3Com 4007 Release Note

http://www.3com.com/

Part No. 10013820

Published August 2000

Switch 4007

Release Notes

Management Module, Release 3.0.5

Layer 2 Switch Fabric Modules, Release 3.0.5

Layer 2 Switching Modules, Release 3.0.5

Multilayer Switching Modules, Releases 3.0.5

3Com Corporation

5400 Bayfront Plaza

Santa Clara, California

95052-8145

Copyright © 2000, 3Com Corporation. All rights reserved. No part of this documentation may be reproduced in any form or by any means or used to make any derivative work (such as translation, transformation, or adaptation) without written permission from 3Com Corporation.

3Com Corporation reserves the right to revise this documentation and to make changes in content from time to time without obligation on the part of 3Com Corporation to provide notification of such revision or change.

3Com Corporation provides this documentation without warranty, term, or condition of any kind, either implied or expressed, including, but not limited to, the implied warranties, terms, or conditions of merchantability, satisfactory quality, and fitness for a particular purpose. 3Com may make improvements or changes in the product(s) and/or the program(s) described in this documentation at any time.

If there is any software on removable media described in this documentation, it is furnished under a license agreement included with the product as a separate document, in the hard copy documentation, or on the removable media in a directory file named LICENSE.TXT or !LICENSE.TXT. If you are unable to locate a copy, please contact 3Com and a copy will be provided to you.

UNITED STATES GOVERNMENT LEGEND

If you are a United States government agency, then this documentation and the software described herein are provided to you subject to the following:

All technical data and computer software are commercial in nature and developed solely at private expense.

Software is delivered as “Commercial Computer Software” as defined in DFARS 252.227-7014 (June 1995) or as a “commercial item” as defined in FAR 2.101(a) and as such is provided with only such rights as are provided in 3Com’s standard commercial license for the Software. Technical data is provided with limited rights only as provided in DFAR 252.227-7015 (Nov 1995) or FAR 52.227-14 (June 1987), whichever is applicable.

You agree not to remove or deface any portion of any legend provided on any licensed program or documentation contained in, or delivered to you in conjunction with, these Release Notes.

Unless otherwise indicated, 3Com registered trademarks are registered in the United States and may or may not be registered in other countries.

3Com, CoreBuilder, SuperStack, and Transcend are registered trademarks of 3Com Corporation. The 3Com logo is a trademark of 3Com Corporation. 3Com Facts is a service mark of 3Com Corporation.

Netscape is a registered trademark of Netscape Communications Corporation in the United States and other countries. Java is a registered trademark of Sun Microsystems, Inc. UNIX is a registered trademark in the United

States and other countries, licensed exclusively through X/Open Company, Ltd.

All other company and product names may be trademarks of the respective companies with which they are associated.

C

ONTENTS

1 O

VERVIEW

Scope 11

Important Notices 12

System Requirements 14

2 R ELEASE H IGHLIGHTS

Accelerated Address Aging 16

Purpose 16

How It Works 16

Recommendations 17

Interaction with Transcend Enterprise VLAN Manager 17

Advanced Autonegotiation 18

Sample Problem Description 18

How Advanced Autonegotiation Works 18

Configuration Options 19

Configuring Standard Autonegotiation 19

Configuring a Fixed Port State 19

Configuring Advanced Autonegotiation or Port Monitoring 19

Sample Configurations and System Behavior 20

Configuration A 20

Configuration B 20

Configuration C 21

Configuration D 21

Configuration E 21

Information Display 22

Important Considerations 22

LX 70 GBIC Transceiver 23

Trap Filtering 24

Available Trap Filters 24

Managing Traps 24

3 R ELEASE I SSUES F OR M ANAGEMENT M ODULES

Corrected Problems in Release 3.0.0

28

Operating Considerations for Release 3.0.5

29

Diagnostic events 29

Chassis Reset 29

Software Transfers 29

Password Access 30

System and Module Naming 30

Web Management 30

Known Problems in Release 3.0.5

30

Diagnostics 30

SLIP Interfaces 31

System Time 31

Software Downloads 31

4 R ELEASE I SSUES F OR S WITCH F ABRIC M ODULES

Corrected Problems in Release 3.0.0

34

System 34

Ethernet 34

Bridging 34

VLAN 34

Corrected Problems in Release 3.0.5

35

Bridging 35

Management 35

Problems Fixed in v3.0.5 Address Table 36

Problem Overview 36

Problem Detail 36

Work Around 37

Recommendations 37

Guidelines from FR003 TRANSCEND UNIX 5.0 UPDATE 37

Important Notes 39

Operating Considerations for Release 3.0.5

41

Staging 41

System 41

Management 41

Web-based Management 41

Ethernet 42

Transmit and Receive Ports on GBIC Modules 42

Bridging 42

Trunking 42

Resilient Links 43

VLANs 43

Known Problems for Release 3.0.5

44

System 44

Management 44

Ethernet 45

Trunking 45

VLANs 46

5 R

ELEASE

I

SSUES FOR

L

AYER

2 S

WITCHING

M

ODULES

Corrected Problems in Release 3.0.0

48

System 48

Management 48

Ethernet 48

Bridging 48

VLANs 49

Corrected Problems in Release 3.0.5

49

Bridging 49

Management 50

NVRAM 51

Problems Fixed in v3.0.5 Address Table 51

Problem Overview 51

Problem Detail 52

Work Around 52

Recommendations 53

Guidelines from FR003 TRANSCEND UNIX 5.0 UPDATE 53

Important Notes 54

Operating Considerations for Release 3.0.5

57

System 57

Web-based Management 57

SNMP 57

57

Ethernet 57

Bridging 58

Trunking 59

VLANs 59

Known Problems in Release 3.0.5

60

System 60

Management 60

Ethernet 61

Bridging 61

Trunking 62

6 R

ELEASE

I

SSUES FOR

M

ULTILAYER

S

WITCHING

M

ODULES

Corrected Problems in Release 3.0.0

66

System 66

Bridging 66

Virtual LANs (VLANs) 66

IP 66

QoS 66

Roving Analysis 66

Web Console 66

Ethernet 66

4-port Gigabit Ethernet Layer 3 Switching Module (GBIC) 67

Corrected Problems in Release 3.0.5

67

QoS 67

RMON 67

RMON-2 Network Management Applications 67

IP Policies 67

OSPF 68

VRRP 68

Problems Fixed in v3.0.5 Address Table 68

Problem Overview 68

Problem Detail 69

Work Around 69

Recommendations 70

Guidelines from FR003 TRANSCEND UNIX 5.0 UPDATE 70

Important Notes 71

Operating Considerations in Release 3.0.5

74

System 74

SNMP 74

RMON 75

Ethernet 75

Bridging 75

Backplane Port Trunking 76

Trunking 76

Packet Filtering 77

Virtual Router Redundancy Protocol (VRRP) 77

IP 77

IP Multicast 77

AppleTalk 77

QoS 78

Roving Analysis 79

Filter Builder 79

Transmit and Receive Ports on GBIC Modules 79

Web-based Management 79

VLANs 79

4-port GBIC Multilayer Switching Module (3CB9RG4) 80

Known Problems in Release 3.0.5

80

System 80

SNMP 80

Ethernet 81

Bridging 81

Trunking 82

VLANs 83

Internet Protocol (IP) 83

OSPF 83

Virtual Router Redundancy Protocol (VRRP) 84

IP Multicast 84

IPX 84

QoS 84

RSVP 85

RMON 85

Web-based Management 86

4-port GBIC Multilayer Switching Module (Model

Number 3CB9RG4) 86

7 D OCUMENTATION C HANGES

Terminology 88

Ethernet 88

Internet Packet Exchange (IPX) 88

Important Considerations 89

Quality of Service (QoS) 90

Web-based Management 90

Resetting Modules 91

SNMP Traps 92

EME Command Changes in Release 3.0.5

92

Old command structure in software release 3.0.0: 92

New command structure in software release 3.0.5: 92

Examples 93

Setting SNMP Trap Destinations 94

8 S OFTWARE I NSTALLATION

Verifying Module Software Releases and Compatibility 96

Verifying Software Releases on an Existing System 96

Verifying Software Releases on a New System and Modules 98

How to Obtain the Latest Software 100

Where to Place the Software 100

Terminology for the EME Module 101

Clarifying Acronym Use 101

Terms for Redundant EME Modules 101

Prerequisites for Software Installation 102

Upgrading the EME Module to Release 3.0.5

104

Upgrading a Single EME Module 105

Upgrading Two EME Modules 106

Installing EMC Boot Software 108

Installing EMC Operational Software 108

Installing EME Boot Software 109

Installing EME Operational Software 109

Upgrading Layer 2 Switching Modules to Release 3.0.5

110

Upgrading Multilayer Switching Modules From 2.1 to 3.0.5

111

Upgrading Multilayer Switching Modules From 3.0 or 3.0.1 to 3.0.5

114

Transition Issues from Release 2.1 to 3.0

114

Upgrade Procedure 116

Updating Emergency Download Parameters 117

Downgrading Modules From 3.0.5 to 3.0.0 or 3.0.1

119

9 R EFERENCE

MIB Files 121

Entering Commands 122

Year 2000 Compliance 122

10 T ECHNICAL S UPPORT

Online Technical Services 123

World Wide Web Site 123

3Com Knowledgebase Web Services 123

3Com FTP Site 124

3Com Bulletin Board Service 124

Access by Analog Modem 124

Access by Digital Modem 125

3Com Facts Automated Fax Service 125

Support from Your Network Supplier 125

Support from 3Com 125

Returning Products for Repair 127

O

VERVIEW

1

Scope

These release notes summarize operational requirements and issues for the following Switch 4007 modules and software releases:

Table 1 Software Release Numbers Addressed in These Release Notes

Description

Management Module (EME)

Model Number Software Release

3CB9EME Management software —

Boot code 3.0.0

3CB9FG9

Operational code 3.0.5

Controller software —

Boot code 2.1.0

Operational code 3.0.0

3.0.5

9-port Gigabit Ethernet Layer 2

Switching Fabric

24-port Gigabit Ethernet Layer 2

Switching Fabric

4-port Gigabit Ethernet Interface

Module

3CB9FG24T

3CB9LG4

3.0.5

3.0.5

20-port 100BASE-FX Layer 2

Switching Module

36-port 10/100BASE-TX Layer 2

Switching Module

9-port Gigabit Ethernet Layer 2

Switching Module

12-port 10/100BASE-TX

Multilayer Switching Module

10-port 100BASE-FX

Multilayer Switching Module

4-port Gigabit Ethernet

Multilayer Switching Module

3CB9LF20MM 3.0.5

3CB9LF36R

3CB9LG9MC

3CB9RF12R

3CB9RF10MC

3CB9RG4

3.0.5

3.0.5

3.0.5

3.0.5

3.0.5

Filenames eme30000.bt

eme30005.op

emcv20100.bt

emc30000.op

fg930005.all

fga24t30005.all

Not applicable; module passes traffic directly to switch fabric.

lf20mm3005.all

lf36r30005.all

lg9mc30005.all

rf12r30005.all

rf10mc30005.all

rg430005.all

12 C HAPTER 1: O VERVIEW

Important Notices

Read these important notices before you begin.

The Switch 4007 management interfaces refer to the Switch 4007

Management Module as the Enterprise Management Engine (EME). In these release notes, the Switch 4007 Management Module is referred to as the EME for brevity.

The term Administration Console refers to the menu-driven, command line interface on switching modules; each switching module has its own

Administration Console. The terms EME prompt , EME console , or system prompt refer to the command line interface that you first encounter when you connect to the system. It is from this interface that you connect to a switching module’s Administration Console, modify system settings such as power and temperature, modify settings on the EME module, or arrange software downloads.

The 10-port 100BASE-FX Multilayer Switching Module (3CB9RF10MC) that is listed in Table 1 is new to the Switch 4007 product line at this release.

Although the heritage of the Switch 4007 product line is the

CoreBuilder 9000 product line, these Switch 4007 release notes do not cover all aspects that apply to the CoreBuilder 9000. These release notes include only information about the subset of modules that are officially supported under the Switch 4007 brand and about configurations that apply to the Switch 4007 (for example, issues that deal with dual switch fabrics are not included because the Switch 4007 chassis only has one fabric slot). For a list of modules that are officially supported in a

Switch 4007 system, see Table 1. If you have a CoreBuilder 9000 system or if you are operating unsupported modules in a Switch 4007 system, we recommend that you download from the 3Com Web site the suite of

CoreBuilder 9000 Release Notes for Software Release 3.0.5.

To obtain later software maintenance releases or upgrades and associated release notes for the Switch 4007, visit the 3Com Web site: http://support.3com.com/

For comprehensive, up-to-date information about the Switch 4007 and other 3Com products, see the online user guides on the 3Com Web site: http://support.3com.com/nav/switches.htm

Important Notices 13

The Switch 4007 software and management interfaces are built from

CoreBuilder ® 9000 switch technology. In Switch 4007 software, the interface prompts and displays indicate this heritage. The examples, tables, and figures in the Switch 4007 documentation may also indicate this heritage.

CAUTION: Except for the EME which operates four different files on mixed release numbers (see Table 1), all modules in the Switch 4007 chassis must run software release 3.0.5 for functional reasons. The software for CoreBuilder 9000 modules and Switch 4007 modules is identical. Upgrade the EME first, followed by the switching modules. For software download and installation instructions, see the chapter

“Software Installation” later in these release notes.

CAUTION: The Switch 4007 Setup Wizard, supplied on the Software

CD-ROM, is designed to simplify installation of the switch. Release 1.0 of the Setup Wizard enables you to enter an invalid IP address or subnet mask, which prevents Telnet access to the EME management module.

3Com recommends that you do not use the Setup Wizard until an updated version is available from the 3Com Web site.

CAUTION: To run the Transcend® Network Supervisor application (for managing and monitoring Switch 4007 activity), the Microsoft® Virtual

Machine and Microsoft WinSock2 applications may be required on your

PC. During installation, Transcend Network Supervisor will check your

PC's configuration and prompt you if these files are needed. If required, you should exit from the Transcend Network Supervisor installation utility and proceed to download (from the Microsoft web site) and install each of these applications separately. After they are installed, you can restart the installation utility for Transcend Network Supervisor.

CAUTION: This release requires that the EME have 20 MB of CPU RAM.

To determine the current CPU RAM size, enter show eme

from the EME command line interface. The

CPU Ram Size (MB)

field on the left side of the screen indicates the amount of CPU RAM:

8

or

20

. If your EME has

8 MB of CPU RAM, you must replace one of the two 4 MB DRAM cards with a 16 MB DRAM card (order number 3CB9EE). After you install it, you will have the required 20 MB of CPU RAM. To order a 16 MB DRAM card, contact your network supplier or 3Com representative.

14 C HAPTER 1: O VERVIEW

System

Requirements

CAUTION: Before you attempt to download any module software, save the configuration file for the modules that you are upgrading to an external device using the upload

command from the EME prompt. Name the files so that they uniquely identify the versions of the module code that were in use at the time they were uploaded, as well as the chassis and slot where the module is located. See the Switch 4007

Implementation Guide for an explanation about how to use this command. You must save the configuration files because they are not guaranteed to work with revisions of code other than the ones with which they were created. If at some point you want to downgrade the modules to preexisting code, you must use the configuration file that corresponds to that version of code.

Each Switch 4007 chassis has these minimum system requirements:

■

One power supply.

■

■

■

■

To increase device and network availability, you can optionally install a second power supply.

One fan assembly.

One switch fabric module with software that is compatible with other modules in the chassis.

Depending on the other modules that are installed in the chassis and their associated throughput requirements, choose either the 24-port

Gigabit Ethernet Switch Fabric Module (Model Number 3CB9FG24T) or the 9-port Gigabit Ethernet Switch Fabric Module (Model

Number 3CB9FG9).

One Management Module (Model Number 3CB9EME) with 20 MB of

CPU RAM and software that is compatible with other modules in the chassis.

To increase device availability and network availability, you can optionally install a second Management Module (EME).

An appropriate number of Interface Modules, Layer 2 Switching

Modules, or Multilayer Switching Modules with software that is compatible with other modules in the chassis.

2

R

ELEASE

H

IGHLIGHTS

Software Release 3.0.5 is a maintenance release, however it also includes a few new features. The applicability of new features to the Switch 4007 module types is outlined in Table 2.

Table 2 New Features on Switch 4007 Modules

Accelerated

Address Aging

Advanced

Autonegotiation

Support for LX 70

GBIC Transceiver

Trap Filtering

Layer 2

Switching

Modules

YES

YES

YES

NO

Layer 2

Switch Fabric

Modules

YES

Multilayer

Switching

Modules

NO

Management

Modules

(EME)

NO

NO

YES

NO

YES

YES

NO

NO

NO

YES

The remaining sections in this chapter provide descriptions of the features.

16 C HAPTER 2: R ELEASE H IGHLIGHTS

Accelerated

Address Aging

For all Layer 2 Switching Modules in Release 3.0.5, a third option has been added to the bridge spanningTree stpState

menu called agingOnly

(other options are enable

and disable

).

Purpose If you select the agingOnly

option, the Spanning Tree Protocol (STP) is disabled on the module but the address table’s aging interval will change

(becomes shorter) whenever the module detects that one of its links has gone down. The objective is to flush addresses as quickly as possible.

Although STP is not operating on the module, this temporary, alternate aging process draws upon a couple of the STP parameter values. These can be system defaults or customized values.

How It Works Under normal operating conditions, the Layer 2 module uses the aging interval value that you set in the bridge agingTime

command. However, with the agingOnly

option selected, whenever a module detects that one of its links has gone down, the aging interval temporarily changes to the value that you set in the bridge spanningTree stpForwardDelay option. Usually, this interval is much shorter than the bridge agingTime and thus addresses age more quickly. After a period of time equal to bridge spanningTree stpForwardDelay

+ bridge spanningTree stpMaxAge

seconds has transpired, a module then reverts to using the value in bridge agingTime

as the aging interval.

Essentially, the agingOnly

option mimics the action that STP takes when ports go down, but does so without having STP enabled on the module.

The accelerated aging process affects all addresses in the Layer 2 module’s address table; it is not limited to the addresses that are associated with the down port.

This feature was not implemented on Layer 3 modules because it was not needed. Whether STP is enabled or disabled, whenever a port has gone down or is disabled, Layer 3 modules immediately flush all addresses that are associated with that specific port.

Accelerated Address Aging 17

Recommendations If you select the agingOnly

option, 3Com recommends that you also set the stpForwardDelay

value to the lowest number in the allowed range.

This will ensure that addresses flush as quickly as possible.

If you are using Transcend Enterprise VLAN Manager on your network but are not running STP on your module, 3Com recommends that you select the agingOnly

option.

Interaction with

Transcend Enterprise

VLAN Manager

The agingOnly

option in Release 3.0.5 is especially helpful if you use

3Com Transcend Enterprise VLAN Manager software because it helps to prevent incorrect MAC address-based VLAN assignments.

Prior to Release 3.0.5, if you did not have STP enabled, a Layer 2 module would not flush MAC addresses upon a link down event. It would only age addresses according to the value set in bridge agingTime

. This situation can cause VLAN-assignment problems when endstations are frequently plugged and unplugged into ports on the module and the

Transcend Enterprise VLAN Manager (EVM) application is used on the network. That is, if a new endstation (with a new MAC address) is plugged into a port prior to the aging out of the old MAC address on that port, EVM may assign the endstation to the VLAN to which the old MAC address belonged, which could be an incorrect VLAN for the new endstation.

In summary, prior to Release 3.0.5, if you wanted addresses to age more quickly after a link down event, you would have to enable STP. Now, with the agingOnly

option, you do not have to do that.

18 C HAPTER 2: R ELEASE H IGHLIGHTS

Advanced

Autonegotiation

Advanced Autonegotiation is an enhancement to the standard autonegotiation feature that is defined in the IEEE 802.3u specification, which addresses how two ends of a 10/100-TX Ethernet connection become synchronized in speed and duplex mode. Due to limitations of the specification, certain configurations produce duplex mismatches that generate errors, reduce network performance, and may even “lock up” an entire 802.3u-compliant switch. 3Com developed Advanced

Autonegotiation to overcome these limitations and correct duplex mismatches on 10/100 links. In addition, although the 802.3u specification does not apply to 100FX ports, Advanced Autonegotiation also helps correct duplex mismatches on these links.

Sample Problem

Description

To illustrate the problem with the 802.3u specification:

Imagine a switch-to-switch connection (or switch-to-PC NIC connection) where port A has standard autonegotiation enabled but port B does not and instead is fixed at 100 Mbps and full duplex mode. The auto-detection process in standard autonegotiation allows port A to detect and match the speed of port B but port A cannot detect port B's duplex mode and therefore port A defaults to half duplex mode per the specification. Thus, a duplex mismatch exists and causes problems for the link, switch, and network.

This problem affects any vendor’s switch that complies with the

IEEE 802.3u specification. It is not unique to 3Com.

In this example, if port B had been fixed in half duplex mode, then a duplex mismatch would not have occurred.

How Advanced

Autonegotiation

Works

New to Software Release 3.0.5, Advanced Autonegotiation is related in functionality to the Ethernet Port Monitoring feature that was introduced in Software Release 3.0.

Port Monitoring monitors the error counters on specified ports and provides notification via traps and messages on the management interfaces. Port monitoring can also partition (temporarily disable) ports that exceed a system error threshold.

Advanced Autonegotiation 19

As with Port Monitoring, Advanced Autonegotiation also monitors port error levels and provides notification, but goes further by changing a port’s mode into full duplex mode. However, if errors continue to exceed the system threshold, Advanced Autonegotiation partitions the port.

Configuration

Options

To provide flexibility, the Advanced Autonegotiation and Port Monitoring features are enabled and disabled separately from standard autonegotiation. The benefits of this arrangement include:

„

„

T he Advanced Autonegotiation and Port Monitoring features can also work for 100FX ports, which do not have an autonegotiation process.

Port Monitoring can be enabled for ports that have been set to operate a fixed speed and duplex mode (autonegotiation is disabled).

