16:00:02 <acabot> #startmeeting watcher 16:00:03 <openstack> Meeting started Wed Oct 21 16:00:02 2015 UTC and is due to finish in 60 minutes. The chair is acabot. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:08 <openstack> The meeting name has been set to 'watcher' 16:00:17 <acabot> hi 16:00:52 <vincentfrancoise> Hi 16:00:58 <dtardivel_> hi 16:01:02 <jwcroppe> hello 16:01:19 <edleafe> o/ 16:01:36 <sballe> o/ 16:01:41 <jwcroppe> o/ 16:01:47 <acabot> our meeting agenda #link https://wiki.openstack.org/wiki/Watcher_Meeting_Agenda 16:02:46 <jwcroppe> acabot: ready... thanks for sending the email out 16:02:49 <tpeoples> hi 16:02:57 <acabot> ok lets start with announcements 16:03:09 <acabot> #topics announcements 16:03:20 <acabot> #topic announcements 16:04:09 <acabot> I'd like to introduce vincentfrancoise who is joining our watcher-drivers team 16:04:24 <sballe> cool! 16:04:28 <jwcroppe> welcome :) 16:04:46 <sballe> vincentfrancoise: are you from b-com? 16:05:10 <vincentfrancoise> Yeah I'm from b-com :) 16:05:11 <tpeoples> howdy vincentfrancoise 16:05:34 <sballe> welcome 16:05:41 <acabot> vincentfrancoise has been working on our metering chain for 3 months now and will jump on Watcher core components now 16:05:49 <dtardivel_> acabot: do we need to add vincent into Launchpad "watcher drivers" team ? 16:06:07 <acabot> dtardivel_: yes please do 16:06:26 <jwcroppe> vincentfrancoise: welcome aboard 16:06:44 <acabot> #action dtardivel_ add vincentfrancoise in watcher drivers 16:07:00 <acabot> any other annoucement ? 16:07:19 <jwcroppe> acabot: nothing here - ready for actions 16:07:33 <acabot> jwcroppe: ok 16:07:45 <acabot> #topic Review Action Items 16:08:29 <acabot> mailing to the community has been sent today 16:08:36 <jwcroppe> saw that - thanks 16:08:49 <sballe> yeah that was great! 16:08:49 <acabot> we already have an answer from a ceilometer contributor 16:09:28 <acabot> http://lists.openstack.org/pipermail/openstack-dev/2015-October/077627.html 16:09:39 <jwcroppe> yep, will be good to continue socializing Watcher 16:09:55 <acabot> I think we should definitely add Watcher to https://wiki.openstack.org/wiki/Ceilometer#Ceilometer_Extensions 16:10:01 <jwcroppe> +1 16:10:04 <sballe> also we need to look opnfv and the doctor project so we have an answer to the quesiton how is wathcer different from doctor 16:10:11 <sballe> acabot: +1 16:10:46 <jwcroppe> #link https://wiki.opnfv.org/doctor 16:11:32 <jwcroppe> sballe: they look to be quite different - we can add this to Wiki FAQ 16:12:01 <acabot> #action acabot sballe look at opnfv doctor project before the summit 16:12:07 <sballe> jwcroppe: yes and no. I think we just need to have a written down answer around why they are different 16:12:17 <jwcroppe> sballe: +1 16:12:24 <acabot> sballe: +1 16:13:18 <jwcroppe> "The goal of this project is to build fault management and maintenance framework for high availability of Network Services on top of virtualized infrastructure" ... I think the core missions are different. We can get pen to paper on that. 16:13:26 <acabot> doctor is not part of openstack so even if its close, they can live in both openstack and opnfv 16:14:05 <sballe> agree with both of you. We just need to make sure we understand the difference 16:14:07 <acabot> opnfv is deeply related to NFV and I think we do not want to address this right now in Watcher 16:14:42 <jwcroppe> acabot: +1 ... we've got the action above to write that down, next topic? 16:14:47 <sballe> the question is doctor going ot be a sub-set of watcher? 16:14:56 <sballe> jwcroppe: +1 16:15:08 <jed56> +1 16:15:10 <jed56> a) Scheduled evacuation and Live migration would be considered to be actions to move VM from faulty host. If VM does not want to support this, no action. 