Linksys WET11 not acquiring DHCP address from ISC DHCPd server

(I also posted this in alt.networking.wireless earlier today, as my
DHCP server is running Ubuntu 6.06 Linux, I thought this group might
be able to help. Thank you.)


I am trying to get one of the Linksys WET11's to bridge a couple
desktops to the rest of my home network. The problem is that the
WET11 does not acquire a DHCP address from my Linux server running ISC
DHCPd. (Diagram of my network below)

While watching the syslog, I can see the WET11's MAC address appear
along with the DHCP requests, like this:

Jul 17 12:46:08 piglet dhcpd: DHCPDISCOVER from 00:06:25:45:78:d5 via
eth1
Jul 17 12:46:09 piglet dhcpd: DHCPOFFER on 192.168.0.8 to
00:06:25:45:78:d5 via eth1

The problem is that the WET11 never makes its DHCPREQUEST back to the
server. I can set a static IP address for the WET11 just fine, but
the clients connected to the WET11 have the same behavior as the
above.

The WET11 is configured in infrastructure mode and otherwise seems to
be working. It is version 1, and the latest firmware. The WAP11 is
also v.1 and the latest firmware. I have two of the WET11's and they
both exhibit the same behavior.

any ideas?
K


Diagram Legend
| denotes wired link
} denotes wireless link

Internet
|
Motorola Cable Modem
|
XUbuntu Linux server
2 Ethernet Interfaces
Shorewall / IPTables
ISC DHCP server
|
Linksys 10/100 Hub
|
Linksys WAP11 v 1.0 (not running DHCP)
}
Linksys WET 11
|
10/100 switch
|
Desktop clients

Try enabling BOOTP to see if that works.

I've noticed a few network appliances that need BOOTP, which is a pain
because I don't want that enabled and can't have it in some cases.


Kevin T. Neely wrote:
> (I also posted this in alt.networking.wireless earlier today, as my
> DHCP server is running Ubuntu 6.06 Linux, I thought this group might
> be able to help. Thank you.)
>
>
> I am trying to get one of the Linksys WET11's to bridge a couple
> desktops to the rest of my home network. The problem is that the
> WET11 does not acquire a DHCP address from my Linux server running ISC
> DHCPd. (Diagram of my network below)
>
> While watching the syslog, I can see the WET11's MAC address appear
> along with the DHCP requests, like this:
>
> Jul 17 12:46:08 piglet dhcpd: DHCPDISCOVER from 00:06:25:45:78:d5 via
> eth1
> Jul 17 12:46:09 piglet dhcpd: DHCPOFFER on 192.168.0.8 to
> 00:06:25:45:78:d5 via eth1
>
> The problem is that the WET11 never makes its DHCPREQUEST back to the
> server. I can set a static IP address for the WET11 just fine, but
> the clients connected to the WET11 have the same behavior as the
> above.
>
> The WET11 is configured in infrastructure mode and otherwise seems to
> be working. It is version 1, and the latest firmware. The WAP11 is
> also v.1 and the latest firmware. I have two of the WET11's and they
> both exhibit the same behavior.
>
> any ideas?
> K
>
>
> Diagram Legend
> | denotes wired link
> } denotes wireless link
>
> Internet
> |
> Motorola Cable Modem
> |
> XUbuntu Linux server
> 2 Ethernet Interfaces
> Shorewall / IPTables
> ISC DHCP server
> |
> Linksys 10/100 Hub
> |
> Linksys WAP11 v 1.0 (not running DHCP)
> }
> Linksys WET 11
> |
> 10/100 switch
> |
> Desktop clients

On 2006-07-18, Ken Roberts <(E-Mail Removed)> wrote:
> Try enabling BOOTP to see if that works.
>
> I've noticed a few network appliances that need BOOTP, which is a pain
> because I don't want that enabled and can't have it in some cases.

