and the Alcatel logo are registered trademarks of Alcatel. Xylan®, OmniSwitch®, OmniStack®,
®
are registered trademarks of Alcatel Internetworking, Inc.
OmniAccess™, Omni Switch/Router™, PolicyView™, RouterView™, SwitchManager™, VoiceView™,
WebView™, X-Cell™, X-Vision™, and the Xylan logo are trademarks of Alcatel Internetworking, Inc.
This OmniSwitch product contains components which may be covered by one or more of the following
U.S. Patents:
•U.S. Patent No. 6,339,830
•U.S. Patent No. 6,070,243
•U.S. Patent No. 6,061,368
•U.S. Patent No. 5,394,402
•U.S. Patent No. 6,047,024
•U.S. Patent No. 6,314,106
•U.S. Patent No. 6,542,507
26801 West Agoura Road
Calabasas, CA 91301
(818) 880-3500 FAX (818) 880-3505
info@ind.alcatel.com
US Customer Support—(800) 995-2696
International Customer Support—(818) 878-4507
Internet—http://eservice.ind.alcatel.com
iiOmniSwitch Troubleshooting GuideSeptember 2005
Page 3
Contents
About This Guide .........................................................................................................xv
Supported Platforms ......................................................................................................... xv
Who Should Read this Manual? ...................................................................................... xvi
When Should I Read this Manual? .................................................................................. xvi
What is in this Manual? ................................................................................................... xvi
What is Not in this Manual? ...........................................................................................xvii
How is the Information Organized? ...............................................................................xvii
Related Documentation ..................................................................................................xvii
Before Calling Alcatel’s Technical Assistance Center .................................................... xx
Chapter 1Troubleshooting the Switch System ......................................................................1-1
In This Chapter ................................................................................................................1-1
This OmniSwitch Troubleshooting Guide describes how to use Command Line Interface (CLI) and Dshell
commands available on the OmniSwitch 6600 Family, OmniSwitch 6800 Series, OmniSwitch 7700/7800,
and the OmniSwitch 8800 to troubleshoot switch and network problems.
Supported Platforms
This information in this guide applies to the following products:
• OmniSwitch 6624 (OmniSwitch 6600-24)
• OmniSwitch 6648 (OmniSwitch 6600-48)
• OmniSwitch 6600-P24
• OmniSwitch 6600-U24
• OmniSwitch 6602-24
• OmniSwitch 6602-48
• OmniSwitch 6800
• OmniSwitch 7700
• OmniSwitch 7800
• OmniSwitch 8800
Note. All references to OmniSwitch 6624 and 6648 switches also apply to the OmniSwitch 6600-P24,
OmniSwitch 6600-U24, OmniSwitch 6602-24, and OmniSwitch 6602-48 unless specified otherwise.
Unsupported Platforms
The information in this guide does not apply to the following products:
• OmniSwitch (original version with no numeric model name)
Note. Troubleshooting documentation for legacy products (e.g., Omni Switch/Router) can be downloaded
at http://support.ind.alcatel.com/releasefiles/indexpage.cfm.
Who Should Read this Manual?
The principal audience for this user guide is Service and Support personnel who need to troubleshoot
switch problems in a live network. In addition, network administrators and IT support personnel who need
to configure and maintain switches and routers can use this guide to troubleshoot a problem upon advice
from Alcatel Service and Support personnel..
However, this guide is not intended for novice or first-time users of Alcatel OmniSwitches. Misuse or failure to follow procedures in this guide correctly can cause lengthy network down time and/or permanent
damage to hardware. Caution must be followed on distribution of this document.
When Should I Read this Manual?
Always read the appropriate section or sections of this guide before you log into a switch to troubleshoot
problems. Once you are familiar with the commands and procedures in the appropriate sections you can
use this document as reference material when you troubleshoot a problem.
What is in this Manual?
The principal sections (i.e., the chapters numbered numerically) use CLI and Dshell commands to analyze
and troubleshoot switch problems. Each section documents a specific switch feature (e.g., hardware, server
load balancing, routing).
Note. Dshell commands should only be used by Alcatel personnel or under the direction of Alcatel.
Misuse or failure to follow procedures that use Dshell commands in this guide correctly can cause lengthy
network down time and/or permanent damage to hardware.
Appendix A provides an architecture overview for the OmniSwitch 6600 Family, OmniSwitch 7700/7800,
and the OmniSwitch 8800.
Appendices B and C provide the following for debug and technical support CLI commands:
• Command description.
• Syntax.
• Description of keywords and variables included in the syntax.
• Default values.
• Usage guidelines, which include tips on when and how to use the command.
• Release history, which indicates the release when the command was introduced.
Appendix D provides a list of useful VI editor commands and a sample VI session that modifies the
boot.params file.
What is Not in this Manual?
This guide is intended for troubleshooting switches in live networks. It does not provide step-by-step
instructions on how to set up particular features on the switch or a comprehensive reference to all CLI
commands available in the OmniSwitch. For detailed syntax on non debug CLI commands and comprehensive information on how to configure particular software features in the switch, consult the user
guides, which are listed in “Related Documentation” on page xvii.
How is the Information Organized?
Each chapter in this guide includes troubleshooting guidelines related to a single software feature, such as
server load balancing or link aggregation.
Related Documentation
The following are the titles and descriptions of all the Release 5.1 and later OmniSwitch user guides:
• OmniSwitch 6600 Family Getting Started Guide
Describes the hardware and software procedures for getting an OmniSwitch 6624 or 6648 up and
running. Also provides information on fundamental aspects of OmniSwitch software and stacking
architecture.
• OmniSwitch 6800 Series Getting Started Guide
Describes the hardware and software procedures for getting an OmniSwitch 6800 up and running. Also
provides information on fundamental aspects of OmniSwitch software and stacking architecture.
• OmniSwitch 7700/7800 Getting Started Guide
Describes the hardware and software procedures for getting an OmniSwitch 7700 or 7800 up and
running. Also provides information on fundamental aspects of OmniSwitch software architecture.
• OmniSwitch 8800 Getting Started Guide
Describes the hardware and software procedures for getting an OmniSwitch 8800 up and running. Also
provides information on fundamental aspects of OmniSwitch software architecture.
• OmniSwitch 6600 Family Hardware Users Guide
Complete technical specifications and procedures for all OmniSwitch 6624 and 6648 chassis, power
supplies, fans, uplink modules, and stacking modules.
Complete technical specifications and procedures for all OmniSwitch 6800 chassis, power supplies,
fans, uplink modules, and stacking modules.
• OmniSwitch 7700/7800 Hardware Users Guide
Complete technical specifications and procedures for all OmniSwitch 7700 and 7800 chassis, power
supplies, fans, and Network Interface (NI) modules.
• OmniSwitch 8800 Hardware Users Guide
Complete technical specifications and procedures for all OmniSwitch 8800 chassis, power supplies,
fans, and Network Interface (NI) modules.
• OmniSwitch CLI Reference Guide
Complete reference to all CLI commands supported on the OmniSwitch 6624/6648, 7700/7800, and
8800. Includes syntax definitions, default values, examples, usage guidelines and CLI-to-MIB variable
mappings.
• OmniSwitch 6600 Family Switch Management Guide
Includes procedures for readying an individual switch for integration into a network. Topics include the
software directory architecture, image rollback protections, authenticated switch access, managing
switch files, system configuration, using SNMP, and using web management software (WebView).
• OmniSwitch 6800 Series Switch Management Guide
Includes procedures for readying an individual switch for integration into a network. Topics include the
software directory architecture, image rollback protections, authenticated switch access, managing
switch files, system configuration, using SNMP, and using web management software (WebView).
Includes procedures for readying an individual switch for integration into a network. Topics include the
software directory architecture, image rollback protections, authenticated switch access, managing
switch files, system configuration, using SNMP, and using web management software (WebView).
• OmniSwitch 6600 Family Network Configuration Guide
Includes network configuration procedures and descriptive information on all the major software
features and protocols included in the base software package. Chapters cover Layer 2 information
(Ethernet and VLAN configuration), Layer 3 information (RIP and static routes), security options
(authenticated VLANs), Quality of Service (QoS), and link aggregation.
• OmniSwitch 6800 Series Network Configuration Guide
Includes network configuration procedures and descriptive information on all the major software
features and protocols included in the base software package. Chapters cover Layer 2 information
(Ethernet and VLAN configuration), Layer 3 information (RIP and static routes), security options
(authenticated VLANs), Quality of Service (QoS), and link aggregation.
Includes network configuration procedures and descriptive information on all the major software
features and protocols included in the base software package. Chapters cover Layer 2 information
(Ethernet and VLAN configuration), Layer 3 information (routing protocols, such as RIP and IPX),
security options (authenticated VLANs), Quality of Service (QoS), link aggregation, and server load
balancing.
• OmniSwitch 6600 Family Advanced Routing Configuration Guide
Includes network configuration procedures and descriptive information on the software features
included in the advanced routing software package (OSPF).
• OmniSwitch 6800 Series Advanced Routing Configuration Guide
Includes network configuration procedures and descriptive information on the software features and
protocols included in the advanced routing software package (OSPF, DVMRP, PIM-SM).
Includes network configuration procedures and descriptive information on all the software features and
protocols included in the advanced routing software package. Chapters cover multicast routing
(DVMRP and PIM-SM) and OSPF.
• Technical Tips, Field Notices
Includes information published by Alcatel’s Service and Support group.
• Release Notes
Includes critical Open Problem Reports, feature exceptions, and other important information on the
features supported in the current release and any limitations to their support.
These user guides are included on the Alcatel Enterprise User Manual CD that ships with every switch.
You can also download these guides at http://www.ind.alcatel.com/library/manuals/index.cfm?cnt=index.
Before Calling Alcatel’s Technical Assistance
Center
Before calling Alcatel’s Technical Assistance Center (TAC), make sure that you have read through the
appropriate section (or sections) and have completed the actions suggested for your system’s problem.
Additionally, do the following and document the results so that the Alcatel TAC can better assist you:
• Have a network diagram ready. Make sure that relevant information is listed, such as all IP addresses
and their associated network masks.
• Have any information that you gathered while troubleshooting the issue to this point available to
provide to the TAC engineer.
