I need some suggestions/help with a Windows 2003 server issue. I haven’t been to the customers site yet, but their complaint is that their windows workstations (Windows XP workstations, mainly) are very slow to log in and make the initial connection to the network. The office is connected to the main headquarters via (2) T1’s, and I’ve checked the network traffic (they are averaging 32K/sec with a maximum of 188K/sec both inbound and outbound…IOW, they are hardly using their pipe at all) and this doesn’t seem to be a bandwidth issue (that is of course what they thought the problem was and why they brought me in to look at).
I’ve done some remote ping testing as well as some remote speed tests and file upload tests and there doesn’t seem to be any issue with speed or throughput, so that leads me to think it’s a systems issue of some kind. I’ve checked with some of the other remote sites and no one else seems to be having this issue.
What the people working at the troubled site are reporting is 5+ minute log in problems. Once they log in there doesn’t seem to be any issues, though they claim that they lose connection sometimes, and that the machines seem to lose the trust relationship with the domain fairly often.
Any ideas or suggestions would be a help as I’m not much of a systems guy and don’t have a lot of experience with Windows server/workstation environments.
Hi. Your first step should be to enable userenv logging on a couple of the problematic clients. This records the steps taken during the computer startup and user logon and is mainly about the application of GPOs (Group Policy Objects). This MS articles explains how to enable logging - don’t forget to turn it off once you’ve fixed it
What will be very useful is the time stamp that’s recorded against each action - you’ll be able to see when the system hangs for long periods without doing anything.
Using my psychic debugging, I’d suggest the following likely causes:
Somebody has made a GPO change recently. Probably one which runs a script from a non-local server or runs a custom vbscript logon script which is poorly coded (as in not site aware).
The sites & subnets config has been changed and the clients are no longer looking to their local domain controller (if they have one). Low WAN traffic suggests this is less likely, but even if the clients aren’t pulling much over the WAN they could be suffering from poor RTT.
Sorry, didn’t know anyone responded to this thread. My thanks for both of your input. I’m actually headed out to the site right now and will look at the situation and see if either of your suggestions are applicable. Again, thank you both.