Configuring Standard Autonegotiation

To enable or disable standard autonegotiation, at the CLI prompt, enter: ethernet autoNegotiation

Then select a port or range of ports and enter either enable or disable .

Configuring a Fixed Port State

To set a fixed port state (standard autonegotiation is disabled), at the CLI prompt, enter: ethernet portMode

Then select a port or range of ports and enter the appropriate speed and duplex mode.

Configuring Advanced Autonegotiation or Port Monitoring

To enable or disable either Advanced Autonegotiation or Port Monitoring, at the CLI prompt, enter: ethernet monitoring mode

Then select a port or range of ports, and enter one of these options:

„ disabled

— neither Port Monitoring nor Advanced Autonegotiation is enabled.

„

„ enabled

— Port Monitoring is enabled advanced

— Advanced Autonegotiation is enabled (includes the Port

Monitoring function)

20 C HAPTER 2: R ELEASE H IGHLIGHTS

Sample

Configurations and

System Behavior

This section describes how the system behaves in certain feature configurations. Refer to Table 3 when reading the corresponding paragraphs below.

Table 3 Sample Configurations

Configuration A

Configuration B

Configuration C

Configuration D

Configuration E

Standard

Autonegotiation

Enabled

Yes

Port

Monitoring

Enabled

No

Yes

Yes

No

No

Yes

No

Yes

No

Advanced

Autonegotiation

Enabled

No

No

Yes

No

Yes

Configuration A

Ports autonegotiate connections according to the IEEE 802.3u specification. Error conditions are not monitored. No alerts are sent and preventive actions are not taken. If the remote port (the port to which the switch is connected) is fixed in full duplex mode or uses a non-compliant version of autonegotiation, duplex mismatches are likely to occur, causing problems for the link, switch, and portions of the network that rely on that switch. However, if standard autonegotiation is enabled on the switch and the remote port, the speed and duplex modes will be set correctly.

Configuration B

Ports autonegotiate connections according to the IEEE 802.3u specification. The system monitors specified ports for errors and sends alerts through traps and console messages. If the remote port is fixed in full duplex mode or uses a non-compliant version of autonegotiation, duplex mismatches are possible. However, the Port Monitoring feature can partition ports with high error rates (which, incidentally, may also be caused by factors other than duplex mismatches such as damaged cabling).

The error threshold is fixed in system software; it is not user-configurable.

The initial time period in which a port is partitioned (disabled) is 30 seconds. When the system reenables the port, it also immediately monitors the error conditions. If the error threshold is reached again, the port is partitioned again. The time that the port is partitioned increases with each incidence, up to a maximum of 5 minutes.

Advanced Autonegotiation 21

Configuration C

Ports autonegotiate connections according to the IEEE 802.3u specification. The system monitors specified ports for errors and sends alerts through traps and console messages. If a high error rate is detected on a given port, Advanced Autonegotiation forces that port into full duplex mode. The system continues to monitor the error rate to determine if that action solved the problem or not. If errors above the system threshold continue to exist, then Advanced Autonegotiation partitions the port.

Configuration D

Autonegotiation does not occur because it is not enabled or does not exist (100 FX). Thus, for 10/100 TX ports, the port speed and duplex mode are fixed in some combination of 10 or 100 Mbps and half or full duplex. 100 FX ports are fixed in either 100 half duplex or 100 full duplex. The Port Monitoring feature monitors for errors and generates alerts through traps and console messages. If the number of errors exceed the system threshold, the system partitions the port.

Configuration E

Autonegotiation does not occur because it is not enabled or does not exist (100 FX). Thus, for 10/100 TX ports, the port speed and duplex mode are fixed in some combination of 10 or 100 Mbps and half or full duplex. 100 FX ports are fixed in either 100 half duplex or 100 full duplex. The Advanced Autonegotiation feature responds differently, depending on the type of link:

„ 10/100 TX links — The system monitors for errors and generates alerts. If error thresholds are exceeded on 10/100 links, the system can only partition the port because the duplex mode has been fixed. It cannot change the mode from half duplex to full duplex or vice versa.

If you want the complete functionality of Advanced Autonegotiation on a given 10/100 port, ensure that you also enable standard autonegotiation on that port.

„ 100 FX links — The system monitors for errors and generates alerts. If error thresholds are exceeded, it can change the duplex mode from half duplex to full duplex and vice versa. The system continues to monitor the error rate to determine if the duplex change solved the problem or not. If the error rate continues to exceed the system threshold, then Advanced Autonegotiation partitions the port.

22 C HAPTER 2: R ELEASE H IGHLIGHTS

Information Display From the ethernet monitoring

menu, the summary

option provides a display of information.

Some fields, such as portNumber

, portLabel

, and various error counter fields, are repeated from the ethernet detail

display for your convenience if you are troubleshooting.

However, there are three new fields:

„ monitorMode

— Displays the current setting from the ethernet monitoring mode

option ( disabled, enabled,

or advanced

).

„

„

„

„ monitorStatus

— Indicates the status of Port Monitoring or

Advanced Autonegotiation on the port. There are three possible entries in this field:

Normal

— The port state has not been modified.

Advanced

— The port has been forced into the full duplex state.

„ Partitioned

— The port is temporarily disabled. (The word

“partitioned” also appears in the portState

field in the ethernet summary

and ethernet detail

displays.)

MonitorErrors

– Indicates the number of times that the port has transitioned into the partitioned state.

Important

Considerations

„

„

„

„

„

The initial time period in which a port is partitioned (disabled) is 30 seconds. When the system reenables the port, it also immediately monitors the error conditions. If the error threshold is reached again, the port is partitioned again. The time that the port is partitioned increases with each incidence, up to a maximum of 5 minutes.

„

The error thresholds are fixed in system software; they are not user-configurable.

Partitioned ports are disabled at the link level. That is, both the PHY and MAC components are shut down so that the switch no longer forwards traffic to that port. This action is taken for the following reasons:

It prevents a switch from being congested due to the condition known as “head-of-line blocking”

It allows resilient link connections (see the bridge link

menu) to switchover and thus minimizes network disruption

The attached device may self-heal during renegotiation or reboot

LX 70 GBIC

Transceiver

LX 70 GBIC Transceiver 23

„

„

„

Partitioned ports may cause the Spanning Tree Protocol (STP) to calculate and establish new paths. Then, when the ports are released from their partitioned state, this process may recur.

You cannot enable any form of autonegotiation on existing trunked links. Conversely, if you have autonegotiation enabled on ports and then create a trunk with those ports, the system automatically disables autonegotiation on those ports. Trunked links must always operate at the same speed and duplex mode. 3Com recommends that you fix the port speed and duplex mode at both ends of the connection prior to creating trunks.

When Advanced Autonegotiation forces a port into full duplex mode, that setting is not saved into nonvolatile memory. Thus, this setting does not survive a system reset or reboot.

Software Release 3.0.5 provides support for the Long Haul (70km) GBIC

Transceiver (Model Number 3CGBIC97).

The Gigabit Interface Converter (GBIC) ports on your system connect to

Gigabit Ethernet networks through a GBIC transceiver, providing a high-speed connection over fiber-optic cable. The transceiver connects to the network using a fiber-optic duplex Subscriber Connector (SC).

The Long Haul (70km) GBIC Transceiver can handle transmission rates of up to 1.25 Gbit/sec and distances of up to 70km over single mode fiber.

To ensure optimal performance, compatibility, and regulatory compliance, use only GBIC transceivers that 3Com supports. For a list of GBIC devices that are qualified for the CoreBuilder 3500, CoreBuilder 9000 (Switch

4007), and CoreBuilder 9400, see this 3Com Web site: http://www.3com.com/gigabit_ethernet/gbics

24 C HAPTER 2: R ELEASE H IGHLIGHTS

Trap Filtering

In Software Release 3.0.5, the command interface for managing SNMP traps has changed and new features have been added for filtering

(enabling or disabling) link state, topology change, and authentication traps.

Available Trap Filters These trap filters are available:

„ set snmp trap filter link_state

Enables or disables the generation of link state traps. A bridge sends a link state trap when one of its configured ports transitions from the

Up state to the Down state, or from the Down state to the Up state.

„ set snmp trap filter topology_change

Enables or disables the generation of topology change traps, which are sent by a bridge when one of its ports moves from the Learning

State to the Forwarding state, or from the Forwarding state to the

Blocking state.

The trap is not sent if a newRoot

trap is sent for the same transition. If you disable topology change traps, you will still see newRoot

traps which indicate that the sending agent has become the new root of the Spanning Tree.

„ set snmp trap filter authentication

Enables or disables the feature that sends an alert to the management workstation when someone tries to gain access to the EME, and the IP address or community name is not valid for the attempted read or write operation.

Managing Traps In Software Release 3.0.5, the command interface for managing SNMP traps has changed. These commands are affected:

„

„

„ set snmp trap show snmp clear snmp

Trap Filtering 25

The set snmp trap

command enables you to configure trap destinations, filters, and trap receive settings. set snmp trap destination

<community> filter link_state enable disable topology_change enable disable authentication enable disable receive enable disable

Set the trap destination. The system prompts for an IP address.

Enables or disables the generation of link state traps.

Enables or disables the generation of topology change traps.

Enables or disables the generation of authentication traps.

As a trap receiver, the EME receives traps from other SNMP devices, including other CoreBuilder 9000 modules, that have the EME IP address in their community table.

You can specify the network management stations that receive alerts from the EME. To do so, you must create a new community table entry for each network management station.

Note that SNMP traps are transmitted only through an in-band

IP routing interface.

26 C HAPTER 2: R ELEASE H IGHLIGHTS

The show snmp

command displays the status of the SNMP trap configuration and extensions. show snmp trap Displays the status of the trap receive setting and link state, topology change, and authentication trap filters. Displays an index of trap destinations extensions

The clear snmp

command clears SNMP trap destinations. clear snmp trap_destination all

<index>

Clears trap destinations. You can clear all trap destinations or use an index number from the table displayed by show snmp trap

.

3

R

ELEASE

I

SSUES

F

OR

M

ANAGEMENT

M

ODULES

This chapter addresses the following issues for the Switch 4007

Management Module:

■

Corrected Problems in Release 3.0.0

■

■

Operating Considerations for Release 3.0.5

Known Problems in Release 3.0.5

The numbers that appear in parentheses at the end of some descriptions are for internal 3Com use only.

The Switch 4007 software and management interfaces are built from

CoreBuilder ® 9000 switch technology. In Switch 4007 software, the interface prompts and displays indicate this heritage. The examples, tables, and figures in the Switch 4007 documentation may also indicate this heritage.

The Switch 4007 management interfaces refer to the Management

Module as the Enterprise Management Engine (EME). In these release notes, the Switch 4007 Management Module is referred to as the EME for brevity.

Although the heritage of the Switch 4007 is the CoreBuilder 9000, these

Switch 4007 release notes do not cover all aspects that apply to the

CoreBuilder 9000. These release notes include only information about the subset of modules that are officially supported under the Switch 4007 brand and about configurations that apply to the Switch 4007

(for example, issues that deal with dual switch fabrics are not included because the Switch 4007 chassis only has one fabric slot). For a list of modules that are officially supported in a Switch 4007 system, see

Table 1. If you have a CoreBuilder 9000 system or if you are operating unsupported modules in a Switch 4007 system, we recommend that you download the suite of CoreBuilder 9000 Release Notes for Software

Release 3.0.5 from the 3Com Web site.

28 C HAPTER 3: R ELEASE I SSUES F OR M ANAGEMENT M ODULES

Corrected Problems in Release 3.0.0

This section describes problems that were corrected in EME software

Release 3.0.0.

■

The condition that caused the EME to excessively ARP from its

10BASE-T port when you inserted it into a multinetted subnetwork has been corrected. (27009)

■

■

■

The EME no longer deletes the module’s BladeConfig.X.Y file from the file system when you use the default filename with the upload command. (23456)

If you have two switch fabric modules in the chassis and one is not functioning properly, you can use the set power slot <slot> mode disable

command to power off the switch fabric module that is not functioning correctly instead of physically removing the module from the chassis. (23135)

When you enter the show module <slot>.<subslot> verbose command and the supported files field displays none

, you no longer have to reset the module. The supported file types appear after the module comes up.

■

■

■

■

■

■

■

After you reset the chassis, and during the system start-up procedures, the system may display the following message: os socket err

. You can ignore this message. The system continues the start-up procedures as usual.

You no longer see an os socket err

message while the EME is processing multiple commands over a Telnet session and the time to process the command has been reduced.

You no longer see an os socket err message after you reset the chassis and during the system start-up procedures. (27650 and

29672)

The EME now recognizes all file types. In previous releases, if you were going to hot swap a module, or download software to a module, the

EME sometimes did not recognize certain file types.

The EME deletes a software image from its file system after successfully downloading software to a switching module from an external file server.

Transcend ® network management software now correctly interprets

SNMP trap information. (29657)

Bursty multicast or broadcast traffic no longer causes EME performance problems. (24014 and 27383)

Operating Considerations for Release 3.0.5

29

■

■

The EME no longer redirects broadcast frames to its default gateway.

(27142)

The EME password no longer reverts back to the default password after a reset of the switch fabric module and a switchover to the redundant EME. (29077)

Operating

Considerations for

Release 3.0.5

This section describes the operating considerations that exist for an EME that is operating with software Release 3.0.5.

Diagnostic events

■

Diagnostic events that are logged to the Enterprise Management

Engine (EME) show the slot number for the module at the time of failure. If you move the module to a different slot before these events get logged, the diagnostic events correspond to the previous slot number.

Chassis Reset

■

The Chassis Reset button functions only on the management module with the active EMC. The active EMC is indicated by the module with the illuminated Active LED.

Software Transfers

■

Do not attempt to perform any management tasks while you are downloading software.

■

While you are downloading or uploading a file, do not close the console window that you used to execute the command. If you use a script to execute the command, do not halt the script.

■

If you are planning to transfer several software load image files consecutively with the download

command, ensure that adequate space is available in the EME file system before you begin each of the file transfers.

To determine whether there is adequate file space, enter the show file command.

If the value in the Number of bytes available field is less than the size of the file to be transferred, do not begin the file transfer. Use the clear file

command to delete unneeded files until the value in the number of bytes available field is adequate. If you attempt the transfer without adequate space available, the transfer will fail.

30 C HAPTER 3: R ELEASE I SSUES F OR M ANAGEMENT M ODULES

Password Access

■

To gain access to the EME command interface, the system prompts you for a username and password. The default username is admin

(lowercase), but there is no default password. The first time that you log in, simply press Enter at the password prompt.

■

To define a password for your EME, use the set login password command. After you enter the password, the EME user interface prompts you to reenter it to verify it.

System and Module

Naming

■

When you create the system name through the Administration

Console command line interface, you can use spaces in the name and no quotation marks are required. However, when you create a module name that contains spaces, then you must use quotation marks around the module name. 3Com recommends that you not use spaces for the module name.

Web Management

■

You need to click some security windows in the Web Management interface twice for the browser to accept the information. (24806)

■

If you use Netscape as your browser, you must install the Help files on an external server, not on a local PC. For instructions on how to install and configure Help for the Netscape browser, see the

“Documentation Changes” chapter later in these release notes.

(29490)

Known Problems in

Release 3.0.5

This section describes the problems that are known to exist for an EME that is operating with software Release 3.0.5.

Diagnostics

■

The EME incorrectly displays the following message when you set the servdiag loop count to 0 with the set servdiag 2.1 loop_cnt 0 command:

Test may take up to 2 minutes and 0 seconds

Do you wish to continue (y/n):

The servdiag test runs forever, as it is designed to do when you set the loop count to 0 (not up to 2 minutes) .

(30015)

Known Problems in Release 3.0.5

31

SLIP Interfaces If you attempt to enable a Serial Line Internet Protocol (SLIP) interface before you have properly configured the SLIP interface and the terminal port, the EME returns

UT_ERR

. This is not a standard 3Com error message.

System Time

■

The EME may incorrectly display the following error message after you reset the EME with SNTP enabled:

SNTP client error

■

■

This is not a valid error message under these conditions; there is no problem with the SNTP client. (29808)

After you set the clock to

24:00:00

to indicate midnight, the EME sets the clock to

00:00:00

of the same day rather than the next day.

You can specify the next day in the date field with the EME set clock date_time

command. (29835)

The EME command line interface does not accept a user-defined GMT offset when you select the user-defined time zone setting ( set clock time_zone 28-Input_an_offset_from_GMT

command). You can set a user-defined time zone with a GMT offset through the WebConsole.

(29853)

Software Downloads

■

When downloading v3.0.0 code or higher to switching modules, you must use the all_image

command. Selecting any other option causes an incomplete download. See Chapter 8 in these release notes for software installation procedures. (30404)

■

If your Switch 4007 has a redundant EME configuration that has both of the EME Ethernet ports connected to a SuperStack II Switch 3900,

Switch 9300, CoreBuilder 9400, or Layer 2 module on another

Switch 4007 system, in the event of a failover the secondary EME cannot immediately send traffic. It must wait until the old primary

EME’s MAC address has aged out of the switch’s address table or clears the PC ARP cache. If you select either enabled

or agingOnly from the bridge spanningTree stpState

command on the switch to which the dual EMEs are connected, the new address will be learned within a few seconds. (30199)

32 C HAPTER 3: R ELEASE I SSUES F OR M ANAGEMENT M ODULES

4

R

ELEASE

I

SSUES

F

OR

S

WITCH

F

ABRIC

M

ODULES

This chapter addresses the following topics for Switch 4007 switch fabric modules that are operating with software Release 3.0.5:

■

Corrected Problems in Release 3.0.0

■

■

■

Corrected Problems in Release 3.0.5

Operating Considerations for Release 3.0.5

Known Problems for Release 3.0.5

The numbers that appear in parentheses at the end of some descriptions in this chapter are for internal 3Com use only.

The Switch 4007 management interfaces refer to the Management

Module as the Enterprise Management Engine (EME). In these release notes, the Switch 4007 Management Module is referred to as the EME for brevity.

The Switch 4007 software and management interfaces are built from

CoreBuilder 9000 switch technology. In Switch 4007 software, the interface prompts and displays indicate this heritage. Examples, tables, or figures in the Switch 4007 documentation may also indicate this heritage.

Although the heritage of the Switch 4007 is the CoreBuilder 9000, these

Switch 4007 release notes do not cover all aspects that apply to the

CoreBuilder 9000. These release notes include only information about the subset of modules that are officially supported under the Switch 4007 brand and about configurations that apply to the Switch 4007

(for example, issues that deal with dual switch fabrics are not included because the Switch 4007 chassis only has one fabric slot). For a list of modules that are officially supported in a Switch 4007 system, see Table 1 in Chapter 1 of these release notes. If you have a CoreBuilder 9000 system or if you are operating unsupported modules in a Switch 4007 system, we recommend that you download the suite of CoreBuilder 9000

Release Notes for Software Release 3.0.5 from the 3Com Web site.

34 C HAPTER 4: R ELEASE I SSUES F OR S WITCH F ABRIC M ODULES

Corrected Problems in Release 3.0.0

This section discusses problems that pertain to switch fabric modules and that were corrected in software Release 3.0.0.

System

■

The system no longer displays os socket err after you reset the chassis and during the system start-up procedures. (27650 and

29672)

■

The 24-port switch fabric modules will no longer cold reset when you disconnect the fiber cable on the 2-port 1000BASE-LX Gigabit

Ethernet (GEN) Interface Module (Model Number 3CB9LG2SC).

(27672 and 25814)

Ethernet

■

The 24-port Gigabit Ethernet (GEN) Switch Fabric Module

(Model Number 3CB9FG24) no longer experiences a delay due to a link state change. (23982 and 25748)

■

Whenever a Gigabit Ethernet Interface module (Model Numbers

3CB9LG2MC, 3CB9LG2SC, 3CB9LG4) is reset (with the module nvData reset

command, module insertion, warm reset, service diagnostics, and so forth) or a backplane port is disabled and then reenabled, the links to the switch fabric module are now reestablished, and you no longer need to reset the primary switch fabric module. (30063)

Downloading Release 3.0.1 code to the 9-port Gigabit Ethernet Switch

Fabric Module (Model Number 3CB9FG9) or 24-port Gigabit Ethernet

Switch Fabric Module, 12 trunks (Model Number 3CB9FG24T) resolves this problem (30063).

Bridging

■

You can now list the bridge port address table and walk through MIBs without causing Trunk Control Message Protocol (TCMP) enabled trunk ports to go down, or causing Spanning Tree Protocol (STP) topology changes. (23076 and 24547)

■

When a port enters the Spanning Tree forwarding state, the root bridge no longer erroneously sends a new root trap instead of a topology change trap. (28665)

VLAN

■

Removing the default virtual LAN (VLAN) before you create additional

VLANs no longer causes modules to go down after a module reset.

(26223)

Corrected Problems in Release 3.0.5

35

Corrected Problems in Release 3.0.5

This section discusses problems that pertain to switch fabric modules and that have been corrected in software Release 3.0.5.

Bridging

■

Spanning tree bridges that are attached to a Switch 4007 via Gigabit

Ethernet interface module ports no longer experience topology changes during a switch fabric module failover event. (29865)

■

If you want addresses to age more quickly after a link down event, you can now enable the agingOnly

option from the bridge spanningTree stpState

menu. This option does not enable the

Spanning Tree Protocol (STP); rather, it uses one of the STP values to temporarily age addresses more quickly after link down events.

