I'm not sure if I understand this correctly, but I believe we have a
smilar setup:
You have existing VLANs which have their own DHCP server and existing
instances running. And you want to add new instances having their
network interfaces in these VLANs, therefor you intend to disable
neutron's dhcp server for that network. But your statement
neutron subnet-create ext_net --gateway 10.35.1.254 10.35.1.0/24 --
--enable_dhcp=True
enables dhcp on that subnet, I think you want to use the
router:external option and disable dhcp:
neutron subnet-create ext_net --gateway 10.35.1.254 10.35.1.0/24
--disable-dhcp --router:external True
See [1] for details on external networks.
This way your new instances won't get any ip address and your existing
instances won't lose their network config.
Now if you have that setup, you can use --config-drive with nova boot,
the instance will receive its ip address that neutron has created.
I hope I got this right.
[1] http://docs.openstack.org/mitaka/networking-guide/scenario-classic-lb.html
Zitat von Satish Patel <satish....@gmail.com>:
I am planning to upgrade liberty to mitaka now so hope i will have
latest and greatest feature to solve this problem. why openstack
community doesn't have any straightforward solution.
When i was trying to use DHCP in openstack i found openstack DHCP
start provide ip address to my existing LAN machines ( we are using
flat VLAN with neutron), that is why i disable openstack DHCP, Is it
common or i am doing something wrong?
We have existing DHCP in out LAN i don't want openstack DHCP take over
and start assigning own IPs to Provide LAN.
neutron subnet-create ext_net --gateway 10.35.1.254 10.35.1.0/24 --
--enable_dhcp=True
On Thu, Aug 25, 2016 at 9:26 AM, Satish Patel <satish....@gmail.com> wrote:
Eugen,
I think config-drive make sense when you don't have initial network in
place. In my case i don't care about fixed-IP for instance. I only
need to setup network using whatever IP neutron provide in that case
how do i query neutron port to find out what IP address is available
or neutron going to provide so i can take that information and pass to
userdata. Its sounds tricky any idea how to do that?
On Thu, Aug 25, 2016 at 2:53 AM, Eugen Block <ebl...@nde.ag> wrote:
Hi,
we've been trying to learn how to feed cloud-init with ip
addresses, too. If
DHCP is disabled in your network, the instance won't get it's eth0
configured and won't be able to query the metadata server.
Creating a port before attaching it to a booting instance also doesn't work
if no dhcp is running on that network, I just tried that to be sure.
I've tried several ways but I only found one working option. For external
networks (or networks without dhcp) we are using config-drive now.
Depending
on the OpenStack version it could be possible that you'll need
cloud-init-0.7.7, we had to fix two issues ourselves in version
0.7.6 to get
it working, one of them was a missing default route.
With enabled config-drive the instance doesn't need a configured interface,
it's a temporarily mounted drive from where the required
information is read
by cloud-init.
You can either add the option "--config-drive true" in your nova boot call
or check the checkbox in Horizon.
To answer your question about ports, you can create a new port either in
Horizon, but there you won't be able to assign a specific ip
address. If you
want a specific ip address you have to call neutron port-create (port-name
is optional):
neutron port-create <NETWORK-ID> --fixed-ip
subnet_id=<SUBNET-ID>,ip_address=<IP> --name <PORT-NAME>
The resulting ID of that port can be used in nova boot call:
nova boot --flavor 2 --image <IMAGE> --nic port-id=<PORT-ID>
<INSTANCE-NAME>
Another way to assign a specific ip address to a booting instance without
port-creation (but DHCP has to be enabled) would be:
nova boot --flavor 2 --image <IMAGE> --nic
net-id=<NET-ID>,v4-fixed-ip=<IP> <INSTANCE-NAME>
for example:
nova boot --flavor 2 --image dc05b777-3122-4021-b7eb-8d96fdab2980 --nic
net-id=4421e160-d675-49f2-8c29-9722aebf03b2,v4-fixed-ip=192.168.124.6 test1
Hope this helps!
Zitat von Satish Patel <satish....@gmail.com>:
My question is how to query ports and pass info to cloud-init? is
there any document or api which i can call using script and setup
network ifcfg-eth0 file
On Wed, Aug 24, 2016 at 5:38 PM, Kaustubh Kelkar
<kaustubh.kel...@casa-systems.com> wrote:
You can create the ports beforehand and plug them in while creating the
instance. As for assigning IP addresses, you can query the ports and pass
the information to cloud-init. I am not sure if there is any
other way to do
this.
Even if DHCP is disabled, OpenStack assigns IP information to ports when
a VM is created, and you can see this in your dashboard. The MAC and IP
information is used to configure iptables rules within security
groups. Here
is the archived thread that provides this information:
http://lists.openstack.org/pipermail/openstack-dev/2014-December/053069.html.
-Kaustubh
-----Original Message-----
From: Satish Patel [mailto:satish....@gmail.com]
Sent: Wednesday, August 24, 2016 5:05 PM
To: James Downs <e...@egon.cc>
Cc: openstack <openstack@lists.openstack.org>
Subject: Re: [Openstack] Guest VM IP configuration script
I am using neutron networking with vlan ( its provider VLAN). We are not
using
DHCP but i need some kind of hack to inject IP address in instance using
cloud-
init.
We are using cloud-init but i don't know how does it work and get IP
from
neutron. I am new with neutron stuff.
On Wed, Aug 24, 2016 at 4:29 PM, James Downs <e...@egon.cc> wrote:
> On Wed, Aug 24, 2016 at 03:25:26PM -0400, Satish Patel wrote:
>> I enabled following in nova.conf on compute node but didn't work :(
>>
>> flat_injected=true
>>
>> Do i need to do anything else?
>
> Are you using flat networking?
> Nova-networks or Neutron?
>
> At this point, if you're not using DHCP, your only option is to
> arrange to feed
the networking information into the metadata for the VM at creation
time, and
use someting like cloud-init to configure the networking. The ancient
networking injection stuff has either been removed, or been broken for
years.
>
> Cheers,
> -j
>
> _______________________________________________
> 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
_______________________________________________
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
_______________________________________________
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
--
Eugen Block voice : +49-40-559 51 75
NDE Netzdesign und -entwicklung AG fax : +49-40-559 51 77
Postfach 61 03 15
D-22423 Hamburg e-mail : ebl...@nde.ag
Vorsitzende des Aufsichtsrates: Angelika Mozdzen
Sitz und Registergericht: Hamburg, HRB 90934
Vorstand: Jens-U. Mozdzen
USt-IdNr. DE 814 013 983
_______________________________________________
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
--
Eugen Block voice : +49-40-559 51 75
NDE Netzdesign und -entwicklung AG fax : +49-40-559 51 77
Postfach 61 03 15
D-22423 Hamburg e-mail : ebl...@nde.ag
Vorsitzende des Aufsichtsrates: Angelika Mozdzen
Sitz und Registergericht: Hamburg, HRB 90934
Vorstand: Jens-U. Mozdzen
USt-IdNr. DE 814 013 983
_______________________________________________
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
