on SCM front panel
1/5 ACT 1/5 STAT 1/6 ACT 1/6 STAT 1/7 1/7 1/8 1/8 ACT ACT ACT STAT
32
States Yellow or amber = OOS Off = normal operation Green = power is good Yellow flashing = booting Yellow on = functioning Off = inactive Blue = ready for hot swap Green = link Off = no link Yellow blink = activity Off = no link or port disabled Green = link is good Off = no link or port disabled Green = enabled Green blink = activity Off = no link or port disabled Green = link is good Off = no link or port disabled Green = enabled Green blink = activity Off = no link or port disabled Green = link is good Green blink = activity Off = no link or port disabled Green = enabled Off = port disabled
LED part of the RJ-45 connector - indicates PHY port activity. Note: The COM Express connections are reserved for future implementation.
2 Yellow COM Express and LMP serial console port Activity
LED part of the RJ-45 connector - indicates PHY port activity. Note: The COM Express connections are reserved for future implementation.
4 green Fabric Ethernet XFP Port Status
LEDs part of the RJ-45 connector indicates GE PHY port status, driven through the PHY LED signals.
4 green Base Ethernet SFP Port Status
LEDs part of the RJ-45 connector indicates GE PHY port status, driven through the PHY LED signals.
33
3
LED controller Base assignments
Table 8. LED controller Base assignments Eth switch Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Base Bit stream position 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Faceplate port label 1/6 1/6 1/6 1/6 1/7 1/7 1/7 1/7 1/8 1/8 1/8 1/8 1/9 1/9 1/9 1/9 1/5 1/5 1/5 1/5 Switch port GE18 GE18 GE18 GE18 GE19 GE19 GE19 GE19 GE20 GE20 GE20 GE20 GE21 GE21 GE21 GE21 XG0 XG0 XG0 XG0 Indicator label Act Act Stat Stat Act Act Stat Stat Act Act Stat Stat Act Act Stat Stat Act Act Stat Stat Position color Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow Green Yellow
34
35
3
Figure 10. Network timing subsystem
EXT CLK IN 4 EXT CLK IN 3 EXT CLK IN 2 EXT CLK IN 1 TO2 EXT CLK OUT DS26502 Framer / Deframer DS26502 Framer / Deframer DS26502 Framer / Deframer CLK1 -A TO4 12.8MHz TCXO REFCLK 12.8MHz OCXO 12.8MHz XO I3 I4 CLK3 -A CLK3 -B TO11 TO1 SYNC2K I11 8 KHz CLK1 -B CLK2 -A TO5 19.44 MHz 2 KHz OUT 6.48 MHZ OUT 2 KHz IN 6.48 MHz IN CLK2 -B UPDCH3Tx UPDCH4Tx UPDCH3Rx UPDCH4Rx BITS/SSU-3 REF-3 DS26502 Framer / Deframer BITS/SSU-4 REF-4
I5 I6 I9 I10
TO3
19.44 MHz
RTM
(Illustration)
SCM
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 83.
36
A shelf peripheral module (SPM) acts as the shelf interface when installed in a 14-slot RadiSys shelves. The SPM slides into the rear of the platform and connects to the SCM in the corresponding front module slot (i.e., hub slot) using Zone 3 connectors. SPMs are optional for SCMs installed in RadiSys 6-slot shelves or non-RadiSys shelves. Figure 11 shows a block diagram example of an SPM designed for the 14-slot RadiSys shelf.
37
3
Figure 11. SPM block diagram
Local Power Conversion
Zone 3 Conn
Buf Buf
ID ID & Temp
SFP*
SFP
SFP
SFP
SFP
SFP
BITS/ SSU Clk BITS/ SSU Clk BITS/ SSU Clk BITS/ SSU Clk
RJ -45
RJ -45
RJ -45
RJ -45
SPM (Illustration)
* Reserved for future implementation.
38
SPI_SLVA[0..2] BLED_CLK BLED_DAT FLED_CLK FLED_DAT RTM_RST~ SPI_CLK SPI_SDO SPI_SDI BASE_MDC
3
Table 10. RTM interface signals (continued) Signal name BASE_MDIO BASE_XMDC BASE_XMDIO FABRIC_XMDC FABRIC_XMDIO EXTCLK1_IN+/EXTCLK2_IN+/EXTCLK3_IN+/EXTCLK4_IN+/EXTCLK_OUT+/RTM_LMP_INTR5~ RTM_LMP_INTR6~ NTS_8K_OUT+/NTS_19M_OUT+/PCI_E_TX+/PCI_E_RX+/BASE_GE_TX[0..1]+/BASE_GE_RX[0..1]+/BASE_10G_TX[0..3]+/BASE_10G_RX[0..3]+/FABRIC_10G_TX[0..3]+/FABRIC_10G_RX[0..3]+/Direction from SCM to RTM Out I/O Out Out I/O In In In In Out In In Out Out Out In Out In Out In Out In Pins Description 1 1 1 1 1 2 2 2 2 2 1 1 2 2 2 2 164 16 16 16 32 32 Base 1Gb interface MDIO data. 3.3 V bidirectional signal from Base switch. Base 10 Gb interface MDIO clock. 2.5 V signal from Base switch. Base 10 Gb interface MDIO clock. 2.5 V bidirectional signal from Base switch. Fabric 10 Gb interface MDIO clock. 2.5 V signal from Base switch. Fabric 10 Gb interface MDIO data. 2.5 V bidirectional signal from Base switch. Recovered clock 1 in, LVDS differential Recovered clock 2 in, LVDS differential Recovered clock 3 in, LVDS differential Recovered clock 4 in, LVDS differential Clock out, LVDS differential Interrupt (low-true) from the RTM to the LMP. Pulled up on SCM. Interrupt (low-true) from the RTM to the LMP. Pulled up on SCM. 8KHz external clock out, LVDS differential 19.44MHz external clock out, LVDS differential One PCI Express Tx lane, differential One PCI Express Rx lane, differential Base interface 1Gb Tx port, differential. Base interface SERDES signals. The number of Base ports available depends on SCM component loading (stuffing). Base interface 1Gb Rx port, differential. Base interface SERDES signals. The number of Base ports available depends on SCM component loading (stuffing). Base interface 10 Gb Tx port, differential. Base interface XAUI signals. The number of Base ports available depends on SCM component loading (stuffing). Base interface 10 Gb Rx port, differential. Base interface XAUI signals. The number of Base ports available depends on SCM component loading (stuffing). Fabric interface 10 Gb Tx port, differential. Fabric interface XAUI signals. The number of Fabric ports available depends on SCM component loading (stuffing). Fabric interface 10 Gb Rx port, differential. Fabric interface XAUI signals. The number of Fabric ports available depends on SCM component loading (stuffing).
40
Figure 12 shows how the IPMC communicates with other subsystems like the LMP and the RTM.
