Monday, 2014-04-14

*** matsuhashi has joined #openstack-infra00:01
*** Guest41065 is now known as StevenK00:06
*** e0ne has joined #openstack-infra00:17
*** CaptTofu has joined #openstack-infra00:18
*** adalbas has quit IRC00:26
*** ildikov_ has joined #openstack-infra00:29
*** fifieldt has joined #openstack-infra00:30
*** ildikov has quit IRC00:31
*** shakamunyi has joined #openstack-infra00:31
*** shakamunyi has quit IRC00:37
*** adalbas has joined #openstack-infra00:42
*** mestery has joined #openstack-infra00:48
openstackgerritDavid Pursehouse proposed a change to openstack-infra/jenkins-job-builder: Add support for multiple Gerrit servers in Gerrit trigger
*** wchrisj has joined #openstack-infra00:56
*** unicell has joined #openstack-infra00:57
*** ildikov has joined #openstack-infra00:59
*** mriedem1 has quit IRC01:01
*** ildikov_ has quit IRC01:01
*** mriedem has joined #openstack-infra01:05
*** nati_ueno has joined #openstack-infra01:12
*** derekh has quit IRC01:15
*** ianw has quit IRC01:16
*** ianw has joined #openstack-infra01:16
*** CaptTofu has quit IRC01:29
*** CaptTofu has joined #openstack-infra01:29
*** wchrisj has quit IRC01:32
*** yaguang has joined #openstack-infra01:36
*** nosnos has joined #openstack-infra01:41
*** saschpe has quit IRC01:42
*** slagle has joined #openstack-infra01:43
*** saschpe has joined #openstack-infra01:43
openstackgerritDavid Pursehouse proposed a change to openstack-infra/jenkins-job-builder: Add support for triggering on unreviewed patches
*** miqui has joined #openstack-infra01:53
*** e0ne has quit IRC02:01
*** mriedem has quit IRC02:04
*** dstanek has quit IRC02:14
*** nati_ueno has quit IRC02:21
*** dstanek has joined #openstack-infra02:22
*** e0ne has joined #openstack-infra02:27
*** jcoufal has quit IRC02:28
*** e0ne has quit IRC02:30
*** e0ne_ has joined #openstack-infra02:30
*** pcrews__ has joined #openstack-infra02:33
*** CaptTofu has quit IRC02:35
*** CaptTofu has joined #openstack-infra02:35
*** coolsvap|afk is now known as coolsvap02:38
*** amotoki has quit IRC02:38
*** CaptTofu has quit IRC02:40
*** zhiyan_ is now known as zhiyan02:41
*** Longgeek has joined #openstack-infra02:48
*** adalbas has quit IRC02:51
*** talluri has quit IRC03:14
*** nati_ueno has joined #openstack-infra03:23
*** asettle is now known as alex-lunch03:29
*** matsuhashi has quit IRC03:31
*** nati_ueno has quit IRC03:35
*** CaptTofu has joined #openstack-infra03:36
*** pcrews__ has quit IRC03:40
*** CaptTofu has quit IRC03:40
*** aconrad has joined #openstack-infra03:47
*** nosnos has quit IRC03:49
openstackgerritJoshua Hesketh proposed a change to openstack-infra/zuul: Allow zuul-server to run a merger process
*** nati_ueno has joined #openstack-infra03:50
*** talluri has joined #openstack-infra04:03
*** nati_ueno has quit IRC04:05
*** zhiyan is now known as zhiyan_04:07
*** alex-lunch is now known as asettle04:13
*** thuc has joined #openstack-infra04:19
*** msabramo has joined #openstack-infra04:20
*** thuc_ has joined #openstack-infra04:20
*** aconrad has quit IRC04:24
*** thuc has quit IRC04:24
*** e0ne_ has quit IRC04:27
*** Ryan_Lane has joined #openstack-infra04:30
*** matsuhashi has joined #openstack-infra04:38
*** nosnos has joined #openstack-infra04:48
*** thuc_ has quit IRC04:51
*** chandan_kumar has joined #openstack-infra04:52
*** thuc has joined #openstack-infra04:52
*** thuc has quit IRC04:56
*** ildikov has quit IRC04:58
*** medieval1 has quit IRC05:11
*** skraynev_afk is now known as skraynev05:12
uvirtbotLaunchpad bug 1306967 in openstack-ci "thstack-ci service account" [Undecided,New]05:19
*** CaptTofu has joined #openstack-infra05:24
*** CaptTofu has quit IRC05:29
*** unicell has quit IRC05:31
*** om has joined #openstack-infra05:35
*** nosnos_ has joined #openstack-infra05:36
*** nosnos has quit IRC05:36
*** unicell has joined #openstack-infra05:40
*** nati_ueno has joined #openstack-infra05:42
*** zhiyan_ is now known as zhiyan05:43
*** zul has quit IRC05:46
zarorcarrillocruz: gerrit upgrade is scheduled for may 1st-7th.05:55
*** nati_ueno has quit IRC05:56
*** ildikov has joined #openstack-infra05:58
*** matsuhas_ has joined #openstack-infra06:00
*** zul has joined #openstack-infra06:00
*** thuc has joined #openstack-infra06:03
*** matsuhashi has quit IRC06:03
*** thuc has quit IRC06:08
rcarrillocruzk, thx zaro06:08
*** afazekas has joined #openstack-infra06:12
*** unicell has quit IRC06:12
*** skolekonov has joined #openstack-infra06:13
*** unicell has joined #openstack-infra06:22
*** Ryan_Lane has quit IRC06:23
*** amotoki has joined #openstack-infra06:24
*** nosnos has joined #openstack-infra06:32
*** nosnos_ has quit IRC06:32
*** _nadya_ has joined #openstack-infra06:34
*** msabramo has quit IRC06:36
*** matsuhas_ has quit IRC06:37
*** matsuhashi has joined #openstack-infra06:37
*** matsuhashi has quit IRC06:37
*** matsuhashi has joined #openstack-infra06:38
*** sdake_ has joined #openstack-infra06:41
*** sdake_ has quit IRC06:41
*** sdake_ has joined #openstack-infra06:41
*** e0ne has joined #openstack-infra06:42
*** _sweston_ has joined #openstack-infra06:45
*** _nadya_ has quit IRC06:46
*** _sweston_ has quit IRC06:49
*** sweston has joined #openstack-infra06:49
*** mrmartin has joined #openstack-infra07:06
*** milki has quit IRC07:08
*** andreaf has joined #openstack-infra07:08
*** e0ne_ has joined #openstack-infra07:11
*** CaptTofu has joined #openstack-infra07:12
*** e0ne has quit IRC07:14
*** jgallard has joined #openstack-infra07:17
*** CaptTofu has quit IRC07:17
*** flaper87|afk is now known as flaper8707:17
*** dizquierdo has joined #openstack-infra07:17
*** reed has joined #openstack-infra07:17
*** yolanda_ has joined #openstack-infra07:17
*** jamielennox is now known as jamielennox|away07:18
*** jcoufal has joined #openstack-infra07:22
*** e0ne_ has quit IRC07:32
*** e0ne has joined #openstack-infra07:33
*** asettle has quit IRC07:35
*** rallegre has joined #openstack-infra07:37
*** e0ne has quit IRC07:37
*** talluri has quit IRC07:42
*** e0ne has joined #openstack-infra07:43
*** matrohon has joined #openstack-infra07:44
*** rallegre has quit IRC07:46
*** ihrachyshka has joined #openstack-infra07:53
*** morganfainberg is now known as morganfainberg_Z08:00
*** ildikov_ has joined #openstack-infra08:02
*** rallegre has joined #openstack-infra08:02
*** praneshp has quit IRC08:04
*** yassine has joined #openstack-infra08:04
*** ihrachyshka has quit IRC08:05
*** ildikov has quit IRC08:05
*** ihrachyshka has joined #openstack-infra08:05
*** simonluo has joined #openstack-infra08:09
*** jpich has joined #openstack-infra08:10
*** mrda is now known as mrda_away08:14
rcarrillocruzzaro: the gerrit upgrade raises a question on two bugs i was working on08:15
rcarrillocruzone of them there's already a change for it08:15
*** alexpilotti has joined #openstack-infra08:16
rcarrillocruzshould I rewrite the code to make it work with  the current gerrit and with gerrit 2.8?08:16
rcarrillocruzor should I dump the code and make it work with just 2.8 since it will be released in two weeks ( and await for testing/reviews then)08:17
rcarrillocruzin Gerrit 2.8, the change-abandoned hook passes the --commit argument, whereas on current gerrit does not, so I had to write a method to pull that info from the gerritdb08:19
*** talluri has joined #openstack-infra08:24
*** ildikov has joined #openstack-infra08:25
*** dizquierdo has quit IRC08:25
*** jooools has joined #openstack-infra08:28
*** ildikov_ has quit IRC08:28
*** achuprin_ has joined #openstack-infra08:29
*** achuprin has quit IRC08:31
*** ominakov has joined #openstack-infra08:31
*** talluri_ has joined #openstack-infra08:33
*** talluri has quit IRC08:35
*** che-arne has joined #openstack-infra08:39
rcarrillocruzfrom the refs/changes/ git log rather than DB, sorry08:41
*** rallegre has quit IRC08:56
*** rallegre has joined #openstack-infra08:58
*** jlibosva has joined #openstack-infra08:58
*** CaptTofu has joined #openstack-infra09:00
*** pblaho has joined #openstack-infra09:03
*** CaptTofu has quit IRC09:05
*** talluri_ has quit IRC09:06
*** fbo_away is now known as fbo09:18
*** katyafervent_awa is now known as katyafervent09:20
*** derekh has joined #openstack-infra09:21
*** jp_at_hp has joined #openstack-infra09:25
*** zehicle has quit IRC09:28
*** zhiyan is now known as zhiyan_09:30
*** jp_at_hp has quit IRC09:34
*** rallegre has quit IRC09:35
*** StevenK has quit IRC09:44
*** JoshNang has quit IRC09:46
*** soren has quit IRC09:46
*** spredzy has quit IRC09:46
*** soren has joined #openstack-infra09:47
*** spredzy has joined #openstack-infra09:47
*** JoshNang has joined #openstack-infra09:48
*** yassine has quit IRC09:48
*** jp_at_hp has joined #openstack-infra09:52
*** dizquierdo has joined #openstack-infra09:56
*** rallegre has joined #openstack-infra09:59
*** rallegre has quit IRC10:03
*** rallegre has joined #openstack-infra10:08
*** fbo is now known as fbo_away10:10
*** sahumada has quit IRC10:12
*** rallegre has quit IRC10:12
*** sahumada has joined #openstack-infra10:12
*** jgallard has quit IRC10:16
*** ianw has quit IRC10:17
*** rallegre has joined #openstack-infra10:35
*** rallegre has quit IRC10:36
*** ihrachyshka has quit IRC10:36
*** matsuhashi has quit IRC10:40
*** matsuhashi has joined #openstack-infra10:41
*** CaptTofu has joined #openstack-infra10:46
*** david-lyle has joined #openstack-infra10:46
*** simonluo has quit IRC10:49
*** matsuhashi has quit IRC10:50
*** matsuhas_ has joined #openstack-infra10:52
*** coolsvap is now known as coolsvap|afk10:52
*** ihrachyshka has joined #openstack-infra10:53
*** Longgeek has quit IRC10:54
*** Longgeek has joined #openstack-infra10:54
openstackgerritSean Dague proposed a change to openstack-infra/devstack-gate: normalize the services for a post grizzly world
*** chandan_kumar has quit IRC10:55
*** rallegre has joined #openstack-infra10:57
*** ihrachyshka has quit IRC10:57
*** rallegre has quit IRC10:58
*** Longgeek has quit IRC10:58
*** rallegre has joined #openstack-infra10:59
*** matsuhas_ has quit IRC11:00
*** yassine has joined #openstack-infra11:01
*** Longgeek has joined #openstack-infra11:01
*** rallegre has quit IRC11:03
*** matsuhas_ has joined #openstack-infra11:04
*** rallegre has joined #openstack-infra11:08
*** chandan_kumar has joined #openstack-infra11:08
*** slagle has quit IRC11:09
*** rallegre has quit IRC11:11
*** slagle has joined #openstack-infra11:11
*** CaptTofu has quit IRC11:15
*** matsuhas_ has quit IRC11:15
*** CaptTofu has joined #openstack-infra11:15
*** lcostantino has joined #openstack-infra11:15
*** chandan_kumar has quit IRC11:16
*** matsuhashi has joined #openstack-infra11:16
*** CaptTofu has quit IRC11:19
*** mbacchi has joined #openstack-infra11:21
*** Longgeek has quit IRC11:26
*** fbo_away is now known as fbo11:26
*** Shrews has joined #openstack-infra11:27
*** amotoki has quit IRC11:28
*** jhesketh has quit IRC11:29
*** weshay has joined #openstack-infra11:29
*** chandan_kumar has joined #openstack-infra11:30
*** weshay has quit IRC11:32
*** e0ne has quit IRC11:36
*** e0ne has joined #openstack-infra11:37
*** sandywalsh has joined #openstack-infra11:38
*** e0ne has quit IRC11:41
*** david-lyle has quit IRC11:41
*** yaguang has quit IRC11:44
*** matsuhashi has quit IRC11:46
*** StevenK has joined #openstack-infra11:48
sdaguefungi: it seems that nodepool in hp region-b still isn't getting the fedora x86_64 image. Any ideas there?11:49
*** om has quit IRC11:50
*** om has joined #openstack-infra11:51
*** matsuhashi has joined #openstack-infra11:51
*** msabramo has joined #openstack-infra11:52
*** CaptTofu has joined #openstack-infra11:55
*** nosnos has quit IRC12:01
*** enikanorov has quit IRC12:04
*** enikanorov has joined #openstack-infra12:04
*** e0ne has joined #openstack-infra12:07
*** e0ne_ has joined #openstack-infra12:09
*** e0ne has quit IRC12:09
*** Longgeek has joined #openstack-infra12:13
*** dkranz has joined #openstack-infra12:15
*** rfolco has joined #openstack-infra12:16
*** pdmars has joined #openstack-infra12:17
*** aysyd has joined #openstack-infra12:20
*** rallegre has joined #openstack-infra12:26
YorikSarjamespage: ping12:27
YorikSarjamespage: I've finally reproduced that bug in Python 2.7.3 we talked about on Friday12:27
*** e0ne_ has quit IRC12:27
*** e0ne has joined #openstack-infra12:27
YorikSarjamespage: bug #130747912:27
uvirtbotLaunchpad bug 1307479 in python2.7 "multiprocessing.managers doesn't work due to Python stdlib bug" [Undecided,New]
*** zhiyan_ is now known as zhiyan12:27
YorikSar(or was it Thursday?..)12:28
jamespageYorikSar, do you have a nice test case for this? something say that someone who does not use openstack can reproduce?12:29
jamespageit will make the SRU process alot easier - I can ping doko to get his attention on this (python maintainer)12:29
*** matsuhashi has quit IRC12:29
YorikSarjamespage: You can take devstack and run it with the lates patchset in
*** matsuhashi has joined #openstack-infra12:30
YorikSarjamespage: It's not very nice... I can try to write some simple scripte demonstrating the problem though.12:30
jamespageYorikSar, that would be awesome12:30
jamespageYorikSar, what's the criticality of this bug? its it blocking juno work atm?12:31
*** om has quit IRC12:31
YorikSarjamespage: Oh, wait. There's example in bug description in Python's bugtracker.12:31
YorikSarjamespage: Basically, yes. We want to push rootwrap daemon work as early in Juno as possible and this bug won't let tests pass.12:32
YorikSarAnd my workarounds somehow won't work :(12:32
jamespageYorikSar, ok - understand - but might not be this week - as 14.04 release is thurs as well12:33
*** weshay has joined #openstack-infra12:33
*** BobBall_ has quit IRC12:33
*** adalbas has joined #openstack-infra12:33
*** matsuhashi has quit IRC12:34
YorikSarjamespage: I wonder if our gate will switch to 14.04 as well...12:34
YorikSar"as well" => "after that"12:35
*** zhiyan is now known as zhiyan_12:36
*** zhiyan_ is now known as zhiyan12:36
YorikSarjamespage: I've updated bug description to point to the example.12:37
YorikSarjamespage: Can I do anything to help the process besides (stopping) poking you about it?12:38
