Open Stack

Okay thanks for that, a few moments of darkness is fine.

Is this right :

On the existing and new network nodes

1. Apply this patch to Quantum: https://gist.github.com/jaypipes/8135839
2. Set the router_scheduler_driver in nova.conf to
quantum.scheduler.l3_agent_scheduler.LeastUtilizedScheduler
Reference:https://github.com/stackforge/cookbook-openstack-network/blob/mas
ter/templates/default/neutron.conf.erb#L232
3. Put this script into a cron job:
https://github.com/stackforge/cookbook-openstack-network/blob/master/files/
default/neutron-ha-tool.py

What do I need to do to make the DHCP and Metadata agent schedulable ?

"Depending on if you use namespaces or not you will have to make sure the
router_id in the l3_agent.ini on the new node is properly set to the same
id)² I am assuming if I silly SCP the configs from the first network node
to the second this will take care of that ? (We don¹t use namespaces
anyhow)


-- 
Geraint Jones

Director of Systems & Infrastructure
Koding 
https://koding.com
geraint at koding.com
Phone (415) 653-0083





On 17/02/14 8:06 pm, "Robert van Leeuwen"
<Robert.vanLeeuwen at spilgames.com> wrote:

>> We currently have a single network node in our deployment, we need to
>>upgrade it to 10gb
>> So I thought, I would add a second network node and do a swing
>>migration and thus avoid an outage.
>>
>> Is this possible in havana ? Whats the process to do it ?
>
>Yes and no.
>Yes, you can migrate the l3 node to a different machine but no, it will
>not be 100% instantaneous.
>Depending on how many networks are connected to the l3 agent it will take
>some time for each of those to be up & running although it will probably
>be running again within a minute.
>I would certainly not do this without announcing some kind of maintenance
>window.. 
>
>The steps depend a bit on the configuration.
>I would not do this without doing a test with a dev setup first so you
>get comfortable with how this works.
>Do not forget that the l3 agent is also NATing for floating IPs and the
>metadata agent so make sure that also keeps working.
>( usually it entails running "neutron l3-agent-router-remove $AGENTID
>$ROUTERID"  which should be schedule the router on a new l3 node.
>Depending on if you use namespaces or not you will have to make sure the
>router_id in the l3_agent.ini on the new node is properly set to the same
>id)
>
>Cheers,
>Robert van Leeuwen