Monday, 2015-02-23

*** claudiub has quit IRC00:00
*** yamamoto has joined #openstack-infra00:05
*** MarkAtwood has joined #openstack-infra00:08
*** yamamoto has quit IRC00:08
morganfainberggod. that is awful ^^00:11
morganfainbergbut yeah it mostly can work00:12
dstufftiterating is sad though00:12
*** armax has joined #openstack-infra00:14
dstufftyou either have to iterate over ints, or do some c style for i in range(len(bytes_obj)): char = bytes_obj[i:i+1]00:14
*** markvoelker has joined #openstack-infra00:16
clarkbclearly you need a def char_iter(bytes_obj): for i in range(len(bytes_obj): yield bytes_obj[i:i+1] iterator in stdlib00:18
*** markvoelker has quit IRC00:20
*** armax has quit IRC00:32
*** rm_work is now known as rm_work|away00:35
*** camunoz_ has joined #openstack-infra00:37
*** camunoz has quit IRC00:39
*** hdd has joined #openstack-infra00:39
mordredclarkb: we have the "split a repo" cantrip documented somewhere, right?00:47
clarkbin the puppet split out spec00:48
clarkbmordred ^00:48
mordredkk. thanks00:48
*** garyh has joined #openstack-infra00:48
*** yamamoto has joined #openstack-infra00:48
*** tmcpeak has joined #openstack-infra00:49
*** markvoelker has joined #openstack-infra00:50
*** amotoki has joined #openstack-infra00:53
*** tmcpeak has quit IRC00:53
*** markvoelker has quit IRC00:55
*** rkukura has quit IRC00:56
*** rkukura has joined #openstack-infra00:58
*** rkukura has quit IRC01:03
*** harlowja_at_home has joined #openstack-infra01:06
*** sdake_ has quit IRC01:07
*** boris-42 has joined #openstack-infra01:13
openstackgerritClint 'SpamapS' Byrum proposed openstack-infra/gear: Fix byte related python3 behaviors
*** ghostpl_ has joined #openstack-infra01:19
*** achanda has quit IRC01:21
*** salv-orlando has quit IRC01:22
*** ghostpl_ has quit IRC01:23
*** sdake_ has joined #openstack-infra01:27
*** baoli has joined #openstack-infra01:28
*** wenlock has joined #openstack-infra01:30
*** sdake__ has joined #openstack-infra01:31
*** tmcpeak has joined #openstack-infra01:32
*** ociuhandu has quit IRC01:32
*** sdake_ has quit IRC01:33
*** dimsum__ has joined #openstack-infra01:38
openstackgerritClint 'SpamapS' Byrum proposed openstack-infra/gear: Fix byte related python3 behaviors
*** dims_ has joined #openstack-infra01:39
*** dkranz has quit IRC01:39
openstackgerritClint 'SpamapS' Byrum proposed openstack-infra/gear: Fix byte related python3 behaviors
SpamapSwoot, py34 passing01:40
* SpamapS should probably squash that into one instead of 3 patches01:41
*** baoli has quit IRC01:41
*** armax has joined #openstack-infra01:41
*** dimsum__ has quit IRC01:43
*** amotoki has quit IRC01:44
*** hdd has quit IRC01:44
openstackgerritMerged openstack-dev/hacking: Updated from global requirements
openstackgerritWayne Warren proposed openstack-infra/jenkins-job-builder: Fix bug in template job yaml expansion.
*** markvoelker has joined #openstack-infra01:51
openstackgerritWayne Warren proposed openstack-infra/jenkins-job-builder: Support the use of lists in trigger-build publisher.
*** markvoelker has quit IRC01:57
openstackgerritWayne Warren proposed openstack-infra/jenkins-job-builder: Support the use of lists in trigger-build builder.
openstackgerritWayne Warren proposed openstack-infra/jenkins-job-builder: Support the use of lists in trigger-build builder.
openstackgerritWayne Warren proposed openstack-infra/jenkins-job-builder: Add Parameterized Remote Trigger Plugin support.
*** markvoelker has joined #openstack-infra02:00
*** baoli has joined #openstack-infra02:02
*** MarkAtwood has quit IRC02:03
*** ddieterl_ has joined #openstack-infra02:06
*** MarkAtwood has joined #openstack-infra02:08
*** ddieterly has quit IRC02:09
*** signed8bit has joined #openstack-infra02:13
*** tmcpeak has quit IRC02:13
*** MarkAtwood has quit IRC02:15
*** baoli has quit IRC02:16
*** MarkAtwood has joined #openstack-infra02:21
*** stevemar has joined #openstack-infra02:21
*** achanda has joined #openstack-infra02:22
*** wenlock has quit IRC02:26
*** markvoelker has quit IRC02:26
*** markvoelker has joined #openstack-infra02:27
*** achanda has quit IRC02:27
*** achanda has joined #openstack-infra02:27
*** markvoelker has quit IRC02:31
*** yamahata has quit IRC02:32
*** yamahata has joined #openstack-infra02:32
*** vilobhmm has quit IRC02:36
*** MarkAtwood has quit IRC02:37
*** achanda has quit IRC02:38
openstackgerritBrian Curtin proposed openstack/requirements: Avoid httpretty 0.8.7 release
*** koolhead17 has joined #openstack-infra02:43
openstackgerritBrian Curtin proposed openstack/requirements: Avoid httpretty 0.8.7 release
*** asettle is now known as asettle-luncheon02:45
*** yamahata has quit IRC02:46
openstackgerritMerged openstack-infra/system-config: Add Elizabeth K. Joseph to servers and storyboard
*** Ryan_Lane has joined #openstack-infra02:58
*** sdake_ has joined #openstack-infra02:58
clarkbpleia2 ^ have fun :)02:58
*** sdake__ has quit IRC03:02
*** Ryan_Lane has quit IRC03:02
*** mriedem has quit IRC03:06
anteayalooks like mordred's woot! also confirms your assumption clarkb03:06
*** ghostpl_ has joined #openstack-infra03:07
*** wenlock has joined #openstack-infra03:09
*** ghostpl_ has quit IRC03:12
*** otter768 has quit IRC03:14
*** MarkAtwood has joined #openstack-infra03:19
*** amotoki has joined #openstack-infra03:19
*** baoli has joined #openstack-infra03:27
*** oomichi_ has joined #openstack-infra03:30
*** baoli has quit IRC03:42
*** bhunter71 has quit IRC03:43
*** ddieterl_ has quit IRC03:47
*** harlowja_at_home has quit IRC03:53
jheskethoh clarkb, are you still around?03:53
clarkbnot really03:54
clarkbwhats up?03:54
jheskethheh, all good, I'll catch you tomorrow03:54
*** asettle-luncheon is now known as asettle03:55
clarkbdidnt get around to the swift thing. or zuul reviews :(03:55
jheskethno trouble :-)03:56
*** weshay has quit IRC04:02
*** _nadya_ has joined #openstack-infra04:10
*** wenlock has quit IRC04:10
*** vilobhmm has joined #openstack-infra04:12
openstackgerritMonty Taylor proposed openstack-infra/project-config: Add ansible role for running puppet
openstackgerritMonty Taylor proposed openstack-infra/system-config: Install standalone ansible roles
openstackgerritMonty Taylor proposed openstack-infra/puppet-ansible: Put the library file into the role
openstackgerritMonty Taylor proposed openstack-infra/puppet-ansible: Remove embedded ansible role
mordredclarkb: ^^ those are precursors to the other stuff, but I wanted to be able to submit the next set of patches on top of the new repo04:16
SpamapSmordred: only 4? LA to NYC is a long flight.. ;)04:17
mordredSpamapS: :)04:18
mordredI have some more in waiting04:18
*** deepakcs has joined #openstack-infra04:19
*** rwsu-afk has quit IRC04:19
*** armax_ has joined #openstack-infra04:19
*** signed8bit has quit IRC04:20
*** armax has quit IRC04:20
*** armax_ is now known as armax04:20
*** armax has quit IRC04:22
*** ivar-lazzaro has quit IRC04:22
*** vilobhmm1 has joined #openstack-infra04:28
*** BharatK has joined #openstack-infra04:28
*** vilobhmm has quit IRC04:30
*** MarkAtwood has quit IRC04:31
*** bswartz has quit IRC04:34
*** _nadya_ has quit IRC04:34
*** bswartz has joined #openstack-infra04:35
*** unicell has quit IRC04:46
*** unicell has joined #openstack-infra04:47
*** koolhead17 has quit IRC04:52
openstackgerritDavid Pursehouse proposed openstack-infra/jenkins-job-builder: Add support to customize trigger-builds block thresholds
*** ghostpl_ has joined #openstack-infra04:56
*** deepakcs has quit IRC05:00
*** ghostpl_ has quit IRC05:01
notmynameclarkb: jhesketh: I just got in from a long weekend away. I saw "swift" mentioned a few times in here. anything I should be paying attention to?05:01
jheskethnotmyname: nothing worrying ;-). We've just been working on making the uploads a little faster05:03
jhesketh if you're interested05:04
notmynamecool. ok, thanks05:04
notmynameoh, nice (from the commit message)05:04
*** garyh has quit IRC05:09
*** vigneshvar has joined #openstack-infra05:13
*** otter768 has joined #openstack-infra05:15
openstackgerritRussell Bryant proposed openstack-infra/project-config: Add stackforge/networking-ovn project
*** baoli has joined #openstack-infra05:16
*** sdake_ has quit IRC05:18
*** otter768 has quit IRC05:20
*** HeOS has joined #openstack-infra05:23
*** baoli has quit IRC05:28
*** hdd has joined #openstack-infra05:29
*** baoli has joined #openstack-infra05:32
*** achanda has joined #openstack-infra05:35
*** yfried has quit IRC05:44
*** pcaruana has quit IRC05:45
*** baoli has quit IRC05:46
*** ddieterly has joined #openstack-infra05:49
*** ddieterly has quit IRC05:53
*** Ryan_Lane has joined #openstack-infra05:59
*** camunoz_ is now known as camunoz_gone05:59
*** Ryan_Lane has quit IRC06:03
*** baoli has joined #openstack-infra06:04
*** garyh has joined #openstack-infra06:10
*** baoli has quit IRC06:18
*** achanda has quit IRC06:25
openstackgerritAdrian Moisey proposed openstack-infra/jenkins-job-builder: Add support for github commit status pending
*** cnesa has joined #openstack-infra06:25
*** coolsvap_ is now known as coolsvap06:27
*** ildikov has quit IRC06:38
*** MarkAtwood has joined #openstack-infra06:39
*** woodster_ has quit IRC06:40
*** tnovacik has joined #openstack-infra06:40
*** cnesa has quit IRC06:43
*** FallenPegasus has joined #openstack-infra06:44
*** MarkAtwood has quit IRC06:45
*** FallenPegasus is now known as MarkAtwood06:45
*** ghostpl_ has joined #openstack-infra06:45
*** deepakcs has joined #openstack-infra06:45
*** ddieterly has joined #openstack-infra06:50
*** ghostpl_ has quit IRC06:50
*** hdd has quit IRC06:54
*** ddieterly has quit IRC06:54
*** achanda has joined #openstack-infra06:56
*** jcoufal has joined #openstack-infra06:57
*** cnesa has joined #openstack-infra07:00
*** vilobhmm1 has quit IRC07:01
*** mrunge has joined #openstack-infra07:04
*** unicell has quit IRC07:06
*** afazekas_ has joined #openstack-infra07:06
*** scheuran has joined #openstack-infra07:08
*** belmoreira has joined #openstack-infra07:10
*** unicell has joined #openstack-infra07:10
*** Murad has joined #openstack-infra07:13
*** yfried has joined #openstack-infra07:13
MuradIt seems that devstack now requires python-novaclient 2.21 to work07:13
Muradapt-get installs 2.17, how can I get 2.2107:13
*** vilobhmm has joined #openstack-infra07:14
*** otter768 has joined #openstack-infra07:16
*** yfried_ has joined #openstack-infra07:17
*** yfried has quit IRC07:17
*** stevemar has quit IRC07:19
*** mrmartin has joined #openstack-infra07:20
*** otter768 has quit IRC07:21
*** ildikov has joined #openstack-infra07:32
*** tnovacik has quit IRC07:40
*** vilobhmm1 has joined #openstack-infra07:42
*** achanda has quit IRC07:42
*** vilobhmm has quit IRC07:44
*** yfried_ is now known as yfried|afk07:44
*** hrubi has joined #openstack-infra07:47
*** achanda has joined #openstack-infra07:48
*** sergsh has joined #openstack-infra07:48
*** jlibosva has joined #openstack-infra07:48
*** e0ne has joined #openstack-infra07:49
*** e0ne is now known as e0ne_07:49
*** ddieterly has joined #openstack-infra07:50
*** baoli has joined #openstack-infra07:52
*** doude has joined #openstack-infra07:52
*** luqas has joined #openstack-infra07:52
*** e0ne_ is now known as e0ne07:53
openstackgerrityolanda.robla proposed openstack-infra/storyboard: Add tools to run integration tests
*** ddieterly has quit IRC07:55
*** yfried|afk is now known as yfried_07:56
*** MarkAtwood has quit IRC07:58
*** MarkAtwood has joined #openstack-infra08:00
*** baoli has quit IRC08:06
*** yfried_ is now known as yfried|afk08:13
openstackgerritRadomir Dopieralski proposed openstack/requirements: Add ironic-discoverd
*** ttx2 has joined #openstack-infra08:15
*** ttx2 is now known as ttx08:15
*** vilobhmm1 has quit IRC08:16
*** yfried|afk is now known as yfried_08:16
*** ghostpl_ has joined #openstack-infra08:16
*** ttx has quit IRC08:17
*** ttx has joined #openstack-infra08:17
*** ttx has quit IRC08:17
*** ttx has joined #openstack-infra08:18
*** ttx has quit IRC08:19
*** ttx has joined #openstack-infra08:19
*** ghostpl_ has quit IRC08:20
*** Ala has joined #openstack-infra08:21
*** achanda has quit IRC08:21
*** achanda has joined #openstack-infra08:22
*** jcoufal has quit IRC08:22
*** jcoufal has joined #openstack-infra08:23
openstackgerrityolanda.robla proposed openstack-infra/shade: Add caching interval to mutable caches
*** luqas has quit IRC08:31
*** achanda has quit IRC08:32
*** dtantsur|afk is now known as dtantsur08:35
*** achanda has joined #openstack-infra08:45
*** MaxV has joined #openstack-infra08:46
*** mpavone has joined #openstack-infra08:46
*** ildikov has quit IRC08:50
*** ddieterly has joined #openstack-infra08:51
*** jistr has joined #openstack-infra08:55
*** ddieterly has quit IRC08:56
*** e0ne has quit IRC08:57
*** viktors|afk is now known as viktors08:58
*** tnovacik has joined #openstack-infra09:03
*** hashar has joined #openstack-infra09:04
*** hashar has quit IRC09:05
*** sergsh has quit IRC09:06
*** ildikov has joined #openstack-infra09:06
*** mpaolino has joined #openstack-infra09:08
*** devvesa has joined #openstack-infra09:12
*** devvesa has quit IRC09:14
*** devvesa has joined #openstack-infra09:14
*** achanda has quit IRC09:16
*** otter768 has joined #openstack-infra09:17
*** bmwiedemann has left #openstack-infra09:19
*** sergsh has joined #openstack-infra09:20
*** otter768 has quit IRC09:22
*** yamamoto has quit IRC09:23
*** hichihara has quit IRC09:24
*** sergsh has quit IRC09:27
*** ghostpl_ has joined #openstack-infra09:29
*** salv-orlando has joined #openstack-infra09:33
*** sergsh has joined #openstack-infra09:33
*** andreykurilin_ has joined #openstack-infra09:34
*** oomichi_ has quit IRC09:34
*** mpavone has quit IRC09:36
*** koolhead17 has joined #openstack-infra09:36
*** koolhead17 has quit IRC09:37
*** koolhead17 has joined #openstack-infra09:37
*** koolhead17 has joined #openstack-infra09:37
*** mpavone has joined #openstack-infra09:40
*** jp_at_hp has joined #openstack-infra09:43
*** bogdando has joined #openstack-infra09:45
*** pblaho has joined #openstack-infra09:47
*** lucasagomes has joined #openstack-infra09:48
lucasagomessdague, created the patch using DEVSTACK_LOCAL_CONFIG, apparently it's very new and I couldn't find any example of use of this var in devstack. So lemme know if the usage is correct on my patch. Thanks09:49
*** yamamoto has joined #openstack-infra09:51
*** ddieterly has joined #openstack-infra09:52
*** luqas has joined #openstack-infra09:53
*** ddieterly has quit IRC09:56
*** ssam2 has joined #openstack-infra09:58
*** luqas has quit IRC09:58
*** dizquierdo has joined #openstack-infra09:58
*** e0ne has joined #openstack-infra09:59
*** ParsectiX has joined #openstack-infra10:02
*** claudiub has joined #openstack-infra10:03
*** salv-orlando has quit IRC10:04
*** ihrachyshka has joined #openstack-infra10:07
*** sergsh has quit IRC10:14
*** salv-orlando has joined #openstack-infra10:14
*** marcusvrn has joined #openstack-infra10:18
*** e0ne is now known as e0ne_10:21
openstackgerritJohn Vandenberg proposed openstack-infra/git-review: Update project links
*** marcusvrn1 has joined #openstack-infra10:23
*** andreykurilin_ has quit IRC10:23
*** marcusvrn has quit IRC10:25
*** marcusvrn1 has quit IRC10:25
*** sergsh has joined #openstack-infra10:26
*** e0ne_ is now known as e0ne10:28
*** marcusvrn has joined #openstack-infra10:29
*** koolhead17 has quit IRC10:29
*** yfried_ is now known as yfried|afk10:41
*** vigneshvar has quit IRC10:42
*** rlandy has joined #openstack-infra10:45
*** ddieterly has joined #openstack-infra10:53
*** armax has joined #openstack-infra10:53
*** ddieterly has quit IRC10:57
*** yamamoto has quit IRC10:58
*** teran has joined #openstack-infra11:05
*** jaypipes has joined #openstack-infra11:05
*** yfried|afk is now known as yfried_11:06
*** pc_m has joined #openstack-infra11:07
openstackgerritIhar Hrachyshka proposed openstack/requirements: Don't use git diff --quiet to detect local change
ihrachyshkasdague, ^^ we'll see whether it fixes the issues that adam_g and mriedem saw yesterday.11:11
*** armax has quit IRC11:12
yolandahi, jhesketh, thx for all reviews11:12
sdaguelucasagomes: it's not part of devstack, it's part of the tooling around devstack11:14
sdaguelook at openstack-infra/project-config and openstack-infra/devstack-gate11:14
lucasagomessdague, ohh damn! gotcha, lemme take a look11:14
lucasagomessdague, sorry, I'm not familiar with those bits at all11:15
*** yfried_ is now known as yfried|afk11:16
*** yfried|afk is now known as yfried_11:17
*** otter768 has joined #openstack-infra11:18
*** salv-orlando has quit IRC11:19
*** andreaf_ has joined #openstack-infra11:20
sdaguelucasagomes: those are the parts that actually do all the setup and configure for devstack for all the tests11:20
lucasagomessdague, yeah found some example in the project-config. Putting a patch up in a second11:21
lucasagomesthanks for pointing it out11:21
*** andreaf_ has quit IRC11:22
*** otter768 has quit IRC11:22
*** bogdando has left #openstack-infra11:24
*** MarkAtwood has quit IRC11:24
openstackgerritLucas Alvares Gomes proposed openstack-infra/project-config: Ironic: Bump memory size for non-concurrency job
