On 26 November 2012 20:38, Gary Kotton <span dir="ltr"><<a href="mailto:gkotton@redhat.com" target="_blank">gkotton@redhat.com</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF" text="#000000">If I understand correctly the IP address of the DHCP server is
passed by Nova to the VM. Which IP address will this be? If a load
balancer is used the address can be the same - that is a virtual IP.<br></div></blockquote><div><br>Hm, define 'passed'. Nova knows what it is for firewalling purposes. It's probably not so hard to change the DHCP address into a list of permitted addresses. The VM itself first receives the address as part of the response; it certainly shouldn't be passed as config data of any form.<br>
<br>Assuming the two DHCP servers back on to the same database (and obviously assuming the database itself is HA) then leases would be co-ordinated between the two even with LB at work. Generally we're not picking a random address out of a pool and promising not to give it to anyone else; we're providing an address we decided upon when the VM boot command was first received, and we're just retrieving it at this point (albeit from a local cache, as it's pushed into dnsmasq rather than retrieved from the DB on demand).<br>
-- <br>Ian.<br></div></div></div>