• If the problem appears to be with only a few-fewer than four-switches, capture the output from the
show tech-support CLI command on these switches. (See Appendix C, “Technical Support
Commands,” for more information on show tech-support CLI commands.)
When calling Alcatel TAC in order to troubleshoot or report a problem following information can be helpful to get a quick resolution:
• All the dump files that were created, if any
• Output of switch log or the switch log files swlog1.log and swlog2.log
• Configuration file boot.cfg
• A capture of the show microcode command
• A capture of the show module long command
• A capture of the show tech-support command from CLI.
• If a CMM fail over to the redundant CMM happened because of this failure then include this informa-
tion from both of the CMMs.
Dial-in or Telnet access can also helpful for effective problem resolution.
In order to troubleshoot the system, a basic understanding of the operation of Chassis Management
Modules (CMMs) and their interaction with Network Interface (NI) modules is required. Some concepts
are covered in this chapter:
• Understanding of the “Diagnosing Switch Problems” chapter in the appropriate OmniSwitch Switch
Management Guide.
• Understanding of the “Using Switch Logging” from the appropriate OmniSwitch Network Configura-
tion Guide is highly recommended.
In This Chapter
“Introduction” on page 1-2
“Troubleshooting System for OS-6624/6648 and OS-7/8XXX” on page 1-3
“Advanced Troubleshooting” on page 1-9
“Dshell Troubleshooting” on page 1-11
“Using AlcatelDebug.cfg” on page 1-26
“Troubleshooting IPC on OS-6/7/8XXX Series of Switches” on page 1-27
The CMM is the Management Module of the switch. All of the critical operations of the switch including
the monitoring is the responsibility of the CMM. CMM not only provides monitoring but also synchronizes all of the NI for different operations. The operation of the CMM is the same in OS-6/7/8XXX
switches. The only difference is that OS-6/7XXX has the switching fabric inherent to the module whereas
OS-8800 has fabric at the back of the chassis.
NI has a build in CPU. Each NI has its own CPU, which acts independently of the CMM. The CPU of the
NI has to interact with the CPU on the CMM for certain operations. If this operation becomes out of sync
then it can create problems specific to that NI.
In order to troubleshoot the system, an understanding of the CMM and NI operation is essential.
Troubleshooting the Switch SystemTroubleshooting System for OS-6624/6648 and OS-7/8XXX
Troubleshooting System for OS-6624/6648 and
OS-7/8XXX
If the switch is having any problems the first place to look for is the CMM. All the task are supervised by
CMM. Any in coherency between CMM and the NI can cause problems to appear.
1 The first step for troubleshooting problems with the switch is to look at the overall general health of the
switch.
OmniSwitch 7700/7800/8800
Verify that all of the modules in the chassis are up and operational, using the command:
-> show module status
Operational Firmware
Slot Status Admin-Status Rev MAC
------+-------------+------------+---------+----------------CMM-A UP POWER ON 36 00:d0:95:6b:09:40
NI-1 UP POWER ON 5 00:d0:95:6b:22:5c
NI-3 UP POWER ON 5 00:d0:95:6b:23:2e
NI-5 UP POWER ON 5 00:d0:95:6b:3a:20
OmniSwitch 6624/6648
If the switch is having any problems the first place to look for is the CMM. All the task are supervised by
CMM. Any in coherency between CMM and the NI can cause problems to appear. For OS-6600 with 8
units stacked together, the CMM will be:
• Primary
• Secondary
• Idle
The switch with the lowest ID will become the primary CMM.
The first step for troubleshooting problems with the switch is to look at the overall general health of the
switch.
Verify that all of the modules in the chassis are up and operational, using the command:
Troubleshooting System for OS-6624/6648 and OS-7/8XXXTroubleshooting the Switch System
NI-2 UP POWER ON N/A 00:d0:95:84:3d:26
NI-3 UP POWER ON N/A 00:d0:95:86:50:f4
NI-4 UP POWER ON N/A 00:d0:95:84:49:be
NI-5 UP POWER ON N/A 00:d0:95:84:39:be
NI-6 UP POWER ON N/A 00:d0:95:84:4a:90
NI-7 UP POWER ON N/A 00:d0:95:84:39:f4
NI-8 UP POWER ON N/A 00:d0:95:84:3c:44
OmniSwitch 6600 with 8 stackable switches show up. Notice that the switch with ID 1 is the primary
CMM and the switch with ID of 2 is the secondary. All the switch also show up as NI because each switch
has a CPU and is also a NI.
To verify the stacking topology, use the following command:
-> show stack topology
Link A Link A Link A Link B Link B Link B
NI Role State RemoteNI RemoteLink State RemoteNI RemoteLink
The above command shows the stacking topology. Switch 1 is the primary connected to Switch 8 on port
51 and Switch 2 on port 52. The state of CPUs for all the switches in the stack is shown by the output of
this command.
2 Verify the power supply (or supplies).
OmniSwitch 7700/7800/8800
Omni Switch 7/8XXX has build-in mechanism to power off the modules if the power supply is not
enough. Switching off a power supply in a chassis which does not have redundant power supply will result
in power off of the modules. Make sure that there is no power involvement.
Check the power supply status, using the command:
-> show power supply 1
Module in slot PS-1
Model Name: OSR-PS-06,
Description: OSR-PS-06,
Part Number: 901750-10,
Receive800000000
Transmit/Receive8000000000
Memory8043434343
Cpu8002060507
Temperature Cmm5038373737
Temperature Cmm Cpu5032323132
The above command shows the receive, transmit/receive, memory, CPU, temperature CMM and temperature CMM CPU statistics for current, 1 minimum average, 1 hour average and 1 hour maximum. All the
values should be within the threshold. Anything above the threshold depicts that some abnormal behavior.
Normally 1 hour average maximum might be high if the switch was booted up in the last hour but it should
be fairly steady during normal operation.
If none of the above are above the threshold then the next step is to try to isolate the problem to a particular NI. Due to the distributed architecture every NI has it own CPU to perform some operations locally. It
is possible that a particular NI might be at high CPU utilization at a time when other NI as well as the CPU
are within the thresholds.
If none of the above are above the threshold then the next step is to try to isolate the problem to a particular NI (or a switch within an OmniSwitch 6624/6648 stack) with the show health slot_number CLI
command:
-> show health 5
* - current value exceeds threshold
Slot 051 Min 1 Hr 1 Hr
ResourcesLimit CurrAvgAvgMax
Troubleshooting the Switch SystemTroubleshooting System for OS-6624/6648 and OS-7/8XXX
The average on one minute is calculated from the average of 12 samples. Each sample is an average of the
CPU utilization during 5 seconds. Those values are stored in a table. The current minute (1 Min Avg or
“min”) displays the average of the last 12 samples.
Every 60 seconds the average of the 12 samples is recorded into the average value for this minute. Those
values are stored in a form of 60 samples which represent one hour.
Most probably one of the above would help to localize the problem to a particular NI or to CMM. For
more details see, Section “Monitoring Switch Health” in the chapter titled “Diagnosing Switch Problems”
in the appropriate OmniSwitch Network Configuration Guide.
4 Check the switch log.
OmniSwitch 6624/6648 and 7700/7800/8800
Now, one of the most important things to check is the switch log. Switch log would contain the error
messages depending on the settings of the log levels and applications configured to generate error
messages. Default settings of the log switch log can be view using the command:
-> show swlog
Switch Logging is:
- INITIALIZED
- RUNNING
Log Device(s)
----------------
flash
console
Only Applications not at the level ‘info’ (6) are shown
Application IDLevel
----------------------------
CHASSIS (64)debug3 (9)
By default, log devices are set to be flash and console. This can be changed and specific log servers can be
used to log the messages, please refer to the Switch Management Guide for further details. The application trace level is set for ‘info’. Any error messages or informational messages would be logged in the
switch log.
Switch log should be viewed to see if any errors messages were generated by the switch. The command to
use is:
-> show log swlog
Displaying file contents for 'swlog2.log'
FILEID: fileName[swlog2.log], endPtr[32]
MON AUG 21 23:09:57 2023HSM-CHASSISinfo== HSM == GBIC extraction detect
ed on NI slot 1, GBIC port 2
MON AUG 21 23:28:33 2023HSM-CHASSISinfo== HSM == GBIC Insertion detecte
d on NI slot 1, GBIC port 1
MON AUG 21 23:28:33 2023HSM-CHASSISinfo== HSM == GBIC Insertion detecte
Troubleshooting System for OS-6624/6648 and OS-7/8XXXTroubleshooting the Switch System
d on NI slot 1, GBIC port 2
MON AUG 21 23:28:39 2023HSM-CHASSISinfo== HSM == GBIC extraction detect
ed on NI slot 5, GBIC port 2
MON AUG 21 23:30:39 2023HSM-CHASSISinfo== HSM == GBIC Insertion detecte
d on NI slot 5, GBIC port 2
MON AUG 21 23:30:41 2023HSM-CHASSISinfo== HSM == GBIC extraction detect
ed on NI slot 1, GBIC port 1
MON AUG 21 23:30:45 2023HSM-CHASSISinfo== HSM == GBIC extraction detect
ed on NI slot 1, GBIC port 2
TUE AUG 22 00:05:45 2023CSM-CHASSISinfo== CSM == !!!ACTIVATING!!!
TUE AUG 22 00:05:45 2023CSM-CHASSISinfo== CSM == !!! REBOOT !!!
TUE AUG 22 00:05:53 2023SYSTEMalarm System rebooted via ssReboot(),
restart type=0x2 (COLD)
The log message are kept in two log files: swlog1.log and swlog2.log in flash. In the above example, log
messages show that some GBICs were extracted and inserted at a particular time. In addition, the switch
was rebooted. This information helps to relate the time of the problem together with the events happening
at the switch. In addition, it also provides an idea about if the source of the problem was external or internal to the switch.
If the log messages do not show enough information then they can be changed for specific applications to
a higher log level or for all the applications running in the switch. For setting up different log levels in
switch log, please refer to the “Using Switch Logging” chapter in the appropriate OmniSwitch Network
Configuration Guide.
