Worldwide [change] |Account |Register |About Cisco
Guest

Hierarchical Navigation

Cisco WAN Manager

Capturing Debug Information from Cisco WAN Manager

Downloads

Document ID: 27026


Contents

Introduction
Prerequisites
      Requirements
      Components Used
      Conventions
Turn CWM Processes into Debug Mode
General Procedure to Log the Debug
      Step-by-Step Instructions
Debug Information
Verify
Troubleshoot
Related Information

Introduction

This document explains how to capture the log information you need to send to the Technical Assistance Center (TAC) to troubleshoot. Most Cisco WAN Manager (CWM) log files required by the TAC are in the /usr/users/svplus/log directory. You can increase the log level into debug mode. However, you must stop and restart the CWM core. If you do not want to warm start the core, you must send either the full content of svplus/log or a particular subset of that directory to TAC. This document provides examples of how to capture debug information from the CWM processes.

Prerequisites

Requirements

There are no specific prerequisites for this document.

Components Used

The information in this document is based on these software and hardware versions.

  • CWM 11.0.00

  • CWM 10.5.10

  • CWM 10.4.10

  • CWM 9.2

Conventions

For more information on document conventions, refer to the Cisco Technical Tips Conventions.

Turn CWM Processes into Debug Mode

Some configuration files in the /usr/users/svplus/config directory contain the debug level field for particular processes. You can begin to log debug information into log files when you turn this variable from the normal log mode of two, to a higher value and restart a process. This is a list of some of those configuration files:

  • CWMGateway.conf

  • ILMITopoc.conf

  • dbglevel.conf

  • emd.conf

  • emsd.conf

The dbglevel.conf. file, is the basic file for most of the processes. In order to cause most of the CWM processes to generate debug information, set the Debug Level from two to five in this file and restart the CWM core. If the debug must be specific to one process, you can change its configuration file and restart the process.

caution Caution: Debug information can quickly swamp the file system. Monitor the disk space and turn CWM back to the regular log mode once you complete the test.

Remove the content of svplus/log before you capture new debug information. This helps TAC to quickly locate the correct log messages in the log files. You must stop the CWM core in order to remove log files.

General Procedure to Log the Debug

Step-by-Step Instructions

Complete these steps to capture debug information.

  1. Change /usr/users/svplus/config/dbglevel.conf from Debug Level 2 to Debug Level 5.

  2. Stop core: Option 2 in CWM script.

  3. Remove the log files under svplus/log (you can back up the current ones if necessary): Option C in CWM script

  4. Start core: Option 1 in CWM script.

  5. Reproduce the problem.

  6. Stop the core: Option 2 in CWM script.

  7. In order to tar up the log directory, issue these commands:

    cd /usr/users/svplus/log 
tar cvf /tmp/log.tar *.log 
compress /tmp/log.tar

  8. Return the debug level to two in config/dbglevel.conf file.

  9. Restart the core.

  10. Email the /tmp/log.tar.Z file to TAC.

Debug Information

Element Managers

In order to perform a more sophisticated procedure to capture debug information for the processes that deal with nodes (such as configuration update and filter alarms), complete these steps:

  1. Select a switch where you have problems with traps or update messages.

  2. Stop the core.

  3. Turn debug level to five in config/emd.conf or config/emsd.conf files.

  4. Remove the log files in the log directory.

  5. Start the core.

  6. Issue the kill -USR1 <emsd pid> command and determine what EMSC child manages that switch. For example, from svplus/tmp/emsd.dump, locate the child id. This is part of that file in our lab:

      NODES  nodeList_g: 
    ID   name       ipaddr   (int)       child 
    4    wanpeach   c0a865c5 -1062705723 3 
    3    wanbush    c0a86603 -1062705661 2 
    2    wankriek   c0a86642 -1062705598 4 
    1    wankwak    a3047c3  170936259   5 
    0    wanduvel   a304740  170936128   1

    BPX wanduvel is child number 1.

    Once you have that child id, kill -HUP the pid of that child. For example:

     slavefish% psg emsc 
    9274 ?        S  5:23 emsc -c 1 -e 9010 -D stratacom -t 8999 
    9275 ?        S  3:25 emsc -c 2 -e 9010 -D stratacom -t 8999 
    9276 ?        S  0:45 emsc -c 3 -e 9010 -D stratacom -t 8999 
    9277 ?        S  0:35 emsc -c 4 -e 9010 -D stratacom -t 8999 
    9278 ?        S  0:45 emsc -c 5 -e 9010 -D stratacom -t 8999

    In this example, you issue the command: kill -HUP 9274.

  7. Reproduce the problem with severity 0.

  8. Tar up the log directory and send that file to the TAC.

Alarm Subsystems

Two important processes in the CWM Alarm subsystem are the rtm and the eventd processes. The rtm process handles traps from the MGXs node. The eventd process deals with alarms from BPX or IGX (SWSW world). In order to turn those processes into debug mode, kill them because they do not read the debug level from dbglevel.conf at startup.

rtm

  1. Set RTM debug level to five in the ~svplus/conf/rtm.conf file:

    ps -deaf | grep rtm 
kill -HUP <eventd pid>

  2. Email the rtm.log file to TAC. Issue this command to check that file:

    <rtmTrapInput> trap from agent <ip address> trap# <trap_number> total traps <Total_traps>
(  18581:  1) 10:16:51 WARNING: <rtmTrapInput> trap from agent 172.29.15.144 trap# 50134 total traps 39685

  3. Reset the rtm default debug level to two. Otherwise, it compromises the rtm performance.

eventd

  1. Change the debug level in dbglevel.conf to five.

    ps -deaf | grep eventd 
kill -HUP <eventd pid>

  2. Email eventd.log to TAC.

  3. Reset the eventd default debug level to two.

Statistics Collection Manager

Complete these steps:

  1. As svplus user, issue these commands to capture the information from the main log for TAC:

    cd /usr/users/svplus
echo 'select node_id, node_name, mode from node' | dbaccess stratacom > log-node.txt
ls -lRt incoming purge spool > log-ls.txt
echo 'select * from coll_summ_list' | dbaccess scmdb > log-coll_summ_list.txt
tar cvf /tmp/scm.tar log-*.txt log/scm*.log log/stats*.log* config/scm*.conf config/*.stats etc/*
compress /tmp/scm.tar

  2. Send the scm.tar.Z file to the TAC engineer in charge of your case.

Verify

There is currently no verification procedure available for this configuration.

Troubleshoot

There is currently no specific information available to troubleshoot this configuration.

Related Information

Updated: Apr 21, 2006 Document ID: 27026