[Openstack-operators] nova-network in all compute nodes?????
Sergio Ariel de la Campa Saiz
sacampa at gmv.com
Wed May 23 08:12:10 UTC 2012
I have been reading the conversation between Lorin and Christian about nova-network HA, and I have a question that maybe one of them can answer me... or maybe any member of this group :-) :-)
This is my environment:
2 ubuntu servers, essex and VLAN networking:
- 1 controller with nova-network, glance, keystone, nova-api, nova-scheduler but NOT nova-compute
- 1 node: nova-compute only
In this configuration all my VM are going to be launched in the node since there is no nova-compute running in my controller. When I try to launch a VM appears a problem whith my network. I think that is because there is no nova-network installed in the node. Bridge and VLAN interfaces are created in the controller but they are not created in the node.
My question is: it is necessary to install nova-network in all compute nodes??
Thanks a lot.
de la Campa Saiz
GMV-SES Infraestructura /
Isaac Newton, 11
P.T.M. Tres Cantos
+34 91 807 21 00
+34 91 807 21 99
De: openstack-operators-bounces at lists.openstack.org [openstack-operators-bounces at lists.openstack.org] En nombre de Lorin Hochstein [lorin at nimbisservices.com]
Enviado el: lunes, 21 de mayo de 2012 0:32
Para: Christian Parpart
CC: openstack-operators at lists.openstack.org
Asunto: Re: [Openstack-operators] Nova (in VLAN mode) and the high availability of a "central" nova-network
Have you seen the HA docs? http://docs.openstack.org/trunk/openstack-compute/admin/content/existing-ha-networking-options.html
If you set multi_host=True in your nova.conf, then nova will work properly with multiple nova-network services. But I've only done this with one nova-network service running on each compute host, I don't know how/if it would work if you had only two nova-network services running.
Lead Architect - Cloud Services
Nimbis Services, Inc.
On May 20, 2012, at 5:57 PM, Christian Parpart wrote:
due to the fact, that I cannot provide all compute nodes with public IPs (allowing us
to run nova-network on each compute node), we're having one (or two in fact) frontend
node doing the public network traffic work.
So, how can I make a single nova-network node highly available?
It would be fine to have some vrrpd/keepalived enabled node, that then
starts/stops the sensible services (like nova-network) in question.
However, does this work well with nova-network?
Will this produce some noticable delay in availability when the active master goes down
and until the current slave has fully become the master (with nova-network being started up)?
How does this affect the nova environment.
Especially, thinking of the `nova-manage service list` output, where only one nova-network instance
ought to be then, but in HA mode, we are to have at least 2 then. How well does Nova play with that?
Thanks in advance,
Openstack-operators mailing list
Openstack-operators at lists.openstack.org
This message including any attachments may contain confidential
information, according to our Information Security Management System,
and intended solely for a specific individual to whom they are addressed.
Any unauthorised copy, disclosure or distribution of this message
is strictly forbidden. If you have received this transmission in error,
please notify the sender immediately and delete it.
Este mensaje, y en su caso, cualquier fichero anexo al mismo,
puede contener informacion clasificada por su emisor como confidencial
en el marco de su Sistema de Gestion de Seguridad de la
Informacion siendo para uso exclusivo del destinatario, quedando
prohibida su divulgacion copia o distribucion a terceros sin la
autorizacion expresa del remitente. Si Vd. ha recibido este mensaje
erroneamente, se ruega lo notifique al remitente y proceda a su borrado.
Gracias por su colaboracion.
More information about the Openstack-operators