Novell Home

How To Troubleshoot GroupWise Abends - Part I

Novell Cool Solutions: Tip

Digg This - Slashdot This

Posted: 20 Dec 2001
 

Versions: GroupWise 6

GroupWise Abend Troubleshooting Guide

For details or updates on this tip, see TID-10021981.

PART I

Steps to follow when troubleshooting a GroupWise Abend:

(See Part II of this guide for a theoretical explanation of GroupWise Abends and Abend definitions.)

STEP 1: PERFORM `REGULAR' TROUBLESHOOTING STEPS (Do these steps if there are good reasons for doing them):
Load POA with /noconfig switch
Rebuild PO / Dom databases
Run Vanilla: Unload all but the GW NLM that's causing the abend
Move GW agents off to another server
Rename the WPCSOUT/WPCSIN directory, then load GW Agent
Rename the OFFILES directory, then load GW Agent
Rename Index subdirectories
Load the POA with /nomf /noqf /norecover /notcpip /noconfig /nosmp.
Follow TID 2951111 suggestions (Resolving POA Abends)
Follow TID 2953249 suggestions (Resolving MTA Abends)

STEP 2: LOOK AT CONFIG.TXT
Look at NLM dates. Even if you have applied the latest Support Packs (OS, GW, CLIB, etc.), it may be a later TCP/IP Support Pack available. Check Config reader's Suggestions' tab. Current ConfigReader NLM is found in CFGRD6B.EXE - NetWare Config Reader Ver 2.67 at support.novell.com

STEP 3: LOGGING (MTA, ADA, POA) Turn on Verbose Logging (Diagnostic for the POA). Then bring the Agent up, let it abend the server, then go into the debugger and toggle to the agent screen to get additional information. Often you can see that the MTA or POA is trying to process a message or the event that is causing the abend. See TID 2920074 if you need help using the NetWare Debugger utility.

STEP 4: LOOK AT ABEND.LOG

  1. Check `Running process'
  2. Check EIP
  3. What is at the top of the stack
  4. Are all Abends identical ?

Review abend log to check for a history of the abends. Does the EIP always point to the same address? Is GroupWise always the running process?
Look for commonalities and trends with EIP and Stack information.
Look at Patch Level (CLIB, GW, TCP, etc)
Look at Break Line (# before the colon) indicates which Abend # it is
Running Process: Is this a GW NLM?
Where is EIP: Even if it is in TCP.NLM GW may have passed it a bad parameter (EIP in ___________.NLM)
Go through each Abend, list each, and how many times each Abend happens
IE. Server.nlm abended 10 times, GWPOA abended 15 times '..focus on the one that happens most frequently first
How often does an abend occur
Look at the stack information in the Abend.log

The next step is to `walk' the stack to determine which process passed the bad parameter and caused the abend. This is usually done by Novell Technical Services.

STEP 5: APPLY LATEST PATCHES (GW, NW, NT)
A high percentage of Abends are corrected by applying current versions of the Operating System and GroupWise patches. Check http://support.novell.com for latest patch information.

STEP 6: UPDATE NIC AND DISK DRIVERS
The only way to assure that you have the latest version of these drivers is to download them from the respective vendor. The compare size and date against the same on your server.

STEP 7: CHECK SERVER STATISTICS
IE. LRU Stat ' `Last Resource Used.' The higher the number the better. Shows the last time the system ran out of cache and had to swap information to the harddrive. If this number is low (IE. Less than 15 minutes) it indicates that more memory is needed. Check utilization. Understand what the `baseline' is prior to making a judgement. On the average, utilization above 60% is considered high. Off-load NLMs to another server, upgrade CPU.

STEP 8: APPLY GW FIELD TEST FILES
These files are 'pre-release' files that Novell has developed to increase functonality and to fix known issues. This requires calling Novell Technical Services and open up an incident.

STEP 9: CHECK ABEND TID FOR ADDITIONAL SUGGESTIONS
TID 2917538 (Troubleshooting Abends)

STEP 10.: STREAMLINE GROUPWISE
Confirm that the server has enough resources to service GroupWise requests by implementing the following 14 SET parameters found in Appendix B in TID 10016883 (GroupWise Sizing Recommendations).

STEP 11: CHECK CONSOLE LOG
Console.log is found in sys:\etc directory and is a log of everything that happens on the server console.

STEP 12: GET A CORE DUMP
Different methods for getting core dumps: Cacheless vs. with cache Writing the IMG file to the local DOS partition vs. across the LAN to another server vs. to a different DOS drive See TIDs 2926501 and 2917547 for getting a coredump

For details or updates on this tip, see TID-10021981.


Novell Cool Solutions (corporate web communities) are produced by WebWise Solutions. www.webwiseone.com

© 2014 Novell