If the switch is running in redundant configuration make sure that the two CMMs are completely synchronized. This can be done using the command:
-> show running-directory
Running CMM : PRIMARY,
Running configuration : WORKING,
Certify/Restore Status : CERTIFIED,
Synchronization Status : SYNCHRONIZED
If the two CMMs are not synchronized and the problem leads to the failure of Primary CMM then it will
result in re-initialization of all of the modules. If the two CMMs are properly synchronized and primary
CMM failed, the take over mechanism will be transparent to the end user. So, for complete redundancy
keep the two CMMs synchronized.
Look for any post-mortem dump files that may be created due to the problem with the switch. Post
Mortem Dump files have an extension of *.dmp and are created in /flash directory of the CMM (be sure to
check the secondary CMM, if running in redundant mode). System dump files are normally named as
“cs_system.dmp”, Memory related dump files are normally created as “MemMon000.dmp” and NI related
dump files are named as “SloXSliYver1.dmp”, where X is the slot number and Y is the slice number.
The creation of a dump file indicates a problem with the switch. System related dump files can be viewed
through CLI but other dump files cannot. For system related dump files use the command:
-> show log pmd cs_system.pmd
Capture the output of this command. In addition, if there are any dump files created in the switch, they
should be downloaded through FTP to forward them to technical support. Technical Support can have
them analyzed to find the source of the problem.
Troubleshooting the Switch SystemAdvanced Troubleshooting
Advanced Troubleshooting
One level of switch logging is stored in the two log files located in the /flash directory. There is another
low level debug that can be enabled and used for diagnosing the problems. This debug is known as
“systrace”, meaning system trace. The information in this trace is stored in NVRAM on the CMM, so it is
valid until powered off. Soft reboot of the switch will retain the trace information but powering off the
switch will result in loosing all of the information. This is less CMM intensive so can be used to collect all
the background information about the different tasks running in the switch.
The command to look at the default settings for systrace is
-> debug systrace show
sysTrace is:
- INITIALIZED
- RUNNING
- configured to TRACE CALLERS
- configured to NOT WATCH on stdout
All Applications have their trace level set to level ’info’ (6)
Systrace is set for the level of “info” for all the applications. Any application with trace level other than 6
is displayed in the above command output. Notice that it is initialized by default and is running in the
background. By default it is configured not to display messages on the console. The purpose of systrace is
to track all the system processes called and the caller.
Application log levels can be changed and specific applications can also be set for the logging purposes.
The commands are similar to switch log.
The applications and the log levels are the same as switch log applications. Please refer to the “Section
Switch Logging Commands Overview” section in the “Using Switch Logging” chapter in the appropriate
OmniSwitch Network Configuration Guide.
Systrace can be enabled using the command:
-> debug systrace enable
To look at the systrace log file use the following command:
Advanced TroubleshootingTroubleshooting the Switch System
_SM_IPCUP_TIMEOUT
3349118948 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> CS_TIMEOUT
3345200526 CSM-CH info tCsCSMtask ***HELLO FSM TRACE***
3342928783 CSM-CH info tCsCSMtask ***HELLO FSM TRACE***
3342928661 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> Event = CS_CSM_HELLO
_SM_IPCUP_TIMEOUT
3342928628 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> CS_TIMEOUT
3336738410 CSM-CH info tCsCSMtask ***HELLO FSM TRACE***
3336738287 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> Event = CS_CSM_HELLO
_SM_IPCUP_TIMEOUT
3336738256 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> CS_TIMEOUT
3334849145 CSM-CH info tCsCSMtask ***HELLO FSM TRACE***
3330548020 CSM-CH info tCsCSMtask ***HELLO FSM TRACE***
3330547902 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> Event = CS_CSM_HELLO
_SM_IPCUP_TIMEOUT
3330547869 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> CS_TIMEOUT
3324495309 CSM-CH info tCsCSMtask ***HELLO FSM TRACE***
3324357940 CSM-CH info tCsCSMtask ***HELLO FSM TRACE***
3324357816 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> Event = CS_CSM_HELLO
_SM_IPCUP_TIMEOUT
3324357782 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> CS_TIMEOUT
3318167293 CSM-CH info tCsCSMtask ***HELLO FSM TRACE***
3318167171 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> Event = CS_CSM_HELLO
_SM_IPCUP_TIMEOUT
3318167139 CSM-CH info tCsCSMtask csCsmHelloReceptio - -> CS_TIMEOUT
This information is useful to analyze the different processes taking place in the switch.
Other useful command to use in case of problem is:
-> show tech-support
This command captures all of the information from the chassis, including the hardware information,
configuration, software release active and some other statistics about the number of buffers being used at
the time of the use of command. The output of the command is saved in /flash as “tech_support.log”.
Other variation of this command is:
Troubleshooting the Switch SystemDshell Troubleshooting
Dshell Troubleshooting
To further diagnose the task consuming the CPU on the CMM one needs to use the following Dshell
commands:
Note. Dshell commands should only be used by Alcatel personnel or under the direction of Alcatel.
Misuse or failure to follow procedures that use Dshell commands in this guide correctly can cause lengthy
network down time and/or permanent damage to hardware.
Working: [Kernel]->spyReport
NAME ENTRY TID PRI total % (ticks) delta % (ticks)
In the above example, the OSPF task is suspended. Typically when a task is suspended, the system will
automatically reboot and generate a system dump file. In the event that the system does not reboot, then
try to gather the task trace and memory dump for that specific task using the following command:
To troubleshoot a CPU or memory spike with 5.1.5.X, you can start a software routine in dshell and it will
log the task name to the swlog whenever there is a spike in CPU or memory usage.
Switch/> dshell
Certified: [Kernel]->lkup "Hog"
catchCpuHog 0x00152700 text => to turn on CPU watch
catchMemHog 0x0013fa80 text => to turn on Memory watch
releaseCpuHog 0x00152720 text => to turn off CPU watch
releaseMemHog 0x0013fb60 text => to turn off Memory watch
value = 58685232 = 0x37f7730
Certified: [Kernel]->
To troubleshoot a problem related to stack overflow:
The NI Debugger software can be launched in Dshell using the following command:
Working: [dshell]-> <nidebug
This will launch the NI Debugger. To change to a specific slot and slice (Coronado) the following
command can be used:
changeSlot slot,slice
Now the processor on that slot can be accessed just like CMM to see all tasks (running or suspended),
tasks consuming the CPU the most, and other commands like task trace(tt) or task info(ti).
This will result in generating a PMD file for slot 1 slice 0 in /flash directory, which can then be forwarded
to Engineering for analysis. In addition, there is a software available known as “ni_pmdexploit” which can
be used on UNIX OS to exploit the PMD files in VI format. The OMD files generated on the switch for NI
are in binary format and cannot be viewed by switch log commands on the switch. These files need to be
converted to VI format to be analyzed.
The format to exploit a NI pmd file is “ni_pmdexploit <filename> < <new filename”. Once it is exploited,
it can be viewed using normal UNIX editors.
elements seen by linkNumber of elements seen by the link with the global/local port number
as 1a, in the order they are seen and the role of each stack
NINI number of the switch in the stack.
CMMCMM number of the switch in the stack. CMM number can be 65 (Pri-
mary), 66 (Secondary) or 0 (Idle). Role can be 1 (Primary), 2 (Secondary) or 3 (Idle).
state_linkStatus of link A and B which can be 1 if up or 0 if down.
remote_slotRemote slot number.
remote_linkRemote link number.
Accessing Dshell on Idle Switches
OS6600 in standalone environment is like one NI for OmniSwitch 7000 and 8000 series switches. Just
going into Dshell will allow the use of normal Vx Works commands.
There are two ways to access Dshell. One is using the dshell command from CLI or pressing control-w, control-w (twice). The second method is used when the console or telnet is not accessible. However,
before doing so, it must be enabled by following the steps below on the primary and secondary switches:
Using AlcatelDebug.cfgTroubleshooting the Switch System
In stacking environment, only the primary and secondary switches have console enabled whereas console
is disabled for the idle switches. To enable the Dshell access to the idle switches use the following
command on primary stack:
Nisup_control_WW_onslot
You must execute this command on each idle switch in the stack. Please note that these switches will not
allow to exit with the exit command. To restore normal Dshell access you will need to reboot the switch.
Using AlcatelDebug.cfg
When you are using IPMS/DVMRP with 802.1Q it is recommended that debug interfaces set backpressure enable be used. This command can be put in the boot.cfg file, but it is overwritten as soon as write
memory is issued, since it is a debug command and the setting is lost after a reboot. To retain the debug
settings after a system reboot, put debug commands into a file called AlcatelDebug.cfg in both the working and certified directories. Use Notepad or VI editor to create the AlcatelDebug.cfg file.
Example:
-> vi AlcatelDebug.cfg
-> debug set WWON 1 => to allow dshell access in the event of the console lockup
-> debug set esmDebugLevel 4 => see port up/down event on swlog
-> debug interfaces set backpressure enable => to enable system backpressure
Troubleshooting the Switch SystemTroubleshooting IPC on OS-6/7/8XXX Series of Switches
Troubleshooting IPC on OS-6/7/8XXX Series of
Switches
IPC (Inter Process Communication) is should by the system to communicate between different software
modules. This communication can be between different processes in the same software module or between
two entirely separate modules. This process can be between NI and CMM or between CMM to CMM.
Burst Bus commonly known as BBUS (management bus) is used for the IPC communication. IPC uses
connectionless build-in Vx Works sockets to communicate.
Typical problems that can arise because of the problems with IPC can cause the following symptoms:
• Loss of access to the console of the switch
• Loss of messages between CMM and NI resulting in switching and routing problems.
• High CPU utilization on CMM
Debugging IPC
IPC has 5 different buffer pools:
• Urgent Pools
• Control Pools for control messages
• Normal Pools for some control messages as well as other messages
• Jumbo Pools
• Local Pools
Each of these pools have some dedicated buffers available. Once any of these processes initiates a socket
to communicate, it is suppose to tear the socket down after the communication is done. If it does not tear
the socket then it might result in occupying the buffer space which will not be available for other
processes.
