14:00:09 #startmeeting watcher 14:00:10 Meeting started Wed Feb 24 14:00:09 2016 UTC and is due to finish in 60 minutes. The chair is acabot__. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:11 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:13 The meeting name has been set to 'watcher' 14:00:16 hi 14:00:18 \o 14:00:21 hi all 14:00:25 (O_O) 14:00:30 hi 14:00:36 hi 14:00:43 meeting agenda #link https://wiki.openstack.org/wiki/Watcher_Meeting_Agenda#02.2F24.2F2016 14:00:47 o/ 14:00:57 o/ 14:01:00 o/ 14:01:20 hi 14:01:36 #topic Announcements 14:02:06 hi 14:02:22 #info dtardivel nominates vincentfrancoise as a Watcher core contributor 14:02:37 we had a couple of +1 on the ML 14:03:02 even a +2 14:03:10 :) 14:03:11 +1 14:03:11 does anybody "dont want" vincentfrancoise as a core ? :-) 14:03:38 vincent did a great job. 14:04:10 I also think so 14:04:13 +1 14:04:48 #action dtardivel set vincentfrancoise as core contributor on watcher, python-watcherclient & watcher-dashboard 14:05:14 acabot__: done :) 14:05:21 #info mid-cycle highlights have been shared through the ML 14:05:21 I was already core on watcher-dashboard since I was the one creating the project :) 14:05:48 if I missed something important in the list please reply to it 14:06:00 #link http://lists.openstack.org/pipermail/openstack-dev/2016-February/087174.html 14:06:15 BTW, maybe we should discuss on who is to be core on watcher-dashboard at some point 14:07:11 true, do we have any candidate to become core on watcher-dashboard ? 14:07:11 or we can do it through the ML 14:07:50 today we only have vincentfrancoise right ? 14:07:51 FYI, dtardivel and I are the only core on watcher-dashboard 14:07:56 ok 14:08:41 tpeoples gzhai : do you want to be part of the watcher-dasboard core ? 14:09:07 we can wait for more contributors related to horizon stuff 14:09:18 I want more time for watcher... 14:09:23 doubtful, i'm not all that interested in UI side / JS. let me have more of a look and get back to you 14:10:39 ok lets continue with vincentfrancoise & dtardivel as core for now 14:11:05 any other annoucement from the team ? 14:11:18 It's just that we need to make sure we take this new project into account whenever writing new BPs 14:12:04 for sure vincentfrancoise 14:12:27 I'll attend openstack hackathon in Chengdu, China. And ask for a slot for watcher presentation:) 14:12:46 gzhai : great ! 14:13:02 Will use some previous slides to introduce watcher to China openstack community. 14:13:05 hi 14:13:17 watcher-dashboard tagged 0.0.1 available ( + online doc, + docker container image watcher/dashboard) 14:13:34 i apologize for my lack of activity in the past week and a half. that is going to change.. jfyi if you were wondering where the heck i've been :( 14:13:38 #info gzhai will try to introduce Watcher to an openstack hackathon in Chengdu, China 14:14:05 tpeoples : yes we thought you were in vacation :-D 14:14:06 gzhai: +1 14:14:49 i had to move offices, lots of other stuff going on. like i said, it will change 14:15:08 if not kick me out 14:15:10 :) 14:15:16 dtardivel shouldn't we align versions between watcher, watcher-pythonclient and watcher-dashboard at some point ? 14:15:33 tpeoples: no worries, I did the code review in the meantime and added some comments 14:16:17 gzhai: what will be the number of your attendees ? 14:16:36 30~40 in last year. 14:16:50 #topic Review Action Items 14:16:54 big company in china, like Huawei, 99cloud. 14:17:04 acabot__: not necessary ... 14:17:18 gzhai: nice! 14:17:34 gzhai : I'm on the Huawei booth right now at Mobile World Congress ;-) 14:17:45 :) 14:17:50 dtardivel: because they should be able version agnostic? e.g., python-watcherclient should be able to interact with a watcher backend of a different version? 14:17:59 Huawei put a lot of resources for openstack. 14:20:18 tpeoples : as think that until we are in version 0.x, you need to get the latest version of each component to run Watcher, right ? 