cannot change ip address in pool (fixed question)

[Simon Hobson]

Jan nowak top-posted:

>Sep 13 07:32:35 debian dhcpd: ICMP Echo reply while lease 172.16.20.203 valid.
>Sep 13 07:32:35 debian dhcpd: ICMP Echo reply while lease 172.16.20.203 valid.
>Sep 13 07:32:35 debian dhcpd: ICMP Echo reply while lease 172.16.20.203 valid.
>Sep 13 07:32:35 debian dhcpd: ICMP Echo reply while lease 172.16.20.203 valid.
>Sep 13 07:32:37 debian dhcpd: Reclaiming abandoned lease 172.16.20.203.
>Sep 13 07:32:37 debian dhcpd: Reclaiming abandoned lease 172.16.20.203.
>Sep 13 07:32:37 debian dhcpd: Reclaiming abandoned lease 172.16.20.203.
>Sep 13 07:32:37 debian dhcpd: Reclaiming abandoned lease 172.16.20.203.

Not seen that first message before, but basically it means the server 
has pinged the address and got a reply - when it shouldn't have. So 
it's marked the lease as abandoned - abandoned leases are only used 
as last resort if nothing else is available.

But this is a different address to the one of interest ...

>Sep 13 07:32:41 debian dhcpd: data: "leased-address" configuration 
>directive: there is no lease associated with this client.

That would be correct - the lease has been abandoned and there's 
nothing leased to the client.

>Sep 13 07:32:58 debian dhcpd: DHCPDISCOVER from 00:14:5e:7f:35:3c 
>via eth0: network xxx: no free leases



>Sep 13 07:32:58 debian dhcpd: DHCPDISCOVER from 00:14:5e:7f:35:3c 
>via eth0: network xxx: no free leases

This is what you expect to see when using restricted lease ranges and 
old leases are still active. There isn't anything for the server to 
assign to the client. Again, this isn't the client (MAC) in the 
config snippet.

It's not that helpful to see this unrelated log extract. If it's a 
network with too much DHCP traffic to extract only the relevant 
entries, then you probably need to either temporarily disable DHCP 
service for all but one client, or setup a test environment with only 
the one client.

>shared-network xxx{
>
>subnet 172.16.0.0 netmask 255.255.0.0 {
>next-server 192.168.0.1;
>option domain-name-servers 217.30.129.149, 217.30.137.200;
>option routers 172.16.0.1;
>option subnet-mask 255.255.0.0;
>option broadcast-address 172.16.255.255;
>}

Could I suggest better formatting would be a good idea ? Something 
like this would make the nesting more readily viewable :

shared-network xxx{

   subnet 172.16.0.0 netmask 255.255.0.0 {
     next-server 192.168.0.1;
     option domain-name-servers 217.30.129.149, 217.30.137.200;
     option routers 172.16.0.1;
     option subnet-mask 255.255.0.0;
     option broadcast-address 172.16.255.255;
   }

   subnet ...

This is especially true when we get to your host statements.


>#########################################
>host m69 { hardware ethernet 00:0a:73:f9:e6:3a; fixed-address 
>172.16.10.69; filename "<http://plik.cm>plik.cm";}
>class "M069" { match if binary-to-ascii(16, 8, ":", option 
>agent.remote-id) = "0:a:73:f9:e6:3a"; spawn with option 
>agent.remote-id; lease limit 1;}
>pool { allow members of "M069"; range 172.16.20.69; }
>##############################################
>}

Not only is this hard to read, it's mixing scopes as well.

host m69 {
   hardware ethernet 00:0a:73:f9:e6:3a;
   fixed-address 172.16.10.69;
   filename "plik.cm";}
   class "M069" {
     match if binary-to-ascii(16, 8, ":", option agent.remote-id) = 
"0:a:73:f9:e6:3a";
     spawn with option agent.remote-id; lease limit 1;
   }
   pool {
     allow members of "M069";
     range 172.16.20.69;
   }
}

So what *exactly* are you trying to achieve with this ?
You've correctly put the host statement in the global scope, but then 
you've defined a class and pool within it.
I'm not sure about the class (do these need to be in the global scope 
?), but the pool really belongs within the subnet.

And there should be no need to have both the class/pool AND a fixed 
address - either one or the other. So either :

host m69 {
   hardware ethernet 00:0a:73:f9:e6:3a;
   fixed-address 172.16.10.69;
   filename "plik.cm";}
}

or :

class "M069" {
   match if option agent.remote-id) = 0:a:73:f9:e6:3a;
   spawn with option agent.remote-id; lease limit 1;
}
and
   subnet 172.16.0.0 netmask 255.255.0.0 {
     next-server 192.168.0.1;
     option domain-name-servers 217.30.129.149, 217.30.137.200;
     option routers 172.16.0.1;
     option subnet-mask 255.255.0.0;
     option broadcast-address 172.16.255.255;
     pool {
       allow members of "M069";
       range 172.16.20.69;
     }
   }

Ahh, now I've split that up, it becomes clearer. 172.16.10.69 is for 
the device with MAC address 00:0a:73:f9:e6:3a and 172.16.20.69 is for 
the downstream device.
As you've defined it, I don't think the class would be matched 
because the definition (and pool statement) are within a host 
statement that will not apply to the downstream client.
By definition, the downstream client will not be the host with MAC 
address 0:a:73:f9:e6:3a, so the host statement will not be matches, 
and nothing inside it will be acted upon.

-- 
Simon Hobson

Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed
author Gladys Hobson. Novels - poetry - short stories - ideal as
Christmas stocking fillers. Some available as e-books.