lucasagomessdague, ^ will wait for the tests see if works. Again, thanks for pointing that out and sorry for the previous misunderstanding11:25
*** salv-orlando has joined #openstack-infra11:25
*** SumitNaiksatam has quit IRC11:26
*** SumitNaiksatam has joined #openstack-infra11:26
*** cnesa has quit IRC11:27
*** ldnunes has joined #openstack-infra11:27
*** salv-orlando has quit IRC11:29
*** salv-orl_ has joined #openstack-infra11:29
openstackgerritLucas Alvares Gomes proposed openstack-infra/project-config: Ironic: Bump memory size for non-concurrency job
*** tnovacik has quit IRC11:40
*** baoli has joined #openstack-infra11:41
*** dhritishikhar_ has joined #openstack-infra11:44
openstackgerritSebastian Kalinowski proposed openstack-infra/project-config: Add fuel-tasks-validator project to stackforge
*** mpaolino has quit IRC11:50
*** ddieterly has joined #openstack-infra11:53
*** prmtl has joined #openstack-infra11:55
*** thelexx is now known as lxsli11:56
*** ddieterly has quit IRC11:58
*** mugsie has joined #openstack-infra12:03
*** spredzy is now known as spredzy|afk12:07
*** dkliban_afk is now known as dkliban12:09
*** rfolco has joined #openstack-infra12:09
*** jcoufal has quit IRC12:10
*** jcoufal_ has joined #openstack-infra12:10
*** dmsimard_away is now known as dmsimard12:10
*** yamamoto_ has joined #openstack-infra12:14
*** Kiall has quit IRC12:17
*** Kiall has joined #openstack-infra12:19
*** vhoward has quit IRC12:19
*** lucasagomes is now known as lucas-hungry12:24
*** tnovacik has joined #openstack-infra12:25
*** adalbas has joined #openstack-infra12:29
*** aysyd has joined #openstack-infra12:30
*** AJaeger has joined #openstack-infra12:33
*** AJaeger has joined #openstack-infra12:33
*** mugsie has quit IRC12:33
*** mugsie has joined #openstack-infra12:33
*** mugsie is now known as Guest9442312:34
prmtlhi, I kindly ask for review:
*** Guest94423 has quit IRC12:35
*** Guest94423 has joined #openstack-infra12:35
*** Guest94423 is now known as muguie12:36
*** muguie is now known as mugsie_12:36
*** spredzy|afk is now known as spredzy12:37
*** salv-orl_ has quit IRC12:38
*** salv-orlando has joined #openstack-infra12:39
*** weshay has joined #openstack-infra12:40
*** cdent has joined #openstack-infra12:41
*** salv-orlando has quit IRC12:42
*** salv-orlando has joined #openstack-infra12:43
*** dims_ has quit IRC12:44
*** derekh has joined #openstack-infra12:45
*** sdake_ has joined #openstack-infra12:48
KiallHey folks - The latest release of eventlet is causing some issues for Designate due to a broken monkey patch of dns.reversename.from_address() .. What's the deal with capping the requirement? We've never needed to before...12:49
*** koolhead17 has joined #openstack-infra12:50
*** dimsum__ has joined #openstack-infra12:51
*** zigo has quit IRC12:52
*** ZZelle has joined #openstack-infra12:52
*** marcusvrn1 has joined #openstack-infra12:53
*** ddieterly has joined #openstack-infra12:54
*** zigo has joined #openstack-infra12:56
*** marcusvrn has quit IRC12:57
*** ddieterly has quit IRC12:58
*** pelix1 has joined #openstack-infra12:59
*** sdake_ has quit IRC13:00
*** dprince has joined #openstack-infra13:00
*** deepakcs has quit IRC13:00
*** BharatK has quit IRC13:01
*** nicodemos has joined #openstack-infra13:02
*** salv-orlando has quit IRC13:04
openstackgerritKiall Mac Innes proposed openstack/requirements: Block eventlet 0.17.0
*** sdake_ has joined #openstack-infra13:08
*** dkranz has joined #openstack-infra13:09
*** kgiusti has joined #openstack-infra13:09
*** ParsectiX has quit IRC13:10
*** mwagner_lap has joined #openstack-infra13:13
*** ParsectiX has joined #openstack-infra13:16
*** otter768 has joined #openstack-infra13:19
*** bswartz has quit IRC13:20
*** dimsum__ has quit IRC13:22
*** dimsum__ has joined #openstack-infra13:22
*** mugsie_ is now known as mugsie13:23
*** salv-orlando has joined #openstack-infra13:23
*** lucas-hungry is now known as lucasagomes13:23
*** otter768 has quit IRC13:24
openstackgerritMerged openstack/requirements: Don't use git diff --quiet to detect local change
*** marcusvrn has joined #openstack-infra13:28
*** salv-orlando has quit IRC13:28
*** mbacchi has joined #openstack-infra13:29
*** gordc has joined #openstack-infra13:30
*** marcusvrn1 has quit IRC13:31
*** salv-orlando has joined #openstack-infra13:33
*** rfolco has quit IRC13:36
*** rfolco has joined #openstack-infra13:36
*** dimsum__ is now known as dims13:39
*** salv-orlando has quit IRC13:41
*** bknudson has quit IRC13:41
fungiKiall: i have a vague memory of eventlet and dnspython monkey patching each other out of existence... maybe you need something like keystone did in ?13:44
KiallNah, we really do need the eventlet patching of dnspython .. They've just managed to break a function which doesn't even do I/O with 0.17..13:45
KiallYep :)13:45
*** salv-orlando has joined #openstack-infra13:46
fungiahh, okay, followed the maze of links to the upstream bug report now13:46
KiallLOL - The commit that broke us, is the commit that fixes the issue Keystone mentioned in their commit message ;)13:47
*** woodster_ has joined #openstack-infra13:47
*** Guest30952 has joined #openstack-infra13:48
*** lucasagomes has quit IRC13:50
*** bhunter71 has joined #openstack-infra13:50
*** bswartz has joined #openstack-infra13:50
*** tnovacik has quit IRC13:53
*** salv-orlando has quit IRC13:53
*** ddieterly has joined #openstack-infra13:55
fungisuch irony13:58
*** ayoung has joined #openstack-infra14:00
KiallYep, I might have a workaround that doesn't require the cap... But there might be more breakage.. What's the deal with accepting a cap like that for a non-core project? :)14:00
*** ddieterly has quit IRC14:00
*** bknudson has joined #openstack-infra14:05
*** dtantsur is now known as dtantsur|brb14:06
*** dkranz has quit IRC14:06
fungiit's up to the requirements-core reviewers to decide, generally14:09
*** mfink_ has joined #openstack-infra14:10
*** dhritishikhar_ has quit IRC14:10
*** sdake_ has quit IRC14:10
fungithe infra team usually stays out of the discussion unless it's a review around the tooling for requirements validation/sync or something causing widespread breakage14:11
Kiallfungi: makes sense :)14:11
*** mjturek1 has joined #openstack-infra14:12
*** Guest28285 is now known as mgagne14:13
*** mgagne has joined #openstack-infra14:13
sdagueafazekas: that's a pretty basic service start problem14:17
mordredKiall: is there any version which works for both of you?14:18
Kiallmordred: yep, anything other than the version eventlet released today ;)14:18
mordredthat's so awesome14:18
*** dustins has joined #openstack-infra14:19
KiallI submitted this o/r cap to block 0.17, ;)14:19
*** dkranz has joined #openstack-infra14:19
sdagueafazekas: yeh, that service hasn't started on any job for some long amount of time14:19
*** sdake_ has joined #openstack-infra14:19
mordredI mean, I'd be inclined to block 0.17 just on the basis of we know it's provably broken in some way - but i'm still coffeeing14:19
mordredKiall: that said, I'd probably express that as !=0.17.0 so that we don't have to do another round when they release a fix14:20
afazekas_sdague: it fails because the dependency is not installed, it is specified as test time only dependency.14:20
Kiallmordred: Yea, I've got a workaround that unblocks us while jeeping 0.17.. I just have no idea if anything else is broken and we just haven't noticed.. I'm having trouble tracing the root casue even though I have the eventlet commit that broke things ..14:20
*** tiswanso has joined #openstack-infra14:20
*** sdake__ has joined #openstack-infra14:20
sdagueafazekas_: right, which means ceilometer is broken, because if it's an optional driver why is it always installing?14:20
afazekas_sdague: I am not sure is it optional or not, but AFAIK we have tempest+swift+ceilometer test and that module  is for swift support14:22
sdagueafazekas: it's not for swift support, it's for ceph support14:22
*** sdake_ has quit IRC14:24
Kialllooks like eventlet will do a bugfix release real-soon :) I love responsive upstreams!14:25
afazekas_sdague: rgw :) ok.  Looks like there is no option for turning it off.14:28
sdagueafazekas: yeh, that's a bug14:28
sdaguein ceilometer14:28
sdagueit should be listed as a critical bug14:29
*** e0ne is now known as e0ne_14:30
*** e0ne_ is now known as e0ne14:31
afazekas_sdague: If someone would like to use the rgw module, but the pip forgets to install the dependency it is also a bug, especially if the import is unconditional.14:32
*** stevemar has joined #openstack-infra14:33
sdagueit's a driver, we've always put the burden on the user to do that install14:33
sdagueafazekas: is there a bug?14:34
afazekas_sdague: I am do not know do have any generic guideline/strategy for how to handle optional stuff at openstack and what is the ceilometer policy. Usually there is 3 option. A: the python package depends on all possible runtime dependency. B, conditional import. C, splitting the package to multiple packages.  For strategy A the requirement change is the right step.14:36
gordcjust an fyi: we're looking at ceilo issue now.14:36
openstackLaunchpad bug 1424657 in Ceilometer "ceilometer unconditionally requires rados gateway support" [Critical,New]14:36
sdaguegordc: yeh, that needs to be fixed ASAP, currently blocking devstack14:36
*** mrmartin has quit IRC14:36
sdagueafazekas: A is not what we do14:37
*** ddieterly has joined #openstack-infra14:37
sdagueif it's a driver, it should be something in a config file that's specified14:37
*** amotoki has quit IRC14:37
*** claudiub has quit IRC14:38
*** jcoufal_ has quit IRC14:38
*** jcoufal has joined #openstack-infra14:38
afazekas_sdague: At the moment the issue breaks devstack changes, the B or C requires more time.14:38
gordcsdague: yeah, we're looking more to to B.14:39
Kiallsdague: digging into eventlet a few mins ago, and I spotted their tox.ini does some interesting stuff for drivers which I've not seen before:
Kiall(semi related, but no use to devstack ;))14:40
*** mriedem has joined #openstack-infra14:40
sdaguegordc: yeh, it seems like this is a driver, so should be specified in a config file and only loaded if specified14:41
gordcsdague: agreed (main reason why we forced it into test-reqs)14:41
*** Guest30952 is now known as annegentle14:42
sdaguegordc: I am kind of curious though, 0 tests for ceilometer see if services can start with just requirements?14:43
*** dhritishikhar_ has joined #openstack-infra14:43
sdaguebecause, the fact that this service is straight dead and nothing noticed, seems .... odd14:43
*** e0ne is now known as e0ne_14:44
gordcsdague: we don't have any tempest test relating to pollster because i don't think we every fleshed out how to capture data without modifying polling interval (iirc)14:45
*** lucasagomes has joined #openstack-infra14:46
sdaguegordc: you should make sure this is something that gets tested for in your functional tests14:46
*** sabeen has joined #openstack-infra14:46
gordcsdague: sure. seems like a reasonable suggestion.14:47
*** mrmartin has joined #openstack-infra14:47
lucasagomeshi, if someone have a time please take a look at, in Ironic we this memory bump so that we can start tests for the new ramdisk in gate. thank you14:48
*** beekneemech is now known as bnemec14:48
odyssey4melooking for a +2 +w in this small review which has been sitting for a few days:
HenryGHi folks! How do we get the "Rebase Change" button enabled for a stackforge repo?14:51
odyssey4meHenryG the stackforge repo I do a lot of work in (os-ansible-deployment) has it already enabled14:52
odyssey4meperhaps you could use it for comparison14:52
HenryGodyssey4me: remind me where the repo config is?14:52
odyssey4meHenryG I'm not sure if it's in project-config or somewhere else... I've only been involved in patches for tests - we had the rebase button right from the initiation of the project!14:53
*** superdan is now known as dansmith14:53
fungiHenryG: odyssey4me: it should already be enabled for all projects globally14:54
*** e0ne_ has quit IRC14:54
fungiHenryG: odyssey4me: it will only be displayed if gerrit thinks it can trivially rebase the change (and the jgit gerrit uses is not nearly so robust at figuring out how to handle merge conflicts as cgit, so just because a rebase seems trivial doesn't mean gerrit is going to be able to figure out how to make it happen)14:55
*** e0ne has joined #openstack-infra14:56
HenryGfungi: thanks, yes it is enabled, I was looking at a review that gerrit did not think was trivial :)14:56
*** signed8bit has joined #openstack-infra14:56
*** dtantsur|brb is now known as dtantsur14:57
*** deepakcs has joined #openstack-infra14:57
*** patrickeast has joined #openstack-infra14:58
*** wenlock has joined #openstack-infra14:59
*** tmcpeak has joined #openstack-infra14:59
*** wenlock has quit IRC14:59
*** wenlock has joined #openstack-infra15:00
*** tnovacik has joined #openstack-infra15:02
deepakcsfungi, Hi GM. Can u help with one more manul run, we seem to have some insight into the oom issue15:03
fungideepakcs: sure, i'll set one going in a few minutes15:03
deepakcsfungi, thanks, i am hoping this would be the last i would be troubling u :)15:04
clarkbdeepakcs have you tried reproducing this on a node with 8GB of ram?15:05
deepakcsclarkb, we tried with 6GB and it doesn't re-crate on my local setup, even copied the same tempest conf and local rc15:05
afazekas_gordc: so what to now quickly ?  (A, dep change, B, revert the ceilo change, C simple conditional import, D make the devstack job non voting , E wait until some makes it properly configurable )15:05
*** erikmwil_ has joined #openstack-infra15:06
*** erikmwilson has quit IRC15:06
*** erikmwil_ is now known as erikmwilson15:06
fungiyeah, confirmed last week that their local tests were with 8gb ram15:06
*** erikmwil- has joined #openstack-infra15:06
deepakcsclarkb, but doing some analsysi it seem before the oom, it was the cinder backup tests that were running, and those are the most IO intensive15:06
*** BharatK has joined #openstack-infra15:06
deepakcsclarkb, so disabling c-bak is what we are trying now15:06
gordcafazekas_: easy patch is revert15:06
*** prad has joined #openstack-infra15:06
*** patrickeast_ has joined #openstack-infra15:06
*** dimtruck is now known as zz_dimtruck15:06
gordcafazekas_: or we can wait 15minutes to try this other option which hopefully works?15:06
*** patrickeast has quit IRC15:07
*** patrickeast_ is now known as patrickeast15:07
*** sdake__ has quit IRC15:08
afazekas_gordc, The 'check-devstack-dsvm-updown' is not voting on ceilo15:08
gordcyes, i was just made aware.15:09
*** lascii is now known as alaski15:09
gordcafazekas_: reverting change as quickfix.15:11
*** prmtl has quit IRC15:12
clarkbdeepakcs and you are testibg with 64bit centos7 install? seems like something that OOMs this hard should be reproduceable15:12
clarkbbut maybe its related to the number of concurebt tests and test order15:13
deepakcsclarkb, yes, centos7 and i am surprised too as to why it doesn't reocreate15:13
*** dangers_away is now known as dangers15:13
deepakcsclarkb, i tried to use the same pretty_tox cmdlien that the failing logs have (which are set to concurrency=4)15:13
deepakcsclarkb, but yes, i had just 1 cpu, but looks like the infra VM has 8, which i realised just today, not sure if that adds to the mix ?15:13
*** nelsnelson has joined #openstack-infra15:13
mriedemjogo: sdague: mtreinish: clarkb: do you know if we have /proc/cpuinfo dumped anywhere in the logs for the host on a test run? looking at jogo's live migration multi-host change
deepakcsclarkb, fungi any other ideas to re-create locally are welcome :)15:14
clarkbmriedem setup host log file15:14
openstackgerritLucas Alvares Gomes proposed openstack-infra/project-config: Ironic: Bump memory size for non-concurrency job
clarkbdeepakcs yes I think the 8 cores is factoring in, it determines the number of service daemons to start among other things15:15
deepakcsclarkb, q on test order, i see the PYTHONHASHSEED is dumped in the logs, how do i set the same seed for my local run ?15:15
clarkbdeepakcs just set that env var to that value15:16
*** sigmavirus24_awa is now known as sigmavirus2415:16
deepakcsclarkb, ok, can try with 4 max :( my laptop has 4 phy cpus only. I am working to see if i can get a server class machine for this15:16
fungi(prior to calling tox, so potentially in the tox.ini)15:16
fungier, prior to calling python i mean15:17
deepakcsclarkb, ah ok, then i should just say  shouldn't append bash otherwise the env var won't be carried foarward, rigfht ?15:17
clarkbPYTHONHASHSEED=12345 pretty_tox.sh15:18
deepakcsclarkb, ah ok, that was simple :)15:18
gilliardmriedem: so is this a question of the actual CPU flags on the machines that run the tests?15:18
afazekas_sdague: is it planned to add check-devstack-dsvm-updown job to all project ?15:19
deepakcsclarkb, i will try with 4 cpus and hasseed tomorrow, but i hope (99% sure) with the c-bak disabled, we should not hit oom15:19
mriedemgilliard: i think so15:19
sdagueafazekas_: no15:19
*** imcsk8 has quit IRC15:19
*** imcsk8 has joined #openstack-infra15:19
deepakcsclarkb, fungi there isn't /proc/cpuinfo in the logs, so had to determine total cpus from syslog indirectly, any reason cpuinfo is excluded ?15:20