IPC pools can be looked in dshell using the command:
Working: [Kernel]->ipc_pools
UrgentPool: Full size is 1024, remaining: 1024
In socket queues: 0 Not queued: 0:
In DMA queues: 0
ControlPool: Full size is 5096, remaining: 5090
In socket queues: 1 Not queued: 3:
In DMA queues: 2
NormalPool: Full size is 2024, remaining: 2022
In socket queues: 0 Not queued: 2:
In DMA queues: 0
JumboPool: Full size is 256, remaining: 255
In socket queues: 1 Not queued: 0:
In DMA queues: 0
Troubleshooting IPC on OS-6/7/8XXX Series of SwitchesTroubleshooting the Switch System
LocalPool: Full size is 64, remaining: 64
In socket queues: 0 Not queued: 0:
In DMA queues: 0
Each type of pool has the following listed in the command output:
• Maximum size of buffers available
• Currently available buffers
• Socket Queues being used
• Not Queued in pool
• Direct Memory Access Queues
Currently available buffers should always be around the maximum available in normal operation. In some
scenarios, it might happen that the remaining pools are decrementing at a fast rate and are never freeing up
the buffers. This can lead to problem with IPC.
Iterative use of the command will help to identify the situation.
An example is as follows:
Working: [Kernel]->ipc_pools
UrgentPool: Full size is 1024, remaining: 1024
In socket queues: 0 Not queued: 0:
In DMA queues: 0
ControlPool: Full size is 5096, remaining: 5062
In socket queues: 4 Not queued: 20:
In DMA queues: 10
NormalPool: Full size is 2024, remaining: 2022
In socket queues: 0 Not queued: 2:
In DMA queues: 0
JumboPool: Full size is 256, remaining: 255
In socket queues: 1 Not queued: 0:
In DMA queues: 0
LocalPool: Full size is 64, remaining: 64
In socket queues: 0 Not queued: 0:
In DMA queues: 0
Working: [Kernel]->ipc_pools
UrgentPool: Full size is 1024, remaining: 1024
In socket queues: 0 Not queued: 0:
In DMA queues: 0
ControlPool: Full size is 5096, remaining: 5060
In socket queues: 6 Not queued: 20:
In DMA queues: 10
NormalPool: Full size is 2024, remaining: 2022
In socket queues: 0 Not queued: 2:
In DMA queues: 0
JumboPool: Full size is 256, remaining: 255
In socket queues: 1 Not queued: 0:
In DMA queues: 0
Troubleshooting the Switch SystemTroubleshooting IPC on OS-6/7/8XXX Series of Switches
LocalPool: Full size is 64, remaining: 64
In socket queues: 0 Not queued: 0:
DMA queues: 0
In the above two outputs it seems that the control pool is stuck and the socket queues are incrementing. In
order to find out which task is using these queues we need to look at the socket information.
To look in detail about these pools the following commands can be used in Dshell:
• Ipc_urgent_pools_detail number
• Ipc_control_pools_detail number
• Ipc_normal_pools_detail number
• Ipc_jumbo_pools_detail number
• Ipc_local_pools_detail number
The above commands have an option to specify the number of sockets to be displayed in Dshell. If no
number is specified then it will display all the sockets in use which can be real problem in case of thousands of sockets being used.
Troubleshooting IPC on OS-6/7/8XXX Series of SwitchesTroubleshooting the Switch System
value = 10 = 0xa
Working: [Kernel]->
The above command displays a lot of information but we are interested in the most repeating socket ID. In
the above example it is 0x8. To look into what does this socket means the following command can be used
in Dshell:
Troubleshooting IPC on OS-6/7/8XXX Series of SwitchesTroubleshooting the Switch System
LocalPool: Full size is 0, remaining: 1024
In socket queues: 0 Not queued: 0:
In DMA queues: 0
value = 6 = 0x6
The above display of the command does not show the Full size of any of the pools. This indicates that
CMM is unable to view the IPC pools of the NI. In this scenario, one needs to load the NI Debugger and
go to the NI and look at the IPC Pools. One of the NI would be generating many IPC messages that would
result in IPC sockets to be eaten up by that NI resulting in flooding of enormous amount of IPC messages
and in turn loosing communication with the CMM.
The following is an example of using the NiDebug command to display the IPC pools of all NIs.
Multiple task trace of the task with IPC Pools should be taken. This process might have to be repeated on
multiple NI in order to find out the cause of the problem and identify the NI causing the problem to
happen.
OmniSwitch 6624/6648 Example
Follow the steps below for an example of displaying IPC pool data on an OmniSwitch 6624/6648;r
1 Check the In socket queues and Not queued fields for all the pools and identify the pool that has the
highest value with the ipc_pools command as shown below:
Working: [Kernel]->ipc_pools
ipc_pools
UrgentPool: Full size is 1024, remaining: 1024
In socket queues: 0 Not queued: 0:
In DMA queues: 0
ControlPool: Full size is 4096, remaining: 3451
In socket queues: 640 Not queued: 5:
In DMA queues: 0
NormalPool: Full size is 1024, remaining: 377
In socket queues: 620 Not queued: 16:
In DMA queues: 0
JumboPool: Full size is 256, remaining: 256
In socket queues: 0 Not queued: 0:
In DMA queues: 0
LocalPool: Full size is 1024, remaining: 1024
In socket queues: 0 Not queued: 0:
In DMA queues: 0
value = 1 = 0x1
2 Find the most repeated socket ID ipc_normal_pools_detail command as shown below:
Working: [Kernel]->ipc_pools_detail 1,0
NormalPool: Full size is 1024, remaining: 377
Socket ID = 0x7, dest slot = 1, remote addr = 0x60001, ipc status = G
Task ID = 0x756ba38, PayLoad Len= 20, ipc priority = 0x1, data ptr = 0x6cfcba
Port Numbering Conversion OverviewTroubleshooting the Switch System
Port Numbering Conversion Overview
The sections below document how to convert port number parameters.
Note. Dshell commands should only be used by Alcatel personnel or under the direction of Alcatel.
Misuse or failure to follow procedures that use Dshell commands in this guide correctly can cause lengthy
network down time and/or permanent damage to hardware.
ifindex to gport
To convert from ifindex to global port (gport) number use the findGlobalPortFromIfIndex Dshell
command as shown below:
-> dshell
Working: [Kernel]->findGlobalPortFromIfIndex 16011
value = 505 = 0x1f9
gport to ifindex
To convert from global port (gport) to ifindex use the findIfIndexFromGlobalPort Dshell command as
shown below:
-> dshell
Working: [Kernel]->findIfIndexFromGlobalPort 505
value = 16011 = 0x3e8b
Converting from lport
The lport numbering process varies on each platform type (e.g., Falcon/Eagle or Hawk), as well as
module type (e.g., ENI-C24, GNI-C2, GNI-U12, GNI-U8, GNI-C24, GNI-U24, etc.). To determine the
lport value use two Dshell commands: dmpValidPorts and dmpAbsPort.
The following subsections describe conversions based on platform type. You need to be careful that both
commands can be used on either Dshell or Nidebug based on platform type. In addition, input values for
dmpAbsPort vary depending on platform type.
OmniSwitch 7700/7800/8800 (Falcon/Eagle) Example
The following displays all valid lport values with the dmpValidPorts command from NiDebug. Afterwards, you should do a dump for each slice.
Find all valid lports values with the dmpValidPorts command from Dshell on each element (i.e., each
slot in a stack). Afterwards, you should do a dump for each slot.
This chapter assumes that it has been verified that the connectivity problem is across Ethernet media and
the connection between the non-communicating devices is switched/bridged not routed (i.e., Devices are
in the same IP Subnet).
For configuration assistance in designing and configuring switched Ethernet connectivity, please refer to
the “Configuring Ethernet Ports” chapter in the appropriate OmniSwitch Network Configuration Guide.
For known specifications and limitations, Please refer to the appropriate Release Notes Revision.
In This Chapter
“Overview of Troubleshooting Approach” on page 2-2
“Verify Physical Layer Connectivity” on page 2-3
“Verify Current Running Configuration” on page 2-5
Verify that there is valid link light along the entire data path between the devices that can not switch to
each other. Make sure to include all interswitch links. Verify LED’s on all involved CMMs and NIs are
Solid OK1, Blinking OK2. If this is not the case, contact technical support.
Use the show interfaces command to verify operational status is Up, speed and duplex are correct and
match the other side of the connection. Run this command on the same interface multiple times to verify
errors (Error Frames, CRC Error Frames, Alignment Errors) are not incrementing. If the error counts are
incrementing verify the health of the cabling as well as the NIC involved. Also note that if the Collision
Frames is incrementing, this is normal for a half duplex connection. If the port is set to full duplex and
these errors are still incrementing, verify the duplex setting on the other side of the connection. Finally, if
these commands were run while the end stations were trying to ping each other, verify Bytes Received is
incrementing. If is not, verify the NIC card.
Note. Remember to do this for each port along the data path, not just the ports that directly attached to the
end stations.
-> show interfaces 5/1
Slot/Port 5/1 :
Operational Status : up,
Type : Fast Ethernet,
MAC address : 00:d0:95:7a:63:87,
BandWidth (Megabits) : 100, Duplex : Full,
Long Accept : Enable, Runt Accept : Disable,
Long Frame Size(Bytes) : 1553, Runt Size(Bytes) : 64
If the port reports operational status down, verify the physical link, but also verify the necessary NIs and
CMM are receiving power and are up and operational. Use the show ni command followed by the slot
number and the show cmm command to verify this.
Troubleshooting Switched Ethernet ConnectivityVerify Current Running Configuration
Verify Current Running Configuration
If the physical layer looks OK, then verify the configuration. Use the show configuration snapshot all to
display the current running configuration. Use this command to verify the ports that are involved are in the
correct VLAN. Also review the output of the command to verify there is nothing explicit in the configuration that would cause the problem, such as a deny ACL that could be found under the QoS subsection.
-> show configuration snapshot all
! Chassis :
system name OS7800
! Configuration:
! VLAN :
vlan 7 enable name "VLAN 7"
vlan 7 port default 5/1
vlan 7 port default 5/2
! 802.1Q :
! Spanning tree :
! Bridging :
! IPMS :
! AAA :
aaa authentication console "local"
! QOS :
qos apply
! Policy manager :
! Session manager :
! SNMP :
! IP route manager :
ip router router-id 127.0.0.1
ip router primary-address 127.0.0.1
! RIP :
! OSPF :
! BGP :
! IP multicast :
! Health monitor :
! Interface :
! Link Aggregate :
! Port mirroring :
! UDP Relay :
! Server load balance :
! System service :
! VRRP :
! Web :
! AMAP :
! GMAP :
! Module :
To further verify the ports are in the correct VLAN and that they are in spanning tree forwarding instead of
blocking use the show vlan port command. Also note that the port type must match what it is connecting
to. If the port is 802.1Q tagged enabled for the required vlan, then the device it attaches to must also be Q
tagged enabled for that vlan. Remember to run this command on all ports in the data path.