41
3
Figure 12. IPMC control interface to payload and RTM
+12V RTM_+12V_ENA~ +3.3V_IPMI +12V_RTM
RTM:+12V_FAIL~
RTM_+3.3V_ENA~ RTM:+3.3V_FAIL~ RTM_INTR~ RTM_PRESENT~ RTM_EJECT_IN~ RTM_RST~ RTM:HOTSWAP RTM_ENA~ IPMC:I2C_SCL IPMC:I2C_SDA
+3.3V_RTM
+12V +3.3V COME:PWR_ENA~ COME:PWR_FAIL~ COM_E_PRST~ COM_E_TYPE2 COM_E_TYPE1 COM_E_TYPE0 COME:RESET# CECOM1_TX CECOM1_RX CE_UC_ENA CE_GBE1 +12V_CE +3.3V_CE
RTM
Power switch
1 GbE Ena
COM Express*
Update channel
GPIO5
Flip Flop
IPMC:MSG_FLAG IPMC:MSG_FLAG_CLR
-48 V
+3.3V_IPMC PAYLOAD_PWR#
LMP
48 V DC/DC
IPMC
*Reserved for future implementation
+12V
42
Power supplies
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.
Power supplies
Information on the SCMs power consumption and emissions is provided in Appendix A: Specifications on page 125.
43
3
-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 PowerQUICCIII, 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 PowerQUICCIII, 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.
44
PHYSICAL INTERFACES
Mechanical specification
The SCM dimensions conform to the PICMG AdvancedTCA 3.0 R2.0 Specification. Table 11. Base board layout
COM Express site* Clock/ calendar SAS/SATA site*
SFP sockets
45
Physical Interfaces
Thermal design
Heat sinks are used in the SCMs 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 PowerQUICCIII processor. The heat sinks for the XFPs are a pin fin configuration and are compliant with the specifications defined in the XFP MSA.
ATCA - 2210
COM-E USB COM-E SER
FABRIC BASE
For descriptions of the front panel LEDs and their various states, see LED controller interface subsystem on page 31.
H/S
RESET
AdvancedTCA
Reset
46
Push Buttons/Switches
The 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.
47
Physical Interfaces
See the ATCA-2210 SCM Installation Guide for the list of SFP configurations and products that have been tested and confirmed by RadiSys to work with the SCM. The connector pinouts are included in the following table. Table 14. SFP Base Ethernet connector pinouts Pin 1 2 3 4 5 6 7 8 9 10 11 Description GND TXFAULT TXDIS MODDEF2 MODDEF1 MODDEF0 RATE SEL LOS GND GND GND Pin 11 12 13 14 15 16 17 18 19 20 20 Description GND RDRD+ GND +3,3V +3.3V GND TD+ TDGND GND
48
Backplane interfaces
Backplane interfaces
The following SCM interfaces are available through the backplane: Twelve 10/100/1000Base-T Base interface node ports One inter-LMP link to a redundant SCM through the update channel One Base Ethernet switch-to-switch link port through Base channel 2 Dual 10/100Base-T Base interface port to one or two external Shelf Managers through Base channel 1 Twelve 10 Gb XAUI Fabric interface node ports Six synchronization channel clock input/outputs to the NTS Master/slave link to NTS module clock through the update channel Two -48V power rails IPMB interface (two I 2C ports) with programmable pull-ups Eight hardware address lines, seven address / one parity
Update channels
Update 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 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.
49
Physical Interfaces
Table 16. Zone 1 contact assignments P10 (continued) Contact number 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Designation SDA_A SCL_B SDA_B MT1_TIP MT2_TIP -RING_A -RING_A MT1_RING MT2_RING RRTN_A RRTN_B SHELF_GND LOGIC_GND ENABLE_B VRTN_A VRTN_B EARLY_A EARLY_B ENABLE_A -48V_A -48V_B Unused pin on SCM Description IPMB data, port A IPMB clock, port B IPMB data, port B Metallic test # 1 tip Metallic test # 2 tip Ringing generator A negative ringing Ringing generator B negative ringing Metallic test #1ringing Metallic test #2 ringing Ringing generator A return Ringing generator B return Shelf ground and safety ground Logic ground (reference and return for front module-tofront module logic signals) Enable B (short pin for power sequencing, tied to VRTN_B on backplane) Voltage return A Voltage return B -48 volt early (pre-charge) A -48 volt early (pre-charge) B Enable A (short pin for power sequencing, tied to VRTN_A on backplane) -48 volt A, uses ENABLE_A to enable converters -48 volt B, uses ENABLE_B to enable converters Mating sequence Third Third Third Third Third Third First Third Third Third Third First Fourth First First First First First Fourth Second Third
50
Backplane interfaces
1 CLK1A+ CLK1ACLK1B+ CLK1B2 Tx4(UP)+ Tx4(UP)- Rx4(UP)+ Rx4(UP)3 Tx2(UP)+ Tx2(UP)- Rx2(UP)+ Rx2(UP)4 Tx0(UP)+ Tx0(UP)- Rx0(UP)+ Rx0(UP)5 Fabric channel 15 Tx2[15]+ Tx2[15]Rx2[15]+ Rx2[15]6 Tx0[15]+ Tx0[15]Rx0[15]+ Rx0[15]7 Fabric channel 14 Tx2[14]+ Tx2[14]Rx2[14]+ Rx2[14]8 Tx0[14]+ Tx0[14]Rx0[14]+ Rx0[14]9 Fabric channel 13 Tx2[13]+ Tx2[13]Rx2[13]+ Rx2[13]10 Tx0[13]+ Tx0[13]Rx0[13]+ Rx0[13]Note: Each differential pair has an individual L-shaped ground contact (not shown).
1 Tx2[12]+ Tx2[12]Rx2[12]+ Rx2[12]2 Tx0[12]+ Tx0[12]Rx0[12]+ Rx0[12]3 Fabric channel 11 Tx2[11]+ Tx2[11]Rx2[11]+ Rx2[11]4 Tx0[11]+ Tx0[11]Rx0[11]+ Rx0[11]5 Fabric channel 10 Tx2[10]+ Tx2[10]Rx2[10]+ Rx2[10]6 Tx0[10]+ Tx0[10]Rx0[10]+ Rx0[10]7 Fabric channel 9 Tx2[9]+ Tx2[9]Rx2[9]+ Rx2[9]8 Tx0[9]+ Tx0[9]Rx0[9]+ Rx0[9]9 Fabric channel 8 Tx2[8]+ Tx2[8]Rx2[8]+ Rx2[8]10 Tx0[8]+ Tx0[8]Rx0[8]+ Rx0[8]Note: Each differential pair has an individual L-shaped ground contact (not shown).
