SV: DISCOVER bursts

[Glenn Satchell]

Maybe after sending an offer, the server won't re-use that address
until it either times out in 2 minutes, or it gets a DHCPREQUEST for
that address? Even if it is the same client - perhaps in some cases it
might be legal for the client to request several addresses, eg like a
microsoft RAS server? (As an aside does a MS-RAS use different
client-ids for each address it requests, I don't know?)

Once it has committed the 2 minute lease, that address is in a sort-of
pending state waiting for a request and can't be re-used?

regards,
-glenn

>Subject: RE: SV: DISCOVER bursts
>Date: Tue, 9 Jan 2007 12:37:23 +0100
>From: <Staffan.Ungsgard at teliasonera.com>
>To: <dhcp-users at isc.org>
>
>My original analysis was somewhat off, but my race theory still remains.
>
>This is what happened:
>
>The client did a discover and got an offer. In respons to the offer the client 
erroneously answered with a new discover.
>This second discover arrived 20 milliseconds after the first discover. The dhcp 
server regarded this a totally new discover.
>I have sniffed the traffic and diffed the discovers - they are indeed the same.
>It does seem like the server still at the second discover has recorded the 
original one and hence offers a new address.
>
>I know it's single threaded - this is very strange.
>
>-- Staffan
>
>-----Original Message-----
>From: dhcp-users-bounce at isc.org [mailto:dhcp-users-bounce at isc.org] On Behalf Of 
Simon Hobson
>Sent: den 9 januari 2007 12:21
>To: dhcp-users at isc.org
>Subject: Re: SV: DISCOVER bursts
>
>Thinking about this thread a bit more, if the server is single threaded 
shouldn't that mean that it's not possible for it to respond to another client 
request until it's handled the first - and hence created the lease record ?
>
>If that is the case, I'm not sure how the server could be handing out different 
addresses to the same client if all it's requests have the same mac/client-id.
>
>