*** matsuhashi has joined #openstack-infra12:42
*** ildikov_ has joined #openstack-infra12:44
*** om has joined #openstack-infra12:45
*** lascii is now known as alaski12:46
*** ociuhandu has quit IRC12:46
*** ildikov has quit IRC12:46
*** zul has quit IRC12:48
*** zul has joined #openstack-infra12:51
*** freyes has joined #openstack-infra12:54
*** skolekonov has quit IRC12:54
*** mfer has joined #openstack-infra12:55
*** freyes has quit IRC12:58
*** thuc_ has joined #openstack-infra12:59
*** fifieldt has quit IRC12:59
*** dprince has joined #openstack-infra13:05
*** bknudson has quit IRC13:06
*** matsuhashi has quit IRC13:06
*** blamar has joined #openstack-infra13:07
*** thomasem has joined #openstack-infra13:07
*** jeckersb_gone is now known as jeckersb13:08
*** matsuhas_ has joined #openstack-infra13:10
*** jpeeler has joined #openstack-infra13:12
*** jpeeler has quit IRC13:12
*** jpeeler has joined #openstack-infra13:12
*** mriedem has joined #openstack-infra13:18
*** HenryG has quit IRC13:21
*** om has quit IRC13:28
*** Longgeek has quit IRC13:28
*** om has joined #openstack-infra13:29
*** jgrimm has quit IRC13:30
*** Longgeek has joined #openstack-infra13:31
*** rallegre has quit IRC13:33
*** msabramo has quit IRC13:34
*** rallegre has joined #openstack-infra13:35
*** ociuhandu has joined #openstack-infra13:38
*** dkliban_afk is now known as dkliban13:38
*** yfried__ has quit IRC13:39
*** mwagner_lap has quit IRC13:39
slaglehi, i don't *think* this change has been applied to gerrit yet,
slagleis there a way to check?13:41
slaglei don't see the acl overrides when i look at,access13:42
*** matsuhas_ has quit IRC13:42
*** salv-orlando has joined #openstack-infra13:43
*** basha_ has joined #openstack-infra13:49
*** basha_ has left #openstack-infra13:49
*** rlandy has joined #openstack-infra13:49
*** prad has joined #openstack-infra13:51
*** freyes has joined #openstack-infra13:51
*** matsuhashi has joined #openstack-infra13:52
*** yfried__ has joined #openstack-infra13:54
*** matsuhashi has quit IRC13:54
*** skraynev is now known as skraynev_afk13:55
*** eharney has joined #openstack-infra13:57
*** enikanorov_ has joined #openstack-infra13:57
*** mestery_ has joined #openstack-infra13:58
*** basha_ has joined #openstack-infra13:58
*** mestery has quit IRC13:58
*** wchrisj has joined #openstack-infra13:58
*** enikanorov has quit IRC13:59
*** thuc has joined #openstack-infra13:59
*** zns has joined #openstack-infra14:00
*** zns has quit IRC14:00
*** signed8bit has joined #openstack-infra14:00
*** thuc__ has joined #openstack-infra14:00
*** zns has joined #openstack-infra14:00
SergeyLukjanovslagle, hey14:01
SergeyLukjanovLabel Approved Exclusive14:01
SergeyLukjanov   diskimage-builder-core14:01
SergeyLukjanovLabel Code-Review Exclusive14:01
SergeyLukjanov   diskimage-builder-core14:01
SergeyLukjanovslagle, everything is ok now14:01
SergeyLukjanovslagle, oops, you're talking about stable/*14:02
*** thuc_ has quit IRC14:02
slagleSergeyLukjanov: yes :)14:02
slaglei heard the gerrit acl configs are applied manually on fridays14:03
slaglei don't know if that happened last week or not14:03
*** thuc has quit IRC14:04
*** basha_ has quit IRC14:05
*** jgrimm has joined #openstack-infra14:06
*** basha_ has joined #openstack-infra14:06
*** basha_ has left #openstack-infra14:06
SergeyLukjanovslagle, looks like changes aren't applied, so, you need to ask someone from infra root to check14:07
slagleSergeyLukjanov: ok, thanks14:08
*** rwsu has joined #openstack-infra14:09
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: added query for neutron migrations issue
*** dkehn has quit IRC14:13
*** dkehnx has quit IRC14:14
*** om has quit IRC14:17
*** beekneemech is now known as bnemec14:18
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: added query for neutron migrations issue
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: added query for neutron migrations issue
openstackgerritA change was merged to openstack-infra/elastic-recheck: added query for neutron migrations issue
*** wchrisj has quit IRC14:31
*** zns has quit IRC14:31
*** wchrisj has joined #openstack-infra14:32
*** mestery_ is now known as mestery14:32
*** zns has joined #openstack-infra14:33
*** dims has quit IRC14:35
sdaguesalv-orlando: I made the bug critical as well, because I think right now it's probably the most pressing issue to address in neutron, as it has a lot of systemic issues14:37
*** MarkAtwood has joined #openstack-infra14:37
salv-orlandosdague: it's ok to say it critical… but I'd say that probably this is more than a simple bug. Anyway, it's good to track it somehow.14:38
sdaguesalv-orlando: agreed, as I think the unwind is going to be more extensive14:38
sdagueI think there will need to be a blueprint to address the bug, but the bug is at least useful for tracking (and things like elastic recheck need a bug)14:39
*** rallegre has quit IRC14:39
*** Longgeek has quit IRC14:40
*** ameade has joined #openstack-infra14:40
*** thuc has joined #openstack-infra14:40
*** pcrews__ has joined #openstack-infra14:40
*** bknudson has joined #openstack-infra14:41
*** dims has joined #openstack-infra14:42
*** thuc__ has quit IRC14:43
*** thuc has quit IRC14:44
*** rallegre has joined #openstack-infra14:47
*** amotoki has joined #openstack-infra14:47
*** HenryG has joined #openstack-infra14:48
*** wenlock has joined #openstack-infra14:53
*** zz_gondoi is now known as gondoi14:54
*** zhiyan is now known as zhiyan_14:54
anteayasdague: you were looking for fungi earlier14:54
anteayahe is away for two weeks14:54
anteayahe is getting married this weekend14:55
sdagueoh, nice14:55
anteayayeah, sorry you didn't know14:55
anteayahe doesn't make much fuss14:55
*** miqui has quit IRC14:55
*** miqui has joined #openstack-infra14:55
sdagueok, I'll leave that as post release week issue14:56
Guest38269sdague: good morning14:57
Guest38269oh no!14:57
*** Guest38269 is now known as jeblair14:57
jeblairsdague, anteaya: good morning! :)14:57
sdaguemorning :)14:57
*** jeblair is now known as Guest3464214:57
Guest34642i'm back from pycon; sleep deprived but hopefully should be able to start catching up14:57
Guest34642okay gimme a sec14:58
*** mwagner_lap has joined #openstack-infra14:58
sdagueman, your irc client doesn't like you this morning :)14:58
*** gondoi is now known as zz_gondoi14:58
Guest34642it gets grumpy when i ignore it for too long :)14:58
*** Guest34642 is now known as corvus14:58
*** thedodd has joined #openstack-infra14:59
*** thuc has joined #openstack-infra15:00
anteayacorvus: morning15:00
corvusstoryboard meeting time in #openstack-meeting-315:00
openstackgerritRandy Coulman proposed a change to openstack-infra/jenkins-job-builder: Support the Store SCM plugin
*** zz_gondoi is now known as gondoi15:03
*** doddstack has joined #openstack-infra15:06
phschwartzMorning infra15:07
*** thedodd has quit IRC15:08
*** basha_ has joined #openstack-infra15:10
*** chandan_kumar has quit IRC15:11
*** aconrad has joined #openstack-infra15:12
*** basha_ has quit IRC15:12
*** jpich has quit IRC15:12
*** dims has quit IRC15:13
SergeyLukjanovphschwartz, morning15:14
openstackgerritRandy Coulman proposed a change to openstack-infra/jenkins-job-builder: Support the Store SCM plugin
krotscheckHey, clarkb: ElasticSearch for storyboard: Good idea/bad idea?15:14
*** nijaba has quit IRC15:14
*** nijaba has joined #openstack-infra15:15
*** nijaba has joined #openstack-infra15:15
*** basha_ has joined #openstack-infra15:18
*** jpich has joined #openstack-infra15:19
openstackgerritDavid Caro proposed a change to openstack-infra/jenkins-job-builder: Added parallelization options
*** om has joined #openstack-infra15:21
*** basha_ has quit IRC15:21
*** kmartin has joined #openstack-infra15:24
*** zul has quit IRC15:24
*** coolsvap|afk is now known as coolsvap15:25
*** zul has joined #openstack-infra15:25
*** branen has joined #openstack-infra15:27
*** dkehn has joined #openstack-infra15:29
*** dizquierdo has quit IRC15:31
*** chuck_ has joined #openstack-infra15:31
*** zul has quit IRC15:31
*** chuck_ is now known as zul15:32
*** gyee has joined #openstack-infra15:32
*** matjazp has joined #openstack-infra15:33
*** dkehn_ has joined #openstack-infra15:35
*** dkehn is now known as dkehnx15:35
*** dkehn_ is now known as dkehn15:35
openstackgerritA change was merged to openstack-infra/reviewday: Quit smoking.
*** dkehn has quit IRC15:36
*** atiwari has joined #openstack-infra15:36
*** dkehn has joined #openstack-infra15:36
*** melwitt has joined #openstack-infra15:37
openstackgerritDirk Mueller proposed a change to openstack-dev/hacking: Open 0.9.0 development
*** dims has joined #openstack-infra15:39
*** BadCub01 has joined #openstack-infra15:39
*** pblaho has quit IRC15:39
*** thuc has quit IRC15:39
*** thuc has joined #openstack-infra15:39
*** gokrokve has joined #openstack-infra15:40
*** freyes has quit IRC15:43
*** thuc has quit IRC15:44
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Add cross-project unit tests for oslo libraries
*** enikanorov_ has quit IRC15:45
*** basha_ has joined #openstack-infra15:46
zarorcarrillocruz: could you please provide a pointer to the changes?15:46
dhellmannSergeyLukjanov: regarding the non-voting jobs in ^^, should I split those out? I want to make them voting very soon, but I don't want to break everything when that commit lands15:46
zaroaconrad: hi15:47
SergeyLukjanovdhellmann, hey, looking15:47
aconradhi zaro15:47
dhellmannSergeyLukjanov: and as far as the number of jobs, I can't do much about that. jeblair, fungi, and I talked through some alternatives a while back, but all of them fail to meet some of our requirements15:48
*** asselin has joined #openstack-infra15:49
SergeyLukjanovdhellmann, yup, re number of jobs the only idea is complex "oslo" job to test all oslo libs in one job15:50
dhellmannSergeyLukjanov: that was one idea, but it introduces gate blocking for projects on libraries they don't use15:50
SergeyLukjanovdhellmann, re non-voting gate jobs, I'm ok with adding them if you're going to make them voting soon15:51
dhellmannSergeyLukjanov: it also doesn't support the libraries gating on changes in the apps15:51
SergeyLukjanovdhellmann, yup, it sounds not very good15:51
dhellmannSergeyLukjanov: yeah, I would love to have this all sorted out before the summit15:51
SergeyLukjanovdhellmann, ok, I'll take a look one more time and +2 if everything is ok15:52
*** dims has quit IRC15:52
dhellmannSergeyLukjanov: I'm open to suggestions on alternative implementations; I just haven't found one that works. :-/15:53
dhellmannSergeyLukjanov: cool, thanks!15:53
*** gothicmindfood has quit IRC15:53
SergeyLukjanovdhellmann, yup, I don't see better solutions for it atm15:54
dhellmannSergeyLukjanov: does jjb support passing a list of values to a job template (and the template receiving them as a list)?15:55
dhellmannI could define 2 jobs, one for checking projects on libs and one for checking libs on projects, and pass the right list of inputs to it15:55
*** gondoi is now known as zz_gondoi15:55
SergeyLukjanovdhellmann, I don't know jjb enough well to answer this question15:55
*** mrmartin has quit IRC15:55
dhellmannI would hate to learn that the unit tests for one app break the tests for another, though, so I'm not sure I want to combine those into a single job15:56
*** gothicmindfood has joined #openstack-infra15:57
dhellmannanyway, I'm completely happy to revisit and reimplement this if we come up with something better, but I'd like to at least start with this version for now until the better idea comes along :-)15:57
*** e0ne has quit IRC15:59
*** e0ne has joined #openstack-infra15:59
*** sabari2 has joined #openstack-infra16:00
*** jcoufal has quit IRC16:01
*** coolsvap1 has joined #openstack-infra16:02
*** zul has quit IRC16:02
*** e0ne has quit IRC16:03
*** coolsvap1 has quit IRC16:03
*** zul has joined #openstack-infra16:04
*** coolsvap1 has joined #openstack-infra16:04
*** zul has quit IRC16:04
*** dims has joined #openstack-infra16:05
*** jergerber has joined #openstack-infra16:07
*** coolsvap1 has quit IRC16:08
*** zul has joined #openstack-infra16:08
sdaguekrotscheck: what's the problem that elastic search is trying solve in storyboard, maybe we can start there and move backwards16:09
*** thuc has joined #openstack-infra16:09
sdaguebecause ES is a lot of infrastructure, and a storyboard data set seems small in comparison16:09
corvussdague: full text searhc16:10
*** thuc has quit IRC16:10
corvusor something like that, possibly with letters in the right order.  i don't know if we're picky though.  :)16:10
krotscheckcorvus sdague: Which, to be fair, we probably won’t get to for another week or three - we’ve got a lot of things we can do with a pure filter approach.16:11
sdagueI locally managed lucene would be a ton less effort and weight I'd think16:11
krotschecksdague: I thought that ES was built on Lucene?16:11
*** thuc has joined #openstack-infra16:11
sdagueit is16:11
krotschecksdague: Another alternative is Sphinx16:11
krotschecksdague: Or MySQL fulltext indexes.16:11
corvussdague: we _definitely_ don't want to have to run a logstash-sized cluster for openstack's storyboard; so yeah, ideally i think a storyboard instance of our size should not be outweighed by its full-text indexer16:12
corvusi don't know what a 'small' es installation looks like, so that's part of the open question for me16:12
sdagueright, I think ES only really makes sense if you need the E part16:12
*** dims has quit IRC16:13
sdagueI think for small data sets - might actually make more sense. Then it could be fully managed by the storyboard server. Probably easier to deploy and run.16:14
corvussdague: or run solr and use the api?16:15
sdagueyou just add weight to run another service like that16:16
krotscheck…..I get a strange satisfaction by seeing that pylucene is actually java :)16:16
*** jaypipes has joined #openstack-infra16:16
sdaguewell everything that talked to lucene at function level needs java somewhere :)16:16
*** rallegre has quit IRC16:17
*** rallegre has joined #openstack-infra16:18
sdaguehonest at this stage full text search in mysql is probably totally good enough16:18