51
Physical Interfaces
1 Tx2[7]+ Tx2[7]Rx2[7]+ Rx2[7]2 Tx0[7]+ Tx0[7]Rx0[7]+ Rx0[7]3 Fabric channel 8 Tx2[6]+ Tx2[6]Rx2[6]+ Rx2[6]4 Tx0[6]+ Tx0[6]Rx0[6]+ Rx0[6]5 Fabric channel 5 Tx2[5]+ Tx2[5]Rx2[5]+ Rx2[5]6 Tx0[5]+ Tx0[5]Rx0[5]+ Rx0[5]7 Fabric channel 4 Tx2[4]+ Tx2[4]Rx2[4]+ Rx2[4]8 Tx0[4]+ Tx0[4]Rx0[4]+ Rx0[4]9 Fabric channel 3 Tx2[3]+ Tx2[3]Rx2[3]+ Rx2[3]10 Tx0[3]+ Tx0[3]Rx0[3]+ Rx0[3]Note: Each differential pair has an individual L-shaped ground contact (not shown).
Tx2[2]+ Tx2[2]Rx2[2]+ Rx2[2]Tx3[2]+ Tx0[2]+ Tx0[2]Rx0[2]+ Rx0[2]Tx1[2]+ Fabric channel 1 Tx2[1]+ Tx2[1]Rx2[1]+ Rx2[1]Tx3[1]+ Tx0[1]+ Tx0[1]Rx0[1]+ Rx0[1]Tx1[1]+ ShMC Port BI_SHMC BI_SHMC BI_SHMC BI_SHMC BI_SHMC _Tx0+ _Tx0_Rx0+ _Rx0_Tx1+ 6 Base channel 2 BI_DA2+ BI_DA2BI_DB2+ BI_DB2BI_DC2+ 7 Base channel 3 BI_DA3+ BI_DA3BI_DB3+ BI_DB3BI_DC3+ 8 Base channel 4 BI_DA4+ BI_DA4BI_DB4+ BI_DB4BI_DC4+ 9 Base channel 5 BI_DA5+ BI_DA5BI_DB5+ BI_DB5BI_DC5+ 10 Base channel 6 BI_DA6+ BI_DA6BI_DB6+ BI_DB6BI_DC6+ Note: Each differential pair has an individual L-shaped ground contact (not shown). Unused pin on SCM
52
Backplane interfaces
SPI_RTM_ SLVA2 SLVA1 SLVA0 SEL~ 4 RTM_RST~ SPI_CLK SPI_SDO SPI_SDI BASE_MDC 5 EXTCLK_ EXTCLK_ RTM_LMP_ RTM_LMP_ EXTCLK2_IN+ OUTOUT+ IRQ5 IRQ6 6 NTS_8K_ NTS_8K_ NTS_19M_ NTS_19M_ EXTCLK4_IN+ OUT+ OUTOUT+ OUT7 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTMxTX0+ RTMxTX0RTMxTX1+ RTMxTX1RTMxTX2+ 8 FAB:XE FAB:XE FAB:XE FAB:XE FAB:XE RTMxRX0+ RTMxRX0RTMxRX1+ RTMxRX1RTMxRX2+ 9 BASE:GE BASE:GE BASE:GE BASE:GE BASE:GE RTM6TX+ RTM6TXRTM6RX+ RTM6RXRTM7TX+ 10 BASE:GE BASE:GE BASE:GE BASE:GE BASE:GE RTM4TX+ RTM4TXRTM4RX+ RTM4RXRTM5TX+ Note: Each differential pair has an individual L-shaped ground contact (not shown).
EXTCLK3_IN+ EXTCLK3_INFAB:XE RTMxTX3+ FAB:XE RTMxRX3+ BASE:GE RTM7RX+ BASE:GE RTM5RX+ FAB:XE RTMxTX3FAB:XE RTMxRX3BASE:GE RTM7RXBASE:GE RTM5RX-
53
Physical Interfaces
54
Backplane interfaces
55
Physical Interfaces
Internal interfaces
PowerQUICCIII COP/JTAG header
The common on-chip debug processor/Joint Test Action Group (COP/JTAG) port provides a full duplex serial interface for communications between the internal PowerQUICCIII 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. Table 25. PowerQUICCIII COP/JTAG connector J3
1 3 5 7 9 11 13 15 2 4 6 8 10 12 14 16
Pin # 1 3 5 7 9 11 13 15
Pin # 2 4 6 8 10 12 14 16
Table 26. PowerQUICCIII COP/JTAG pin functions Pin # 1 2 3 4 5 6 7 8 9 10, 14 11 12 13 14 15 Name TDO QACK TDI TRST QREQ~ +3.3V TCK CKSTP_IN~ TMS N.C. SOFTREST~ GND HREST# Key CKSTP_OUT~ Description Development serial data out Quiescent Acknowledge (GND) Development serial data in (Pull-up) Test (soft) reset (Pull-down) Quiescent request (Pull-up) +3.3V supply through 1K) Development serial clock (Pull-up) Check Stop Input (Pull-up) Test mode select input (Pull-up) No Connect Soft Reset on PowerQUICCIII (Pull-up) Digital GND Hard Reset (Pull-up) Mechanical Key (No connect) Check stop output (Pull-up)
56
Internal interfaces
IPMC headers
The following headers on the SCM are embedded in the Intelligent Platform Management Controller (IPMC) entity, and are documented in the IPMC Reusable Entity Hardware Specification. J6, H8 Serial Debug and Programming Header J8, H8 JTAG Debugger Header J10, Auxiliary Function and IO FPGA Programming Header J11, Watchdog and Power Header
2 4 6 8 10 12 14
Pin # 1 3 5 7 9 11 13
Pin # 2 4
12 14
57
Physical Interfaces
58
Configuration checklist
This 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. Task: _____ Verifying Base interface operation _____ Verifying Fabric interface operation _____ Verifying Shelf Manager operation _____ Configuring IP addresses for SCM interfaces _____ Adding user accounts and setting passwords Configuring SNMP trap destinations and security _____ access (optional) _____ Enabling the SNMP agent (optional) _____ Enabling the SCMs DHCP server (optional) _____ Enabling and disabling the Telnet server and TFTP server (optional) Enabling and disabling syslog and other services _____ (optional) Copying a configuration from existing CLI _____ configuration files (optional) Copying CLI configuration files to a TFTP server _____ (optional) _____ Copying any file to the SCM (optional) This chapter begins with the following essential procedures: Connecting to the SCM on page 60 Logging in to the SCM on page 60 Rebooting the SCM on page 60 Where to find information: SCM Installation Guide SCM Installation Guide SCM Installation Guide page 61 page 63 page 63 page 64 page 65 page 66 page 67 page 68 page 68 page 69
59
Telnet connection
1. Connect an Ethernet cable to the SCMs Ethernet maintenance port (labeled LMP ETH). 2. Connect the cables 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). You should receive a login prompt.
If you have not yet set passwords, no password is required for the root and admin accounts.
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 73.
60
Configuring IP addresses
Configuring IP addresses
The SCM IP interfaces are configured by default to minimize your work in setting them up. Some 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 61. 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 61. 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 62. For more information and other possibilities, see Pre-defined IP addressing scheme on page 79.
61
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.
