Tuesday, 2016-04-12

*** suro-patz has quit IRC00:12
*** fredli has quit IRC00:21
*** markvoelker has joined #openstack-performance00:35
*** markvoelker has quit IRC00:41
*** asti_ has joined #openstack-performance01:02
*** arnoldje has quit IRC01:13
*** zz_dimtruck is now known as dimtruck01:17
*** asti_ has quit IRC01:26
*** asti_ has joined #openstack-performance01:27
*** asti_ has quit IRC01:52
*** bapalm has quit IRC02:06
*** bapalm has joined #openstack-performance02:13
*** markvoelker has joined #openstack-performance02:37
*** markvoelker has quit IRC02:42
*** arnoldje has joined #openstack-performance02:49
*** furlongm has quit IRC03:07
*** dimtruck is now known as zz_dimtruck03:44
*** furlongm has joined #openstack-performance03:45
*** zz_dimtruck is now known as dimtruck04:14
*** dimtruck is now known as zz_dimtruck04:23
*** markvoelker has joined #openstack-performance04:38
*** markvoelker has quit IRC04:42
*** zz_dimtruck is now known as dimtruck04:44
*** dimtruck is now known as zz_dimtruck04:55
*** zz_dimtruck is now known as dimtruck05:06
*** dimtruck is now known as zz_dimtruck05:16
*** arnoldje has quit IRC05:23
*** markvoelker has joined #openstack-performance05:38
*** suro-patz has joined #openstack-performance05:39
*** markvoelker has quit IRC05:43
*** zz_dimtruck is now known as dimtruck06:07
*** dimtruck is now known as zz_dimtruck06:17
*** suro-patz has quit IRC06:51
*** zz_dimtruck is now known as dimtruck07:08
*** dimtruck is now known as zz_dimtruck07:19
*** nihilifer has quit IRC07:38
*** markvoelker has joined #openstack-performance07:39
*** markvoelker has quit IRC07:44
*** haukebruno has joined #openstack-performance08:01
*** zz_dimtruck is now known as dimtruck08:11
*** dimtruck is now known as zz_dimtruck08:20
*** dims has quit IRC09:04
*** dims has joined #openstack-performance09:05
*** zz_dimtruck is now known as dimtruck09:11
*** openstackgerrit has quit IRC09:17
*** openstackgerrit has joined #openstack-performance09:18
*** dimtruck is now known as zz_dimtruck09:21
*** markvoelker has joined #openstack-performance09:40
*** markvoelker has quit IRC09:45
*** openstack has joined #openstack-performance10:00
*** zz_dimtruck is now known as dimtruck10:12
odyssey4meHi everyone - in some recent testing we're finding that performance is significantly enhanced when the rpc_thread_pool_size, rpc_conn_pool_size and db_max_pool_size are adjusted. I'm wondering if any tests have been done to try and figure out whether the performance improvements made by adjusting these are somehow connected to CPU/Memory available? I'm thinking that it may be useful to have these dynamically set based on resources10:13
odyssey4me available.10:13
*** dimtruck is now known as zz_dimtruck10:22
*** tinbar has joined #openstack-performance10:33
*** rook has quit IRC10:58
*** zz_dimtruck is now known as dimtruck11:13
*** rook has joined #openstack-performance11:22
rookodyssey4me which services did you tune those in?11:22
rookodyssey4me I have played with those - however it was to get better db utilization11:22
*** dimtruck is now known as zz_dimtruck11:23
*** markvoelker has joined #openstack-performance11:41
*** markvoelker has quit IRC11:45
DinaBelovaodyssey4me I remember some experiments were done around the topic you mentioned. We do not have any results published yet, but I'm really recommending you to share *your* results - this will be super helpful for everyone here11:52
DinaBelova++ to rook about what services you sere playing with - that's interesting question, as all OpenStack services have different MQ/DB usage patterns11:53
*** rfolco has joined #openstack-performance11:53
odyssey4meyeah, so at the moment I only have small AIO test environments, so it's difficult for me to determine anything usefully conclusive12:04