sdagueas long at the plug point is clean enough to plug in other search in the future16:19
sdaguethat's basically how mediawiki and drupal function16:19
*** zz_gondoi is now known as gondoi16:19
*** moted has quit IRC16:19
*** moted has joined #openstack-infra16:20
*** jpich has quit IRC16:24
*** dims has joined #openstack-infra16:27
*** mrmartin has joined #openstack-infra16:28
*** basha_ has quit IRC16:28
ruhefolks, lucene is too low-level16:29
corvussdague: have a link to your zuul depends-on etherpad?  i thought of a concern we need to keep in mind16:30
*** ominakov has quit IRC16:30
*** thuc has quit IRC16:30
*** thuc has joined #openstack-infra16:31
corvusheh i never would have guessed that.  :)16:32
*** chandan_kumar has joined #openstack-infra16:33
*** thuc_ has joined #openstack-infra16:33
*** thuc__ has joined #openstack-infra16:34
*** thuc has quit IRC16:35
*** thuc has joined #openstack-infra16:36
*** thuc has quit IRC16:36
*** mrmartin has quit IRC16:36
*** thuc has joined #openstack-infra16:37
*** thuc_ has quit IRC16:38
*** om has quit IRC16:38
*** thuc__ has quit IRC16:39
mordredsdague: there is no full text search in mysql16:40
mordredthat is usable16:40
*** yassine has quit IRC16:40
mordredit's only there on myisam tables16:40
mordredand I employ every ounce of my ability to veto things to veto any use of myisam16:40
mordredlucene sounds great16:41
mordredkrotscheck to me16:41
mordredto me16:41
mordredkrotscheck: ^^16:41
mordredis what I meant to type16:41
mordreddoesn't new gerrit also use lucene for searching?16:41
ruheif you compare Lucene with ES or Solr, it'll be like a socket compared to REST API over HTTP. It is a very low-level thing16:43
corvussdague: okay, added some food for thought at the bottom16:44
krotscheckmordred: Are you sure it’s only available on MyISAM?16:44
sdagueruhe: yeh, low level is fine in this case I think16:44
mordredkrotscheck: oh! wow. I don't know anything about mysql anymore16:45
* mordred retracts statements16:45
mordredwe're using mysql via cloud servers16:45
krotscheckmordred: Yeah, no idea what version that dropped on.16:46
ruhesdague: i don't think so. with lucene you need to write additional code, with ES you just send your data to it to persist. then send http queries for data16:46
corvusmordred: it's that oracley goodness16:46
mordredand it's running 5.1.7316:46
ruhesdague: and the amount of additional code overweights ES maintenance сosts16:46
krotscheckmordred: Looks like it’s new in 5.616:47
openstackgerritSean Dague proposed a change to openstack-infra/config: make postgresql job not use config drive
sdagueruhe: running an ES cluster is non trivial16:48
*** mrmartin has joined #openstack-infra16:48
ruhesdague: yes it is. but figuring out how to adopt lucene for a search engine is also non trivial. and from my experience i say that running an ES cluster is easier compared to writing your own code around Lucene :)16:50
*** rallegre has quit IRC16:50
*** che-arne has quit IRC16:50
ruhei didn't maintain ES cluster, but i've used both Lucene and Solr (which is very similar to ES)16:50
sdagueyeh, I've used lucene before, it's not that bad :)16:51
sdaguemordred: honestly, iteration 1 with LIKE strings is probably fine for the size data set that storyboard has16:52
corvusmaybe solr is easier? :)16:52
ruheand *maybe* it is OK to run ES on a single node16:52
sdaguemaybe, managing a lucene indexer for a small set of data isn't that bad. That being said, is "really good search" really a near term need?16:53
corvussdague: yeah, i think we're on the same page about the priority here; it's not at the top of the list; i think it's just something we want to start thinking about because we will probably need it for the 'all the openstack bugs' stage16:53
corvusopenstack has a lot of bugs ;)16:53
sdaguesure, though some of the other categorization needs probably happen first. Like being able to see artifacts created by you, or assigned to you16:53
*** harlowja_away is now known as harlowja16:54
anteayacan someone with ssh access to the git farm run `ps aux | grep salt` on any git server and paste the output, please?16:54
sdaguethat actually raise a meta question, should I report bugs in storyboard as stories?16:54
corvussdague: yeah; check out the storyboard meeting log for the full context.  that's where we're spending most of our time.  just trying to back-burner this one so we don't completely leave it out of the design16:54
corvussdague: absolutely16:54
sdagueok, will do16:54
corvussdague: also future-work can go there16:55
*** afazekas has quit IRC16:55
*** SumitNaiksatam has quit IRC16:56
*** fbo is now known as fbo_away16:56
mordredanteaya: [mordred@git01 ~]$ ps aux | grep salt16:56
mordredmordred  24100  0.0  0.0 103244   864 pts/0    S+   16:56   0:00 grep salt16:56
anteayamordred: thanks16:57
anteayathis confirms corvus' belief that salt-minions are not currently running on git servers16:57
openstackgerritA change was merged to openstack-infra/devstack-gate: normalize the services for a post grizzly world
anteayaso ssh it is16:57
mordredanteaya: I have a patch for you to check out16:58
*** rallegre has joined #openstack-infra16:58
anteayamordred: looking at it now16:59
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Drive puppet from the master over ssh
mordredanteaya: I have an update that's betterer16:59
anteayaah new hotness16:59
* anteaya moves to the new hotness16:59
mordrednibalizer: ^^ you may also want to look at that, you had some comments on the puppet kick version17:00
*** jdurgin has quit IRC17:00
mordredcorvus: btw - I LOVE the idea of making this a non-slave zuul worker17:00
mordredcorvus: it makes several places where this one is designed like poo much better17:00
*** jdurgin has joined #openstack-infra17:01
corvusmordred: ++17:01
mordredcorvus: but I think that can potentially be a step 217:01
*** derekh has quit IRC17:03
*** nati_ueno has joined #openstack-infra17:04
anteayamordred: does this patch do anything with the return code? so far I see it returned but that is it17:04
mordredanteaya: no, it does not. it should17:05
mordredthere are two more changes I think I want to make in this17:06
nibalizermordred: so I still don't think you'll be able to run every machine in 15 minutes...17:07
mordredcorvus: you know - although I Do not think that ansible playbooks will help with this task at all and would introduce needless complexity - the "run parallel invocations" feature might be useful to address nibalizer's concern ^^17:07
* mordred goes to make a quick poc patch to show a thought there17:07
*** dkliban is now known as dkliban_afk17:07
*** dkliban_afk is now known as dkliban17:07
nibalizermordred: i think an xargs -P 10 -n 1 is the correct solution17:08
* anteaya pulls up mordred's paste from a week ago:
mordrednibalizer: ooh. I hadn't thought of that17:08
nibalizerthat way you run 10 at a time(so you dont kill the master) but they all get done asap17:08
nibalizerand of course '10' is a number to adjust17:08
mordrednibalizer: yup. love it17:09
nibalizeri mean i love devops tools, dont get me wrong17:09
nibalizerbut sometimes its time to use the shell17:09
mordrednibalizer: can I xargs a shell function?17:09
nibalizerheck yea17:09
chmouelclarkb: hey clark i was wondering what was the status tox.ConfigError: ConfigError: substitution key 'posargs' not found17:09
corvuschmouel: i talked to holger at pycon; he's going to look more closely at clarkb's merge request17:10
corvuschmouel: but the problem is still there in the latest tox, so we pin to an older vesion17:10
chmouelcorvus: ah nice, (i am at pycon at the moment as well)17:10
corvuschmouel: i'm back at home now (flew back yesterday)17:10
*** amcrn has joined #openstack-infra17:11
corvuschmouel: (i'm jeblair, in case you don't know my alternick)17:11
chmouelcorvus: ah yeah i didn't know your altenick :)17:11
chmouelcorvus: so yeah i def saw you ;)17:11
*** corvus is now known as jeblair17:11
jeblairthere we go; freenode likes me now17:12
*** SumitNaiksatam has joined #openstack-infra17:12
mordrednibalizer: how do I remove OVERRIDE_LIST from FULL_LIST?17:12
mordredhang on - I got it17:12
*** mrodden has quit IRC17:12
*** Ryan_Lane has joined #openstack-infra17:13
*** jooools has quit IRC17:13
sdaguejeblair / mordred: - turns out we broke our expected testing config back in jan. We should get that through soonish to make sure we don't regress17:13
*** yolanda_ has quit IRC17:13
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Drive puppet from the master over ssh
mordrednibalizer: how about that ^^17:16
*** gokrokve_ has joined #openstack-infra17:16
*** thuc has quit IRC17:17
zaroaconrad: what app are you building on top of jjb?17:17
*** thuc has joined #openstack-infra17:18
aconradzaro: a command line interface that prompts the user for questions and generates a YAML file out of it.17:19
nibalizermordred: i like it17:19
nibalizeran idea: do you think it would be advantageous to use ssh -o Connectiontimeout=something versus the way you're doing it with shell timeout?17:19
aconradI'm trying to get my company on Jenkins, but Jenkins is messy and I'm trying to make it easier for our developers to hook up their github project onto Jenkins17:19
*** gokrokve has quit IRC17:19
zaroaconrad: a wizard to build jjb jobs? is it on github?17:20
aconradno, not yet17:20
aconradI'm trying to think how I can open source it17:20
mordredaconrad: patches to jjb?17:20
aconradzaro: because the questions asked are somewhat specific to my company, so I'm not sure how to make it very generic17:20
mordredoh. in that case - me shuts up17:21
nibalizermordred: looks good!17:21
mordrednibalizer: awesome17:21
nibalizeri left a comment on the authorized key thing17:21
mordrednibalizer: re: ssh connectiontimeout - if puppet hangs and ssh just timeouts the connection, could se leave a stale hung puppet on the box?17:21
aconradzaro: specific, but also quite generic. Like "do you want to build pull requests?", "do you want to prevent code coverage regression?", "what is the command to run your tests", ...17:22
aconradI'd say the flow is specific.17:22
zaroaconrad: what is a 'file-like' object?17:22
*** MarkAtwood has quit IRC17:22
*** thuc has quit IRC17:22
aconradzaro: ... or anything that quacks like a file (duck-typing)17:22
*** MarkAtwood has joined #openstack-infra17:23
aconradzaro: basically, in our case, an object that has a .read() method.17:23
aconradzaro: are you at PyCon?17:23
anteayaaconrad: no, he isn't17:24
*** UtahDave has joined #openstack-infra17:24
zaroaconrad: missed it this time. i was at confoo.17:24
nibalizermordred: im not sure... googling isn't telling me the answer17:24
aconradzaro: ah right17:24
nibalizerbut I do know that the timeout command you're using sends a TERM by default, so maybe we'd want to upgrade it to timeout -s 9 puppet agent --test17:24
*** mrodden has joined #openstack-infra17:24
aconradI met with James B., very nice guy17:25
mordrednod. also, is 60m too long?17:25
*** thuc has joined #openstack-infra17:25
*** thuc has quit IRC17:26
mgagnezaro: they wish to build some kind of wizard abstracting Jenkins and JJB for a bunch of developers who never heard about Jenkins or CI. The tool asks a bunch of predefined questions and build the appropriate JJB configs based on those answers. My understanding is this tool will integrate with the JJB "internal api" instead of relying on the CLI.17:26
*** denis_makogon_ has joined #openstack-infra17:26
*** thuc has joined #openstack-infra17:26
nibalizeralso might be considered useful to run puppet with --logdest=syslog because I think when we use 'puppet agent --test' we have to specify that17:26
openstackgerritTrent Geerdes proposed a change to openstack/requirements: Adding default branch and remote in .gitreview file.
mgagnezaro: that's why they are proposing to add the ability to interact with "file like objects"17:26
nibalizermordred: im okay with 60m, its super long but it fits my definition of $BIG_NUMBER, which I think is its intended purpose17:27
*** thuc has quit IRC17:27
aconradI heard there was an openstack room porting OpenStack to Python3 @ PyCon17:27
mgagnezaro: file like object being the instance returned by open() or anything derivated17:27
anteayaaconrad: yes, 518B17:27
*** thuc has joined #openstack-infra17:27
aconradI met with mgagne, too! ;)17:27
*** hogepodge has joined #openstack-infra17:27
mgagneaconrad: I have no recall of meeting you :D /jk17:27
*** thuc has quit IRC17:27
*** jeckersb is now known as jeckersb_gone17:28
*** thuc has joined #openstack-infra17:28
anteayamordred: so how do you want to deal with the return code from run_ssh?17:29
anteayasince taht is teh magic piece for automating manage-projects17:29
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Drive puppet from the master over ssh
mordrednibalizer: I updated to use the ssh_authorized_keys type per your suggestion17:30
mordredanteaya: in the latest vesrion - it actually does respond to the return code in the for loop17:30
*** khyati has joined #openstack-infra17:30
mordredwith if ! ; then break17:30
zaroaconrad: that sounds neat,  the wizard seems useful but kinda targeted. thanks for the explanation mgagne,  i was wondering what the use case would be.17:31
anteayamordred: it breaks17:32
*** ominakov has joined #openstack-infra17:32
*** jeckersb_gone is now known as jeckersb17:32
anteayacan it log an error?17:32
mgagnezaro: yes, it's kinda targeted however the changes proposed are generic enough IMO17:32
anteayaor perhaps sleep and try again?17:33
mordredanteaya: it will sleep and try again in about 15 minutes - as for logging - we might want to write a local log just in case we want to track down problems later17:33
mordredfor the most part I think the intent is that we can see puppet errors in puppetboard17:34
*** ominakov has quit IRC17:34
*** ominakov has joined #openstack-infra17:34
anteayaif it hits the break, what will show up in puppetboard, just a failure?17:34
*** ominakov has quit IRC17:35
anteayawould it show up in this list of reports?
