I manage a very simple setup for a small organization. Recently the mail server crashed due to the system partition running out of space. This was due to excess files being stored on the C: drive that shouldn't have been there in the first place, I moved these files and cleared up about 15GB of space, restarted the mail server and everything seemed to be fine. Shortly after I started getting reports of users having problems connecting to their email. Some were having troubles connecting via Outlook (2003 and/or 2007), some were having troubles getting email through the OWA. And usually users having issues connecting to email via Outlook, also lost connecting to their network drivers as well. The issues are random, it doesn't happen to certain users, it happens to everyone at random times and it usually only lasts an hour or so. The only exception is user that have admin rights seem to be fine (except with the occasional OWA issue). Some things I have checked but I'm not sure if they are related or if I"m even on the right track. 1. When the user is having an issue with email and network drives, I cannot see the server by DNS name but I can by IP. Manually addeing the DNS record to the HOSTS file did nothing. Simply will not ping by DNS name, but will by IP address. (Admin accounts don't seem to ever have this issue) 2. Problems are not workstation specific, for example if a user is having problems, another user could log into that station and be fine. (again stating the problems are random) 3. When a user cannot connect to the OWA, they receive a 500 internal server error. If they check it again a few mins later it might be fine. This seems to happen to everyone, including my domain admin account. 4. The active directory quota table may be corrupt, but quotas are not even setup for the user accounts, so I'm not sure if this is an issue. I discovered this when running checks via ntdsutil -> semantic database analysis -> check quota. I did run the rebuild quota but I'm not sure if it did anything. 5. The WINS server seems to have stopped working, but I'm not sure how recent of an issue this is, or if it should even be considered related. All workstations have the server listed as their primary DNS server. Other thoughts: none of the problems existed until after the C: drive ran out of space and the mail server crashed. No updates have been installed, no software has been installed/removed. No changes were made other than moving some non-system files from the C: partition to the data partition. A server restart seems to fix everything temporarily, but only for a few hours. Other than the quota table, the AD database seemed fine. I've ran a check disk and a defrag on the C: partition since the mail server crash but neither seemed to help. I'm hoping someone can point me in the right direction to find my problem. This is becoming a very inconvenient issue. Oh one more thing about the setup, there is only 1 server, it serves as the mail server, file server, domain server, and dns server. There are 5 offices that use this server, 4 connect via a VPN connection, the issues occur both on the local network and the remote networks.

I would say it has to do with the workstations not being able to resolve the server via name. Everything you mention would confirm that as the root issue, except for the 500 error with OWA. That 500 error means it communicated with the server, but the server had an error. You should look in the error logs for details. It would also be a good idea to check the server error logs to try to nail down this DNS issue. Have you confirmed that the workstations have the correct network settings (ip, dns, etc.) while they are experiencing the issue? Does a ipconfig /renew and/or ipconfig /flushdns help? -Ryan Adams Free Computer Tips and more:http://RyanTAdams.com Paid Tech Support: Black Diamond

Well I had considered the DNS problem but I've somewhat ruled it out. I have confirmed the workstations have the correct settings and the DNS issues are not a problem if the user account is @ an admin level. Most workstations use a static IP address, but ipconfig /flushdns does not help. What's also weird is to troubleshoot the DNS problem, I manually added the server in the hosts file which should have fixed the issue, but even then I could not ping by DNS name, but could by IP address. Server name is 2k3server1, in normal circumstances I can ping this DNS name just fine. When I noticed users having a problem connecting and noticed I could not ping 2k3server1 but could ping it's IP address of 192.168.3.100 I added the following record to C:\windows\system32\drivers\etc\hosts 2k3server1 192.168.3.100 After adding that record I would still receive time outs when pinging 2k3server1 but could still ping 192.168.3.100, which I found very odd. When I logged the user off and logged in as myself, I could ping via DNS name 2k3server1 just fine. That almost makes it seem like a permissions issue, but I don't see what would cause it and why it works fine one minute, and doesn't another, but does again a few minutes later.

What if you try to ping using the full name such as 2k3server1.mydomain.local? Are the static IPs assigned via a reservation on the DNS server, or are the IPs manually entered on each client? What about on the server? -Ryan Adams Free Computer Tips and more:http://RyanTAdams.com Paid Tech Support: Black Diamond

Sorry I didn't mention that before, but I tested the full name as well when testing the DNS issues. It didn't matter if I was trying to ping the full name or the short version, same results. I also added the full DNS name into the hosts file. Also the only DNS issue seems to be the server, ping other locations such as other workstations or internet sites such as google.com come back fine. NSLOOKUP works fine and it shows the DNS record coming from the server, but again it will not ping via DNS name. As for the workstations, those that are statically set are still set to register themselves to the DNS server. I have looked at the DNS records on the server, nothing seemed to be out of place.

Though a pain, it may be a good idea to redo the whole DNS/DHCP setup. I'm pretty sure the inability to resolve the server by name is the root issue, so you need to get that fixed. -Ryan Adams Free Computer Tips and more:http://RyanTAdams.com Paid Tech Support: Black Diamond

Well I can give that a shot, believe me re-setting up the DNS/DHCP is not near as much hassle as some of the other fixes I have tried. The only reason I haven't and the only reason I don't feel like it's not going to resolve the issue is because the problem doesn't not exist if the user account is an admin. If I log in with my account I have no DNS issues on any workstation, the issue only applies to regular user accounts. And the problem is not consistent, although users may not be able resolve the server by DNS one mine, it's fine the next. But I will give that a shot later this evening when everyone is off the network.