Wednesday, 2014-09-10

jeblairjaypipes: lgtm, cool, thanks!00:00
*** oomichi has quit IRC00:00
clarkbfungi: will do00:00
*** nelsnelson has quit IRC00:01
jaypipesjeblair: merged. thank YOU!00:01
*** yamahata has quit IRC00:01
*** MaxV has quit IRC00:04
*** david-lyle has quit IRC00:05
*** todd_dsm has quit IRC00:05
jeblairso i think the puppet changes are ready to merge, but i'm not going to be around this evening, so i'm not going to aprv right now00:05
jeblairclarkb, nibalizer: ^00:06
clarkbjeblair: ok probbaly something to do first thign tomorrow00:06
*** bhuvan_ has quit IRC00:07
nibalizerim willing to be on for the next few hours if someone wants to punch it00:07
nibalizerbut tomorrow is fine00:07
*** gokrokve has quit IRC00:10
*** baohua has quit IRC00:11
*** bhuvan has joined #openstack-infra00:11
*** sdake_ has quit IRC00:12
jogoAlex_Gaynor: ping00:15
*** tomoe has joined #openstack-infra00:16
jogoAlex_Gaynor re:
*** camunoz has joined #openstack-infra00:17
*** prad has joined #openstack-infra00:17
*** gyee is now known as zuul00:18
*** zuul is now known as gyee00:18
*** prad has quit IRC00:19
*** signed8bit has joined #openstack-infra00:23
*** signed8bit is now known as signed8bit_ZZZzz00:23
openstackgerritAdam Gandelman proposed a change to openstack-infra/config: Adds support for Ironic parallel Tempest testing
*** yamamoto has joined #openstack-infra00:23
*** sdake_ has joined #openstack-infra00:24
*** signed8bit_ZZZzz has quit IRC00:24
openstackgerritA change was merged to openstack-infra/config: Install Zuul onto workers
adam_gany devstack-gate cores around to take a look at getting this into the gate while the night is young?
*** Longgeek has quit IRC00:26
*** yamamoto has quit IRC00:28
*** zz_dimtruck is now known as dimtruck00:31
*** esker has joined #openstack-infra00:32
*** unicell has joined #openstack-infra00:33
*** tomoe has quit IRC00:34
*** tomoe has joined #openstack-infra00:34
*** daya_k has quit IRC00:34
*** sdake_ has quit IRC00:34
*** esker has quit IRC00:38
jeblairadam_g: why is 3 not the default in devstack?00:38
*** dangers is now known as dangers_away00:39
*** tomoe has quit IRC00:40
adam_gjeblair, its meant to be tunable based on system resources, with the default = 100:40
jeblairadam_g: what happens if we run the smoke tests or javelin on the base side?00:42
jeblairadam_g: is marked as in-progress, but it isn't really, is it?00:43
uvirtbotLaunchpad bug 1365105 in devstack "grenade fails to upgrade keystone when mod wsgi is enabled" [Undecided,In progress]00:43
adam_gjeblair, the base side is deployed with virt_driver=fake, so tempest will fail00:43
*** stevemar has joined #openstack-infra00:43
jeblairadam_g: okay.  what about javelin?  i thought it did some setup that was used later (eg, ensuring that tenants/resources still exist)00:44
*** Guest39956 has quit IRC00:46
*** russellb has quit IRC00:46
adam_gjeblair, its a devstack bug, in progress @
*** tomoe has joined #openstack-infra00:47
*** wendar has quit IRC00:47
*** wendar has joined #openstack-infra00:47
jeblairadam_g: ok, that didn't get linked in lp for some reason00:47
*** dimtruck has quit IRC00:47
*** tomoe has quit IRC00:48
adam_gjeblair, it does create some stuff but 1) doesnt yet have the smarts to only create the things that will actually work on base 2) it does not work when neutron is enabled, which is required in the devstack+ironic case00:48
adam_git being the new javelin2 in tempest00:48
