14:03:31 <enikanorov__> #startmeeting neutron lbaas
14:03:32 <openstack> Meeting started Thu Dec 12 14:03:31 2013 UTC and is due to finish in 60 minutes.  The chair is enikanorov__. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:03:33 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:03:35 <openstack> The meeting name has been set to 'neutron_lbaas'
14:03:59 <enikanorov__> #topic announcements
14:04:21 <enikanorov__> yesterday agent unification patch was merged
14:05:01 <enikanorov__> which means that any vendor plugin driver that want's to use agent should do it with our framework
14:05:40 <enikanorov__> HAProxy driver is the one and only plugin driver at the moment, which utilizes that
14:05:58 <evgenyf> Hi All
14:06:09 <enikanorov__> but it should be fairly easy now to ad drivers for similar solutions like nginx
14:06:13 <enikanorov__> hi evgenyf
14:07:03 <enikanorov__> any questions on unified agent?
14:08:07 <enikanorov__> ok, lets move on to the next topic
14:08:26 <enikanorov__> #topic third party testing
14:08:52 <songole> Would the agent be doing appliance VM instantiation in future?
14:09:09 <enikanorov__> songole: it is possible
14:09:47 <enikanorov__> ok, so back to the testing, we have 2 tempest scenarios on review that we hope will land soon
14:10:07 <enikanorov__> they already could be used to perform end-to-end testing of the service
14:10:26 <enikanorov__> and actually helper us to identify the issues with our patches
14:10:53 <enikanorov__> does anyone want to update on integration with jenkins? evgenyf?
14:11:21 <Vijay_> I have Eugene
14:11:34 <Vijay_> One question regardin Testing infra setup. We were able to subscribe to events inspect the files on the patch submitted.
14:11:37 <enikanorov__> the deadline for this was I-2, but i don't see much activity on vendor side
14:11:41 <evgenyf> We are working on OpenStack testing env. establishment
14:11:56 <evgenyf> Will update about the Jen. integration in a minute
14:12:49 <enikanorov__> Vijay_: so what's the question?
14:13:07 <Vijay_> Will just refreshing fetching latest Neutron code and re-starting neutron service will do to refresh the code
14:13:40 <Vijay_> or is it advisable to run "stack.sh" and do the setup again?
14:14:10 <enikanorov__> Vijay_: yeah, stack.sh should be run. fetaching and restarting the service is not sufficient
14:14:38 <Vijay_> We are halfway through, next week we will have more progress.
14:14:44 <enikanorov__> good
14:14:47 <obondarev> BTW are you guys aware of today's meeting on third party testing at 17 UTC?
14:14:54 <evgenyf> enikanorov__, What you mean by integration with Jenkins?
14:15:01 <obondarev> I believe similar questions will be discussed there
14:15:22 <enikanorov__> evgenyf: your system should post back the testing result, as far as i understand the requirement
14:15:23 <Vijay_> ok. will try to join. Thanks obondarev.
14:16:08 <evgenyf> We are not there yet, Working on environment and running tests
14:16:43 <enikanorov__> ok, good
14:17:42 <enikanorov__> did you try those tempest scenarios in your lab yet?
14:18:13 <Vijay_> We will try it next week and update.
14:18:19 <enikanorov__> ok
14:18:54 <evgenyf> enikanorov__, Not yet. BTW, as far as I know, we should push results back to gerrit, right? not Jenkins.
14:18:54 <enikanorov__> any queestions on testing?
14:19:17 <enikanorov__> evgenyf: oh right, mistyped my thought :)
14:19:33 <evgenyf> enikanorov__, ok , good
14:20:15 <Vijay_> No we can move on
14:20:21 <enikanorov__> ok
14:20:36 <enikanorov__> #topic L7 rules
14:20:52 <enikanorov__> evgenyf: any update?
14:21:01 <enikanorov__> i saw avishay has pushed some patch on gerrit, but have not yet looked at it
14:21:28 <evgenyf> enikanorov__, Avishay will not join today. He asked me to update that he pushed first patch, inluding data layer modifications
14:22:14 <enikanorov__> i see. in order to add this feature we'll need consistent patches: API change+ model change+ migration+ ut's
14:22:25 <enikanorov__> so it seems like lot of work yet to be done
14:22:40 <evgenyf> OK. Will tell him
14:23:07 <enikanorov__> also there should be cli for that and tempest API tests (at least)
14:23:50 <enikanorov__> so any major feature require lots of work...
14:23:55 <enikanorov__> ok, lets moe to the next one then
14:23:58 <enikanorov__> loadbalancer instance
14:24:24 <enikanorov__> #topic loadbalancer instance
14:24:56 <enikanorov__> the initial implemetation is on review. the patch is not fully ready, but it passes gate tests including tempest API tests
14:25:19 <enikanorov__> which means that new API is fully backwards compatible
14:25:27 <Vijay_> cool!
14:25:28 <enikanorov__> cli-side is also on the gerrit
14:25:48 <enikanorov__> now i'm working on tempest API tests for the new resource
14:26:11 <enikanorov__> that's all on loadbalancer instance at the moment
14:26:15 <enikanorov__> any questions?