*** otter768 has joined #openstack-infra15:20
openstackgerritMerged openstack-infra/project-config: Change os-ansible-deployment checks to use scripts
fungideepakcs: no reason i'm aware of. seems like that would be a reasonable improvement to the openstack-infra/devstack-gate repo15:20
clarkbit is logged...15:21
deepakcsfungi, ok, i will work on patch :)15:21
clarkbthe problem is that jobs dies so hard we dont get the logs15:21
clarkbsee mriedems question above15:21
deepakcsclarkb, in devstack setup host log file , u mean ?15:22
openstackgerritMerged openstack-infra/shade: Don't compare images when image is None
*** claudiub has joined #openstack-infra15:22
anteayafungi: when do you begin holidaying?15:22
mriedemi don't suppose /usr/share/libvirt/cpu_map.xml is logged anywhere15:23
fungideepakcs: i'll include the cpuinfo in this log tarball once the manual run completes15:23
sdagueafazekas: I consider projects knowing whether or not their services start to be kind of table stakes, and a responsibility for that project. We can try to put a little more sanity checking inside devstack normal runs, but it wasn't on my short list15:23
anteayafungi: was thinking you would be in the warm today15:23
deepakcsfungi, thanks and gluster logs too (/var/log/glusterfs).15:23
clarkbdeepakcs yes in the setup host log file15:23
deepakcsfungi, if this doesn't resolve oom, i would request providing ssh access to the env15:24
fungianteaya: i'm heading up to norfolk tomorrow afternoon and working from car/hotel, then flying out wednesday morning and probably working from airline wireless until early afternoon est15:24
deepakcsclarkb, ok, but logging it separately as cpuinfo.txt would be more evident and useful i feel.15:24
anteayaworking travel and warm15:24
*** otter768 has quit IRC15:24
deepakcsfungi, how long will u be gone ? have nice holidays :)15:25
fungideepakcs: ~10 days15:25
AJaegerfungi, 10 days? We'll miss you! Have a great vacation!15:26
fungithanks AJaeger15:27
*** sdake_ has joined #openstack-infra15:28
*** ssam2 has quit IRC15:32
*** jlvillal|afk is now known as jlvillal15:32
openstackgerritMonty Taylor proposed openstack-infra/shade: Don't try to add an IP if there is one
openstackgerritMonty Taylor proposed openstack-infra/shade: Return extra information for debugging on failures
*** ssam2 has joined #openstack-infra15:33
*** achanda has joined #openstack-infra15:34
*** zigo_ has joined #openstack-infra15:36
*** mestery has quit IRC15:36
*** dkranz has quit IRC15:36
*** alexpilotti has joined #openstack-infra15:37
*** zz_dimtruck is now known as dimtruck15:38
*** radez_g0n3 is now known as radez15:38
*** yfried_ has quit IRC15:39
*** mrmartin has quit IRC15:39
*** dkranz has joined #openstack-infra15:40
*** asselin_ has joined #openstack-infra15:40
*** mattfarina has joined #openstack-infra15:41
*** radez is now known as radez_g0n315:43
*** achanda has quit IRC15:44
*** sergsh has quit IRC15:45
*** jawed has joined #openstack-infra15:45
dimsfungi: safe travels!15:48
fungithanks dims15:48
*** ChuckC has joined #openstack-infra15:49
*** mrmartin has joined #openstack-infra15:52
*** HeOS has quit IRC15:52
*** ChuckC_ has quit IRC15:52
*** MarkAtwood has joined #openstack-infra15:55
openstackgerritMonty Taylor proposed openstack-infra/shade: Make image processing work for v2
*** xyang1 has joined #openstack-infra15:57
krotscheckStoryboard Meeting in #openstack-meeting-316:00
*** patrickeast has quit IRC16:00
*** tiswanso has quit IRC16:01
*** tiswanso_ has joined #openstack-infra16:02
*** dhritishikhar_ has quit IRC16:02
*** dkranz has quit IRC16:03
*** amotoki has joined #openstack-infra16:05
*** carl_baldwin has joined #openstack-infra16:06
*** tmcpeak has quit IRC16:06
anteayaasselin: let me know how you make out with the mirror config issue16:07
*** prmtl has joined #openstack-infra16:07
*** dkranz has joined #openstack-infra16:08
*** che-arne has joined #openstack-infra16:09
*** mpaolino has joined #openstack-infra16:11
*** yfried_ has joined #openstack-infra16:12
*** Kiall has quit IRC16:13
asselinanteaya, sure16:13
*** prmtl has quit IRC16:14
*** harlowja_at_home has joined #openstack-infra16:15
*** jcoufal has quit IRC16:15
*** tmcpeak has joined #openstack-infra16:16
*** erikmwilson has quit IRC16:16
*** erikmwil- is now known as erikmwilson16:16
*** belmoreira has quit IRC16:17
*** tmcpeak has quit IRC16:19
*** teran has quit IRC16:19
*** dannywilson has joined #openstack-infra16:20
jogomriedem: one of the issues I think we are seeing with which cpu to pick is xen vs kvn16:20
jogosince rax has xen (not libvirt xen at that)16:20
*** devlaps has joined #openstack-infra16:21
*** mtanino has joined #openstack-infra16:21
*** dhritishikhar_ has joined #openstack-infra16:21
mriedemjogo: can we determine that at runtime in d-g when it's setting up?16:23
mtreinishjogo: can't we just restrict the multinode slaves to a single provider?16:23
mtreinishclarkb: ^^^ ?16:23
*** dannywil_ has joined #openstack-infra16:23
*** dannywilson has quit IRC16:24
*** xgerman has joined #openstack-infra16:24
jogomriedem: so a few  things: I tried using pentium16:24
jogoand that worked, but some tempest tests failed and thought that may be related16:25
*** andreykurilin_ has joined #openstack-infra16:25
jogomriedem: and I think sorting this out at runtime makes the most sense16:25
jogomriedem: on a side note, aiopcpu is working with tempest full, just without live migration. So that isn't too bad16:26
*** spzala has joined #openstack-infra16:28
jeblairmtreinish: no, we can never restrict slaves to a single provider.  the whole system is fault-tolerant across both clouds16:29
*** nelsnelson has quit IRC16:30
jeblairmtreinish: we had a couple of days last week where we could not get any nodes from hpcloud; did you notice? :)16:30
mtreinishjeblair: no, I realize that. I just meant having both slaves for a multinode job allocated from the same place16:30
jeblairmtreinish: oh, they already are16:30
jeblairmtreinish: same AZ even16:30
*** mwagner_ has joined #openstack-infra16:31
*** Kiall has joined #openstack-infra16:31
jogomtreinish: but single providers have multiple CPU types16:31
mtreinishjogo: yeah, I remember the problem now16:32
mtreinishfor some reason I thought it was across some region boundry that was the issue16:32
mtreinishchalk it up to being before noon16:32
*** e0ne has quit IRC16:32
*** mestery has joined #openstack-infra16:33
*** e0ne has joined #openstack-infra16:33
*** amitgandhinz has joined #openstack-infra16:33
jogomtreinish: hehe16:33
*** dkranz has quit IRC16:36
clarkbjogo is it possible to set explicit cpu flags rather than model?16:38
clarkbif so we can just union the flags of all nodes at runtime and use that16:38
clarkbif nova cant do that today it may make a good L feature16:39
*** deepakcs has quit IRC16:39
*** ddieterly has quit IRC16:39
openstackgerritDarragh Bailey proposed openstack-infra/jenkins-job-builder: Add tests for yaml anchor behaviour
jogoclarkb: great question, no idea16:40
openstackgerritDarragh Bailey proposed openstack-infra/jenkins-job-builder: support aliases defined in parent inside included yaml
jogobut I don't think nova supports that, not sure about libvirt16:41
clarkbpretty sure libvirt and qemu/kvm do16:41
*** yjiang5 has joined #openstack-infra16:42
clarkbmodel is a shortcut for more verbosely setting flGs16:42
*** tonytan4ever has joined #openstack-infra16:42
*** devlaps has quit IRC16:42
clarkbproblem here is the models we get and the flags exposed arent quite sane because we are nested16:42
clarkbeg core 2 never had sse4_2 but its in our core 2 processor flags16:43
sdagueclarkb: yeh, so this is a really extreme edge case16:43
sdagueso it's not something that was considered16:44
sdaguealso, what happens when a new n-cpu comes online with different flags?16:44
clarkbnested virt isnt an edge case imo16:44
sdagueno, not nested virst16:44
clarkbsdague that is a problem regardless16:44
clarkbit wont be for us if we union flags though16:44
sdagueexpecting live migration to work across a heterogenous hardware set16:44
clarkbif you ask me live migration shouldnt be a cloud feature at all :) but we support it16:45
sdaguewell, we support it as an admin command16:45
clarkband heterogenous hardware isnt really a problem if you can describe a subset to use16:46
jeblair(if you ask me, i'm really happy our production servers can get live-migrated when the cloud needs to perform maintenance)16:46
clarkbproblem is we may not be able to because nova only accepts cpu models16:47
*** yfried_ has quit IRC16:47
*** nelsnelson has joined #openstack-infra16:47
*** devlaps has joined #openstack-infra16:48
clarkbjeblair ya but that will go away if we config drive16:48
*** mpavone has quit IRC16:48
clarkbconfig drive breaks live migration16:48
*** amotoki has quit IRC16:48
SpamapSI think live migration is necessary in small-medium scale clouds16:48
*** tmcpeak has joined #openstack-infra16:49
SpamapSIf you have only, say, 36 compute hosts supporting a few thousand VMs, you can't afford to leave some running outdated software forever.16:49
*** harlowja_at_home has quit IRC16:49
clarkbI wonder if nova does support this and that is how hpcloud adds the sse4_2 flag16:50
clarkbjogo ^16:50
SpamapSIf you have 3600 compute hosts, then you just stop sending vms to the old ones till they drain a bit, then send the unlucky few running pets there that they'll be rebooted in a few weeks.16:50
jeblairSpamapS: i (personally) think it's more about use-case/audience.  a cloud that is providing "virtual servers" would do well to have live-migration, otherwise they compete poorly with "real servers".  but one providing "compute resources" perhaps can afford to ignore it16:50
SpamapSs/there/there notices/16:50
clarkbmordred this may be a reason to nova agent after all :/16:51
*** deepakcs has joined #openstack-infra16:51
*** dkranz has joined #openstack-infra16:51
SpamapSjeblair: well said. I think we agree, because I think a cloud running virtual servers is bound to be small - medium. :)16:51
jeblairheh :)16:52
SpamapSjeblair: on another note, in case you missed it, presentty is in Debian NEW16:52
jeblairSpamapS: neato! thanks!16:53
*** hdd has joined #openstack-infra16:53
jeblairi should do more work on it :)16:53
*** amitgandhinz has quit IRC16:53
SpamapSjeblair: also, I took a stab at making gear py3 compatible.16:54
SpamapSOh wait, I stand corrected!16:55
*** ildikov has quit IRC16:55
*** gyee has joined #openstack-infra16:55
SpamapSjeblair: presentty _cleared_ NEW this morning.16:56
mordredclarkb: what?16:56
jeblairSpamapS: what does that mean?16:56
* mordred reading scrollback16:56
fungijeblair: means its pending landing in sid now16:57
*** reed has joined #openstack-infra16:57
fungiprobably show up in sid/unstable sometime today depending on scheduling16:57
*** ChuckC_ has joined #openstack-infra16:57
mordredSpamapS: rax provides "virtual servesr" - it's why they're useful16:57
SpamapSdefinitely within 24 hours16:57
mordredclarkb: when you say "config drive breaks live migration" - do you have more info so I can understand the issue?16:58
SpamapSmordred: I don't mean to deride virtual server clouds. I mean to say why I think people get confused and say live migration shouldn't exist in a cloud.16:58
fungithough britney won't consider it for testing until after jessie releases16:58
openstackgerritBoris Pavlovic proposed openstack/requirements: Add flake8-double-quotas to requirements
mordredSpamapS: gotcha. I'm just happy that rax disagrees with those people16:58
*** ChuckC_ has quit IRC16:58
clarkbmordred nova cannot live migrate with config drives attached16:58
clarkbthere is a long standing open bug about it. something about how your device goes away iirc16:59
SpamapSdoesn't rax do config drives?16:59
mordredclarkb: so, if I "nova boot --config-drive=1" that vm can never be live-migrated?16:59
mordredclarkb: or only if the config-drive is mounted in some manner?16:59
clarkbmordred aiui correct. this is why our multinode test envs use metadata server and not config drive16:59
jeblairclarkb: can it live-migrate with cinder volumes attached?16:59
jogoclarkb: I think that is considered a bug, and not a feature at least17:00
mordredwell, there are two issues here:17:00
mordreda) what we do for nodepool nodes17:00
clarkbjeblair I do not know about cinder specifically but I think iscsi may make that work better17:00
mordredb) what we do for our servers17:00
*** ChuckC has quit IRC17:00
mordreda) we do not care about live migration17:00
mordredb) we do17:00
*** ChuckC_ has joined #openstack-infra17:00
jeblairmordred: agreed17:00
clarkbmordred correct17:00
mordredI'm more opposed to having nova-agent running in nodepool nodes, so I think that's easy17:00
mordredI _do- have a nova-agent element for dib17:00
mordredso if we need to do that route for long-lived servers but still want to make our own base images, we're set17:01
* jogo notes that nova-agent isn't a nova thing17:01
mordredhowever, I'd suggest that we poke jogo in the face with a stick until he fixes the config-drive live migration :)17:01
*** vigneshvar has joined #openstack-infra17:01
clarkbalso mikal knows all about this bug so we should just ask him in a few hours17:01
mordredjogo: yah. it's a rax thing17:01
mordredjogo: that talks to xenstore17:01
mordredit's their way of doing dhcp without running dhcp17:02
*** viktors is now known as viktors|afk17:02
mordredof course, if rax would just dhcp, none of this would really be necessary17:02
jeblairmordred, clarkb: i'd actually like to not run "nova-rootkit" on our servers if we can avoid it :)17:03
jogoclarkb: yeah so next step is to follow the instructions danpb has and do automatic cpu type selection in d-g17:03
mordredjeblair: ++17:03
*** scheuran has quit IRC17:03
clarkbjogo we should also look into directly setting flags17:04
clarkbjogo I think that may be easier long run17:04
jogoclarkb: hmm yeah, that is worth looking into17:04
*** rwsu has joined #openstack-infra17:05
mordredjeblair: totally agree. it's also possible that something at rax is different and live migration would work there - so it might be worth poking mikal to see if he knows any more informatino17:06
*** tiswanso_ has quit IRC17:06
mordredI would honestly be very surprised if rax let something like that out into the wild, since it would drastically confuse the customers who are used to running multi-year servers there17:07
mordredlike, it may be one of those things that's easy to fix for the rax xen case but harder to fix broadly or something17:07
*** mrmartin has quit IRC17:07
jeblairyeah, worth digging into17:08
*** HeOS has joined #openstack-infra17:08
*** yfried_ has joined #openstack-infra17:09
fungiSpamapS: rax does support config drive, but i'm not sure whether they do live migration (they stopped supporting online resize a while back at least)17:09
clarkbmordred yup we should definitel have mikal clue bat me/us17:10
*** mpaolino has quit IRC17:11
*** tiswanso_ has joined #openstack-infra17:11
bhunter71I'm drawing a blank on something-  when nodepool adds nodes to my jenkins01, they're currently getting through the whole gearman thing, and to the final jenkins-slave-song-and-dance.  I'm getting "hudson.AbortException: Cannot find SSH User credentials with id: null".   I wanted global credentials 'jenkins' here.   what did I do wrong?17:11
fungior have him delegate cluebatting to jhesketh/mattoliverau17:11
*** xgerman has quit IRC17:11
*** dkranz has quit IRC17:11
*** jawed has quit IRC17:12
pleia2good morning :)17:12
pleia2clarkb: yay17:12
*** BobBall is now known as BobBall_AWOL17:12
fungibhunter71: the credentials-id is the uuid of the jenkins credentials object. i usually obtain it from config.xml17:13
bhunter71ah, thanks fungi.   (necessitated by a jenkins master getting corrupted/rebuilt)17:14
fungibhunter71: essentially, configure a ssh credentials set in jenkins for nodepool's use, then go fishing in the jenkins config on disk to discover the corresponding uuid for it17:14
clarkbits one of my big gripes with jenkins move to the credentials manager17:15
clarkbusing the key in $HOME/.ssh worked fine before :/17:15
*** carl_baldwin has quit IRC17:16
fungiyeah, the admin experience there is sort of lacking17:16
*** MarkAtwood_ has joined #openstack-infra17:16
*** JayJ has joined #openstack-infra17:17
*** ihrachyshka has quit IRC17:19
*** teran has joined #openstack-infra17:20
*** tkelsey has joined #openstack-infra17:20
*** unicell has quit IRC17:20
*** otter768 has joined #openstack-infra17:21
*** thedodd has joined #openstack-infra17:21
*** baoli has quit IRC17:22
*** baoli has joined #openstack-infra17:22
*** marun has joined #openstack-infra17:23
*** mpaolino has joined #openstack-infra17:23
*** baoli has quit IRC17:23
*** sdake_ has quit IRC17:23
*** baoli has joined #openstack-infra17:24
*** teran has quit IRC17:25
*** MaxV has quit IRC17:25
*** jlk has quit IRC17:25
*** jlk has joined #openstack-infra17:25
*** dkranz has joined #openstack-infra17:25
*** otter768 has quit IRC17:25
*** johnthetubaguy has joined #openstack-infra17:25
*** baoli has quit IRC17:26
*** baoli has joined #openstack-infra17:26
openstackgerrityolanda.robla proposed openstack-infra/shade: Add a method to create image snapshots from nova.