If ports that should be in forwarding are in blocking, or vice versa, please consult Chapter 4, “Trouble-
shooting Spanning Tree.”
Verify Source Learning
If the configuration looks correct, source learning should be examined. If connectivity exists but is slow,
or intermittent source learning could be the root cause, since data packets would be flooded. However, if
there is no packet throughput between the devices the problem is likely not due to a source learning problem.
To verify that the MAC addresses are being learned correctly use the show mac-address-table slot
command. Verify that the correct mac address is being learned of the correct port, in the correct vlan.
-> show mac-address-table slot 5
Legend: Mac Address: * = address not valid
Vlan Mac Address Type Protocol Operation Interface
Troubleshooting Switched Ethernet ConnectivityVerify Switch Health
Verify Switch Health
If source learning appears to be not working correctly, verify the health of the switch with the show health,
and show health slot commands. Be sure to run the latter command on all necessary NIs. Any variables
that have reached or exceeded their limit value could cause forwarding problems on the switch. In this
case please contact Technical Support. For more detailed source learning trouble shooting, please see
Chapter 3, “Troubleshooting Source Learning.”
-> show health
* - current value exceeds threshold
Device 1 Min 1 Hr 1 Hr
Resources Limit Curr Avg Avg Max
-----------------+-------+------+------+-----+---Receive 80 00 00 00 00
Transmit/Receive 80 00 00 00 00
Memory 80 39 39 39 39
Cpu 80 02 02 02 03
Temperature Cmm 50 39 39 39 39
Temperature Cmm Cpu 50 31 31 31 31
-> show health 5
* - current value exceeds threshold
Slot 05 1 Min 1 Hr 1 Hr
Resources Limit Curr Avg Avg Max
If everything checked appears to be valid, verify that this is not an ARP problem. On the end stations
involved, enter a static mac address for the device it is trying to communicate with. If connectivity is
restored, please see Chapter 11, “Troubleshooting ARP.”
Using the Log FileTroubleshooting Switched Ethernet Connectivity
Using the Log File
If none of the above suggest a reason as to why Ethernet switching is not properly working, look into the
log file and see if there are any messages that may suggest why switching is not working properly. Use the
show log swlog command to view the system log file. Look for evidence of a system or interface problem
around the time the problem began.
-> show log swlog
Displaying file contents for ’swlog2.log’
FILEID: fileName[swlog2.log], endPtr[32]
configSize[64000], currentSize[64000], mode[2]
Displaying file contents for ’swlog1.log’
FILEID: fileName[swlog1.log], endPtr[48903]
configSize[64000], currentSize[64000], mode[1]
Time Stamp Application Level Log Message
------------------------+--------------+-------+-------------------------------THU DEC 12 08:13:51 2002 SYSTEM info Switch Logging device ’swlog1.lt
THU DEC 12 08:13:53 2002 SYSTEM info Switch Logging device ’swlog2.lt
THU DEC 12 08:13:56 2002 SYSTEM info Switch Logging device ’/dev/cont
THU DEC 12 08:13:56 2002 CSM-CHASSIS info == CSM == start up
THU DEC 12 08:13:56 2002 CSM-CHASSIS info == CSM == Activating a new vers
THU DEC 12 08:13:56 2002 CSM-CHASSIS info == CSM == The working version i
THU DEC 12 08:13:56 2002 CSM-CHASSIS info == CSM == MONITORING ON
THU DEC 12 08:13:56 2002 CSM-CHASSIS info == CSM == This CMM is primary
After following the troubleshooting steps via CLI for physical connection, configuration validation,
system health and source learning, here are the additional commands in dshell to troubleshoot problems
related connectivity problem:
Checking the 7700/7800 Nantucket Fabric
nanlistB04
Certified: [Kernel]->nanListB04
No SOP Interrupt: 0
Multicast FIFO Full Interrupt: 0
Multicast Buffer Full Interrupt: 0
Unicast Buffer Full Interrupt: 0
Multicast Dump Interrupt: 0
Unicast Dump Interrupt: 0
Unicast Attempt Count: 8a620
Multicast Attempt Count: acecf
Unicast In Count: 8a627
Multicast In Count: acecf
Unicast Out Count: 8a634
Multicast Out Count: 3600e
Dummy Count: 61578
Total FLength Count: 0
value = 0 = 0x0
Certified: [Kernel]->
The total Flengtlh Count value should be 0 or a small value, a large value indicating that there are frames
being back up in the fabric queue.
Troubleshooting Switched Ethernet ConnectivityUsing the Log File
Checking the 7700/7800 Nantucket Fabric for Interrupts, Data
Counts and Error Counts
Working: [Kernel]->nanListB02
HB Out of Sync Interrupts: 0
Error Count Exceeded Interrupts: 0
Framing Error Interrupts: 0
Parity Error Interrupts: 0
B02 Data Port 0 Frame Count = 690dbd37
B02 Data Port 1 Frame Count = 0
B02 Data Port 2 Frame Count = 542e70d9
B02 Data Port 3 Frame Count = 0
B02 Data Port 4 Frame Count = 0
B02 Data Port 5 Frame Count = 0
B02 Data Port 6 Frame Count = 0
B02 Data Port 7 Frame Count = 9e75d47
B02 Data Port 8 Frame Count = 690dbd39
B02 Data Port 9 Frame Count = 0
B02 Data Port 10 Frame Count = 542e70d9
B02 Data Port 11 Frame Count = 0
B02 Data Port 12 Frame Count = 0
B02 Data Port 13 Frame Count = 0
B02 Data Port 14 Frame Count = 0
B02 Data Port 15 Frame Count = 9e75d47
Checking the Traffic Queue on the NI
Working: [Kernel]->FindBuffer 3,0 => where 3 is the slot number
Queue = 0x62 length = 0x40, Address 0x6881880
Queue = 0x63 length = 0x40, Address 0x68818c0
value = 3 = 0x3
The above capture shows one of the queues is backed up on the NI. Check if the queue is sending traffic
using the following command syntax:
In order to troubleshoot Source Learning problems, a basic understanding of the process is required.
A review of the “Managing Source Learning” chapter from the appropriate OmniSwitch Network Configu-ration Guide is required. The following RFC and IEEE standards are supported:
RFCs supported2674 - Definitions of Managed Objects for Bridges
with Traffic Classes, Multicast Filtering and
Virtual LAN Extensions
IEEE Standards supported802.1Q - Virtual Bridged Local Area Networks
802.1D - Media Access Control Bridges
In This Chapter
“Introduction” on page 3-2
“Troubleshooting a Source Learning Problem” on page 3-3
When a packet first arrives on NI source learning examines the packet and tries to classify the packet to
join its correct VLAN. If a port is statically defined in a VLAN, the MAC address is classified in the
default VLAN. Otherwise, if Group Mobility is being used the MAC address is classified into the correct
VLAN based on the rules defined.
As soon as the MAC address is classified in a VLAN, an entry is made in Source Address Pseudo-CAM
associating the MAC address with the VLAN ID and the Source Port. This Source Address is then relayed
to the CMM for management purposes.
If an entry already exists in MAC address database with the same VLAN ID and the same source port
number then no new entry is made. If VLAN ID or the source port is different from the existing entry in
MAC address database then the previous entry is aged out and a new entry is made in the MAC address
database. This process of adding a MAC address in the MAC address database is known as Source Learning.
A MAC address can be denied to learn on a port based on different policies configured through QOS or
Learned Port Security. A MAC address may be learned in a wrong VLAN based on the policies defined
for the port.
Note: This document does not discuss the basic operation of Source Learning. To learn about how Source
Learning works, refer to the “Managing Source Learning” in the appropriate OmniSwitch Network Config-uration Guide.
Troubleshooting Source LearningTroubleshooting a Source Learning Problem
Troubleshooting a Source Learning Problem
In order to troubleshoot a source learning problem the first step is to verify that the physical link is up and
the port has correctly auto-negotiated with the end-station.
The next thing is to verify that the port is a member of the right VLAN, if a port is statically configured
for a VLAN, or the Group Mobility policies are correctly defined. The workstation configuration should
also be verified.
The first thing to look for is the MAC address table to verify that the MAC address is being learned:
-> show mac-address-table
Vlan Mac Address Type Protocol Operation Interface
This does show that the MAC address 00:c0:4f:12:f7:1b is learned on port 8/23, see the figure on page 3-2.
So, the source learning process for this workstation has been completed successfully.
Now, a single MAC address can be a member of multiple VLANs based on different protocols. To verify
that the MAC address has been learned in all of the VLANs, the above command can be used. The protocol field will be different based on different protocols being used and classified into different VLANs.
MAC addresses can also be viewed based on VLAN ID, using the following command:
->show mac-address-table 114
Legend: Mac Address: * = address not valid
Vlan Mac Address Type Protocol Operation Interface
The above command shows the two workstations learned in VLAN 114 on NI 8 and 16.
Whether it be a Layer 3 packet or layer 2, the first step is to have the source MAC address learned in the
MAC address table. Layer 3 involves resolution of ARP, for more details on ARP see troubleshooting
section of ARP, and then the available routes to the destination which involves routing, for more details on
Routing see troubleshooting section of Routing.
By default the MAC address aging time is set to 300 seconds. This can be viewed:
->show mac-address-table aging-time
Mac Address Aging Time (seconds) for Vlan 1 = 300
Mac Address Aging Time (seconds) for Vlan 114 = 300
This can be changed using the command:
->mac-address-table aging-time 500
Mac Address Aging Time (seconds) for Vlan 1 = 500
Mac Address Aging Time (seconds) for Vlan 114 = 500
This can also be changed on a particular VLAN:
->mac-address-table aging-time 600 vlan 114
It may be required to change the aging timer to a higher value to prevent the aging time of silent devices.
Another method by which silent devices can be accommodated is to use the permanent/static MAC
address assigned to a port using the command:
Once, the MAC addresses are learned on the ports then the devices should be able to communicate
depending on the upper layers. Variations of MAC-related commands can be viewed in the “Managing
Source Learning” chapter from the appropriate OmniSwitch Network Configuration Guide.
