It is not a detailed diagram. It is just an attempt to explain the idea of having a bunch of drivers which all have the same stevedore interface. My main message is to avoid having "modes" and any other complicated structures in agent, but just plain pluggable drivers (any piece of functionality is implemented in an equal manner). The question appeared because the comments on etherpad made me think that some people'd like to have hard coded stages like prepare or deploy. That is why I wrote this letter. Vladimir Kozhukalov On Fri, Mar 21, 2014 at 9:52 PM, Jay Faulkner <jay at jvf.cc> wrote: > On 3/21/14, 10:18 AM, Vladimir Kozhukalov wrote: > > And here is scheme > > > https://drive.google.com/a/mirantis.com/file/d/0B-Olcp4mLLbvRks0eEhvMXNPM3M/edit?usp=sharing > > > > Vlamimir, can you recreate this drawing in a format that doesn't require > an additional browser plugin? Thanks. > > -Jay > > > > > _______________________________________________ > OpenStack-dev mailing list > OpenStack-dev at lists.openstack.org > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140322/c0c6e0c2/attachment.html> -------------- next part -------------- A non-text attachment was scrubbed... Name: Ironic agent driver architecture.png Type: image/png Size: 56953 bytes Desc: not available URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140322/c0c6e0c2/attachment.png>