14:27:23 <enikanorov__> ok
14:27:42 <enikanorov__> lets move to SSL
14:27:42 <s3wong> enikanorov: this is the reivew, right? https://review.openstack.org/#/c/60207/
14:27:55 <enikanorov__> #topic SSL termination
14:28:14 <enikanorov__> s3wong: right
14:28:54 <evgenyf> Vijay commented my WIKI updates
14:29:20 <evgenyf> Modified it accordingly. Waiting for you review guys
14:29:28 <enikanorov__> ok
14:30:00 <evgenyf> Will try to catch Mark just to be sure the direction is good
14:30:19 <evgenyf> All these new entities etc..
14:30:32 <enikanorov__> ok, good
14:31:02 <enikanorov__> would be good to see any implementation proposal in I-2
14:31:05 <Vijay_> Hi Evg, default policy is not a UI only feature right?
14:31:14 <enikanorov__> or most certainly it will not land in icehouse
14:31:47 <evgenyf> Vijay, No
14:31:52 <Vijay_> ok thanks!
14:31:55 <evgenyf> I mean, right:)
14:32:09 <Vijay_> ?? :-(
14:32:28 <evgenyf> I'm working on it. Will push initial work soon.
14:32:46 <enikanorov__> ok, cool
14:32:48 <evgenyf> Just want to be sure Mark is aware of the design
14:33:00 <enikanorov__> yeah, that is a good approach
14:33:24 <evgenyf> Vijay, It's not onlu UI feature
14:33:46 <enikanorov__> ok, another topic i would like to discuss is a various HA proposals we had
14:33:51 <enikanorov__> obondarev: could you update?
14:33:55 <obondarev> sure
14:34:03 <obondarev> HA for agents is on review
14:34:12 <obondarev> https://review.openstack.org/#/c/59743/
14:34:29 <obondarev> still struggling with random unit test failures
14:34:41 <obondarev> there is a race I guess
14:35:06 <obondarev> need to add some more unit tests and fix the race
14:35:26 <Vijay_> Evg: In that case, lets add the default policy description in the "Resources change" title in the place where we are describing "SSL policy "
14:35:35 <obondarev> this patch only works for plugins that are not host specific
14:35:38 <enikanorov__> i still think the design could be different for the case when agent manages the device and not the process
14:36:23 <obondarev> enikanorov__: can you please clarify?
14:36:54 <enikanorov__> as we discussed, it could be solved with a common queue instead of scheduling
14:36:56 <evgenyf> Vijay, will do
14:37:19 <obondarev> enikanorov__: yeah, I put some thoughts on it to the bp description
14:37:21 <enikanorov__> I'd sent an email to oos-dev wtih those two options
14:37:30 <enikanorov__> and ask for opinions
14:37:50 <obondarev> there is a common problem with common queue
14:37:59 <enikanorov__> which one?
14:38:02 <obondarev> not specific for lbaas even
14:38:25 <obondarev> that some action should be processed consequntly
14:38:34 <obondarev> actions*
14:38:43 <enikanorov__> can you give an example?
14:39:03 <obondarev> with many agents listening to one queue there might be races
14:39:24 <enikanorov__> not really
14:39:24 <obondarev> not sure I'm ready with an example now
14:39:37 <enikanorov__> it is not different as with 1 agent
14:39:43 <obondarev> we discussed it with sgran
14:39:54 <enikanorov__> which processes messages in multithreading fashion
14:40:41 <enikanorov__> ok, lets discuss it offline
14:41:01 <obondarev> ok , I agree second approach is possible, need to collect pros and cons
14:41:35 <enikanorov__> ok, another topic is ha for haproxy
14:41:39 <obondarev> yeah
14:41:48 <obondarev> not started yet thiugh
14:42:05 <enikanorov__> do you have a design in mind?
14:42:08 <obondarev> I look toward haproxy + keepalived approach
14:42:22 <s3wong> I can look into HA for haproxy, since it should be done via haproxy + keepalived
14:42:53 <obondarev> that is what I was plannin to start tomorrow or next week
14:43:05 <enikanorov__> ok, no hurry i think
14:43:36 <enikanorov__> but anyway, keepalived approach seems to me quite complex, considering os environment
14:44:06 <enikanorov__> whoever takes care of that, don't forget to write a wiki page with the design
14:44:20 <obondarev> sure
14:45:07 <enikanorov__> any other items to discuss?
14:45:16 <Vijay_> i have one quick quesetion regarding testing.
14:45:29 <Vijay_> You said that I-2 is the deadline
14:45:32 <Vijay_> this is for what?
14:46:06 <enikanorov__> this is to provide a system for testing vendor plugins
14:46:12 <enikanorov__> that is integrated with gerrit
14:46:29 <enikanorov__> if your plugin is not yet in the tree, than you don't need to hurry
14:47:26 <Vijay_> ok
14:48:07 <Vijay_> anything else?
14:48:20 <enikanorov__> not from my side
14:48:35 <enikanorov__> i think we can end the meeting
14:48:46 <Vijay_> ok. Bye Everyone!
14:48:53 <s3wong> Thanks! Bye!
14:48:55 <enikanorov__> thanks for joining
14:48:59 <enikanorov__> #endmeeting