I inherited a network that frequently loses connectivity. For example, I could ping a workstation and get 4 packet responses, and do it again and get timeouts. Basically, ALL or NOTHING network connectivity. Then, I have to unplug the cable for a few seconds and then plug it back into the computer and I have connectivity again (for a very short period). Sometimes EventViewer logs have MasterBrowser election alerts, but not everytime there is a connection problem. These things happen on every workstation and server on network. Link lights appear on computer and full-duplex 100mbps network switches. No hardware conflicts listed in ComputerManagement snap-in. Edited RegKey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Browser\Parameters so Server is the ONLY Master Browser on network. Able to browse all workstations and server in MyNetworkPlaces/Network Neighborhood. Server's Network Monitor driver does not indicate heavy traffic or packet storms or anything strange. Here are specs: 1 P4/256MB Windows 2000 Server and 25 PIII/128MB Windows 2000 workstations running TCP/IP (no NetBEUI,etc.). All machines part of one workgroup, with local accounts (since network doesn't really work yet). IP, WINS and DNS addresses are static on workstations. Server also has static IP and DNS addresses. WINS Server running on server with static entries. Machines all networked with Cat5 and 2 network switches connected by uplink. What do you think? Could it be a Cat5 or network switch hardware problem? Your suggestions in troubleshooting these things are very welcome.

This problem could be anywhere, Initally I would say that if you are pinging using ip address and losing those, then the loss of the server as the master browser will not affect the binding of the IP address on the NIC card and thus is not related to your issue. I would start from the physical layer checking all connections are ok. Because you describe your issue as coming and going, then I would think if the problem were at a physical layer and all your cabling is ok (assuming you are not using a wireless lan or some form of lan to lan routing) then I would expect the issue is relating to the nic or nic driver of the server. At a higher level at the data link layer, you would need to make sure that if you are using layer 2 switches or bridges, that the network connections are not being moved. Brides and switch create ARP tables which remember the mac address of the device that are connected to them. If you have moved the connections on your switch or bridge , the possibilty would be that your switch is remembering the location of the servers mac address and not flushing its arp cache properly. IN that case , you usually can manually flush the arp cache by command line or by turning off the switch for a few secoonds. At the network layer, I would check for IP address conflicts, it sounds as if address are static on your network, so I would check the server to ensure that someone is not using that IP address and causing a conflict. Usually your machine will indicated an IP address conflict but I have seen cases were this doesnt happen. In short, I think you need to check, is your network on one lan, if not and there is some dialup or wirless connection to the other location. Find out is this problem only confined to the server or are there other machines which have this issue. Is there any pattern to these timeouts? e.g. Time of day etc. Since you inherited the network have you made changes? If this issue is only occuring on the server then I would assume if its not an address conflict then it is the access method to it. Lastly, if this issue happens again, you should do a traceroute to see where this is failing, if this is just a lan then this is irelevant. The post is written on the assumption that the issue is a tcp/ip problem. Hope this is of use jonob

Yes, thank you for your insight. *The network switches were reset. *Static IPs checked- no conflicts in WINS or on local machines. *WINS, DNS, Default Gateway, Subnet Mask all ok too. *Cables will be tested Monday. I've noticed if I ping anywhere on network from single remote VPN client, everything is fine. If I ping internally from Server (i.e., 192.168.0.2) to any workstation (i.e., 192.168.0.10), I get two packets response, then timeouts and then network goes down again. This has also happened from workstations to other workstations/server. Network is wired-only with Cat5 and runs TCP/IP only. No Problem Devices found in Computer Management MMC. Do you have any other thoughts? Here is network structure (arrows are uplinks): INTERNET -> firewall (192.168.0.1) -> Network Switch 1 -> Network Switch2

Jonob, Oh do I have an update. The old Administrator tried to change the server IP back to 192.168.0.1 and deleted the default gateway (which is a router with ip address 192.168.0.1). So pings from server to workstations would not all return correctly. Now everything is fine. I shall document this as PEBCAK (Problem Exists Between Chair and Keyboard). Thank you for your time and assistance. Kahlib