15:30:09 <DinaBelova> #startmeeting Performance Team
15:30:10 <openstack> Meeting started Tue Feb  7 15:30:09 2017 UTC and is due to finish in 60 minutes.  The chair is DinaBelova. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:30:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:30:14 <openstack> The meeting name has been set to 'performance_team'
15:30:17 <DinaBelova> hello folks :)
15:30:19 <spn> o/
15:30:25 <rcherrueau> o/
15:30:48 <DinaBelova> spn rcherrueau ad_rien_ rook o/
15:31:01 <spn> DinaBelova: o/
15:31:06 <rook> DinaBelova o/
15:31:09 <rook> in and out today :)
15:31:14 <DinaBelova> :D
15:31:18 <ad_rien_> ad_rien_ o/
15:31:38 <DinaBelova> ok, it looks like not everyone has joined yet, but still let's get started
15:31:42 <DinaBelova> #topic Action Items
15:31:47 <DinaBelova> last time we had several ones
15:32:09 <DinaBelova> let's start with rook yours :)
15:32:10 <DinaBelova> rook Neutron API Performance results to performance docs ? :)
15:32:22 <rook> ha, work is complete just need to generate the rst
15:32:24 <rook> :*(
15:32:28 <rook> it is on the backburner
15:32:41 <DinaBelova> rook heh, so let's keep it for one more week
15:32:47 <rook> we can drop it
15:32:59 <rook> as i don't see that happening any time soon, ill get to when soon /me hopes
15:33:15 <DinaBelova> last time spn was wandering what test plan is those data based on
15:33:19 <DinaBelova> rook ack
15:33:45 <DinaBelova> #info rook will add Neutron API Performance results to performance docs once the nearest schedule will be clear for him
15:34:00 <rook> DinaBelova: what was the question?
15:34:30 <DinaBelova> rook "what actually is measured for Neutron API performance - what calls and what tools are used"
15:34:45 <DinaBelova> spn is that correct?
15:34:53 <spn> DinaBelova: yes
15:35:08 <spn> I wanted to see if I can follow the test plan and try it in-house
15:36:09 <DinaBelova> some time ago we used http://docs.openstack.org/developer/performance-docs/test_plans/neutron_control_plane/plan.html but I don't know is rook used the same approach
15:36:55 <rook> https://review.openstack.org/#/c/412554/
15:36:59 <rook> was the tst plan i wrote
15:37:02 <rook> it is more generic
15:37:26 <spn> rook: great.. thank u
15:37:34 <DinaBelova> rook ack, so we just cover all possible API calls with browbeat :) that's fair enough :)
15:37:38 <rook> right
15:38:09 <DinaBelova> ok, so it looks like we may proceed with my action items
15:38:35 <DinaBelova> regarding reviewing the https://pad.inria.fr/p/sfBMUvWOpKtmfXT8 - ad_rien_ I took a look on abstract at the end of last week and everything was ok, so I just did not ping you
15:39:04 <DinaBelova> I think it's safe to say it's a good proposal with good chances for being chosen
15:39:12 <ad_rien_> sorry
15:39:24 <ad_rien_> I'm back (was tchating with rcherreau)
15:39:27 <DinaBelova> for what? :)
15:39:29 <DinaBelova> ah :)
15:39:37 <DinaBelova> ad_rien_ np, it's ok.
15:39:40 <ad_rien_> we submitted the proposal yesterday
15:39:43 <ad_rien_> as expected
15:40:10 <DinaBelova> sure, let's see how it'll go
15:40:11 <DinaBelova> from my perspective this proposal looks cool
15:40:28 <ad_rien_> do not hesitate to vote for the proposal ;)
15:40:29 <ad_rien_> BTW
15:40:31 <ad_rien_> :-P
15:40:34 <DinaBelova> :D sure :D
15:40:36 <ad_rien_> you can put an action on that :D
15:40:56 <ad_rien_> more seriously we would like to put additional efforts on ENOS
15:41:09 <ad_rien_> (We are checking with Orange to see whether we can get additional engineering task force)
15:41:17 <ad_rien_> our idea would be to put ENOS in the CI
15:41:30 <ad_rien_> in order to test performance regression tests
15:41:47 <ad_rien_> s/test/execute performance regression tests
15:41:58 <ad_rien_> we can discuss that point later
15:42:10 <DinaBelova> ad_rien_ ack
15:42:21 <DinaBelova> so regarding last action item on me - share status of workloads testing methodology/results sharing
15:42:28 <tovin07> hello, sorry for being late
15:42:52 <DinaBelova> sadly I have to announce that Intel folks made whole this methodology proprietary :( so nothing will be shared in terms of workloads testing :(
15:42:57 <DinaBelova> tovin07 np :)
15:43:25 <ad_rien_> DinaBelova:  can we at least know what was the technology used ?
15:43:35 <ad_rien_> i.e. kolla based/Kubernetes based ?
15:43:43 <ad_rien_> or even that point is confidential?
15:44:14 <DinaBelova> ad_rien_ the thing was to deploy OpenStack resources in very specific configuration with very specific applications covering almost all ares of usual workloads  (CPU, RAM, disk intensive)
15:44:21 <DinaBelova> and then run workloads through those applicaitons
15:44:26 <DinaBelova> this actually leads us even closer to the SPEC benchmark usage - as it's producing similar results
15:44:44 <ad_rien_> ok
15:45:16 <DinaBelova> spn after a conversation with Intel SPEC folks I got a message, that they are not able to share this benchmark somehow
15:45:43 <spn> oh..
