14:00:54 <jed56> #startmeeting watcher 14:00:55 <openstack> Meeting started Wed Jul 27 14:00:54 2016 UTC and is due to finish in 60 minutes. The chair is jed56. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:56 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:58 <openstack> The meeting name has been set to 'watcher' 14:01:07 <jed56> hi everybody 14:01:14 <vincentfrancoise> o/ 14:01:57 <jed56> woah 14:03:08 <jed56> many peoples! 14:03:27 <jed56> agenda for today #link https://wiki.openstack.org/wiki/Watcher_Meeting_Agenda#04.2F06.2F2016 14:03:34 <jed56> #topic Announcements 14:03:40 <vincentfrancoise> seems like everyone is on holiday :) 14:03:54 <jed56> #info our talks are open to votes 14:04:03 <jed56> We submitted 6 talks, this is great ! 14:04:06 <hvprash> Hi 14:04:08 <jwcroppe> Hi 14:04:14 <gzhai-mobl> hi 14:04:22 <jed56> This is the list of presentations 14:04:31 <jed56> Watcher, the Infrastructure Optimization service for OpenStack: Plans for the O-release and beyond with Susanne Balle, Charlotte, Taylor Peoples 14:04:33 <sballe_> Hi 14:04:41 <jed56> Optimizing Cloud Resources leveraging OpenStack Watcher with Prashanth Hari 14:04:41 <jed56> Susanne Balle, Sameer Adhikari 14:04:49 <jwcroppe> o/ 14:04:49 <jed56> Automatic network efficiency adjustment to optimize power saving 14:04:49 <jed56> David Slama,Tal Anker, Joe Cropper 14:04:54 <jed56> Installing Watcher and demo of a Watcher optimization with Vincent francoise, David TARDIVEL Gang Zhai 14:05:01 <jed56> Towards to an Intelligent Cloud Gang Zhai, Susanne Balle 14:05:08 <jed56> Balance workloads for optimal performance with Watcher Vincent Mahe, Ananth Narayan Sankaranarayanan 14:05:23 <jed56> so please vote for us ;) 14:05:25 <jed56> #link https://www.openstack.org/summit/barcelona-2016/vote-for-speakers/presentation/15931/?q=Watcher 14:05:42 <sballe_> +1 14:06:34 <jed56> #info The mascot for watcher is a jellyfish (mid-cycle) 14:06:54 <jed56> The watcher mid-cycle was a great meeting, thanks everybody ! 14:07:03 <sballe_> Is that settled ? The jellyfish 14:07:18 <sballe_> We gave them 3 proposals 14:07:41 <jed56> I'm not sure 14:07:47 <jed56> this is the winner of our pool 14:07:49 <jed56> poll 14:08:12 <jed56> we will need to wait acabot to have this information 14:08:13 <sballe_> Yeah I hope for the jelly fish 14:08:33 <michaelgugino> hello all 14:08:33 <jed56> The etherpad of the meeting is available here : https://etherpad.openstack.org/p/newton-watcher-midcycle 14:09:03 <jwcroppe> jellyfish ... +100000 :) 14:09:09 <sballe_> Lol 14:09:20 <jed56> jwcroppe: ;-) 14:09:24 <hvprash> :) 14:09:58 <jed56> jwcroppe: :) 14:10:12 <jed56> jwcroppe: can you share with us your meeting with nova 14:10:34 <jwcroppe> yes, there was a lot of discussion on the watcher use cases 14:11:44 <jwcroppe> jaypipes was in favor of opening up a placement API that would, given an instance ID, would return the resource providers (from which you could determine the hosts) that satisfied the request spec 14:12:06 <jed56> jwcroppe: every possible hosts ? 14:12:32 <jwcroppe> yes (possibly there could be some item to pass in that would limit the # of return results) 14:12:43 <jed56> ok 14:12:53 <jwcroppe> but there were also some others that had other ideas of having nova support a general weighing function... 14:13:14 <jwcroppe> and then the nova placement API would just return the top-ranked one that satisfied the request spec 14:13:39 <jed56> do you know how nova will decided to rank the list ? 14:13:49 <jwcroppe> but... there was concern that having external parties (e.g., Watcher) having to constantly update the weights for the resources would overwhelm the placement API 14:14:40 <jwcroppe> So those were essentially the two ideas - I think more discussion is needed to reach consensus 14:15:27 <jed56> what are the next actions ? do we need to write a backlog spec ? 