afazekas_ianw, ping17:27
jogojohnthetubaguy: so your a rax person17:27
johnthetubaguyjogo: hey17:27
johnthetubaguywhats up?17:27
jogowe are having issues with the CPU types for live migration17:27
johnthetubaguyhmm, interesting17:28
jogojohnthetubaguy: see
jogomriedem: ^17:28
*** ddieterly has joined #openstack-infra17:28
jogoclarkb mriedem: btw got some feedback from danpb
*** jistr has quit IRC17:29
mriedemjogo: shit creek17:29
*** tiswanso_ has quit IRC17:29
johnthetubaguyOK, so you want to get a list of minimum CPU requirements in rax, or am I missing something?17:30
mriedemis this why we generally assume live migration doesn't work?17:30
*** radez_g0n3 is now known as radez17:30
openstackgerritMerged openstack-infra/puppet-apparmor: Update metadata.json
*** tiswanso_ has joined #openstack-infra17:30
rcarrillocruzhey folks17:31
jogojohnthetubaguy: just sanity checking if what we are seeing is right17:31
jogojohnthetubaguy: we can't even use 'Conroe'17:31
jogowhich is very old at this point17:32
rcarrillocruzi've seen in our internal infra that some Gerrit projects have a 'config' branch with metadata from Gerrit ( a groups file and a project.config, that contain acls about that Gerrit project)17:32
rcarrillocruzanyone seen this?17:32
clarkbjogo right I think you need newer to match up with the flags that are set17:32
jeblairrcarrillocruz: i think every project should have it17:33
clarkbjogo OR set all capabilities directly17:33
jeblairrcarrillocruz: "refs/meta/config"17:33
rcarrillocruzreally? i've noticed that in some17:33
rcarrillocruznot in all projects17:33
*** sabeen3 has joined #openstack-infra17:33
*** sabeen has quit IRC17:34
zarorcarrillocruz: i think some may be empty because it inherits from All-Projects17:34
rcarrillocruzthx zaro, that may be the reason17:34
rcarrillocruzthx jeblair17:34
*** mestery has quit IRC17:34
rcarrillocruzi was worried we had cruft on our Gerrit instance17:34
jeblairrcarrillocruz: i think jeepyb should put something there in all cases17:35
*** mestery has joined #openstack-infra17:35
*** SumitNaiksatam has quit IRC17:35
jeblairzaro: ^17:35
*** devvesa has quit IRC17:35
jogoclarkb: I think setting capabilities directly would be good17:35
fungijeblair: governance repo change idea
* anteaya goes for a walk, in the lovely sunny day17:35
rcarrillocruzjeblair: so your take is that regardless of inheritance, jeepyb should put a config branch for all repos it creates?17:36
deepakcsfungi, ping, any update on the manual run ?17:36
*** SumitNaiksatam has joined #openstack-infra17:36
*** derekh has quit IRC17:36
*** ChuckC_ has quit IRC17:36
jeblairfungi: seems reasonable, at least until we finish up docs.o.o/infra17:37
zarorcarrillocruz: i believe gerrit provides a refs/meta/config for every project, just that some may not contain anything because of inheritance17:37
fungideepakcs: yes, it completed and i'm writing a follow-up e-mail now, logs are at
zarorcarrillocruz: but i think wht jeblair is saying is that jeepyb will always add something to project.config for every subproject?17:38
*** e0ne has quit IRC17:38
SpamapSoh my17:38
clarkbthe way we use it it will because we always have acls17:38
SpamapS  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND17:38
jeblairzaro: i think so, yeah.17:38
SpamapS 1006 clint     20   0 14.204g 0.010t   2736 S 202.6 68.7   1578:51 python17:39
SpamapS 1006 pts/0    Sl   1579:07 python -m discover -t ./ gear/tests/ --load-list /tmp/tmpFdXUhm17:39
SpamapSrogue unit test17:39
rcarrillocruzhmm ok, let me check inheritance on projects showing that branch and projects without it...17:39
jeblairSpamapS: neat!  oh, there are two outstanding gear patches, i don't know if they would have an effect.17:39
SpamapSjeblair: it's spinning on this:17:40
SpamapSpid  1030] sendto(15, "\0REQ\0\0\0\36\0\0\0\0", 12, 0, NULL, 0 <unfinished ...>17:40
SpamapS[pid  1025] epoll_ctl(5, EPOLL_CTL_DEL, 16, {EPOLLRDBAND|EPOLLERR|EPOLLHUP|EPOLLRDHUP|EPOLLET|0x39ca9000, {u32=32730, u64=24297708030296026}}) = 017:40
zaropleia2: threw a logo at you
SpamapSI only noticed because my box slowed to a crawl :-P17:41
pleia2zaro: good idea, thanks17:41
zaropleia2: can you help create a publication branch for me?17:41
pleia2zaro: I can! let me know what you need and I'll see about doing it17:43
*** sdake_ has joined #openstack-infra17:43
johnthetubaguyjogo: mriedem: I think I am missing context, this is getting KVM live working when running on rax?17:44
*** ChuckC has joined #openstack-infra17:45
*** Ala has quit IRC17:45
zaropleia2: would like a "jenkins-job-builder" branch.17:45
*** Sukhdev has joined #openstack-infra17:46
clarkbjohnthetubaguy yes we are trying to test it in our multi node test envs17:46
pleia2zaro: someone can correct me if I'm wrong, but I *think* you create the branch yourself and then when you need a version of it - like you gave the talk at a conference, is when we need to do tagging17:46
*** salv-orlando has joined #openstack-infra17:46
johnthetubaguyclarkb: so they are quite new xeon chips in performance, so I guess xen might be hiding some CPU flags17:47
zaropleia2: instructions in README says you need to contact infra to create a branch.17:47
zaropleia2: i mean README from template branch17:48
*** dizquierdo has quit IRC17:48
marcusvrnhi guys! is there any way to change the ci account name? for example: hds-hnas-ci to hds-ci?17:48
pleia2zaro: ah, branch name in gerrit17:48
johnthetubaguyclarkb: jogo: I don't think we tell xen to hide anything at this point, but I should check really17:48
zaropleia2: yep, that's what i need17:48
pleia2zaro: got it, looking into that now17:48
*** AJaeger_ has joined #openstack-infra17:50
*** zul has quit IRC17:51
pleia2clarkb: re: publication branches in gerrit, what do I use for "Initial Revision" ?17:51
*** coolsvap is now known as coolsvap_17:52
jeblairpleia2: i'd use 'template'17:52
pleia2jeblair: ok, thanks17:52
clarkbpleia2 ya, since thats what the changes should be based on17:53
*** zul has joined #openstack-infra17:53
pleia2zaro: done!17:53
jeblairnibalizer: 151098 safe to aprv?17:53
clarkbavoids rebasing to push for review17:53
pleia2hm, when I added it through the gerrit UI it added a "master" subsection that the others don't have17:54
pleia2can someone else take a look?17:54
nibalizerjeblair: yea doesn't affect anything yet17:54
nibalizerexcept of course, my heard *thumpthump*17:54
jeblairpleia2: not sure what you are referring to17:55
openstackgerritMerged openstack-infra/puppet-gerrit: Make sure Gerrit sites are using the proper libraries.
