XGS1930 dhcp client strange behavior

Alberto77

I have a XGS1930-28 with fw 4.50 (ABHT.7) connected to linux pc with debian 9 os (with dnsmasq versione 2.80 as dhcp server).
Sometime, when dhcp lease time expire, happen something very strange, dhcp negotiation fail, this is linux log:

 Dec 14 00:00:14 localhost dnsmasq-dhcp[1698]: DHCPREQUEST(br0) 192.168.127.166 bc:cf:4f:f2:bc:2f 

Dec 14 00:00:14 localhost dnsmasq-dhcp[1698]: DHCPACK(br0) 192.168.127.166 bc:cf:4f:f2:bc:2f XGS1930-v1

Dec 14 00:00:14 localhost dnsmasq-dhcp[1698]: DHCPDISCOVER(br0) bc:cf:4f:f2:bc:2f 

Dec 14 00:00:14 localhost dnsmasq-dhcp[1698]: DHCPOFFER(br0) 192.168.127.166 bc:cf:4f:f2:bc:2f 

Dec 14 00:00:18 localhost dnsmasq-dhcp[1698]: DHCPREQUEST(br0) 192.168.127.166 bc:cf:4f:f2:bc:2f 

Dec 14 00:00:18 localhost dnsmasq-dhcp[1698]: DHCPACK(br0) 192.168.127.166 bc:cf:4f:f2:bc:2f XGS1930-v1

Dec 14 00:00:18 localhost dnsmasq-dhcp[1698]: DHCPDISCOVER(br0) bc:cf:4f:f2:bc:2f 

Dec 14 00:00:18 localhost dnsmasq-dhcp[1698]: DHCPOFFER(br0) 192.168.127.166 bc:cf:4f:f2:bc:2f 

Dec 14 00:00:22 localhost dnsmasq-dhcp[1698]: DHCPREQUEST(br0) 192.168.127.166 bc:cf:4f:f2:bc:2f 

And this operations continue infinite loop.
The very stange situation is that the dnsmasq.leases contains old ip address 192.168.127.165 but in arp table there is new ip 192.168.127.166.

XGS1930 answer on 80 port on new ip 192.168.127.166 but answer with ping on both ip (????).

I have solve with force new ip negotation on  XGS1930 (from web browser interface).

Do you now if there is this kind of problem and this is fix in new fw (if this kind of problem is due at XGS1930 obviously)?

Thanks

Zyxel_Melen

Hi @Alberto77,

Have you tried another DHCP server and got the same result?
Also, how did you determine the ICMP response of IP 192.168.127.165 comes from XGS1930 but not other devices? Have you used some tools to confirm the MAC address?

I will try to replicate this problem. Could you send your configuration via private message?