host-identifier and option 82

[Michal Vančo]

The bigest "advantage" of using static host decls (in my opinion) is that you can create multiple entries with the same IP address and different MAC. Client then will be free to change MACs and still be bound to single IP without any intervention (imagine an ISP which provides "static" IP addresses for its clients - actually, this is how I'm using it now).

I want to achieve the same goal by replacing MACs with "host-identifier option agent.remote-id" in order to not force clients to register their MACs by calling helpdesk. The only problem I've seen so far is that the client fails to renew its IP after T1 passes (because RENEWING DHCPREQUEST is sent directly to server bypassing agent relay). If I understand the RFC correctly, client should "succeed" in REBIND (after T2 passes) because REBINDING DHCPREQUEST is broadcasted hence seen by relay agent. Relay agent shall add opt 82 into DHCPREQUEST in this stage. With properly inserted opt 82, server should be able to match static host decl again.

Is this the way it is supposed to work? Feel free to correct me.

thanks
michal

On 27.1.2012, at 12:06, David Forrest wrote:

> On Fri, 27 Jan 2012, Michal Vančo wrote:
> 
>> Hi Glen,
>> 
>> yes I know about stash-agent-options flag. And I also know what its purpose is.
>> Despite this fact, what happens when you restart DHCP server? What I know is, that no information about
>> static leases (static host entries) is written into lease file. Hence server will forget all stashed agent options
>> and the client will not be able to renew its address. This is what I probably see when my
>> logfile starts to fill with "unknown lease" messages after I restart server.
>> 
>> regards
>> -michal
> 
> Some time ago I had problems with the treatment of static leases and started using the "reserved" statement in the problematical assignment added to the leases file.  .  It did take one of the pool addresses up permanently, but so did the static assignment anyway.  For my static servers, it hasn't caused a problem and it allows whatever queuing/aliasing to occur naturally.  Such a tactic might work for you as it has to be done on a closed dhcpd anyway.  I added the rDNS and also added the option to turn allow updates off as it failed anyway because it existed.
> 
> Dave Forrest St. Louis, Missouri_______________________________________________
> dhcp-users mailing list
> dhcp-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/dhcp-users