16:16:03 <sballe> jed56: This could sound like watcher as well 16:16:13 <jed56> yes you can see that here https://wiki.opnfv.org/_media/doctor-opnfv-proposal.pptx 16:16:51 <acabot> as stated we will have a deeper look at it before the summit 16:16:53 <edleafe> but that's an aspect of live migration in general 16:16:58 <jwcroppe> I propose that we do some research on this and start compiling notes and write some notes 16:17:01 <tpeoples> seems like the difference is fault tolerance vs optimization, two kind of related but not quite the same goals 16:17:04 <jwcroppe> before summit 16:17:22 <sballe> jwcroppe: yes I believe that is the action above 16:17:28 <jwcroppe> sballe: agree 16:17:30 <acabot> lets move to stackforge code migration 16:17:39 <jed56> tpeoples i agree 16:17:42 <sballe> so did it happen? 16:17:48 <sballe> the stackforge migration/ 16:18:01 <dtardivel_> watcher and python-watchercient projects have migrated to openstack namespace 16:18:07 <sballe> +1 16:18:28 <jwcroppe> dtardivel_: assume we also need watcher-specs created ? 16:18:40 <dtardivel_> but there is an issue on watcher project, due to bad oslo config version used 16:18:58 <wznoinsk> tpeoples: +1, watcher would be doing fault tolerance moves as well? 16:19:08 <dtardivel_> we alreay fixed this issue internaly 16:19:36 <jed56> we are waiting for a core review for watcher-specs 16:19:48 <dtardivel_> We will put a patch before the summit in order to finalize the migration 16:19:52 <sballe> can you send us the review 16:20:14 <jwcroppe> jed56: sounds good - do we have our core names loaded into specs and watcher-core ? 16:20:50 <jed56> yes : https://review.openstack.org/#/c/233679/ 16:20:53 <sballe> thx 16:21:09 <dtardivel_> sballe: yes of course 16:21:15 <jed56> the document was not always clear between openstack and stackforge 16:21:17 <jwcroppe> jed56: IIRC, for watcher-specs it was sballe, acabot and jwcroppe? For watcher-core it was jed56, dtardivel_ and tpeoples, correct? 16:21:33 <tpeoples> wznoinsk: i don't think that is something we would try to handle with the first iteration of watcher, but it is something we can discuss and keep in mind. you could think of "faults" as just some other stat that could reach a threshold to be optimized, so the framework watcher provides should allow that in the future 16:21:59 <acabot> jwcroppe: jed56 is talking about creating the watcher-specs repo 16:22:19 <acabot> jwcroppe: as it is still in review, we cannot add contributors yet 16:22:30 <jwcroppe> acabot: +1 16:22:47 <wznoinsk> tpeoples: yes, optimization is one thing and fault tolerance is another, it tend to treat tehse two separate, we can take any discussion off-meeting/in open discussion part ... 16:22:53 <sballe> tpeoples: we will be looking at sla and risk minization so we will be looking at risk for component failure after our initial poc around nergy comsumption 16:22:55 <jwcroppe> acabot: I think watcher-core should be ready, correct? 16:24:43 <acabot> jwcroppe: today watcher-core is the watcher-drivers team 16:25:15 <acabot> jwcroppe: we are still waiting for watcher-specs creation to start a dedicated launchpad with contributors 16:25:24 <jwcroppe> ok, sounds good 16:25:30 <acabot> jwcroppe: if its needed 16:27:20 <acabot> we also add a github repo for watcher toold 16:27:22 <acabot> tools 16:27:30 <acabot> helping people deploying Watcher 16:27:44 <acabot> #link https://github.com/b-com/watcher-tools 16:28:05 <dtardivel_> we are internaly on review on docker part 16:28:36 <dtardivel_> we will push also ansible playbook later (before the summit I hope) 16:28:43 <sballe> acabot: yeah I started doing the deploy the otehr day. Still working at it. 16:28:47 <jwcroppe> acabot: yeah and hopefully watcher can easily build on devstack as well (eventually) 16:29:06 <sballe> I have devstack deployed and was tryingot get watcher to work. 16:29:22 <sballe> I could use some more documentation :-) 16:29:52 <jwcroppe> sballe: code == documentation ;-) 16:29:58 <jwcroppe> just kidding - agree 16:30:07 <acabot> sballe: I think we will do it together next week and will improve doc as well 16:30:22 <dtardivel_> please wait for next commit on docker part, because we updated container images (located on dicker hub), and we need to update documentation as well 16:31:19 <dtardivel_> for information we use gerrithub.io for reviews on this project 16:31:21 <jwcroppe> acabot: ok, so sounds like progress has been made on the repos 16:31:37 <jwcroppe> acabot: and we just need watcher-specs to complete review 16:32:11 <acabot> jwcroppe: I think everything will be OK before Tokyo 16:32:50 <jwcroppe> sounds good - shall we move on 16:32:51 <jwcroppe> ? 