odyssey4meI thought that perhaps we could collaborate on some sort of test plan and make use of the scale lab to execute on it12:04
odyssey4mefor the moment my focal point is just on nova/neutron - but I am thining that I need to isolate the testing to each service and I also think that each service's performance profile is quite different12:04
*** openstack has quit IRC12:04
*** openstack has joined #openstack-performance12:07
*** anevenchannyy has joined #openstack-performance12:07
*** markvoelker has joined #openstack-performance12:12
*** zz_dimtruck is now known as dimtruck12:14
*** dimtruck is now known as zz_dimtruck12:23
*** asti_ has joined #openstack-performance12:35
*** regXboi has joined #openstack-performance12:48
*** arnoldje has joined #openstack-performance12:57
*** aditi has joined #openstack-performance12:59
*** zz_dimtruck is now known as dimtruck13:05
rookodyssey4me so in our automation we have integrated a tool called Connmon13:05
rookto monitor the usage of the db connections13:05
rookhttps://bitbucket.org/zzzeek/connmon13:06
rook^ might be helpful in this work13:06
*** asti_ has quit IRC13:07
*** aditi has quit IRC13:12
*** asti_ has joined #openstack-performance13:16
*** asti_ has quit IRC13:28
*** rfolco has quit IRC13:39
DinaBelovaodyssey4me - you can propose the test plan to the performance-docs and we may figure out together what env this can be possibly run against - as some of the services do not require full big scale lab, and some - do. Can you propose the test plan so we can review it?13:41
*** rfolco has joined #openstack-performance13:54
*** arnoldje has quit IRC14:00
*** asti_ has joined #openstack-performance14:14
odyssey4meDinaBelova sure, I'd like to collaborate on this rather than do it iin isolation14:18
odyssey4methis is what the performance team is about, after all :)14:18
*** asingh has joined #openstack-performance14:29
*** asti_ has quit IRC14:32
*** spyderdyne has joined #openstack-performance14:32
*** arnoldje has joined #openstack-performance14:36
*** asti_ has joined #openstack-performance14:39
*** dimtruck is now known as zz_dimtruck14:52
*** dougs1 has joined #openstack-performance15:02
*** tinbar has quit IRC15:04
*** dougs has quit IRC15:05
*** asti_ has quit IRC15:14
*** openstackgerrit has quit IRC15:18
*** openstackgerrit has joined #openstack-performance15:18
*** zz_dimtruck is now known as dimtruck15:33
*** yarkot_ has joined #openstack-performance15:38
*** dimtruck is now known as zz_dimtruck15:43
*** bsilverman has joined #openstack-performance15:50
bsilvermanHi DinaBelova boris-42 dims et al.15:52
dimshey bsilverman15:52
DinaBelovao/15:52
bsilvermanDinaBelova: I'm going to send you a meeting invite for this week re: performance-docs and the test-plan template. We want to get this done while we are waiting for other things.15:54
DinaBelovabsilverman ack15:54
*** zz_dimtruck is now known as dimtruck15:55
openstackgerritMerged openstack/performance-docs: Add KloudBuster to the data plane testing tool page  https://review.openstack.org/30323015:57
bsilverman“First of all, do not try to scale-test clouds with 3 controllers15:59
bsilvermanrunning, for example,  >=10000 users' requests per a second. Use 5 controllers instead. And for >=100000 rps, use 7 perhaps.”15:59
bsilverman15:59
*** rohanion has joined #openstack-performance15:59
bsilvermanDid you see that DinaBelova?15:59
rookHeya15:59
DinaBelovabsilverman is that some email thread?15:59
DinaBelovahey guys :)15:59
* DinaBelova starting15:59
DinaBelova#startmeeting Performance Team15:59
openstackMeeting started Tue Apr 12 15:59:57 2016 UTC and is due to finish in 60 minutes.  The chair is DinaBelova. Information about MeetBot at http://wiki.debian.org/MeetBot.15:59
