Linux box is running a DHCP server. When a winXP client tries to get an IP address it just keeps trying with no timeouts and no errors. If I try to repair the connection on the winXP system I immediately get an error saying "Windows could not finish repairing the problem because the following action cannot be completed - Renewing your IP address".
Looking at my dhcpd.leases file (using tail -f) I can see it is continually being updated every second with the message below being repeated (with just the timestamps changing by one second each time):
lease 192.168.80.150 {
starts 5 2008/06/20 16:15:57;
ends 6 2008/06/21 16:15:57;
cltt 5 2008/06/20 16:15:57;
binding state active;
next binding state free;
hardware ethernet 00:16:d4:b1:cf:82;
uid "\001\000\026\324\261\317\
202";
client-hostname "rental-02";
}
So if I read that correctly the winXP is getting the request to the DHCP server, the DHCP server is assigning an IP address but for whatever reason the winXP client is not accepting it and keeps re-requesting it every second or perhaps the DHCP server keeps sending it every second until it gets a response.
Okay, now for the very complicated part. All was working fine until yesterday when we replaced a switch on our network and we turned our servers off as well to re-route power cables while we cleaned up some network patch cables. The switches do have DHCP flitering options but this is disabled and we have to vlan groups each with a linux box that runs dhcp & samba and the other server (with identical dhcpd.conf file is identical except the ip addresses reflecting their subnets) runs fine and all PCs are assigned an IP address correctly from it.
I do notice that sometimes when I run netstat that it shows the dhcpd process listening on both 0.0.0.0:67 and 0.0.0.0:32xxx with the Recv-Q line for the port 67 line reading 1800 (that field is zero for all other ports). I recompiled dhcp (I use gentoo linux) and that seemed to go away and it is just bound to 0.0.0.0:67
Additional info:
I run dhcpd in chrooted mode on both servers.
Here is a snippet of my dhcpd.conf file:
option ntp-servers 192.168.80.254;
option domain-name "xxx.ca";
option domain-name-servers 192.168.80.253;
option netbios-name-servers 192.168.80.253;
option netbios-node-type 8; ### Node type = Hybrid ###
option subnet-mask 255.255.255.0;
ddns-updates on; ### Dynamic DNS enabled ###
ddns-update-style interim;
authoritative;
default-lease-time 86400;
max-lease-time 172800;
log-facility syslog;
subnet 192.168.60.0 netmask 255.255.255.0 {
}
subnet 127.0.0.0 netmask 255.0.0.0 {
}
subnet 192.168.80.0 netmask 255.255.255.0 {
range dynamic-bootp 192.168.80.150 192.168.80.200;
option routers 192.168.80.254;
allow unknown-clients;
}
host computerlab01 {
hardware ethernet 00:1a:a0:3f:32:a1;
fixed-address 192.168.80.1;
}
host computerlab02 {
hardware ethernet 00:1a:a0:3f:34:16;
fixed-address 192.168.80.2;
}
... and this continues for each computer in the computer lab
the dhcp portion is basically only used for the spare network ports in the room for when people plug in their notebooks.
The PCs in the computer lab appear to be getting the static assigned IP address via DHCP but then perhaps because they've had the address before they are going off some winXP cache or something.
Any thoughts?
Start Free Trial