openstackgerritMerged openstack-infra/puppet-gerrit: Add the Gerrit replication plugin paramater 'defaultForceUpdate'
pleia2see "jenkins-job-builder" that I just added for zaro's new slidedeck17:55
openstackgerritAndreas Jaeger proposed openstack/requirements: Update openstackdocstheme
*** marun has quit IRC17:55
jeblairpleia2: where's the master subsection you're talking about?17:56
*** yamahata has joined #openstack-infra17:56
*** unicell has joined #openstack-infra17:56
*** bhuvan has joined #openstack-infra17:56
pleia2jeblair: oh, nevermind, it's just below jjb, but it's the actual master being alphabetical17:56
*** sdake__ has joined #openstack-infra17:56
pleia2more coffee \o/17:56
jeblairah :)17:56
*** unicell has quit IRC17:57
pleia2I'll get used to this UI soon :)17:57
*** unicell has joined #openstack-infra17:57
*** dimtruck is now known as zz_dimtruck17:57
*** ghostpl_ has quit IRC17:59
*** unicell has quit IRC17:59
*** zz_dimtruck is now known as dimtruck18:00
*** unicell has joined #openstack-infra18:00
*** sdake_ has quit IRC18:00
*** nelsnelson has quit IRC18:00
jeblairnibalizer, clarkb: is there a reason we're calling our modules "openstackci-foo" instead of "openstackinfra-foo"?18:00
jeblairin general, we're trying to s/ci/infra/18:00
*** nelsnelson has joined #openstack-infra18:01
*** Ryan_Lane has joined #openstack-infra18:01
nibalizerwe already had the openstackci name and user on puppetforge, but I don't care18:01
openstackgerritMerged stackforge/python-jenkins: Add get_job_info_regex method to get info about a set of jobs
openstackgerritMerged openstack-infra/puppet-httpd: Rename apache2 to httpd globally
jeblairnibalizer: yeah, i think a legacy username is fine, but if it affects how the modules are named / namespaced, it's probably worth another look18:02
nibalizerits ~60 one line commits to change it18:03
nibalizerand much easier to change it now than after we start forge publishing18:03
jeblairnibalizer: is module name completely independent from the user who uploads?18:05
jeblairnibalizer: added it to tomorrow's agenda, just to make sure we're all in agreement before we patch bomb18:06
*** dhritishikhar_ has quit IRC18:06
*** ildikov has joined #openstack-infra18:07
*** BadCub has quit IRC18:07
*** beecee has joined #openstack-infra18:07
*** afazekas_ has quit IRC18:07
*** dustins_ has joined #openstack-infra18:09
*** david-lyle_afk is now known as david-lyle18:09
*** unicell has quit IRC18:10
*** ChuckC has quit IRC18:11
*** dustins has quit IRC18:12
jeblairkrotscheck: the bottom of your puppet-storyboard stack (156755) has a commit message that says do not merge until an abandoned change (156764) merges18:12
*** jlibosva has quit IRC18:13
*** lucasagomes is now known as lucas-dinner18:14
*** tonytan4ever has quit IRC18:14
*** tonytan4ever has joined #openstack-infra18:15
SpamapSjeblair: Hey, so I've been poking at gear a bit, and I can't seem to find how to make a blocking client submitJob call that waits for the job response, or a non-blocking way to go ahead and wait for the response if I have nothing else to do.18:15
SpamapSjeblair: I looked and zuul just spins with time.sleep(1)18:15
*** lucas-dinner has left #openstack-infra18:15
*** dtantsur is now known as dtantsur|afk18:16
SpamapSNot sure the protocol has something that helps here, one might just need to go ahead and select/poll on the socket.18:16
*** deepakcs has quit IRC18:16
fungiclarkb: i have a feeling isn't working for devstack-centos7 images. free says ~24gb18:16
*** achanda has joined #openstack-infra18:16
*** mrmartin has joined #openstack-infra18:17
*** achanda has quit IRC18:18
krotscheckjeblair: I need to fix that.18:18
* krotscheck just got into the office18:18
*** achanda has joined #openstack-infra18:18
jeblairSpamapS: you can subclass Worker and implement "handleWorkComplete" (and friends)18:19
*** mpaolino has quit IRC18:19
*** tonytan4ever has quit IRC18:19
SpamapSjeblair: yes, however, I still have to spin on time.sleep() until they are called.18:20
SpamapSunless there's a thread I can join?18:20
*** ssam2 has quit IRC18:20
openstackgerritMerged openstack-infra/puppet-storyboard: Add metadata.json to puppet-storyboard
*** ivar-lazzaro has joined #openstack-infra18:20
openstackgerritPaul Michali proposed openstack-infra/project-config: Setup non-voting VPNaaS func test for StrongSwan
fungiclarkb: er, says ~30gb i mean18:20
fungiclarkb: on hpcloud obviously18:21
*** teran has joined #openstack-infra18:21
*** amrith has joined #openstack-infra18:21
jeblairSpamapS: yeah, there isn't an out-of-the box way to do that...18:21
jeblairSpamapS: we could probably put an event on the job and wait on that18:21
*** ivar-lazzaro has quit IRC18:22
amrithhi ... could I get someone on the infra team to help get merged please. It is holding up a number of changes in Trove.18:22
*** ivar-lazzaro has joined #openstack-infra18:22
AJaeger_amrith: +2ed.18:24
*** mestery has quit IRC18:24
amrithAndreas, thanks!18:24
SpamapSjeblair: might just need an equivalent of this
jeblairSpamapS: so we could have the default packet handler for those packets call job.wait_event.set().  then you could just do "job.wait()"18:24
*** ryanpetrello has quit IRC18:25
*** patrickeast has joined #openstack-infra18:25
arosenfungi:  have you seen this nodepool issue? I have the directory: /home/nodepool/nodepool_dib/devstack-trusty-dib-1424651675.d but not  /home/nodepool/nodepool_dib/devstack-trusty-dib-142465167518:25
*** ryanpetrello has joined #openstack-infra18:25
*** zul has quit IRC18:25
*** mestery has joined #openstack-infra18:25
openstackgerritBharat Kumar Kobagana proposed openstack-infra/project-config: Adding glusterfs job to glusterfs plugin's check pipeline
*** zul has joined #openstack-infra18:25
*** teran has quit IRC18:26
*** rkukura has joined #openstack-infra18:26
*** yfried_ has quit IRC18:26
*** vilobhmm has joined #openstack-infra18:27
fungiarosen: looks like maybe your image-create never finished so image-upload doesn't find it?18:27
*** jp_at_hp has quit IRC18:27
arosenfungi:  i was wondering if its actually missing the file extension ? and should be passing in : devstack-trusty-dib-1424651675.qcow218:28
arosenbecause I have that file as well.18:28
arosenfungi:  changing the code to this makes the image upload, i'll see if it works:
*** vilobhmm has quit IRC18:29
fungiarosen: oh, maybe... it does look like it wants that to be a file it can read from18:29
arosenyea i'm pretty sure it wants to be that disk file.18:29
nibalizerjeblair: no i believe module-author-name and logged in name must be identical18:30
fungiarosen: i think filename is supposed to include the appropriate suffix18:30
*** tiswanso_ has quit IRC18:30
jeblairnibalizer: do you know who set up the user?18:30
*** vilobhmm has joined #openstack-infra18:31
nibalizerjeblair: the credentials for that user are in hiera18:31
nibalizeri think mordred set it up18:31
fungiyeah, mordred added it to hiera according to the git log18:31
*** dustins_ has quit IRC18:31
fungii recall seeing it in there18:32
*** carl_baldwin has joined #openstack-infra18:32
*** dustins has joined #openstack-infra18:32
*** vilobhmm1 has joined #openstack-infra18:32
mordredwhat did I do?18:32
*** vilobhmm_ has joined #openstack-infra18:32
jeblairmordred: i was thinking that we probably want our modulenames to be "openstackinfra-foo"18:32
mordredI set it up AGES ago when things were still called openstackci18:32
mordredjeblair: ++18:33
jeblairmordred: so i reckon we need a new user, based on what nibalizer just said18:33
*** harlowja has joined #openstack-infra18:33
jeblairmordred: you think you could take care of that?18:34
*** tiswanso_ has joined #openstack-infra18:34
*** ashleighfarnham has joined #openstack-infra18:34
*** mestery has quit IRC18:34
mordredjeblair: sure18:35
*** dustins has quit IRC18:35
*** vilobhmm has quit IRC18:35
*** vilobhmm_ is now known as vilobhmm18:35
*** dustins has joined #openstack-infra18:36
*** mestery has joined #openstack-infra18:38
fungiclarkb: on further investigation, is only working on devstack-precise, not devstack-trusty nor devstack-centos718:38
*** ivar-laz_ has joined #openstack-infra18:38
clarkbdib related possibly?18:39
*** melwitt has joined #openstack-infra18:39
*** ivar-laz_ has quit IRC18:39
clarkbthough dib rebuilds grub rules18:39
fungiclarkb: maybe...18:39
SpamapSjeblair: I'll think about it some as I make this little custom client thing I've been poking at. Thanks for your thoughts.18:39
clarkbbut precise is grub1 the others are grub218:39
SpamapSmorganfainberg: Hi, so, broadcasting..18:39
fungiclarkb: this is based entirely on the output of free on representatives of those node types in hpcloud18:39
*** ivar-laz_ has joined #openstack-infra18:40
fungiclarkb: also i can't find where our cgroups enforcement is getting applied any longer18:40
*** ivar-laz_ has quit IRC18:40
SlickNikPlease can I get some eyeballs on when folks get a chance? Thanks!18:41
*** ivar-laz_ has joined #openstack-infra18:41
*** tkelsey has quit IRC18:41
*** koolhead17 has quit IRC18:41
*** ivar-lazzaro has quit IRC18:41
sdagueclarkb: config drive preventing live migration is a bug, and I think tonyb might be working on that one now18:42
*** ivar-laz_ has quit IRC18:42
*** zul has quit IRC18:42
*** marun has joined #openstack-infra18:42
clarkbno cgroups on single use nodes18:42
clarkbfungi ^18:42
mordredSlickNik, sdague, jogo: ^^ why would we exempt trove from the stable-client rules?18:42
mordredI just want to make sure I understand18:42
*** ivar-lazzaro has joined #openstack-infra18:42
fungiclarkb: we were apparently including jenkins::cgroups from openstack_project::slave but that doesn't get used by openstack_project::single_use_slave18:42
fungiclarkb: yeah18:42
sdaguethe clients are all capped on stable branches now18:42
sdaguemordred: ^^^18:42
sdagueso it doesn't do anything18:43
fungiclarkb: so, intentional?18:43
mordredsdague: ok. so this isn't special casing trove out of something, it's just reflecting the reality we want?18:43
*** ddieterly has quit IRC18:43
clarkbfungi  I think idea was we didnt need it for single use nodes18:43
clarkbfungi but the glusterfs job is evidence for otherwise18:43
sdagueit's reflecting that there was 0 direct testing of client compatibility, and capping on stable branches removed all our indirect testing18:43
*** ddieterly has joined #openstack-infra18:44
*** zul has joined #openstack-infra18:44
mordredsdague: thanks18:44
fungiclarkb: well, the worker i randomly selected to rerun it on today was in hpcloud, and apparently had 30gb ram available to it (which is what sent me down the path of figuring out why and jenkins::cgroups weren't coming into play and whether that was what caused the job to not oom)18:44
sdagueI think if you look at those juno / icehouse jobs you'll see the pip freeze shows it's using an old version even with the LIB_FROM_GIT18:44
clarkbfungi the memory restriction to 8gb is done by kernel boot params18:45
SlickNikmordred — no this isn't special casing for trove — what sdague said.18:45
clarkbfungi not cgroups18:45
clarkbcheck grub config18:45
fungiclarkb: right, so free should be reporting 8gb rather than 30gb there i think18:46
fungii'll poke around in them and see what they have configured18:46
sdagueclarkb: cpu_mode & cpu_model from nova might give you what you need here18:46
clarkbfungi yup18:46
*** asselin_ has quit IRC18:47
clarkbsdagye not quite the trouble is our rax/hp VMs are a wierd combo of flags that doesnt really overlap with libvirt presets18:47
sdagueclarkb: and there is no dumbest common denomentator?18:47
greghaynesSpamapS: Trying to find on storyboard that backlog we created, got any pointers?18:47
clarkbsdague not yet no18:48
*** dangers is now known as dangers_away18:48
clarkbeven pentium doesnt work18:48
SpamapSgreghaynes: it's in the list of stories ;)18:48
greghaynesya dont say?18:48
sdagueclarkb: so that's all just an xml file in libvirt18:49
sdaguemy guess is hpcloud just has a modified one of those18:49
fungiclarkb: looking on a devstack-trusty-hpcloud node i see that /etc/default/grub is missing mem=8G even though the sed pattern as written in does correctly add it. could this be getting reset somehow?18:49
sdagueand that if you *really* wanted to here, you could do the same thing there18:50
fungiclarkb: i'll go digging in nodepool image logs to see if it mentions getting run at all18:50
SpamapSgreghaynes: I can't seem to find it actually18:50
greghaynes\O/ its not total pebkac18:51
openstackgerritPaul Michali proposed openstack-infra/project-config: Setup non-voting VPNaaS func test for StrongSwan
SpamapSgreghaynes: Had to go to the openstack-infra project, then list stories.. last one listed. :-P18:51
greghaynesah, ty18:52
openstackgerritBharat Kumar Kobagana proposed openstack-infra/project-config: Adding glusterfs job to glusterfs plugin's check pipeline
clarkbsdague we may have to in order to test live migration18:52
*** tonytan4ever has joined #openstack-infra18:54
sdagueclarkb: does 486 work as a cpu model?18:54
clarkbsdague I dont think that has been tried18:54
sdague      <feature name='fpu'/>18:54
sdague      <feature name='vme'/>18:54
sdague      <feature name='pse'/>18:54
clarkbhost, pentium, and core2 dobt work18:54
sdaguewhere are the 2 cpu flag sets?18:54
openstackgerritMerged openstack-infra/project-config: Do not run stable-compat jobs against trunk python-troveclient
fungiclarkb: does nodepool dib not log to the regular image log?18:54
clarkbrax has heterogenous nodes, hpcloud is homogenous but different than the rax options18:55
*** yfried_ has joined #openstack-infra18:55
clarkbfungi it foes18:55
fungii feel like every time i pull a loose thread to look into a bug i unravel the entire sweater18:55
*** dhritishikhar has joined #openstack-infra18:55
*** SumitNaiksatam has quit IRC18:56
sdagueclarkb: well, actually, rax has heterogenous nodes, but has created a cpu model that works for all of them18:56
sdaguesorry, hp cloud18:56
*** marun has quit IRC18:57
*** tonytan4ever has quit IRC18:58
sdagueclarkb: pentium doesn't work>?18:59
sdagueare you sure?18:59
sdaguebecause all those flags look like they are in there18:59
clarkbsdague according to jogo it failed19:00
sdaguedo we have a test run to look at?19:00
morganfainbergSpamapS, hi so broadcasting19:00
clarkbjogo has bunches on his change let me find it19:00
morganfainbergSpamapS, and i've been looking at oslo.messaging, i think i have a line in on it now. but i need to stand up a strawman server/client to do this. shouldn't be too bad.19:01
morganfainbergSpamapS, but i started putting some basic structure together for this.19:01
SpamapSmorganfainberg: cool. I am poking at my idea for efficient streaming of broadcasts here:
* morganfainberg looks19:02
sdagueclarkb: thats cpu_model = Conroe19:02
SpamapSmorganfainberg: still just tiny bits of library and tests that kind of imagine how it will work19:02
morganfainbergSpamapS, oh cool19:02
clarkbsdague look at patchsets I think several were tried19:02
SpamapSmorganfainberg: but the basic idea is to buffer these things up in workers and then fling them out to subscribers somewhat efficiently via the coalesce functionality.19:03
morganfainbergSpamapS, makes sense to me.19:03
fungiclarkb: so, looking in the dib log, i think maybe 50-restrict-memory is happening too early. it doesn't find either /etc/default/grub or /boot/grub/grub.conf19:03
SpamapSI did get a little side tracked with python3.4 because I'm tired of always waving my hands that "we'll get to that later" ;)19:04
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Added a Vagrant module and manifest for easier development.
*** zul has quit IRC19:04
clarkbfungi that sounds plausible19:04
morganfainbergSpamapS, shouldn't be too hard to19:04
*** zul has joined #openstack-infra19:04
clarkbfungi is it postinstall.d?19:04
morganfainbergSpamapS, get all that in for something useful.  py3.4 yeah19:04
morganfainbergSpamapS, i tend to wave my hands on that too :P19:04
fungiclarkb: nope, install.d19:04
*** dustins has quit IRC19:04
fungiclarkb: should be post i guess?19:05
*** tonytan4ever has joined #openstack-infra19:05
sdagueclarkb: live migration tests passed on the pentium version19:05
SpamapSdamnit. Sometimes F7 runs flake8, and sometimes it doesn't.19:05
sdaguethe fails were network ones19:05
clarkbfungi that would be my guess, maybe spamaps or greghaynes can confirm19:05
sdagueon different tests19:05
*** bhunter71 has quit IRC19:06
clarkbsdague ok, I am just rrpeating what jogo has said, is it possible cirros wont boot on pentium?19:06
clarkbdoes that affect arch.19:06
SpamapSfungi: grub gets installed _really_ late.19:06
krotscheckopenstackci/httpd isn't a puppetforge thing, is it.19:06
clarkbSpamapS any reason for that?19:06
sdagueclarkb: maybe19:06
*** tonytan_brb has joined #openstack-infra19:06
fungiSpamapS: will sticking something in post-install.d be assured to happen after grub is installed?19:07
sdaguebut... live migration tests pass if the guest doesn't start?19:07
*** dustins has joined #openstack-infra19:07
SpamapSclarkb: I believe because we don't have our final block device until very late. Might be legacy.19:07
fungiSpamapS: or alternate suggestions for altering the grub configuration?19:07
clarkboh right the wholr cp everything issue :(19:07
SpamapSfungi: no, finalise.d19:07
greghaynesfungi: looks like it happens in finalise.d19:07
greghaynes51-finalise.d to be specific19:07
SpamapSit would make sense to install them earlier, and just try to prevent them from poking at the root dev on install19:08
SpamapSbut really, what you need is a .d you can drop things in19:08
fungiso if i do a finalise.d/90-whatever that ought to work?19:08
*** bhunter71 has joined #openstack-infra19:08
SpamapSfungi: 52? ;)19:09
greghaynesWell, is there any reason we have to install grub that late? if its just looking at configs that seems fine, we really just need to run the grub-install or w/e its called nowdays that late19:09
*** tonytan4ever has quit IRC19:09
*** ddieterly has quit IRC19:10
clarkbgreghaynes thats my question19:10
greghaynesI think we used to do that, actually19:10
openstackgerritAndreas Jaeger proposed openstack-infra/project-config: Limit manual jobs
fungiSpamapS: just in case the dib maintainers decide to bump grub to something higher than 52 ;)19:10
*** ddieterly has joined #openstack-infra19:10
SpamapSIt's simpler if we do it late when everything is already on the final block device.19:10
*** mrunge has quit IRC19:10
SpamapSfungi: aaaaand we're back in 8th grade with GW-BASIC ;)19:10
greghaynesyes, IIRC we used to have to install, then force an uninstall/reinstall when we tried it19:11
fungiSpamapS: always increment line numbers by 10!19:11
greghaynesbecause of some handwavy reason I cant remember19:11
*** zul has quit IRC19:11
*** zul has joined #openstack-infra19:11
*** tmcpeak has quit IRC19:12
*** cdent has quit IRC19:14
*** tmcpeak has joined #openstack-infra19:15
*** dhritishikhar has quit IRC19:15
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Corrected metadata.json
clarkbSpamapS not sure its simpler, the only diffedence is one command right?19:17
*** Rockyg has joined #openstack-infra19:17
sdagueclarkb / fungi - can you guys look at this - so we can change a devstack default and not impact gate log collection19:17
SpamapSclarkb: yes, and I'm not sure either.19:18
*** SumitNaiksatam has joined #openstack-infra19:19
fungisdague: on 157057 i guess we don't see any benefit to making that configurable at the job level right now?19:19
*** peterstac has joined #openstack-infra19:19
*** sputnik13 has joined #openstack-infra19:19
openstackgerritBharat Kumar Kobagana proposed openstack-infra/project-config: Adding glusterfs job to glusterfs plugin's check pipeline
sdaguefungi: right now, I don't think so19:20
sdaguewe're taking the default atm19:20
peterstacHi, would anyone here be able to look at ?19:20
peterstacpython-troveclient is dead in the water until we get that merged ...19:21
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Added new OAuth Parameters
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Updated storyboard configuration for pymysql.