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:00
openstackThe meeting name has been set to 'performance_team'16:00
bsilvermanyes, in mos-dev, we will discuss later.16:00
rooko/16:00
DinaBelovao/16:00
rohaniono/16:00
DinaBelovabsilverman please write here the email topic, I'll take a look16:00
*** haukebruno has quit IRC16:00
DinaBelovaokay, folks :)16:01
DinaBelovalet's wait for a couple of moments16:01
DinaBelovato see if others will join16:01
bsilvermanPossible issues with RPC-related communications/messaging on scale "16:01
bsilvermanFrom aleksander16:01
DinaBelovabsilverman ack16:02
*** lezbar has joined #openstack-performance16:02
openstackgerritMerged openstack/performance-docs: Add MariaDB performance report  https://review.openstack.org/30055316:03
openstackgerritMerged openstack/performance-docs: Add tests results for marathon health check  https://review.openstack.org/30014916:03
DinaBelovaokay, so going to the action item16:03
DinaBelova#topic Action Items16:03
DinaBelova#link http://eavesdrop.openstack.org/meetings/performance_team/2016/performance_team.2016-04-05-16.00.html16:03
DinaBelovawe had several last time, one related to the review of rohanion 's code16:04
DinaBelovarohanion - btw there were some comments left by harlowja16:04
DinaBelovaplease take a look16:04
DinaBelovaand others are related to the keystone testinig16:04
DinaBelovaI had to rerun my measurements for Mitaka16:05
rohanionI've seen them, yes16:05
DinaBelovadue to sporadic connection issue I've observed on the env during my first run16:05
rohanionI will change this to a thread-local driver instance16:05
DinaBelova#info now http://docs.openstack.org/developer/performance-docs/test_results/keystone/index.html contains the valid results16:05
DinaBelovarohanion ack16:05
dimsbsilverman : please encourage aleksander to post to openstack-dev@ ML. not sure folks here have any context16:05
bsilvermandims: yes, just discussing that with Dina, thanks.16:06
DinaBelovadims - bsilverman just typed me he'll drive it16:06
DinaBelovawe just realized this topic is not covered in the openstack-dev yet16:06
DinaBelova:)16:06
bsilvermandims: I was confused as to where I sourced it after I pasted it.16:06
DinaBelova:) bsilverman no worries :)16:06
dimsthanks bsilverman and DinaBelova16:06
DinaBelovaso after keystone tests rerunning two bugs were observed16:07
DinaBelova#link https://bugs.launchpad.net/keystone/+bug/156740316:07
openstackLaunchpad bug 1567403 in OpenStack Identity (keystone) newton "Local context cache seems to work improperly" [Medium,In progress] - Assigned to Morgan Fainberg (mdrnstm)16:07
DinaBelova#link https://bugs.launchpad.net/keystone/+bug/156741316:07
openstackLaunchpad bug 1567413 in oslo.cache "Keystone fetches data from Memcache even if caching is explicitly turned off" [Undecided,Confirmed] - Assigned to Morgan Fainberg (mdrnstm)16:07
DinaBelovahappily Morgan did really great and quick analysis16:08
DinaBelovaso fixes are already on review16:08
dimsyay morgan!16:08
DinaBelova:)16:08
DinaBelovawe need to ask him to join :) as we're saying nice thingsa bout him :)16:08
DinaBelovaI'm going to continue my Keystone research on multinode16:09
DinaBelovawith failover cases coverage16:09
DinaBelovalet's see where it'll lead16:09
DinaBelova#action DinaBelova share status of Keystone multinode testing16:09
DinaBelovaso that's all I guess about action items we had last time16:10
* DinaBelova going to Summit Topic if no objections16:10
DinaBelova#topic Upcoming Summit discussion16:10
DinaBelovaas mentioned last time16:11
DinaBelova#info we're having WG discussion slot on the summit16:11
DinaBelova#link https://www.openstack.org/summit/austin-2016/summit-schedule/events/739916:11
