I've got a pair of VMWare virtual machines that I work with regularly. One machine is the dev environment for a smart client app. The other machine is the business logic server (web services) for the smart client app. The database machine is actually on a third machine - an actual physical box. Because they have to talk to each other an talk to a physical box on the network, I have both of them set up with bridged networks, which means that they're publicly available and talking to the actual domain controller for DHCP and DNS purposes
Normally I boot these VMs every morning and shut them down every night. However, in that I was in the middle of some testing I just paused both of them and resumed them this morning. Unfortunately, I forgot to resume the server machine until after a few hours of using the client machine. This meant that my client somehow got confused and thought that my server's IP address was in the 192 range rather than the 10.10 range it really is in. I couldn't get the two of them to talk.
To resolve the issue, I tried clearing the DNS cache on the client with an 'ipconfig /flushdns' command. This did not work. I tried clearing the NBT Remote Cache Name table on the client with a 'nbtstat -R' command. This didn't work. What did work was renewing the IP address on the server with a pair of 'ipconfig /release' and 'ipconfig /renew' commands.
Now my machines are happily humming along again.