I use Win Server 2003 SP2 to connect about 100 users to printers using IP connections. Intermittintly, about 5-10 users will drop their connection to all printers their print status says "acces denied, unable to connect". I have tried reseting the TCP/IP printer server as well as the print spoller but this does not fix the problem. THe problem actually goes away, sometimes after a client reboot & sometimes after about 10-20 mins. Some of the same intermittent connection problems have been occuring during logon and with mapped network drives. During logon I have been getting errors stating the domin controller is unreachable. The work around I've been using is remove the computer from the domain and then rejoin. Mapped network drives have been giving permission denied errors intermittently as well. Its been rather crazy but all the problems seem to be connected to Keberos. In event viewer I have a bunch of event id error 4 The kerberos client received a KRB_AP_ERR_MODIFIED...

I'm just guessing here, hopefully someone smarter can give you a definitive answer. Is the license logging service running? It might be restricting access due to exceeding the maximum permitted number of CALs. If it is, you can disable the service (although this will make it harder to demonstrate compliance). Alternatively, still assuming it is the license logging service that's causing you're issue (which is by no means certain) you could change the time that Windows waits before disconnecting idle sessions, that way once a user has printed their document their connection will drop and their 'slot' will be free for someone else to use. .D

The License logging service is turned off. Thanks for the input.

Hmmmm I suppose it's possible that the network is simply overloaded. Are you on a 100MB LAN or 1GB? Use perfmon.msc to monitor the network cards on the server, there's all sorts of counters that you can add to track usage. Best of luck .D

Does sound like a overloaded server as Davehelps suggests. When was the last time you defragged the server drive(s)? Have you reviewed your dns event viewer log? Please post the specific event id in the errors you are seeing. Liberal: broad-mindedness, person who favors a political philosophy of progress, reform & protection of civil liberties Neocon:someone who believes in America but not in dissent, core of Democracy

We do defrag the server weekly and only use it for filesharing, DNS, DHCP, and IP Printing. I checked the task mgr and the usage is consistantly less that 2%. However I did find more information yesterday while troubleshooting a DNS issue. When I ping workstations I get different IP address than whats listed in the DHCP Manager (under Administrative tools). I think this whole problem is related to bad DNS pointers. Users IP address are not being resolved properly from the DC. I tried to RDP into a couple of stations today and was routed to the wrong machines. Things are getting messy. Is there anyway to reset the DNS pointers?

Open the DHCP console, right-click your DHCP scope and choose Properties. On the General tab, change the lease duration to unlimited. This will stop clients from even bothering to change addresses. On the DNS tab, tick Enable DNS dynamic updates... and choose to always dynamically update the records. This should force any change in a DHCP-allocated address to be updated in DNS. Better still, set up as many workstations as practical to use static IPs rather than DHCP. Then the pointers will always be accurate. If they're not moving anywhere then DHCP is just a waste of network bandwidth - save it for laptops. .D