[Openstack-operators] HA solutions for L3-Agent and metadata-agent ?

Henry Gessau gessau at cisco.com
Thu Mar 6 19:21:25 UTC 2014

On Thu, Mar 06, at 11:41 am Daneyon Hansen (danehans) (danehans at cisco.com)

> On 3/6/14 7:50 AM, "gustavo panizzo <gfa>" <gfa at zumbi.com.ar> wrote:
>>On 03/06/2014 09:54 AM, Alvise Dorigo wrote:
>>> Hi,
>>> I opened some time ago a thread about this argument (more in general
>>> about HA for
>>> OpenStack): 
>>> Now I'm definitely focusing on making the Neutron's agents highly
>>> available.
>>> As stated in the mentioned thread there's no solution for making L3
>>> agent HA with the active/active paradigm. But I have to use the
>>> active/active paradigm.
>>there is no active/active route protocol to route ip packages, you can
>>have only one default gw. that's true in virtual as physical world.
>>i remember a eNovance guy giving a presentation during HK summit, they
>>have patched neutron in order to use VRRP. you may want to search that
>>if you use it, pls share your results. i haven't a chance to test it
> I believe routing protocols and a first hop redundancy protocol are two
> separate, but complimentary HA topics.  Most routing protocols have the
> ability to load share across multiple links. If one of the links go down,
> the routing protocol will detect it and remove the path from its
> forwarding table.  The ability for the L3 agent to support a first hop
> redundancy protocol such as VRRP would be a great step in the right
> direction for HA.  Zookeeper could be another option (approach to solving
> Ceilometer's central agent HA issue. By using the Service Group API and a
> backend such as ZooKeeper.  If eNovance has this code, can they send it
> out for review? I know L3 agent HA has been a point of interest for a long
> time and the current limitations force HA deployments to use provider
> networking. 

Check the wiki meetings page for the Neutron L3 subteam:

See section 1.1 for the blueprints and code reviews for HA router work.
There is a weekly IRC meeting where these things are discussed.

>>> In the mentioned thread, Jay suggests to use a tool to replicate l3
>>> agent's status to the other node upon failure of the current node (I
>>> haven't used this tool yet).
>>> Reading the HA guide by Openstack
>>> there's not a solution active/active-based also for the metadata agent.
>>> I wonder if there's a similar tool to replicate the metadata agent's
>>> status on the other node, as for the l3 agent.
>>neutron's metadata agent is just a proxy to nova-api, it is stateless
>>1AE0 322E B8F7 4717 BDEA BF1D 44BB 1BA7 9F6C 6333
>>OpenStack-operators mailing list
>>OpenStack-operators at lists.openstack.org
> _______________________________________________
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators

More information about the OpenStack-operators mailing list