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
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
62
10. When prompted to save changes, say yes (y) to save the changes to the SCMs startup configuration.
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
63
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. 8. Save the file. The changes take place when you enable or restart the SNMP agent, or cause it to re-read its configuration file. See Enabling the SNMP agent on page 64 or the CLI command service snmpd on page 106. Changes to the file are saved persistently, so they will be restored after the SCM is rebooted. If you have installed the Net-SNMP 5.2.1 man pages onto a system, you can find more details on the snmpd.conf file options using the command:
man snmpd.conf
4. Enable the master agent by entering: service snmpd 5. Exit blade-management mode by entering: exit 6. Make the change to the agents 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 106 and the copy system:running-config description on page 102.
64
4. Copy the example configuration files or your replacement files to the required file names as follows:
cp pri.dhcpd.conf.example pri.dhcpd.conf cp sec.dhcpd.conf.example sec.dhcpd.conf cp 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. Be sure to make the address and 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 79 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
65
10. Make the lease database file persistent by copying it to the bind mount directory:
cd /rsys/onboot.data/bindmount mkdir -p mnt/RAM cp /mnt/RAM/dhcpd.leases mnt/RAM
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
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
66
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
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.
67
Executing the command returns you to the master CLI. 5. Retrieve the Fabric Ethernet switch configuration file and make it the running configuration:
fabric-ethernet copy tftp://<IP address>/<file-2> nvram:startup-config
Executing the command returns you to the master CLI. 6. Retrieve the blade-management configuration file and make it the running configuration (from the master CLI):
copy tftp://<IP address>/<file-3> nvram:startup-config copy nvram:startup-config system:running-config
The running configuration and the persistent startup configuration has changed to reflect the content of the configuration files.
68
3. Save the Base Ethernet switch configuration file to persistent storage and copy it to the server:
base-ethernet copy system:running-config nvram:startup-config copy nvram:startup-config tftp://<IP address>/<file-1> exit
4. Save the Fabric Ethernet switch configuration file and copy it to the server:
fabric-ethernet copy system:running-config nvram:startup-config copy 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-config copy system:running-config nvram:startup-config copy 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.
69
70
SOFTWARE FEATURES
This chapter describes the major software features of the SCM. In some cases, it describes usage of features. It also indicates where to find additional information. The SCM supports the management interfaces and features described below. Management Interface Command Line Interface (CLI) SNMP agent APIs (application programming interfaces) HPI (Hardware Platform Interface) IPMI (Intelligent Platform Management Interface) Feature PICMG-compliant Shelf Manager Ethernet switching (VLANs, spanning tree, link aggregation, Class of Service queuing, storm control, and more) IP routing Described in This manual, starting in Using the Command Line Interface on page 99. Details on the some portions of the CLI appear in other manuals as identified on page 100. This manual in SNMP Agent Support on page 89. In addition, the Ethernet Switching Software Reference describes controls for Layer 2 trap notifications. This manual in API Overview on page 95. The Shelf Management Software Reference. The Shelf Management Software Reference.
Description page 72 page 72 and Ethernet Switching Software Reference page 72 and IP Routing Software Reference page 73 page 74 page 77 page 79 page 82 page 83 page 83 and page 96 page 86 page 86 and page 97 page 86 and page 97
Procedures and Command Reference Shelf Management Software Reference Ethernet Switching Software Reference IP Routing Software Reference Linux and U-Boot documentation N/A N/A page 61 and page 79 page 63 N/A page 106 N/A API Reference page 101 and API Reference
Management Interfaces CLI, HPI, IPMI CLI, API, SNMP MIBs for certain aspects CLI
Linux and boot loader RAM and file system usage (JFFS, persistent files) IPv4 interface and Ethernet switch port identifiers Pre-defined IP addressing scheme (DHCP client, defaults) DHCP server Telnet and SSH Network timing subsystem (NTS) software (system clock) NTP server (time of day) Module state management Flash and file system management (moving files, persistency)
standard Linux and U-Boot standard Linux N/A standard Linux standard Linux standard Linux CLI, API standard Linux API CLI, API
71
6
Feature
Software Features
Software image management (versions) Firmware and software upgrades Logging service (syslog errors and messages) E-Key authorization and notification
Procedures and Command Reference API Reference Firmware and Software Upgrades page 67 and page 87 N/A
Management Interfaces API Linux utilities, API standard Linux, CLI None
Shelf Manager
The 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, such as controlling 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 switching
The SCM provides Ethernet switching software that runs on both the Base and Fabric switches. It 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 on page 89 and the Ethernet Switching Software Reference. For information about the API, see API Overview on page 95.
IP routing
The 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.
72
Linux services
All 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 client FTP client TFTP server and client Watchdog daemon BootP
73
Software Features
Linux prompt
The default Linux shell prompt can be very informative. It 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.
File persistence
Because 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 101 and Flash and file system management on page 86.) Other files must be included in a persistency directory.
74
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. Its contents 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. Its 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 filesgenerated each time the SCM bootsin /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 101.
75
Software Features
After reboot, the file appears in RAM. After this point, any changes to a bind mounted file are saved persistently. Any changes to an overlayed file are not.
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
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.
76
For example, 1003 for slot 3. 1021 1022 1023 1024 1025 1033 1034 1036
CLI Slot/Port Description 0/<slotnum> Backplane Base Ethernet connections to front slots, where For example, <slotnum> is the physical chassis slot number. 0/3 for slot 3. 1/5 1/6 1/7 1/8 1/9 2/1 2/2 2/4 10 Gbps Base Ethernet connection to front panel port 1/5. 1 Gbps Base Ethernet connection to front panel port 1/6. 1 Gbps Base Ethernet connection to front panel port 1/7. 1 Gbps Base Ethernet connection to front panel port 1/8. 1 Gbps Base Ethernet connection to front panel port 1/9. Backplane Base Ethernet connection to the other, redundant SCM (AdvancedTCA 3.0 Base channel 2). Base Ethernet connection to the LMP eth2 interface. Base Ethernet connection to a non-RadiSys Shelf Manager (AdvancedTCA 3.0 Base channel 1), to be used when the SCM is in a non-RadiSys shelf. Base Ethernet connections to the RTM through the zone 3 connector.
To LMP
Base pseudo-Ethernet connection to the LMP dtl0 interface, implemented over PCI.
77
Software Features
Table 28. Base Ethernet switch port identifiers and descriptions Base Switch Port SNMP ifIndex Link aggregation 1066 connections 1067 1068 1069 1070 1071 1072 1073 CLI Slot/Port 4/1 4/2 4/3 4/4 4/5 4/6 4/7 4/8 Description Base Ethernet link aggregation (LAG) logical portchannels. These exist only if created by users and are numbered in the order in which they were created.
