We have a Win 2008 DHCP server handing out addresses to our L2 switch environment. There are multiple scopes for each switch type with the third octet changed for each one. For example Switch type 1 gets 192.168.2.x and switch type 2 gets 192.168.3.x. We utilize the vendor class option to shunt the types into the correct scope. This results in varying scope subnets of which all of the are either /23 or /24. Looking at a wireshark it appears that the /16 on the server NIC is handed out to all devices who receive an address and everything happily communicates.
I've exported and imported that scope into a new 2016 server. In testing activating the new scope I was losing management to devices, which would cause me to backout of my test and back to the old server. I compared the wireshark capture and the device that I lost management to received the correct IP address, but was handed the scope subnet of /23 rather then the server subnet. I discovered I can reach it from another switch in the same scope that has a /16.
Is this a behavior change in 2016 DHCP from the original server set up (inherited by me)? The overall preference would be to keep the 2008 behavior for the 11 scopes we have for the devices all in the same management network.
No comments:
Post a Comment