Advanced Troubleshooting
The advanced troubleshooting for Source learning related problems is to look whether the traffic is
coming in from a port and the NI is not learning the MAC, if not prevented by using any other rules.
->debug ip packet board ni 8 start
R 8/23 00c04f12f71b->00d0957962c4 IP 10.40.114.50->10.40.114.2 ICMP 8,0
seq=58460.
8 S 8/23 00d0957962c4->00c04f12f71b IP 10.40.114.2->10.40.114.50 ICMP 0,0
seq=58460.
ebug ip 8 R 8/23 00c04f12f71b->00d0957962c4 IP 10.40.114.50->10.40.114.2 ICMP
8,0 seq=58716.
8 S 8/23 00d0957962c4->00c04f12f71b IP 10.40.114.2->10.40.114.50 ICMP 0,0
seq=58716.
packet 8 R 8/23 00c04f12f71b->00d0957962c4 IP 10.40.114.50->10.40.114.2 ICMP 8,0
seq=58972.
8 S 8/23 00d0957962c4->00c04f12f71b IP 10.40.114.2->10.40.114.50 ICMP 0,0
seq=58972.
stop8 R 8/23 00c04f12f71b->00d0957962c4 IP 10.40.114.50->10.40.114.2 ICMP 8,0
seq=59228.
8 S 8/23 00d0957962c4->00c04f12f71b IP 10.40.114.2->10.40.114.50 ICMP 0,0
seq=59228.
->debug ip packet stop
This command shows that the packets are coming into the switch and a reply is being sent by the switch to
the end station.
Various combinations of debug ip packet command can be used to find out the incoming traffic. The
combinations possible are as follows:
debug ip packet [start] [timeout seconds] [stop] [direction {in | out | all}] [format {header | text | all}]
[output {screen | switchlog}] [board {cmm | ni [1-16] | all | none} [ether-type {arp | ip | hex [hex] |
all}] [ip-address ip_address] [ip-pair [ip1] [ip2]] [protocol {tcp | udp | icmp | igmp | num [integer] |
all}] [show-broadcast {on | off}] show-multicast {on | off}]
start Starts an IP packet debug session.
timeout Sets the duration of the debug session, in seconds. To specify a dura-
tion for the debug session, enter timeout, then enter the session length.
secondsThe debug session length, in seconds.
stop Stops IP packet debug session.
direction in Debugs incoming packets
direction outDebugs outgoing packets.
direction allDebugs both incoming and outgoing packets.
output switchlogOutput will be saved to a log file.
board cmmDebugs CMM packets.
board niDebugs packets for a Network Interface (NI). To debug a specific inter-
face, enter ni, then enter the slot number of the NI.
board allDebugs packets for all CMMs and NIs on the switch
board noneClears the previous board settings.
If the problems are associated with the source learning on a specific NI then the limitations of the Number
of MAC addresses learned should also be considered. Current limitations are:
Number of learned MAC
32K
addresses per network interface
(NI) module
Number of learned MAC
64K
addresses per switch
The total number of MAC addresses learned per switch can be viewed using the command:
-> show mac-address-table count
Mac Address Table Count:
Permanent Address Count = 0,
DeleteOnReset Address Count = 0,
DeleteOnTimeout Address Count = 0,
Dynamic Learned Address Count = 36,
Total MAC Address In Use = 36
If the problem is still not resolved then kindly contact Tech Support for further troubleshooting.
The OmniSwitch 6/7/8XXX has a distributed architecture. Source Learning is specific to a NI. Each NI
has a layer 2 pseudo-cam which is which can hold 64K entries. 32K entries are reserved for L2 Source
Addresses which are local to that NI in L2SA table and the rest of 32K entries are reserved for L2 Destination Addresses which can be from local or remote NI in L2DA table.
Note. Dshell commands should only be used by Alcatel personnel or under the direction of Alcatel.
Misuse or failure to follow procedures that use Dshell commands in this guide correctly can cause lengthy
network down time and/or permanent damage to hardware.
If a problem is specific to a NI and the MAC address is not being learned by the switch, then the first step
is to verify from the pseudo-cam of that NI that the MAC address has been learned. There can be a possibility that the NI has learned the MAC but CMM is not reporting that MAC because of IPC messages lost
between the CMM and NI.
The commands available to troubleshoot this problem are:
slcDumpL2SA: Display all the SA PseudoCAM entries on one slot/slice.
• Format: slcDumpL2SA slot_num, slice_num
slcDumpL2DA: Display all the Destination Address (DA) PseudoCAM entries on one slot/slice.
• Format: slcDumpL2DA slot_num, slice_num
slcLkupL2SA: Display the SA PCAM entries with MAC, VLAN) tuple on a slot/slice, the high 4 bytes of
MAC are MacHi, other 2 bytes are macLo, VLAN non-significant value is 0.
slcLkupL2DA: Display the DA PCAM entries with (MAC, VLAN} tuple on a slot/slice, the high 4 bytes
of MAC are MacHi, other 2 bytes are macLo, VLAN non-significant value is 0.
Now, if device A connected on slot 8 is unable to communicate to device B in slot 16 then the following
steps can be taken to verify configuration on the NI
First look at the source MAC on slot 8 using the command:
Both of the MAC addresses are learned in the correct VLANs on the right NI.
Now, if device A is trying to communicate to device B then the next thing to look for is the destination
MAC address table. This is to verify that the destination MAC address table has the information about the
device B.
So, the two devices should be able to communicate.
The L2SA and L2DA tables will be different for each slot. L2SA table will be based on the MAC address
learned on that slot. This will not be synchronized to all the other modules. Only the CMM will know
about it. When the request comes in from device A for device B, first a lookup is done on the local L2SA
and L2DA tables to see if there is a matching entry. If there is no matching entry then a request is sent on
the BBUS to all the other Coronados, if any Coronado has the matching entry in its L2SA table it responds
back with the Global port number of that entry. L2DA table is updated on the originating Coronado and
the packet is forwarded to the Global port to reach the destination.
If no other Coronado responds back to the request then the packet is sent over the flood queue to all the
other Coronado to be flooded out of the ports in the same VLAN. If a device responds back on the flooded
request, L2SA for that NI is updated and the Global port number is send to the originating device using the
same lookup as the response will be a unicast packet.
To see Source learning in action on an NI, set the debug level higher (levels are 1-6):
-> Sl_NiDebug=4
To see Source Learning in action on a CMM, set the debug level higher (levels are 1-6):
-> Sl_CmmDebug=5
To view the messages on the console, disable systrace:
To look at the forwarding database on OS-6600 in Dshell use the slcDumpSlotSlice command., which
displays which slot/slice is considered to be up and operational by the source learning software:
Certified: [Kernel]->slcDumpSlotSlice
Source Learning Slice Up List:
slot/slice 2/0, type = 838930434, firstgport = 64, lastgport = 123
value = 68 = 0x44 = ’D’
To look at the forwarding database on OS-6600 in Dshell use the dumpL2 command:
In order to troubleshoot spanning tree related problems an understanding of the protocol and its features
are needed. The OmniSwitch supports two Spanning Tree Algorithms; 802.1D (standard) and 802.1w
(rapid reconfiguration). In addition, the Omniswitch supports two Spanning Tree operating modes: flat
(single STP instance per switch) and 1x1 (single STP instance per VLAN).
Spanning Tree Protocol is defined in the IEEE 802.1D standard.
The 802.1w amendment to that standard, Rapid Reconfiguration of Spanning Tree, improves upon STP by
providing rapid reconfiguration capability via Rapid Spanning Tree Protocol
For configuration assistance please read the “Configuring Spanning Tree Parameters” in the appropriate
OmniSwitch Network Configuration Guide.
In This Chapter
“Introduction” on page 4-1
“Troubleshooting Spanning Tree” on page 4-2
“Dshell” on page 4-5
“Generic Troubleshooting in Dshell” on page 4-10
“CMM Spanning Tree Traces” on page 4-25
Introduction
The primary purpose for spanning tree is to allow for physical redundancy in a bridged network, while
assuring the absence of data loops. The protocol allows for dynamic fail-over as well.
One of the most important tools needed in troubleshooting a STP problem, is to be prepared before it
happens. It is essential to have a network diagram that depicts both the physical (cables) and logical
(VLANs) configurations. It also very useful to know which ports are normally in blocking/forwarding
prior to any problem.
Troubleshooting Spanning TreeTroubleshooting Spanning Tree
Troubleshooting Spanning Tree
A failure of the Spanning Tree Protocol (STP) will usually cause either a bridge loop on the LAN or
constant reconvergence of STP. This in turn can cause several resultant problems.
• If there is a bridge loop on the LAN, there can appear to be a broadcast storm since broadcast packets
will continuously loop the network. In addition, unicast traffic can be affected as the port a unicast
address is learned off of, can toggle from one port to another in a very short time period.
• If STP is constantly reconverging, this can cause temporary network outages as ports could through the
30 seconds of listening and learning as defined by 802.1D. One can see if STP is constantly reconverging that the LAN could be perpetually down.
In determining the cause of the STP problem, its useful to first verify the configuration, especially if the
network having problems has recently been installed.
Use the show spantree command to verify that STP is enabled and that both sides of the link are running
the same STP protocol.
-> show spantree
Vlan STP Status Protocol Priority
-----+----------+--------+--------
1 ON 802.1D 32768
10ON802.1D32768
Use the show spantree command and specify a VLAN to verify the correct mode, designated root ID, root
port, and configurable timers. The timers need to be consistent across a physical link running STP. Also
very useful to note in this command are Topology changes and Topology age. If topology changes are
incrementing quickly, the LAN can not agree who is root. This can be caused by dropped BPDUs (which
will be discussed later), a bridge that insists it is root regardless of received BPDUs, or a physical link
going in and out of service.
