Question : NTBACKUP scheduled backup fails - VSS problem?

After upgrading server from Windows 2000 Server SP4 to Windows 2003 Server SP2 ntbackup scheduled jobs were initially OK, but now my weekly scheduled Full Backup (C: 92Gb and System State) fails. Other scheduled differential backups of the same data usually run without a problem, but have been known to fail occasionally. One possibly relevant factor is that my weekly full backup runs with the /um switch, which means it is an unmanaged backup, and will overwrite any tape in the drive. Nevertheless, the backups were working with this switch, and still do on another W2003 Server without issue.

The System log on a working backup normally show the following events:
1. Removable Storage service was successfully sent a start control
2. Removable Storage service entered the running state
3. The Volume Shadow Copy service was successfully sent a start control
4. The Volume Shadow Copy service entered the running state
5. The Microsoft Software Shadow Copy Provider service was successfully sent a start control
6. The Microsoft Software Shadow Copy Provider service entered the running state

When the backup fails, I only get 1 & 2, then about 31-33 minutes later
3. RSM was stopped
4. Removable Storage service entered the stopped state

I also get a UPHCLEAN event in the Application log at the same time as 3 & 4 on failed backups.

The backup log for ntbackup is empty, and no ntbackup application log events are recorded. I have seen many other similar problems on the web, but no working solution. I have tried the following:

1. Re-registered VSS DLLs
2. Installed VSS Update Rollup Package KB940349

Was going to move to backup exec but that is too costly and ntbackup has been great in the past.

Answer : NTBACKUP scheduled backup fails - VSS problem?

Well for the first time in weeks the backup ran (courtesy of not using the /um switch). So I guess I will set up separate jobs for my 4 full backup tapes. Interestingly, one of my other servers works fine with the /um switch, but it is not backing up so much data (22Gb). The current server was working OK, and data is constantly increasing, so perhaps the amount of data combined with the /um switch is causing a timeout somewhere?

Anyway, hopefully all is fixed. I will leave this question open for now, and post next week after another full backup. If there is a solution and I can use the /um switch that would be great. Thanks for your input Gary.