Thanks for the pointer. However, using those outputs to inject user data seems
to create a circular dependency:
myserver:
type: OS::Nova::Server
properties:
name: test-vm
image: ubuntu-16.04
flavor: m1.small
networks:
- network: { get_param: [ mgmt, id ] }
- network: { get_param: no_dhcp_net }
user_data_format: RAW
user_data:
str_replace:
template: |
#!/bin/bash
echo -e "\n" >> /etc/network/interfaces
echo "auto ens4" >> /etc/network/interfaces
echo "iface ens4 inet static" >> /etc/network/interfaces
echo "address $IPADDR" >> /etc/network/interfaces
echo "netmask 255.255.255.0" >> /etc/network/interfaces
ifdown ens4 2>/dev/null
ifup ens4 2>/dev/null
params:
$IPADDR: { get_attr: [ myserver, addresses, no-dhcp-net, 0, addr ] }
For now, I am going back to creating port and pass in that information as user
data.
-Kaustubh
From: Brent Troge [mailto:brenttroge2...@gmail.com]
Sent: Wednesday, July 20, 2016 7:27 PM
To: Kaustubh Kelkar <kaustubh.kel...@casa-systems.com>
Cc: openstack@lists.openstack.org
Subject: Re: [Openstack] [OpenStack][Nova] Configuring IP address on a DHCP
disabled network
Isn't that exposed through outputs ?
And then use the output combined with configdrive cloudinit to configure your
NICS
outputs:
instance_ip:
description: The IP address of the deployed instance
value: { get_attr: [my_instance, first_address] }
On Jul 20, 2016 5:58 PM, "Kaustubh Kelkar"
<kaustubh.kel...@casa-systems.com<mailto:kaustubh.kel...@casa-systems.com>>
wrote:
My use case involves booting up an instance attached to two networks – one
having Neutron DHCP service running and one without. Looking within the
documentation [1], I was hoping to use configuration drive to provide IP
information for the second interface. But, if I understand correctly, one could
theoretically pass any IP information through it; which may or may not match
the IP that is assigned to the Neutron port. I wonder if there is any way the
actual IP assigned to the Neutron port is exposed in some way within the VM.
One crude workaround I can think of is to create a port beforehand with a fixed
IP, and use this information to configure static networking within the VM.
However this might not be manageable if a user launches numerous VMs since
he/she will have to ensure that IPs do not overlap.
[2] looks promising, but I am not clear how I can just specify the second
interface only.
[1] http://docs.openstack.org/user-guide/cli_config_drive.html
[2] http://blog.oddbit.com/2015/06/26/openstack-networking-without-dhcp/
Thanks,
Kaustubh
_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to :
openstack@lists.openstack.org<mailto:openstack@lists.openstack.org>
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