peterstacoops, nevermind - just checked and it's merged already, thx!19:21
*** otter768 has joined #openstack-infra19:21
*** teran has joined #openstack-infra19:22
greghaynes is what im thinking with how we used to do that19:22
greghaynesalthough that was redhat specific19:22
greghaynesclarkb: oh, looks like we do explicitly remove grub2 in pre-install.d in the ubuntu element. Youall dont use that though right?19:23
*** pelix1 has quit IRC19:23
clarkbwe do19:24
*** alexpilotti has quit IRC19:24
greghayneswhich is interesting beacause I believe the redhat-common element installs it in that phase19:24
mordredpeterstac: I like to service requests before they are made19:24
greghaynesso at a minimum I think we should be consistent between those two19:25
mordredgreghaynes, clarkb: but I want us to move to the ubuntu-minimal element once it's all sorted19:25
mordredjust fyi19:25
greghaynesmordred: ah, so ATM youre still not using ubunut-minimal?19:25
krotscheckmordred: What was that channel for the ansible things again?19:25
mordredgreghaynes: not yet19:25
peterstacmordred: Great philosophy, I like it! ;)19:25
AJaeger_We haven't used ERROR_ON_CLONE yet - introduces it. Anybody here that can confirm that this is doing the right thing?19:26
*** otter768 has quit IRC19:26
mordredkrotscheck: ##ansible-shade19:26
*** tmcpeak has quit IRC19:26
*** teran has quit IRC19:27
*** hdd has quit IRC19:28
openstackgerritJeremy Stanley proposed openstack-infra/project-config: Run restrict-memory element later in image build
fungiclarkb: SpamapS: greghaynes: ^19:28
anteayaAJaeger_: I see ERROR_ON_CLONE in
*** unicell has joined #openstack-infra19:28
AJaeger_anteaya: so do I - since 2012. But we're not using it in project-config at all, so wanted some confirmation about setting it.19:29
anteayaI don't see it used anywhere else, so I don't know what happens if the variable it set to FALSE19:30
asselinHi, we're seeing pypi errors recently causing our 3rd party ci builds to fail. Is anyone aware of any releases of pip etc that could affect it?19:30
*** unicell has quit IRC19:30
*** unicell has joined #openstack-infra19:30
asselinreleases of pip that e.g. would require updates to our .pip/pip.conf and/or .pydistutils.cfg19:31
anteayaAJaeger_: I see more in devstack/19:31
fungiasselin: what errors are you seeing?19:31
AJaeger_anteaya: looking at devstack, it should be fine to set it....19:31
*** amitgandhinz has joined #openstack-infra19:31
*** amitgandhinz has quit IRC19:31
asselinfungi, connection timeouts19:31
fungiasselin: intermittent?19:31
openstackgerritMonty Taylor proposed openstack-infra/shade: Add support for volume attach/detach
*** amitgandhinz has joined #openstack-infra19:32
*** tmcpeak has joined #openstack-infra19:32
sdagueAJaeger: that is definitively the wrong thing to do19:32
fungiasselin: to or some other host?19:32
asselinfungi, I have it setup to use our mirror and fallback to pypi19:32
asselinfungi, intermiitent, but enough to cause everything to fail:
*** marun has joined #openstack-infra19:33
openstackgerritSwaminathan Vasudevan proposed openstack-infra/project-config: Create Neutron DVR experimental job for aiopcpu
AJaeger_sdague: could you comment on the change and tell vponomaryov how to do it properly, please?19:33
openstackgerritMonty Taylor proposed openstack-infra/shade: Add a method to create image snapshots from nova.
asselinfungi, I noticed lots of warnings, deprecations, etc recently. I'm wondring if we're actually hitting issues best guess19:34
AJaeger_sdague: ah, you commented - thanks19:34
anteayaAJaeger_: thanks for asking19:34
anteayasdague: for my education why is this the wrong approach?19:34
*** johnthetubaguy is now known as zz_johnthetubagu19:35
sdagueERROR_ON_CLONE is a safety measure in devstack that was created for the gate19:35
sdaguebecause zuul has to be the one that sets up all the git trees that you intend to test with19:35
AJaeger_anteaya, sdague: I found some ways to run less manual jobs - review of would be welcome.19:35
sdagueotherwise you won't be testing what you think it is19:35
dstufftasselin: does your proxy support CONNECT19:35
asselindstufft, ?? not sure. how do you check?19:36
dstuffttry to connect to PyPI using just plain old requests via HTTPS19:36
anteayasdague: thank you, should I add in a comment in devstack-gate perhaps that states that this variable should never be set to FALSE for our needs?19:36
AJaeger_sdague: so, the clone in line 454 of is wrong?19:37
sdagueAJaeger: honestly, I don't know what they did wrong19:37
sdaguebut I know if they think ERROR_ON_CLONE is the fix, they are incorrect :)19:37
mordredclarkb, jeblair: I updated 156088 to fix your comments - when you get a sec, could I get you to look at it, 156247 and 153623 ? They're the last three I need before I can get the PR merged upstream19:37
sdaguebecause it's never the fix19:37
*** mestery_ has joined #openstack-infra19:38
fungiAJaeger_: git_clone is a function which just uses the local cache if there ERROR_ON_CLONE is set19:38
sdagueit just means you are bypassing the testing system, and no longer testing anything meaningful19:38
*** mestery has quit IRC19:38
openstackgerritMerged openstack-infra/project-config: Added Fuel Mellanox plugin
*** andreykurilin_ has quit IRC19:38
dstufftasselin: also there was a bug awhile back where requests wouldn't forget the proxy after ~50 some requests or so19:38
sdaguefungi: actually, it's not even that19:38
dstufftso if you're not on pip 6.0.8 you might want to upgrade, I forget what version fixed it19:38
AJaeger_vponomaryov: please read discussion above for your manila change19:39
fungiAJaeger_: or, rather, which just uses the local cache, so that you won't trip the ERROR_ON_CLONE safeties19:39
sdagueif the directory is there, it never gets far enough to touch ERROR_ON_CLONE19:39
sdagueand that's the point, zuul has to be the one that sets up all the git trees, it's the only thing that has the right global view19:39
fungiyeah, i incorrectly oversimplified19:39
asselindstufft, I'm on 6.0.8
fungibut anyway point being that git_clone is not necessarily running "git clone"19:39
fungiin our case it pretty much never is19:40
*** salv-orlando has quit IRC19:40
AJaeger_sdague, fungi thanks for the explanations. I don't understand everything yet but this helps!19:40
dstufftasselin: OK, well I can't stay here, I jsut had to check something real quick. I have a migraine so I'm going back to lay down. An important thing to figure out is if you can access PyPI with requests. That'll tell you if it's a pip problem or a requests problem (pip uses requests to do HTTP, pip 6.0.8 uses requests 2.5.1).19:41
fungidhellmann is releasing all the things19:42
openstackgerritPaul Michali proposed openstack-infra/project-config: Setup non-voting VPNaaS func test for StrongSwan
* fungi braces for impact19:42
*** dhritishikhar has joined #openstack-infra19:42
asselindstufft, ok thanks. hope you feel better.19:42
dhellmannfungi: I'm done for today.19:42
fungidhellmann: that was a nice large batch19:43
fungiearly and often!19:43
*** marun has quit IRC19:43
dhellmannfungi: most of the actual changes are small. The messaging and db libs will come tomorrow or wed, and those have some bigger changes so I'll space them out more.19:43
openstackgerritAnita Kuno proposed openstack-infra/devstack-gate: Add a comment for ERROR_ON_CLONE
fungicool! i'm excited19:44
*** marun has joined #openstack-infra19:44
*** Sukhdev has quit IRC19:45
anteayasdague fungi can you look at ^^ to see if this is helpful and relevant?19:46
openstackgerritSwaminathan Vasudevan proposed openstack-infra/project-config: Create Neutron DVR experimental job for aiopcpu
anteayafungi: thanks I'll respin19:47
*** zul has quit IRC19:48
fungisdague: wow, yeah...
fungisdague: their devstack plugin is trying to explicitly clone from github19:48
mordredfungi, sdague we may want to figure out how to hard-fail things that try to do that19:49
openstackgerritAnita Kuno proposed openstack-infra/devstack-gate: Add a comment for ERROR_ON_CLONE
mordredin a systemic way19:49
fungimordred: well, we did19:49
*** amrith is now known as _amrith_19:49
mordredfungi: oh yeah? what did we do?19:49
mordredfungi: what if we put a fake hostname for in /etc/hosts ...19:50
sdaguemordred: well, we do19:50
fungimordred: die 622 'Cloning not allowed in this configuration'19:50
sdaguemordred: no, we explode super hard19:50
fungipoint of the story was, that's why they want to work around it by turning off that safety net19:50
mordredoh - they do?19:51
sdagueright, they saw an error, turned off this flag, and thought that was the fix19:51
sdaguewhich I -2ed and said "this is never the fix"19:51
*** e0ne has joined #openstack-infra19:51
*** openstackgerrit has quit IRC19:51
*** dhritishikhar has quit IRC19:52
*** openstackgerrit has joined #openstack-infra19:52
sdagueanteaya: I proposed alternate text19:52
anteayasdague: awesome thanks19:52
* anteaya prepares to copy/paste whatever sdague wrote19:52
* anteaya respings19:53
anteayathat too19:53
*** ivar-lazzaro has quit IRC19:54
*** ivar-lazzaro has joined #openstack-infra19:54
sdaguegerritbot no longer posting in here?19:55
openstackgerritAnita Kuno proposed openstack-infra/devstack-gate: Add a comment for ERROR_ON_CLONE
sdagueoh, there it is19:55
sdaguejust slow19:55
anteayaI was slow19:55
*** bdpayne has joined #openstack-infra19:56
anteayaI posted I was respinning and then edited the patch19:56
anteayamostly this is for me and AJaeger_ for when we hit this again in reviews19:56
*** yjiang5 is now known as yjiang5_away19:56
AJaeger_anteaya: thanks19:56
anteayaAJaeger_: thanks for asking about it19:57
sdaguewell, it might also be useful because people start reading d-g code when their job fails, and do things like turning it off (like we saw)19:57
*** achanda has quit IRC19:57
sdaguehopefully this will slow a little of that down as well19:57
anteayamaybe a little19:57
AJaeger_worth trying ;)19:58
openstackgerritMerged openstack/requirements: Add toposort test requirement
*** achanda has joined #openstack-infra19:59
*** tmcpeak has quit IRC20:01
*** zul has joined #openstack-infra20:01
*** tmcpeak has joined #openstack-infra20:01
*** tkelsey has joined #openstack-infra20:05
*** IAX has joined #openstack-infra20:06
*** rm_work|away is now known as rm_work20:07
*** stevemar has quit IRC20:07
*** IAX has left #openstack-infra20:07
*** markvoelker has joined #openstack-infra20:07
*** mestery_ has quit IRC20:09
*** Rockyg has quit IRC20:09
*** BharatK has quit IRC20:11
*** mestery has joined #openstack-infra20:11
openstackgerritMerged openstack-infra/project-config: Add stackforge/networking-ovn project
anteayarussellb: whom do you want as the first person added to the networking-ovn gerrit groups? yourself?20:13
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Added a Vagrant module and manifest for easier development.
*** tmcpeak has quit IRC20:15
*** amitgandhinz has quit IRC20:15
openstackgerritMerged openstack-infra/project-config: Adding bgpvpn project to stackforge
*** marcusvrn has quit IRC20:15
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Corrected metadata.json
*** thedodd has quit IRC20:17
*** kgiusti has quit IRC20:17
*** zz_johnthetubagu is now known as johnthetubaguy20:17
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Added new OAuth Parameters
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Updated storyboard configuration for pymysql.
*** teran has joined #openstack-infra20:19
*** andreykurilin_ has joined #openstack-infra20:19
*** ivar-lazzaro has quit IRC20:20
*** ivar-lazzaro has joined #openstack-infra20:21
*** baoli has quit IRC20:22
*** kgiusti has joined #openstack-infra20:22
*** ivar-lazzaro has quit IRC20:22
*** ivar-lazzaro has joined #openstack-infra20:23
russellbanteaya: me, yes.20:23
russellbanteaya: and mestery was the 2nd person i was going to seed it with20:23
*** vigneshvar has quit IRC20:23
mesteryrussellb anteaya: ++20:24
openstackgerritMichael Krotscheck proposed openstack-infra/system-config: Added Valid OAuth client configuration to storyboard manifest.
*** unicell has quit IRC20:29
*** unicell has joined #openstack-infra20:29
*** stevemar has joined #openstack-infra20:30
*** unicell1 has joined #openstack-infra20:30
*** unicell has quit IRC20:30
*** erlon has joined #openstack-infra20:30
*** unicell has joined #openstack-infra20:32
*** unicell1 has quit IRC20:32
*** sdake__ has quit IRC20:32
*** ddieterly has quit IRC20:32
*** ddieterly has joined #openstack-infra20:33
zaroAJaeger_: take another look?
*** harlowja has quit IRC20:34
*** harlowja_ has joined #openstack-infra20:34
*** ZZelle_ has joined #openstack-infra20:34
*** mriedem has quit IRC20:35
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Moved working_directory configuration element into correct block.
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Added new OAuth Parameters
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Updated storyboard configuration for pymysql.