Prior to Release 3.0.5, if you did not have STP enabled, a Layer 2 module would not flush MAC addresses after a link down event; it would only age addresses according to the value set in bridge agingTime

. This situation can cause VLAN-assignment problems when endstations are frequently plugged and unplugged into ports on the module and the Transcend Enterprise VLAN Manager (EVM) application is used on the network. That is, if a new endstation (with a new MAC address) is plugged into a port prior to the aging out of the old MAC address on that port, EVM may assign the endstation to the

VLAN to which the old MAC address belonged, which could be an incorrect VLAN for the new endstation. Now, with the agingOnly option, this problem is resolved. For more information about the agingOnly

option, see the “Release Highlights” chapter of these release notes. (28875, 30198, 29697)

Management

■

You can now configure or enable SNMP trap destinations from the

Administration Console with the snmp trap addModify command on switch fabric modules or interface modules. However, we strongly recommend that you enable trap destinations only via the EME interface. (29770)

■

The value of entPhysicalClass that is returned for the instance of the module itself (entPhysicalClasss.4) now correctly returns the value 9, which represents the value for the module.

36 C HAPTER 4: R ELEASE I SSUES F OR S WITCH F ABRIC M ODULES

Problems Fixed in v3.0.5 Address Table

The “instability” issue caused by TCMP messages not being processed properly in MPLA configurations, has been fixed. Stimulus such as

Address Table MIB walks, or Address Tracker application queries, on any of the v3.0.0 devices had caused a problem where other “high availability” features would be forced to adjust as a reaction to unprocessed TCMP messages. MPLA with TCMP configurations and allClosed VLANs had problems in this area.

The problem with SNMP GetNext commands not being processed properly has been partially fixed, such that MAC addresses now return upon command from the CLI and MIB walk queries. However, general response time to these queries is slow due to the limitations of the ASIC and various processes vying for shared resources.

Problem Overview

If SNMP tools like SNMX or GetMany (Jeff Case Tools), are used to execute address table MIB walks against v3.0.5 code, these requests will time out at times and not return all MAC addresses requested. The

Transcend Address Tracker application exacerbates this condition and causes traffic to slowdown, which is recoverable once Address Tracker query processes are shut down.

Manual requests (from the CLI or Web interfaces) and non-automated requests do not cause this problem. MAC addresses are returned smoothly from requests from the CLI or Web interface.

Problem Detail

SNMP applications that query a3ComSysBridgeVlanPortAddressRemoteAddress and the other “allClosed VLAN” address table MIBS may cause traffic slowdowns if issued asynchronously with multiple variable binds as described below:

■

■

Asynchronous requests — an SNMP request that is sent by an SNMP program before the previous request has been processed, and responded to, by the EME in the system.

Multiple Variable Binds — SNMP packets should not contain more than one Get request per SNMP packet. There are no issues with

Synchronous SNMP with one variable bind per SNMP packet.

However, the processing capability of some modules prohibits a timely

SNMP response to the gathering agent, in which case the SNMP

Corrected Problems in Release 3.0.5

37 gathering agent will time out and not report any MAC addresses have been returned by the EME.

Work Around

There is a work around for the issues of timing out while gathering MAC addresses. That is, you must extend the timeout interval of the SNMP gathering agent to allow for up to 15 seconds of delay when your network is using high availability features such as VRRP and MPLA.

Unfortunately, the applications Transcend Address Tracker and Enterprise

VLAN Manager cannot modify their timeout intervals sufficiently to accommodate the EME, which is the cause of an incomplete number of

MAC addresses returning to Address Tracker.

In some configurations involving MPLA and Closed VLANs, traffic may be slowed altogether while querying the switch for MAC and IP addresses, but the device and network do recover smoothly from these instances.

Stability issues have been fixed with respect to proper TCMP processing, however the return of MAC addresses upon query in these configurations is slow.

Recommendations

Follow the guidelines listed below for disabling background processes of

Transcend Address Tracker and Enterprise VLAN Manager if you do not want to have these applications querying your network devices on the default time interval.

Transcend applications within FR003 UNIX 5.0 update are supported with software release 3.0.5. Transcend NT 1.2 applications are also supported by software release 3.0.5.

Guidelines from FR003 TRANSCEND UNIX 5.0 UPDATE

The following known problem reference is from the Functional

Limitations section of the FR003S README file from JANUARY 2000

TRANSCEND UNIX 5.0 UPDATE RELEASE: General NMDww04250.

Some high-availability features from software release 3.0.0 including

MPLA, STP and VRRP, may become unstable when the associated packet switching devices are polled by Enterprise VLAN Manager, Address

Tracker, and other applications that poll the bridge address MIB tables, perform a MIB walk of the bridge address table, or use a command line interface to retrieve bridge address table data.

38 C HAPTER 4: R ELEASE I SSUES F OR S WITCH F ABRIC M ODULES

If applications involving address table polling are run on an active production network, the SNMP polling can contribute to unexpected events in the high-availability configuration, such as MPLA links down and STP topology changes. It is recommended these applications that involve address table polling be used during initial configuration or during problem isolation instead of during active, busy network periods. For more information about this issue, please visit the 3Com Knowledgebase on the 3Com Web site.

Applications such as Enterprise VLAN Manager and Address Tracker have background daemons that continue to poll the network even when the

GUI (or front-end) applications are not running. The devices that may be affected are the Switch 4007, CoreBuilder 9000, CoreBuilder 9400,

CoreBuilder 3500, SuperStack II Switch 9300, and SuperStack II Switch

3900 (all running 3.0.0 software).

If you are not using the Enterprise VLAN Manager and Address Tracker applications you should un-install these applications with the Transcend un-install utility. When applying the Transcend patch, the patch installation utility looks to see which applications are installed and patches those applications only.

1 Proceed to the /usr/3Com/install/bin directory

2 Run the

./deinst.scr

script and select the VLAN Manager and Address

Tracker applications for de-installation.

Be aware that even though the application GUI is not running, there are several background daemons that are running. These daemons handle

SNMP polling to the devices. Network problems (as described above) may occur if these daemons continue to poll the device address tables.

Stopping VLAN Manager and Address Tracker specific daemons prevents

SNMP polling. To stop these daemons from running, you need root access to the workstation.

Corrected Problems in Release 3.0.5

39

If you are using the Enterprise VLAN Manager and Address Tracker applications, do one of the following:

Under the /usr/3Com/diagnostics directory run: ./tncs_diagnostics.sh

1 Choose #1 to administer TNCS servers

2 Choose #1 to start/stop specific TNCS servers

3 Choose each of the application servers and follow the directions on the screen to finish the task

Important Notes

Transcend Central standalone installation will start AddressTracker and

Enterprise VLAN Manager daemons—be aware of this and shut them down if needed.

VLAN Manager daemons and HPOV NNM: The VLAN poller daemons medp

and pdp/java

are background daemons that are integrated with

NNM; therefore they can be started/stopped by the HPOV NNM commands – ovstart/ovstop. If you reboot your workstation for any reason or start NNM with the ovstart command, the Enterprise VLAN

Manager background daemons are invoked automatically. Please use the tncs_diagnostics.sh

script after you restart NNM so that you can stop the Enterprise VLAN Manager daemons once they are started.

Transcend daemons and restarting your workstation: Transcend daemons are started automatically if your workstation is rebooted for any reason. Please be aware of this if you plan to shutdown and restart your

UNIX workstation.

Be aware of the polling cycles you set for VLAN Manager, Address

Tracker, and other Transcend applications that do SNMP polling. Setting shorter polling intervals for these applications can increase the chance of instability in your network. By staggering the polling intervals for all of

Transcend applications that use polling and have customizable polling rates, you can help minimize SNMP traffic. Also, planning your polling cycles during off-peak hours can help alleviate extra traffic on your network as well.

40 C HAPTER 4: R ELEASE I SSUES F OR S WITCH F ABRIC M ODULES

If you do not wish to un-install these applications and plan to use them in the future, you can Start and Stop VLAN Manager and Address Tracker specific daemons by doing the following:

1 Proceed to /usr/3Com/diagnostics directory and run:

./tncs_diagnostics.sh

This requires that you have root access to the workstation. Stopping these daemons prevents any background SNMP polling on the network.

2 Verify the daemons have shutdown by using the ps command as follows: a For Enterprise VLAN Manager: ps –ef |grep NCD

should not list any

Enterprise VLAN Manager daemons. b For Address Tracker:

ps –ef |grep atservice

should not list atservice daemon.

Enterprise VLAN Manager background daemons are invoked automatically. Please use the tncs_diagnostics.sh

script after you restart NNM so that you can stop the VLAN Manager daemons once they are started.

Another way to prevent VLAN Manager from polling: The

/usr/3Com/ncd/runtime/sav directory contains all files comprising the

Enterprise VLAN Manager database. These files are

127.0.0.1.ppp

,

{NMS

IP}.6790.PDP

, and

{NMS IP}.ppp

and should be moved somewhere else or renamed with a

*.old

extension. Without a network database no

SNMP polling is executed even if the background daemons restart.

Clearing the Address Tracker database: Address Tracker polling is independent of the database ( atdatastore.dat

). If it does not find this file at application startup, it is created automatically. Therefore, clearing the atdatastore.dat

file does not stop Address Tracker from polling.

Operating Considerations for Release 3.0.5

41

Operating

Considerations for

Release 3.0.5

This section describes operating considerations that apply to switch fabric modules and interface modules (no on-board switching) operating with software Release 3.0.5.

Staging

■

You cannot stage a module in one chassis type and then move this module to another chassis type due to the variations in the backplane port mapping. When you move a switch fabric module from one chassis type to another chassis type, you must reset the module with the module nvData reset

command from the Administration

Console. Failure to reset the configurations on the switch fabric module in this situation can result in unpredictable behavior. Note however, that if you set the staging bit, this reset happens automatically.

System

■

When you create the system name, you can use spaces in the name and no quotation marks are required. However, when you create a module name that contains spaces, then you must use quotation marks around the module name. 3Com recommends that you not use spaces for the module name. (25677)

Management

■

You may experience slow performance (response time) when you perform a walk of the a3ComSystems MIB, particularly when you request the address tables:

■ a3ComBridgePortAddressBridgeIndex

■

■

■ a3ComBridgePortAddressPortIndex

Performing a walk of this MIB may also affect the performance of the

Administration Console interface and other management functions.

There is no SNMP support for creating or deleting trunks.

To prevent timeouts of SNMP requests, 3Com recommends that you increase the default timeout for the network management station.

Web-based

Management

■

If you use Netscape as your browser, you must install the Help files on an external server, not on a local PC. For instructions on how to install and configure Help for the Netscape browser, see the chapter

“Documentation Changes” later in these release notes. (29490)

42 C HAPTER 4: R ELEASE I SSUES F OR S WITCH F ABRIC M ODULES

Ethernet

■

Each gigabit link in the system sends congestion control frames every two seconds to detect and recover from traffic congestion. Flashing

LEDs on the front panel and the corresponding statistics counters indicate this traffic.

Transmit and Receive

Ports on GBIC

Modules

■

■

The positions of the Transmit and Receive ports on modules with GBIC ports are as follows:

Receive (RX) is the left port

■

Transmit (TX) is the right port

Modules that do not use GBICs have the Transmit and Receive ports in the opposite positions: Receive (RX) is the right port, and Transmit (TX) is the left port.

Bridging

■

The time that it takes for addresses to age out of address tables may be up to twice as long as the bridge aging time attribute indicates.

■

Bridge port addresses in the address list display are not sorted numerically. If you need to find a specific address, use the bridge port address find

command.

■

The bridge port address count, which is the count of the actual number of addresses in the address table, may not match the bridge port address list under certain conditions. The bridge port address list command does not list control addresses (MAC addresses for control frames that go directly to the processor).

■

Control addresses are installed in the address table under the following circumstances:

Spanning tree for the bridge is enabled.

■

TCMP is enabled on any of the defined trunk groups.

■

An IP VLAN interface is defined (there will be a control address in the address table for each port within the associated VLAN).

These control addresses are included in the bridge display address count field, but are not listed in the bridge port address list.

Trunking

■

Invalid trunking configurations may occur if you manually reset modules that are involved in backplane port trunking definitions, or make modifications before the configuration change takes effect.

Do not reset these modules for 5 minutes after you make configuration changes.

Operating Considerations for Release 3.0.5

43

■

■

3Com recommends that you enable the Trunk Control Message

Protocol (TCMP) on each end station of a trunk. If this protocol is not enabled on both end stations, TCMP cannot correctly validate trunk configuration rules. ( 29581)

When moving a staged module the trunking configuration is learned from the fabric when automapping is used. (27113)

Resilient Links

■

You can define resilient links only on switch fabric modules for ports that are connected to GEN interface modules. However, if a resilient link on a switch fabric module is connected to a GEN interface module, and you replace that module with a Fast Ethernet Layer 2 module, the switch fabric module incorrectly maintains the resilient link. The resilient link does not work properly in this case. (26308)

VLANs

■

The following bridge commands may take a few minutes to complete when several VLANs are configured in the system:

■

■ bridge vlan mode bridge vlan define

■ bridge vlan remove

To minimize the delay, configure the VLANs in order of increasing

VLAN ID (VID) and remove them in order of decreasing VID.

44 C HAPTER 4: R ELEASE I SSUES F OR S WITCH F ABRIC M ODULES

Known Problems for Release 3.0.5

This section describes problems that are known to exist for switch fabric modules and interface modules (no onboard switching) operating with software Release 3.0.5.

System

■

You must manually update module emergency download parameters the first time that a module is updated to the Release 3.0.5 operational code. After this, the parameters are updated automatically when the operational code is downloaded.

■

■

At the Administration Console, entering a seemingly valid command string (two or more commands taken from menus and nested submenus) may not work. If this situation occurs, enter commands separately from one menu prompt at a time.

The reset chassis command does not work if EME diagnostics are disabled. As a guideline, do not leave EME diagnostics disabled.

(30097)

Management

■

Do not install GEN interface modules while a GEN switch fabric module is resetting or if power to any of the switch fabric modules in a chassis is disabled.

■

■

The MIB object a3ComSysEthernetPortCount returns the value of the combined number of front panel ports, backplane ports, and the management port.

This MIB-II counter does not increment: snmplnASNParseErrs.

■

The following MIB objects have write access according to the corresponding RFC:

■

■

■

■

■ ifAdminStatus ipDefaultTTL etherStatsDataSource ipRouteAge

However, the 3Com implementation does not allow these parameters to be changed from their default values.

The Administration Console interface incorrectly displays the following unsupported trap:

■

MIBII: Authentication Failure

Known Problems for Release 3.0.5

45

■

■

Because the 64-bit counters in the ifXTable are read-only counters, you cannot reset them. Attempting to do so results in no response from the SNMP agent, and the values are not set.

Through SNMP the object a3ComSysEthernetPortRequestedPortMode can be set to values out of range for the port, only use the values that are for 10 or 100. (30022)

Ethernet

■

The following Ethernet statistics are not reset after the

module baseline set command:

■ rxPeakFrameRate

■

■

■ txPeakFrameRate rxPeakByteRate txPeakByteRate

■

■

■

■

TCMP Trunk Statistics

The ethernet summary and ethernet detail

displays incorrectly show a portType of 1000BaseSX(SC) for the 1000BASE-LX Gigabit

Ethernet (GEN) Interface module (Model Number 3CB9LG2SC).

(29689)

If you disable an Ethernet port on the switch fabric module with the ethernet portState disable command, and later re enable that port, the bridge state for the port remains down. The system does not forward network traffic on that port. You must reset the switch fabric module from the EME command line interface with the reset module command to re enable the bridge state and allow network traffic to be forwarded on that port.

If a bridging loop exists on a switching module or switch fabric module, output from the ethernet detail

command can take up to

30 minutes to display. (30080)

Trunking

■

These TCMP trunk statistics in the bridge trunk detail display are not reset after you enter the module baseline set

command: rxFrames, rxHellos, txFrames, txHellos, rxBadVersion, rxBadType, rxSameTrunkld, rxOverflow

■

When you add ports to an existing trunk, and change the port mode in the same modify operation, you need to reset the module and modify the trunk twice to add the additional ports to the trunk. The port mode is updated after the first reset operation. You can add a

46 C HAPTER 4: R ELEASE I SSUES F OR S WITCH F ABRIC M ODULES

■ port to a trunk without changing the port mode in a single modify operation. (29804)

When you insert a switching module that supports trunks on the backplane into an unpopulated slot that has AutoMap enabled, only the first link of the trunk for that slot is in the Up state, and other links are in the Down state. To avoid this problem, either populate the slot before you insert the module, or enable all backplane links in that slot that are in the Down state. (30041)

VLANs

■

When defining multiple VLANs after a VLAN mode change, you should wait a short period (5 minutes) after you change the mode before defining additional VLANs. Moving too quickly causes the memory on the switch fabric module to become exhausted which further causes the module to reboot and loose the settings that have just been changed. (30586)

5

R

ELEASE

I

SSUES FOR

L

AYER

2

S

WITCHING

M

ODULES

This chapter addresses the following topics for Switch 4007 Layer 2

Switching Modules:

■

Corrected Problems in Release 3.0.0

■

■

■

Corrected Problems in Release 3.0.5

System Issues

Known Problems

The numbers that appear in parentheses at the end of some descriptions are for internal 3Com use only.

The Switch 4007 management interfaces refer to the Management

Module as the Enterprise Management Engine (EME). In these release notes, the Switch 4007 Management Module is referred to as the EME for brevity.

The Switch 4007 software and management interfaces are built from

CoreBuilder 9000 switch technology. In Switch 4007 software, the interface prompts and displays indicate this heritage. Examples, tables, or figures in the Switch 4007 documentation may also indicate this heritage.

Although the heritage of the Switch 4007 is the CoreBuilder 9000, these

Switch 4007 release notes do not cover all issues that apply to the

CoreBuilder 9000. These release notes only include information about the subset of modules that are officially supported under the Switch 4007 brand and about configurations that apply to the Switch 4007

(for example, issues that deal with dual switch fabrics are not included because the Switch 4007 chassis only has one fabric slot). For a list of modules that are officially supported in a Switch 4007 system, see Table 1 in Chapter 1 of this document. If you have a CoreBuilder 9000 system or if you are operating unsupported modules in a Switch 4007 system, we recommend that you download the suite of CoreBuilder 9000 Release

Notes for Software Release 3.0.5 from the 3Com Web site.

48 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

Corrected Problems in Release 3.0.0

This section describes problems that relate to Layer 2 Switching Modules and that were corrected in software Release 3.0.0.

System

■

The system no longer displays os socket err

after you reset the chassis and during the system start-up procedures. (27650 and

29672)

■

The 36-port 10/100BASE-TX Layer 2 modules (Model Numbers

3CB9LF36R and 3CB9LF36T) no longer experience port lockup due to high port errors. (28668)

Management

■

Uploading a configuration file that already exists on the EME to a TFTP server no longer removes the file from the EME file system, as shown in the show file

display. (23456)

Ethernet

■

The 36-port 10/100BASE-TX Layer 2 Fast Ethernet RJ-45 module

(Model Number 3CB9LF36R) no longer asserts because of port duplex mode mismatch and TX underrun errors. (27365, 27362, and 29195)

■

■

Current autonegotiation status is now correctly reflected the first time that you view the Ethernet summary and detail displays. (27864)

Whenever a non-switch fabric module is reset (with the module nvData reset command, by inserting the module into a chassis, by doing a warm reset, by running service diagnostics, and so forth) or a backplane port is disabled and then reenabled, the links to the switch fabric module no longer go down on some occasions and are now always reestablished. (30063)

Bridging

■

You can now list the bridge port address table and walk through MIBs without causing Trunk Control Message Protocol (TCMP)-enabled trunk ports to go down, or causing Spanning Tree Protocol (STP) topology changes. (23076 and 24547)

■

■

When a port enters the Spanning Tree forwarding state, the root bridge no longer erroneously sends a new root trap instead of a topology change trap. (28665)

Enabling the Spanning Tree Protocol (STP) on a port that already had

STP enabled no longer results in an erroneous Spanning Tree topology change trap. (28655)

Corrected Problems in Release 3.0.5

49

■

■

The 36-port 10/100BASE-TX Layer 2 modules (Model Numbers

3CB9LF36R and 3CB9LF36T) now correctly populate the address table for Transcend ® Address Tracker. (26575)

Transcend Address Tracker polling no longer causes STP convergence.

(26879)

VLANs

■

Removing the default VLAN before you create additional VLANs no longer causes modules to go down after a module reset. Previously, you had to move the module to another slot in the chassis to reinitialize. (26223)

■

Previously, in some configurations, the system did not allow the proper number of VLANs to be defined, causing an out-of-resources message. The maximum number of VLANS are now allowed on regular switching modules, but not on the fabric. (27137)

Corrected Problems in Release 3.0.5

This section describes problems that have been corrected for Layer 2

Switching Modules operating with software Release 3.0.5.

Bridging

■

If you want addresses to age more quickly after a link down event, you can now enable the agingOnly

option from the bridge spanningTree stpState

menu. This option does not enable the

Spanning Tree Protocol (STP); rather, it uses one of the STP values to temporarily age addresses more quickly after link down events.

Prior to Release 3.0.5, if you did not have STP enabled, a Layer 2 module would not flush MAC addresses after a link down event; it would only age addresses according to the value set in bridge agingTime

. This situation can cause VLAN-assignment problems when endstations are frequently plugged and unplugged into ports on the module and the Transcend Enterprise VLAN Manager (EVM) application is used on the network. That is, if a new endstation (with a new MAC address) is plugged into a port prior to the aging out of the old MAC address on that port, EVM may assign the endstation to the

VLAN to which the old MAC address belonged, which could be an incorrect VLAN for the new endstation. Now, with the agingOnly option, this problem is resolved. For more information about the agingOnly

option, see the “Release Highlights” chapter of these release notes. (28875, 30198, 29697)

50 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

Management

■

The following request now returns correct and complete values for the

9-port 1000BASE-SX Gigabit Ethernet Layer 2 Switching Module:

■

■

The value of entPhysicalClass that is returned for the instance of the module itself (entPhysicalClass.4) now returns 9, which represents the value for the module.

The module agent now returns the ipNetToMediaPhysAddress MIB-II to the ipNetToMediaTable.

■

■

The object cb9000BackplaneConnectIfIndex now lists backplane port numbers in increasing order for the 9-port 1000BASE-SX Gigabit

Ethernet Layer 2 Switching Module.