DinaBelovathis will be two parts planned16:11
DinaBelova1) Mitaka results16:11
DinaBelova2) Newton plans16:11
DinaBelovaI've created an etherpad for you guys to participate :)16:12
DinaBelova#link https://etherpad.openstack.org/p/newton-performance-team16:12
DinaBelovaI'm really encouraging you folks to go through it and propose ideas of what you're going to go through during Newton16:13
DinaBelovaor possibly just ideas of testing that i may try to fit into our (Mirantis) testing schedule16:13
DinaBelovarook, rohanion, bsilverman, dims, odyssey4me ^^16:13
rookreviewing.16:14
bsilvermanDinaBelova: Where would we send, compile ideas? Is there a etherpad?16:14
rookDinaBelova Do we have a Trello to track work?16:14
DinaBelovabsilverman yes sir :) https://www.openstack.org/summit/austin-2016/summit-schedule/events/739916:14
DinaBelovabsilverman not yet, but possibly will be a nice idea16:14
bsilvermanDinaBelova: Thanks :)16:14
DinaBelovaor, sorry16:14
DinaBelovarook ^^16:14
DinaBelovathe last message referred to you16:14
rookoh :D16:15
DinaBelovaso again - please feel free to add your ideas, comments, whatever16:17
DinaBelovalet's collect as much info as possible16:17
DinaBelovaprior summit :)16:17
DinaBelovato have good basement for the discussion :)16:17
DinaBelovainteresting if harlowja is around16:17
rohanionwill take a look, tnx Dina16:18
DinaBelovaharlowja promised to check if he'll be able to give us some time for osprofiler discussion in terms of oslo track :)16:18
lezbarme too16:18
DinaBelovarohanion u'r welcome :)16:18
DinaBelovalezbar indeed16:18
*** dimtruck is now known as zz_dimtruck16:19
DinaBelovait looks like harlowja is still sleeping :)16:19
DinaBelovaokay, any questions regarding summit folks?16:19
boris-42DinaBelova: but I am here16:20
DinaBelovaboris-42 sorry?16:20
DinaBelovaboris-42 I'm happy you're here :)16:20
boris-42DinaBelova: finally* =)16:20
DinaBelovaahaha :)16:20
DinaBelovawell, it just means you're more brave today than harlowja :)16:20
DinaBelovaboris-42 btw16:21
boris-42DinaBelova: btw did you hear about team called opnfv-testperf16:21
DinaBelovaboris-42 nope :)16:21
boris-42DinaBelova: so they are testing16:21
boris-42DinaBelova: and you know what they are testing?16:21
DinaBelovaboris-42 - just in background - please take a look on https://review.openstack.org/#/c/247005/16:22
DinaBelovaboris-42 :)16:22
DinaBelovaboris-42 you may just tell me :)16:22
boris-42DinaBelova: oh going to run away=)16:22
DinaBelovalol16:22
boris-42DinaBelova: so they are testing OpenStack16:22
DinaBelovawell, too vague :)16:23
DinaBelovaany links, sir?16:23
boris-42DinaBelova: #opnfv-testperf - is their irc chat (it's opnfv community)16:23
DinaBelovaboris-42 ack16:23
DinaBelovaI'll take a look16:23
DinaBelovait looks like we may go to the next topic16:23
DinaBelova#topic OSProfiler weekly update16:23
rohanionokay16:24
rohanionso I've almost finished writing the ceilo driver16:24
DinaBelovaboris-42 I need some your efforts sir :) I need release to be cut after https://review.openstack.org/#/c/294516/ will land (btw strange it's still not landed)16:24
DinaBelovarohanion cool!16:24
rohanionHope to push the changes to the review today16:24
DinaBelovarohanion sadly I had no chance to try your code myself yet16:24
DinaBelovareally hope to do it this week16:25
rohanionno problem, take your time, it's possible that I'll run it myself16:25
DinaBelovarohanion ack16:25
rohanionafter finishing with the ceilo driver and the testing stuff16:25
DinaBelovaboris-42 - also one question :) I remember you wanted to add job to test osprofiler functionality - for other services. Like nova. That time lack of such tests was a reason for -216:26
DinaBelovarohanion thank you16:26