*** gokrokve_ has quit IRC17:37
*** denis_makogon_ is now known as denis_makogon17:37
*** SumitNaiksatam_ has joined #openstack-infra17:39
*** SumitNaiksatam has quit IRC17:39
*** SumitNaiksatam_ is now known as SumitNaiksatam17:39
*** yolanda_ has joined #openstack-infra17:40
*** yolanda_ has quit IRC17:40
*** yolanda_ has joined #openstack-infra17:40
anteayahow does @override_list get its value?17:44
mordredanteaya: it _may_17:44
mordredanteaya: it gets passed in from the thign - one sec - I'll give a link17:44
*** ildikov_ has quit IRC17:45
nibalizermordred: awesome17:45
anteayaah yes, thanks17:46
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Move to using ssh_authorized_key type
mordrednibalizer: there ^^ is a patch to update all of our creation of such files17:47
anteayaso I am looking here:
anteayathe reports seem circular17:47
anteayaI am having trouble finding out what actually ran17:47
anteayaso I probably don't know how to read the reports17:48
mordredanteaya: well - it doesn't report a lot on success17:48
mordredwhich we like17:48
mordredanteaya: for instance17:48
mordred(the first report linked)17:48
mordredshows that puppet ran and was happy and nothing changed17:49
mordredthis is what we expect to happen most times17:49
anteayathat is what that tells me17:49
mordredif you look at a report where something changed:
anteayaokay thanks17:49
mordredyou'll see what it looks like when it does take actions and is happy about it17:49
mordredand if you look here:
mordredyou'll see what it looks like when it does something and it's unhappy about it17:50
anteayawhat is v'0'17:50
mordrednibalizer: btw ^^ I have NO IDEA why runs on storyboard alternate between success and failure in that way17:50
anteayasorry u'0'17:50
mordredanteaya: the return code for teh command was 017:50
anteayaah okay17:50
mordrednibalizer:  it only seems to break on storyboard in that particular way - pip and virtualenv ARE on that system17:51
anteayaif it is unhappy about it, why did it report success for everything?17:51
anteayasorry, wrong link, I see the failure message now17:51
*** enikanorov_ has joined #openstack-infra17:52
anteayajenkins doesn't like the linting for 80976 yet17:53
anteayaso this is to get this in place and then there can be a further patch that says if override list runs without breaking, trigger manage-projects17:54
*** coolsvap is now known as coolsvap|afk17:54
*** rallegre has quit IRC17:55
anteayaand honestly since override_list is just the git_farm I would be in favour of calling the variable the git_farm17:55
anteayaunless you want to keep it open to adding more servers in the future17:55
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Move to using ssh_authorized_key type
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Drive puppet from the master over ssh
mordredanteaya: I may want to add more servers in the future - it's realy about "run these in order and fail if any of them fail" and the second part is "run all of tehse and we dont' care"17:57
anteayaso paint the bridge that says, this ran successfully on all git servers fire manage-projects17:59
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Revert "Revert "Revert "Stop running manage-projects"""
*** matjazp has quit IRC17:59
*** matjazp has joined #openstack-infra18:00
*** marun has joined #openstack-infra18:00
jeblairi love that, btw.  :)18:00
*** gema has quit IRC18:01
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Revert "Revert "Revert "Stop running manage-projects"""
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Move to using ssh_authorized_key type
*** pcrews_ has joined #openstack-infra18:02
*** zehicle_at_dell has joined #openstack-infra18:02
mordredanteaya: and actually - thanks for reminding me - ^^ there is the patch (the revert revert revert one) that adds manage-projects back in18:02
anteayamordred: so where is the bit that says only run manage-projects _after_ override_list has run successfully?18:03
*** gema has joined #openstack-infra18:04
nibalizermordred: gets my +118:04
*** pcrews__ has quit IRC18:04
nibalizermordred: I don't have time to debug storyboard right now, maybe this evening18:04
mordredanteaya: it doesn't - is in override_list18:04
jeblairoh the ssh_autherized_key is lovely!18:04
mordredanteaya: so puppet will only run on after successful runs of puppet on git*18:05
*** matjazp has quit IRC18:05
anteayaso that is the race we have been looking to address18:05
mordredanteaya: yup18:05
mordredwe address it not by associating manage-projects with pupept on git* - but by making pupept on review run m-p and by associating puppet on review with puppet on git*18:05
anteayaI am not seeing that puppet will only run on reviw.o.o after success on puppet on git18:05
anteayaI will look again18:05
jeblairnibalizer: thanks for enlightening us on that :)18:05
mordredanteaya: look in run_remote_puppet18:05
mordredjeblair: right? pretty isnt' it? (we'll need to land that one by hand because we'll need to change a few keys in hiera at the same time)18:06
mordredjeblair: since they have the header and comment baked in in hiera18:06
jeblairmordred: what's the 'disable puppet' story?  i saw a comment from nibalizer but didn't see the answer18:07
anteayaoverride_list has to be run in order and review.o.o is at the bottom of the list18:07
mordredjeblair: oh - how to deal with puppet being disabled on a host?18:07
jeblairmordred: yeah, or phrased slightly differently, what's the process to disable puppet in the nwo?18:07
*** _nadya_ has joined #openstack-infra18:08
*** nati_ueno has quit IRC18:08
mordredjeblair: for the hosts that arent' git* and review, I think "puppet agent --disable 'insert comment here'" should still dtrt.18:08
mordredjeblair: hrm - wait - I may be wrong about something...18:08
mordredthat still won't let us avoid the remote trigger while we're wanting to locally test18:09
mordredlet me noodle on it18:09
*** nati_ueno has joined #openstack-infra18:09
* mordred REALLY wished that there was a way to run puppet agent --test with puppet agent --disable in place18:10
*** pcrews__ has joined #openstack-infra18:10
jeblairmordred: does 'puppet agent --disable' write a file that we could check for?18:10
mordredlike "puppet agent --test --force"18:10
mordredjeblair: it does18:10
jeblairmordred: so maybe "if file dne puppet agent --test"18:10
jeblairshould be the ssh command18:10
mordredjeblair: /var/lib/puppet/state/puppetdlock18:11
jeblairhehe i read that as "puppet deadlock"18:11
mordredjeblair: there will be a minor race condition for people doing local testing18:11
jeblairmordred: why?18:11
mordredin that you can't run puppet yourself if agent is disabled18:11
clarkbeyes still dilated. hope to be useful soonish18:12
mordredso you'll have to do "puppet agent --enable ; puppet agent --test ; puppet agent --disable" and hope the the global run_all doesn't also try to run in the window18:12
jeblairmordred: you can't run 'puppet agent --test' if it is disabled?18:12
mordredjeblair: sadly, no18:12
*** praneshp has joined #openstack-infra18:12
jeblairthen we don't need to do anything special for the ssh trigger18:12
anteayaclarkb: rest your eyes18:13
*** pcrews_ has quit IRC18:13
jeblairmordred: maybe we should just make another file and ignore puppet agent --test18:13
jeblairignore puppet agent --disable18:13
anteayathey heal fast but they need to rest while they are healing18:13
mordredjeblair: that actually isnt' a terrible idea - since puppet agent --disable is pretty terrible18:14
mordredbtw - you can give a message to --disable to list a reason18:14
mordredbut puppet does not print that reason if you try to run --test while it's disabled18:15
*** tjones has joined #openstack-infra18:15
*** morganfainberg_Z is now known as morganfainberg18:15
mordredjeblair: oh - biggest problem - our ssh key would also have to have access to run the check for the flag18:15
nibalizermordred: puppet agent --enable && puppet agent --test ; puppet agent --disable 'message'18:16
mordrednibalizer: right - but there's a race there18:16
mordredwhat we WANT is "puppet agent --test --force" which will allow a one-time manual run even if disabled18:16
nibalizeroh there si a way18:17
nibalizeryou do puppet agent --test --lockfile=someotherfile18:17
nibalizerits not actually that argument tho18:18
nibalizergive me a sec to figure it out again18:18
tjoneshi - i am trying to run tox on mavericks and I am getting this error clang: error: unknown argument: '-mno-fused-madd' [-Wunused-command-line-argument-hard-error-in-future].  i think i know how to fix it but when i patch .tox/py27/lib/python2.7/site-packages/setuptools/command/ it appears to get overwritten with something without my fix.  anyone already fixed this issue?18:18
mordredtjones: hi! well, you can't patch anything in the .tox dir like that - those are all files genearted as part of the virtualenv18:18
*** doddstack has quit IRC18:19
mordredthe bug is probably either in easy_install upstream in its support for clang on mavericks, or in the particular package it's trying to install18:19
*** doddstack has joined #openstack-infra18:19
mordrednibalizer: puppet agent --test --puppetdlockfile=/tmp/foo18:20
*** andreaf has quit IRC18:20
mordredjeblair: ok - if we just use puppet agent --disable18:20
*** e0ne has joined #openstack-infra18:20
mordredthen as we're testing, we can use the above to skip being blocked by disable18:20
tjonesmordred: yeah that is what i suspected.  ugh18:21
jeblairmordred: wfm; a ci doc patch for that would probably be good (along with maybe 'splainin the ssh thing)18:21
nibalizermordred: awesome18:21
jeblair(cause i'll forget that)18:22
nibalizeri found agent_disabled_lockfile but --puppetdblockfile sounds more like what I remember18:22
mordredjeblair: nod18:22
mordredjeblair: I'll do that in a bit18:23
lcostantinoqq,  is there a channel for devstack?18:24
anteayalcostantino: for getting going with devstack #openstack-dev is probably your best bet18:25
anteayafor devstack development, #openstack-qa is the one you want18:25
mordredanteaya: does the patch/patches make sense to you?18:25
lcostantinoanteaya, great thanks18:25
anteayamordred: I am working though them18:25
anteayathe --disable bit I am catching up on18:26
anteayaI don't understand why it needs a file passed to the command to enable it18:26
anteayalcostantino: welcome18:26
mordredanteaya: so, sometimes we want to disable automatic runs of puppet on a server, because we're doing manual maintainence or testing of something18:26
* anteaya nods18:27
mordredso when we do that, if we just run "puppet agent --disable" on the server, it will happily not run puppet on the host from the script18:27
mordredbut we still need to be able to run puppet on the host ourselves so we can do our testing18:27
mordredso for that, we would run "puppet agent --test --puppetdlockfile=/tmp/foo" - which will fool puppet into letting us run it even though it's generally enabled18:28
* anteaya notes a foggy understanding18:28
anteayacan I hear it again in different words?18:28
anteayawhat is the relationshipw between --test and --disable18:29
anteaya--test is used when puppet has been disabled?18:29
*** jlibosva has quit IRC18:29
*** peddamat has joined #openstack-infra18:30
mordredno - --test is the poorly-named option which says "puppet, please run one time and do not start a daemon"18:30
*** peddamat has quit IRC18:30
mordredso "puppet agent --test" is just the command one uses to run puppet18:30
anteayasomething just clicked18:30
openstackgerritA change was merged to openstack-infra/jenkins-job-builder: add ability to parse from a file object
anteaya--disable stops the daemon18:30
*** sdake_ has quit IRC18:31
jeblairwe use --test in both cases (from the ssh script and when locally testing) because in both cases we want to wait until the end of the run and see if everything worked18:31
anteaya--test runs puppet without starting the daemon18:31
* anteaya nods18:31
anteayaI'm still not a 100% on why you need to pass a file as an argument18:32
anteayawhat is expected to be the contents of the file?18:32
anteayaan empty file18:32
mordredit's a path to where puppet will make a lock file18:32
mordredso passing it is a hack18:32
anteayawhy does it need to make a lock file?18:33
mordredaroudn the fact that puppet has already made a "puppet is disabled" lockfile at the default location18:33
anteayagot it18:33
anteayaa disabled puppet has a lockfile18:33
nibalizeranteaya: --test is also 'please give me output'18:34
clarkbmordred: have we tested deleting the ec2 facter file yet? I think I am going to kick off more nodepool image builds now if necessary18:35
anteayayes, I felt that was an important bit fitting into the whole we need a return code18:35
*** matjazp has joined #openstack-infra18:35
clarkbI guess that is actually an easy thing for me to test on a VM of my own18:35
*** matjazp has quit IRC18:35
anteayanibalizer: do you happen to have a sample paste of what kind of output --test might offer?18:36
nibalizersure, sec18:36
anteayanibalizer: thanks18:36
anteayaclarkb: how are your eyes?18:36
*** e0ne has quit IRC18:37
mordredclarkb: no, I have not testied it18:37
nibalizermordred: anteaya in puppet 3 at least you can do puppet agent --onetime --detailed-exit-codes --no-daemonize and it won't go to the background, produce output, or start the daemon18:37
clarkbanteaya: I have the foil sunglasses things they give you on :) and that helps a ton18:38
nibalizeranteaya: ^18:38
clarkbmordred: ok, I will take a stab at that now18:38
anteayaclarkb: k, yes my sister had the dialeted eyes once, she was blind most of the rest of the day18:39
clarkbanteaya: last time it went for a few hours. I figure I am about half way done18:39
*** gokrokve has joined #openstack-infra18:39
anteayaclarkb: cool18:39
anteayanibalizer: looking18:39
openstackgerritDerek Higgins proposed a change to openstack-infra/config: Add a check for nova-docker
*** thuc has quit IRC18:40
*** thuc has joined #openstack-infra18:40
nibalizeranteaya: so thats most puppet runs18:41
nibalizerusually puppet doesn't make changes18:41
anteayanibalizer: how could we get this output in a log file should --test fail to exit 0?18:41
nibalizeranteaya: --logdest=filename or --logdest=syslog will put it to syslog18:41
anteayalots of work18:41
anteayawhere could that go in mordred's patch?18:41
*** e0ne has joined #openstack-infra18:42
*** aysyd has quit IRC18:42
clarkbmordred: so that just totally worked on my precise test box. Going to spin up a centos6 machine now18:43
anteayanibalizer: that would put everything to syslog even runs with exit code 0, would it not?18:43
*** thuc has quit IRC18:45
nibalizeranteaya:  in the run_puppet function i believe18:45
anteayatoo noisy methinks18:45
*** MarkAtwood has quit IRC18:45
anteayaI just made a comment on patch 618:45
mordredclarkb: awesome18:46
mordrednibalizer: I don't think that'll do what you're thinking it will do18:46
mordredthere are two things  we need - one is logs of the puppet run if puppet failed - those are already going to puppetmaster, pupeptdb and then puppetboard18:47
*** ildikov has joined #openstack-infra18:47
mordredwe don't need to do anything with logdest to get that18:47
mordredthe other thing we _might_ need is a log of the run_all script itself - so taht we can catch a timeout failure18:47
*** hogepodge has quit IRC18:48
*** _nadya_ has quit IRC18:51
*** zns has quit IRC18:51
*** zns has joined #openstack-infra18:53
nibalizermordred: i agree that logdest is probably not needed, just mentioning its existence. Right now logs go to syslog i believe and we'd have to set logdest=syslog to maintain that behavior... which I agree we probably don't need to do18:53
mordrednibalizer: nod18:54
nibalizermordred: right and you have a TODO in there with a case to perform some action if the puppet node failed18:54
*** aysyd has joined #openstack-infra18:55
*** khyati has quit IRC18:56
*** jcoufal has joined #openstack-infra18:57
*** mrda_away is now known as mrda18:58
*** _nadya_ has joined #openstack-infra18:58
*** dstanek has quit IRC18:59
*** zehicle_at_dell has quit IRC19:00
*** MarkAtwood has joined #openstack-infra19:00
*** nati_ueno has quit IRC19:01
*** pdmars_ has joined #openstack-infra19:01
*** nati_uen_ has joined #openstack-infra19:02
clarkbpleia2: you don't happen to know where fedora20 hides facter's libs do you?19:03
clarkbpleia2: I am digging around on a tripleo-f20 image and learning I don't know how to yum19:03
*** markmc has joined #openstack-infra19:03
*** khyati has joined #openstack-infra19:04
anteayaclarkb: she is flying19:04
*** pdmars has quit IRC19:05
anteayashe might get wifi when she hits chicago, in an hour or so I think19:05
*** dstanek has joined #openstack-infra19:05
*** dripton is now known as dripton_afk19:05
clarkbah thanks19:05
*** _nadya_ has quit IRC19:06
*** jswarren_ has joined #openstack-infra19:06
anteayamordred: I'm waiting on a new patchset for 80976 that addresses nibalizer's comments19:07
anteayaI'm good with the patches taht depend on it19:07
*** jswarren has quit IRC19:08
*** gondoi is now known as zz_gondoi19:09
*** dstanek has quit IRC19:09
*** doddstack has quit IRC19:10
*** aysyd has quit IRC19:10
clarkbpleia2: anteaya: found it. If you install yum-utils you can `repoquery --list packagename` to get a list of files19:10
*** thedodd has joined #openstack-infra19:10
anteayaclarkb: \o/19:11
*** ociuhandu has quit IRC19:13
nibalizeralso rpm -ql facter19:13
*** aysyd has joined #openstack-infra19:16
aconradzaro: thanks for reviewing my patch19:17
*** dripton_afk is now known as dripton19:17
aconradzaro: what's the release cycle of jjb? when you I expect a new release of the package?19:17
aconradwhen should I expect*19:18
*** Ajaeger1 has joined #openstack-infra19:18
*** thedodd has quit IRC19:19
*** thedodd has joined #openstack-infra19:20
*** zz_gondoi is now known as gondoi19:22
*** thedodd has quit IRC19:22
*** reed has quit IRC19:22
*** chandan_kumar has quit IRC19:22
openstackgerritMatthew Treinish proposed a change to openstack-infra/config: Change names for duplicate neutron jobs
*** pcrews_ has joined #openstack-infra19:23
openstackgerritClark Boylan proposed a change to openstack-infra/config: Remove facter ec2 metadata facts.