*** ddieterly has quit IRC20:37
*** ddieterly has joined #openstack-infra20:38
*** zul has quit IRC20:40
clarkbon the new project creation front, remember puppet is still broken with that20:40
*** zul has joined #openstack-infra20:40
clarkbsorry I haven been mostly afk today silly doctors20:41
*** salv-orlando has joined #openstack-infra20:41
*** baoli has joined #openstack-infra20:44
*** achanda has quit IRC20:45
*** tnovacik has quit IRC20:46
*** ddieterly has quit IRC20:46
*** marun has quit IRC20:49
*** markvoelker has quit IRC20:51
ianwafazekas: here, if you are20:52
*** marun has joined #openstack-infra20:52
*** yjiang5_away has quit IRC20:53
*** yfried_ has quit IRC20:57
*** hichihara has joined #openstack-infra20:59
*** tnovacik has joined #openstack-infra20:59
*** ghostpl_ has joined #openstack-infra20:59
*** tkelsey has quit IRC21:00
*** sandywalsh_ has joined #openstack-infra21:02
mesteryHowdy infra folks! I'm debugging an issue running OpenDaylight in the gate. I'm getting ECONNREFUSED when neutron tries to talk to ODL on port 8088, this works fine locally, wondering if there is an issue with using htis port in the gate?21:03
mesteryFor reference, see the logs here:
*** nicodemos has quit IRC21:03
*** ghostpl_ has quit IRC21:04
mesteryThe ODL tomcat server uses 8080 be default, which swift also uses, so thus the reason for using 8088 instead.21:04
mesteryI had to pick a new port21:04
*** sandywalsh has quit IRC21:05
*** bknudson has quit IRC21:05
*** garyh has quit IRC21:06
*** garyh has joined #openstack-infra21:06
mordredjogo, sdague: question on the NUMA thread ...21:07
sdaguenuma thread?21:07
mordredsdague: on the mailing list21:08
mordredsdague: [openstack-dev] [nova] Ubuntu, qemu, NUMA support21:08
ianwmestery: is tomcat started and running?  it's not clear from the devstack log21:08
mesteryianw: Yes, see here:
mesteryianw: That's the OpenDaylight logs21:08
*** aysyd has quit IRC21:08
ianwmestery: Listening on port 12001 bound on address <- is that something else?21:09
mesteryAdmitedly, there's not much there ... but it's up and running. I even have a (!!!) 5 minute delay built in to let tomcat "settle" before moving on. Shudder ....21:09
russellbclarkb: ah, i'm getting an error pushing commits to a new empty project ... perhaps a known problem?   ! [remote rejected] HEAD -> refs/publish/master (no common ancestry)21:09
mesteryianw: Yes21:09
mordredsdague: does that imply that if I wanted to run a kvm openstack on trusty, that because of qemu compile options I would not be able to?21:09
mordredwithout a backport package21:09
mordredsdague: this is really just me trying to learn how some things hang together21:09
clarkbrussellb thats the broken project creation I think21:10
russellbclarkb: OK, thanks21:10
*** mestery has quit IRC21:10
sdaguemordred: the answer should be no21:10
sdaguelet me go refresh on the thread21:11
*** mestery has joined #openstack-infra21:11
russellbclarkb: worth filing a bug for?21:11
mesteryianw: My VPN bounced back now21:11
mesteryianw: You can see I put in some debug to look at the tomcat config file here:
mordredrussellb: I'm working on it - it just opened up a rabbit hole21:11
russellbmordred: ah ok, thanks!  i'll sit tight then21:11
mordredrussellb: and I think I took the blue pill by accident21:11
russellbit happens21:12
mordredrussellb: I mean, I'm not working on your particular project this instant - which project is it?21:12
*** kgiusti has quit IRC21:12
*** ayoung has quit IRC21:12
russellbmordred: stackforge/networking-ovn ... not urgent by any stretch21:12
*** salv-orlando has quit IRC21:12
mordredkk. thanks21:12
sdaguemordred: so basically, if you want nodes to do numa'y things, you have to give it some extra info21:13
mordredsdague: and the mechanism to do that is via qemu?21:13
*** achanda has joined #openstack-infra21:13
sdaguethe way this *should* work is that if the nodes can't do that, it's explicit about it, or fails. Right now the numa code paths are largely untested, so it explodes a lot21:13
*** marun has quit IRC21:13
sdaguemordred: flavor and image attributes21:14
*** dustins has quit IRC21:14
*** rlandy has quit IRC21:14
mordredsdague: awesome21:14
mordredsdague: I love exploding things21:14
*** ddieterly has joined #openstack-infra21:14
ianwmestery: nothing obvious ... i'd start the job again and ask carkb or jeblair to put the host on hold for you, and log in to see what's going on21:15
*** markvoelker has joined #openstack-infra21:15
*** sdake_ has joined #openstack-infra21:15
fungirussellb: you said to add you as the initial member of networking-ovn-core and networking-ovn-release? i'll take care of that now21:15
russellbfungi: yes, and mestery while you're at it if you don't mind21:15
russellbfungi: thank you!21:15
mesteryianw: Good call, I'll do that and see.21:15
mesteryrussellb fungi: Thanks!21:15
*** markvoelker has quit IRC21:15
fungirussellb: mestery: done and done21:16
*** markvoelker has joined #openstack-infra21:16
openstackgerritMerged openstack-infra/jenkins-job-builder: adds support for not condition in conditional_step
mesteryfungi: thank you sir!21:16
openstackgerritDouglas Mendizábal proposed openstack-infra/project-config: Add a new devstack-gate job to Barbican's tests
*** dustins has joined #openstack-infra21:16
*** salv-orlando has joined #openstack-infra21:17
*** ldnunes has quit IRC21:17
*** sdake__ has joined #openstack-infra21:17
mesteryclarkb fungi: I'm trying to debug an ODL thing in the gate, any chacne you can keep the -dvsm node from this review around for me for a bit:
mesteryfungi clarkb: I'm headed to a Meetup in 30 minutes or so, but will be online when I land there and can look at this during the Meetup21:18
fungirussellb: i don't see any indication that you should be getting an error. did you base your commit on the initial state of that repo?21:18
sdaguemtreinish was working on a tempest test that called the numa code kind of on a null path to try to reduce the random explody, but at least at the nova midcycle that wasn't landable21:18
russellbfungi: oh, no ... it's a repo i initialized locally.  d'oh.21:18
fungirussellb: that would entirely explain the error message21:18
mtreinishsdague: well I was helping w/ reviews and stuff, but I didn't write it21:18
russellbfungi: quite ...21:18
sdaguemtreinish: is that passing yet?21:19
russellbfungi: i'm very jet lagged, that's my excuse21:19
mtreinishsdague: last I looked they finally got it to a point where it wouldn't fail randomly21:19
fungimestery: local firewall rules blocking access on that port on the worker maybe?21:19
sdagueok, that would be good to land then21:19
sdagueto prevent regressions21:19
mtreinishsdague: but we can't land it because it's still a 100% tied to some env config21:19
sdagueyou got the link?'21:19
mesteryfungi: That was my thinking too! Where to look for that? Any pointers?21:19
sdagueoh, ok21:19
mtreinishlike it broke xenapi ci21:19
* fungi goes looking for where devstack iptables gets tweaked21:19
* mestery wanders behind fungi with a flashlight21:19
russellbfungi: mordred yeah, project works fine, nevermind.  thanks21:20
sdaguemtreinish: hmmm.... so nova won't reject a numa flavor create if no driver in the cloud supports it?21:21
*** sdake_ has quit IRC21:21
sdaguethat seems like the safety lever we need there actually21:21
fungimestery: i do see at least lib/ironic in openstack-dev/devstack adding rules to iptables to permit additional traffic it needs for its features21:21
mtreinishsdague: yeah that's what it looks like21:21
mordredrussellb: yay!21:21
mesteryfungi: I'll give that a shot, thanks for the pointer!21:21
fungiperhaps other external devstack plugins do somilar things21:22
fungisimilar too21:22
*** otter768 has joined #openstack-infra21:22
*** radez is now known as radez_g0n321:23
*** ChuckC has joined #openstack-infra21:25
*** vilobhmm2 has joined #openstack-infra21:25
*** vilobhmm has quit IRC21:25
*** vilobhmm1 has quit IRC21:25
*** _amrith_ is now known as amrith21:26
*** otter768 has quit IRC21:27
*** baoli has quit IRC21:27
openstackgerritMerged openstack-infra/devstack-gate: Explicitly turn on database query logging
*** hdd has joined #openstack-infra21:30
jeblairclarkb, fungi: can you think of a way to indicate what kind of CPU resources we would need from a potential new nodepool provider cloud?21:30
jeblairclarkb, fungi: i tell people "8 vcpus" as a ballpark, but i mean, that's highly variable depending on what kind of cpu we're talking about, and what kind of oversubscription21:31
*** yjiang5 has joined #openstack-infra21:31
fungioversubscription aside, aggregate bogomips?21:33
fungi(we want the machine to be able to do absolutely nothing this many times a second)21:33
mordredjeblair: I was just talking to sdague - and apparently we're over 50% idle for large swaths of time21:35
sdagueyeh, we have dstat numbers on every job21:35
SpamapSget to the idle times faster? ;)21:35
sdagueso you can data mine it21:35
jeblairfungi: they are surprisingly consistent.  5186 on hpcloud and 5200 on rax.21:35
*** tmcpeak has joined #openstack-infra21:35
sdagueSpamapS: yeh, but if we are going for agregate through put, I'm not convinced21:36
openstackgerritKhai Do proposed openstack-infra/jenkins-job-builder: The email publisher needs a list of recipients
sdagueif it's node pool nodes, oversubscribed, but not cpu rate limitted actually works nicely21:36
sdaguethat's what hp cloud old was21:36
fungijeblair: is that taking the vcpu count into consideration?21:36
jeblairfungi: i don't know, that's the raw value printed in dmesg21:36
fungilooks like that's per vcpu21:37
*** unicell has quit IRC21:39
*** unicell1 has joined #openstack-infra21:39
sdaguejeblair: maybe we should run a linpack run on our nodepool nodes21:40
sdagueand say "at least this fast on linpack"21:40
*** dizquierdo has joined #openstack-infra21:40
mordredoh - actually21:40
mordredyou know what?21:40
mordredthere's that new google cloud profiler thing21:40
sdagueor something equiv21:40
nibalizerya that21:40
nibalizeri thought of running it on hpcloud21:41
nibalizerbut i didn't want to cry21:41
nibalizeralso I think we'd have to kick it a bit to work on openstack21:41
fungioh, neat21:41
sdagueyeh, wasn't that a suggested summer of code project by ttx21:41
mordredyah - well, porting it to shade and then figuring out how to regularly run it on our clouds or something might be informative21:41
*** AJaeger_ has quit IRC21:42
jeblairalso, i guess, "runs devstack-gate in < 1hr" is a benchmark too, just not super easy to run right now :)21:42
clarkbjeblair the neutron unittests are much easier21:43
*** mrmartin has quit IRC21:43
jeblairclarkb: good point21:43
pleia2ok, I think I managed finally to dig myself out of my email backlog, running off to dr appt now, should finally be able to get back to reviews and all things infra after that21:43
jeblairclarkb: the're around 1 hour now?21:43
clarkbpleia2 thats what I keep saying21:43
pleia2clarkb: hah :)21:43
*** tiswanso_ has quit IRC21:43
*** dustins has quit IRC21:43
clarkbjeblair I think 40minutes21:44
clarkbpleia2 in second doctors office of the day right now21:44
pleia2clarkb: oh, fun21:44
jeblairclarkb: huh, i found one that's 10 minutes.21:44
jeblair2015-02-23 20:18:26.670 | Ran: 10967 tests in 510.0000 sec.21:44
*** sdake_ has joined #openstack-infra21:45
sdagueyeh, with the drivers out of tree they got shorter21:45
*** btran has joined #openstack-infra21:45
clarkbjeblair 4478 seconds according to random log I looked at21:45
*** dustins has joined #openstack-infra21:46
*** achanda has quit IRC21:46
asselinhi, anyone know where ubuntu_mysql.qcow2 is supposed to be cached in 3rd party ci?
clarkbasselin into the devstack images dir21:47
jeblairclarkb: yeah, that's the linearized sum, but still shorter in parallel.21:47
*** achanda has joined #openstack-infra21:47
clarkbwait gah21:47
clarkbso many different subunit formatters now they make no sense to me21:47
clarkbya that one was only 612 secinds21:48
clarkbmaybe nova is the long unittest project now21:48
jeblairclarkb: i looked at one which was 12min21:48
*** dustins has quit IRC21:48
mtreinishclarkb: arew you complaining about:
*** sdake__ has quit IRC21:49
clarkbmtreinish I just find it all too noisy21:49
asselinclarkb, I see them here: /opt/stack/new/devstack/files, but which nodepool image script is supposed to save it? maybe I need to search some more.21:49
*** MaxV has joined #openstack-infra21:50
clarkbasselin the devstack nodepool scripts do it21:50
mtreinishclarkb: sure, it's a delicate balance we want some debug info but not too much21:50
clarkbmtreinish also hard to read those logs on my phone :)21:50
mtreinishclarkb: if you've got a better suggestion for the output push it up, nothing there is really set in stone21:50
mtreinishclarkb: that's why I got a nexus 6...21:51
clarkbmtreinish I prefer default testr21:51
mtreinishbigger screen :)21:51
*** mestery has quit IRC21:51
clarkbshow me failures, total tests run and run time21:51
clarkbsimple and only the info I need21:51
*** abhirc has joined #openstack-infra21:52
sdagueclarkb: I hate default testr, any program that runs for > 20 seconds with no status progress is problematic. Especially when you are running code that can deadlock.21:52
*** bswartz has quit IRC21:53
clarkbsdague the cases wherr that happens are infrequent enough to not default to always having necessary output to debug deadlocks21:53
fungicarefully going back over the packages we preinstall on thick slaves right now, are we unable to pip install any of python-libvirt, python-lxml and python-zmq? thinking we should see about trying to leave as many of those as possible up to per-project virtualenvs21:53
clarkbcan just turn it on when needed21:53
sdagueor all the times where it incorrectly skipped things and passed21:54
clarkbsdague that info is written21:54
sdagueand because it wasn't verbose, it took us weeks to realize that we missed tests21:54
mordredfungi: I know we can pip lxml21:54
mtreinishclarkb: from what I've seen you're actually the minority, most people like to see that it's doing something as it runs21:54
clarkbmtreinish yes I am21:54
mordredfungi: I _think_ libvirt is pip installin gnow?21:54
clarkbI know this I just dont grok it21:54
clarkbnordred no they stopped21:55
clarkbmordred ^21:55
mordredclarkb: really? they gave up on that?21:55
clarkbso we dont need libvirt-python at all21:55
clarkbyup using fakes now21:55
mordredthat's even better21:55
clarkbsure it just wouldve been nice to do years ago :)21:55
jeblairmtreinish, clarkb: am also in that minority.  i would be okay with some output to see deadlocks when run interactively, but i don't actually find 10,000 lines of "everything is okay" to be helpful in that department.21:56
fungiwell, i'm 99% of the way to having a complete cross-platform bindep manifest now which will turn "devstack" workers into "bare" workers package-wise21:56
jeblairmtreinish, clarkb: (i also don't find no output to be helpful)21:56
clarkbpython-zmq is pip installable too but it grabs C zmq from an external source which may not be good21:56
fungioh, that a'splains why we don't use it then21:57
*** rkukura has quit IRC21:57
clarkbit will link against local C if it finds it21:57
clarkblibzmq-dev or somesuvh21:57
mtreinishjeblair: yeah, it comes back to that delicate balance. If you guys have a better suggestion feel free to push up a patch :)21:57
fungiahh, we should probably add that to the bindep manifest when we get around to making improvements to the packageset then21:58
fungifirst i'm trying to replicate what we have as closely as possible, so as to avoid unpleasant surprises21:59
*** rkukura_ has joined #openstack-infra22:01
sdagueclarkb: well, as someone that spends a ton of my time running tests locally, the progress part is really important, especially when I'm making cross cutting changes, because lots of things can make these tests go until the 300s timeout. And waiting 5 minutes for first fails to show up because something went terrible is hugely non productive.22:01
*** ddieterly has quit IRC22:01
fungione thing i've noticed this is going to require a change in though... we'll need to start running the gerrit-git-prep builder before the revoke-sudo builder if we want projects to be able to specify the system packages to add for a particular job22:01
jeblairfungi: yeah, i think that should be safe22:02
clarkbsdague would a simple: no tests reported data in 30 seconds (last file tests were xyz) message fix that problem?22:02
*** ddieterly has joined #openstack-infra22:03
sdaguenot really, you're still putting an extra 27 second delay in my iteration loop in that case22:03
openstackgerritMerged openstack-infra/gear: Fix SSL non-blocking IO support
clarkb27 seconds?22:04
sdaguemost of these nova tests should complete in 3 seconds22:04
openstackgerritRicardo Carrillo Cruz proposed openstack-infra/shade: Add initial compute functional tests to Shade
clarkbthen set the nova timeout to 4 seconds22:05
openstackgerritMerged openstack-infra/gear: Revert "Pin python-daemon to <2.0"
jeblairsdague: do you run tests serially locally?22:05
sdaguethat first line of good output means I can get up and get a cup of coffee22:05
clarkboverride for longer tests22:05
sdaguejeblair: nope22:05
*** baoli has joined #openstack-infra22:05
jeblairsdague: how do you pick out a stuck test then?22:05
*** tkelsey has joined #openstack-infra22:06
rcarrillocruzmordred: , if you like the approach i can add more tests in bulk...22:06
rcarrillocruzjeblair: ^22:06
sdaguewe're displaying worker id, so you can figure out if a worker is stuck22:06
jeblairsdague: (i'm genuinely curious; for zuul, i run the suite in parallel, get "coffee", then if anything failed or timed out, run that one test in the foreground)22:06
clarkbsdague (timeout is overrideable on the cli too)22:06
jeblairsdague: so you watch the stream and when you don't see a worker id for a while, you figure it's stuck?22:06
*** Ryan_Lane has quit IRC22:06
*** tonytan_brb is now known as tonytan4ever22:07
sdaguejeblair: yeh, or more importantly, know that all the workers seem to be off and running22:07
jeblairsdague: oh, you're only concerned that they have run one test successfully to completion?  what does that tell you?  can't any test hang?22:07
fungii start tests and then do
sdagueclarkb: the argument we are having is that you are actually telling me that I develop software incorrectly, and that I should not develop software the way that works for me.22:08
jeblairsdague: i guess it tells you that the system isn't so broken that all tests hang?22:08
mordredrcarrillocruz: LOVE IT22:08
sdaguejeblair: yeh, and honestly, it sometimes is22:08
clarkbno I am only suggesting that your way maybe isnt the best default for everyone22:08
rcarrillocruzi'll build on that then22:09
mordredrcarrillocruz: although I'd put the cloud = openstack_cloud() in the setUp and make it = - just so that you don't have to do it every time22:09
mordredbut that might just be me22:09
sdagueclarkb: sure, I accept that22:09
rcarrillocruzgood call22:09
jeblairsdague: (to be clear, i'm not participating in an argument, i am genuinely trying to understand your workflow which requires me asking a lot of simple sounding questions)22:09
*** rkukura_ is now known as rkukura22:11
sdaguejeblair: personally without the progress displays my mental iteration loop just doesn't work. And if I can't fast iterate on this stuff my throughput goes way down.22:11
sdaguealso, after we added the subunit-trace function, which always displays the subunit stderr, we found a bunch of tests that were doing terrible things, stack tracing components and still passing22:12
jeblairSpamapS: see comments on 15914622:12
jeblairSpamapS: correction: 15814622:13
sdaguewhich is why I feel that possitive output is a really important part of creating good tests22:13
*** rkukura has left #openstack-infra22:13
sdagueand test runnings that hide everything except final error code make it easy to make bad mistakes in your test suite so it doesn't do what you think it's doing22:13
*** radez_g0n3 is now known as radez22:16
SpamapSjeblair: ty22:16
*** JayJ has quit IRC22:16
clarkbmy favorite is sudo fails but test passes (dont think tracing would catch that though but maybe depends on how its logged)22:17
SpamapSjeblair: I may also squash the series since it's really all the same thing: make python3 work.22:17
*** JayJ has joined #openstack-infra22:17
sdagueclarkb: yeh, I know we've still got a bunch of things mocked wrong in nova functional tests where we're running of of ips in tests, stack traces, tests still pass. Haven22:17
sdague't gotten around to pulling all that appart yet22:18
*** vilobhmm1 has joined #openstack-infra22:18
*** vilobhmm2 has quit IRC22:18
openstackgerritRicardo Carrillo Cruz proposed openstack-infra/shade: Add initial compute functional tests to Shade
afazekasianw: Is anybody working on switching the kernel in the f21 job ?22:20
ianwafazekas: yeah, i'm digging into it now22:20
ianwafazekas: it seems that the extlinux.conf doesn't get rewritten correctly by a new kernel install22:20
SpamapSjeblair: also I may stop caring much about the py3 work.. there's some demons in the threading there... I keep finding dead backgrounded unit tests running py3.4 tests on my machine.22:21
openstackgerritRicardo Carrillo Cruz proposed openstack-infra/shade: Add initial compute functional tests to Shade
ianwafazekas: at least on rax22:21
afazekasianw: The original image has different extlinux.conf then the installer expects22:21
SpamapSone of those "I will debug this when py3.4 is the alternative to a large wall of flame." moments22:21
ianwafazekas: there's two it seems ... /boot/extlinux.conf and /boot/extlinux/extlinux.conf, the latter is an anaconda one, the first seems hand-written22:22
ianwafazekas: it doesn't seem possible to regenerate a new one from scratch ... or if it is it's not obvious22:22
jeblairSpamapS: oh wow :(  i'd really like it to work with python 3.  i can't pick that up right now, but if you give up i will try to look into it at some point.22:23
afazekasianw: the installer uses the grubby command, but it may have issues if you use spaces in the name22:23
*** sarob has quit IRC22:23
*** JayJ has quit IRC22:23
*** JayJ has joined #openstack-infra22:24
SpamapSjeblair: they've fixed a lot of GIL things in 3.4, so it's entirely possible things are running better and there are unfound races.22:24
afazekasianw: extlinux support is not in the grubby's man page, it is kind of new thing..22:24
SpamapSjeblair: it's also possible python-daemon 2.0.x is still awful.22:24
clarkbafazekas: ianw why would we switch the kernel?22:25
ianwclarkb: because the one in the image doesn't work22:25
clarkbwhat is wrong with it? and should we also be working on getting the base images fixed?22:26
afazekasclarkb: The f21 job is very likely fails because of kernel issue22:26
*** tmcpeak has quit IRC22:26
clarkbgertty syncing over tethered connection always makes me wish I synced prior to leaving home22:27
afazekasclarkb: the kernel itself is updated in the image, but the old one gets booted22:27
ianwclarkb: you'd think "yum update kernel" would update the kernel, but it doesn't on rax ... yes it's a underlying issue with the rax image22:27
*** tmcpeak has joined #openstack-infra22:27
clarkbafazekas: ianw gotcha22:27
clarkbthat sounds ... unpleasant22:27
ianwclarkb: it looks like someone hand-wrote a extlinux.conf that confuses the update mechanism22:28
jeblairclarkb: it should be fast!  also, if you are running with 155615, you can open a change directly, refresh, and it will prioritize that over the bg sync22:28
afazekasThere is two way for setting default image in extlinux22:28
afazekasone is the default option, the other is the timeout22:28
clarkbjeblair: its much better than before I just think that my connection is sad22:28
afazekasf21 expects the default, as remember the rax image usaes the timeout22:28
*** JayJ has quit IRC22:28
*** mattfarina has quit IRC22:28
clarkbjeblair: I will have to grab 15561522:28
*** harlowja_ has quit IRC22:29
*** JayJ has joined #openstack-infra22:29
jeblairclarkb: 155773 is my current tip and has that as a parent22:29
* afazekas sleep22:29
jeblairclarkb: (also starred changes)22:29
*** harlowja has joined #openstack-infra22:30
ianwafazekas: thanks for you help, i'll email to update.  i think solution might be to copy anaconda config and then update kernel22:30
*** pc_m has quit IRC22:30
clarkbjeblair: thanks will grab that22:31
clarkbianw: should we poke rax too about hteir image?22:31
clarkbianw: if so jroll and JayF hav ebeen super handy22:31
jrollwhat'd we break now22:32
ianwjroll: on the rax f21 image, there's a /boot/extlinux.conf that looks like
clarkbjroll: the f21 base image supplied by rax apparently is configured in a away that prevents its kernel form being updated with yum22:33
clarkbjeblair: woah that was a fast sync, I think there must be even newer magic than what I was expecting on the fast sync front22:33
*** JayJ has quit IRC22:33
ianwjroll: when you yum update, it messes things up.  the APPEND line gets copied (so you get the wrong initramfs) and the default boot isn't updated22:34
*** JayJ has joined #openstack-infra22:34
*** JayJ has quit IRC22:34
jrollianw: mind emailing details to and I'll pass it along to people that deal with that stuff?22:34
ianwjroll: i think you can probably recreate this with grubby --some --series --of --not --well --documented --commands22:34
ianwjroll: sure22:35
clarkbjroll: you sir are awesome22:35
jrollianw: clarkb: thanks, I do what I can22:35
*** tonytan4ever has quit IRC22:35
*** dkranz has quit IRC22:36
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Apache Config is now an optional module.