16:32:58 <sballe> +1 16:32:58 <acabot> yes Tokyo agenda 16:33:18 <acabot> sballe sent us an invit for tuesday morning 16:33:30 <sballe> yes let me know if you need to be added 16:33:47 <sballe> it is Tuesday at 10:45 during the coffee break 16:33:50 <jwcroppe> acabot: I started working on a doc I'd like to discuss when we get there with the team... it's not complete, but it will get the discussion going 16:33:52 <jwcroppe> #link https://docs.google.com/document/d/1xcLQY6532Ed2FS2_6d4s6CF4ijeTqr43k5GMSurFepM/edit#heading=h.vae16077waj6 16:34:26 <acabot> wznoinsk: do you want to join ? 16:34:33 <sballe> ok. Should we add all our session to our wiki? I am assuming we can't add them to the official list 16:35:03 <wznoinsk> sballe: yes, thanks 16:35:38 <jwcroppe> the doc I was using a 'scratch pad' of some thoughts we should explore during the PoC phase - with the goal of making things as simple as possible and start getting the framework built out 16:35:39 <sballe> done 16:35:50 <acabot> I think we can add them here https://wiki.openstack.org/wiki/Design_Summit/Mitaka/Etherpads 16:36:03 <sballe> ok let's do thta 16:36:27 <acabot> so I will add the tuesday session but we still have a problem with rooms for next days 16:36:38 <sballe> acabot: jwcroppe Do tou know who Federico Michele Facca is? 16:36:51 <jwcroppe> sballe: no 16:37:11 <jwcroppe> sballe: that's a new name t me 16:37:20 <acabot> sballe: yes sorry I met Federico at a conference on Monday and he is looking at Watcher 16:37:27 <sballe> ok 16:37:32 <jwcroppe> cool 16:37:59 <jwcroppe> acabot: So we have our first meeting setup by sballe, but we should carve out some other focused design slots 16:38:00 <acabot> #action acabot add watcher session to https://wiki.openstack.org/wiki/Design_Summit/Mitaka/Etherpads 16:38:02 <sballe> acabot: do oyu remember the name of the oher intel person that was interested in joining? 16:38:42 <acabot> sballe: isnt it wznoinsk ? 16:39:09 <sballe> no but I think I can find her name 16:39:54 <acabot> sballe: oh yes sorry Mrittika Ganguli 16:40:08 <sballe> yes taht was it. I will invite her too 16:40:36 <acabot> jwcroppe: +1 but we still have a problem with rooms, we should probably define on our tuesday meeting the next design slots 16:41:31 <acabot> everyone is OK with groupme ? 16:42:08 <sballe> feel free to invite people tot he GroupMe app 16:42:14 <tpeoples> yep 16:42:17 <acabot> ok 16:42:23 <jwcroppe> acabot: yeah, I'm not sure if we'll get a room this summit - we might have to take over some corners or some such :) 16:42:55 <acabot> jwcroppe: so we will ;-) 16:42:56 <sballe> +1 16:43:05 <acabot> jwcroppe: next to the coffee machine please 16:43:24 <sballe> Last year a lot of the LBaaS/FWaas discussions happened in the hallway on chairs 16:43:29 <sballe> acabot: +1 16:43:30 <jwcroppe> acabot: +1 16:43:32 <wznoinsk> during breaks the booths on the corridors are usually busy but we can try to use one of them (with couches) 16:43:41 <acabot> ok lets move to our last topic 16:43:43 <sballe> wznoinsk: +1 16:43:51 <acabot> #topic open discussions 16:44:07 <sballe> Just a reminder: Can people please sign in to GroupMe? 16:44:10 <jwcroppe> if people get a chance - take a look at the google doc I started above. it's very rough admittedly, but it's a good straw man for some discussion 16:44:38 <acabot> jwcroppe: ok I will do it in the plane ;-) 16:44:42 <jed56> jwcroppe: i will 16:44:58 <jwcroppe> cool 16:45:24 <jwcroppe> and yes - to echo sballe - join GroupMe :) 16:46:23 <acabot> OK can we close our meeting and just say see you in Tokyo ? 16:46:32 <sballe> sounds good to me 16:46:37 <jwcroppe> Sounds like a plan 16:46:51 <sballe> bye 16:46:52 <tpeoples> safe travels 16:46:52 <acabot> #endmeeting