■

■

The object a3ComSysBridgePortAddressStaticPort now returns one of the following values:

The local bridge port index when a static address is configured for the port

The value 0 when a dynamic address is configured for the port

■

■

■

■

■

■

■

■

■

If there is a trunk on the switching module, an SNMP getnext request on the MIB object a3ComTrunkTcmpMacState now correctly returns the correct value. For example, a Gigabit Ethernet switching module that has two trunks:

Ports 1 and 9 — MIB objects a3ComTrunkTcmpMacState.14.1 and

3ComTrunkTcmpMacState.14.9

Ports 4 and 5 — MIB objects a3ComTrunkTcmpMacState.15.4 and

3ComTrunkTcmpMacState.15.5

A getnext request on a3ComTrunkTcmpMacState.14.9 returns a3ComTrunkTcmpMacState.15.4

A getnext request on a3ComTrunkTcmpMacState.15.4 returns the next MIB object

The objects in the a3ComTrunkMacTable are now accessible via SNMP.

SNMP now returns an error if the object a3ComSysEthernetPortRequestedPortMode is set to values out of range for the port. (30022)

An SNMP getnext ifIndex request with a very large instance ID now returns the correct value ifDescr.1, instead of the incorrect value ifIndex.1.

An SNMP get request on entLPPhysicalIndex is now implemented.

Corrected Problems in Release 3.0.5

51

■

■

■

The following requests now return correct and complete values for the

9-port 1000BASE-SX Gigabit Ethernet Layer 2 Switching Module:

The value of entPhysicalVendorType that is returned for the instance of the module itself (entPhysicalVendorType.4) now contains the correct description.

The value of entPhysicalDescr that is returned for the module now contains slot, subslot, and revision numbers as it should.

NVRAM

■

Port Monitor Mode and Resilient Link Port configuration is properly initialized when upgrading a 36-port Layer 2 Switching Module from software release 2.x to release 3.0 or 3.0.5. (30432)

Problems Fixed in v3.0.5 Address Table

The “instability” issue caused by TCMP messages not being processed properly in MPLA configurations, has been fixed. Stimulus such as

Address Table MIB walks, or Address Tracker application queries, on any of the v3.0.0 devices had caused a problem where other “high availability” features would be forced to adjust as a reaction to unprocessed TCMP messages. MPLA with TCMP configurations and allClosed VLANs had problems in this area.

The problem with SNMP GetNext commands not being processed properly has been partially fixed, such that MAC addresses now return upon command from the CLI and MIB walk queries. However, general response time to these queries is slow due to the limitations of the ASIC and various processes vying for shared resources.

Problem Overview

If SNMP tools like SNMX or GetMany (Jeff Case Tools), are used to execute address table MIB walks against v3.0.5 code, these requests will time out at times and not return all MAC addresses requested. The

Transcend Address Tracker application exacerbates this condition and causes traffic to slowdown, which is recoverable once Address Tracker query processes are shut down.

Manual requests (from the CLI or Web interfaces) and non-automated requests do not cause this problem. MAC addresses are returned smoothly from requests from the CLI or Web interface.

52 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

Problem Detail

SNMP applications that query a3ComSysBridgeVlanPortAddressRemoteAddress and the other “allClosed VLAN” address table MIBS may cause traffic slowdowns if issued asynchronously with multiple variable binds as described below:

■

■

Asynchronous requests — an SNMP request that is sent by an SNMP program before the previous request has been processed, and responded to, by the EME in the system.

Multiple Variable Binds — SNMP packets should not contain more than one Get request per SNMP packet. There are no issues with

Synchronous SNMP with one variable bind per SNMP packet.

However, the processing capability of some modules prohibits a timely

SNMP response to the gathering agent, in which case the SNMP gathering agent will time out and not report any MAC addresses have been returned by the EME.

Work Around

There is a work around for the issues of timing out while gathering MAC addresses. That is, you must extend the timeout interval of the SNMP gathering agent to allow for up to 15 seconds of delay when your network is using high availability features such as VRRP and MPLA.

Unfortunately, the applications Transcend Address Tracker and Enterprise

VLAN Manager cannot modify their timeout intervals sufficiently to accommodate the EME, which is the cause of an incomplete number of

MAC addresses returning to Address Tracker.

In some configurations involving MPLA and Closed VLANs, traffic may be slowed altogether while querying the switch for MAC and IP addresses, but the device and network do recover smoothly from these instances.

Stability issues have been fixed with respect to proper TCMP processing, however the return of MAC addresses upon query in these configurations is slow.

Corrected Problems in Release 3.0.5

53

Recommendations

Follow the guidelines listed below for disabling background processes of

Transcend Address Tracker and Enterprise VLAN Manager if you do not want to have these applications querying your network devices on the default time interval.

Transcend applications within FR003 UNIX 5.0 update are supported with software release 3.0.5. Transcend NT 1.2 applications are also supported by software release 3.0.5.

Guidelines from FR003 TRANSCEND UNIX 5.0 UPDATE

The following known problem reference is from the Functional

Limitations section of the FR003S README file from JANUARY 2000

TRANSCEND UNIX 5.0 UPDATE RELEASE: General NMDww04250.

Some high-availability features from software release 3.0.0 including

MPLA, STP and VRRP, may become unstable when the associated packet switching devices are polled by Enterprise VLAN Manager, Address

Tracker, and other applications that poll the bridge address MIB tables, perform a MIB walk of the bridge address table, or use a command line interface to retrieve bridge address table data.

If applications involving address table polling are run on an active production network, the SNMP polling can contribute to unexpected events in the high-availability configuration, such as MPLA links down and STP topology changes. It is recommended these applications that involve address table polling be used during initial configuration or during problem isolation instead of during active, busy network periods. For more information about this issue, please visit the 3Com Knowledgebase on the 3Com Web site.

Applications such as Enterprise VLAN Manager and Address Tracker have background daemons that continue to poll the network even when the

GUI (or front-end) applications are not running. The devices that may be affected are the Switch 4007, CoreBuilder 9000, CoreBuilder 9400,

CoreBuilder 3500, SuperStack II Switch 9300, and SuperStack II Switch

3900 (all running 3.0.0 software).

54 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

If you are not using the Enterprise VLAN Manager and Address Tracker applications you should un-install these applications with the Transcend un-install utility. When applying the Transcend patch, the patch installation utility looks to see which applications are installed and patches those applications only.

1 Proceed to the /usr/3Com/install/bin directory

2 Run the

./deinst.scr

script and select the VLAN Manager and Address

Tracker applications for de-installation.

Be aware that even though the application GUI is not running, there are several background daemons that are running. These daemons handle

SNMP polling to the devices. Network problems (as described above) may occur if these daemons continue to poll the device address tables.

Stopping VLAN Manager and Address Tracker specific daemons prevents

SNMP polling. To stop these daemons from running, you need root access to the workstation.

If you are using the Enterprise VLAN Manager and Address Tracker applications, do one of the following:

Under the /usr/3Com/diagnostics directory run: ./tncs_diagnostics.sh

1 Choose #1 to administer TNCS servers

2 Choose #1 to start/stop specific TNCS servers

3 Choose each of the application servers and follow the directions on the screen to finish the task

Important Notes

Transcend Central standalone installation will start AddressTracker and

Enterprise VLAN Manager daemons—be aware of this and shut them down if needed.

VLAN Manager daemons and HPOV NNM: The VLAN poller daemons medp

and pdp/java

are background daemons that are integrated with

NNM; therefore they can be started/stopped by the HPOV NNM commands – ovstart/ovstop. If you reboot your workstation for any reason or start NNM with the ovstart command, the Enterprise VLAN

Manager background daemons are invoked automatically. Please use the tncs_diagnostics.sh

script after you restart NNM so that you can stop the Enterprise VLAN Manager daemons once they are started.

Corrected Problems in Release 3.0.5

55

Transcend daemons and restarting your workstation: Transcend daemons are started automatically if your workstation is rebooted for any reason. Please be aware of this if you plan to shutdown and restart your

UNIX workstation.

Be aware of the polling cycles you set for VLAN Manager, Address

Tracker, and other Transcend applications that do SNMP polling. Setting shorter polling intervals for these applications can increase the chance of instability in your network. By staggering the polling intervals for all of

Transcend applications that use polling and have customizable polling rates, you can help minimize SNMP traffic. Also, planning your polling cycles during off-peak hours can help alleviate extra traffic on your network as well.

If you do not wish to un-install these applications and plan to use them in the future, you can Start and Stop VLAN Manager and Address Tracker specific daemons by doing the following:

1 Proceed to /usr/3Com/diagnostics directory and run:

./tncs_diagnostics.sh

This requires that you have root access to the workstation. Stopping these daemons prevents any background SNMP polling on the network.

2 Verify the daemons have shutdown by using the ps command as follows: a For Enterprise VLAN Manager: ps –ef |grep NCD

should not list any

Enterprise VLAN Manager daemons. b For Address Tracker:

ps –ef |grep atservice

should not list atservice daemon.

Enterprise VLAN Manager background daemons are invoked automatically. Please use the tncs_diagnostics.sh

script after you restart NNM so that you can stop the VLAN Manager daemons once they are started.

56 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

Another way to prevent VLAN Manager from polling: The

/usr/3Com/ncd/runtime/sav directory contains all files comprising the

Enterprise VLAN Manager database. These files are

127.0.0.1.ppp

,

{NMS

IP}.6790.PDP

, and

{NMS IP}.ppp

and should be moved somewhere else or renamed with a

*.old

extension. Without a network database no

SNMP polling is executed even if the background daemons restart.

Clearing the Address Tracker database: Address Tracker polling is independent of the database ( atdatastore.dat

). If it does not find this file at application startup, it is created automatically. Therefore, clearing the atdatastore.dat

file does not stop Address Tracker from polling.

Operating Considerations for Release 3.0.5

57

Operating

Considerations for

Release 3.0.5

This section describes operating considerations that apply to Layer 2

Switching Modules that operate with software Release 3.0.5.

System

■

After you enter an nvdata reset

command (or initiate a trunk group definition or deletion), there is a delay before the module initiates the reset and returns to the Administration Console.

■

When you first insert and initialize a Gigabit Ethernet Layer 2

Switching Module, the switch fabric module generates the trap sequence:

Link Up... Link Down... Link Up

. This is normal.

Web-based

Management

■

If you use Netscape as your browser, you must install the Help files on an external server, not on a local PC. For instructions on how to install and configure Help for the Netscape browser, see “Documentation

Changes” later in these release notes. (29490)

SNMP

■

You may experience slow performance (response time) when you perform a walk of the a3Com.29 MIB, particularly when you request the following fields:

■ a3ComBridgePortAddressBridgeIndex

■ a3ComBridgePortAddressPortIndex

Performing a walk of this MIB may also affect the performance of the

Administration Console interface and other management functions.

■

■

■

■

The MIB object a3ComSysEthernetPortCount returns a value that is the combined number of front panel ports, the backplane ports, and the management ports.

There is no SNMP support for creating or deleting trunks.

To prevent timeouts of SNMP requests, 3Com recommends that you increase the default timeout for the network management station.

The value returned by ifType for backplane interfaces (117) are specified as defined by the Internet Assigned Numbers Authority

(IANA) and not as specified in RFC 1213 or RFC 1573.

Ethernet

■

Although you can change the PaceAccess setting to enable

, the modules do not currently support PACE and the setting actually remains disabled.

58 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

■

■

The rxFrameRate and rxBytes statistics in the ethernet detail display are incorrect. These statistics incorrectly add an extra 8 bytes to every packet that a 100 Mbps port receives. For example, when a 64 byte frame is sent into a 100 Mbit port, the value

72 is displayed in the ethernet detail

rxBytes field. Also, when 148,880 frames per second are sent into a 100 Mbit port, the value

151,143

frames per second is displayed in the ethernet detail

rxFrameRate field.

For non-connected front panel ports, the Autonegotiation state displays failed

. This is a display issue only. This condition applies to the 9-port 1000BASE-SX Gigabit Ethernet Layer 2 Switching Module

(Model Number 3CB9LG9MC). (26528)

Bridging

■

The following bridge commands may take a few minutes to complete when several VLANs are configured in the system in allClosed mode:

■

■ bridge vlan mode bridge vlan define

■

■

■

■

■ bridge vlan remove

■ bridge trunk define

To minimize this delay, configure VLANs in order of increasing VID and remove them in order of decreasing VID.

The time that it takes for addresses to age out of address tables may be up to twice as long as the bridge aging time attribute indicates.

Bridge port addresses in the address list display are not sorted numerically. If you need to find a specific MAC address, use the bridge port address find

command.

The txFrames statistic in the bridge port summary

display increments on ports with a down link state. (27962)

■

■

■

The bridge port address count, which is the count of the actual number of addresses in the address table, may not match the bridge port address list under certain conditions. The bridge port address list command does not list control addresses (MAC addresses for control frames that go directly to the processor). Control addresses are installed in the address table under the following circumstances:

Spanning tree for the bridge is enabled.

TCMP is enabled on any of the defined trunk groups.

An IP VLAN interface is defined (there is a control address in the address table for each port within the associated VLAN).

Operating Considerations for Release 3.0.5

59

These control addresses are included in the bridge display address count field, but are not listed in the bridge port address list.

Trunking

■

When moving a staged module, the trunking configuration is learned from the fabric when automapping is used. (27113)

■

Make trunking definitions and modifications only from the

Administration Console command line interface, and not through

SNMP.

■

■

When you add ports to an existing trunk, and change the port mode in the same modify operation, you need to reset the module and modify the trunk twice to add the additional ports to the trunk. The port mode is updated after the first reset operation. You can add a port to a trunk without changing the port mode in a single modify operation. (29804)

3Com recommends that you enable the Trunk Control Message

Protocol (TCMP) on each end station of a trunk. If this protocol is not enabled on both end stations, TCMP cannot correctly validate trunk configuration rules. (29581)

VLANs

■

The maximum number of virtual LANs (VLANs) that you can define in a Switch 4007 system is 127. You may not be able to add every bridge port in the system to every VLAN that you have defined.

■

When the vlan remove all

command is executing, it may take several minutes to remove all VLANs when the system is operating in allClosed mode.

60 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

Known Problems in

Release 3.0.5

This section describes problems that are known to exist in Layer 2

Switching Modules operating with software Release 3.0.5.

System

■

At the Administration Console, entering a seemingly valid command string (two or more commands taken from menus and nested submenus) may not work. If this situation occurs, enter commands separately from one menu prompt at a time.

■

Some Release 3.0.x features may become unstable when the associated switch devices are polled by Enterprise VLAN Manager,

Address Tracker, and other Transcend applications that poll the bridge address MIB tables. These features and protocols include Trunk

Control Message Protocol (TCMP) and Spanning Tree Protocol (STP) topology changes.

3Com recommends that you use these Transcend management applications only for configuration and problem isolation purposes.

When you use these applications, SNMP polling can contribute to unexpected events in the configuration associated with TCMP and

STP. (29371 and 29797)

For information about controlling the Transcend Enterprise VLAN

Manager and Address Tracker management applications, see the

“General Functional Limitations” section, problem NMDww04250, in the

Transcend Network Control Services Readme files for the February 2000

Update.

Management

■

When you are prompted to “wait” for a reset to take place, do not attempt to start a new connection during this delay. Wait until control returns you to the Administration Console prompt before you start a new connection.

■

■

■

After you enter an nvdata reset command (or initiate a trunk group definition or deletion), there is a delay before the module resets.

The following MIB-II counters do not increment: inReceives, tcpOutRsts, udpOutDatagrams, udpNoPorts, snmplnASNParseErrs, snmplnBadVersion, snmplnBadCommunityNames, snmplnBadCommunityUses

Known Problems in Release 3.0.5

61

■

■

The following MIB objects have write access according to the RFC’s.

However, 3Com implementation does not let you change these values to anything other than the default value:

■ ifAdminStatus, ipDefaultTTL, etherStatsDataSource, ipRouteAge

The Administration Console interface incorrectly displays the following unsupported trap:

■

■

■

MIB-II: Authentication Failure

Because the 64-bit counters in the ifXTable are read-only counters, you cannot reset them. Attempting to do so results in no response from the SNMP agent, and the values are not set.

If you reset the 9-port 1000BASE-SX Gigabit Ethernet Layer 2

Switching Module from the EME, you see MIB II Cold Start and Link

Up traps from the module, but you do not see a Link Down trap. The normal

Link Up... Link Down... Link Up

sequence has probably occurred and the module has reset, even though you did not see a

Link Down trap.

Ethernet

■

The following Ethernet statistics are not reset after you enter the module baseline set

command:

■ rxPeakFrameRate, txPeakFrameRate, rxPeakByteRate, txPeakByteRate

■

If a bridging loop exists on a switching module or switch fabric module, output from the

ethernet detail command can take up to

30 minutes to display. (30080)

Bridging

■

The bridge incorrectly forwards frames that are sent to their own port

MAC address.

■

Once the bridge address table times out, if the MAC address was learned by a loopback port, ping packets are dropped and the bridge port stats same segment discard

counter increases. If the MAC address was learned by the sourceport, ping packets go through successfully.You can tell if you have a bad cable by monitoring the same segment discard

counter. (30250)

62 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

Trunking

■

Do not use special characters in trunk names. (25715)

■

The Administration Console of the 36-port 10/100BASE-TX Fast

Ethernet RJ-45 Layer 2 Switching Module (Model Number 3CB9LF36R) incorrectly allows you to add a roving analysis port when that port has already been defined as part of a trunk group. When you do this, the system removes the analyzer port from all VLANs.

■

■

To recover from this situation, you must remove the roving analysis port with the analyzer remove

command and modify the affected

VLANs to add the trunk port back to the VLAN with the bridge vlan modify

command. (29802)

These TCMP trunk statistics in the bridge trunk detail display are not reset after you enter the module baseline set

command:

■ rxFrames, rxHellos, txFrames, txHellos, rxBadVersion, rxBadType, rxSameTrunkId, rxOverflow

If a trunk is configured on a module, the following fields in the ethernet detail

display may show rates that are higher than the actual rates:

■ rxFrameRate, rxByteRate, rxPeakFrameRate, rxPeakByteRate, txFrameRate, txByteRate, txPeakFrameRate, txPeakByteRate

In addition, the port for which the report is requested does not have to be involved in the trunk for this situation to occur.

■

■

■

When you insert a switching module that supports trunks on the backplane into an unpopulated slot that has AutoMap enabled, only the first link of the trunk for that slot is in the Up state, and other links are in the Down state.

There are two ways to avoid this problem:

Do not enable AutoMap on an unpopulated slot. Enable AutoMap after you insert the switching module.

Enable all backplane links in that slot that are in the Down state.

(30041)

Known Problems in Release 3.0.5

63

■

Each module is allowed a certain number of member ports in a trunk

(L2=6 and L3=8). When you modify a trunk so that the sum of existing ports plus the added ports is greater than the number of ports allowed, an error is generated which states that the internal resource threshold has been exceeded and the ports have not been added to the trunk.

To workaround this, do not modify a trunk so that the sum of the existing trunk member ports and the added ports are greater than 6 on Layer 2 modules and 8 on Multilayer modules. Instead, perform a modification to remove the old ports, and then another modification to add the new ports. This involves rebooting the module twice.

(30638)

64 C HAPTER 5: R ELEASE I SSUES FOR L AYER 2 S WITCHING M ODULES

6

R

ELEASE

I

SSUES FOR

M

ULTILAYER

S

WITCHING

M

ODULES

This chapter addresses the following topics for Switch 4007 Multilayer

Switching Modules:

■

Corrected Problems in Release 3.0.0

■

■

■

Corrected Problems in Release 3.0.5

Operating Considerations in Release 3.0.5

Known Problems in Release 3.0.5

The numbers that appear in parentheses at the end of some descriptions in this chapter are for internal 3Com use only.

The Switch 4007 management interfaces refer to the Management

Module as the Enterprise Management Engine (EME). In these release notes, the Switch 4007 Management Module is referred to as the EME for brevity.

The Switch 4007 software and management interfaces are built from

CoreBuilder 9000 switch technology. In Switch 4007 software, the interface prompts and displays indicate this heritage. Examples, tables, or figures in the Switch 4007 documentation may also indicate this heritage.

Although the heritage of the Switch 4007 is the CoreBuilder 9000, these

Switch 4007 release notes do not cover all aspects that apply to the

CoreBuilder 9000. These release notes include only information about the subset of modules that are officially supported under the Switch 4007 brand and about configurations that apply to the Switch 4007

(for example, issues that deal with dual switch fabrics are not included because the Switch 4007 chassis only has one fabric slot). For a list of modules that are officially supported in a Switch 4007 system, see Table 1 in Chapter 1 of these release notes. If you have a CoreBuilder 9000 system or if you are operating unsupported modules in a Switch 4007 system, we recommend that you download the suite of CoreBuilder 9000

Release Notes for Software Release 3.0.5 from the 3Com Web site.

66 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

Corrected Problems in Release 3.0.0

This section describes the problems that were corrected for Multilayer

Switching Modules in software Releases 3.0.0 and 3.0.1.

System

■

Multilayer switching modules no longer return erroneous MBA diagnostic failure messages. (29211)

Bridging

■

Spanning Tree Bridge Protocol Data Units (BPDUs) are no longer dropped during high broadcast levels. (26230)

Virtual LANs (VLANs)

■

Previously, in some configurations, the system did not allow you to define the proper number of VLANs, causing out-of-resource messages. The system now allows correct VLAN definitions. (27137)

IP

■

Multilayer switching modules now advertise statically defined routes in

RIP updates. (27000)

■

■

User Datagram Protocol (UDP) Helper now works for ports 137 through 139 WINS packets. (27359 and 27929)

In rare instances, Dynamic Host Configuration Protocol (DHCP) clients were unable to get an IP address. This has been fixed. (25748)

QoS

■

The problems that caused Quality of Service (QoS) drop controls to fail have been fixed. (26950)

Roving Analysis

■

The system forwards IEEE 802.1Q tagged frames properly to the roving analysis monitor ports.