DinaBelovaboris-42 ? :)16:27
boris-42DinaBelova: so16:28
odyssey4methanks DinaBelova I'll take a peek and try to join the sessions, assuming they don't clash with the OpenStack-Ansible sessions :)16:28
DinaBelovaodyssey4me ack16:28
boris-42DinaBelova: add recheck to the patch16:29
boris-42DinaBelova: I will cut release16:29
DinaBelovaboris-42 - well, I was mostly pinging about the functional job :)16:29
DinaBelovabut thanks :)16:29
harlowjaDinaBelova hey16:29
harlowja:-P16:29
DinaBelovaharlowja good morning :D16:29
boris-42DinaBelova: so about func job.. hmm let me see what is missing16:29
harlowjayo yo16:30
boris-42DinaBelova: so but I already finished everything I remember, so we merged patches everywhere...16:30
boris-42DinaBelova: that should work khmkhm16:30
harlowjaso DinaBelova do u want a work session for oslo, i can free one up16:30
DinaBelovaboris-42 - hm... the job that really tests profiler functionality?16:30
DinaBelovaharlowja - I guess we may try :)16:30
harlowjaif u think it would be useful, then i'm fine with that16:31
DinaBelovawill be a good idea to understand what preferences people have - if any16:31
harlowjak16:31
harlowjathe fishbowl sessions i think i got filled with other oslo ones, but i think i can free up a workroom session16:31
DinaBelovain fact i do not know if this session will be productive - it depends much on the people's interest16:31
DinaBelovaharlowja ^^16:32
harlowjaya, that's always a question :-P16:32
DinaBelovaso if you're having some doubts.. heh16:32
DinaBelovaI do have them as well16:32
harlowjaboris-42 and DinaBelova and harlowja will be interested16:32
harlowjahopefully at least those 3 will show up :-P16:32
DinaBelovawell, that's true :)16:32
boris-42DinaBelova: so there is DSVM job16:32
boris-42DinaBelova: what we need is some tempest tests?)16:33
boris-42DinaBelova: and to enable it16:33
boris-42-)16:33
harlowjaDinaBelova ok, maybe think it over, let me know16:33
harlowjai'm ok with doing this16:33
DinaBelovaharlowja ack16:33
DinaBelovaboris-42 probably yes16:34
DinaBelovado you think you'll be able to work on them?16:34
boris-42DinaBelova: because job already exists16:34
boris-42DinaBelova: so not sure how to the magic with libs again16:34
boris-42DinaBelova: that's the major blocker16:35
DinaBelovaboris-42 may I ask you to investigate this field?16:35
DinaBelovaas we all really want nova change to land :)16:35
boris-42DinaBelova: so really not sure about this week...16:35
odyssey4memeteorfox ping - note https://www.openstack.org/summit/austin-2016/summit-schedule/events/7399 and https://etherpad.openstack.org/p/newton-performance-team16:36
DinaBelovaboris-42 - well, this week is not mandatory16:36
DinaBelovaboris-42 I'll try to look myself as well16:36
DinaBelova#info boris-42 and DinaBelova to take a look on the functional osprofiler tests to enable nova change merge16:36
DinaBelovaokay, so let's try to find time slot16:37
DinaBelovaand work on this16:37
DinaBelovaalso some news about OPW16:37
DinaBelovathe voting for candidates is almost finalized16:37
DinaBelovaand most probably we'll have one intern next cycle :)16:37
DinaBelovarohanion ^^16:38
DinaBelovafew days left for final decision16:38
rohanionI've received your email, yes16:38
DinaBelovarohanion ack16:38
DinaBelovaokay, it looks we closed OSprofiler topic16:38
DinaBelova#topic Open Discussion16:38
DinaBelovaanything else to go through folks?16:39
DinaBelovait looks nope16:39
DinaBelovaokay, thanks everyone for coming :)16:40
harlowjaDinaBelova when u coming to states16:40
harlowjalol16:40
*** asti_ has joined #openstack-performance16:40
DinaBelovasee you next week - it'll be our last meeting before summit :)16:40
harlowjaboris-42 where are u these days :-P16:40