Table 29. Fabric interface Ethernet interface identifiers and descriptions Fabric Switch Port SNMP ifIndex To front slots 2000 +
<slotnum>
For example, 2005 for slot 5. To front panel ports 2017 2018 2019 2020 To RTM 2049 2050 2051 To LMP 2065 Link aggregation connections 2066 2067 2068 2069 2070 2071 2072 2073 N/A
CLI Slot/Port Description 0/<slotnum> Backplane Fabric Ethernet connections to front slots, where For example, <slotnum> is the physical chassis slot number. 0/5 for slot 5. 1/1 1/2 1/3 1/4 3/1 3/2 3/3 N/A 4/1 4/2 4/3 4/4 4/5 4/6 4/7 4/8 5/1 to 5/28 10 Gbps Fabric Ethernet connection to front panel port 1/1. 10 Gbps Fabric Ethernet connection to front panel port 1/2. 10 Gbps Fabric Ethernet connection to front panel port 1/3. 10 Gbps Fabric Ethernet connection to front panel port 1/4. Fabric Ethernet connections to the RTM.
Fabric pseudo-Ethernet connection to the LMP dtl1 interface, implemented over PCI. 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.
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.
78
eth0:0
eth0:1 eth1
10.1.<LCHAS>.<SLOT>/ Subinterface. 16 (where <LCHAS> is the logical chassis and <SLOT> is the physical slot) None, interface not created Subinterface. 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. PCI bus connection between the Base switch and the LMP. Subinterface.
dtl0
dtl0:0
79
Software Features
Table 30. Default IP addresses and address usage for IP interfaces Interface Default IPv4 Address/ Name Subnet Mask dtl0:1 <ShelfMgrIPAddr>/ 24, which is 192.168.16.17/24 by default Description Used by the LMP to communicate with the Shelf Manager through the Base interface. Interface and IP Address Usage 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. This subinterface is configured through the Base interface CLI using the network parms command. See page 62. This is a pseudo-Ethernet interface.
dtl0:2 dtl1
None, interface not created Subinterface. Dynamically assigned by DHCP PCI bus connection between the Fabric switch and the LMP. None, interface not created Subinterface.
dtl1:0
This subinterface is configured through the Fabric interface CLI using the network parms command. See page 62.
80
The physical slot location uniquely identifies this SCMs location in the shelf. The logical slot is an easy way for The logical slot identifies the position of the SCM in the chassis the SCM to distinguish itself from in relation to the other SCM. The lower-numbered physical slot the other SCM. is logical slot 1. The higher numbered slot is logical slot 2.
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 client
Some 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/networkdevices/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. By default the DHCP client waits 300 seconds for a DHCP server to respond before timing out. If a timeout occurs, the interface is not initialized. The blade continues to boot up, and the DHCP client attempts to acquire an IP address in the background. The timeout can be configured in the /etc/template.dhclient.conf file. A value of 0 represents an infinite timeout. If you have Linux man pages installed on a system, refer to the dhclient.conf man page for more information.
81
Software Features
To request addresses for the first time, see Requesting IP addresses through DHCP on page 61.
Customization options
Users 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 81. 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.
DHCP server
The 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 SCMs DHCP server on page 65.
82
83
Software Features
Source priorities
The 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.
84
If the SCM's primary reference clock fails, 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. It 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.
85
Software Features
NTP server
The 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 67.
86
Logging service
Logging service
The 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 blade-wide syslog service is disabled, but logging is enabled for several components. Messages from the components are ignored unless the blade-wide service is enabled. The blades syslog configuration also determines the destination for the messages. See Enabling and disabling syslog and other services on page 67. 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.
87
Software Features
88
The SCM supports an SNMP agent that enables remote SNMP management of the SCM. 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 communicates with the subagents using the AgentX protocol. The master agent receives requests from the SNMP manager and completes the requested action, coordinating with its subagents when necessary. The master agent also generates notifications and relays notifications generated by the subagents to the registered notification destinations. The SNMP agent must be configured and enabled as described in Configuring SNMP trap destinations and security access on page 63 and Enabling the SNMP agent on page 64. The SCM implements its SNMP agent by incorporating Net-SNMP open-source software, version 5.2.1. Many usage details are covered in Net-SNMP man pages and on the http://www.net-snmp.com Web site. The SCM currently supports SNMPv2c.
89
90
Notification support
Notification support
The SCM supports the following notifications: coldStart (SNMPv2-MIB) authenticationFailure (SNMPv2-MIB) linkDown (IF-MIB) linkUp (IF-MIB) newRoot (BRIDGE-MIB) topologyChange (BRIDGE-MIB)
91
92
These files are located on the blade at /usr/share/snmp/mibs. To make it easy to copy these files elsewhere, the same directory contains a zipped archive of these files called mibs.tgz. A README.txt file located there lists these files and their sources.
93
94
API 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 SCMs LMP. The APIs introduced here are the: Ethernet switching API on page 96 NTS API on page 96 Module state management API on page 97 Flash and file system management API on page 97 Software image management API on page 98 Management plane framework (MPF) API on page 98
Reference information for the Module state management API, Flash and file system management API, and Flash and file system management API is located in the Blade Management API Reference provided in the software distribution. 95
NTS API
The network timing subsystem API allows you to configure the SCMs clock generation and select the network timing sources. It also allows you to retrieve the status and receive asynchronous NTS events. These events include: Loss of signal on any input Change of reference source state between valid and invalid Change in operating state, such as locked or holdover Temporary interruption to the currently selected reference source A few relevant details are as follows: Function name prefix: Shared library: Header file to include: RSYS_Nts librsysntsapi.so nts/rsys_nts_api.h
96
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.
97
98
The SCMs command line interface (CLI) allows you to interactively configure and monitor a wide range of services. This chapter describes how the CLI is divided and provides basic usage information that is useful throughout the different areas of the CLI. It refers to other locations for reference material on specific commands.
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.
99
Mode Description Configure the Base Ethernet switch. Configure the Fabric Ethernet switch, including optional IP routing features. provides, including Linux services, SNMP, and network timing. Access the Linux shell, which is not part of the CLI. When you exit the shell, you return to the CLI. Access shelf management, alarm, and FRU information.
blade-mgmt
Described in Ethernet Switching Software Reference and IP Routing Software Reference This manual on page 105 Standard Linux documentation Shelf Management Software Reference
linux-shell
Platform management
platform-mgmt
platform-mgmt#
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 102. 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 commands
Many 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
100
Show commands
Show commands
A 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.
Not necessary. The configuration is saved persistently when you execute the commands.
101
show running-config
Shows 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 tftp
Copies 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 blademanagement mode configuration. Syntax: Options:
nvram:startup-config copy tftp:<url> { nvram:startup-config | system:running-config }
copy system:running-config
Saves 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 blade is rebooted. This command currently saves only the blade-management mode configuration. Syntax: Options:
tftp:<URL> copy system:running-config {tftp:<URL> | nvram:startup-config}
Saves the configuration to the SCMs non-volatile memory. The saved configuration is reloaded when the SCM is rebooted.