-> show spantree 10
Spanning Tree Parameters for Vlan 10
Spanning Tree Status : ON,
Protocol : IEEE 802.1D,
mode : 1X1 (1 STP per Vlan),
Priority : 32768 (0x8000),
Bridge ID : 8000-00:d0:95:79:62:8a,
Designated Root : 8000-00:d0:95:79:62:8a,
Cost to Root Bridge : 0,
Root Port : None,
Next Best Root Cost : 0,
Next Best Root Port : None,
Hold Time : 1,
Topology Changes : 0,
Topology age : 0:0:0
Current Parameters (seconds)
Max Age = 20,
Forward Delay = 15,
Hello Time = 2
Parameters system uses when attempting to become root
System Max Age = 20,
System Forward Delay = 15,
System Hello Time = 2
Troubleshooting Spanning TreeTroubleshooting Spanning Tree
Use the show spantree ports command to determine if the port is in forwarding or blocking and are in the
correct VLAN. Remember that in any LAN with physical redundancy there must be at least one port in
blocking status. If it is known which ports are usually in blocking, those ports can be a good place to start
to verify they are still in blocking status.
-> show spantree ports
Vlan Port Oper Status Path Cost Role
-----+-----+------------+---------+-----
10 5/10 FORW 100 DESG
If ports that should be in blocking are now in forwarding, there are two likely causes. The first is that there
was a physical failure in a link that was previously in forwarding. The second is that the BPDUs from the
root are being dropped. If it appears that BPDUs are being dropped, troubleshoot this as if it were any
other packet being dropped.
Use the show interfaces command to look for errors incrementing on the port as well as to verify duplex
settings match on either side of the link.
-> show interfaces 5/10
Slot/Port 5/10 :
Operational Status : up,
Type : Fast Ethernet,
MAC address : 00:d0:95:7a:63:90,
BandWidth (Megabits) : 10, Duplex : Half,
Long Accept : Enable, Runt Accept : Disable,
Long Frame Size(Bytes) : 1553, Runt Size(Bytes) : 64
Troubleshooting Spanning TreeTroubleshooting Spanning Tree
Since STP is run in a distributed fashion it is important to verify that each NI that is involved is not having
a resource problem. Use the show health command to verify the resources available on an NI.
-> show health 5
* - current value exceeds threshold
Slot 05 1 Min 1 Hr 1 Hr
Resources Limit Curr Avg Avg Max
If the problem has been ascertained to be layer 2 data loop, and it is needed to restore network connectivity quickly, it is recommended to disable all redundant links either administratively or by disconnecting
cables.
As mentioned previously, it is important to verify the health of the NI as well as the CMM. Please refer to
Chapter 1, “Troubleshooting the Switch System,” for directions.
Note. Dshell commands should only be used by Alcatel personnel or under the direction of Alcatel.
Misuse or failure to follow procedures that use Dshell commands in this guide correctly can cause lengthy
network down time and/or permanent damage to hardware.
The commands run above to verify STP configuration on a particular port give the CMM perspective.
Since STP is run on the NI it is important to query the NI to verify what was seen from the CMM. To
verify a ports forwarding status use the esmDumpCoronado slot,slice, 0x6608000+vlan_id*4,32
command. This will indicate if the port as the NI sees it is in forwarding/blocking. The 32 in the above
command shows 32 register values starting from the vlan_id specified. If the vlan_id used is 1 then the
above command will display the values from VLAN 1 to VLAN 31. The bits are dedicated to the ports in
the following order, starting from least significant bit. The bits are set (value=1) to indicate that the ports
are forwarding for that VLAN. If 0 then the port is blocking for that VLAN.
Please note that the examples in this section have the following assumptions:
• Ports 1-12: First 12 Ethernet ports.
• Port 13: First Gigabit port.
• Ports 14,15,16: Not used.
• Ports 17-28: Second half of 12 Ethernet ports.
• Port 29: Second Gigabit port.
• Port 1/1 is a member of VLANs 1,140,141,150, and 511.
For VLAN 1 the bits set are 203 which are equivalent to binary 0000 0000 0011. Bits 1 and 2 are set indicating that ports 1 and 2 have the spanning tree vector set for VLAN 1. The next register value is for
VLAN 2, hex value is 8000000.
Binary: 1000 0000 0000 0000 0000 0000 0000
Binary value indicates that bit 28 is set which means that port 24 is set for VLAN 2. The next register
value will indicate the value for VLAN 3. Hex value is c00.
Binary: 1100 0000 0000
Bits 11 and 12 are set indicating that spanning tree has been set for ports 11 and 12. These ports are
forwarding.
Each NI when boots up sends a message to every other NI indicating that it is up and running. This
message is critical for setting up the port Queues to transfer data as well as for Spanning tree. If an IPC
message is lost by a particular NI then other NI will not see that NI as being a part of spanning tree
domain. This may result in split spanning tree leading to a layer 2 loop. This kind of scenario might
happen in the case of hot swaps.
To verify that each NI known about every other NI the following command should be used in NI Debugger, This should be run on all NIs that are used in STP.
Working: [Kernel]->NiDebug
1:0 nidbg> stpNISock_boardupprint
1:0
1:0 STP boards up :
1:0 board in slot : 2 slice : 0 is up
1:0 board in slot : 4 slice : 0 is up
1:0 board in slot : 5 slice : 0 is up
1:0 board in slot : 6 slice : 0 is up
1:0 board in slot : 7 slice : 0 is up
1:0 board in slot : 8 slice : 0 is up
1:0 board in slot : 9 slice : 0 is up
1:0 board in slot : 10 slice : 0 is up
1:0 board in slot : 11 slice : 0 is up
1:0 board in slot : 12 slice : 0 is up
1:0 board in slot : 13 slice : 0 is up
1:0 board in slot : 14 slice : 0 is up
1:0 board in slot : 16 slice : 0 is up
1:0 value = 0 = 0x0
This command will show all the other slots except for itself.
To look at all the BPDUs being received and transmitted on a particular slot and slice the following
command can be used in NiDebug command. This will display, BPDUs as well as notifications when
there is a topology change in real time.
1:0 nidbg> stp_printf_flag=1
1:0 *** stpkern_bpduIn stp_id=511 portid=c type=2
1:0 PIM port c state 4 1024 0
1:0 Message age of received BPDU : 0
1:0 PIM port c state 5 1024 0
1:0 recordProposed operPointToPointMAC=1
1:0 PIM port c state 7 1536 0
1:0 PIM port c state 4 1536 0
1:0 port 12 is forward (5)
1:0 tick (tack) time is now 701603
1:0
1:0 RSTBPDU transmitted on port 33 on STP 57
1:0 Root bridge ID = 3200d0 95820514
1:0 Path to Root cost = 3
1:0 Designated bridge ID = 800000d0 957962aa
1:0 Designated portId = 29697
1:0 Bridge portId = 29697
1:0 Message age : 256
1:0 Proposing
1:0
1:0 RSTBPDU transmitted on port 33 on STP 51
1:0 Root bridge ID = 3200d0 95820514
1:0 Path to Root cost = 3
1:0 Designated bridge ID = 800000d0 957962aa
1:0 Designated portId = 29697
1:0 Bridge portId = 29697
1:0 Message age : 256
1:0 Proposing
1:0 tick (tack) time is now 701628
1:0 tick (tack) time is now 701634
1:0 tick (tack) time is now 701635
1:0
1:0 RSTBPDU transmitted on port 33 on STP 60
1:0 Root bridge ID = 3200d0 95820514
1:0 Path to Root cost = 3
1:0 Designated bridge ID = 800000d0 957962aa
1:0 Designated portId = 29697
1:0 Bridge portId = 29697
1:0 Message age : 256
1:0 Proposing
1:0 tick (tack) time is now 701636
1:0
1:0 RSTBPDU transmitted on port 12 on STP 140
1:0 Root bridge ID = c800d0 957962aa
1:0 Path to Root cost = 0
1:0 Designated bridge ID = c800d0 957962aa
1:0 Designated portId = 29196
1:0 Bridge portId = 29196
1:0 Message age : 0
1:0 tick (tack) time is now 701637
1:0
1:0 RSTBPDU transmitted on port 33 on STP 52
1:0 Root bridge ID = 3200d0 95820514
1:0 Path to Root cost = 3
1:0 Designated bridge ID = 800000d0 957962aa
1:0 Designated portId = 29697
1:0 Bridge portId = 29697
1:0 Message age : 256
1:0 Proposing
1:0 RSTBPDU transmitted on port 33 on STP 61
1:0 Root bridge ID = 3200d0 95820514
1:0 Path to Root cost = 3
Generic Troubleshooting in DshellTroubleshooting Spanning Tree
Generic Troubleshooting in Dshell
Note. Dshell commands should only be used by Alcatel personnel or under the direction of Alcatel.
Misuse or failure to follow procedures that use Dshell commands in this guide correctly can cause lengthy
network down time and/or permanent damage to hardware.
The stp_help command (executed from the NiDebug Dshell command prompt) displays the trace menu
for the Spanning Tree algorithm on NIs. Enter stpNI_help at ???? at what? Text missing here. ????
-> dshell
Working: [Kernel]->NiDebug
NiDebug>>stp_help
stpNISock_globals : Global variables
stpNISock_warningprint : warning trace
stpNISock_totraceprint : time-out trace
stpNISock_traceprint : event trace
stpNISock_intraceprint : inter-NI trace
stpNISock_boardupprint : boards up
stpNISock_printon : activates STP Socket Handler printf
stpNISock_printoff : desactivates STP Socket Handler printf
stpni_printStaFied : status field description trace
stpni_debugPport : Physical Port editing trace
stpni_debugLport : Logical Port editing trace
stpni_debugport : Physical & Logical Port editing trace
stpni_traceprint : event and warning trace
stpni_printon : activates STP NI printf
stpni_printoff : desactivates STP NI printf
These NI spanning tree trace utilities are described in the subsections that follow.
Event Trace (stpni_traceprint)
This trace includes the events received and generated by the Spanning Tree and the warning detected
while processing an event. A warning entry contains the name of the C source file and a line number. The
explanation of the warning can be given by Engineering.
Each event trace entry is built as follows:
• An ASCII pattern reflecting the event.
• Up to 4 parameters (a -1 (or 0xffffffff) indicates that the parameter is not significant).
The following is an example of the stpni_traceprint command printout:
Event names displayed by the stpni_traceprint command are described in the subsections that follow.
PORTATCH
This corresponds to a port attached event received from the Spanning Tree CMM. The Spanning Tree
CMM generates this event when it receives a Port attach indication from the Port Manager.