*** tomoe has joined #openstack-infra00:48
*** zeus has joined #openstack-infra00:48
*** zeus is now known as Guest319200:48
Alex_Gaynorjogo: pong00:49
*** russellb has joined #openstack-infra00:49
*** dimtruck has joined #openstack-infra00:49
jeblairadam_g: thanks.  lgtm.  i'd like sdague to see it though, so not aprving.00:49
*** wenlock has joined #openstack-infra00:49
jeblairsdague: 116761 ^00:49
adam_gjeblair, thanks, appreicated00:50
*** bdpayne has quit IRC00:50
*** gokrokve has joined #openstack-infra00:50
*** maxbit has joined #openstack-infra00:52
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard: Event actor now added to message payload
openstackgerritEmilien Macchi proposed a change to openstack-infra/config: Introduce tripleo-puppet-elements into Stackforge
*** tomoe has quit IRC00:54
*** yamahata has joined #openstack-infra00:54
*** gyee has quit IRC00:55
*** torandu has quit IRC01:01
*** yaguang has joined #openstack-infra01:02
*** tomoe has joined #openstack-infra01:02
*** signed8bit has joined #openstack-infra01:02
*** signed8bit is now known as signed8bit_ZZZzz01:02
*** signed8bit_ZZZzz has quit IRC01:03
*** markmcclain has quit IRC01:04
*** SumitNaiksatam has quit IRC01:04
*** maxbit has quit IRC01:09
*** dims_ has joined #openstack-infra01:14
*** hdd has joined #openstack-infra01:14
openstackgerritJamie Lennox proposed a change to openstack-infra/config: Add keystoneclient-kerberos repository
anteayalifeless: so this repo names tripleo-release and diskimage-builder-core in its acl file:
lifelessanteaya: yes01:17
anteayalifeless: first question, are you aware and are you okay with this?01:17
lifelessbroadly, yes01:18
anteayaokay good enough01:18
anteayanext and maybe this is an EmilienM question, why does this repo need to create proposed branches01:18
anteayaor have permissions on stable branches?01:18
anteayabecause if it doesn't we could get rid of lines 6 - 17 in the above acl file01:19
*** yamamoto_ has joined #openstack-infra01:23
*** hdd has quit IRC01:24
*** armax_ has quit IRC01:24
*** todd_dsm has joined #openstack-infra01:24
*** changbl has joined #openstack-infra01:27
*** yamamoto_ has quit IRC01:28
*** baohua has joined #openstack-infra01:29
jogoAlex_Gaynor: open question in
*** bhuvan has quit IRC01:29
jogoAlex_Gaynor: that I thought maybe you can answer01:30
jogoAlex_Gaynor: see the Note in the commit message01:30
Alex_Gaynorjogo: I don't recognize that issue, dstufft is more likely than I to know01:30
jogoAlex_Gaynor: thanks01:31
Alex_Gaynorsorry I'm not more helpful01:31
*** Ryan_Lane has quit IRC01:32
*** amotoki has joined #openstack-infra01:33
*** dmsimard_away is now known as dmsimard01:33
*** Daisy has joined #openstack-infra01:33
dstufftI don't recognize that issue either01:35
*** oomichi has joined #openstack-infra01:35
jogodstufft: thanks, fungi ^01:35
*** yamamoto_ has joined #openstack-infra01:36
lifelesssdague: next time it times out01:36
lifelesssdague: can you get the OOPS id please? its the ticket for getting a detailed report01:37
lifeless(same as our request id)01:37
jogofungi: so it looks like your concern about pip cache may not be an issue01:37
lifelessX-Lazr-OOPSID IIRC01:37
*** amotoki has quit IRC01:37
clarkbjogo: its easy to test. pip install foo==oldversion, pip install foo01:38
jogofungi: in which case caching some pip stuff in the images won't break things (although its unclear to me if  that is the best route to take to fix things)01:38
clarkbjogo: check what version you get01:38