clarkbmordred: jeblair ^ is tested on precise and centos6 but not fedora20 yet, but I don't expect facter to act differently on fedora19:23
*** reed has joined #openstack-infra19:23
Ajaeger1Hi clarkb: could you remove your -1 from , please ? the translation team really likes to have these in - see Akhiro and Sascha's comments19:23
*** thedodd has joined #openstack-infra19:24
clarkbAjaeger1: looking19:24
clarkbAjaeger1: I would really rather I left my -1 there so others can add their comments19:24
clarkbI defer to you guys on what actually merges, but without the -1 other reviewers may not see there is some conflict19:25
slagleanyone around who can apply this change to gerrit?
*** pcrews__ has quit IRC19:26
*** enikanorov_ has quit IRC19:26
*** vhoward has left #openstack-infra19:26
*** enikanorov_ has joined #openstack-infra19:27
slaglei think maybe they're supposed to get applied manually on fridays? but i don't see that it has been applied when i look at the project acl's in gerrit19:27
*** pcrews has joined #openstack-infra19:27
*** pcrews_ has quit IRC19:27
Ajaeger1clarkb: understood19:27
Ajaeger1anybody else that wants to comment on in either way?19:28
*** amcrn is now known as notamrith19:28
clarkbslagle: last week was a pretty terrible week for us and I am pretty sure we simply didn't run the script on friday19:28
clarkbslagle: it turns out when the entirety of the Internet becomes untrusted it makes for a terrible week19:28
slagleclarkb: sure, totally understood, i'm just asking politely :)19:29
Ajaeger1clarkb: I hope and are less controversial - could you review those as well as time permits, please? The first is an optimization that will make some projects happy, the second is avoiding the duplicated IDs19:30
clarkbslagle: I can try to get to it tomorrow19:30
slaglei'm not sure if it was applied, but didn't do what i expect, or wasn't applied is all19:30
slagleclarkb: thanks19:30
clarkbslagle: I don't think it would've applied at all19:30
clarkbAjaeger1: looking19:30
*** lcostantino has quit IRC19:31
*** notamrith is now known as amcrn19:32
clarkbAjaeger1: reviewed the first one. One small command issue otehrwise lgtm19:36
*** gondoi is now known as zz_gondoi19:36
*** _nadya_ has joined #openstack-infra19:36
Ajaeger1clarkb: thanks, just reading your comments and will update quickly19:37
rcarrillocruzhey guys, could anyone please review ?19:37
rcarrillocruznot sure if i should write the code for Gerrit 2.8 hooks given that we will upgrade presumably in May19:37
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Move to using ssh_authorized_key type
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Revert "Revert "Revert "Stop running manage-projects"""
clarkbrcarrillocruz: I will look in a bit19:38
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Drive puppet from the master over ssh
rcarrillocruzthx clarkb19:38
nibalizerrevert revert revert19:38
clarkbyou haven't lived until you have 3 nested reverts19:39
mordredjeblair, nibalizer: I have added docs19:39
rcarrillocruzclarkb: Gerrit 2.8 does contain --commit in the change-abandoned hook, the change above is against Gerrit 2.4.219:39
nibalizermordred: awesome! /me loves docs19:39
mordredI also added a puppet defaults file to turn "start on boot" off for the pupept agent19:39
rcarrillocruzi'm doubtful if i should write the change against 2.4.2 AND 2.819:39
rcarrillocruzland it as is19:39
rcarrillocruzor rewrite it completey against 2.819:40
*** nkinder has quit IRC19:40
mordredclarkb: also - you may wan tto look at those three patches - nibalizer, anteaya, jeblair and I talked through them a little bit while you were having your eyes gouged19:40
nibalizerya, might also want to put a puppet agent --disable on '' since the ubuntu package will start the daemon up on install19:40
nibalizeror not disable19:40
clarkbAjaeger1: has the --ignore-obsoluete chaneg been tested? should it just be --ignore-obsolete instead of --ignore-obsolete=true?19:40
nibalizeri mean disable the service19:40
clarkbAjaeger1: reading the babel docs it isn't quite clear to me19:40
clarkbanteaya: looking at
Ajaeger1clarkb: I tested it - and it needed --ignore-obsolee=true19:40
mordrednibalizer: I think it installs with the start to fase though19:41
clarkbAjaeger1: ok thanks19:41
mordrednibalizer: I just had to remove a bunch of docs about how to enable the agent to run19:41
nibalizermordred: oh that would be amazing, puppetlabs++ for packaging right19:41
clarkbAjaeger1: approved that one19:41
mordredclarkb: awesome19:42
Ajaeger1clarkb: thanks!19:42
Ajaeger1clarkb: will keep an eye on the jobs19:42
clarkbmordred: can you chime in on for Ajaeger1 ?19:42
*** lcostantino has joined #openstack-infra19:42
mordredclarkb: only if you'll review my patches while I do :)19:43
clarkbmordred: okay19:43
Ajaeger1clarkb, mordred: Thanks!19:43
mordredAjaeger1: so - still looking - but to set the stage here19:44
mordredAjaeger1: often times people prefer to have a thing a certain way to solve a problem19:44
mordredwe find many times that the problem trying to be solved by other communities is framed different in the face of the infra automation machine19:45
mordred(which is just my thoughts based on "There has been quite some discussion on the opensuse-i18n mailing list and for some projects, translators prefer to have the files in - and thus we need to handle this.")19:45
*** jcoufal has quit IRC19:46
anteayamordred: looks good to me19:46
Ajaeger1mordred: history of this patch is that the jobs where failing since we had .tx/config in the repos, then it was manually updated, removed, added after complaints,...19:46
mordredAjaeger1: yah - so, why not change it to git reset --  ?19:47
mordredinstead of git add?19:47
*** nati_uen_ has quit IRC19:47
*** e0ne has quit IRC19:47
mordredI can see why it's possible someone might want it tin the rpo - is there a valid use case where our update scripts might want to change its value?19:47
Ajaeger1mordred: would work as well but then it needs manual update eventually...19:47
*** e0ne has joined #openstack-infra19:48
*** dprince has quit IRC19:48
Ajaeger1mordred: Yes, whenever a new file gets added. For example in openstack-manuals, when we create a new manual19:48
*** ociuhandu has joined #openstack-infra19:48
Ajaeger1Since this is a post job, failures are not really seen ;(19:48
mordredAjaeger1: oh - wow, really?19:48
Ajaeger1unless somebody monitors them.19:48
adam_gare the devstack-gate nodes setup with any special local iptables rules that would affect connections to $HOST_IP?19:49
clarkbadam_g: yes19:49
Ajaeger1mordred: yeah, that's part of the underlying problem why I'd like this done foolproof ;)19:49
clarkbadam_g: iptables is fairly locked down19:49
mordredAjaeger1: ok. I think I can be on board with the add19:49
adam_gclarkb, ah!19:49
mordredI mean, the world sucks and all19:49
clarkbadam_g: let me show you what heat does to fix it19:49
*** e0ne has quit IRC19:50
mordredAjaeger1: +2 - you'll need to convince SergeyLukjanov or jeblair too - since fungi is out this week19:50
clarkbadam_g: heat specific rule additions. YOu can also run iptables directly in your test (just be sure to put rules at the beginning of the ruleset and not the end if you do that)19:50
mgagneis there any schedule available for Friday 16th?19:51
Ajaeger1mordred: thanks, will do at another time.19:51
mordredmgagne: for what?19:51
Ajaeger1Now fixing the problems that clarkb pointed out...19:51
Ajaeger1... on my other patch19:51
clarkbmgagne: the 16th isn't friday19:51
mgagnemordred: whatever there is left19:51
mordredoh - at the summit19:52
mgagnemordred: yep, May19:52
mgagneclarkb: sorry, context: openstack summit19:52
clarkbmgagne: gotcha19:52
adam_gclarkb, awesome. thanks.19:52
clarkbadam_g: running iptables direclty in your test might be a good place to test the fix then we can formalize it in puppet19:53
*** sarob has joined #openstack-infra19:53
clarkbnice thing about test specific is then you aren't waiting for us when images need to update19:53
*** mrmartin has quit IRC19:53
clarkbmordred: so I udnerstand, we are running all puppet jobs from this cron after your change?19:54
clarkbmordred: the number of servers isn't raelly big so should be fine. Just making sure I grok the intent19:54
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Reduce number of files send for propose_translation
mordredclarkb: yes19:54
Ajaeger1clarkb: updated the patch - the nit you noticed was a real bug19:54
mgagneclarkb: so, is there a schedule available for Friday? mostly design sessions?19:54
adam_gclarkb, yeah--ill put it in the devstack deployment to see if it helps19:54
mordredclarkb: then, follow up step - we turn run_all/run_remote_puppet into python and make it a non-jenkins-slave zuul worker19:55
mordredclarkb: so that we can trigger runs of it in the post pipeline19:55
clarkbmgagne: I am not involved in summit planning but my understanding is it will be all design sessions19:55
clarkbmgagne: conference runs monday to thursday and design session tuesday to friday19:55
openstackgerritA change was merged to openstack-infra/config: Do not generate obsolete translations
zaroaconrad: i don't have the power to release jjb (*yet*)19:55
mordredbut for now, having a script in a cron seems more better19:56
mordredzaro: weren't we working on that?19:56
mgagneclarkb: cool, trying to see if it's worth it staying until Saturday or not. Some people aren't technical so it's hard to justify it =)19:56
zaromordred: not approved yet,
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Drive puppet from the master over ssh
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Move to using ssh_authorized_key type
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Revert "Revert "Revert "Stop running manage-projects"""
*** zz_gondoi is now known as gondoi19:58
* mordred missed rocketships19:58
zaroaconrad: i'll probably try to do every 3months if given the power.19:58
clarkbAjaeger1: lgtm19:59
*** fbo_away is now known as fbo19:59
mordredclarkb: is there a reason we've not approved the jjb patch?19:59
mordredclarkb: it has 5 +2's on it20:00
clarkbmordred: which jjb patch?20:00
Ajaeger1clarkb: thanks for the review!20:00
clarkbmordred: ah, I didn't realize jeblair had changed his position20:00
*** e0ne has joined #openstack-infra20:00
clarkbmordred: in that case no, I think it can be approved20:00
aconradzaro: Ok cool. Well, I'll try to keep an eye on the release. Once it's released, I'll patch my wizard tool to pass file objects to JJB rather than filenames20:01
tjonesmordred: i think i have "solved" the issue i have with clang and mavericks.  is there a good place to document this so others don't suffer as much?20:01
aconrad(it's already patched, I'll just need to wait for jjb to get released before merging it upstream)20:02
mordredtjones: hrm. not sure - I thnk there is a "getting started" section on the wiki about getting dev workstations set up20:02
mordredanteaya: ^^ do you know where taht section of the wiki is?20:02
clarkbtjones: mordred: also individual projects have docs for getting tests running20:02
zarorcarrillocruz: i would make 86382 work with gerrit 2.8.  your change is unlikely to land before upgrading anyways.20:02
clarkbin doc/source/somethingsometing20:02
aconradwhat are +2 code reviews?20:02
mordredaconrad: reviews from core reviewers20:03
aconradah, weighting20:03
mordredaconrad: we require as a general policy for patches to have 2 +2 reviews before landing them20:03
aconradmordred: I see. Gerrit should have a total count of upvotes20:04
aconrad(and down-votes)20:04
anteayamordred: I will look20:04
zaroaconrad: we don't think there's value in that.20:05
aconradand what's the green check?20:05
zaroaconrad: same as +220:05
aconradI thought these were core-reviewers20:05
aconradah ok20:05
Ajaeger1Can anybody help me where I can find the jenkins logs easily for the latest runs of the translation jobs? I used to look at but api-site-manuals-propose-translation-update was run there 5 days ago - and no other jenkins has the logs...20:05
*** _nadya_ has quit IRC20:05
*** _nadya_ has joined #openstack-infra20:05
*** markmc has quit IRC20:05
Ajaeger1here's the link:
zaroaconrad: although i think you can make gerrit have total up votes with some config settings20:06
anteayathis is the only wikipage I see that deals with setting up devstack:
anteayaand it is dated20:07
Ajaeger1something is broken - shows me a running job for openstack-manuals with job openstack-manuals-manuals-upstream-translation-update:  but clicking on the link, I get a 404.20:07
Ajaeger1This is the link:
Ajaeger1is broken?20:08
anteayathe gerrit workflow page:
anteayalinks to this page for devstack:
anteayawhich doesn't show up in the wiki search20:08
clarkbAjaeger1: is the last run job there with a console log20:09
clarkbAjaeger1: and that console log looks suspiciously wrong. Not sure what is happening there20:09
clarkbhrm, could this be related to my reload config off disk stuff that happened last week?20:09
*** UtahDave has quit IRC20:09
anteayait doesn't show up for "installing a devstack" search20:09
Ajaeger1clarkb: five days ago?20:09
*** e0ne has quit IRC20:09
Ajaeger1clarkb: exactly, there're the jobs missing.20:10
clarkbAjaeger1: yeah that is about when te jenkins config loading happened20:10
*** e0ne has joined #openstack-infra20:10
Ajaeger1clarkb: the job was run, see
annegentlemordred: three reverts in a row in a sentence? Wowsa20:10
Ajaeger1but jenkins does not know about it20:10
*** e0ne has quit IRC20:11
sdagueannegentle: mordred is the king of the triple negative20:11
clarkbAjaeger1: that is amazing20:11
clarkbAjaeger1: I am looking on disk now to see if maybe permissions are breaking jenkins or similar20:11
annegentlesdague: two reverts make it a vert?20:11
*** e0ne has joined #openstack-infra20:11
sdagueannegentle: or a horiz?20:11
anteayacurrently the devstack documentation is in a github repo, dtroyer is working on some steps to bring it into the fold20:11
clarkbAjaeger1: it looks fine on disk20:12
clarkbmordred: jeblair should I do a full restart of jenkins.o.o and see if that fixes it? I have never seen this before20:12
mordredclarkb: fine by me?20:13
*** pcrews has quit IRC20:13
clarkbputting it in shutdown mode now20:13
clarkband no jobs running so restarting it20:14
rcarrillocruzok zaro, will do that then, thx20:14
*** BadCub01 has quit IRC20:15
openstackgerritA change was merged to openstack-infra/config: proposal to allow additional people to release jjb
*** e0ne has quit IRC20:16
clarkbAjaeger1: that appears to have fixed it20:16
clarkbAjaeger1: jenkins bug I guess20:16
*** e0ne has joined #openstack-infra20:17
anteayachmouel: great, thanks for submitting that20:17
Ajaeger1clarkb: indeed. Now I see many run jobs ;) Too many to review tonight ;)20:17
annegentlejeblair: hey what's an example of a crippled open-core project? (I need a history lesson in open source I think)20:19
*** e0ne_ has joined #openstack-infra20:20
*** e0ne has quit IRC20:20
clarkbannegentle: nginx, mysql, puppet(sort of) I can probably think of more if given time20:20
mordredannegentle: Eucalyptus20:21
aconradhey jeblair, can you send me in PM your email? I think I left your card at my hotel room. I'd like to send you this email about the painful process of submitting a patch to openstack20:21
annegentleclarkb: so in this case, the meaning is, unless you use all the "commercial" plugins it's sort of useless standalone?20:21
mordredin that there are features you only get if you buy something20:21
mordredand mainly, that those features are REALLY something that the main thing should have20:21
aconrador maybe I can write a blog post20:21
annegentlegot it, thanks mordred and clarkb20:21
*** gokrokve has quit IRC20:21
mordredand that they've been held out of the main thing so that they can be offered as a for-purchase add-on20:21
annegentlejeblair: no need to define :)20:21
mordredannegentle: yah20:21
*** mwagner_lap has quit IRC20:22
mordredclarkb: actually, thus far mysql has amanzingly managed to remain not open core20:22
mordredclarkb: the management at the time I was there kept wanting to make it so, and we kept successfully fending them off20:23
clarkbmordred: huh, I thought you paid money to sun/oracle for the good stuff20:23
clarkbwhat do you pay for then?20:23