102
copy nvram:startup-config
copy nvram:startup-config
Copies the persistently saved startup configuration to the specified URL, or reloads it so it becomes the running configuration. This command currently copies only the blademanagement mode configuration. Syntax: Options:
tftp:<URL> copy nvram:startup-config {tftp:<URL> | system:running-config}
erase nvram:startup-config
Erases the saved CLI configuration files from persistent storage. This returns aspects of the blade that are controlled through the CLI to the factory defaults. Syntax:
erase nvram:startup-config
exit
Leaves 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 blade. Syntax:
exit
logout
Exits the CLI and logs out from the blade. This command works only from the master CLI (top level). Syntax:
logout
help
Shows the editing and history keys. This is similar to what is shown on page 101. Syntax:
help
103
104
10
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. The command types include: SNMP commands on page 106 Network timing subsystem (NTS) commands on page 106
When the command reference pages below do not indicate a mode, use the top-level blademanagement mode.
Config
To 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.
105
10
SNMP commands
service snmpd
Enables the SNMP master agent by enabling it to listen on the UDP port 161 (without any options or with the start option), or disables it by keeping it from listening on any port (stop option). The restart option causes the agent to reset itself by performing both a stop and start, deleting any log files and starting with the current configuration files. The reconfigure option forces the agent to re-read its configuration files and change its operation based on their contents, without fully re-initializing. Syntax:
service snmpd [ start | stop | restart | reconfigure ]
Tip To ensure seamless communication with the subagents, the snmpd daemon is always running, and the stop option does not stop the daemon. Do not check for running processes to determine whether the SNMP agent is running. Instead, 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 64, SNMP Agent Support on page 89
106
10
Description The a half of the CLK3 bus, which is a backplane bus that is part of the PICMG-defined synchronization clock interface. A node blade can input a signal on CLK3. The b half of the CLK3 bus.
Frequency Input 8 kHz, 1.544 MHz (T1 and J1 modes only), 2.048 MHz (E1 mode only), or 19.44 MHz 8 kHz, 1.544 MHz (T1 and J1 modes only), 2.048 MHz (E1 mode only), or 19.44 MHz 8 kHz 8 kHz 8 kHz 6.48 MHz
clk3b
The first input from the rear transition module (RTM). The third input from the rear transition module (RTM). The fourth input from the rear transition module (RTM). 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. The second update channel clock received from the partner SCM.
The second input from the rear transition module (RTM). 8 kHz
xlink2*
2 kHz
xlink1 and xlink2 here could be called xlink1(rx) and xlink2(rx) to indicate that they are received clocks. The same names are used for transmitted clocks, but the paths are electrically distinct. Both SCMs transmit and receive a pair of xlink clocks at all times.
Description
Frequency Output
The a half of the CLK1 bus, which is a backplane bus that 8 kHz is part of the PICMG-defined synchronization clock interface. The b half of the CLK1 bus. 8 kHz The a half of the CLK2 bus, which is a backplane bus that 19.44 MHz is part of the PICMG-defined synchronization clock interface. The b half of the CLK2 bus. 19.44 MHz
107
10
sysclk1
The first output to the rear transition module (RTM). The outputs to the RTM are always being driven. The second output to the RTM. The third output to the RTM. 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.
2.048 MHz in E1 mode or 1.544 MHz in T1 or J1 mode 8 kHz 19.44 MHz 6.48 MHz
sysclk2 sysclk3
2 kHz
** xlink1 and xlink2 here could be called xlink1(tx) and xlink2(tx) to indicate that 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.
network-clock-select
Sets 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: Mode: Defaults:
network-clock-select <priority> { clk3a | clk3b | extclk1 | extclk2 | extclk3 | extclk4 }
Config Source:
xlink1 clk3a clk3b extclk1 extclk2 extclk3 extclk4
See also:
show network-clocks priority, show network-clocks source priority, Input clock definitions
108
network-clock-select mode
10
network-clock-select mode
Selects 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: Mode: Default: See also:
network-clock-select mode [ automatic | force { clk3a | clk3b | extclk1 | extclk2 | extclk3 | extclk4 } ]
no network-clock-select mode
Sets the reference clock source to automatic, which selects the highest priority clock. Syntax: Mode:
no network-clock-select mode
Config
network-clock-select role
Configures the SCM to act as the master, slave, or autonomous in relationship to the other SCM in the shelf. The master SCM provides the system clocks, while the slave stays synchronized with the master and outputs the same system clocks, if so configured, providing a seamless failover source. In the autonomous role, the SCMs do not communicate their timing or configuration information to each other, and the SCM may be under manual control of a management application such as a system manager. Syntax: Mode: Default: See also:
network-clock-select role [ master | slave | autonomous ]
Config the SCM in lower-numbered slot is the master, and the other SCM is the slave show network-clocks brief
no network-clock-select role
Sets the SCM roles to the default, making the SCM in the lower-numbered slot the master, and the other the slave. Syntax: Mode:
no network-clock-select role
Config
109
10
network-clock-select revertive
Sets the reference clock to revert to the original source if the signal is lost, the SCM fails over to another source, and the signal is restored. With revertive mode enabled, when a newly validated or revalidated source has a higher priority than the current reference source, the SCM transitions to the higher priority source. With revertive mode disabled, a source transition does not occur in that same situation, and the number of source transitions is minimized. Syntax: Mode: Default: See also:
network-clock-select revertive
Config Does not revert, stays with current source show network-clocks configuration
no network-clock-select revertive
Sets the reference clock to stay with the current source, even if the original source is restored after a failover. Syntax: Mode:
no network-clock-select revertive
Config
network-clock-select manage-bused-resource
Enables 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 cant 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: Mode: Default: See also:
network-clock-select manage-bused-resource
no network-clock-select manage-bused-resource
Disables E-Key management of the clock bus resources. Syntax: Mode:
no network-clock-select manage-bused-resource
Config
110
network-clock-select request-bused-resource
10
network-clock-select request-bused-resource
Requests 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: Mode: See also:
network-clock-select request-bused-resource {clk1 | clk2 | clk3}
no network-clock-select request-bused-resource
Releases 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: Mode:
no network-clock-select request-bused-resource { clk1 | clk2 | clk3 }
Config
network-clock-select output
Enables 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: Mode: Default: Example:
network-clock-select output { clk1 | clk2 } [ a | b ]
Config No output To configure the partner SCMs to provide redundancy as described in System clock redundancy on page 85, use this setup: On the master SCM:
network-clock-select network-clock-select network-clock-select network-clock-select request-bused-resource clk1 request-bused-resource clk2 output clk1 a output clk2 a
111
10
Either SCM can request the bus tokens and share them with the other SCM. See also: show network-clocks
no network-clock-select output
Disables output of the system clock signal on the selected clock bus. Syntax: Mode:
no network-clock-select output { clk1 | clk2 } [ a | b ]
Config
network-clock-select framer
Sets 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: Mode: Default: See also:
network-clock-select framer [ e1 | t1 | j1 ]
no network-clock-select framer
Resets the framing format expected from the framer clock to the default. Syntax: Mode:
no network-clock-select framer
Config
network-clock-select frequency
Specifies 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: Mode: Default: See also:
network-clock-select frequency { clk3a | clk3b } [ 8-kHz | 1544-kHz | 2048-kHz | 19.44-MHz ]
112
show network-clocks
10
no network-clock-select frequency
Sets the selected input clock back to the default frequency. Syntax: Mode:
no network-clock-select frequency { clk3a | clk3b }
Config
show network-clocks
Shows the frequency and enable status of each clock that the SCM may output. Syntax: Output:
Network Clock Output show network-clocks
The output clock name. The output clocks are described in Output clock definitions on page 107.
