Wrong gateway when using class in a subnet

[Glenn Satchell]

Hi Adam

That is perfectly valid, and is the typical way classes are meant to be 
used.

The earlier example was defining the class inside the subnet, so in that 
case the class inherits the subnet settings (like option routers) even 
when it is used in a different subnet or scope.

You can also put statements inside the class definition and those 
settings will apply to members of that class.

Note that once you allow a class in a pool, all other requests are 
denied from that pool. Likewise if you deny a class, then all others are 
allowed. So you will probably want two pools (at least) in each subnet, 
one where you allow EndUserStuff; and one where you deny it, to allocate 
addresses for all the other devices.

regards,
-glenn

On 06/02/12 05:37, Adam Moffett wrote:
> Sorry to post again so soon, but I'm still trying to wrap my head around
> the full implications of my misunderstanding about classes.
>
> Can I use the same class in multiple shared networks and expect to get
> the right IP out to somebody?
>
> Example:
>
> class "EndUserStuff" {
> match if not substring(hardware,1,3) = 00:04:f2;
> }
>
> shared-network "A" {
> [...]
> subnet 204.80.232.0 netmask 255.255.255.0 {
> [...]
> pool {
> allow members of "EndUserStuff";
> }
> }
> }
>
> shared-network "B" {
> [...]
> subnet 174.47.201.0 netmask 255.255.255.0 {
> [...]
> pool {
> allow members of "EndUserStuff";
> }
> }
>
> If that's expected to do the right thing, then it would tremendously
> simplify my dhcpd.conf.
>