jogoclarkb: you mean pip install foo==oldversion, check if file is cached in a new venv, pip install foo>oldversion01:39
clarkbjogo: sure01:39
clarkbjogo: actually no01:39
clarkbdon't specify >oldversion01:39
*** dmsimard is now known as dmsimard_away01:39
clarkbthat should wrok01:39
clarkbthe issue is whether or not pip install foo gets you old version or new version01:39
*** gokrokve_ has joined #openstack-infra01:40
jogoclarkb: ahh so if there is an old version taht does satisfy the reqs but it  isn't the newest01:40
*** wenlock has quit IRC01:40
*** gokrokve has quit IRC01:41
jogoclarkb: lets say that in that test, old is installed and not new. Is that an issue?01:41
clarkbjogo: it may be. I think you just want old cached01:41
clarkbso you will probably need two venvs01:41
jogoclarkb: so lets say we do this with a wheelhouse01:42
lifelessthe precondition is01:42
*** JadnG has joined #openstack-infra01:42
lifelessold cached01:42
lifelessnew not cached01:42
lifelessinstall 'thing'01:42
lifelesscheck new is installed.01:42
jogolifeless: right. The question is if old is installed, is that a problem?01:42
clarkbyes old should not be installed01:42
lifelessbefore you start? Yes, its a problem01:42
clarkbotherwise you get into the pip upgrade logic01:43
jogolifeless clarkb: if we pre-populate the wheelhouse pip download dir on image build01:43
jogoclarkb: ahh01:43
clarkbwhich is likely going to do the thing you want, but uninstalled first it may not01:43
lifelessyou can pip install foo==old;pip uninstall foo; pip install foo.01:43
clarkbalso don't use wheels01:43
lifelesswheels need to be tested separately I suspect01:43
*** rmcall has quit IRC01:47
*** hdd has joined #openstack-infra01:47
*** gokrokve has joined #openstack-infra01:48
*** gokrokve_ has quit IRC01:48
*** gus has quit IRC01:49
*** mattoliverau has quit IRC01:49
*** mattoliverau has joined #openstack-infra01:50
*** gus has joined #openstack-infra01:51
anteayajamielennox: hello01:53
jamielennoxanteaya: hi01:53
anteayajamielennox: rather than create a new acl file for kerberos, how would you feel about using the middleware acl file?
jamielennoxanteaya: keystoneclient probably makes more sense - but sure i can do that01:54
anteayaI highly doubt you need to create proposed branches for this new project01:54
pleia2hm, how do you quit gertty?01:54
anteayaI don't see a keystoneclient.config01:54
anteayaI'm guessing it uses something else01:54
*** nosnos has joined #openstack-infra01:54
*** amcrn has quit IRC01:55
*** chenglch has joined #openstack-infra01:56
anteayajamielennox: there it is
jogohmm the cached file list is odd01:56
jogowill have to resume this later01:56
anteayabut it too has permissions for a proposed branch which is not necessary for a client repo01:56
jamielennoxanteaya: ah, was looking - i figured it must be there somewhere01:56
anteayathe python- at the front01:56
clarkbpleia2: ^Q01:57
anteayaI always just ignore that mentally01:57
pleia2clarkb: aha, thanks :)01:57
anteayajamielennox: so yeah, let's just point to that file for simplicity01:57
jamielennoxanteaya: so is it a mistake to have a proposed branch for keystoneclient?01:57
anteayajamielennox: not a mistake01:58
anteayabut it isn't necessary01:58
anteayayou don't do release candidates for clients01:58
*** todd_dsm has quit IRC01:58
anteayaso you don't need the abiltiy to create proposed branches01:58
anteayaproposed branches are specifically for our release workflow:
anteayaso the services need permissions to create proposed branches01:59
anteayabut clients and other repos don't01:59