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Parameterized working directories.
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: StoryBoard is now installed into a Virtualenv
openstackgerritMichael Krotscheck proposed openstack-infra/puppet-storyboard: Moved all module parameters into one manifest.
openstackgerritMichael Krotscheck proposed openstack-infra/system-config: Use StoryBoard optional https manifests.
openstackgerritMichael Krotscheck proposed openstack-infra/system-config: Centralized storyboard module configuration.
openstackgerritMichael Krotscheck proposed openstack-infra/system-config: Hostname and CORS domains are now configurable.
clarkbmordred: 158173 is jenkins sad22:37
clarkbmordred: looks like its near the first of the changes needed for the ansible work22:37
mordredclarkb: oh - piddle22:37
mordredclarkb: I was hacking on that one mildly blind on the plane - but yes, it is22:38
mordredI'll jump on it22:38
mordredoh - I get what's wrong22:38
mordredclarkb: patch in just a sec22:39
clarkbmordred: thanks, will review as soon as its available22:39
*** mtanino has quit IRC22:40
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Add an index on accesstokens table for access_token column
clarkbif others are able to review and it would be a great help for the swift logs22:42
clarkbone is a bug fix for zuul's handlingof layout values and the other is sets up swift uploads in parallel22:42
*** achanda has quit IRC22:43
*** johnthetubaguy is now known as zz_johnthetubagu22:44
jeblairclarkb: ack, will do22:45
*** achanda has joined #openstack-infra22:45
*** tkelsey has quit IRC22:46
openstackgerritMonty Taylor proposed openstack-infra/project-config: Add ansible role for running puppet
mordredclarkb: ^^22:47
*** dmsimard is now known as dmsimard_away22:47
*** tkelsey has joined #openstack-infra22:50
*** lnr has joined #openstack-infra22:51
clarkbjhesketh: can you point out the zuul change that allows reporting as not jenkins?22:51
clarkbjhesketh: you have a bunch of changes and want to make sure I orientate myself properly22:51
clarkbmordred: you need to rebase :022:52
clarkber :)22:52
*** lnr has left #openstack-infra22:52
mordredclarkb: :(22:53
*** sarob has joined #openstack-infra22:55
*** radez is now known as radez_g0n322:55
clarkbjhesketh: question for you on the zuul merger run in zuul server change22:56
clarkbjeblair: ^ you may find that interesting too22:56
openstackgerritSwaminathan Vasudevan proposed openstack-infra/project-config: Create Neutron DVR experimental job for aiopcpu
*** hdd has quit IRC22:57
*** tkelsey has quit IRC22:58
*** tmcpeak has quit IRC22:59
*** e0ne has quit IRC23:00
jeblairclarkb, jhesketh: yeah, i think that's what i was getting at in my original comment.  i think it would be good to be consistent.23:01
*** tmcpeak has joined #openstack-infra23:02
jogoclarkb sdague: re using pentium for live migration23:02
jheskethclarkb: (and all its dependencies)23:02
clarkbjhesketh: thank you23:02
jheskethwill take a look at that merger process23:03
openstackgerritMonty Taylor proposed openstack-infra/project-config: Add ansible role for running puppet
mordredclarkb: ok - that's actually nicer and smaller now23:03
*** harlowja has quit IRC23:04
*** rwsu is now known as rwsu-afk23:04
jogosdague: ahh I see you found the results, I wonder if those were random failures or not23:05
jheskethclarkb: on the swift logs, when we have time, we should have a discussion to what level we want to support source/destination type notation that jenkins uses23:05
*** hamzy has quit IRC23:06
*** hamzy has joined #openstack-infra23:06
*** Ryan_Lane has joined #openstack-infra23:06
clarkbjhesketh: for job configuration?23:06
clarkbjhesketh: I think python has stdlib glob support, if we can just do that for source then hvae the bit for preserve hierarchy thats probably sufficient23:07
*** stevemar has quit IRC23:07
*** stevemar has joined #openstack-infra23:07
openstackgerritJoe Gordon proposed openstack-infra/devstack-gate: Set custom cpu_model for live_migrate
*** camunoz_gone is now known as camunoz23:08
jogoretrying pentium to get some another data point to check for consistent failures (I suspect it will fail the same way)23:08
jheskethclarkb: hmm, that shouldn't be too hard23:08
jheskethclarkb: the alternative would be to copy all the files you want into a staging dir and uploading that.. it'll make the indexes more straight forward anyway23:09
jheskethclarkb: the glob stuff doesn't seem too difficult though so I'll tackle that today23:10
clarkbjhesketh: cool23:10
*** baoli has quit IRC23:11
*** mfink_ has quit IRC23:12
*** achanda has quit IRC23:12
clarkbjogo: sdague I have a hunch its doing something funny to arch that breaks us23:13
clarkbjogo: that said, what is in the kvm64 cpu type? I wonder if that is super generic and would work well23:13
*** devlaps has quit IRC23:14
*** devlaps has joined #openstack-infra23:14
*** sarob has quit IRC23:14
jogoclarkb: once the pentium one fails, we can try kvm64 again23:15
jogobut rax is xen ...23:16
*** harlowja has joined #openstack-infra23:16
*** sarob has joined #openstack-infra23:16
*** baoli has joined #openstack-infra23:17
asselindstufft, my coworker noticed that the pip packages installed aren't the latest. and don't match what you said they should:
*** MaxV has quit IRC23:17
clarkbjogo: right but we are using qemu on the nested level23:17
jogolooks pretty generic to me :)23:18
openstackgerritKhai Do proposed openstack-infra/project-config: Refactor Gerrit plugin build jobs
*** sarob has quit IRC23:19
clarkbjogo: ya I think that may actually be what we need but we shall see23:19
*** sarob has joined #openstack-infra23:20
openstackgerritTim Kelsey proposed openstack-infra/project-config: Adding anchor gate tests
dstufftasselin: pip doesn't depend on stuff like a normal package, we bundle our dependencies inside of pip._vendor.*23:20
dstufftso you can't use pip freeze to see what version of requests we're using23:20
dstufft(unless youy're using python-pip from a Linux vendor that undoes that)23:20
asselindstufft, I'm using what -infra/devstack is using...23:21
mordredthat should be the latest from git-pip23:21
clarkb hopefully I am not too far off base on that23:22
jeblairmordred: i'm pretty sure if git-pip were a thing, it would break us up and down23:22
clarkbjogo: I don't think we want i38623:22
dstufftjogo: does setuptools-git count? :V23:23
asselinI am seeing this recent change that looks related :
jeblairclarkb: i am -2 on having our zuul use two accounts23:23
dstuffttoo many j names23:23
*** otter768 has joined #openstack-infra23:23
clarkbjeblair: how would you approcah that problem then?23:23
jogoclarkb: that seems to be where kvm64 and pentium are defined23:23
clarkbjeblair: eg how do we run tests that provide useful info that are not gating and actually looked at by humans23:23
jeblairclarkb: check jobs23:24
clarkbjogo: oh, I should read into paths less then23:24
clarkbjeblair: except we clean check23:24
clarkbjeblair: and jgriffith already blew up at us for even suggesting that route23:24
jeblairclarkb: why not gate?23:24
jeblairblew up?23:24
clarkbjeblair: that is a question for jgriffith23:24
clarkbjeblair: yes23:24
clarkbhold on23:24
jeblairclarkb: i must have missed that meeting23:24
*** sarob has quit IRC23:24
clarkbjeblair: precipitated much of this23:25
* mordred is also confused23:25
*** achanda has joined #openstack-infra23:25
*** oomichi has joined #openstack-infra23:25
mordredoh. hrm. interesting23:25
clarkbI have no problems with having these tests check/gate23:26
clarkbbut others do23:26
mordredso - the rule of what can run in the gate has always been "is it open source"23:26
mordredit's just that most projects do not have multiple open source options23:26
clarkb(I responded to that email too)23:26
StevenKmordred: I am getting awfully close to teaching delorean about building .debs successfully23:28
*** otter768 has quit IRC23:28
StevenK2015/02/24 10:27:55 Error response from daemon: Conflict, The name builder-ubuntu is already assigned to 31c8b7c0465d. You have to delete (or rename) that container to be able to assign builder-ubuntu to a container again.23:28
StevenKThat ^ turning up every single time is getting annoying23:28
jeblairclarkb: (to complete a thought from earlier, a check job that should not gate should be non-voting because of clean check)23:28
clarkbjeblair: ok23:29
clarkbjeblair: I think the concern there is non voting is often ignored, BUT we can probably fix that as a UI thing?23:29
jeblairclarkb: i think we already have23:29
*** dprince has quit IRC23:29
*** sabeen3 has quit IRC23:30
clarkbit does make me happy that that is the route you would like because its less upfront work to get this going so win win :)23:30
mordredStevenK: neat!23:30
mordredStevenK: and blech23:30
jeblairclarkb: yay!  i am also happy you are happy :)23:30
*** baoli has quit IRC23:30
* mordred is happy that jeblair is happy that clarkb is happy, and also that clarkb is happy23:31
fungii'm still wondering how we reconcile the fact that, to take this specific example, cinder tells the sheepdog driver maintainers they must have integration testing reported on cinder changes to stay supported, infra tells the sheepdog driver maintainers that we're able to run their tests for them no need for an external ci rig, but cinder doesn't want job results from "jenkins" for this23:34
asselinanyone know how this recent devstck change in tools/ may ignore  pip.conf configurations?
jeblairfungi: i believe the problem is over-application of bureaucracy.  let's use less of it.  :)23:34
fungiit almost seems like projects are getting stockholm syndrome from third-party testing23:35
*** krtaylor has quit IRC23:35
asselin* ignore pip.conf configurations for virtual environments23:36
*** prad has quit IRC23:36
*** mfisch has quit IRC23:36
fungiwhere third-party testing was supposed to be an exceptional workaround for when we can't support the tests needed for some component, it's now been turned into some panacea23:36
*** alexpilotti has joined #openstack-infra23:36
*** carl_baldwin has quit IRC23:37
*** gordc has quit IRC23:37
fungiasselin: global /etc/pip.conf? something else?23:37
clarkbfungi: yes, because projects told everyone you must run third party testing. not "hey work with infra to get your test workload tested"23:37
clarkbfungi: it really set up the solution space in one way23:38
asselinfungi, I updated that too. chatting with dtroyer in -qa23:38
openstackgerritKhai Do proposed openstack-infra/publications: Update logos on System slide
openstackgerritMax Rydahl Andersen proposed openstack-infra/jenkins-job-builder: Initial work for supporting downstream-ext
jeblairclarkb: will respond to your email23:40
clarkbjeblair: thanks23:40
*** abhirc has quit IRC23:41
*** armax has joined #openstack-infra23:41
*** abhirc has joined #openstack-infra23:41
*** ivar-laz_ has joined #openstack-infra23:42
*** andreykurilin_ has quit IRC23:43
*** ivar-lazzaro has quit IRC23:45
*** salv-orlando has quit IRC23:45
*** carl_baldwin has joined #openstack-infra23:47
*** krtaylor has joined #openstack-infra23:47
*** che-arne has quit IRC23:47
*** ddieterly has quit IRC23:50
clarkbmordred: jenkins still doesn't like you :)23:54
clarkbmordred: I think you can tox -elayout locally to run it on your laptop23:54
jheskethclarkb: is there a meaning to '**' within jenkings?23:55
clarkbjhesketh: ya its a recursive *23:55
clarkbjhesketh: so foo/**/bar matches foo/stuff/here/toarbitrary/depth/bar23:55
jheskethah I see23:56
*** MarkAtwood_ has quit IRC23:56
*** MarkAtwood has quit IRC23:57
*** marcusvrn has joined #openstack-infra23:57
jheskethclarkb: does't look like that is part of the python glob library23:57
*** melwitt has quit IRC23:57
jheskethwill probably have to walk the tree23:57
clarkbjhesketh: its probably ok to not support * and be more specific in jobs23:58
clarkbjhesketh: I doubt there are any places where we really need **23:58
*** ivar-laz_ has quit IRC23:58
jheskethclarkb: true, but we'd need to check nearly every job as most of them use test-results which look like "**/*testr_results.html.gz".. not sure if we can assume they are always in the same place23:59
*** ivar-lazzaro has joined #openstack-infra23:59
*** dizquierdo has quit IRC23:59
clarkbjhesketh: ya, I think since we end up templating and cargo culting so much it won't be that much work23:59
clarkbjhesketh: I can try digging into that if you like23:59

Generated by 2.14.0 by Marius Gedminas - find it at!