16:00:41 <mlavalle> #startmeeting neutron_performance
16:00:42 <openstack> Meeting started Mon Feb 11 16:00:41 2019 UTC and is due to finish in 60 minutes.  The chair is mlavalle. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:45 <openstack> The meeting name has been set to 'neutron_performance'
16:01:26 <mlavalle> Hi there!
16:01:35 <haleyb> o/
16:01:41 <bcafarel> o/
16:01:52 <rubasov> hello
16:02:18 <mlavalle> bcafarel: I didn't copy in an email I sent last nigh.... forwarding it to you now, though
16:02:47 <mlavalle> done
16:03:08 <bcafarel> mlavalle: no problem, I am still catching up on some of last week's email
16:03:24 <bcafarel> thanks got it!
16:03:57 <mlavalle> let's wait 1 more minute and then we start
16:04:11 <njohnston_> o/
16:04:18 <mlavalle> jrbalderrama: bienvenido. We are glad you shwed up
16:05:39 <mlavalle> ok, let's get moving
16:05:58 <mlavalle> #topic Action items from last meeting
16:06:44 * mlavalle scared jrbalderrama of, apparently. oh well
16:07:17 <mlavalle> So as you might remember, during our last meeting we has call with the EnOS team
16:07:41 <jrbalderrama_> (just having some issues with my account)
16:08:18 <mlavalle> during that meeting I agreed to create in EnOS (using the Vagrant provider) a first approach for a Neutron performance test
16:08:42 <mlavalle> the idea is to start prototyping and refine the test that we ant to eventually perform at scale
16:09:05 * mlavalle is glad jrbalderrama_ no se rajo
16:09:54 <mlavalle> so I deployed all the toold necessary in my personal system: vagrant, virtualbox, EnOS
16:10:43 <mlavalle> I mostly followed the steps shown here: https://enos.readthedocs.io/en/stable/installation.html
16:11:16 <mlavalle> I had a few bumps on the raod but eventually I got the "enos deploy" command going....
16:11:21 <mlavalle> until I hit a wall
16:11:40 <mlavalle> so I sent an email to jrbalderrama_ and company
16:12:14 <mlavalle> I copied almost everybody in that email and added bcafarel earlier today
16:12:35 <jrbalderrama_> Yes indeed, @msimonin will reply that mail
16:12:39 <mlavalle> jrbalderrama_: any suggestions as to what mistake I might be making?
16:13:08 <mlavalle> ah ok, I'll wait for that reply
16:13:28 <mlavalle> does it look like I am on the right path, jrbalderrama_?
16:14:08 <slaweq> hi, sorry for being late
16:14:10 <jrbalderrama_> yes you are, but we tested enos using py2.x because of kolla, we realised you use v3
16:15:31 <mlavalle> ahhh. I think that Saturday evening I had an issue using python 2.7 and that is where I changed to python 3
16:15:49 <mlavalle> ok, I'll give it another try with 2.7 and share issues if I find them
16:15:59 <mlavalle> I have another question jrbalderrama_ ....
16:17:35 <mlavalle> I installed EnOS from source, because according to the docs, that's the way to be able to test OpenStack master
16:18:18 <jrbalderrama_> We also have another potential issue coming from our side (we will get back to you with more details in the mail)
16:19:12 <mlavalle> however, in the reservation.yaml generated by the "enos new" kolla queens is suiggested. Should I change that to master as well?
16:19:37 <mlavalle> jrbalderrama_: see line 41 here: http://paste.openstack.org/show/744817/
16:21:02 <jrbalderrama_> I see, normally a release of enos is associated to a "stable" tag of OS
16:21:31 <jrbalderrama_> in that example was queens but the most recent supported version is rocky
16:22:19 <mlavalle> jrbalderrama_: ok so maybe install in my venv from package and let it deploy OS Rocky?
16:23:20 <jrbalderrama_> that is a better idea. We can even do better, we will release a new (minor) version targeting properly to OS Rocky
16:23:48 <jrbalderrama_> and you can install that version using the packages
16:23:51 <mlavalle> jrbalderrama_: when is that release going to take place?