anteayajamielennox: probably what happened is the acl file got copy/pasted from the keystone acl file02:00
anteayavery common02:00
openstackgerritJamie Lennox proposed a change to openstack-infra/config: Add keystoneclient-kerberos repository
jamielennoxmakese sense, hows that ^ ?02:01
* anteaya looks02:01
*** r-daneel has quit IRC02:02
anteayalooks good for the acl file02:02
anteayais python-keystoneclient-kerberos registered on pypi?02:02
*** fifieldt_ is now known as fifieldt02:03
jamielennoxanteaya: not yet, i was going to wait until the infra review went through02:03
jamielennoxthere's a governance one to file as well02:03
anteayawe need to have the project registered on pypi for the infra patch to be approved02:04
jamielennoxanteaya: ok, i'll register that now02:04
anteayajamielennox: you don't need to have a release there yet, just claim the namespace02:04
morganfainberganteaya, ++ thanks for looking at the patch!02:04
morganfainberganteaya, :)02:04
anteayamorganfainberg: of course :D02:04
morganfainberganteaya, also Hi!!!02:04
* morganfainberg wavers02:04
anteayahello there02:04
* anteaya wavers back02:04
anteayamorganfainberg: haven't talked to you in a long while02:05
morganfainberganteaya, it's been kinda busy!02:05
anteayano kidding02:05
anteayahow you doing?02:05
morganfainbergnot bad02:05
anteayaglad to hear it02:05
morganfainbergtrying to figure out travel to paris.02:05
anteayaif I could, I would sail02:06
morganfainbergoooh i like it02:06
jamielennoxanteaya: done:
morganfainbergprobably be a rough sail that time of year02:06
anteayamight never make it to Paris but meh02:06
anteayamorganfainberg: no kidding02:06
anteayajamielennox: thanks02:06
jamielennoxit's scary how little prompting there is with setuptools and pypi02:06
morganfainbergjamielennox, hah i register through the web page02:06
morganfainbergjamielennox, less questions02:06
jamielennoxwell i had the cookiecutter project done02:07
morganfainbergor uncertainties02:07
jamielennoxthen python register02:07
jamielennoxno questions, no are you sure you want to regisetr a new public project02:07
morganfainbergyeah i still prob. would have gone with web page02:07
*** yamamoto_ has quit IRC02:07
jamielennoxjust 200 OK done02:07
clarkbdon't use register :) it http's too02:07
clarkbnot https02:07
clarkbregister is just bad02:07
jamielennoxah - it's listed as the number 1 option on the pypi site02:07
morganfainbergjamielennox, for that patch add the gerrit bot to the -keystone channel please (if you don't mind)02:08
morganfainbergjamielennox, for the new repo02:08
anteayawell it is good to go as is, yay02:08
anteayaand yeah if you want the bot reporting add that too02:09
morganfainbergi mean we can add that separately too02:09
morganfainbergbut if we're waiting for approvals... might as well add it right?02:09
anteayaadd it now, yes02:09
anteayafor new projects jeblair waits for me to +102:09
anteayaas long as I am keeping up with patch reviews02:10
morganfainbergoh should gate-python-keystoneclient-kerberos-python33 that be python34?02:10
anteayawell I don't know02:10
openstackgerritJamie Lennox proposed a change to openstack-infra/config: Add keystoneclient-kerberos repository
anteayaI wondered that myself02:10
morganfainbergclarkb, ^02:10
anteayabut I don't think we are ready for 34 program wide yet02:10
anteayafungi has patches up02:10
jamielennoxthe python3-jobs task in layout only has 3302:11
morganfainberganteaya, right - figure we might as well ask prior to needing *another* patch just to change that02:11