Web Console

■

The WebConsole now supports the new Administration Console password

menu item in IP RIP. (25663)

Ethernet

■

Whenever a non-switch fabric module is reset (with the module nvData reset

command, by inserting the module into a chassis, by doing a warm reset, by running service diagnostics, and so forth) or whenever a backplane port is disabled and then reenabled, the links to the switch fabric module no longer go down on some occasions and are now always reestablished. (30063)

Corrected Problems in Release 3.0.5

67

4-port Gigabit

Ethernet Layer 3

Switching Module

(GBIC)

■

At Release 3.0.1 on the 4-port Gigabit Ethernet Multilayer Switching

Module, the traffic from the same source MAC address that traverses different ports of a trunk is now learned in the address cache for that port. In earlier releases, the module sometimes dropped Layer 2 traffic that came from the same source MAC addresses. This problem had no effect on routed traffic on the module.

Corrected Problems in Release 3.0.5

This section describes problems that have been corrected in software

Release 3.0.5 for Multilayer Switching Modules.

QoS

■

If you modify a QoS control classifier to a control number of 1, the system no longer returns error number 37. (25804)

■

Previously, QoS NonFlow classifier and control settings stopped working after the module was reset. To work around this problem you needed to redefine NonFlow control to get it working again after a reset. You no longer need to do this. (28174 and 28842)

RMON

■

When RMON memory resources have been fully depleted, problems no longer occur when you begin monitoring with certain groups.

(30016)

RMON-2 Network

Management

Applications

■

■

You can now use the Discover Agents Wizard of Traffix Manager v3.0.

You no longer need to manually configure Release 3.0.5 agents. Also you no longer need to limit other RMON-2 management applications monitoring to two ports per Release 3.0.0 agent. (30016)

When you use a management application such as Traffix, you no longer need to disable monitoring on all ports before you select a new port to monitor. Leaving one port enabled during the disable/enable process no longer causes the RMON agent to crash. (30016)

IP Policies

■

The command line interface now accepts a single line definition of an import/export policy. You also may use valid minimum abbreviations as well as progressing through the menus. For complete information on valid command abbreviations, see the Switch 4007 Command

Reference Guide .(29565)

68 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

OSPF

■

The convergence problem resulting from a 12-port Fast Ethernet

Multilayer Switching Module not forming an adjacency with a

CoreBuilder 3500 switch has been fixed. (30323)

■

The 12-port Fast Ethernet Multilayer Switching Module no longer asserts when you collect the OSPF routerID via WebManager. (30508)

VRRP

■

The switch fabric module no longer gives continuous reports of VRRP newMasterTraps on all interfaces. (30438)

Problems Fixed in v3.0.5 Address Table

The “instability” issue caused by TCMP messages not being processed properly in MPLA configurations, has been fixed. Stimulus such as

Address Table MIB walks, or Address Tracker application queries, on any of the v3.0.0 devices had caused a problem where other “high availability” features would be forced to adjust as a reaction to unprocessed TCMP messages. MPLA with TCMP configurations and allClosed VLANs had problems in this area.

The problem with SNMP GetNext commands not being processed properly has been partially fixed, such that MAC addresses now return upon command from the CLI and MIB walk queries. However, general response time to these queries is slow due to the limitations of the ASIC and various processes vying for shared resources.

Problem Overview

If SNMP tools like SNMX or GetMany (Jeff Case Tools), are used to execute address table MIB walks against v3.0.5 code, these requests will time out at times and not return all MAC addresses requested. The

Transcend Address Tracker application exacerbates this condition and causes traffic to slowdown, which is recoverable once Address Tracker query processes are shut down.

Manual requests (from the CLI or Web interfaces) and non-automated requests do not cause this problem. MAC addresses are returned smoothly from requests from the CLI or Web interface.

Corrected Problems in Release 3.0.5

69

Problem Detail

SNMP applications that query a3ComSysBridgeVlanPortAddressRemoteAddress and the other “allClosed VLAN” address table MIBS may cause traffic slowdowns if issued asynchronously with multiple variable binds as described below:

■

Asynchronous requests — an SNMP request that is sent by an SNMP program before the previous request has been processed, and responded to, by the EME in the system.

■

Multiple Variable Binds — SNMP packets should not contain more than one Get request per SNMP packet. There are no issues with

Synchronous SNMP with one variable bind per SNMP packet.

However, the processing capability of some modules prohibits a timely

SNMP response to the gathering agent, in which case the SNMP gathering agent will time out and not report any MAC addresses have been returned by the EME.

Work Around

There is a work around for the issues of timing out while gathering MAC addresses. That is, you must extend the timeout interval of the SNMP gathering agent to allow for up to 15 seconds of delay when your network is using high availability features such as VRRP and MPLA.

Unfortunately, the applications Transcend Address Tracker and Enterprise

VLAN Manager cannot modify their timeout intervals sufficiently to accommodate the EME, which is the cause of an incomplete number of

MAC addresses returning to Address Tracker.

In some configurations involving MPLA and Closed VLANs, traffic may be slowed altogether while querying the switch for MAC and IP addresses, but the device and network do recover smoothly from these instances.

Stability issues have been fixed with respect to proper TCMP processing, however the return of MAC addresses upon query in these configurations is slow.

70 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

Recommendations

Follow the guidelines listed below for disabling background processes of

Transcend Address Tracker and Enterprise VLAN Manager if you do not want to have these applications querying your network devices on the default time interval.

Transcend applications within FR003 UNIX 5.0 update are supported with software release 3.0.5. Transcend NT 1.2 applications are also supported by software release 3.0.5.

Guidelines from FR003 TRANSCEND UNIX 5.0 UPDATE

The following known problem reference is from the Functional

Limitations section of the FR003S README file from JANUARY 2000

TRANSCEND UNIX 5.0 UPDATE RELEASE: General NMDww04250.

Some high-availability features from software release 3.0.0 including

MPLA, STP and VRRP, may become unstable when the associated packet switching devices are polled by Enterprise VLAN Manager, Address

Tracker, and other applications that poll the bridge address MIB tables, perform a MIB walk of the bridge address table, or use a command line interface to retrieve bridge address table data.

If applications involving address table polling are run on an active production network, the SNMP polling can contribute to unexpected events in the high-availability configuration, such as MPLA links down and STP topology changes. It is recommended these applications that involve address table polling be used during initial configuration or during problem isolation instead of during active, busy network periods. For more information about this issue, please visit the 3Com Knowledgebase on the 3Com Web site.

Applications such as Enterprise VLAN Manager and Address Tracker have background daemons that continue to poll the network even when the

GUI (or front-end) applications are not running. The devices that may be affected are the Switch 4007, CoreBuilder 9000, CoreBuilder 9400,

CoreBuilder 3500, SuperStack II Switch 9300, and SuperStack II Switch

3900 (all running 3.0.0 software).

Corrected Problems in Release 3.0.5

71

If you are not using the Enterprise VLAN Manager and Address Tracker applications you should un-install these applications with the Transcend un-install utility. When applying the Transcend patch, the patch installation utility looks to see which applications are installed and patches those applications only.

1 Proceed to the /usr/3Com/install/bin directory

2 Run the

./deinst.scr

script and select the VLAN Manager and Address

Tracker applications for de-installation.

Be aware that even though the application GUI is not running, there are several background daemons that are running. These daemons handle

SNMP polling to the devices. Network problems (as described above) may occur if these daemons continue to poll the device address tables.

Stopping VLAN Manager and Address Tracker specific daemons prevents

SNMP polling. To stop these daemons from running, you need root access to the workstation.

If you are using the Enterprise VLAN Manager and Address Tracker applications, do one of the following:

Under the /usr/3Com/diagnostics directory run: ./tncs_diagnostics.sh

1 Choose #1 to administer TNCS servers

2 Choose #1 to start/stop specific TNCS servers

3 Choose each of the application servers and follow the directions on the screen to finish the task

Important Notes

Transcend Central standalone installation will start AddressTracker and

Enterprise VLAN Manager daemons—be aware of this and shut them down if needed.

VLAN Manager daemons and HPOV NNM: The VLAN poller daemons medp

and pdp/java

are background daemons that are integrated with

NNM; therefore they can be started/stopped by the HPOV NNM commands – ovstart/ovstop. If you reboot your workstation for any reason or start NNM with the ovstart command, the Enterprise VLAN

Manager background daemons are invoked automatically. Please use the tncs_diagnostics.sh

script after you restart NNM so that you can stop the Enterprise VLAN Manager daemons once they are started.

72 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

Transcend daemons and restarting your workstation: Transcend daemons are started automatically if your workstation is rebooted for any reason. Please be aware of this if you plan to shutdown and restart your

UNIX workstation.

Be aware of the polling cycles you set for VLAN Manager, Address

Tracker, and other Transcend applications that do SNMP polling. Setting shorter polling intervals for these applications can increase the chance of instability in your network. By staggering the polling intervals for all of

Transcend applications that use polling and have customizable polling rates, you can help minimize SNMP traffic. Also, planning your polling cycles during off-peak hours can help alleviate extra traffic on your network as well.

If you do not wish to un-install these applications and plan to use them in the future, you can Start and Stop VLAN Manager and Address Tracker specific daemons by doing the following:

1 Proceed to /usr/3Com/diagnostics directory and run:

./tncs_diagnostics.sh

This requires that you have root access to the workstation. Stopping these daemons prevents any background SNMP polling on the network.

2 Verify the daemons have shutdown by using the ps command as follows: a For Enterprise VLAN Manager: ps –ef |grep NCD

should not list any

Enterprise VLAN Manager daemons. b For Address Tracker:

ps –ef |grep atservice

should not list atservice daemon.

Enterprise VLAN Manager background daemons are invoked automatically. Please use the tncs_diagnostics.sh

script after you restart NNM so that you can stop the VLAN Manager daemons once they are started.

Corrected Problems in Release 3.0.5

73

Another way to prevent VLAN Manager from polling: The

/usr/3Com/ncd/runtime/sav directory contains all files comprising the

Enterprise VLAN Manager database. These files are

127.0.0.1.ppp

,

{NMS

IP}.6790.PDP

, and

{NMS IP}.ppp

and should be moved somewhere else or renamed with a

*.old

extension. Without a network database no

SNMP polling is executed even if the background daemons restart.

Clearing the Address Tracker database: Address Tracker polling is independent of the database ( atdatastore.dat

). If it does not find this file at application startup, it is created automatically. Therefore, clearing the atdatastore.dat

file does not stop Address Tracker from polling.

74 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

Operating

Considerations in

Release 3.0.5

This section describes operating considerations that apply to Multilayer

Switching Modules with software Release 3.0.5.

System

■

Modules can be staged in one chassis type then moved to a different chassis type (For example, from a 16-slot CoreBuilder 9000 chassis to a 7-slot Switch 4007 chassis). However, a module’s backplane port trunking configuration (if there is any) may change; it depends on the differences in the backplane port mapping scheme of each chassis and the types of switch fabric modules that are installed.

■

Restoring the nonvolatile data configuration from one Switch 4007 to another can result in duplicate Open Shortest Path First (OSPF) router

IDs and cause OSPF problems. Be sure to change the router ID after you restore nonvolatile data to make the router ID unique in the network. You may need to define preferences after restoring nonvolatile data.

■

■

■

If trunks are defined and QoS controls are defined for these trunked ports, the nonvolatile data conversion from Release 2.0 or 2.1 to

Release 3.0.x does not properly reindex the trunk port numbers in the

QoS controls. You must redefine the controls. (29407)

An administrative weight of 16 for an IP policy is not properly converted. After the conversion, change the administrative weight back to 16. (28641)

Statistics for QoS, IPX, VLAN, TCMP, IP interface, and Ethernet detail for peak bytes and frame rates are not reset when you set a baseline using the module baseline set command. (06374)

SNMP

■

For the MIB object ipDefaultTTL object, write access is allowed but the

3Com implementation does not allow values other than the default value to be set.

■

MIB Object a3ComSysEthernetPortCount gives the value of the combined number of front, backplane, and management ports.

(29430)

■

■

There is no SNMP support for creating or deleting trunks.

To prevent timeouts of SNMP requests, 3Com recommends that you increase the default timeout for the network management station.

Operating Considerations in Release 3.0.5

75

■

■

The MIB a3ComSysBridgePortReceiveMulticastLimit allows setting an out-of-range value. (28995)

You cannot configure more than one routing interface on a multiprotocol VLAN using SNMP. Instead, configure the routing interface through the Administration Console.

RMON

■

You can use the snmp rmonConfiguration

command on the

Administration Console to change the RMON statistics default from collecting receive-only traffic to collecting transmitted and received traffic. Doing so affects the statistics, history, alarm, and event RMON groups. The rest of the RMON groups can only collect received traffic; however, because these groups collect conversation information between hosts, transmitted traffic also is included in the data for these tables.

Ethernet

■

Gigabit Ethernet ports do not autonegotiate with a SUN Ultra 2

Gigabit Ethernet card. If traffic is sent from the SUN during autonegotiation, then autonegotiation completes successfully.

(28448)

■

■

■

If the autonegotiation feature for a 100 Mbps Ethernet link does not function properly, disable autonegotiation and manually set the port speed and duplex mode ( half-duplex

or full-duplex

). Also, on a

1000 Mbps link, autonegotiation must be enabled on both sides of the link.

The noRxBuffers (frames that were discarded due to no available buffer space) counter may increment after an AppleTalk virtual LAN

(VLAN) modify and Address Resolution Protocol (ARP) update.

The rxFrameRate and rxBytes statistics found in the ethernet detail display are incorrect. These statistics incorrectly add an extra 8 bytes to every packet that receives a 100 Mbps port. For example, when a 64 byte frame is sent into a 100 Mbps port, the value

72

is displayed in the Ethernet Detail rxBytes field. Also, when 148,880 frames per second are sent into a 100 Mbit port, the value

151,143

frames per second is displayed in the Ethernet Detail rxFrameRate field.

Bridging

■

Do not set the Bridge aging time to 0 (zero). (21914)

■

Bridged IP multicast frames are not counted in the bridge statistics.

■

The bridge port address count, which is the count of the actual number of addresses in the address table, may not match the bridge

76 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

■

■ port address list under certain conditions. The

bridge port address list

command does not list control addresses. (Control addresses are

MAC addresses for control frames that go directly to the processor.)

Control addresses are installed in the address table under the following circumstances:

Spanning tree for the bridge is enabled.

TCMP is enabled on any of the defined trunk groups.

If the system is running in openVLAN mode, two control addresses are installed, one for spanning tree and one for TCMP.

If the system is running in closedVLAN mode, two control addresses are installed for each VLAN that is defined.

■

An IP VLAN interface is defined, a control address exists in the address table for each port within the associated VLAN.

These control addresses are included in the bridge display address count field, but are not listed in the bridge port address list. (30038)

Backplane Port

Trunking

■

■

Enabling or disabling the Ethernet state of the ports on a backplane link is now done exclusively from the switch fabric module. The switch fabric module enables or disables the ports at both ends of the backplane link including all ports in a trunk. You can no longer change the Ethernet state of a backplane port on an interface module.

The interface modules no longer list the backplane ports when you enter the ethernet portState command. The switch fabric module now completely controls the backplane port state. The

automap command and all backplane trunking commands are only available through the switch fabric modules. However, you can still change the ethernet portstate

of any front panel ports on an interface module.

If a slot in the chassis is autoenabled, the system creates a trunk for that slot (which happens when you reset the switch fabric module) when you use the autoMap command.

Trunking

■

One of the available trunk groups is now used to improve performance to the backplane of the 12-port 10/100BASE-TX Fast

Ethernet Multilayer Switching Module and the 10-port 100BASE-FX

Fast Ethernet Multilayer Switching Module. These modules now support three trunk groups instead of four trunk groups.

Operating Considerations in Release 3.0.5

77

■

■

If you have defined four trunk groups on one of these modules, remove one trunk group before you upgrade to Release 3.0.0.

3Com recommends that you enable the Trunk Control Message

Protocol (TCMP) on each end station of a trunk. If this protocol is not enabled on both end stations, TCMP cannot correctly validate trunk configuration rules. (29581)

When moving a staged module, the trunking configuration is learned from the fabric when automapping is used. (27113)

Packet Filtering

■

Custom packet filters cannot filter for MAC destination address (DA) or source address (SA) for Layer 3 traffic (not only IP but all routed traffic). (27596)

■

■

The system does not filter for MAC addresses if the frame is going to be routed and the filter contains a Destination Port Group Mask

(DPGM). (25028)

When you make multiple port filter assignments through SNMP, an

SNMP operation may return an incorrect value.

Virtual Router

Redundancy Protocol

(VRRP)

■

VRRP is not supported on overlapped, untagged, network-based IP

VLANs. (A “network-based” IP VLAN is a VLAN with a Layer 3 address that is used to determine VLAN membership.) If the overlapped ports are 802.1Q tagged, then VRRP is supported. (28616)

IP

■

On a VLAN, the IP interfaces statistics inReceives, inForwards, and outForward fields are counted one time for each interface. (29295)

■

The Internet Control Message Protocol (ICMP) is enabled by default. If end stations ignore ICMP redirects, the system may generate repeated

ICMP redirects, resulting in performance issues.

■

In allOpen mode, if ICMP redirects are enabled, packets routed through VLANs that contain multiple interfaces are processed in the FP

(slow path). When multiple IP interfaces are defined for a VLAN, set the ICMP redirect to disable

to restore performance (fast path).

IP Multicast

■

To optimize IP multicast performance, do not configure multiple tagged VLANs over backplane ports.

AppleTalk

■

When you remove an AppleTalk interface, you are not prompted with an

Are you sure

message. The change happens immediately. (26041)

78 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

QoS

■

If you add a second time zone control, the first time zone control displays the start date for the second time zone control. This feature confirms that the time is correct. (27272)

■

When you enter a QoS (Quality of Service) control detail command on the CoreBuilder 9000 12-port 10/100BASE-TX Fast

Ethernet Multiprotocol Layer 3 Switching Module and the 10-port

100BASE-FX Fast Ethernet Multiprotocol Layer 3 Switching Module, the rates calculated and displayed are based on power-of-two and not base-10. (That is, one gigabit is defined as 2 30 rather than 10 9 bits.)

■

QoS control aggregate behavior for flow classifiers only works on ports that are in certain groups. (29821)

See the chapter “Documentation Changes” later in these release notes, for additional information about QoS Control aggregate rate limit types.

■

To add a new QoS nonflow classifier, you must first delete one or more

(unrequired) nonflow classifiers in the range of 400 through 498.

■

■

■

■

■

■

You can no longer define an IP multicast rate limit from the IP menu.

Previously, you could use the ip multicast interface enable option and specify a rate limit. Now, use bridge port multicastLimit

or QoS commands to set this type of rate limit.

You cannot set the QoS time of day function from Web Management.

Use the Administration Console instead. (24790)

Changing the system time ahead into the time period that should activate a QoS control timer does not active the timer. The time period of the timer does get updated, but the control itself does not function.

(28231)

Reentering a QoS control from the Administration Console with the control name in quotation marks on one line reprompts for the name.

Reenter the name.

The QoS control 802.1p tag is not used in forwarded frames if a flow classifier is used. A non-flow classifier must be used for this option to take effect. (29896)

If a received IP packet is marked as fragmented and matches a flow classifier that is active under a QoS control, the packet is not classified.

Operating Considerations in Release 3.0.5

79

Roving Analysis

■

Remove a static address from a port before you use it as a roving analysis port. If the Bridge Table has a static address for a given port, and the port is then designated as the “Analyzer” port for RAP, the static address is still functional, which results in the forwarding of unwanted packets through the Analyzer port. (24499)

■

If the module drops packets, these dropped packets are not mirrored to an analyzer port.

Filter Builder

■

Although the system does not support address group filtering, Filter

Builder has controls to create and assign address groups. It also displays and allows the download of the rejdiffaddrgrp

filter that contains the two hardware address group filtering instructions.

(27583)

Transmit and Receive

Ports on GBIC

Modules

■

■

■

The positions of the Transmit and Receive ports on the 4-port Gigabit

Ethernet (GBIC) Multilayer Switching Module (Model Number

3CB9RG4) are as follows:

Receive (RX) is the left port

Transmit (TX) is the right port

Modules that do not use GBICs have the Transmit and Receive ports in the opposite positions: Receive (RX) is the right port, and Transmit (TX) is the left port.

Web-based

Management

■

If you use Netscape as your browser, you must install the Help files on an external server, not on a local PC. For instructions on how to install and configure Help for the Netscape browser, see the chapter

“Documentation Changes” later in these release notes.

VLANs

■

The maximum number of virtual LANs (VLANs) that you can define in a Switch 4007 system is 127. You may not be able to add every bridge port in the system to every VLAN that you have defined.

80 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

4-port GBIC

Multilayer Switching

Module (3CB9RG4)

The following system issue applies only to the 4-port GBIC Gigabit

Ethernet Multilayer Switching Module:

■

If you hot swap this module from one chassis type to a different chassis type and then insert it back into the original slot, the module configuration in NVRAM may reset back to factory default settings. To save the module’s original configuration, perform a backup of the module configuration before you hot swap the module.

Known Problems in

Release 3.0.5

This section describes problems that are known to exist in Multilayer

Switching Modules operating with software Release 3.0.5.

System

■

At the Administration Console, entering a seemingly valid command string (two or more commands taken from menus and nested submenus) may not work. If this situation occurs, enter commands separately from one menu prompt at a time.

SNMP

■

Some Release 3.0.x features may become unstable when the associated Switch devices are polled by Enterprise VLAN Manager,

Address Tracker, and other Transcend applications that poll the bridge address MIB tables. These features and protocols include Trunk

Control Message Protocol (TCMP), Virtual Router Redundancy

Protocol (VRRP), and Spanning Tree Protocol (STP) topology changes.

3Com recommends that you use these Transcend management applications only for configuration and problem isolation purposes.

When you use these applications, SNMP polling can contribute to unexpected events in the configuration associated with TCMP, VRRP, and STP. (29371 and 29797)

For information about controlling the Transcend Enterprise VLAN

Manager and Address Tracker management applications, see the

“General Functional Limitations” section, problem NMDww04250, in the

Transcend Network Control Services Readme files for the February 2000

Update.

