[Openstack-operators] Using Openstack-ansible to deploy hosts on mutliple br-vxlan/br-vlan/br-provider/br-mgmt l3 networks
bpetit at b0rk.in
bpetit at b0rk.in
Tue Nov 22 15:08:24 UTC 2016
On 2016-11-17 11:52, Jean-Philippe Evrard wrote:
> I’ll start by answering the last questions: “Is that kind of
> deployment possible? If not possible, is this or could this be in the
> roadmap of the project?”
> It’s been a while that we discussed the limitations of the
> user-friendliness of our inventory. So, yes it’s part of our long
> term road map, and we discussed it at the summit.
> I don’t expect the inventory to be radically changed in this cycle.
Okay. That's not a problem itself for me. My concerns are only about
the resulting network configuration.
> I think (never tried) it’s technically possible to do what you are
> asking for. But it’s gonna involve a few hacks in the inventory.
That's what I thought (based on my previous tests). The question is for
me: are those hacks going to be problem in a production environment ?
> This is worth a complete conversation on the channel, because it’s
> hard to fully understand your requirements with a brief email.
I will. Thanks for the advice.
> Anyway, let me give you some hints…
> - You could technically assign whatever IP you require in your
> compute/controller nodes, as long as they are properly ssh reachable
> from your deploy node.
Yes, it works.
> - I think you could define many cidr_networks, and change the
> ip_from_q per site (instead of having global_overrides, you’d have
> to define override per host/group of hosts)
I'll test that one.
> - Wherever you place your load balancer, make sure it’s reachable
> from everywhere.
> - Please keep in mind that you most likely (depending on your
> multicast routing configuration?) won’t be able to use haproxy +
> keepalived co-located on your controller nodes, because we are using
> multicast for VRRP. That means that if you have another load balancer,
> or are ready to configure keepalived and haproxy differently, you
> could do it.
> - It’s already possible to inject routes for containers in OSA:
> A simpler approach would be to build overlays between the sites. It
> completely depends on your environment/requirements.
> Hope it helps.
It does. Thank you ! And see you on the irc channel !
> Best regards,
> Jean-Philippe Evrard (evrardjp)
> Rackspace Limited is a company registered in England & Wales (company
> registered number 03897010) whose registered office is at 5 Millington
> Road, Hyde Park Hayes, Middlesex UB3 4AZ. Rackspace Limited privacy
> policy can be viewed at www.rackspace.co.uk/legal/privacy-policy -
> This e-mail message may contain confidential or privileged information
> intended for the recipient. Any dissemination, distribution or copying
> of the enclosed material is prohibited. If you receive this
> transmission in error, please notify us immediately by e-mail at
> abuse at rackspace.com and delete the original message. Your cooperation
> is appreciated.
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
More information about the OpenStack-operators