14:20:20 tpeoples: python-watcherclient should support different version of watcher API. 14:20:22 tpeoples: I guess we just need to stabilize watcher-dashboard a little more before aligning the versions 14:20:36 vincentfrancoise ok 14:20:49 lets move on watcher specs 14:21:01 acabot__: i think that is only true in these early stages where we aren't worrying about backwards compatibility. in the future that shouldn't be necessary 14:21:09 vincentfrancoise : we don't have to align the version 14:21:11 +1 vincentfrancoise 14:21:15 The 0.0.1 tag was put just so we can generate the docker containers, otherwise we wouldn't have tagged it 14:21:16 we can but this is not mandatory 14:21:42 since we have the api compatibility 14:21:52 i'm just concerned about the distance between 0.0.1 & 0.23 :-) 14:22:13 this is a problem ? :) 14:22:25 I reckon we can deal with in Newton if needed 14:22:49 jed56 : not really, just that 0.1 could have been better ;-) 14:23:13 acabot: watcher, python-watcherclient and watcher-dashboard are independent project, with independent life-cycle. We just implemented and pushed a beta PROD version of the dashboard, so this is the reason of version 0.0.1 today. 14:23:29 yes but our 0.1 is quite stable 14:23:30 :) 14:23:57 jed56 dtardivel : Ok i'm fine with that guys 14:24:27 one commit => 0.0.1 : ) 14:24:34 two commit 0.1 14:24:35 lol 14:24:38 #info Efficacy indicators specs has been merged too fast and we realized that it needs improvements 14:24:50 three commit 1.0 stable 14:25:02 :) 14:25:06 so there is a new review open https://review.openstack.org/#/c/283449/, sorry for the confusion 14:25:51 we have 2 specs open for a long time now (ZHAW strategy & workload balance strategy) 14:25:56 acabot__: I'm looking at this. 14:26:16 I suggest to merge them by EOW 14:26:34 seanmurphy vtech : what do you think ? 14:26:41 to merge the 2 strategies? 14:26:45 i saw this in the agenda 14:26:54 we are doing our implementation as you might have seen 14:27:09 there is the point that they seem to be contradictory 14:27:14 at least at the initial glance 14:27:34 no no sorry 14:27:35 one focused on consolidation and one focuses on even distribution of load iiuc 14:27:56 not merge the 2 strategies into one but having both specs merged in gerrit :-) 14:28:09 ok 14:28:20 my sentence is confusing :-) 14:28:28 i don’t understand the advantages/disadvantages of this tbh 14:28:55 i did not read the spec for the workload balance strategy 14:29:15 Who can help to review workload balance? 14:29:16 seanmurphy : could you have a look at it by EOW ? 14:29:27 form our selfish perspective, we would say ‘what is the benefit of this?' 14:29:35 and also ‘does it mean we would have to change our implementation 14:29:42 for sure i can look 14:30:14 seanmurphy : not sure to follow you, are you taking about the process of merging the spec in watcher-spec repo ? 14:30:16 not that we don’t want to be constructive, but i’m just unclear on what the point is/what the benefit is 14:30:33 ah - sorry 14:30:43 confusion - my bad 14:30:52 so you’re not talking abt merging them into each other 14:30:53 right 14:31:21 ok - if it’s just a q of whether our strategy can be merged by eow 14:31:26 then this is doable 14:31:27 seanmurphy: ok back on our feet 14:31:38 woah I was lost :) 14:31:45 +1 14:32:03 regarding the workload balance issue - we can provide a review of this 14:32:25 #action seanmurphy acabot review specs https://review.openstack.org/#/c/271028/ & https://review.openstack.org/#/c/276586/ by EOW 14:33:13 you can add me 14:34:08 gzhai : I think we need more reviews on https://review.openstack.org/#/c/274688/, what do you think ? 14:34:08 #action jed56 review specs https://review.openstack.org/#/c/271028/ & https://review.openstack.org/#/c/276586/ by EOW 14:34:46 +1 14:36:07 vmahe I think we should look at the impact we discussed on efficacy indicators on the input parameters specs 14:36:42 +1 14:37:01 + 14:37:05 +1 14:37:25 #action vmahe review https://review.openstack.org/#/c/274688/ 14:37:48 we have a couple of BPs with high priority that needs specs 14:38:14 tomasz is setting up devstack before starting the scoring module 14:38:35 tpeoples: what about the objects wrapper ? 