clarkbif it works with py34 then sure02:11
anteayajamielennox: you added python-keystone-middleware02:11
clarkbI think new things should just start with py34 if possible otherwise fungi's job becomes impossible02:11
jamielennoxclarkb: there's nothing there02:11
clarkbjamielennox: then yes02:11
morganfainbergclarkb, expirimental so they can trigger it when they're working on it02:11
morganfainbergclarkb, but not consume resources prior to it being actively worked on for py2402:12
morganfainbergerm 3402:12
jamielennoxanteaya: oops :)02:12
jamielennoxso add 34?02:12
jamielennoxor change 33 -> 34?02:12
morganfainbergjamielennox, it should be 34 vs 33 since it's new02:12
morganfainbergjamielennox, if possible02:12
anteayayes if possible02:13
morganfainbergotherwise it'll just need to be changed to py34 *very soon now*02:13
anteayabut I don't know where we are with 3402:13
* morganfainberg nods02:13
openstackgerritJamie Lennox proposed a change to openstack-infra/config: Add keystoneclient-kerberos repository
jamielennoxit doesn't matter so much as it's the experimental job02:13
jamielennoxi made it 34 anyway, we can do a check afterwards and see if it works02:14
*** patrickeast has quit IRC02:14
anteayaokay +1 again02:15
anteayathanks jamielennox02:15
jamielennoxanteaya: thank you02:15
*** jyuso has joined #openstack-infra02:16
anteayaokay EmilienM I am just going to add some comments to your patch and then turn in02:16
anteayahello jyuso02:16
anteayahow are you doing now?02:16
*** jab has quit IRC02:17
*** baohua_ has joined #openstack-infra02:18
*** hdd has quit IRC02:19
*** marcoemorais has quit IRC02:20
*** jab has joined #openstack-infra02:21
*** jab has joined #openstack-infra02:21
*** baohua has quit IRC02:21
*** camunoz has quit IRC02:22
anteayaokay I'm done02:24
anteayareminder I'm not around tomorrow02:24
*** harlowja is now known as harlowja_away02:25
*** SumitNaiksatam has joined #openstack-infra02:33
*** pcrews has quit IRC02:36
*** camunoz has joined #openstack-infra02:38
*** jab has quit IRC02:40
*** arnaud__ has joined #openstack-infra02:40
*** koolhead17 has joined #openstack-infra02:44
*** koolhead17 has joined #openstack-infra02:44
*** _nadya_ has joined #openstack-infra02:46
jyusoanteaya: hi,thx,I'm still debugging the trigger issue:)02:47
*** adalbas has quit IRC02:48
*** wenlock has joined #openstack-infra02:49
*** bswartz has quit IRC02:49
*** bswartz has joined #openstack-infra02:49
*** jpich has quit IRC02:52
lifelessmordred: sdague: info on LP02:52
lifelessmordred: sdague: there was an operational issue- db locks02:52
lifeless14:50 < wgrant> 1. I62602.0 I2 ISQL-main-master I02:52
lifelessthats 62 seconds updating a bug row.02:52
*** yamamoto_ has joined #openstack-infra02:53
lifelessone theory for the cause is the decontention process had died (all writes to some key denormalised data structures are journaled rather than done directly, for liveness...but the journal has to be queried on every read, so there's a thing to aggressively convert the journal to regular rows02:54
*** jpich has joined #openstack-infra02:59
*** KanagarajM has joined #openstack-infra03:00
*** dims_ has quit IRC03:01
*** dims_ has joined #openstack-infra03:01
*** dimtruck is now known as zz_dimtruck03:05
*** jab has joined #openstack-infra03:07
*** jab has quit IRC03:07
*** jab has joined #openstack-infra03:07
*** itzikb1 has quit IRC03:11
*** praneshp has joined #openstack-infra03:11
*** dims_ has quit IRC03:14
*** dims_ has joined #openstack-infra03:15
*** rfolco has quit IRC03:16
*** dims_ has quit IRC03:19
*** weshay has quit IRC03:25
*** otter768 has quit IRC03:28
*** hdd has joined #openstack-infra03:29
*** camunoz has quit IRC03:33
openstackgerritDavid Stanek proposed a change to openstack-dev/pbr: Adds option for excluding files from autodoc trees