*** _nadya_ has quit IRC20:23
mordredthere are add-ons like an external monitoring service20:23
*** Ajaeger1 has quit IRC20:23
mordredbut that's not really part ofa  db server20:23
mordredyou pay for support20:23
annegentleclarkb: mordred: yeah mysql felt like "one of these things is not like the other"20:23
clarkbmordred: I swear you paid and got non gpl extras20:24
clarkbbut maybe tat was just so you could hack on it without the gpl?20:24
mordredclarkb: if you buy support, they will ship you a copy of the "enterprise dashboard" - but you are very unlikely to actually ever insall or use it20:24
mordredclarkb: that is the other thig you can pay for - a non-gpl licensed version20:24
annegentleok, then my next question, is the open core problem what we're running away from by making integration quite wide? I mean, is there a happy medium between killing say docs, infra, qa with a bunch of integration and keeping OpenStack useful?20:25
mordredannegentle: I'd say yes and no20:25
annegentleI think there is but the scaling of integration is really scare20:25
clarkbannegentle: clearly my opinion is wrong here, but I think openstack would be quite useful if we just said only mysql, rabbit, and kvm are supported with openvswitch20:25
annegentleclarkb: hey man, that's useful stuff20:25
mordredannegentle: for me, pulling in docs and infra into the _effort_ called openstack20:26
clarkbthe breadth of supported technology doesn't ahve much to do with usefulness as accomodating particulargroups of people and their expertise imo20:26
mordredis about recognizing that the work going in to those is essential to the project and is no more or less important than work on nova20:26
*** pcrews has joined #openstack-infra20:26
clarkbwhich is a very valid raeson to do it, but doesn't necessarily make openstack more useful20:26
mordredannegentle: on the other hand, I think that additional integrated projects are about providing features that end-users want20:26
sdagueis there a place where this conversation started? I feel like I'm not finding it in the backlog.20:27
*** dizquierdo has joined #openstack-infra20:27
clarkbsdague: I have no idea20:27
*** cody-somerville has joined #openstack-infra20:27
clarkbsdague: but it seems close to my proposed session so I am jumping in :)20:27
annegentlesdague: I read jeblair's TC candidacy email20:27
mordredsdague: annegentle just started asking questions,so I started answering them :)20:27
mordredah - me goes to read20:27
clarkbmordred: ok, for some erason I thought non gpl came with extras20:27
sdagueheh, ok. I didn't read jeblair's TC candidacy :)20:27
mordredclarkb: nope20:27
annegentleyeah context and me, not so much introduced20:27
annegentlejumping in is fun!20:28
mordredI completely agree with jeblair's platform, btw20:29
mordredannegentle: so, to put words in his mouth, one of the fears _I_ have with the core definition stuff is any time I hear the words "OpenStack shouldn't do X because company Y is using that as a vector for monetization"20:29
clarkbmordred: ok, first of your changes has been reviewed, please see comments but +220:30
annegentlemordred: we're not allergic to money20:30
*** e0ne_ has quit IRC20:30
mordredI support company Y doing X for moneteization, but if it makes sense for us to be doing it, I'm not going to not do it just because they are20:30
*** e0ne has joined #openstack-infra20:30
annegentlemordred: yes and I'm only allergic to more upstream work if it's not really meant to be for lots of users20:30
mordredannegentle: right - but just because piston and nebula AND canonical have each written their own bare metal provisioning framework, does not mean that openstack should avoid doing so20:31
annegentlemordred: ah the converse too. Yes same with training guides in the community.20:31
mordredannegentle: in fact, the fact that ALL of them have done it and have not collaborated on it and have used it as a vehicle to lock users in to their product proves to me that it's a needed thing adn that the world will be a better place if we have an opinion20:32
sdagueannegentle: and, another part of the counter to that is this -
mordredannegentle: otoh - I don't think we need to just take on every project just because it's there20:32
*** dkranz has quit IRC20:32
*** nati_ueno has joined #openstack-infra20:32
clarkbsdague: that was funny20:32
sdagueakamai had a patch to openssl that they thought protected them from heartbleed, they kept it to themselves20:32
sdagueas soon as they openned it, the community immediately found a vulnerability in it20:33
clarkbsdague: it was fundamentally flawed20:33
*** melwitt has quit IRC20:33
sdagueit was not openned because "it was assumed to not be interesting for general users"20:33
*** melwitt has joined #openstack-infra20:33
*** e0ne has quit IRC20:33
sdagueclarkb: sure20:34
*** gokrokve has joined #openstack-infra20:35
mordredannegentle: exactly. so the real trick to us to make sure that we accept or reject projects and features on their merits and weigh those against teh cost to us as we have been doing20:35
mordredannegentle: I think it's essentially not our job AT ALL to weigh those against the potential negative impact in teh market it may have on a company who has based a product aroudn a lack of a given feature in openstack20:36
annegentlemordred: yes and I need to let go of the "cost to docs" because of scaling... perhaps not let go but find more ways to enable teams20:37
sdagueannegentle: yeh, I think that's something that all the cross project teams face20:37
sdagueat 5 integrated projects, I think the cross project teams were the primary generators of the content20:38
annegentlesdague: yes that seems to be the crucial breaking point for valuing integration as much as we do20:38
sdaguebut at 11+20:38
sdaguethey have to be more review and mentoring20:38
sdagueand the content functions drive back to the project teams20:38
annegentlesdague: yes the near-doubling is what has me doing the doc contribution survey and considering tool changes and requirements20:38
*** jp_at_hp has quit IRC20:38
mordredannegentle: ++20:39
sdaguebecause the cross project teams are never going to grow as fast at the # of integrated projects20:39
annegentlemordred: glad you like it since I'll surely need help with tooling :)20:39
annegentlesdague: measuring by contributors? Nova still dwarfs us all (ha star terms)20:39
*** mdenny has quit IRC20:40
*** sarob has quit IRC20:40
annegentlesdague: not doubting, just wondering20:41
sdagueannegentle: the way I think about it is by people that are only dedicated to that team20:41
ekarlsoWhat's the discussion about ?20:42
sdaguehonestly, this is something I've definitely thought about as we evolved the QA team20:42
annegentlesdague: oh for sure... same kind of ratios as dev:doc20:42
sdaguebecause it used to be the core team members did most of the test cases. Now, honestly, very few do20:42
*** mdenny has joined #openstack-infra20:42
sdaguethe core team members do plumbling and reviews, and most of the test code comes from non core contributors20:42
annegentleekarlso: I was asking for examples of crippled core from Jim's TC candidacy email, I don't have as much open source project history as this group so they're a good group to ask20:42
annegentleekarlso: plus they put up with my dumb questions20:43
sdaguewho are often part of other project teams20:43
annegentlesdague: ah okay so you're ahead of docs in that way I'd say20:43
clarkbmordred: did you see my comments on my change?20:43
sdagueannegentle: I just assumed all the docs folks were super human. Because honestly I could never imagine keeping up with all of the projects at this point :)20:44
mordredclarkb: me lookie lookie20:44
*** dstanek has joined #openstack-infra20:44
clarkbmordred: and where I said my change I meant your change20:44
*** msabramo has joined #openstack-infra20:44
mordredclarkb: first comment - I added the key just the other day, so I'm fairly certain it's not in use for something else20:44
annegentlesdague: oh it's nuts and we do tend to get super human efforts around release time, but that's the unsustainable model :)20:44
*** cody-somerville has quit IRC20:45
mordredclarkb: on the before- do you mind if we do that as a follow on? or do you think we should do it now?20:45
annegentlesdague: your qa contributions seem sustainable. I've got to write an email about pre-incubating and incubating projects and docs so this is all good research, thanks.20:45
clarkbmordred: well, I think it can happen as a followup20:46
clarkbsince puppet should do it implicitly for us on file resource20:46
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add a requirement on /root/.ssh
mordredclarkb: there ya go20:47
*** MarkAtwood has quit IRC20:47
*** yamahata has joined #openstack-infra20:47
*** nati_ueno has quit IRC20:48
openstackgerritMarc Abramowitz proposed a change to openstack-infra/jenkins-job-builder: Add tox "coverage" target
*** sarob has joined #openstack-infra20:49
mordredok. I've gotta run20:50
openstackgerritRamy Asselin proposed a change to openstack-infra/config: nodepool image update works behind a http proxy
*** nati_uen_ has joined #openstack-infra20:52
* zaro takes a walk.20:52
*** MIDENN_ has joined #openstack-infra20:54
openstackgerritMarc Abramowitz proposed a change to openstack-infra/jenkins-job-builder: `test` command read/write stdin/stdout by default
*** mdenny has quit IRC20:55
*** nati_uen_ has quit IRC20:56
msabramoharlowja: Howdy Josh!20:56
asselinclarkb, fungi, jeblair, mattoliverau thanks for your help getting prepare_node to work behind a http proxy. I just submitted my solution:
*** Sukhdev has joined #openstack-infra20:57
*** fbo is now known as fbo_away20:59
*** markmcclain has joined #openstack-infra21:00
*** nati_ueno has joined #openstack-infra21:00
*** mfer has quit IRC21:01
*** nati_ueno has quit IRC21:01
*** nati_ueno has joined #openstack-infra21:02
openstackgerritRamy Asselin proposed a change to openstack-infra/config: nodepool image update works behind a http proxy
*** sabari2 has quit IRC21:07
dhellmannmordred, clarkb, jeblair : is there any chance I could get one of you to look at tonight or tomorrow morning? SergeyLukjanov left +2 but I lost fungi's when I had to rebase.21:08
harlowjamsabramo howdy!21:12
msabramoharlowja: all you tigers seem to have landed in OpenStack. Amazing.21:14
harlowjamsabramo hey :)21:15
harlowjasome of them ;)21:15
harlowja*not all*21:15
msabramoharlowja: Sure. I see. As soon as I leave Yahoo, *then* you guys all start using Python ;-)21:15
harlowjawell somewhat, lol21:15
msabramowell maybe "as soon" was several years in fact :-)21:16
*** ameade has quit IRC21:16
harlowjawell also python still not super-approved, we have special rights21:16
harlowjaits dumb still21:16
harlowja*still somewhat dumb*21:16
*** ameade has joined #openstack-infra21:16
harlowjanot all perfect yet, lol21:17
*** imcsk8_l1p has quit IRC21:17
*** imcsk8_lap has joined #openstack-infra21:17
harlowjamsabramo u can help make :-P21:17
harlowjamake it even better :-P21:17
*** melwitt has quit IRC21:18
*** marcoemorais has joined #openstack-infra21:18
*** melwitt has joined #openstack-infra21:18
*** jeckersb is now known as jeckersb_gone21:20
harlowjamarcoemorais msabramo more tigers, ha21:20
*** ameade has quit IRC21:22
*** sdake_ has joined #openstack-infra21:27
openstackgerritDerek Higgins proposed a change to openstack-infra/config: Add a check for nova-docker
msabramoCool! marcoemorais too! Hello as well!21:27
*** sarob has quit IRC21:28
marcoemoraisharlowja msabramo howdy!21:30
harlowjamarcoemorais hi!21:30
marcoemoraisharlowja: (as he looks at my cube)21:31
harlowjauma can be found somewhere, though not usually on irc21:31
harlowjaand ed and ken sometimes to21:31
*** dcramer_ has joined #openstack-infra21:32
msabramook so we should just rename this channel to #tiger_reunion :)21:32
*** dkehnx has quit IRC21:33
*** MarkAtwood has joined #openstack-infra21:34
*** aconrad has quit IRC21:35
*** aconrad has joined #openstack-infra21:35
*** dcramer_ has quit IRC21:36
*** nkinder has joined #openstack-infra21:39
clarkbdhellmann: that change is really dense :P so my biggest concern off the bat is it looks like we are adding a ton of new tests21:41
clarkbdhellmann: looks like this is almost doubling the number of tests we run?21:41
clarkband we already fell over when down to one cloud provider...21:42
dhellmannI don't know about doubling, I don't know the base number21:42
clarkbjeblair: mordred SergeyLukjanov ^ is that somethign we should be worried about21:42
clarkbdhellmann: it looks like between 2 and 5 new tests per project21:42
dhellmannthe goal is to avoid the issue we had with oslo.config during icehouse where releases broke unit tests in applications because of subtle internal details changing21:42
clarkbya, I think it tackles a real problem, but I am not sure this is sustainable21:43
*** pcrews_ has joined #openstack-infra21:43
dhellmannyeah, SergeyLukjanov had similar feedback21:43
dhellmannwe talked about figuring out if there's a way to run all of the tests for an application in one job (cross test it against one lib at a time, but all in one job)21:44
dhellmannand vice versa, I guess21:44
*** mrmartin has joined #openstack-infra21:44
dhellmannI'm not sure if that's any better -- the jobs might take a long time that way, does that matter?21:44
*** dkehn_ has joined #openstack-infra21:44
clarkbya people get cranky when tests take a long time21:45
* dhellmann is people21:45
clarkbbut they also get cranky when we go from two cloud providers down to one and tests queue for hours21:45
* dhellmann hates people21:45
*** dkehn_ is now known as dkehnx21:46
harlowjapeopel sux21:46
dhellmannfwiw, I share your concern21:46
dhellmannthis is the best we've been able to come up with, so far :-/21:46
*** pcrews has quit IRC21:46
sdaguewhat if we did what we do with d-g, and have the unit tests run on the projects with the git versions of this libs21:46
dhellmannsome of the libs that don't change often wouldn't need to have every app tested against them, so that's one area we can cut down later on21:46
dhellmannbut the big ones, like config and messaging, I think we want those in the mix21:47
dhellmanni18n maybe less so21:47
clarkbI am reviewing it and willing to give it a shot, but also pointing on the concerns as I go21:47
sdaguethen you'd just be adding a bunch of tests on the oslo* side to by symetric21:47
dhellmannsdague: yeah, these are all symmetric now21:47
sdagueright, but they are symetric by adding jobs to the projects21:47
sdaguevs. changing the existing jobs21:47
dhellmannwhat if the unit test jobs in the gate installed oslo libs from source instead of pypi?21:48
*** nkinder has quit IRC21:48
sdaguethat's just a ton of nodes to burn to block bugs that should be easily resolvable21:49
dhellmannok, I guess I can look at changing the tox job definition to do that21:49
*** chuck_ has joined #openstack-infra21:50
dhellmannit'll have to be smart about finding a list of requirements to update, by knowing what's an oslo lib somehow21:50
dhellmannand I suppose it's going to piss people off if the tests work locally but not in the gate21:50
dhellmannso we need some way for them to do the same thing on their local system21:51
*** zul has quit IRC21:51
dhellmannif they want, that is21:51
*** mrmartin has quit IRC21:52
* dhellmann considers naming the script ("we've replaced their normal coffee with...")21:52
*** jhesketh has joined #openstack-infra21:52
*** mriedem has quit IRC21:53
jeblairif we run all the unit tests in one job, won't it lake a very long time?21:53
jeblairsdague, dhellmann: eg, don't we already have 30-40 minute unit tests runs for a single project?21:53
*** chuck_ has quit IRC21:53
*** zul has joined #openstack-infra21:53
jeblairjhesketh: good morning21:53
clarkbjhesketh: o/21:53
dhellmannjeblair: yes, we would still need separate jobs for the oslo libs21:53
jheskethhow's pycon going?21:54
sdaguejeblair: yes, the neutron unit tests can hit 60 minutes on py26 4x21:54
dhellmannjeblair: but the new idea is to just install oslo libs from source when running the unit tests for an app21:54
jeblairdhellmann: oh in the normal jobs?21:54
dhellmannjeblair: so there's only one unit test job, and it always runs from source -- or maybe we add a 2nd to do that?21:54
mgagnejeblair: any ETA for that one?
