Troubleshooting Dr. Watson Errors in SMSEXEC

• Microsoft Systems Management Server versions 1.0 and 1.1

SUMMARY

SMSEXEC can fail with a Dr. Watson error when either the Inventory Processor, Inventory Dataloader, or Despooler attempts to process corrupt files.

Associated with the Dr. Watson problem are memory read and write errors.

MORE INFORMATION

There are three known reasons for the SMSEXEC process to Dr. Watson. These are a corrupt RAW file, a corrupt MIF file, or a corrupt NIL file.

RAW Files

Located in \sms\logon.srv\inventry.box or \sms\site.srv\inventry.box. The Inventory Agent program reports the inventory for the client by creating a RAW Inventory Agent file (*.raw) and placing this file in the Inventry.box directory of the SMS_SHR share on the client's logon server.

MIF Files

Located in \sms\site.srv\dataload.box\deltamif.col. A management information file (MIF) is an ASCII text file that contains information about a computer component. A MIF is a description of the component that is available to other applications through a standard management interface.

NIL Files

Located in \sms\site.srv\despool.box\receive. There are four possible extensions for an instruction file: *.ins, *.sni, *.ist, and *.nil.

A *.ins or *.sni instruction becomes *.ist or *.nil when the SMS Despooler starts processing the instruction. A *.ist file is for instructions that have an accompanying package file; *.nil is for stand-alone instructions.

Steps to determine what is causing the problem:

1. Verify the existence of a DS_00000.NIL file in \sms\site.srv\despool.box\receive. If this file exists, remove it. Check the despool log file and verify which package ID it was failing on. Then restart SMSEXEC. You may want to remove the problem package and try resubmitting.

2. Create a temporary directory off the root. Move the MIFs from \sms\site.srv\dataload.box\deltamif.col to the temporary directory. Start SMSEXEC and determine whether the crash still occurs. If it doesn't, begin moving the MIFs back to the deltamif.col directory, moving the newest RAWs first. based upon file-creation date. Move them in groups of approximately 10 to 20 MIFs.

1. Create a temporary directory off the root. Move the RAWs from \sms\logon.srv\inventry.box and \sms\site.srv\inventry.box to the temporary directory. Note that we are only concerned about RAW files that have made it to the server that is crashing; we are not concerned about logon servers with backed-up RAWs.

Once the system fails, take the first RAW and move it back to the temporary directory. Restart SMSEXEC. Repeat this process until all BAD RAWs have been isolated. Then edit the RAWs to determine which clients are causing the crash. You will then need to troubleshoot the clients to determine why they are creating bad RAWs (bad hardware, bad UINFO/Custom MIF, and so on).