■

A multilayer switching module that has a port with a physical link configured reports two types of incorrect link up traps when you initialize the module, depending on whether or not the Spanning Tree

Protocol (STP) is enabled. If STP is enabled, the module reports two link up traps when it should only report one. If STP is disabled, the module reports no link up traps when it should report one. (29864)

Known Problems in Release 3.0.5

81

■

■

■

■

■

There is no SNMP support for the AppleTalk Interface Count MIB

(a3ComSysAtInterfaceCount). (05994)

You cannot configure more than one routing interface on a multiprotocol VLAN using SNMP. Instead, configure the routing interface through the Web Management software or the

Administration Console.

You cannot delete and then successfully recreate an hlHostControlStatus entry using SNMP.

IPX Policies are not shown via SNMP. (28352)

The SNMP object a3ComSysEthernetPortRequestedPortMode allows you to set values that are out of range for the port. Only use the values 10 or 100 with this SNMP object. (30022)

Ethernet

■

The ethernet detail display may erroneously report a flow control value for a 10/100 port. A 10/100 port does not support flow control.

■

If a bridging loop exists on a switching module or switch fabric module, output from the ethernet detail

command can take up to

30 minutes to display. (30080)

Bridging

■

The STP pathcost does not reset to a lower value when you remove a trunk. Manually reset the pathcost counter to the desired value.

■

Associated bridge port detail

statistics do not increment when routed frames are filtered using custom or port group filters.

■

■

When you connect a 10/100 Ethernet port to a 10BASE-T Ethernet port, the system does not revise the default path cost to reflect the

10BASE-T Ethernet LAN segment. The 10BASE-T Ethernet port path cost defaults to 100 when it should default to 10.

The following bridge commands may take a few minutes to complete when several VLANs are configured in the system in allClosed mode:

■

■ bridge vlan mode bridge vlan define

■

■ bridge vlan remove bridge trunk define

To minimize this delay, configure VLANs in order of increasing VID and remove them in order of decreasing VID.

82 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

■

Do not enter the bridge port flush all command on any switching module when there are more than 1000 MAC addresses in its bridge address table.

Trunking

■

When you add ports to an existing trunk, and change the port mode in the same modify operation, you need to reset the module and modify the trunk twice to add the additional ports to the trunk. The port mode is updated after the first reset operation. You can add a port to a trunk without changing the port mode in a single modify operation. (29804)

■

■

Each module is allowed a certain number of member ports in a trunk

(L2=6 and L3=8). When you modify a trunk so that the sum of the existing ports plus the added ports is greater than the number of ports allowed, an error is generated which states that the internal resource threshold has been exceeded and the ports have not been added to the trunk. To workaround this, do not modify a trunk so that the sum of the existing trunk member ports and the added ports are greater than 6 on Layer 2 modules and greater than 8 on Multilayer modules.

Instead, perform a modification to remove the old ports, and then another modification to add the new ports. This involves the module rebooting twice. (30638)

On a CoreBuilder 9000 with a single switch fabric module installed, when you insert a switching module that supports trunks on the backplane into an unpopulated slot that has autoMap enabled, only the first link of the trunk for that slot is in the Up state, and other links are in the Down state.

■

■

There are two ways to avoid this problem:

Do not enable autoMap on an unpopulated slot. Enable autoMap after you insert the switching module.

Enable all backplane links in that slot that are in the Down state.

(30041)

Known Problems in Release 3.0.5

83

VLANs

■

VLAN statistics may have non-zero values when you first define a

VLAN.

■

When defining multiple VLANs after a VLAN mode change, you should wait a short period (5 minutes) after you change the mode before defining additional VLANs. Moving too quickly causes the memory on the switch fabric module to become exhausted which further causes the module to reboot and loose the settings that have just been changed. (30586)

Internet Protocol (IP)

■

You cannot create an IP interface using SNMP objects. To create an IP interface, use the Administration Console ip interface define command. (29971)

■

The inReceives statistic in the

ip interface statistics display is incorrect when overlapped IP VLANs are defined.

■

■

The IP interface statistic inDiscards

is always zero; it never increments.

Do not display an Open Shortest Path First (OSPF) routing table through a second console session while you are simultaneously flushing the table.

OSPF

■

A router advertises itself as an Area Border Router (ABR) after it becomes an internal router under the following scenario:

The router has two interfaces, each of which belongs to a different area (which makes this router an ABR). When you look at the linkstate database using

ip ospf linkStateData router

, the router is advertised as an ABR (correctly). If you move one of the interfaces to the area of the second interface and delete the unused area, the system still reports the router as an ABR, although it is actually an internal router now. To correct this problem, disable and then enable

Open Shortest Path First (OSPF) on all interfaces.

84 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

Virtual Router

Redundancy Protocol

(VRRP)

■

■

You can remove a Virtual Router (VR) through the Web Management interface without disabling VRRP. The system does not recognize the change for a few polling intervals. For most networks, the system recognizes the change in only a few seconds. Any possible network outage should last only a few seconds. (29940)

In rare instances VRRP routers transfer between backup and master, and then back to the Backup state. There is no adverse effect on the original Master VRRP router; and after the state changes of the

Backup, the original Master continues to function as expected. This has occurred in test environments only where heavy levels of traffic in a continuous stream are injected. (30410, 30665, 30438)

IP Multicast

■

IP Multicast traceRoute does not work if one router hop further downstream of the router with tunnels defined. (29564)

■

3Com recommends that you use VLAN-based IP interfaces with STP enabled rather than port-based IP interfaces in IP networks that use

DVMRP as the multicast routing protocol when these networks have physical loops in them. Because port-based IP interfaces ignore the

STP state, a port-based IP interface that appears to be blocking traffic is really forwarding traffic. Using DVMRP and STP ensures that the physical loops in the network are removed from the IP Multicast traffic flow.

IPX

■

The ipx server display

screen shows only up to 19 characters of an

IPX server’s name. Novell specifies that a server name can be up to 48 characters. (28021)

QoS

■

The QoS control with 802.1p tag on flow class is not used in forwarded frames. A nonflow classifier must be used for this option.

■

■

■

For cyclical QoS control timers, the output edit displays the start time for the next time period that the control becomes active. (27272)

If you classify traffic as IPX non-custom (for example, classifier 460) or as custom, then adding a drop control has no effect unless an underlying IPX-II VLAN is present. (27678)

In the qos statistics receive display, one-way-flow dropped packets appear in the dropped field, but they also appear erroneously in the NonFlowReserved field. (29297)

Known Problems in Release 3.0.5

85

■

■

■

After you perform a NVRAM conversion, the port numbering of the

QoS controls becomes incorrect if a control is applied to trunked ports.

(29407)

QoS rate limits do not work on packets that are routed in software.

(25036)

Multicast traffic will not be rate limited on a port that is part of overlapped VLANs. (30598

RSVP

■

RSVP supports unicast connections but not multicast connections for software Release 3.0.x.

■

Previous hop data is not recorded in a qos rsvp detail all display.

(27852)

■

■

■

■

The RSVP Integrity object is dropped and not transmitted to the next hop. Therefore, the receiving application rejects the path because the expected Integrity object is missing. (28980)

RSVP may not allow you to utilize the entire available bandwidth.

(06376)

RSVP wildcard filter reservations do not reach the sender. (06415)

Multilayer modules do not display RSVP reserved queue statistics for

RSVP sessions in the qos statistics transmit display when QoS and RSVP are enabled. (29979)

RMON

■

The value -1 may appear as the value for the nlMatrixTopNReversePktRate and nlMatrixTopNReverseOctetRate objects. This is an invalid value. (25889)

■

You cannot set the value for addressMapMaxDesiredEntries even though the interface appears to accept the value. The value remains at the default of 12,000 entries. (28177)

■

■

You cannot delete a subset of the addressMap entries. You must delete all addressMap entries. (You can recreate entries that you did not originally intend to delete.) (29348)

You cannot delete or re enable a subset of the hlMatrixControlEntry entries. You must delete or re enable all entries. (29794)

86 C HAPTER 6: R ELEASE I SSUES FOR M ULTILAYER S WITCHING M ODULES

Web-based

Management

■

■

■

■

■

■

■

■

Using the WebConsole to renumber the 100th QoS flow classifier removes that classifier when you have defined the maximum number of flow classifiers (100 classifiers). You can renumber all other flow classifiers through the WebConsole correctly. To renumber the 100th flow classifier, you must delete it and redefine it. (29845)

If you define an advertisementAddress for an IP interface through the

WebConsole, you cannot modify the directedBroadcast, icmp Redirect, or proxyArp parameters. You can modify these parameters through the Administration Console. (29897)

You cannot remove a control rate limit using the WebConsole Traffic

Policy Control Remove form when only one rate limit has been configured for the control. Use the command line interface to remove a rate limit. (29711)

The WebConsole and SNMP MIB incorrectly allow the setting of reserved IEEE 802.3x full-duplex pause MAC address as the Bridge

Group Address. (28479)

The WebConsole allows the same name for QoS classifiers and controls. Use unique names for classifiers and controls. (29335)

The WebConsole incorrectly accepts a value outside of the valid range for the QoS statistical interval. The valid range is from 0 through

4294967295. (25160)

Setting the module baseline does not reset the IP interface statistics in the WebConsole.

You cannot create multiple IP interfaces with a single VLAN through the WebConsole. (29611)

4-port GBIC

Multilayer Switching

Module (Model

Number 3CB9RG4)

■

■

The 4-port GBIC Multilayer Switching Module allows you to create unsupported RMON-2 groups. As a result, these groups appear to be configurable through LANsentry ® , even though you cannot configure them. The RMON groups that are supported on this module are statistics, history, alarms, and events. (29264)

The bridge port multicast threshold feature on the 4-port GBIC

Multilayer Switching Module forwards two times as many frames as you specify in the threshold limit. For example, if you specify 1000 frames-per-second as the multicast threshold with the bridge port multicastLimit command, the bridge port forwards 2000 frames-per-second. Use a value that is half the frequency of the per-second forwarding rate that you want to specify. (29967)

7

D

OCUMENTATION

C

HANGES

This chapter describes information that is either incorrect or not documented in the indicated Switch 4007 manuals. The topics are:

■

Terminology

■

■

■

■

■

■

Ethernet

Internet Packet Exchange (IPX)

Quality of Service (QoS)

Web-based Management

Resetting Modules

SNMP Traps

Some Switch 4007 manuals use the product name CoreBuilder ® 9000 in their titles. This is because the heritage of the Switch 4007 is the

CoreBuilder 9000 and certain components of the latter product line are supported in the Switch 4007 chassis.

The Switch 4007 software and management interfaces are built from

CoreBuilder 9000 switch technology. In Switch 4007 software, the interface prompts and displays indicate this heritage. Examples, tables, or figures in the Switch 4007 documentation may also indicate this heritage.

The Switch 4007 management interfaces refer to the Switch 4007

Management Module as the Enterprise Management Engine (EME). In these release notes, the Switch 4007 Management Module is referred to as the EME for brevity.

88 C HAPTER 7: D OCUMENTATION C HANGES

Terminology

In any Switch 4007 document, the terms Layer 3 modules , Layer 3 switching module s, or Multilayer Switching Modules refer to the same group of modules:

■

■

■

12-port 10/100BASE-TX Fast Ethernet Multilayer Switching Module

(Model Number 3CB9RF12R)

4-Port Gigabit Ethernet (GBIC) Multilayer Switching Module

(Model Number 3CB9RG4)

10-port 100BASE-FX Fast Ethernet Multilayer Switching Module

(Model Number 3CB9RF10MC)

Ethernet

In the Ethernet chapter of the Switch 4007 Implementation Guide:

The “Port Monitoring” section states that, with the

ethernet monitoring option set to enabled

or advanced

, the system compares

Ethernet error counters against user-defined thresholds. This is incorrect.

The system compares the error counters against its own predetermined threshold. You cannot modify the threshold.

Internet Packet

Exchange (IPX)

The following information should be added to the IPX chapter of the

Switch 4007 Command Reference Guide :

Instead of displaying the entire current IPX route and server databases

(which may contain thousands of entries), the ipx route display

and ipx server display

options prompt you for a regular expression so that you can search only for those route addresses and service names that interest you.

In the following example, only the service names that match the regular expression us(nt)*

are printed:

Select menu option: ipx server display

Enter service type (0x1-0xffff|*) {?} [*]:

Enter service name pattern {?} [*]: us(nt)*

Interface Name Type Network Node Socket Hops Age

1 USNTBOX 4 11111111 00-00-00-00-00-01 451 1 10

2 EUusNTbox 12 1111 00-00-00-00-00-01 225 2 45

2 HQUsNTBOX 1b 1111 00-00-00-00-00-01 50 2 40

Internet Packet Exchange (IPX) 89

The IPX grep facility actually implements a modified subset of the egrep regular expression syntax. Table 4 lists the pattern-matching characters that are implemented.

Table 4 IPX Grep Regular Expression Pattern-Matching Characters

\

+

?

|

()

^

$

[]

.

Pattern

Character Meaning

* Match any number (or none) of the single character that immediately precedes it. The preceding character can be a regular expression.

Match any single character.

Match the following regular expression at the beginning of a word.

Match the following regular expression at the end of a word.

Match any one of the enclosed characters.

■

A hyphen (-) indicates a range of consecutive characters. For example: [a-z] matches any letter.

■

A caret (^) as the first character in the brackets reverses the sense. It matches any one character not in the list. For example: [^a-z] matches any character that is not a letter.

Turn off the special meaning of the character that follows.

Match one or more instances of the preceding regular expression.

Match zero or one instance of the preceding regular expression.

Match the regular expression specified before or after.

Apply a match to the enclosed group of regular expressions.

Important

Considerations

■

■

■

■

The

ipx route display option searches on the route address only, not on everything in the route database.

The ipx server display

option searches on the service name only, not on everything in the service database. Therefore, you cannot enter a regular expression for the service type, even though the prompt looks identical.

The search is not case sensitive.

Single or double quotation marks have no special meaning. They are simply included as part of the search string. If the command line parser encounters a blank, it stops parsing.

90 C HAPTER 7: D OCUMENTATION C HANGES

Quality of Service

(QoS)

In the QoS and RSVP chapter of the Switch 4007 Implementation Guide :

Now that the 10-port 100BASE-FX Multilayer Switching Module has been added to the Switch 4007 product line, the following paragraph should be added to the last point in the “General Guidelines” subsection of the

“Key Guidelines for Implementation” section:

■

The 10-port 100BASE-FX Fast Ethernet Multilayer Switching Module

(Model Number 3CB9RF10MC) allows aggregation on ports 1-3, 4-5,

6-8, and 9-10. Be aware that if you set the rate limit on different ports, for example, 1, 6, and 9, then aggregate functions the same as receivePort rate limit type.

Web-based

Management

■

The * character in a regular expression is not the familiar “wildcard” character that matches any string.

A * character that appears in a regular expression is interpreted in its correct regular expression sense: it matches any number (or none) of the single character that immediately precedes it.

For example, the single line:

Enter server name pattern {?} [*]: us(nt)* actually contains two entirely different * characters. The first * is the default prompt for “everything” with * in its wildcard sense. The second * is a regular expression pattern-matching character to match any number (or none) of the preceding “nt” expression.

In the Management Tools chapter of the Switch 4007 Getting Started

Guide :

The information for Web Management should include some additional notes and some filename changes:

■

This notice should appear at the beginning of the “Windows

Installation” section:

If you use Netscape as your browser, you must install the Help files on an external server, not on a local PC. To install the Help files on an external server, follow the instructions for installing on a PC.

To configure the location of the Help files, use the sample URL syntax in the Help Server Configuration form in the software.

Resetting Modules 91

■

■

In the first paragraph of the section “Installing Supplemental Tools” and in the “Windows Installation” procedure, an incorrect filename is cited. Use the filename

Setup.exe

instead of webmanagev30.exe

.

In the section “Setting Up the Form-Specific Help Files,” this notice should appear:

If you use Netscape as your browser, you must install the Help files on an external server. To configure the location of the Help files, use the sample

URL syntax in the Help Server Configuration form in the software.

Resetting Modules

Use the following command information to supplement the information that exists about the EME in the Switch 4007 Implementation Guide :

■ reset eme cold

— Use this command with diagnostics enabled to reset the EME and execute diagnostics.

Issuing the reset eme cold

command on a Primary EME causes a failover to the Secondary EME. The Secondary EME becomes the Primary

EME.

■ reset eme warm

— Use this command to reset the EME without executing diagnostics. This command is the default if you do not enter anything after reset eme

.

Issuing the reset eme warm

command on a Primary EME in the lower numbered slot does not cause it to fail over. Issuing the reset eme warm command on a Primary EME in the higher numbered slot causes it to fail over. Issuing the reset eme warm

command on a Secondary EME does not cause it to become Primary.

■ reset module <slot>.<subslot> cold

— Use this command to reset any module and execute diagnostics as if the module had been power cycled.

When you reset the Secondary EME from the Primary EME, you must use the reset module <slot>.2 command. To use the reset eme command, you must be directly connected to the Secondary EME.

■

■ reset module <slot>.<subslot> warm

— Use this command to reset any module without executing diagnostics.

reset module all cold

— Use this command to reset all modules, excluding the EME, and execute diagnostics as if the modules had been power cycled.

92 C HAPTER 7: D OCUMENTATION C HANGES

■ reset module all warm

— Use this command to reset all modules, excluding the EME, without executing diagnostics.

SNMP Traps

There are two topics that deal with SNMP traps:

■

EME Command Changes in Release 3.0.5

■

Setting SNMP Trap Destinations

EME Command

Changes in

Release 3.0.5

This section reviews the EME interface command structure for SNMP traps in software release 3.0.0, then shows the modifications that were made in software release 3.0.5.

Old command structure in software release 3.0.0: set eme set snmp show snmp trap_receive disable/enable disable/enable authentication_trap authentication_trap extensions set trap_destination <community> <ip>

New command structure in software release 3.0.5: set snmp trap destination <community> <ip> filter authentication disable/enable link_state disable/enable topology_change disable/enable disable/enable receive show snmp extensions trap clear snmp trap

The show eme

display has been modified in this release. In version 3.0.0, the show eme

command displayed Trap Receive status (ENABLED or

DISABLED). In version 3.0.5, Trap Receive status is shown under the show snmp trap

command.

SNMP Traps 93

Examples

Examples of setting the link state and topology change traps through the

EME console with accompanying displays of the trap outputs to the console are shown here.

CB9000> set snmp trap filter link_state enable

Link state trap set to ENABLE.

CB9000>

Event Received: “SNMP Trap” Event generated:

2000-02-08T16:31:54 Entry: 00812 Slot: 09.02 Id: 00297

Severity: Inform Type: Trap Enterprise: CoreBuilder-9000 Trap source slot: 03.01 Trap info: Link Down Octet string:

Corebuilder 9000-10 Port 100FX Layer 3 Switching Module, manuf: 3Com, Fast-Ethernet Port 1

CB9000> set snmp trap filter link_state disable

Link state trap set to DISABLE.

(No subsequent generation of the link state traps occurs.)

CB9000> set snmp trap filter topology_change enable

Spanning tree topology change trap set to ENABLE.

Event Received: “SNMP Trap” Event generated:

2000-02-08T16:32:31 Entry: 00816 Slot: 09.02 Id: 00297

Severity: Inform Type: Trap Enterprise: dot1dBridge Trap source slot: 03.01 Generic trap: 6 Specific trap: 2

CB9000> set snmp trap filter topology_change disable

Spanning tree topology change trap set to DISABLE.

(No subsequent generation of the topology change traps occurs.)

Only the topology change trap is filtered with this command. You will continue to see any new root traps. A topology change trap has a specific trap of two. A new root trap has a specific trap of one.

94 C HAPTER 7: D OCUMENTATION C HANGES

Setting SNMP Trap

Destinations

The following EME command is not documented in Switch 4007

Implementation Guide :

■ set snmp trap destination

— Use this command to set a Simple

Network Management Protocol (SNMP) trap to a specific destination.

The correct syntax for this command is:

> set snmp trap destination <community> <number>

Where

<community> is the community string of the selected trap where you want to send the trap and

<number> is the IP address of the trap receiver.

8

S

OFTWARE

I

NSTALLATION

Software Release 3.0.5 is only available from the 3Com Web site. That is, modules that ship from the 3Com factory will be pre-loaded with an earlier release of software. If you want to upgrade modules to run release 3.0.5 in your Switch 4007 system, then read this chapter to learn about the prerequisites, as well as how to download the software from the Web and how to install it.

After you upgrade your modules to release 3.0.5, save these instructions in the event that you need to reinstall software for any reason; for example, if it has become corrupted.

This chapter contains the following main topics:

■

Verifying Module Software Releases and Compatibility

■

■

■

How to Obtain the Latest Software

Terminology for the EME Module

Prerequisites for Software Installation

■

■

■

■

■

■

Upgrading the EME Module to Release 3.0.5

Upgrading Layer 2 Switching Modules to Release 3.0.5

Upgrading Multilayer Switching Modules From 2.1 to 3.0.5

Upgrading Multilayer Switching Modules From 3.0 or 3.0.1 to 3.0.5

Updating Emergency Download Parameters

Downgrading Modules From 3.0.5 to 3.0.0 or 3.0.1

CAUTION: Before you begin upgrading module software, read through all sections in this chapter at least once to ensure that you understand the procedures and that you are aware of all important notices and cautions.

96 C HAPTER 8: S OFTWARE I NSTALLATION

Verifying Module

Software Releases and Compatibility

To determine what software release is currently loaded on a Switch 4007 module, follow either of these sets of steps:

■

■

Verifying Software Releases on an Existing System

Verifying Software Releases on a New System and Modules

Verifying Software Releases on an Existing System

1 Connect to a powered-up chassis that is loaded with at least one EME and a variety of switching modules.

2 At the EME prompt, enter: show module all verbose

Scroll through the resulting screens to find the software release number that is loaded on each installed module in the chassis.