*** Sukhdev_ has joined #openstack-infra21:55
jeblairdhellmann: hrm, i worry about that a bit; i think one of our goals is to make sure that projects work with their specified deps21:55
dhellmannjeblair: yeah, the more I think of it the more I think a second job works better21:56
dhellmannone with deps from pypi and a second with oslo deps from source overriding the pypi packages21:56
*** eharney has quit IRC21:56
dhellmanns/pypi/our mirror/g21:56
*** bhuvan has joined #openstack-infra21:57
jeblairjhesketh: i'm back from pycon now; others are still there for the sprints21:57
*** zul has quit IRC21:57
*** Sukhdev__ has joined #openstack-infra21:57
jeblairmgagne: it looks like we were too short-staffed to import that on friday21:57
jheskethjeblair: ah, you're missing the fun! (I'm assuming :-P). How was the conf main?21:57
*** Sukhdev_ has quit IRC21:57
*** dims has quit IRC21:57
*** Ryan_Lane has quit IRC21:57
jeblairmgagne: i think it was just clarkb and mordred on friday, and i guess they didn't get to it21:58
*** asettle has joined #openstack-infra21:58
mgagnejeblair: next friday? =)21:58
clarkbI am going to give it a try tomorrow21:58
jeblairmgagne: so next chance is next friday, or, it's possible mordred's fixes to manage-projects may be close to ready, in which case it can happen as soon as they merge and we start that21:58
clarkbif there are no objects to that21:58
*** Sukhdev has quit IRC21:58
jeblairclarkb: give what a try?21:59
jeblairclarkb: manually, or trying mordred's thing?21:59
dhellmannsdague, jeblair, clarkb : I'll take a stab at defining the oslo-ified-unit-test job tomorrow, if you think that's a better approach?21:59
clarkbinitially was thinking manually but mordreds thing is close too, can probably give the switchover a shot too21:59
jeblairjhesketh: i spent pretty much the entire time talking to people21:59
sdaguedhellmann: it seems like we can get coverage with only 1 extra job per core project, which is good21:59
clarkbdhellmann: not sure, I am willing to give the current change a shot if others are on board21:59
clarkbbut the proposed idea sounds good22:00
dhellmannsdague: yeah, we still need N jobs for the libs themselves, but only for integrated projects22:00
clarkbthis matrix checker is good I think22:01
jeblairdhellmann: i think i have a gap in my conception --22:01
*** markmcclain has quit IRC22:01
clarkbdoesn't add any new config which is nice22:01
sdaguedhellmann: explain? just to make sure we're on the same page22:01
jeblairdhellmann: what can, eg, nova do to break working with oslo.test?22:01
*** UtahDave has joined #openstack-infra22:01
dhellmannsdague: I want oslo.messaging changes gated on the unit tests of every app that uses it, so we don't break those apps with a new release22:01
sdaguedhellmann: sure22:02
jeblair(in the delta from latest-release to master)22:02
*** sarob has joined #openstack-infra22:02
jeblairdhellmann: that part makes perfect sense to me22:02
dhellmannjeblair: if nova makes an assumption about how oslo.test works, we need to prevent changes to oslo.test that would break nova's unit tests22:02
jeblairin fact, we've talked about doing that for certain third party libraries22:02
jeblairdhellmann: so basically, from a really high level view, we're trying to identify/prevent api violations (or identify areas where the api is unclear)?22:03
dhellmanna specific example we had was neutron's tests using a private data member in a cliff class, so when cliff renamed that member their tests started breaking22:03
dhellmannso it was private, and they shouldn't have been mocking it, but they did and that caused a lot of heartburn when we cut a release months after making the change to cliff22:04
dhellmanntook us a while to figure out the cause of the issue22:04
*** om has joined #openstack-infra22:04
*** amcrn has quit IRC22:04
jeblairdhellmann: but that is much more likely to show up in the gate-oslo-on-project side, not the other way around, right?22:05
dhellmannwe have a gap now between the code being tested in an app and the master branch of any given library -- unless we cut a release on every merge, we'll always have a gap22:05
sdaguedhellmann: sure, but there is a risk profile as well22:05
dhellmannsdague: I think every time we release oslo.config it breaks someone22:06
sdagueright, but that's oslo.config releasing22:06
sdaguenot nova unit test adds22:06
jeblairaconrad: <> (it's a rather public address, no need to pm :), and thanks -- very much looking forward to to that22:06
*** sarob has quit IRC22:07
aconradjeblair: I'm actually writing a blog post22:07
aconradI'll send you the link soon22:07
dhellmannsdague: so you think we should just block lib changes on the way the apps are using them? because I'm not ok with that -- too many uses of private stuff in unit test mocks would keep us locked up in oslo22:07
dhellmannonce I change something, I don't want new tests added that use the old form22:07
sdaguedhellmann: ok, that's fair22:08
clarkbdhellmann: ++22:08
dhellmannbecause most of the devs on the other projects aren't paying attention to internals of the libs, and end up copying behavior from other places22:08
sdagueso we're really then talking about py27git22:08
jeblairdhellmann: i'm just thinking that nothing about this actually stops someone from adding use of a private member in a project22:08
clarkbjeblair: it will because it gates changes symetrically22:08
dhellmannjeblair: no, it doesn't, but at least this way I can find them before I break things with a release22:08
clarkbso you can't add them in the first place22:08
sdaguejeblair: that's true, it only stops them adding more uses after it's removed22:08
clarkber oh I see22:08
jeblairdhellmann: if someone adds a ref to oslo.messaging._foo in neutron, it will pass the proposed test22:08
clarkbya you are right22:08
clarkbit prevents code from breaking, not private access22:08
sdaguejeblair: it's the backslide issue22:08
dhellmannyes, that's right, and if I rename _foo to _bar then I'll know to go fix their bad test22:09
sdaguewhere oslo.config tries to remove something22:09
sdagueit breaks neutron22:09
sdaguethey fix neutron22:09
sdagueand remove it22:09
sdaguebut if it's not enforced on the neutron side, it might backslide22:09
sdaguebecause oslo.config isn't released with that change yet22:09
jeblairsdague, dhellmann: yes, but perhaps this happens rarely enough that in such a case you could release?22:10
sdaguejeblair: that would be another option22:10
dhellmannsure, the other thing I thought of doing was weekly alpha releases22:10
dhellmannevery monday, say22:10
*** dims has joined #openstack-infra22:10
dhellmannthat felt onerous, but maybe I can automate that22:10
sdaguedhellmann: why?22:11
dhellmannor delegate22:11
jeblairnew-project-friday and break-everything-monday. ;)22:11
dhellmannjeblair: right22:11
sdagueI mean why do alpha releases22:11
sdagueif you test git, just cut a new minor release22:11
*** prad has quit IRC22:11
sdagueI feel like in the ci world, alpha releases don't do anyone any good22:11
*** gondoi is now known as zz_gondoi22:11
sdaguebecause the machines are consuming it22:11
dhellmannsdague: markmc and I have this discussion periodically. For the oslo.* libs he started going official releases with the apps22:11
dhellmannsdague: yeah, that was another issue; I didn't know if we had alpha wheel publishing to our mirror working properly or not22:12
sdagueyes, and I think that's an idea whose time has passed22:12
sdaguebecause if we are publishing 10 libraries only at the release boundaries...22:12
sdaguestop making it 10 libraries22:12
sdaguejust call it oslo22:13
sdagueand be 1 git tree22:13
jesusauruswhy is config stuck on such an old version of the apache module?22:13
clarkbbtw we may have a fix for the posargs thing. I still think it is the wrong fix (and way over thought, why is shlex not appropriate?) because it isn't a generic fix. It is posargs specific22:13
dhellmannsdague: I need to split up the tree so I can have different expert groups managing the libs22:13
clarkbjesusaurus: because the new apache module versions are terrible and remove features22:13
*** Sukhdev__ has quit IRC22:14
sdaguedhellmann: you can't enforce that by culture?22:14
clarkbjesusaurus: with 0.4 you can provide a vhost template and do whatever you want22:14
jesusaurusclarkb: what features?22:14
clarkbjesusaurus: with new apache you are beholden to whatever the module has implemented22:14
dhellmannsdague: but the idea of just using regular releases would work ok22:14
dhellmannsdague: maybe?22:14
*** msabramo has quit IRC22:14
clarkbjesusaurus: mod rewrite stuff is the big one, we do some non trivial thing with mod rewrite22:14
openstackgerritJoshua Harlow proposed a change to openstack/requirements: Add wrapt library
sdaguedhellmann: there are parts of the nova tree I'll never +2 code on, because I don't feel confident about it. And that's all culture enforced (not rules enforced) and generally works pretty well22:15
*** mdenny has joined #openstack-infra22:15
* dhellmann nods22:15
*** MIDENN_ has quit IRC22:15
sdagueclarkb: there is a config inject in the newest on22:16
* dhellmann hears a drink calling from the other room22:16
jeblairdhellmann: cheers!22:16
clarkbsdague: oh good, maybe it is useable again22:16
*** pcrews_ has quit IRC22:16
sdagueI just switched over my linode to this this weekend, hit a few bumps, but seems ok22:17
*** zul has joined #openstack-infra22:17
clarkbI still think the first release was the bset one :)22:17
clarkbit was simple and worked22:17
sdagueon the upside, they take patches pretty quickly22:18
jeblairannegentle: sorry i was out for a bit; but mordred explained my (our i guess?) position very well i thought.  but let me know if you have any further questions.  after all, i was trying to make sure people understood how i think about things.  :)22:19
sdagueI've submitted 2 PRs, and in both cases they were merged in < 48hrs22:19
jeblairzaro: would you mind preparing an etherpad for tomorrow's infra meeting with the outstanding changes and any decisions we still need to make for gerrit 2.8?22:20
clarkbsdague: I think that may be why 0.4 was the best release :)22:20
clarkbsdague: too much merging and less thinking about useability22:20
jeblairzaro: i think we'll be focusing on that for the next couple of weeks, so it's time to make sure we have everything in order22:20
clarkbwhich is a nice to contributors but not necessary a good thing overall22:20
sdagueyeh, the current one is definitely a world view22:20
mnaserthis place is sure as hell more active than #openstack-dev :p22:20
*** sarob has joined #openstack-infra22:21
sdagueI thought about going my own here, but felt useful to understand it's quirks22:21
*** zul has quit IRC22:21
jeblairmnaser: oh hi!  i saw you took that nova bug :)22:21
*** zul has joined #openstack-infra22:21
mnaserjeblair: hey, actually i got most of the patch done, just need to fix the tests :)22:21
mnaseradded a reported_at field which is relied on instead of updated_at22:21
jeblairmnaser: and i did some fun nova debugging during a devstack walkthrough at pycon22:22
*** slagle has quit IRC22:22
mnaserheh, yeah, the patch turned out to be huge so far lol22:22
clarkbsdague: I think a lot of puppetmodules go overboard in making every little thing expressed in puppet DSL which means you lag in actual features22:22
jeblairi now think that all devstack walkthroughs should have lab equipment that kernel panics just to keep things interesting ;)22:22
mnaserthe bigger the changes, the more worried i get from git review22:22
clarkbsdague: it certainly makes things very puppety but presumably if I need apache I can edit a vhost file22:23
sdagueclarkb: I agree. There are actually some goofy wrong values in here because of it22:23
*** praneshp has quit IRC22:23
jeblairmnaser: one thing you can do is to keep a local branch with lots of little commits and then squash that for git review; that way if it gets out of hand, you still have incremental history22:23
*** slagle has joined #openstack-infra22:23
mnaserjeblair: i'm not as worried about that as much as i'm worried about being told "well this doesn't look like, do it this way" :p22:24
clarkbjeblair: you can install kernel drivers to induce kernel panics :) should be easy to set that up22:24
sdaguejeblair: nice. What was the context of the devstack walkthrough?22:24
*** zul has quit IRC22:24
*** chuck_ has joined #openstack-infra22:25
*** praneshp has joined #openstack-infra22:25
*** sarob has quit IRC22:26
*** rlandy has quit IRC22:26
mnaserjeblair: - far from a "trivial" fix22:26
mnaser(still not done, need more work as some tests are failing because of the extra field added)22:27
jeblairsdague: the rax user exp folks did sort of a drop-in lab environment.  most people were in the 'new to openstack' camp, a few were ready to hack on something so hands-on with devstack was helpful for them22:27
*** jergerber has quit IRC22:27
*** mrmartin has joined #openstack-infra22:27
*** denis_makogon has quit IRC22:27
*** om has quit IRC22:28
mnasersdague: we actually discovered a bug in nova in that small 2 hours22:28
zarojeblair: will do.22:28
jeblairmnaser: yeah, it changes a few things ;)  but from what i've seen (i'm far from a nova expert here), the approach of separating the record update time from downtime detection sounds like it deserves serious consideration22:29
jeblairmnaser runs a cloud provider in montreal22:30
*** pcrews_ has joined #openstack-infra22:30
mnaserjeblair: the point of updated_at is to give the time the record was last updated, not the time it last reported22:30
mnaserprobably one of the biggest changes i've written so let's hope it goes through reviews well, heh (once it's all done)22:30
jeblairmnaser: yep22:31
*** marun has quit IRC22:31
mnasergot busy with other stuff but i'll hack at it later tonight and finish it hopefully22:31
jeblaircool, nice to see you around here! :)22:32
mnaserlikewise! :)22:32
*** chuck_ is now known as zul22:34
sdaguejeblair: you have a moment to ponder the branchless tempest implications?22:35
jeblairsdague: i have been doing that for the past few minutes in fact! :)22:37
*** sarob has joined #openstack-infra22:37
jeblairsdague: just read the spec; agree with everything and anticipate no issues.  i'm worried that means i'm tired.  :)22:37
jeblairi mean, not no issues, but you know, no disagreements on approach or anything...  of course there will be issues....22:38
jeblairsdague: one thing that comes to mind...22:38
openstackgerritClark Boylan proposed a change to openstack-infra/config: Make jenkins proposal jobs use dedicated user.