16:24:49 <jrbalderrama_> I will confirm with @msimonin about working with the master branch of openstack if that is what you want to test
16:25:10 <jrbalderrama_> this week for sure
16:25:59 <mlavalle> jrbalderrama_: super cool. Would you shoot me an email when it is released?
16:26:07 <jrbalderrama_> sure
16:27:12 <mlavalle> jrbalderrama_: once I get EnOS to deploy OpenStack properly, I intend to execute the Rally and shaker worloads here: https://github.com/BeyondTheClouds/enos/tree/master/enos/workload
16:27:43 <mlavalle> to make sure that everything is working properly. Based on that, I will start creating a Rally workload for a future test
16:27:59 <mlavalle> does that make sense? (I'm asking everybody)
16:28:55 <slaweq> make sense for me
16:28:57 <rubasov> it does
16:29:15 <mlavalle> Great!
16:29:33 <mlavalle> that's all the progress I wanted to share with the team about EnOS
16:30:04 <mlavalle> slaweq, rubasov: do you have updates for this meeting?
16:30:19 <rubasov> just a tiny thing
16:30:43 <rubasov> these rally floating ip patches are still up for review: https://review.openstack.org/#/q/topic:rally-neutron-fip+status:open
16:30:58 <rubasov> please add them to your review pile when you can
16:31:50 <slaweq> according to osprofiler and rally integration, I didn't make progress on https://review.openstack.org/#/c/615350/ - it basically works now but rally has to generate separate pages for each scenario, like available in http://logs.openstack.org/50/615350/34/check/neutron-rally-task/c573df5/results/workloads/
16:31:51 <patchbot> patch 615350 - neutron - DNM Try to integrate rally with osprofiler - 34 patch sets
16:31:52 <mlavalle> rubasov: added to my pile. will review today
16:32:13 <rubasov> mlavalle: thank you
16:32:20 <slaweq> what is still missing there is "bad" look of this osprofiler reports
16:32:48 <slaweq> it is basically available in "Scenario Data" tab but without good css and things like that
16:33:01 <slaweq> I hope I will be able to fix that issue on rally side soon
16:33:13 <mlavalle> slaweq: so, do we have what we need from the osprofiler team? or the progress that has to be made now depends on us?
16:33:27 <slaweq> mlavalle: from osprofiler side all is fine
16:33:41 <slaweq> from rally side we need better generation of this reports
16:34:01 <mlavalle> should we ping the really team?
16:34:04 <slaweq> on neutron side we will just need this one patch to be merged then and we should be good
16:34:17 <mlavalle> rally team^^^
16:34:32 <slaweq> I talked with rally team about that and I told them that I will try to help with it
16:34:44 <mlavalle> ok, cool
16:35:06 <slaweq> but it's not highest priority in my backlog now :/
16:35:22 <mlavalle> Great, slowly but surely we are making progress. Given all the things that we all have to do, I think that all in all this is solid progress
16:36:02 <mlavalle> Thanks for all your efforts slaweq and rubasov!!!!
16:37:05 <mlavalle> #topic On demand agenda
16:37:58 <njohnston_> Michael Bayer pushed a change that will help performance, as well as addressing StaleDataErrors on the standard attributes table; please take a look: https://review.openstack.org/635671
16:37:59 <patchbot> patch 635671 - neutron - Run revision bump operations en masse - 6 patch sets
16:38:22 <mlavalle> njohnston_: good point. Thanks for bringing it up
16:39:15 <mlavalle> I also wanted to encourage the team to look at https://www.openstack.org/videos/summits/barcelona-2016/chasing-1000-nodes-scale
16:39:47 <mlavalle> This is a presentation from the Barcelona summit that describes the 1000 node test that has been made previously
16:39:59 <mlavalle> you can download the slides from there
16:40:34 <mlavalle> anything else we should discuss today?
16:41:02 <rubasov> will look at it, thanks
16:41:45 <mlavalle> ok team.... thanks for all your efforts and your attendance to this meeting
16:41:56 <mlavalle> jrbalderrama_: we appreciate a lot your shouwing up
16:42:08 <mlavalle> we know it is late in the day for you
16:42:20 <jrbalderrama_> welcome
16:42:33 <mlavalle> have a great week you all
16:42:37 <mlavalle> #endmeeting