3 Determine which modules need to be upgraded.

CAUTION: Modules in the same chassis must operate at compatible software levels. See Table 1 on page 11 for the required software versions on all modules. Essentially, Table 1 shows that, while the EME module operates with mixed releases as indicated, all other modules must operate with release 3.0.5.

For each EME module, ensure that you check the release numbers on four different images: EMC boot, EMC operational, EME boot, and EME operational. At minimum, you will need to upgrade the EME operational software because it is the only image that operates at release 3.0.5.

On an EME module in slot 8 or 9, notice that a module’s EMC images are always in sub slot 1 and the EME images are always in sub slot 2.

4 Record the slot numbers of the modules that you need to upgrade as well as which type of module.

You will upgrade all EME modules (either one or two) first. Then you can upgrade switching modules in any order.

Verifying Module Software Releases and Compatibility 97

5 If you need to upgrade any piece of software on your EME module, first determine whether the module has the necessary CPU RAM. In the show module all verbose

display, locate this value and consider the following:

■

If the field shows the number 20, you can proceed to upgrade the

EME software.

■

If the field shows the number 8, you need to increase the CPU RAM before you can upgrade the software. You must replace the 4 MB

DRAM card in each EME module with a 16 MB DRAM card. After you install it, it combines with 4 MB of permanent RAM to yield the required 20 MB. To order a 16 MB DRAM card (order number

3CB9EE), contact your network supplier or 3Com representative.

Before you upgrade the RAM, read the related cautions in the section titled “Prerequisites for Software Installation” in this chapter. Installation instructions are provided in the packaging of the 16 MB card.

6 After you have upgraded the CPU RAM on the EME module, upgrade the its software as necessary, following the procedures provided in this chapter:

■

■

Terminology for the EME Module

Upgrading the EME Module to Release 3.0.5

7 Verify that the software on the EME module has been upgraded, then upgrade the software on switching modules (in any order), following the procedures provided in this chapter. For instructions, see the sections:

■

Upgrading Layer 2 Switching Modules to Release 3.0.5

■

■

■

Upgrading Multilayer Switching Modules From 2.1 to 3.0.5

Upgrading Multilayer Switching Modules From 3.0 or 3.0.1 to 3.0.5

Updating Emergency Download Parameters

98 C HAPTER 8: S OFTWARE I NSTALLATION

Verifying Software Releases on a New System and Modules

When you purchase a new chassis and modules or if you have had a system in storage and are just now installing it, you may have a collection of modules with incompatible software. Therefore, unlike the previous procedure, you cannot install all modules and then power up the chassis to check their release numbers. Instead, a step-by-step approach is needed:

1 Install one EME module in an empty, powered-off chassis, according to the instructions in the Enterprise Management Engine Quick Start Guide .

2 Power on the chassis and connect to the EME module via the serial

(RS-232) port. (You cannot Telnet in-band because you do not yet have an

IP address defined for the system.)

3 At the EME prompt, enter: show module all verbose

4 Scroll down to the end of the display and locate the following information for this EME module:

■

Four software images on the module (EMC boot, EMC operational,

EME boot, EME operational)

On an EME module in either slot (8 or 9), notice that the EMC code is always in sub slot 1 and the EME code is always in sub slot 2.

■

CPU RAM size

5 Analyze the EME software and determine your next steps:

■

Do the four different software files on the EME module (EMC boot,

EMC operational, EME boot, EME operational) match the requirements in Table 1 on page 11? If not, note the images that need to be upgraded. (At minimum, you will need to upgrade the EME operational software to release 3.0.5.)

Verifying Module Software Releases and Compatibility 99

6 If you need to upgrade any software on your EME to conform with the requirements in Table 1, before you upgrade it, determine if the EME module has the required amount of CPU RAM:

■

If the display shows the number 20, you can proceed to upgrade the software.

■

If the display shows the number 8, you need to increase the CPU RAM before you can upgrade the software. You must replace the 4 MB

DRAM card in each EME module with a 16 MB DRAM card. After you install it, it combines with 4 MB of permanent RAM to yield the required 20 MB. To order a 16 MB DRAM card (order number

3CB9EE), contact your network supplier or 3Com representative.

Before you upgrade the RAM, read the related cautions in the section titled “Prerequisites for Software Installation” in this chapter. Installation instructions are provided in the packaging of the 16 MB card.

7 Upgrade your EME software, following the procedures in these sections:

■

Terminology for the EME Module

■

Upgrading the EME Module to Release 3.0.5

8 Install all other switching modules in the powered-up chassis, according to the procedure in the module Quick Start Guides .

9 Use the

show module all verbose

command to evaluate their software compatibility with the upgraded EME.

CAUTION: Modules in the same chassis must operate at compatible software levels. See Table 1 on page 11 for compatible software levels.

Essentially, if you have already upgraded the EME, all switching modules must now be upgraded to release 3.0.5.

10 Upgrade the software on each switching module, following the procedures provided in this chapter. For instructions, see the sections:

■

Upgrading Layer 2 Switching Modules to Release 3.0.5

■

■

■

Upgrading Multilayer Switching Modules From 2.1 to 3.0.5

Upgrading Multilayer Switching Modules From 3.0 or 3.0.1 to 3.0.5

Updating Emergency Download Parameters

100 C HAPTER 8: S OFTWARE I NSTALLATION

How to Obtain the

Latest Software

The image files for software release 3.0.5 are available from the 3Com

Web site: http://support.3com.com/infodeli/swlib/index.htm

Refer to Table 1 in Chapter 1 of these release notes to record the software filename for each different type of module. Use this filename to ensure that you download the correct software from the 3Com Web site.

Where to Place the

Software

Download the files to a TFTP server on your network; specifically, to the server directory that is usually named “tftp” or “tftp boot”. In addition, note the IP address of the server. When it is time to install the software on your Switch 4007 modules, you will enter commands from the system prompt that retrieve the software images from this address and directory.

Releases of software prior to Release 2.x.x do not support the EME download operation from a Windows 95 or Windows 98 Telnet session.

You can download software from the EME console interface or through a non-Windows 95 Telnet session.

Terminology for the EME Module 101

Terminology for the

EME Module

An EME module provides controller functions and management functions for the Switch 4007 chassis. Controller functions regulate physical aspects of the chassis such as power and temperature. Management functions regulate more sophisticated aspects such as system configuration, software downloads, and obtaining SNMP data. Because it has these two main functions, an EME module operates with four pieces of software:

■

■

EMC (controller) boot software

EMC (controller) operational software

■

■

EME (management) boot software

EME (management) operational software

Clarifying Acronym

Use

Different uses of the same acronym (EME) can be confusing, so bear this terminology in mind:

■

■

In this chapter, when we refer to the module as a whole, we use the term EME module or, even simpler, the module .

In this chapter, when we refer to EME boot or operational software on the module, we use terms like EME boot software or, even simpler,

EME boot .

Terms for Redundant

EME Modules

In releases prior to 3.0.0, the active EME was called the Master in the CLI displays. Beginning with release 3.0.0, the term changed to Primary .

In releases prior to 3.0.0, the inactive EME was called the Slave in the CLI displays. Beginning with release 3.0.0, the term changed to Secondary .

102 C HAPTER 8: S OFTWARE I NSTALLATION

Prerequisites for

Software

Installation

Read this section for important notices that pertain to Switch 4007 modules prior to upgrading them to software release 3.0.5.

CAUTION: Software installation procedures may also contain prerequisite steps. Be sure to read the complete procedure even if you have updated software on your Switch 4007 in the past.

CAUTION: Modules in the same chassis must operate at compatible software releases. See Table 1 on page 11 for this information.

CAUTION: Before you upgrade software on Switch 4007 switch fabric modules and switching modules to Release 3.0.5, ensure that the EME has already been upgraded with the required amount of CPU RAM and compatible software.

CAUTION: This release requires that each EME module has 20 MB of

CPU RAM. To determine the current CPU RAM size, enter show eme

from the EME command line interface. The

CPU Ram Size (MB)

field on the left side of the screen indicates the amount of CPU RAM. If an EME module has 8 MB of CPU RAM, you must replace the 4 MB DRAM card in each EME module with a 16 MB DRAM card. After you install it, it combines with the 4 MB of permanent RAM to yield the required 20 MB.

To order a 16 MB DRAM card (order number 3CB9EE), contact your network supplier or 3Com representative.

CAUTION: If you do not have dual EMEs installed in a powered up chassis, you will lose power to all modules if you remove the EME module to install the 16 MB card. If you have dual EME modules installed and both of them need RAM upgrades, 3Com recommends that you upgrade the RAM on the Slave/Secondary EME module first. Follow this procedure:

1 Remove the Secondary/Slave EME module and install the 16 MB card.

2 Reinstall the module and wait for it to come back up with

Secondary/Slave status.

3 Remove the Primary/Master EME module and install the 16 MB card.

4 Reinstall the module.

Following this procedure minimizes the disruption to your system (caused by EME switchovers) and network.

Prerequisites for Software Installation 103

CAUTION: Before you attempt to download and install any module software, save the configuration file for the modules that you are upgrading to an external device using the upload

command from the

EME prompt. Name the files so that they uniquely identify the versions of the module code that were in use at the time they were uploaded, as well as the chassis and slot where the module is located (i.e., do not use the default file name). See the Switch 4007 Implementation Guide for an explanation about how to use this command. You must save these configuration files because they are not guaranteed to work with revisions of code other than the ones with which they were created. If at some point you want to downgrade the modules to preexisting code, you must use the configuration file that corresponds to that version of code.

CAUTION: If the EME contains a Release 3.0.x configuration file for a switch fabric module slot or interface module slot, and if you want to install a module in that slot that already has software Release 2.x.x on it, delete the Release 3.0.x configuration file from the EME after you save the 3.0.x configuration file to the external device.

CAUTION: Downloaded software does not become active on a module until you cycle the power to the module slot.

If you are upgrading multiple modules, download software to all modules first and then cycle the power to these module slots.

CAUTION: Do not attempt to perform any management tasks while you are downloading software.

104 C HAPTER 8: S OFTWARE I NSTALLATION

Upgrading the EME

Module to

Release 3.0.5

Release 3.0.5 for the EME module essentially refers to the EME operational image; that is, from the four images on an EME module, the

EME operational image is the only one that has been updated for release 3.0.5. The other three images on the module remain at various previous releases (see Table 1 on page 11 for the required release numbers).

This section describes how to upgrade all four images on the EME module to the required releases. Most likely, however, you will not have to upgrade all four images; but it depends on what images your EME module currently has loaded. At minimum, you will need to upgrade the

EME operational image because it is new at release 3.0.5; however, if the other three images on your EME already conform to the required release numbers, you do not have to reinstall those images.

CAUTION: This section assumes that you have already downloaded the necessary software from the 3Com web site, that your EME has the required 20 MB of CPU RAM, and that you have saved the configuration files that are associated with the current software release on the module.

If you have not completed these tasks, reread all prior sections of this chapter and follow the instructions.

CAUTION: Do not attempt to perform any management tasks while you are downloading software.

This section outlines two different installation procedures — one that applies to a single EME module configuration and one that applies to a dual EME module configuration:

■

Upgrading a Single EME Module

■

Upgrading Two EME Modules

Upgrading the EME Module to Release 3.0.5

105

Upgrading a Single

EME Module

With only one EME module installed in a chassis, the module’s software should be installed in this order:

This is an overview procedure. Each step leads you to a set of specific instructions elsewhere in this chapter. After you complete a set of instructions, return to this procedure and go to the next step.

1 Install EMC boot software, version 2.1. For instructions, see “Installing

EMC Boot Software” later in this chapter. (Do not reinstall if it already exists on your module.) After the module finishes resetting, go to the next step.

2 Install EMC operational software, version 3.0.0. For instructions, see

“Installing EMC Operational Software” later in this chapter. (Do not reinstall if it already exists on your module.) After the module finishes resetting, go to the next step.

3 Install EME boot software, version 3.0.0. For instructions, see “Installing

EME Boot Software” later in this chapter. (Do not reinstall if it already exists on your module.) After the module finishes resetting, go to the next step.

4 Install EME operational software, version 3.0.5. For instructions, see

“Installing EME Operational Software” later in this chapter. After the module finishes resetting, go to the next step.

5 Follow this final step ONLY if you have upgraded any piece of software on your EME module from 2.x to 3.x.x: a Flush the Address Resolution Protocol (ARP) table on the PC management station from which you connected to the EME module. b Flush the ARP table on the TFTP server from which you downloaded software to the EME module. c Ping the server or management station to verify that you are connected properly.

The MAC addresses of the EME were changed in release 3.0.0. The process of flushing the ARP tables on the management station and TFTP server makes these devices relearn the new MAC addresses.

The upgrade process is complete.

106 C HAPTER 8: S OFTWARE I NSTALLATION

Upgrading Two

EME Modules

For a chassis with two EME modules installed, upgrade the software images according to this procedure:

This is an overview procedure. Each step leads you to a set of specific instructions elsewhere in this chapter. After you complete a set of instructions, return to this procedure and go to the next step.

1 At the EME interface prompt, enter show module all

to determine which EME module is on Master/Primary status and which is on

Slave/Secondary status.

For an explanation of these terms, refer to the section “Terminology for the EME Module” earlier in this chapter.

2 Upgrade the software on the Slave/Secondary EME module.

Note the slot number of this module. You will use it in some substeps.

a Install EMC boot software, version 2.1. For instructions, see “Installing

EMC Boot Software” later in this chapter. (Do not reinstall it if this version already exists on your module.) After the module finishes resetting, go to the next step.

b Install EMC operational software, version 3.0.0. For instructions, see

“Installing EMC Operational Software” later in this chapter. (Do not reinstall it if this version already exists on your module.) After the module finishes resetting, go to the next step.

c Install EME boot software, version 3.0.0. For instructions, see

“Installing EME Boot Software” later in this chapter. (Do not reinstall it if this version already exists on your module.) After the module finishes resetting, go to the next step.

d Install EME operational software, version 3.0.5. For instructions, see

“Installing EME Operational Software” later in this chapter. After the module finishes resetting, go to the next step.

3 Upgrade the software on the Master/Primary EME module.

Note the slot number of this module. You will use it in some substeps.

a Install EMC boot software, version 2.1. For instructions, see “Installing

EMC Boot Software” later in this chapter. (Do not reinstall it if this version already exists on your module.) After the module finishes resetting, go to the next step.

b Install EMC operational software, version 3.0.0. For instructions, see

“Installing EMC Operational Software” later in this chapter. (Do not reinstall it if this version already exists on your module.) After the module finishes resetting, go to the next step.

Upgrading the EME Module to Release 3.0.5

107 c Install EME boot software, version 3.0.0. For instructions, see

“Installing EME Boot Software” later in this chapter. (Do not reinstall it if this version already exists on your module.) After the module finishes resetting, go to the next step.

CAUTION: When the module resets during this step, a switchover in EME status occurs. The module that used to be Secondary becomes Primary.

Leave it this way for now, even if you do not want this to be the final configuration; you can switch the EME module hierarchy later, but it is important to complete the entire upgrade process first. d In order to continue the upgrade process, if you have only had one connection to your system (the old Primary EME module), disconnect it now and move the cable to connect to the new Primary EME module. e Follow this step ONLY if you have upgraded any piece of software thus far on either EME module from 2.1 to 3.0.0 or from 2.1 to 3.0.5

(otherwise, skip to the next step):

■

■

■

Flush the Address Resolution Protocol (ARP) table on the PC management station from which you connected to the EME module.

Flush the ARP table on the TFTP server from which you downloaded software to the EME module.

Ping the server or management station to verify that you are connected properly.

The MAC addresses of the EME were changed in release 3.0. The process of flushing the ARP tables on the management station and

TFTP server makes these devices relearn the new MAC addresses. f Install EME operational software, version 3.0.5, onto what used to be the Primary EME module but is now the Secondary EME. For instructions, see “Installing EME Operational Software” later in this chapter. After the module finishes resetting, go to the next step.

g If you upgraded the EME operational software from release 2.x to release 3.0.5, repeat step 3e again (flush the ARP tables on your management station and TFTP server). Otherwise, ignore this step.

The upgrade process is complete.

108 C HAPTER 8: S OFTWARE I NSTALLATION

Installing EMC Boot

Software

To upgrade a module’s EMC boot software to version 2.1, enter the following commands and information:

When downloading this software to either a Master/Primary module or a

Slave/Secondary module, use this command:

CB9000> download module <slot.subslot> boot_image <server address> emc20100.bt

Where

<slot.subslot>

is either 8.1 or 9.1 and

<server address>

is the

IP address of the server where the software image resides.

The system indicates that downloading software will cause the module to reset and asks if you want to continue. Enter y

(yes).

Installing EMC

Operational Software

To upgrade a module’s EMC operational software to version 3.0.0, enter the following commands and information:

When downloading this software to either a Master/Primary module or a

Slave/Secondary module, use this command:

CB9000> download module <slot.subslot> oper_image <server address> emc30000.op

Where

<slot.subslot>

is either 8.1 or 9.1 and

<server address>

is the

IP address of the server where the software image resides.

The system indicates that downloading software will cause the module to reset and asks if you want to continue. Enter y

(yes).

Upgrading the EME Module to Release 3.0.5

109

Installing EME Boot

Software

To upgrade a module’s EME boot software to version 3.0.0, enter the following commands and information:

When downloading to a Slave/Secondary module, use this command:

CB9000> download module <slot.subslot> boot_image <server address> eme30000.bt

When downloading to a Master/Primary module, use this command:

CB9000> download eme boot_image <server address> eme30000.bt

Where

<slot.subslot>

is either 8.2 or 9.2 and

<server address>

is the

IP address of the server where the software image resides.

The system indicates that downloading software will cause the module to reset and asks if you want to continue. Enter y

(yes).

Installing EME

Operational Software

To upgrade a module’s EME operational software to version 3.0.5, enter the following commands and information:

When downloading to a Slave/Secondary module, use this command:

CB9000> download module <slot.subslot> oper_image <server address> eme30005.op

When downloading to a Master/Primary module, use this command:

CB9000> download eme oper_image <server address> eme30005.op

Where

<slot.subslot>

is either 8.2 or 9.2 and

<server address>

is the

IP address of the server where the software image resides.

The system indicates that downloading software will cause the module to reset and asks if you want to continue. Enter y

(yes).

110 C HAPTER 8: S OFTWARE I NSTALLATION

Upgrading Layer 2

Switching Modules to Release 3.0.5

This section explains how to download software release 3.0.5 to a Layer 2

Switch Fabric Module or Layer 2 Switching Module that is currently loaded with any release previous to 3.0.5.

This process assumes that:

■

■

You have already upgraded the software on the EME modules in your system to the release numbers noted in Table 1 on page 11 .

You have downloaded (from the 3Com Web site) the appropriate

Release 3.0.5 images in a TFTP server directory (usually named “tftp” or “tftpboot”). For the filenames to download, consult Table 1 .

1 At the EME prompt, enter the following command: cb9000> download module <slot>.1 all_image <server address>

<filename>

Where

<slot>

is the slot number of the module,

<server address>

is the IP address of the TFTP server where the software image resides, and

<filename> is the filename of the software for the particular module that you want to upgrade.

The software filename has the extension .all

because it includes several images in one file.

The filename that you enter may include a path to the directory within the file server if this directory is not the file server’s default directory.

2 Repeat Step 1 for each Layer 2 module in the chassis.

3 For downloads to take effect on these modules, you must reset them.

To reset all modules simultaneously, at the system prompt, enter: cb9000> reset module all cold

To reset modules individually, at the system prompt, enter: cb9000> reset module <slot>.1 cold

4 Confirm that all of the downloads were successful by viewing a display.

At the system prompt, enter: cb9000> show module <slot>.1 verbose or cb9000> show module all verbose

5 Update the Emergency Download parameters, if necessary. See

“Updating Emergency Download Parameters” later in this chapter.

Upgrading Multilayer Switching Modules From 2.1 to 3.0.5

111

Upgrading

Multilayer

Switching Modules

From 2.1 to 3.0.5

This section explains the process for upgrading a Multilayer Switching

Module to Release 3.0.5 if the module is starting from Release 2.x.

If the module is starting from Release 3.0 or 3.0.1, see “Upgrading

Multilayer Switching Modules From 3.0 or 3.0.1 to 3.0.5” (the next major section in this chapter).

This process assumes that:

■

You have already upgraded the software on the EME modules in your system to the release numbers noted in Table 1 on page 11 .

■

You have downloaded (from the 3Com Web site) the appropriate

Release 3.0.5 images in a TFTP server directory (usually named “tftp” or “tftpboot”). To determine the filenames to download, consult

Table 1 .

1 Refer to the section “Prerequisites for Software Installation” to ensure you have complied with the prerequisites stated there and that you have read important information.

2 Before you upgrade a module, perform the following steps: a On a given module, note which backplane ports are enabled. Use the ethernet detail

command to bring up a display of information.

b If you have four trunk groups defined on a 12-port Multilayer

Switching Module or 10-port Multilayer Switching Module, remove one trunk group.

c If you use QoS on a given module, review the Release 2.x configuration of QoS classifiers and controls. Modify your classifiers and applied controls to reduce the number of classifiers to fewer than

50. To modify or remove a classifier, you must first remove its associated control.

3 To download software to the module, at the system prompt, enter: cb9000> download module <slot>.1 all_image <server address>

<filename>

Where

<slot>

is the slot number of the module,

<server address>

is the IP address of the TFTP server where the new software image resides, and

<filename> is the filename of the software for the particular module that you want to upgrade.

The software filename has the extension .all

because it includes several images in one file.

112 C HAPTER 8: S OFTWARE I NSTALLATION

The filename that you enter may include a path to the directory within the file server if this directory is not the file server’s default directory.

4 Reset the module so that the new software takes effect. At the system prompt, enter: cb9000> reset module <slot>.1 cold

Where

<slot>

is the slot number of the module.

