[Openstack] nova ip

Howard Luckenbaugh hlucken at us.ibm.com
Mon May 12 14:01:12 UTC 2014

This is part of IBM Smart Cloud Orchestrator which is running openstack as
the backend.  I compared my IPtables to a working environment and it looks
fine. I have rebooted and restarted. The VM's are getting a 169 route but
can't ping.
The only place I can ping is on the Region server which is where the 169
address sits.

   Howard Luckenbaugh
   Special Events Infrastructure/IBM.com Infrastructure
   Fax- 1877-210-0298
   hlucken at us.ibm.com
   AIX System  Certified

   Please visit the sites that our team hosts: http://www.ibm.com
   http://www.wimbledon.com  http://www.usopen.org http://www.masters.com
   http://www.austrailianopen.com http://www.usopen.com

   "Vision without Action is just a dream
   Actions without Vision is just passing time
   Vision with Action can change the world"

From:	"Gangur, Hrushikesh (R & D HP Cloud)"
            <hrushikesh.gangur at hp.com>
To:	Robert Collins <robertc at robertcollins.net>, Howard
            Luckenbaugh/Raleigh/IBM at IBMUS
Cc:	"openstack at lists.openstack.org" <openstack at lists.openstack.org>
Date:	05/11/2014 11:29 PM
Subject:	RE: [Openstack] nova ip

Few troubleshooting steps (assuming neutron's metadata agent):
1. Check through VM's console.log whether VMs are acquiring DHCP Ip address
2. If not, check all the bridges (on both controller and compute) up and
running s (br-int or br-tun), generally on reboot some of these bridges do
not come up.
3. If yes to #1, restart neutron-metadata-agent

-----Original Message-----
From: Robert Collins [mailto:robertc at robertcollins.net]
Sent: Saturday, May 10, 2014 3:03 PM
To: Howard Luckenbaugh
Cc: openstack at lists.openstack.org
Subject: Re: [Openstack] nova ip

On 10 May 2014 06:16, Howard Luckenbaugh <hlucken at us.ibm.com> wrote:
> This box was working and listening on to access the
> metadata of the machines. However it has stopped listening. I have
> tried rebooting and comparing networks and etc and nothing I can see
> is stopping it.  Anyone had this happen before.

If you are using neutron's nova metadata agent, it runs a process per
network, then uses a domain socket to escape the network namespace and
communicate with a parent that talks to nova.

If you are using nova-networking, its handled by an iptables DNAT rule to
take incoming requests and forward them to the actual API IP address/port.


Robert Collins <rbtcollins at hp.com>
Distinguished Technologist
HP Converged Cloud

Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to     : openstack at lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20140512/0e27893e/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: graycol.gif
Type: image/gif
Size: 105 bytes
Desc: not available
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20140512/0e27893e/attachment.gif>

More information about the Openstack mailing list