jeblairsdague: is that we should try to keep the feature-flag enabling logic nice and modular/compartmentalized in devstack-gate22:38
*** gokrokve has quit IRC22:38
clarkbjeblair: ^ I apparently missed a comma when I did the first pass of that chang22:38
*** derekh has joined #openstack-infra22:39
clarkbjeblair: everything outside of puppet is ready for it, any chance I can get a rereview?22:39
sdaguejeblair: right, that's actually what I wanted to hit you up about22:39
jeblairsdague: since it's currently doing double-duty as "way to set up lots of projects together" and "way to configure devstack/tempest"22:39
sdaguebecause feature-flag logic is going to include *every* extension (82 extensions in nova as of icehouse)22:39
jeblairi think the long term evolution of that is something like hashar's setup script, but it's not quite ready yet22:39
sdaguewhich means I think we need to move that logic out of bash22:39
clarkbsdague: api extensions?22:39
sdagueclarkb: yes22:39
sdagueapi extensions, by definition, are optional22:40
jeblairsdague: non-bash sounds fine for me22:40
*** Ryan_Lane has joined #openstack-infra22:41
sdaguejeblair: ok, so I had this idea for having a yaml definition of the matrix which we could process in python into a string that we get back into the existing d-g scripts22:41
sdagueall in the d-g repo22:41
jeblairsdague: what happens with that string?22:41
sdagueand a set of vars for the extensions lists22:41
*** reed has quit IRC22:41
sdaguethat pass through to tempest22:42
*** adalbas has quit IRC22:42
jeblairsdague: ok.  should we think about having this script do more of what's already there?  (localrc generation, etc?)22:42
sdaguejeblair: that's possible22:42
sdagueI mostly wanted to solve the more immediate problem first22:42
sdagueso I was going for taking over the minimum amount of function22:43
jeblairsdague: okay.  that sounds good, and i'm open to more of that.  i suspect the long term evolution of d-g is that we move the 'set up git repos' part out of it and into a general purpose tool, and what's in d-g shrinks back down to "set up different ways of running tempest"22:43
jeblairand i think all this fits into that well22:43
sdagueyep, that sounds good22:43
jeblairsdague: +122:44
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard: Hard delte for tasks
jeblairsdague: and thanks!22:44
sdaguethe main reason we'll have to generate an explicit extension list is the feature flag for tempest for compute_extensions defaults to 'all'22:44
*** sabari2 has joined #openstack-infra22:44
sdaguebut the moment nova wants to test a new extension that's juno only, we'll actually have to make stable/icehouse default to the list (- that extension)22:45
*** mrodden has quit IRC22:45
jeblairsdague: it might be worth considering how this affects the "run" base case -- will running devstack with pure defaults still be tested adequately?22:45
clarkbI was thinking taht the flag could be more optparsy too to be less hackish22:45
clarkbthen instead of evaluating some env vars one at a time optparse could set the flags and we could make smarter choices22:46
sdaguejeblair: that's a good question. it would only impact stable/* branches of devstack22:46
*** gokrokve has joined #openstack-infra22:46
sdagueit is possible that we could handle this there instead22:47
sdaguedoing it in d-g might actually solve things like cells testing22:47
sdaguebecause we could just make a much shorter api extension list22:47
sdaguewhich it should pass22:47
sdaguebecause there are only supposed to be a few extensions that it can't work with22:48
jeblairclarkb: the env vars are getting a wee bit out of hand... :)22:48
clarkboptparse won't fix all of the problems with this but may make it a bit more clear and maintainable22:48
*** aconrad has quit IRC22:48
*** gokrokve has quit IRC22:48
*** dizquierdo has quit IRC22:48
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard: Collate story metadata for status
*** nati_ueno has quit IRC22:49
derekhAnybody know why that first 5 jobs in the check-tripleo queue are still there? looks to me like they finished ages ago22:49
sdaguemy thinking about the yaml matrix is to let options both add and *remove* items. So we could start with base cases and go both directions22:49
*** nati_ueno has joined #openstack-infra22:49
clarkbderekh: the tripleo-f20 job is still runnin22:49
sdaguelike defining base case with n-net enabled, and neutron case adds all the q-* services and removes n-net22:49
sdaguewhich would help on some of the else logic22:49
sdagueanyway, I'll provide a sketch22:49
sdagueand we'll review from there22:49
jeblairsdague: yeah that sounds good22:50
clarkbderekh: oh huh, finished 11 hours ao22:50
derekhclarkb: not acording to jenkins22:50
clarkband zuul didn't see it22:50
clarkbjeblair: ^22:50
* clarkb finds logs22:50
sdagueand it's dinner time, so see folks in the morning22:50
*** wenlock has quit IRC22:50
jeblairsdague: bon appetit22:50
jeblairi'm not sure "finished" is the best way to describe that job's current state22:52
jeblairthe console log ends with:22:52
jeblair2014-04-14 12:20:55.547 | Waiting for the overcloud stack to be ready22:52
jeblair2014-04-14 12:20:55.586 | + wait_for_stack_ready 300 10 overcloud22:52
jeblairand lacks the normal jenkins job result boilerplate22:53
jeblairclarkb: so there are no logs from the jenkins master22:54
jeblairi think they may be owned by the wrong user22:54
jeblairwhich may be the result of some previous jenkins deb package fubaring22:55
clarkbjeblair: jenkins.o.o had log issues too, I restarted it and it became happy22:55
*** jerryz has joined #openstack-infra22:55
clarkbbut the log permissions on jenkins.o.o looked fine. I figured it may have been rlated to loading configs off of disk22:56
clarkbbut it may be the upgrade fubar too22:56
jeblairjenkins06 is running no jobs and has no active slaves22:56
clarkboh huh, it must be really wedged then22:56
clarkbit has a lot of offline slaves22:56
*** sarob has quit IRC22:57
jeblairseems it's the only one22:57
jeblairi will stop it, remove all slaves from config, then start22:57
*** dims has quit IRC22:57
clarkbfwiw zuul didn't really get anything new from that test on 86315,222:57
clarkbjeblair: ok22:57
clarkbso zuul's logs are quiet too22:57
*** sarob has joined #openstack-infra22:57
*** asettle is now known as alex-afk22:57
jeblairclarkb: yeah, i suspect everything is still connected and idle, waiting for something to happen on 06; but hard to diagnose what without logs.22:57
*** adalbas has joined #openstack-infra22:57
jeblairwe could try to look at stacktraces, but i don't want to without logs22:58
clarkbit is a lot of noise without logs22:58
jeblairi expect that to cause something to happen with the jobs in zuul22:58
jeblaireither failing or restarting22:58
jeblairbut apparently not22:59
*** yamahata has quit IRC23:01
*** sarob has quit IRC23:01
jeblairwe might need to check on tcp keepalives in gear23:02
*** mbacchi has quit IRC23:02
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard: Collate story metadata for status
*** doddstack has joined #openstack-infra23:03
jeblairclarkb: i'll restart zuul to clean that up, then re-enqueue23:04
jeblairsound good?23:04
*** thedodd has quit IRC23:04
jeblairgood.  :)23:05
*** jamielennox|away is now known as jamielennox23:06
*** mrodden1 has joined #openstack-infra23:06
jeblairrestarting jenkins0623:06
clarkbjeblair: yes sounds good23:06
jeblairwhich is now logging23:06
clarkbI have to say finding the login page for the openstack foundation is hard23:07
clarkbbut its under 'profile' not foundation23:07
*** melwitt has quit IRC23:07
*** doddstack has quit IRC23:08
*** ameade has joined #openstack-infra23:09
*** mwagner_lap has joined #openstack-infra23:11
*** bknudson has quit IRC23:12
openstackgerritClark Boylan proposed a change to openstack-infra/config: Run check-tripleo jobs on stale changes
clarkbI am going to afk soon23:14
clarkbjeblair: need me to help with any of the jenkins stuff?23:14
jeblairclarkb: all done; no worries23:14
jeblairderekh: ^23:15
clarkbthe sun is being sunny and I need to run more errands23:15
openstackgerritA change was merged to openstack-infra/config: Reduce number of files send for propose_translation
clarkbbut hey the spot on the back of my eye isn't melanoma so bring on the sun \o/23:15
jeblairmordred: your run puppet over ssh has 2x+2s; you want to merge that yourself when you're ready to take care of it?23:16
*** lcostantino has quit IRC23:16
jeblairclarkb: yay it's not melanoma!  at least, not yet!  :)23:16
clarkbapparently sun exposure isn't correlated to melanoma of the eyeball so shouldn't worry about the sun at all in that way23:17
*** dcramer_ has joined #openstack-infra23:17
derekhjeblair: clarkb  thanks23:18
*** sabari2 has quit IRC23:18
*** dcramer_ has left #openstack-infra23:19
*** dcramer_ has joined #openstack-infra23:19
*** gokrokve has joined #openstack-infra23:21
*** nati_ueno has quit IRC23:21
*** dims has joined #openstack-infra23:22
nibalizermordred: i stll don't understand why the updatepuppetmaster and run the world cron is a 15 minute cron with a random splay23:22
*** nati_ueno has joined #openstack-infra23:22
nibalizercan you explain that?23:22
jheskethclarkb: if you have a chance to look at this review I'd greatly appreciate it:
*** mrmartin has quit IRC23:23
clarkbjhesketh: looking23:23
jheskethyou, sir, are a good person :-)23:23
*** flaper87 is now known as flaper87|afk23:23
clarkbhrm do we know if fungi swapped those creds yet?23:24
*** mrodden has joined #openstack-infra23:24
*** alex-afk is now known as asettle23:24
jerryzHi folks, what do i need to do to get a stackforge project monitored by jeepyb hook that integrates with launchpad?23:25
*** atiwari has quit IRC23:26
*** pdmars_ has quit IRC23:26
*** mrodden1 has quit IRC23:27
clarkbjhesketh: doesn't look like fungi updated teh swift creds, I am going to +2 it but not approve until someone can swap out those credentials23:27
clarkbjhesketh: I need to run errands now but maybe jeblair can do that?23:28
jheskethclarkb: okay, thanks :-)23:28
clarkbanyways +2 with that one todo item23:29
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard: Added status filter to story API
*** blamar has quit IRC23:31
*** MarkAtwood has quit IRC23:32
*** wenlock has joined #openstack-infra23:33
jerryzI checked the jeepyb code , i believe the hook applies to every project as long as it is on review.o.o. but when i referenced a bug number in my commit message, comment didn't show up in launchpad23:37
*** nati_ueno has quit IRC23:40
clarkbjerryz does the project name on launchpad match the name on review.o.o? and you may need to add an account to the projects bug drivers group on lp23:41
*** nati_ueno has joined #openstack-infra23:42
clarkbI always forget which account it is but you can check other projects23:42
jerryzclarkb: the project does not match precisely with the one on launchpad23:42
jerryzclarkb: there are 2 sub projects on review.o.o and the prefix is the same23:42
clarkbin that case you need to map them in the config projects.yaml file23:43
jerryzclarkb: thank you. i will take a look23:44
*** ianw has joined #openstack-infra23:47
*** ameade has quit IRC23:47
jerryzclarkb: do you mean the group field in review.projects.yaml?23:47
*** yamahata has joined #openstack-infra23:47
*** moted has quit IRC23:48
*** arborism has joined #openstack-infra23:49
*** vhoward has joined #openstack-infra23:51
*** moted has joined #openstack-infra23:52
*** ameade has joined #openstack-infra23:55
*** ameade has quit IRC23:55
*** yamahata has quit IRC23:56
*** Alexandra_ has joined #openstack-infra23:56
*** asettle has quit IRC23:56
*** Alexandra_ has quit IRC23:57
*** asettle has joined #openstack-infra23:57
*** arborism is now known as amcrn23:58
*** wchrisj has quit IRC23:59

Generated by 2.14.0 by Marius Gedminas - find it at!