Running SVR 2008 R2 DC with DHCP and DNS all on the same box. The issue is that some clients will lose internet access suddenly for no apparent reason. Internal addresses (both URL and IP) are still accessible. Sometimes The IP/Hostname combination is consistent across DNS Forward and Reverse lookup zones and in the DHCP list. Most times the IP in the DNS Forward Zone has an older timestamp and is different than in the Reverse Zone or in the DHCP list but even after that has been rectified the issue persists. The problem only applies to generally one machine at a time but it rolls around to different machines (not all machines are affected) and there is no rhyme or reason to it (no specific IPs, Hostnames, Switches, or even times of day). It appears to be completely random amongst the machines that suffer from what we (not so affectionately) call the DHCP Bug. I can set a static IP and that sometimes lasts for days and others only about 20 minutes and then the problem returns to that machine. However, when it returns I can re-enable DHCP on that machine and it will usually (but not always) be fine for an extended period of time. There are certain individuals (regardless of machine) that seem to suffer from this more often, and others whom have never had the issue at all. My first thought was software issues however further investigation negated that too due to the fact at least one individual, that has never had an issue, is running the same software packages as the ones who are having the issue.
DHCP leases expire in 4 hours and DNS scavenging is set for the same time frame. This is due to the number of devices that come and go on the network during the course of a day and the fact that they would eat up all the IP addresses if I leased them longer.
I'm stumped. ANY ideas at this point would be appreciated.