Open Stack

We've been having a persistent problem with our OpenStack (Essex)
cluster.  We are automatically assigning floating ips when systems are
created (auto_assign_floating_ip = True).  When a system boots,
neithre the command line tools nor Horizon seem to know about the
automatically assigned ip address for several minutes (possibly more
than 10 or 15) after the system boots.

The system demonstrably has a floating ip address assigned (if you
initiate an outbound connection from the system, or inspect the
iptables nat rules, you can determine that address and use it to
connect to the system).

Manually assigning a floating ip address will force things to update
(so after manually assigning a floating address you'll see the fixed
address, the automatically assigned address, and the manually assigned
address).

We're running the 2012.1.3 release of things; I've read at least one
bug report that seems to describe this issue that implies the fix
should already be in this release...but we're still having this
problem.

Has anyone else encountered this problem?  Were you able to solve it?
A fix would be great, because right now our documentation is basically
"start an instance...then go do something else for 30 minutes."

-- 
Lars Kellogg-Stedman <lars at seas.harvard.edu>  |
Senior Technologist                           | http://ac.seas.harvard.edu/
Academic Computing                            | http://code.seas.harvard.edu/
Harvard School of Engineering                 |
  and Applied Sciences                        |