Thank you for the response. I tried adding a bootp range to my
/etc/dhcpd.conf file, but it does not seem to have worked. Watching
the syslog after powering up the WET11, I still see only the
DHCPDISCOVER and DHCPOFFER messages. This is what the subnet's dhcp
configuration looks like:

subnet 192.168.0.0 netmask 255.255.255.0 {
range 192.168.0.2 192.168.0.50;
range dynamic-bootp 192.168.0.51 192.168.0.55;
option broadcast-address 192.168.0.255;
option routers 192.168.0.1;
}

thank you,
K

--
http://astroturfgarden.com

Some routing equipment doesn't like a zero octet. Try changing to
192.168.1.x instead.

My Cisco gear needs a special configuration parameter to acknowledge a
zero octet. Not sure why, but I think it's either a bug in the
original tcp stack or an omission in the spec.


Kevin T. Neely wrote:
> On 2006-07-18, Ken Roberts <(E-Mail Removed)> wrote:
> > Try enabling BOOTP to see if that works.
> >
> > I've noticed a few network appliances that need BOOTP, which is a pain
> > because I don't want that enabled and can't have it in some cases.
>
> Thank you for the response. I tried adding a bootp range to my
> /etc/dhcpd.conf file, but it does not seem to have worked. Watching
> the syslog after powering up the WET11, I still see only the
> DHCPDISCOVER and DHCPOFFER messages. This is what the subnet's dhcp
> configuration looks like:
>
> subnet 192.168.0.0 netmask 255.255.255.0 {
> range 192.168.0.2 192.168.0.50;
> range dynamic-bootp 192.168.0.51 192.168.0.55;
> option broadcast-address 192.168.0.255;
> option routers 192.168.0.1;
> }
>
> thank you,
> K
>
> --
> http://astroturfgarden.com

On 2006-07-18, Ken Roberts <(E-Mail Removed)> wrote:
> Some routing equipment doesn't like a zero octet. Try changing to
> 192.168.1.x instead.
>
> My Cisco gear needs a special configuration parameter to acknowledge a
> zero octet. Not sure why, but I think it's either a bug in the
> original tcp stack or an omission in the spec.
>

Still no dice. I did this:

subnet 192.168.0.0 netmask 255.255.254.0 {
range 192.168.1.2 192.168.1.50;
range dynamic-bootp 192.168.1.51 192.168.1.55;
option broadcast-address 192.168.1.255;
option routers 192.168.0.1;
}


Do I need to go "all-in" and change the subnet, as well? I just
extended the netmask and put all the DHCP clients on the the x.x.1.x
half of the subnetwork. I did not want to change all my router and
network settings, and was afraid that if I only changed the DHCP
settings, the client would not be able to reach the router and would
still fail, albeit for a different reason.

thanks,
K


--
http://astroturfgarden.com

If your router doesn't work well with zero octets, that would mean ALL
zero-octet addresses would need to go away, where the zero is in the
network number.

In other words, change all occurrences of 192.168.0 to be 192.168.1.
You'd have to change the router first, so you can get to it after your
other gear switches over.

Keep in mind that this might not be your problem. Some gear has that
problem, other gear does not. As I said earlier, I think it's a bug.

Kevin T. Neely wrote:
> On 2006-07-18, Ken Roberts <(E-Mail Removed)> wrote:
> > Some routing equipment doesn't like a zero octet. Try changing to
> > 192.168.1.x instead.
> >
> > My Cisco gear needs a special configuration parameter to acknowledge a
> > zero octet. Not sure why, but I think it's either a bug in the
> > original tcp stack or an omission in the spec.
> >
>
> Still no dice. I did this:
>
> subnet 192.168.0.0 netmask 255.255.254.0 {
> range 192.168.1.2 192.168.1.50;
> range dynamic-bootp 192.168.1.51 192.168.1.55;
> option broadcast-address 192.168.1.255;
> option routers 192.168.0.1;
> }
>
>
> Do I need to go "all-in" and change the subnet, as well? I just
> extended the netmask and put all the DHCP clients on the the x.x.1.x
> half of the subnetwork. I did not want to change all my router and
> network settings, and was afraid that if I only changed the DHCP
> settings, the client would not be able to reach the router and would
> still fail, albeit for a different reason.
>
> thanks,
> K
>
>
> --
> http://astroturfgarden.com