We are happily running a PVE cluster with several virtual desktops(Windows). The cluster is connected to a LAN with several thinclients, printers and a Windows server. At the moment this Windows server, besides domain controller, is the dhcp server. Thinclients, computers and virtual desktops all get there addresses from this server and all works fine. Recently we've installed a new Windows(2016) server on the PVE cluster to replace the old Windows server in the LAN. This server should become the dhcp server for the complete network replacing the old server. .. and now for the problem. During testing(disabled the old dhcp server) we discovered that dhcp from the new server only works for vm's on the node the new server is installed on. VM's on other nodes or computers in the LAN don't get an address from the new server. Can anybody help us with this problem?
Did you check with Wireshark if the packets arrived or what happend on the network level? Did you use the firewall and have some rules active? Did you configure your switch correctly?
Yes I did with tcpdump(tcpdump -i vmbr0 port 67 or port 68) and no packets arriving. All firewalls (proxmox, windows) disabled for the test. Thinking ... the switch. What is there to configure on the switches? The other DHCP server is working so ... it should be fine ..... ?? Oooppss, suddenly i remembered. A few years ago there was a rogue dhcp server(third party cisco router) in the network which i couldn't disconnect and also didn't manage. To tackle the problem i activated dhcp snooping on the switches allowing only for one specific dhcp server :-( Thank you for triggering my memory! Problem solved.