14:17:23 <jwcroppe> not yet, I want to talk with Jay a bit more... 14:17:54 <jed56> #action jwcroppe talk to Jay Pipes 14:17:55 <jed56> :) 14:17:56 <jwcroppe> the problem is understood and there are two very different ideas about how to solve the prob 14:18:51 <vincentfrancoise> at least that's a start :) 14:18:52 <jwcroppe> and the new placement APIs they're discussing are still to-be-created as well (maybe edleafe can tell us more about those) 14:19:29 <jed56> edleafe: are you here ? :) 14:19:57 <jed56> ok let's continue 14:20:23 <jed56> #action jwcroppe jed56 talk to edleafe Placement APIs nova 14:21:01 <jed56> vincentfrancoise: yes this is the first step 14:21:04 <jed56> quick update regarding newton-3 14:21:17 <jed56> #info : 8 BPs en 12 bugs are targeted for newton-3 14:21:26 <jed56> this is great 14:21:54 <jed56> https://launchpad.net/watcher/+milestone/newton-3 14:22:17 <jed56> any other announcement ? 14:23:21 <jed56> #topic Review Action Items 14:24:12 <jed56> The spec “Graph model describing virtual and physical elements in a data center” need to be merged as soon as possible because the development will start soon. 14:24:26 <jed56> #link https://review.openstack.org/#/c/298891/ 14:24:27 <sballe_> Ok I will review 14:24:48 <jed56> #action sballe_ jed56 review https://review.openstack.org/#/c/298891/ 14:24:49 <hvprash> +1 14:26:38 <jed56> The update define-the-audit-scope (https://review.openstack.org/#/c/339456/) spec need to be overhaul 14:27:07 <jed56> We will only need to filter by AvaibilityZone and aggregatehost 14:27:09 <jwcroppe> I'm also working on the limit-concurrent-actions spec and will have it ready by Fri 14:27:52 <jed56> #action alexchadin "define-the-audit-scope" aggregatehost and avaibilityzone 14:28:22 <jed56> jwcroppe: cool 14:28:36 <jed56> jwcroppe: do you know that this is not possible for now in watcher to have concurrent actions ? 14:28:56 <jed56> we need to implement https://blueprints.launchpad.net/watcher/+spec/planner-storage-action-plan 14:28:58 <jed56> first 14:29:19 <jwcroppe> ok, we can keep as a backlog spec if we need as well 14:29:45 <jed56> yes ! jinquan want to work this blueprint 14:29:58 <jed56> on 14:30:04 <sballe_> Cool 14:30:04 <jwcroppe> I'll still write up the spec for limit-concurrent-actions 14:30:19 <jed56> jwcroppe: yes this is important feature for watcher 14:30:30 <jed56> jwcroppe: can you create a blueprint ? 14:30:42 <jwcroppe> I did 14:30:55 <jwcroppe> https://blueprints.launchpad.net/watcher/+spec/limit-concurrent-actions 14:30:56 <jed56> #action jwcroppe create a blueprint regarding limit-concurrent-actions 14:31:05 <jed56> sorrry :) 14:31:10 <jwcroppe> no prob :) 14:31:41 <jed56> the spec "automatic triggering audit" => https://review.openstack.org/#/c/342473/ need a new patchset 14:32:06 <jed56> #action diga update the spec https://review.openstack.org/#/c/342473/ 14:32:48 <jed56> so now, watcher 14:33:10 <jed56> I would like that the 'cluster-model-objects-wrapper ' and 'scoring-module' blueprints to be merged before the 3 of august. 14:33:18 <jed56> The cluster-model-objects-wrapper blueprint is on the critical path for 'graph-based-cluster-model' blueprint. 14:33:26 <jed56> #link https://review.openstack.org/#/q/status:open+project:openstack/watcher+branch:master+topic:bp/cluster-model-objects-wrapper 14:33:33 <jed56> #link https://review.openstack.org/#/q/status:open+topic:bp/scoring-module 14:33:39 <jed56> Who could take some time to look at that in priority, please ? 14:34:19 <hvprash> jed56, i can take a look 14:34:46 <sballe_> I can look to 14:34:54 <hanrong1> +1 14:34:55 <jed56> #action sballe_ hvprash review https://review.openstack.org/#/q/status:open+project:openstack/watcher+branch:master+topic:bp/cluster-model-objects-wrapper and https://review.openstack.org/#/q/status:open+topic:bp/scoring-module 14:34:58 <danpawlik> +1 14:35:18 <sballe_> I have some time now because I go on vacation Fri 14:35:39 <jed56> #action danpawlik hanrong1 review https://review.openstack.org/#/q/status:open+topic:bp/scoring-module https://review.openstack.org/#/q/status:open+project:openstack/watcher+branch:master+topic:bp/cluster-model-objects-wrapper 14:35:43 <jed56> thanks a lot 14:35:49 <hanrong1> ok 14:36:42 <vincentfrancoise> I need to rebase cluster-model-objects-wrapper <https://review.openstack.org/#/q/status:open+project:openstack/watcher+branch:master+topic:bp/cluster-model-objects-wrapper> due to the recent merges but please do review it :) 14:37:26 <jed56> Add goal_id, strategy_id and host_aggregate CLI options to audit https://review.openstack.org/#/c/332982/ need a new patchset 14:37:38 <hvprash> jed56, working on it today 14:37:55 <jed56> hvprash: super! 