15:45:47 <spn> what is the problem
15:46:09 <DinaBelova> spn yeah
15:46:32 <DinaBelova> I'll try to communicate it a bit further, but the only real option I got is to buy it for Mirantis, that's not really cool
15:46:52 <DinaBelova> as we wanted to make it possible to run it for all Performance Team members
15:47:21 <DinaBelova> and that's bunch of companies - not necessary SPEC members
15:47:54 <spn> oh..
15:48:13 <DinaBelova> spn - so that's most recent news on the SPEC discussion with Intel as well
15:48:43 <DinaBelova> I don't know if communicating about it with some other company will give more results - who knows
15:48:56 <spn> hmm..
15:49:15 <DinaBelova> spn do you have any ideas?
15:49:39 <spn> in terms of getting spec you mean?
15:49:43 <DinaBelova> yeah
15:49:59 <spn> I am not sure how Dell I got it. I will check and update you next week
15:50:09 <DinaBelova> spn ack
15:50:30 <DinaBelova> #action spn find out how Dell was running SPEC Cloud benchmark
15:50:37 <DinaBelova> ok, so it looks like we may proceed
15:50:39 <DinaBelova> #topic Current progress on the planned tests
15:51:03 <DinaBelova> ad_rien_ rcherrueau any news? except the summit talk news? :)
15:51:51 <ad_rien_> rcherrueau is working on the patch for kolla
15:52:06 * ad_rien_ is looking for the url
15:52:42 <ad_rien_> #link https://github.com/rcherrueau/kolla-ansible/commit/e37c87f9e0ea3b2bef7de57f9268d2c913c0c7ff multi-region support in kolla - patch
15:52:58 <ad_rien_> he should send the patch by thursday hopefully
15:53:20 <DinaBelova> cool, so shortly we should have it in kolla itself :) that will be really nice :)
15:53:32 <ad_rien_> hopefully
15:53:34 <ad_rien_> ;)
15:53:50 <DinaBelova> anything else ad_rien_ ?
15:54:00 <ad_rien_> Next actions will be to start WANWide evaluations of OpenStack
15:54:09 <ad_rien_> that's all
15:54:15 <DinaBelova> ack, thank you sir
15:54:16 <ad_rien_> for the moment.
15:54:33 <ad_rien_> you're welcome
15:54:44 <DinaBelova> akrzos are you around? anything to share on your current experiments?
15:56:09 <DinaBelova> in the meanwhile: as for Mirantis - we installed k8s to start with stressing its control plane. although, even without any stress we found a tricky issue with kube proxy - https://github.com/kubernetes/kubernetes/issues/38372
15:56:24 <DinaBelova> its debug actually lead us to the issue in Go language itself 0_o
15:56:29 <DinaBelova> https://github.com/golang/go/issues/18925
15:57:06 <DinaBelova> so it's really funny that issues may lay so deeply in the system
15:57:48 <DinaBelova> therefore no tests were tun yet - we're trying to reproduce this thing again to collect more information for the Go community
15:58:19 <DinaBelova> it looks like akrzos is not yet here
15:58:39 <DinaBelova> so it looks like we may jump to the open discussion
15:58:48 <DinaBelova> #topic Open Discussion
15:59:08 <DinaBelova> ad_rien_ rook spn tovin07 rcherrueau akrzos anything to cover?
15:59:39 <ad_rien_> Is there additional presentations that are "performance team tagged"?
15:59:54 <spn> in the summit?
15:59:57 <ad_rien_> yes
16:00:17 <ad_rien_> I mean does anyone know any other proposal that have been submitted?
16:00:19 <tovin07> from my side: nothing for today
16:00:23 <tovin07> :D
16:00:35 <spn> I submitted one for performance monitoring
16:00:47 <spn> with redhat folks
16:00:51 <ad_rien_> Can we have the link?
16:00:51 <DinaBelova> I've pushed one presentation regarding Perf Team itself with my boss + one more regarding osprofiler with Vipin
16:01:12 <ad_rien_> I think it might be valuable to keep traces of these submissions.
16:01:21 <ad_rien_> Especially because it shows that our team is active
16:01:30 <DinaBelova> afaik there is not yet voting open
16:01:42 <DinaBelova> so we cannot publish links yet :)
16:01:52 <ad_rien_> Moreover it is interesting for others to know what are the current actions
16:02:01 <ad_rien_> (at least from the Inria side we would appreciate to know them)
16:02:15 <DinaBelova> moreover, I just found that call for speakers is prolonged 0_0
16:02:18 <ad_rien_> ok the call has been extended untill the 9th
16:02:24 <DinaBelova> yeah
16:02:25 <ad_rien_> so we can discuss that point next week.
16:02:30 <DinaBelova> ad_rien_ sure
16:02:47 <DinaBelova> #info share links to the perf team presentation once voting will be open for Boston cummit
16:03:41 <DinaBelova> ok, so it looks like we're done for today
16:03:59 <ad_rien_> sounds good
16:04:09 <DinaBelova> thank you folks!
16:04:12 <DinaBelova> see you next week!
16:04:14 <tovin07> thanks
16:04:16 <DinaBelova> #endmeeting