I have inherited a client running SBS 2003 R2 and a mix of WinXP and Win7 clients. All machines have the latest updates.

I started getting complaints that 1 or more computers would 'drop off' the network and be unable to access the server (network drives or exchange) during this time.

When this happens, I can't ping the client pc from the server.

A number of things fix this: - Rebooting the server - Rebooting the switch and router - Rebooting the client But it always returns in a few hours.

We've tried re-cabling, it didn't help. We've replaced the switch and router and patch cables. I've disabled all TCP offload/chimney etc, updated the server's network card drivers and turned off power management on the network card.

What does work, is if I start Ping -T individually to each of the affected clients, the network will run indefinitely without problems. As soon as I stop the ping, I start getting complaints a few hours later.

Any ideas?


You might check the power save settings across the NIC driver and/or system BIOS. Also a log review across the Event Viewer as well (if that has not already taken place) would be another place to check.

  • I have already disabled all the power management on the NIC. I didn't notice anything in the event log, and I've looked at it a number of times. – Dom Sep 6 '11 at 6:35
  • Based on the additional information with DuPie's answer... Are the workstations with the varying success in accessing the server all on the same LAN? You might take a look at the ARP tables across the server, switch and workstations when this occurs. – user48838 Sep 6 '11 at 12:45
  • They're all on the same LAN, but some are in the router and some in the switch. As it's a very small company neither the router or the switch lets me view the ARP table. This is an interesting idea though, as I tried to look up ARP tables last time this happened, but due to the urgency of getting it all working during business hours it fell by the way side. Either way, we've put in a new switch and router and it still happens, so one assumes it's not an ARP issue in either. The server is the common denominator. – Dom Sep 8 '11 at 6:24
  • "we've put in a new switch and router and it still happens, so one assumes it's not an ARP issue in either" That is not a "solid" assumption if the ARP situation is induced by another device on the network. What about the ARP tables of the server and affected workstations during the occurrence? – user48838 Sep 8 '11 at 7:03

I'd likely blame power management settings on the NIC. Replacing the NIC with something non sucky (intel?) is good way to confirm.

Sounds like network link is dropping. Verify that within event viewer on the pc. Should show something like "Network Link down".

If you confirm it's not a network card issue (most likely), then the next likely suspect is a dhcp lease timeout/dhcp server problems. Again, that will be listed in event viewer on local pc. If in doubt, assign a static ip and see if disappears.

  • It's a HP Server NIC, should be non-sucky. When this happens, I can still communicate with some of the workstations, just not all of them. There doesn't seem to be anything in common with the workstations I lose. DHCP is set up pretty default for SBS2003. Lease time is 8 days. There's no way 8 day lease expirations are causing dropouts every 3 hours. – Dom Sep 6 '11 at 6:36

Despite all the connections to 'faulty' computers being replaced and still showing the problem, it appears it was a physical problem after all (who knows how PING fixes a physical problem).

We've just moved offices with all new cabling (same server, routers, switches) and I've turned off the pinging and no more problems (as yet).

It's going down in the book as another "What the?"

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.