I have a Netware 6 Server which has started giving me problems on boot.
The problem began after a long power outage caused by bad weather, and I'm convinced it must be some sort of Btrieve-related corruption or something. I've got a workaround, but I can't figure out how to fix the problem permanently.
The errors are "Unable to successfully initialize Audit Trail Module. Error code=0x20014" and "CSAudit initialization failure, err=2359306". In addition, I also get a Desktop reset error for every AFP volume on the server, and Mac users report vague problems logging in, slow or no login, having to reboot to log in, etc.
Aside from the DHCP problems and the Mac weirdness, I have not noticed any other symptoms.
The main problem I'm having is described in TID 10087843, which says that this is caused by the MKDE losing touch with the sys volume. The fix is to unload the CSAUDIT and microkernel files, then delete the contents of the sys:\system\cslib folder, and restart everything.
This does work. But it doesn't last. If I reboot the server, the error returns immediately, and I have to repeat the process. The server stays stable for days, but then Mac users start complaining,and DHCP starts acting up. At that point, following the TID does not work, and I have to reboot, then apply the TID to restore normalcy.
The server is running SBS 6. When the problem began, it was on SP4, but we have since applied SP5 during the trouble-shooting process. Also, this machine had Pervasive V8 installed in the past, though it's not used any longer. I don't know how significant that is.
Some things I've tried include: restoring the TID-mentioned modules from backups, copying the older PVSW and Btrieve modules over from a working server, replacing the MKDE folder, and numerous applications of the TID. I've also tried changing load order around, and loading in stages, but the problem persists.
I figure there must be some corrupted file or database somewhere, and if I can just identify it and replace it, I can solve the problem. But I don't know what to try replacing.
We have a call into Novell about this, but so far they're just saying "Hmmm" a lot.