*** amotoki has joined #openstack-infra03:34
*** baohua_ has quit IRC03:34
*** baohua has joined #openstack-infra03:35
*** praneshp_ has joined #openstack-infra03:37
*** amotoki has quit IRC03:38
*** praneshp has quit IRC03:39
*** praneshp_ is now known as praneshp03:39
*** signed8bit has joined #openstack-infra03:43
*** signed8bit has quit IRC03:43
*** mbacchi has quit IRC03:44
*** mbacchi has joined #openstack-infra03:44
*** _nadya_ has quit IRC03:49
*** camunoz has joined #openstack-infra03:50
*** mbacchi has quit IRC03:53
*** mbacchi has joined #openstack-infra03:54
*** praneshp_ has joined #openstack-infra03:56
*** praneshp has quit IRC03:57
*** praneshp has joined #openstack-infra03:59
*** praneshp_ has quit IRC04:01
*** lttrl has quit IRC04:06
*** daya_k has joined #openstack-infra04:08
*** arosen1 has joined #openstack-infra04:13
arosen1Hey, I was thinking. When patches enter the gate we don't really need to rerun the pep8 tests since we already ran them in the check queue? Seems like we can save a few resources by removing this test run.04:14
*** Daisy has quit IRC04:22
*** ppai has joined #openstack-infra04:27
*** rmcall has joined #openstack-infra04:31
*** melwitt has joined #openstack-infra04:35
*** arnaud__ has quit IRC04:35
*** yamahata has quit IRC04:37
*** rushiagr_away is now known as rushiagr04:40
*** arnaud__ has joined #openstack-infra04:42
*** melwitt has quit IRC04:43
*** bdpayne has joined #openstack-infra04:46
*** bdpayne_ has joined #openstack-infra04:49
*** reed has quit IRC04:50
openstackgerritIan Wienand proposed a change to openstack-infra/config: Small wsgi application to download tarball of logs
*** bdpayne has quit IRC04:53
*** ildikov has quit IRC04:54
*** mmaglana has quit IRC04:56
*** vhoward has joined #openstack-infra04:58
*** mbacchi has quit IRC05:02
*** baoli has quit IRC05:03
*** JadnG has quit IRC05:05
*** harlowja_at_home has joined #openstack-infra05:06
jogoclarkb lifeless: I confirmed, if you have download cache set it still installs new05:10
jogoclarkb lifeless: (playing with docker)05:11
*** amotoki has joined #openstack-infra05:11
jogo(pip that is)05:13
*** arnaud__ has quit IRC05:14
*** arnaud__ has joined #openstack-infra05:15
*** julim has quit IRC05:17
*** mrmartin has joined #openstack-infra05:17
*** ildikov has joined #openstack-infra05:24
*** arosen1 has quit IRC05:25
*** afazekas_ has joined #openstack-infra05:27
*** cipcosma has joined #openstack-infra05:27
*** mrmartin has quit IRC05:30
*** arnaud__ has quit IRC05:32
*** HeOS has quit IRC05:32
*** amotoki_ has joined #openstack-infra05:34
*** koolhead17 has quit IRC05:37
*** harlowja_at_home has quit IRC05:38
*** yfried__ has joined #openstack-infra05:38
*** hdd has quit IRC05:39
*** jcoufal has joined #openstack-infra05:39
*** amotoki_ has quit IRC05:40
*** k4n0 has joined #openstack-infra05:45
*** stevemar has quit IRC05:49
*** homeless_ has quit IRC05:51
*** Daisy has joined #openstack-infra05:51
*** rlandy has joined #openstack-infra05:53
*** baohua has quit IRC05:53
*** emagana has joined #openstack-infra05:58
*** camunoz has quit IRC06:01
*** loki184 has joined #openstack-infra06:03
*** chenglch|3 has joined #openstack-infra06:03
*** arnaud__ has joined #openstack-infra06:03
*** chenglch has quit IRC06:03
*** yolanda has quit IRC06:04
*** wenlock has quit IRC06:08
openstackgerritDavid Stanek proposed a change to openstack-dev/pbr: Fixes a spelling error
