> Just re-tested in lab the principle behind it using the configuration,
> so without the OMAPI :
> host testing {
> hardware ethernet 00:11:22:33:44:55;
> default-lease-time 30;
> max-lease-time 30;
> }
>
> The device effectively gets a 30 seconds lease.
>
> The switching the configuration to :
> host testing {
> hardware ethernet 00:11:22:33:44:55;
> default-lease-time 300;
> max-lease-time 300;
> }
>
> Makes it that when the device renews, it gets a 5 minute lease.
>
> I will try out the class approach you defined and report if it works.
> I may need help for that too, but I'll at least start by giving it a try.
>
> As for your idea of placing people in another VLAN with a different
> pool configuration, this is how we do it currently but this has the
> downside of creating extra work, especially when doing it after the
> initial deployment phase of the NAC solution.
>
> Thanks !
>
> On 03/04/2016 02:50 AM, Simon Hobson wrote:
>> Julien <
[hidden email]> wrote:
>>
>>> Now, the goal of this is not to affect the current lease but to
>>> affect future leases.
>> I guessed that
>>
>>> Maybe a bit of context will help.
>>>
>>> We have a network which is used for captive portal registration,
>>> when you are done registering, you leave this network (through a
>>> VLAN change).
>>> The lease time is set to 30 seconds so that the device tries to
>>> renew its IP in the new network fast (and then fail, and get the
>>> proper IP)
>>>
>>> Now, some devices connect to the network and stay there for days,
>>> months without doing anything else than getting DHCP and without
>>> trying to register.
>>> We would like these devices to have a lease length that is higher
>>> than the other ones so that it reduces the load on the DHCP server.
>>>
>>> This above is determined at runtime since the detection occurs after
>>> the DHCP server has started, thus the need for it to occur
>>> dynamically through the OMAPI or another mean.
>> As I wrote, I'm not sure that this will help. I'd suggest setting up
>> a test lab and try it - my feeling is that when a client renews, the
>> lease length it's given will depend on the rules for assigning lease
>> lengths, not the length of it's current lease. So when you've
>> lengthened a lease, the client renews, the longer lease is "forgotten
>> about" and a replacement is done for 30s again.
>>
>> Could you use classes instead. Assign these devices to a specific
>> class (via OMAPI) and configure that class to have longer leases ?
>> Or perhaps re-assign the client to a "never never land" VLAN with
>> different lease length rules (but still stuck behind the registration
>> portal) ?
>>
>> _______________________________________________
>> dhcp-users mailing list
>>
[hidden email]
>>
https://lists.isc.org/mailman/listinfo/dhcp-users>
> _______________________________________________
> dhcp-users mailing list
>
[hidden email]
>
https://lists.isc.org/mailman/listinfo/dhcp-users