5 Confirm that the download was successful. At the system prompt, enter: cb9000> show module <slot>.1 verbose

6 Some settings may have not survived the transition from Release 2.x to

Release 3.0.5. Examine the following settings as shown in Table 5 and change them as necessary:

Table 5 Parameters to Change After an Upgrade from Release 2.x to 3.0.5

Release 3.0.5 Command Value Set in Release 2.x

qos control define Trunk controls defined bridge trunk define Backplane trunk port defined bridge trunk define Fast Ethernet or Gigabit

Ethernet trunk defined ipx interface SAPadvertising Disabled

Issue or Value After

Release 3.0.5 Upgrade

Control index is incorrect

Some backplane trunk ports may not be enabled

Requested flow control may be incorrect

Enabled

Action to Take

After Upgrade

See Note 1

See Note 2

See Note 3

See Note 4

Note 1 : If trunked ports are defined, and QoS controls are defined for these trunked ports, they are not properly indexed after the Release 3.0.5 upgrade. You must redefine the controls.

Note 2 : After you upgrade from Release 2.x to Release 3.0.5, all of the backplane links on a trunk may not be enabled after the upgrade. This problem affects all switch fabric modules. Before you downloaded

Release 3.0.5 to a module, you should have noted which backplane ports were enabled. After the upgrade is complete and the module has finished resetting, log into the switch fabric module and reexamine the backplane port status using the ethernet detail

command. Verify that all backplane ports that were enabled under Release 2.x are still enabled. If any port that should be enabled is erroneously disabled, use the ethernet portState enable

command to reenable the port.

Upgrading Multilayer Switching Modules From 2.1 to 3.0.5

113

Note 3: If Fast Ethernet or Gigabit Ethernet trunks are configured in

Release 2.x, the requested flow control value in the displays may appear incorrectly after you upgrade to Release 3.0.5. This has no impact on traffic flow through the device.

Note 4: Before you connect to a multilayer switching module to the network, reset the SAPadvertising parameter to its Release 2.x value.

7 Update the Emergency Download parameters for the module, if necessary. See “Updating Emergency Download Parameters” later in this chapter.

114 C HAPTER 8: S OFTWARE I NSTALLATION

Upgrading

Multilayer

Switching Modules

From 3.0 or 3.0.1 to

3.0.5

This section explains the process for upgrading a Multilayer Switching

Module to Release 3.0.5 if the module is starting from Release 3.0.0 or

Release 3.0.1. This upgrade procedure is simple and straightforward — no feature settings should change as the module transitions from Release

3.0.0 or 3.0.1 to Release 3.0.5.

However, before you begin the upgrade to 3.0.5, if you have previously upgraded the module from Release 2.1 to 3.0.0 or 3.0.1

, 3Com recommends that you first check the settings of some features, as that particular upgrade transition caused many settings to change.

It is highly probable that you would have already noticed such conditions.

However, we thought it prudent to list the issues again, so that you can fix any settings before you upgrade a module to Release 3.0.5.

Transition Issues from

Release 2.1 to 3.0

Table 6 outlines the issues that occur when a Multilayer Switching

Module transitions from Release 2.1 to Release 3.0.0 or 3.0.1.

Table 6 NVRAM Conversion Issues That Result When Upgrading From Release 2.1 to Release 3.0.0 or 3.0.1

Release 3.0 or 3.0.1 Command ip multicast igmp snooping

Value Set in

Release 2.x

Enabled

Issue or Value After

Upgrade to Release

3.0.0 or Release 3.0.1

Disabled

Action to Take

Reset to the Release 2.x value

(see Note 5)

Reset to the Release 2.x value ip multicast igmp query ip multicast igmp interface ttl ip multicast dvmrp mode

Enabled

Enabled

Enabled ip multicast dvmrp interface metric 1

Disabled

Disabled

Disabled

0

Reset to the Release 2.x value

Reset to the Release 2.x value

(See Note 6)

Reset to the Release 2.x value ip multicast dvmrp tunnels define qos control define

Tunnel IP address

Trunk controls defined

Tunnels do not convert Redefine up to 8 DVMRP tunnels per module

Control index is incorrect If trunked ports are defined, and QoS controls are defined for each port, redefine the controls.

qos control define qos classifier define bridge trunk define

Ports assigned to control

TCP/UDP source port range

Backplane trunk port defined

Port numbers may be incorrect

Source port range is incorrect

Some backplane trunk ports may not be enabled.

Modify each QoS control to reselect the desired ports

Change the source range

(0-65535) or a desired start-end range

See Note 7.

Upgrading Multilayer Switching Modules From 3.0 or 3.0.1 to 3.0.5

115

Table 6 NVRAM Conversion Issues That Result When Upgrading From Release 2.1 to Release 3.0.0 or 3.0.1

Release 3.0 or 3.0.1 Command bridge trunk define bridge packetFilter create bridge vlan define ipx interface SAPadvertising

Value Set in

Release 2.x

FEN or GEN trunk defined

Port Group filter defined

NetBIOS protocol

VLAN defined

Disabled

Issue or Value After

Upgrade to Release

3.0.0 or Release 3.0.1

Requested flow control may be incorrect

No filters defined

Enabled

Action to Take

See Note 8.

Recreate the port group filters

VLAN protocol not listed See Note 9.

See Note 10.

Note 5: You must make these IGMP changes regardless of whether you use IP multicast routing on your module. If you cannot change these

IGMP settings and IP multicast traffic is present, remove the source of the multicast traffic, change the settings, and then restore the source of the multicast traffic. This correction is a one-time-only change.

Note 6: After you upgrade to Release 3.0.0, reset your module, and implement the IGMP changes noted in Table 6 , view the DVMRP parameters immediately. The default settings for DVMRP mode, DVMRP metric, and DVMRP tunnels were applied to all IP interfaces at Release

2.x, however they are applied per specified IP interface for Release 3.0.0 and higher. Thus, you may need to make changes in your configuration.

After Release 3.0.0 is in effect, you can apply DVMRP parameters only to the IP interfaces that need them.

Note 7 : After you upgrade from Release 2.x to Release 3.0.5, all of the backplane links on a trunk may not be enabled after the upgrade. This problem affects all switch fabric modules. After you upgrade to Release

3.0, log into the switch fabric module and reexamine the backplane port status using the ethernet detail command. Verify that all backplane ports that were enabled under Release 2.x are still enabled. If any ports that should be enabled are erroneously disabled, use the

ethernet portState enable

command to reenable the port.

Note 8: If Fast Ethernet or Gigabit Ethernet trunks are configured in

Release 2.x, the requested flow control value in the displays may appear incorrectly after you upgrade to Release 3.0.0. This has no impact on traffic flow through the device.

116 C HAPTER 8: S OFTWARE I NSTALLATION

Note 9: The NetBEUI protocol-based VLANs were incorrectly named

NetBIOS in Release 2.x. This naming has been corrected in Release 3.0.

However, the bridge vlan detail

command does not display the protocol type for NetBEUI VLANs that are converted from Release 2.x.

Note 10: Before you connect to a multilayer switching module to the network, reset the SAPadvertising parameter to its Release 2.x value.

Upgrade Procedure To upgrade a multilayer switching module from Release 3.0.0 or 3.0.1 to

Release 3.0.5, follow these steps:

This process assumes that:

■

■

You have already upgraded the software on the EME modules in your system to the release numbers noted in Table 1 on page 11 .

You have downloaded (from the 3Com Web site) the appropriate

Release 3.0.5 images in a TFTP server directory (usually named “tftp” or “tftpboot”). To determine the filenames to download, consult

Table 1 .

1 Refer to the section “Prerequisites for Software Installation” to ensure you have complied with the prerequisites stated there and that you have read important information.

2 To download software to the module, at the system prompt, enter: cb9000> download module <slot>.1 all_image <server address>

<filename>

Where

<slot>

is the slot number of the module,

<server address>

is the IP address of the TFTP server where the new software image resides, and

<filename> is the filename of the software for the particular module that you want to upgrade.

The software filename has the extension .all

because it includes several images in one file.

The filename that you enter may include a path to the directory within the file server if this directory is not the file server’s default directory.

3 Reset the module so that the new software takes effect. At the system prompt, enter: cb9000> reset module <slot>.1 cold

Where

<slot>

is the slot number of the module.

Updating

Emergency

Download

Parameters

Updating Emergency Download Parameters 117

4 Confirm that all of the downloads were successful by viewing a display.

At the system prompt, enter: cb9000> show module <slot>.1 verbose

5 Update the Emergency Download parameters, if necessary. See

“Updating Emergency Download Parameters” next in this chapter.

The procedure in this section allows you to change the filename and IP address of the emergency download parameters that are stored in

NVRAM. These values are used to retrieve a new image if the current operational code on a module becomes corrupted. After the module resets, the boot code detects that the operational code is corrupted and automatically attempts to download a valid operational code image with the stored filename and IP address. This image is not stored in permanent memory. You must perform a normal download of this image at the EME prompt to download this image into flash.

CAUTION: If you do not perform this normal download, the emergency download repeats each time that the module resets until the maximum number of module resets is reached.

This procedure is only required the first time that you upgrade a module to software Release 3.0. From that point, any time that an image is downloaded to the module, these parameters update automatically.

If the IP address of the file server that contains this file changes, or if the directory where it is located changes, or if the actual filename changes, you must use this procedure to update that information.

1 From the EME, connect to the module:

CB9000> connect <slot>.1

The Switch 4007 menu options screen appears.

2 At the prompt, enter module

The Module menu options screen appears.

3 At the prompt, enter nvData

The nvdata menu options screen appears.

118 C HAPTER 8: S OFTWARE I NSTALLATION

4 At the prompt, enter emergencyDownload

The following prompt appears:

Enter file server IP address {?} [<IP address>]:

Where

<IP address>

is the location of the server where the non-corrupted software image file resides.

5 To modify the indicated IP address, enter an IP address, or to keep the existing IP address, press Enter.

The following prompt appears:

Enter image Name {?} [<image name>]:

Where

<image name>

is the stored image that you would want to have downloaded.

6 Enter the < image name

> for the module to which you are connected, i.e., enter the filename with the extension .all

for this module.

The following prompt appears:

Enter file type {?} [file type]:

7 Enter all_image

for file type

.

8 Enter q

and press Enter until you reach the top-level menu options screen.

9 To close the menu options screen, enter disconnect

10 To update the Automatic Emergency Download fields in the EME show module <slot>.1

command display, reset the module using the following EME command:

CB9000> reset module <slot>.1

You must complete step 10 to ensure that the module is updated.

Downgrading Modules From 3.0.5 to 3.0.0 or 3.0.1

119

Downgrading

Modules From 3.0.5 to 3.0.0 or 3.0.1

In the event that you need to downgrade all modules in a Switch 4007 system to Release 3.0.0 or 3.0.1 (except for EMC boot code which remains at Release 2.1), follow these steps:

Releases previous to 3.0.0 are not supported on the Switch 4007.

1 From the 3Com Web site, download the Switch 4007 Release Notes for

Release 3.0.0

.

Review the software compatibility requirements for each type of module

(for example, some Multilayer Switching Modules may use Release 3.0.1) and note the software filenames.

2 From the 3Com Web site, download the appropriate software files.

3 Have ready the Release 3.0.0 or 3.0.1 configuration files that you saved before you upgraded your modules to Release 3.0.5.

You should not have to use them, i.e., the Release 3.0.5 configurations should work fine after the downgrade to 3.0.0 or 3.0.1, but this is a precaution.

4 Save your Release 3.0.5 configuration files to a network server as a precaution, in the event that you want to upgrade your modules back to

Release 3.0.5 at some point.

Your Release 3.0.0 configurations should work fine after an upgrade to

Release 3.0.5, but this is a precaution.

5 Download the appropriate *

.all software files to the switching modules in the system.

After each module’s download is complete, if you are prompted to reset the module, enter n (no).

6 From the system prompt, enter: cb9000> reset module all cold

This command resets all modules simultaneously. After the reset is complete, your switching modules will be operating at Release 3.0.0 or

3.0.1 (depending on the module) and the EME modules remain operating at Release 3.0.5. Continue with the next step.

120 C HAPTER 8: S OFTWARE I NSTALLATION

7 To downgrade an EME module from Release 3.0.5 to Release 3.0.0, you only have to downgrade the EME operational image. The other three images on an EME module should already be the appropriate releases.

■

■

If two EME modules are installed:

Downgrade the image on the Secondary EME module first. At the system prompt, enter:

■

CB9000> download module <slot.subslot> oper_image

<server address> eme30005.op

Where

<slot.subslot>

is either 8.2 or 9.2 and

<server address>

is the IP address of the server where the software image resides.

The system indicates that downloading software will cause the module to reset and asks if you want to continue. Enter y

(yes).

Then, downgrade the image on the Master/Primary module. At the system prompt, enter:

CB9000> download eme oper_image <server address> eme30005.op

Where

<slot.subslot>

is either 8.2 or 9.2 and

<server address>

is the IP address of the server where the software image resides.

The system indicates that downloading software will cause the module to reset and asks if you want to continue. Enter y

(yes).

At this point, a switchover in EME module status will occur. That is, the module that used to be Secondary is now Primary.

■

If a single EME is installed, use this command to downgrade the EME operational image:

CB9000> download eme oper_image <server address> eme30005.op

Where

<slot.subslot>

is either 8.2 or 9.2 and

<server address>

is the IP address of the server where the software image resides.

The system indicates that downloading software will cause the module to reset and asks if you want to continue. Enter y

(yes).

R

EFERENCE

9

MIB Files

This chapter contains the following topics:

■

MIB Files

■

■

Entering Commands

Year 2000 Compliance

The organization of a Management Information Base (MIB) allows a

Simple Network Management Protocol (SNMP) network management application to manage a network device without having a specific description of that device. The Switch 4007 Management Module (EME) supports the following MIBs in software Release 3.0.x:

■

MIB2-MIB.mib

— MIB-II MIB, RFC 1213.

■

■

Unsupported group in this MIB: egp group

IF-MIB-V1SMI.mib

— IF MIB Version 1, RFC 1573.

Unsupported tables in this MIB: ifTestTable

■

■

■

■

■

■

■ ifRcvAddressTable ifHC 64-bit counters

ENTITY-MIB.mib

— Entity MIB, RFC 2037.

Unsupported table and groups in this MIB: entAliasMappingTable entity General group

■ entity Notifications group cb9000.mib

cb9eme.mib

122 C HAPTER 9: R EFERENCE

Entering

Commands

Year 2000

Compliance

MIB names and numbers are usually retained when organizations restructure their businesses; therefore, some of the 3Com Enterprise MIB names may not contain the word “3Com.”

Before you enter any command, 3Com recommends that you:

■

■

Examine all menus and submenus carefully to obtain the complete and correct command string.

Consult the documentation for the valid minimum abbreviation for the command string.

For information on Year 2000 compliance and 3Com products, visit the

3Com Year 2000 Web page: http://www.3com.com/products/yr2000.html

10

T

ECHNICAL

S

UPPORT

3Com provides easy access to technical support information through a variety of services. This chapter describes these services.

Information contained in this chapter is correct at time of publication. For the most recent information, 3Com recommends that you access the

3Com Corporation World Wide Web site.

Online Technical

Services

3Com offers worldwide product support 24 hours a day, 7 days a week, through the following online systems:

■

World Wide Web site

■

■

■

■

3Com Knowledgebase Web Services

3Com FTP site

3Com Bulletin Board Service (3Com BBS)

3Com Facts

SM Automated Fax Service

World Wide Web Site To access the latest networking information on the 3Com Corporation

World Wide Web site, enter this URL into your Internet browser: http://www.3com.com/

This service provides access to online support information such as technical documentation and software, as well as support options that range from technical education to maintenance and professional services.

3Com

Knowledgebase Web

Services

This interactive tool contains technical product information compiled by

3Com expert technical engineers around the globe. Located on the World

Wide Web at http://knowledgebase.3com.com

, this service gives all

3Com customers and partners complementary, round-the-clock access to technical information on most 3Com products.

124 C HAPTER 10: T ECHNICAL S UPPORT

3Com FTP Site Download drivers, patches, software, and MIBs across the Internet from the

3Com public FTP site. This service is available 24 hours a day, 7 days a week.

To connect to the 3Com FTP site, enter the following information into your FTP client:

■

■

■

Hostname: ftp.3com.com

Username: anonymous

Password:

<your Internet e-mail address>

You do not need a user name and password with Web browser software such as Netscape Navigator and Internet Explorer.

3Com Bulletin Board

Service

The 3Com BBS contains patches, software, and drivers for 3Com products.

This service is available through analog modem or digital modem (ISDN)

24 hours a day, 7 days a week.

Access by Analog Modem

To reach the service by modem, set your modem to 8 data bits, no parity, and 1 stop bit. Call the telephone number nearest you:

Country

Australia

Brazil

France

Germany

Hong Kong

Italy

Japan

Mexico

P.R. of China

Taiwan, R.O.C.

U.K.

U.S.A.

Data Rate Telephone Number

Up to 14,400 bps 61 2 9955 2073

Up to 28,800 bps 55 11 5181 9666

Up to 14,400 bps 33 1 6986 6954

Up to 28,800 bps 4989 62732 188

Up to 14,400 bps 852 2537 5601

Up to 14,400 bps 39 2 27300680

Up to 14,400 bps 81 3 5977 7977

Up to 28,800 bps 52 5 520 7835

Up to 14,400 bps 86 10 684 92351

Up to 14,400 bps 886 2 377 5840

Up to 28,800 bps 44 1442 438278

Up to 53,333 bps 1 847 262 6000

Support from Your Network Supplier 125

Access by Digital Modem

ISDN users can dial in to the 3Com BBS using a digital modem for fast access up to 64 Kbps. To access the 3Com BBS using ISDN, call the following number:

1 847 262 6000

3Com Facts

Automated Fax

Service

The 3Com Facts automated fax service provides technical articles, diagrams, and troubleshooting instructions on 3Com products 24 hours a day, 7 days a week.

Call 3Com Facts using your Touch-Tone telephone:

1 408 727 7021

Support from Your

Network Supplier

If you require additional assistance, contact your network supplier. Many suppliers are authorized 3Com service partners who are qualified to provide a variety of services, including network planning, installation, hardware maintenance, application training, and support services.

When you contact your network supplier for assistance, have the following information ready:

■

Product model name, part number, and serial number

■

■

■

A list of system hardware and software, including revision levels

Diagnostic error messages

Details about recent configuration changes, if applicable

If you are unable to contact your network supplier, see the following section on how to contact 3Com.

Support from 3Com

If you are unable to obtain assistance from the 3Com online technical resources or from your network supplier, 3Com offers technical telephone support services. To find out more about your support options, call the

3Com technical telephone support phone number at the location nearest you.

126 C HAPTER 10: T ECHNICAL S UPPORT

When you contact 3Com for assistance, have the following information ready:

■

■

Product model name, part number, and serial number

A list of system hardware and software, including revision levels

■

■

Diagnostic error messages

Details about recent configuration changes, if applicable

Here is a list of worldwide technical telephone support numbers:

Country

Asia, Pacific Rim

Australia

Hong Kong

India

Indonesia

Japan

Malaysia

New Zealand

Pakistan

Philippines

Telephone Number

1 800 678 515

800 933 486

+61 2 9937 5085

001 800 61 009

0031 61 6439

1800 801 777

0800 446 398

+61 2 9937 5085

1235 61 266 2602

Europe

From anywhere in Europe, call: +31 (0)30 6029900 phone

+31 (0)30 6029999 fax

Europe, South Africa, and Middle East

From the following countries, you may use the toll-free numbers:

Austria

Belgium

Denmark

Finland

France

Germany

Hungary

Ireland

Israel

Italy

Latin America

Argentina

Brazil

Chile

Colombia

North America

0800 297468

0800 71429

800 17309

0800 113153

0800 917959

0800 1821502

00800 12813

1800 553117

1800 9453794

1678 79489

AT&T +800 666 5065

0800 13 3266

1230 020 0645

98012 2127

1 800 NET 3Com

(1 800 638 3266)

Country

P.R. of China

Singapore

S. Korea

From anywhere in S. Korea:

From Seoul:

Taiwan, R.O.C.

Thailand

Netherlands

Norway

Poland

Portugal

South Africa

Spain

Sweden

Switzerland

U.K.

Mexico

Peru

Puerto Rico

Venezuela

Enterprise Customers:

1 800 876-3266

Telephone Number

10800 61 00137 or

021 6350 1590

800 6161 463

00798 611 2230

(0)2 3455 6455

0080 611 261

001 800 611 2000

0800 0227788

800 11376

00800 3111206

0800 831416

0800 995014

900 983125

020 795482

0800 55 3072

0800 966197

01 800 CARE (01 800 2273)

AT&T +800 666 5065

800 666 5065

AT&T +800 666 5065

Returning Products for Repair 127

Returning Products for Repair

Before you send a product directly to 3Com for repair, you must first obtain an authorization number. Products sent to 3Com without authorization numbers will be returned to the sender unopened, at the sender’s expense.

To obtain an authorization number, call or fax:

Country

Asia, Pacific Rim

Europe, South Africa, and

Middle East

Telephone Number

+ 65 543 6500

+ 31 30 6029900

Fax Number

+ 65 543 6348

+ 31 30 6029999

Latin America 1 408 326 2927 1 408 326 3355

From the following countries, you may call the toll-free numbers; select option 2 and then option 2:

Austria

Belgium

Denmark

Finland

France

Germany

Hungary

Ireland

Israel

Italy

Netherlands

Norway

Poland

Portugal

South Africa

Spain

Sweden

Switzerland

U.K.

0800 297468

0800 71429

800 17309

0800 113153

0800 917959

0800 1821502

00800 12813

1800553117

1800 9453794

1678 79489

0800 0227788

800 11376

00800 3111206

0800 831416

0800 995014

900 983125

020 795482

0800 55 3072

0800 966197

U.S.A. and Canada 1 800 NET 3Com

(1 800 638 3266)

Enterprise Customers:

1 800 876 3266

1 408 326 7120

(not toll-free)