*** camunoz has joined #openstack-infra06:12
*** _buzz has quit IRC06:19
*** jamespage_ has joined #openstack-infra06:21
*** _buzz has joined #openstack-infra06:21
*** ildikov_ has joined #openstack-infra06:22
*** sc68cal has quit IRC06:22
*** ildikov has quit IRC06:23
*** sc68cal has joined #openstack-infra06:25
*** ildikov has joined #openstack-infra06:29
*** ildikov_ has quit IRC06:29
*** gokrokve has quit IRC06:29
*** flaper87|afk is now known as flaper8706:34
*** emagana has quit IRC06:39
*** pkoniszewski has joined #openstack-infra06:40
*** emagana has joined #openstack-infra06:40
*** emagana has quit IRC06:45
*** rmcall has quit IRC06:45
*** rushiagr is now known as rushiagr_away06:48
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Fix pattern match for translations
*** rushiagr_away is now known as rushiagr06:49
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Fix pattern match for translations
*** Ryan_Lane has joined #openstack-infra06:55
*** gokrokve has joined #openstack-infra06:56
*** mrda is now known as mrda-vacation06:58
*** ashaeron has joined #openstack-infra06:59
*** gokrokve has quit IRC06:59
*** gokrokve has joined #openstack-infra07:00
*** JadnG has joined #openstack-infra07:02
*** gokrokve has quit IRC07:05
*** koolhead17 has joined #openstack-infra07:05
*** k4n0 has quit IRC07:07
*** jpich has quit IRC07:07
*** ildikov has quit IRC07:08
*** _nadya_ has joined #openstack-infra07:11
*** arxcruz has joined #openstack-infra07:12
*** dtantsur|afk is now known as dtantsur07:13
openstackgerritJeffrey Zhang proposed a change to openstack-infra/config: Add keystone-formula project
*** dmellado has joined #openstack-infra07:19
*** zaro has joined #openstack-infra07:20
*** luqas has joined #openstack-infra07:20
*** arxcruz has quit IRC07:21
openstackgerritA change was merged to openstack/requirements: Update oslo.utils to latest version
openstackgerritA change was merged to openstack/requirements: Update oslo.i18n to latest version
*** k4n0 has joined #openstack-infra07:22
openstackgerritJoshua Hesketh proposed a change to openstack-infra/config: Send python-job logs to swift
*** jgallard has joined #openstack-infra07:24
*** jgallard has quit IRC07:25
*** jgallard has joined #openstack-infra07:25
*** luqas has quit IRC07:26
*** luqas has joined #openstack-infra07:27
*** yamamoto_ has quit IRC07:27
*** arnaud__ has quit IRC07:28
*** arxcruz has joined #openstack-infra07:31
openstackgerritJoshua Hesketh proposed a change to openstack-infra/zuul: Display only first 7 chars for git commit id
*** yfried has joined #openstack-infra07:35
*** yfried has quit IRC07:35
*** yfried__ has quit IRC07:35
*** amotoki_ has joined #openstack-infra07:35
*** yfried has joined #openstack-infra07:35
*** alkari has joined #openstack-infra07:37
openstackgerritJeffrey Zhang proposed a change to openstack-infra/config: Add keystone-formula project
*** ndipanov_gone is now known as ndipanov07:39
*** jistr has joined #openstack-infra07:39
*** amotoki_ has quit IRC07:40
*** jlibosva has joined #openstack-infra07:43
*** bo_sh has joined #openstack-infra07:44
bo_shhi there. can someone explain or give me a link to how i use openstack-infra/config in order to use openstack logstash-elastic-search instant on my machines?07:45