DinaBelovaharlowja hhaha for summit ;)16:40
harlowja:)16:40
harlowjawhen's everyone arriving (for summit)16:40
DinaBelovaharlowja I'm going to arrive on Sunday16:41
boris-42harlowja: ah working16:41
boris-42-)16:41
rohanionnext year lol, not going16:41
harlowjaboris-42 u not going to austin?16:41
boris-42harlowja: going16:42
harlowjak16:42
DinaBelovaharlowja he's about tooo long boris-ping16:42
DinaBelova:)16:42
DinaBelovaok, so let's end the meeting if no specific topics - all other stuff we can cover without ticking bot :)16:42
harlowjalol16:42
DinaBelova#endmeeting16:42
harlowjak16:42
openstackMeeting ended Tue Apr 12 16:42:59 2016 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)16:43
openstackMinutes:        http://eavesdrop.openstack.org/meetings/performance_team/2016/performance_team.2016-04-12-15.59.html16:43
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/performance_team/2016/performance_team.2016-04-12-15.59.txt16:43
openstackLog:            http://eavesdrop.openstack.org/meetings/performance_team/2016/performance_team.2016-04-12-15.59.log.html16:43
*** blangston has joined #openstack-performance16:49
*** spyderdyne has quit IRC16:50
*** rohanion has quit IRC16:50
*** asti_ has quit IRC16:54
*** blangston has quit IRC16:57
*** AugieMena has joined #openstack-performance16:58
*** asti_ has joined #openstack-performance17:02
*** asti_ has quit IRC17:06
*** dougs has joined #openstack-performance17:12
*** asti_ has joined #openstack-performance17:13
*** zz_dimtruck is now known as dimtruck17:15
*** dougs1 has quit IRC17:16
*** suro-patz has joined #openstack-performance17:18
*** bsilverman has quit IRC17:19
*** AugieMena has quit IRC17:23
*** asti_ has quit IRC17:28
*** aditi has joined #openstack-performance17:29
*** blangston has joined #openstack-performance17:33
*** blangston has quit IRC17:33
*** aditi has quit IRC17:33
*** rfolco has quit IRC17:38
*** rfolco has joined #openstack-performance17:39
*** suro-patz1 has joined #openstack-performance17:50
*** suro-patz has quit IRC17:51
*** yarkot_ has quit IRC17:57
*** asingh has quit IRC17:57
*** asingh has joined #openstack-performance17:58
*** asti_ has joined #openstack-performance18:05
*** asti_ has quit IRC18:34
*** yarkot_ has joined #openstack-performance18:46
*** rfolco_ has joined #openstack-performance18:48
*** rfolco has quit IRC18:50
*** asti_ has joined #openstack-performance18:54
*** sc68cal has joined #openstack-performance19:44
*** sc68cal is now known as Guest4887319:44
*** asti_ has quit IRC19:46
*** asti_ has joined #openstack-performance19:51
*** Guest48873 has quit IRC19:54
*** asti_ has quit IRC19:54
*** yarkot_ has quit IRC19:55
*** sc68cal_ has joined #openstack-performance19:57
*** dimtruck is now known as zz_dimtruck20:19
*** asti_ has joined #openstack-performance20:33
*** yarkot_ has joined #openstack-performance20:35
*** zz_dimtruck is now known as dimtruck20:40
*** klindgren_ has joined #openstack-performance20:40
*** klindgren has quit IRC20:42
*** dimtruck is now known as zz_dimtruck20:50
*** zz_dimtruck is now known as dimtruck20:54
*** v12aml has quit IRC21:01
*** rfolco_ has quit IRC21:01
*** v12aml has joined #openstack-performance21:03
*** dimtruck is now known as zz_dimtruck21:14
*** asti_ has quit IRC21:23
*** haukebruno has joined #openstack-performance21:53
*** arnoldje has quit IRC22:10
*** asingh has quit IRC22:25
*** regXboi has quit IRC22:27
*** asingh has joined #openstack-performance22:39
*** arnoldje has joined #openstack-performance22:59
*** haukebruno has quit IRC23:07
*** asti_ has joined #openstack-performance23:29
*** rfolco_ has joined #openstack-performance23:38
*** rfolco_ has quit IRC23:50
*** yarkot_ has quit IRC23:55

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!