Output Frequency
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. 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. sources are described in Input clock definitions on page 107.
Role
Selected The source clock currently being used as a network timing reference. The Operating Mode
113
10
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 State
The current alarms for the PLL. The possible values are:
t0-soft-alarm - The PLL is tracking its input reference clock beyond the
problem.
RTM Present
Whether a rear transition module has registered its presence with the NTS API.
Partner Visible
Whether NTS is using the defined priorities to automatically select the input clock. Off indicates the forced selection of a user-specified input. 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. Whether the selected clock source reverts to the original if the original source signal is lost and subsequently restored.
Revertive
114
10
The output clocks the SCM is currently driving. For source clock descriptions, see Input clock definitions on page 107. The framing format expected on RTM or CLK3 source clocks. The bit used to receive and transmit the synchronization status message (SSM). The synchronization status message to transmit. Whether the ability to drive the clock buses (CLK1 - CLK3) is restricted by ownership of E-Key tokens.
Bused Resource
The clock bus E-Key tokens this SCM currently owns, allowing it (and its partner SCM) to drive clock signals on those buses.
network-clock-select For a description of the source clocks, see Input clock definitions on page 107.
The source with the highest priority. This may match the selected clock.
2nd Priority
The source with the 3rd priority. See also: For source clock descriptions, see Input clock definitions on page 107.
115
10
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 Local Partner
The input source, described in Input clock definitions on page 107. The validity status of the input from the perspective of this SCM. 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. The synchronization status message (SSM) value received on the input. The frequency of the input clock signal.
SSM Frequency
See also:
The source clock phase has drifted by more than 11.43 ppm. This alarm does not invalidate the source clock.
hard out-of-band
The source clock phase has drifted by more than 15.24 ppm. This alarm invalidates the source clock.
116
10
no-activity
The source is within the hard out-of-band limits, but shows no activity.
phase-lock
The 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-loss
The xlink1 or xlink2 update channel from the partner SCM no longer has a valid signal. If the slave SCM loses the masters xlink1 signal when it was previously valid, it causes a failover and becomes the master.
loss of frame
The SCM received an alarm indication signal condition (AIS) for this source.
yellow alarm
A backplane clock bus that is part of the PICMG-defined synchronization clock interface.
Token Owner
The 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. Board2Shelf
The last communication this SCM sent to the Shelf Manager regarding this token.
117
10
Shelf Reply
The last reply from the Shelf Manager to this SCM regarding this token, as a standard IPMI return code.
Shelf2Board
The last communication this SCM received from the Shelf Manager.
Board Reply
The last reply from this SCM to the Shelf Manager regarding this token.
118
MAINTENANCE
11
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.
119
11
Maintenance
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, the H/S blue light will never turn solid blue. 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.
Troubleshooting
When 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 59 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.
120
Troubleshooting
11
121
11
Maintenance
Table 32. Troubleshooting actions based on symptoms (Continued) Symptom Intermittently, the SCM and other installed modules experience random data errors. Recommendation Verify that the platforms 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. Verify no cover plates are installed. Empty slots must have air management panels installed to properly maintain airflow and emissions. Use shelf management software to check temperatures at the air intake, on the module, and at the platforms 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.
122
Troubleshooting
11
3. At the prompt to enter U-Boot, enter these key combinations within 2 seconds: Ctrl-x Ctrl-w Tip If your terminal doesnt 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 isnt working properly. Read Loading software from the secondary flash device on page 124. 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> 6. Transfer the software image to the SCMs RAM by entering: tftp 1000000 <image_name> For example: tftp 1000000 /ATCA-2210/release1.7.13_scm3-wr.bin for the file /tftpboot/ATCA-2210/release1.7.13_scm3-wr.bin on the TFTP server.
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 2:0-507 erase 2:0-507 cp 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 doesnt work. 8. Make the new software take effect by rebooting the SCM: reset If the software doesnt boot successfully, try the procedure again, or read Loading software from the secondary flash device.
123
11
Maintenance
124
SPECIFICATIONS
Environmental
RadiSys 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 33. Environmental Characteristic Temperature (ambient) State Operating Short term operating 1 Storage Value +5 C to +45 C 30 C/hr rate of change 5C to +55 C 30 C/hr rate of change 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 5% to 85% RH non-condensing 5% to 90% RH non-condensing at +30 C 5% to 90% RH non-condensing at +40 o C 5% to 95% RH non-condensing at +40 o C Up to 1800 meters (5,905 feet), +55C > 1800 meters up to 4000 meters (13,123 feet), derated linearly to +45C 0 to < 10 kg = 100 mm drop 10 to < 25 kg = 75 mm drop 0 to < 10 kg = 750 mm drop 10 to < 25 kg = 600 mm drop 300 mm free fall drop
Relative humidity Operating Short term operating1 Storage Altitude Short term storage1 Operating2
Shock (drop)
Unpacked Free fall, corners and edges Packaged (Unpalletized) Free fall, corners and edges Palletized
125
Specifications
Table 33. Environmental (continued) Characteristic Vibration State Operating Transportation (packaged) Value 0.1g, 5 to 100 Hz and back, 0.1 octave/min sine sweep 2 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 Seismic Operating Per Zone 4 test method, GR-63-CORE 1 Short term defined as a maximum of 96 consecutive hours. 2 For a maximum of eight consecutive hours.
Safety
The safety specifications are measured under laboratory ambient temperature and humidity (approximately 25C and humidity between 30% and 50%). Testing was performed in partnership with a Nationally Recognized Testing Laboratory (NRTL) accredited to provide the required certifications. Table 34. Safety Characteristic US Canada EU Other Certification Accessory Listing Approval Conformance with the Low Voltage Directive CB Report Standard and Test Criteria UL 60950-1 Safety for Information Technology Equipment CSA 22.2 #60950-1 Safety for Information Technology Equipment EN 60950-1 Safety for Information Technology Equipment IEC 60950-1 Safety for Information Technology Equipment
Mechanical dimensions
Table 35. Mechanical dimensions Characteristic PCB board Dimensions Board thickness Value 322.25mm x 280.0mm +0, -0.3mm (12.687 x 11.023 +0.0, -0.012) 2.0mm 0.2mm (0.079 0.007)
126
Radiated
Operating
Fast transient/burst
Operating
Surge voltages
Operating
Conducted
Operating
Operating
127
Specifications
Power
Table 38. Power emissions Characteristic Nominal operating voltage State -48V Standard and Criteria ETS 300 132-2 Static voltage levels
Table 39. Power consumption summary Supply Total converted from 12V DC/DC Conversion Loss (85%) COM Express 12V RTM 12V Total 12V brick power output Brick Conversion Loss (93%) IPMC and RTM 3.3V & 2.5V Total PIM output PIM Loss (97%)
Total -48V Power Required
Amp
4.47 2.80
Watt 86.9 15.3 26.0 33.6 161.8 12.2 3.0 177.0 5.5
182.5
128
MANAGED SENSORS
The SCM supports the following sensors, 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. The following table provides a summary of sensors that are managed by the SCM. Table 40. Managed sensors Sensor Name # 1 ATCA FRU hot swap 2 ATCA physical IPMB Reading type ATCA FRU hot Sensorswap specific ATCA physical SensorIPMB-0 specific Type Normal reading 0x00 -0x07 0x00-0xFF Notes Return M0 to M7 ATCA hotswap states. Bit [7] = IMPB B Override state Bit [6:4] = IPMB B Local state Bit [3] = IPMB A Override state Bit [2:0} = IPMB A Local status 0 1 Eject latch is open Eject latch is closed Fault status asserted Identity status asserted 0 1 Power supply A Power supply A Presence detected Failure detected 0 1 Power supply B Power supply B Presence detected Failure detected 0 1 Both A & B fuses OK Either A or B fuse blown Presence detected Failure detected 0 1 No SEL overflow SEL overflow detected 0 1 IPMC watchdog not expired IPMC watchdog expired Hardware version number. Firmware is specific to each hardware version.