14:38:46 tomasz: did you success to setup devstack ? 14:38:56 acabot__: i *will* get a spec for that in next 2 days 14:39:37 i saw tomasz' comments on the devstack review... i will make a few more updates there 14:39:40 #action tpeoples submit a drafted spec for https://blueprints.launchpad.net/watcher/+spec/cluster-model-objects-wrapper 14:39:42 thx 14:40:09 tpeoples: I will review it after 14:40:17 I will submit a spec for https://blueprints.launchpad.net/watcher/+spec/select-destinations-filter 14:40:30 jed56 I'm very close. I have controller + compute nodes, able to boot vm, but live migration is not working. troubleshooting it now, got some help from other openstack contributors here in Gdansk 14:40:35 #action acabot submit a drafted spec for https://blueprints.launchpad.net/watcher/+spec/select-destinations-filter 14:41:21 tkaczynski: maybe, we can help you on live migration. Is it on NFS setup ? 14:41:29 tkaczynski: are you using multiple virtualbox VMs? just curious 14:41:34 we can jump to devstack doc improvements ;-) 14:42:40 yes, NFS is working fine. there was some unclear error in the n-cpu logs, we suspect libvirt or qemu problem, but need to investigate 14:42:57 tkaczynski: did you set up SSH keys? 14:43:17 yes, of course. and confirmed with ssh to/from stack/root 14:43:26 hmm ok 14:43:35 and yes, using 2 virtualbox VMs 14:43:41 cool 14:44:23 #info 17 bug fixes are in review, so we need reviewers 14:44:32 once I have it configured, I plan to setup the whole thing again and document it, so I can repeat that process 14:44:42 #action tpeoples review 14:44:57 I assume that with vincentfrancoise as core it will be easier to merge them 14:44:59 tkaczynski: that would be awesome if we add that to our docs 14:45:32 tpeoples: +1 14:46:20 nothing new on python-watcherclient 14:46:35 #info initial commit for watcher-dashboard has been merged 14:46:52 #action acabot set-up launchpad for watcher-dashboard 14:47:15 #topic Blueprint/Bug Review and Discussion 14:48:01 tkaczynski: we have an old blog post about setting live migration in Openstack - it might be helpful #link https://blog.zhaw.ch/icclab/setting-up-live-migration-in-openstack-icehouse/ 14:48:24 we need to start implementing workload balance strategy if we want to demonstrate it in Austin 14:48:44 gzhai : do you have any update on junjie stuff ? 14:48:54 brunograz thanks, I'll definitely have a look there 14:49:22 acabot__: He is working on the specs. 14:49:59 we are preparing the algothm to start working on workload migration spec. 14:50:55 did he starts implementing or not yet ? 14:51:43 acabot__: not yet. We will start soon. 14:52:24 And want to more review for the spec to avoid unnecessary work:) 14:52:49 gzhai i will 14:53:15 jed56: thanks 14:54:15 ping acabot_ 14:54:26 back 14:54:36 sorry wifi disconnection 14:54:59 #topic Open Discussion 14:54:59 #topic Open Discussion 14:55:37 ping jed56 14:55:48 we have a blog post which we’re writing in which we mention watcher and the fact that we’re contributing to it - it’s for an eu proj 14:56:12 seanmurphy: looks great 14:56:34 #info seanmurphy will write a blog post mentionning Watcher contributions 14:56:50 yep - it’s already drafted - it will go out this week 14:56:50 acabot_: is implementation of Watcher policy.json file mandatory for m3 ? 14:57:14 dtardivel: no its not 14:57:37 not in my mitaka-3 list... 14:57:59 you mean this one https://blueprints.launchpad.net/watcher/+spec/watcher-policies ? 14:58:24 its not targeted yet so Newton 14:58:37 acabot_: yes I do 14:59:18 thank you guys 14:59:24 ttyl 14:59:40 ttyl 14:59:42 bye 14:59:45 bye 14:59:46 adiós Antoine 14:59:47 bye 14:59:49 bye 14:59:49 bye 14:59:51 bye 14:59:55 #endmeeting 14:59:58 bye 15:01:32 acabot_: you have to #endmeeting with the same nick you started it with 15:01:52 e.g., 'acabot__' (two underscores) 15:02:00 right 15:02:55 #endmeeting