I have users in a bunch of containers, some users work all the time with
CIFS and some only work once in a while. Iím certain that all contexts are
defined properly. Iím using a Windows 2003 domain controller and a NetWare
6.5 SP2 server with a tree. Iím using domain pass-through for CIFS
and have the correct WINS server defined.

All users that donít work all the time are in these OUs:

All users in these OUs always work properly:

I ran load NMASMON * sys:\public\nmas.txt trunc and the logging shows that
everything is working fine.

I have edited the registry on the PDC to include the setting mentioned in
TID 10079206 to tell the DC to not force a reset after 15 minutes because
I do have Citrix servers.
(HKLM\System\CurrentControlSet\Services\LanmanServ er\Parameters\

I have tweaked the Windows policies (local, domain and domain controller)
as described in TID 10084607 to allow CIFS authentication on the PDC, so
that isnít the problem.

Anytime I do a CIFSSTOP and then CIFSSTRT all users in all OUs work for a
little while but the ones in the problem OUs will break within an hour.

I have closely compared the accounts that always work with the accounts
that only work intermittently and I canít find a difference. I have placed
a forced ZEN application on some of the problem users to test. Even when
the users canít map the drive to CIFS they can run the forced application
which is using the same CIFS share. The problem seems to be a drive
mapping issue or a name resolution problem but I can ping the CIFS share
(SERVER-W) fine since itís in DNS, WINS and the HOSTS file at this point.
The problem is on multiple machines so it not a client workstation issue.

Any help would really be appreciated.