Eject open
Digital
-48V absent A
Digital
-48V absent B
Power supply
Digital
Power supply
Digital
7 8 9
Digital Digital
0 0 0x00
Slot/connector Discrete
129
Managed Sensors
Table 40. Managed sensors (continued) Sensor Name # 10 +3.3_IPMI Type Voltage Reading Normal type reading Threshold Notes This sensor measures voltage in Volts. Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures voltage in Volts Default thresholds LC LNC UNC UC This sensor measures temperature in C Default thresholds LC LNC UNC UC 0.0 0.0 55 75
LNR 0.0 11 +12V Voltage Threshold LNR 0.0 12 +5V Voltage Threshold LNR 0.0 13 +3.3V Voltage Threshold LNR 0.0 14 +2.5V Voltage Threshold LNR 0.0 15 +1.8V Voltage Threshold LNR 0.0 16 +1.25V Voltage Threshold LNR 0.0 17 +1.2V Voltage Threshold LNR 0.0 18 Board temp Temp Threshold 25 LNR 0.0
UNR
UNR
UNR
UNR
UNR
UNR
UNR
UNR
UNR 85
130
Managed sensors
Table 40. Managed sensors (continued) Sensor Name # 19 COM Express temp* Type Temp Reading Normal type reading Threshold 25 Notes This sensor measures temperature in C Default thresholds LNR LC LNC UNC UC UNR 0.0 0.0 0.0 55 75 85 This sensor measures temperature in C Default thresholds LNR LC LNC UNC UC UNR 0.0 0.0 0.0 55 75 85 0 1 PIM alarm not asserted PIM alarm asserted 0 1 COM Express present COM Express not present 0 1 COM Express power COM Express power Fail note asserted Fail asserted 000-111 Module type 0 1 Asserted Not asserted 0 No failover 1 Communication loss with active 2 Active not found / active 3 Active issued M5 4 Set per status command 5 Active / re-arbitrated 6 Forced failover 0x00 Pending 0x01 Standby0 0x02 Standby1 0x03 Standby2 0x04 Active 0 1 RTM is present RTM is not present 0 1 RTM latch is open RTM latch is closed 0 1 2C is not ready 2C is ready RTM I RTM I
20
Fabric/HD temp
Temp
Threshold 25
21 22 23
Power supply
Digital
0 or 1 0 or 1 0 or 1
COM Express present Module/board Digital COM Express power fail COM Express type PB reset Shelf management controller (ShMC) failover Slot or connector Slot or connector Slot or connector Digital
24 25 26
Discrete Digital
110 or 100 0 or 1 0
Sensor-specific Discrete
27
ShMC STATECHG
Sensor-specific Discrete
3 or 4
(ShMC)
2
(Non-ShMC)
28 29 30
0 or 1 1 1
131
Managed Sensors
Table 40. Managed sensors (continued) Sensor Name # 31 RTM temp 0 Type Temp Reading Normal type reading Threshold 25 Notes This temperature measures in C Default thresholds LNR LC LNC UNC UC UNR 0.0 0.0 0.0 55 62 72 This temperature measures in C Default thresholds LNR LC LNC UNC UC UNR 0.0 0.0 0.0 55 62 72 0 1 RTM 12V and 3.3 power is RTM 3.3V or 12V power good failed 0 1 2 No pull-up Half pull-up Full pull-up 0 1 2 No pull-up Half pull-up Full pull-up
32
RTM temp 1
Temp
Threshold 25
33
Digital
34 35
Discrete Discrete
Table 41. Managed sensorsevent generation and board health Sensor Name # 1 2 3 4 5 6 7 8 9 10 11 12 13 ATCA FRU hot swap ATCA physical IPMB Eject open -48V absent A -48V absent B -48V fuse SEL overflow IPMC watchdog timer HW version +3.3V IPMI +12V +5V +3.3V Board healthy Local check scanning No No No No No No No No No No No No No Local event generation disabled No No No No No No No No No No Varies* Varies* Varies* Remote event Remote scanning generation disabled disabled No No No No No No No No No No No No No No No No No Varies* No No No Varies* No No Varies* Varies* Alarm level (if applicable) N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Minor, Major, Critical Minor, Major, Critical Minor, Major, Critical
132
Table 41. Managed sensorsevent generation and board health (continued) Sensor Name # 14 15 16 17 18 19 20 21 22 23 +2.5 +1.8V +1.25V +1.2V Board temp COM Express temp Fabric/HD temp Board healthy Local check scanning No No No No No No No Local event generation disabled Varies* Varies* Varies* Varies* No No No Remote event Remote scanning Alarm level generation disabled (if applicable) disabled No Varies* Minor, Major, Critical No Varies* Minor, Major, Critical No Varies* Minor, Major, Critical No Varies* Minor, Major, Critical No No Minor, Major, Critical No No Minor, Major, Critical No No Minor, Major, Critical No No N/A No Varies* N/A No Varies* N/A No No No No No No No No No No No No Varies No No No Varies** Varies** Varies** Varies** Varies* Varies* No No N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
PIM status No No COM Express present No Varies* COM Express power No Varies* fail 24 COM Express Type No Varies* 25 PB reset No No 26 ShMC failover No No 27 ShMC STATECHG No No 28 RTM PRSNT# No Varies** 29 RTM eject open No Varies** 2C ready 30 No Varies** RTM 1 31 RTM temp0 No Varies** 32 RTM temp1 No Varies* 33 RTM PWR No Varies* Fail_12V_3V 34 IPMB PU Ch A No No 35 IPMB PU Ch B No No * Event generation is enabled when the payload power is turned on. ** Event generation is enabled if an RTM is detected.
133
Managed Sensors
134