14:37:57 <hvprash> there is one unit test i missed and will add to the PS 14:38:09 <vincentfrancoise> good :) 14:38:10 <hvprash> there is also the changes which michaelgugino is working https://review.openstack.org/#/c/330165/ 14:38:13 <jed56> #action hvprash push a new patchset 332982 14:38:14 <sballe_> I can review 14:38:15 <hvprash> which is critical 14:38:42 <hvprash> thx 14:38:52 <jed56> okay 14:39:07 <gzhai-mobl> i can revie 14:39:30 <sballe_> gzhai-mobl: cool 14:39:36 <jed56> #action gzhai-mobl review https://review.openstack.org/#/q/status:open+topic:bp/scoring-module https://review.openstack.org/#/q/status:open+project:openstack/watcher+branch:master+topic:bp/cluster-model-objects-wrapper 14:39:36 <gzhai-mobl> w 14:40:16 <jed56> https://review.openstack.org/#/c/330165/ Implement goal_id, strategy_id and host_aggregate into Audit api need a new patchset 14:40:40 <jed56> this is almost finish 14:41:05 <jed56> michaelgugino: do you think that you will have time to work on it ? 14:41:06 <hvprash> yes 14:41:32 <jed56> okay great 14:41:34 <jed56> #action michaelgugino push a new patchset https://review.openstack.org/#/c/330165/ 14:41:41 <michaelgugino> yes, I'm working on it. I will continue to make changes based on people's reviews 14:41:53 <jed56> michaelgugino: thanks a lot 14:41:55 <hvprash> thx michaelgugino 14:41:57 <michaelgugino> yw 14:42:26 <vincentfrancoise> michaelgugino: do you have a devstack to make sure the integration tests are passing ? 14:43:12 <michaelgugino> not at the moment, I will build one today. 14:43:19 <hvprash> i tested the cli options against the api changes michaelgugino is working and works as expected 14:43:37 <vincentfrancoise> michaelgugino: ok cool :) 14:44:00 <jed56> Add rally-jobs folder to get rally support(https://review.openstack.org/#/c/333371/) patchset need a status 14:44:41 <jed56> #action alexchadin can you give a us a feedback regarding https://review.openstack.org/#/c/333371/ ? 14:45:02 <jed56> https://review.openstack.org/#/c/345214/ need a new patchset 14:46:01 <jed56> #action vincentfrancoise rebase 345214 14:46:51 <jed56> sballe_: do you know if junjie.huang plan to work on https://review.openstack.org/#/c/342594/ ? 14:47:20 <sballe_> Not sure I thought we decided not to move forward with it for now 14:47:31 <sballe_> Is this the airflow one? 14:47:45 <gzhai-mobl> i will ping him 14:47:52 <sballe_> Ok 14:47:55 <sballe_> Thx 14:48:25 <jed56> #action gzhai-mobl contact junjie.huang in order to have status on https://review.openstack.org/#/c/342594/ 14:48:28 <gzhai-mobl> he is waiting for the ci error got fixed 14:48:54 <sballe_> Ok so junjie will finish it? 14:49:05 <sballe_> gzhai-mobl: ^^^ 14:49:09 <jed56> what is the problem with the CI ? 14:49:12 <gzhai-mobl> should do 14:49:18 <sballe_> ok thx 14:49:35 <jed56> IMHO, this is not a CI problem 14:49:55 <gzhai-mobl> some generic error tha 14:50:02 <gzhai-mobl> t time 14:50:18 <jed56> gzhai-mobl: can you rebase the code of junjie ? 14:51:17 <jed56> #action gzhai-mobl rebase https://review.openstack.org/#/c/342594 14:51:28 <gzhai-mobl> i will ask him 14:51:42 <sballe_> jed56: I ll follow up with junjie as well 14:52:19 <jed56> okay 14:52:46 <jed56> #action sballe_ follow https://review.openstack.org/#/c/342594 14:52:58 <jed56> #topic Blueprint/Bug 14:53:09 <jed56> Many bugs have been merged in newton-2 thanks to everybody! 14:53:40 <sballe_> Cool 14:53:46 <jed56> #topic open discussion 14:53:59 <jed56> some body want to discuss something ? :) 14:54:15 <hvprash> would like to start planning for the workload qos. sballe_ do you have sometime before vacation ? 14:54:35 <sballe_> Yes let's talk today or tomorrow 14:54:55 <hvprash> cool, thx 14:54:56 <sballe_> Plus I can IRC while on vacation since you are on the east coast 14:55:11 <sballe_> Only 6 hrs difference 14:55:15 <hvprash> that works too ! but don't want to bother while you are in vacation :) 14:55:17 <jed56> #action sballe_ hvprash start planning for the workload qos 14:55:53 <sballe_> hvprash: no worries 14:56:50 <jed56> Can I close the meeting ? 14:57:10 <vincentfrancoise> I guess so :) 14:57:18 <hvprash> sure :) 14:57:25 <gzhai-mobl> bye 14:57:32 <hvprash> thx jed56 14:57:33 <jed56> thanks everybody have a nice week ! 14:57:34 <vincentfrancoise> bye 14:57:38 <jed56> #endmeeting