The parameters are:
• First parameter: Global port identifier.
• Second parameter: Default VLAN associated to the port.
PORTDELE
This corresponds to a port detach event received from the Spanning Tree CMM. The Spanning Tree CMM
generates this event when either it receives a Port detach indication from the Port Manager or there is
change in the port type (e.g. transition from aggregable to fixed, mobile to fixed).
• First parameter: Global port identifier.
ADDVLAN
This event is generated by the Spanning Tree CMM when it receives a VLAN added event from the
VLAN Manager. This events is sent to all the NI that are up and running by the Spanning Tree CMM.
The parameters are:
• First parameter: The VLAN identifier.
• Second parameter: The Spanning Tree type. A 1 indicates Flat Spanning Tree while a 2 indicates 1x1
Spanning Tree.
• Third parameter: The VLAN administrative state. A 1 indicates Enable while a 2 indicates Disable.
• Fourth parameter: The Spanning Tree administrative state. A 1 indicates Enable while a 2 indicates
Generic Troubleshooting in DshellTroubleshooting Spanning Tree
MODVLADM
This event is received is sent by the Spanning Tree CMM to the NIs when the administrative state of a
VLAN is changed (event generated by the VLAN Manager to the Spanning Tree CMM).
The parameters are:
• First parameter: The VLAN identifier.
• Second parameter: The VLAN administrative state. A 1 indicates Enable while a 2 indicates Disable.
MODVLSTP
This event is received is sent by the Spanning Tree CMM to the NIs when the Spanning Tree state of a
VLAN is changed (event generated by the VLAN Manager to the Spanning Tree CMM).
The parameters are:
• First parameter: The VLAN identifier.
• Second parameter: New Spanning Tree. A 1 indicates Enable while a 2 indicates Disable.
Note. When the Spanning Tree state is Disable, all the ports (Up) are moved to the forwarding state and
are removed from the Spanning Tree scope.
ADDQTAG
This event is received is sent by the Spanning Tree CMM to the NI when a tag is added to a port belonging to that NI. This event is generated on the CMM by the 802.1Q application.
The parameters are:
• First parameter: Global port identifier.
• Second parameter: The 802.1Q tag.
Note. This event is processed by the Spanning Tree NI as a port attach event.
DELQTAG
This event is received is sent by the Spanning Tree CMM to the NI when a tag is removed a port belonging to that NI. This event is generated on the CMM by the 802.1q application.
The single parameters is:
• First parameter: Global port identifier.
Note. This event is processed by the Spanning Tree NI as a port attach event.
Troubleshooting Spanning TreeGeneric Troubleshooting in Dshell
MDEFVLAN
This event is received is sent by the Spanning Tree CMM to the NI when the default VLAN of a fixed or
q-tagged port is change (this also applies to logical port). This event is generated on the CMM by VLAN
Manager application.
The parameters are:
• First parameter: Global port identifier.
• Second parameter: new default VLAN.
PORTAGGR
This event is currently unused.
PORTDISG
This event is currently unused.
AGGR_UP
This event is sent by Link Aggregation NI when it detects that a aggregator comes up; It could be either a
static aggregator (OmniChannel) or a dynamic aggregator (802.3ad). This message is generated when the
first port joins the aggregator only.
The parameters are:
• First parameter: The aggregator identifier (logical port ID value between 0 and 31).
• Second parameter: The global port identifier of the physical port that has joined the aggregator.
• Third parameter: The output QID to be used by the Spanning Tree (not significant).
Note. The output QID is no more used by the Spanning Tree since at the time Link aggregation is asking
for the default queue associated to the physical port, Qdriver might not be ready the provide it. However
Link Aggregation keeps providing this parameter even if now this one is not significant.
AGGRDOWN
This event is sent by Link Aggregation NI when it detects that a aggregator goes down; It could be either a
static aggregator (OmniChannel) or a dynamic aggregator (802.3ad). This message is generated when the
last port has leaved the aggregator.
The single parameter is:
• First parameter: The aggregator identifier (logical port ID value between 0 and 31).
Generic Troubleshooting in DshellTroubleshooting Spanning Tree
PORTJOIN
This event is sent by Link Aggregation NI when a physical port is joining an aggregator; It could be either
a static aggregator (OmniChannel) or a dynamic aggregator (802.3ad). This message is generated after the
first port has joined the aggregator (see “AGGR_UP” on page 4-13).
The parameters are:
• First parameter: The aggregator identifier (logical port ID value between 0 and 31).
• Second parameter: The global port identifier of the physical port that has joined the aggregator.
PORTLEAV
This event is sent by Link Aggregation NI when a physical port is leaving an aggregator; It could be either
a static aggregator (OmniChannel) or a dynamic aggregator (802.3ad). This message is generated after the
first port has joined the aggregator (see “AGGR_UP” on page 4-13). Link aggregation provides the aggre-
gator identifier, the global port identifier of the port which is leaving it and the global port identifier of the
newly primary port
The parameters are:
• First parameter: The aggregator identifier (logical port ID value between 0 and 31).
• Second parameter: The global port identifier of the physical port that has joined the aggregator.
• Third parameter: The global port identifier of the physical port that will have the primary port role.
• Fourth parameter: The output QID of the newly primary port (not significant; see note of “AGGR_UP”
on page 4-13).
BRGPARAM
The is event is generated by the Spanning Tree CMM when a configuration parameter of the Spanning
Tree is changed by the operator. This message is sent to all the NI that are up and running.
The parameters are:
• First parameter: The spanning identifier (i.e., VLAN identifier).
• Second parameter: The type of the parameter. A 1 indicates Spanning Protocol (802.1w(third parame-
ter=4)/802.1D(third parameter=3)), a 2 indicates Spanning Tree (Flat (third parameter=1)/ or 1x1 (third
parameter=2)/), a 3 indicates the bridge priority value, a 4 indicates the Hello timer value, and a 5 indicates the forward delay value, and a 6 indicates the maximum age.
Troubleshooting Spanning TreeGeneric Troubleshooting in Dshell
PTSTPMOD
The is event is generated by the Spanning Tree CMM when the Spanning Tree configuration parameter of
a port is changed by the operator.
The parameters are:
• First parameter: The spanning identifier (i.e., VLAN identifier).
• Second parameter: The global port identifier.
• Third and fourth parameters: The type of the parameter/value. A 0x11 indicates mode of the port
(dynamic(1), blocking(2), forwarding(3)), a 0x12 indicates Spanning Tree administrative state of the
port (enable(1),disable(2)), a 0x13 indicates port administrative state, a 0x14 indicates port priority, a
0x15 indicates port path cost, and a 0x16 indicates port connection type (half-duplex(1),point to point
(2),auto point to point(3),edge(4)).
PORTMOD
The is event is sent by the Spanning Tree CMM to the Spanning Tree NI when the administrative state of
a port is modified by the operator.
The parameters are:
• First parameter: The spanning identifier (i.e., VLAN identifier).
• Second parameter: The global port identifier.
• Third and fourth parameters: The type of the parameter/value. A 0x13 indicates port administrative
state (enable (1),disable(2)).
PORTVLBK
This event is an internal event which generated by the Spanning Tree when STP is processing a Port/
VLAN blocking that can take place at VLAN level or port level.
The parameters are:
• First parameter: The blocking status. A 0x44 indicates blocking already done, a 0x88 indicates nothing
to do, a 0x55 indicates blocking at port level, and a 0xaa indicates blocking at VLAN level.
• Second parameter: The local port identifier.
• Third parameter: The VLAN identifier.
PVLANBLK
This event is registered when the Spanning Tree is generated a Port VLAN Blocking message to Source
Learning NI.
Generic Troubleshooting in DshellTroubleshooting Spanning Tree
The Port VLAN blocking message sent to the Source Learning NI has the following structure:
uint16 VlanId, uint32 PortVector
This event has the following values for the message ID:
• appID: APPID_SPANNING_TREE.
• subMsgNum: STP_PortVlanBlocking.
These event fields are defined below:
• VlanId: A value 1 to 4095 identifies a VLAN (0 means that the message is applied to ports defined by
the PortVector on all VLANs).
• PortVector: A field of bits, one bit by the physical port, which indicates if the port is concerned by the
change of state.
GMBPDU
This message is sent by the Spanning tree NI to the local Group Mobility NI each time a BPDU is received
on a mobile port. Group mobility can take two actions depending on how the mobile port has been configured:
• Ignore BPDU: In this case Spanning Tree will keep on sending GMBPDU each time a BPDU will be
received on the port (there is no Spanning Tree computation for the port).
• Move port to fixed: Group Mobility asks Spanning Tree to revert the mobile port to the fixed state and
the port will be added to Spanning Tree associated to VLAN 1.
The BPDUonMobPort message sent by the Spanning Tree NI has the following format:
This event has the following values for the message ID:
• appID: APPID_SPANNING_TREE
• subMsgNum: STP_BPDUonMobPort
These event fields are defined below:
• LocalPortId: Identifies the physical Port (local reference: 0 to 23) which received the BPDU.
• bpdu_lgth: The length in bytes of the following BPDU.
• bpdu_data: The BPDU.
GMIGBPDU
This message is sent by Group Mobility NI in response to a BPDU on mobile port message sent by the
Spanning Tree. By sending this message group mobility tells to Spanning Tree to ignore BPDU on the
mobile port.
Troubleshooting Spanning TreeGeneric Troubleshooting in Dshell
GM2FIXED
This message is sent by Group Mobility NI in response to a BPDU on mobile port message sent by the
Spanning Tree. By sending this message group mobility tells to Spanning Tree that the mobile port must
be reverted to the fixed state.
The parameters are:
• First parameter: The global port identifier.
• Second parameter: The default VLAN.
VMADDVPA
The event is sent by the VLAN manager NI when a new VLAN needs to be added to a mobile port (no
longer used by the VLAN manager).
The parameters are:
• First parameter: The global port identifier.
• Second parameter: The default VLAN.
VMDELVPA
The event is sent by the VLAN manager NI when a VLAN needs to be removed from a mobile port (no
more used by VLAN manager).
The parameters are:
• First parameter: The global port identifier.
• Second parameter: The default VLAN.
VMDEFVPA
The event is sent by the VLAN manager NI when a the default VLAN of a mobile port needs to be
changed.