*** dizquierdo has joined #openstack-infra07:46
*** koolhead17 has quit IRC07:47
bo_shsomeone? anyone?07:48
*** MaxV has joined #openstack-infra07:49
*** yfried has quit IRC07:54
*** yamamoto_ has joined #openstack-infra07:54
*** alkari has quit IRC07:55
*** yamamoto_ has quit IRC07:55
*** yamamoto_ has joined #openstack-infra07:56
*** ildikov has joined #openstack-infra07:57
*** yfried has joined #openstack-infra07:57
*** e0ne has joined #openstack-infra08:00
*** gokrokve has joined #openstack-infra08:00
*** HeOS has joined #openstack-infra08:02
*** e0ne has quit IRC08:04
*** gokrokve has quit IRC08:05
*** yamamoto_ has quit IRC08:06
*** derekh has joined #openstack-infra08:07
*** jerryz_ has joined #openstack-infra08:11
bo_shseriously? no one answers. :\08:12
ashaeronWell it is in the middle of the night in the states.08:13
bo_shashaeron. uh huh. now that you're saying this, it does make sense a lot!08:14
bo_shSorry for bothering, though. :)08:14
ashaeronNo problem.08:15
*** jerryz_ has quit IRC08:18
*** daya_k has quit IRC08:19
*** cnesa10 has joined #openstack-infra08:21
*** dizquierdo has quit IRC08:24
mika - hm, getting KeyError: 'http' with gertty, any ideas what's going wrong here? jeblair?08:25
*** alexpilotti has joined #openstack-infra08:25
*** yamamoto_ has joined #openstack-infra08:28
*** daya_k has joined #openstack-infra08:30
*** koolhead17 has joined #openstack-infra08:30
*** koolhead17 has joined #openstack-infra08:30
*** koolhead17 has quit IRC08:35
*** amuller has joined #openstack-infra08:42
*** ociuhandu has quit IRC08:47
*** yfried has quit IRC08:47
*** _nadya_ has quit IRC08:48
*** e0ne has joined #openstack-infra08:49
*** mrmartin has joined #openstack-infra08:49
*** bo_sh has left #openstack-infra08:49
*** luqas has quit IRC08:53
*** zz_johnthetubagu is now known as johnthetubaguy08:55
*** jp_at_hp has joined #openstack-infra08:56
*** _nadya_ has joined #openstack-infra08:56
*** koolhead17 has joined #openstack-infra08:57
*** jyuso has quit IRC08:58
*** gokrokve has joined #openstack-infra09:00
*** yamamoto_ has quit IRC09:01
*** mancdaz has quit IRC09:01
*** mrmartin has quit IRC09:02
*** katyafervent is now known as katyafervent_awa09:03
*** [HeOS] has joined #openstack-infra09:04
*** dteselkin has quit IRC09:04
*** HeOS has quit IRC09:04
*** mattymo has quit IRC09:05
*** gokrokve has quit IRC09:05
*** katyafervent_awa is now known as katyafervent09:05
*** dteselkin has joined #openstack-infra09:05
*** mattymo has joined #openstack-infra09:05
*** Daisy has quit IRC09:06
*** koolhead17 has quit IRC09:06
*** koolhead17 has joined #openstack-infra09:07
*** koolhead17 has quit IRC09:09
*** koolhead17 has joined #openstack-infra09:09
*** koolhead17 has quit IRC09:09
*** koolhead17 has joined #openstack-infra09:09
*** lttrl has joined #openstack-infra09:09
*** [HeOS] is now known as HeOS09:09
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Add missing merge-check and noop-jobs to layout.yaml
*** rushiagr is now known as rushiagr_away09:18
*** rushiagr_away is now known as rushiagr09:18
*** dizquierdo has joined #openstack-infra09:19
*** rakhmerov has quit IRC09:20
*** HeOS has quit IRC09:20
*** katyafervent has quit IRC09:20
*** mattymo has quit IRC09:20
*** yamamoto_ has joined #openstack-infra09:20
*** tsufiev has quit IRC09:21
*** tomoe has quit IRC09:21
*** dteselkin has quit IRC09:21
*** luqas has joined #openstack-infra09:21
*** bdpayne_ has quit IRC09:22
*** rakhmerov has joined #openstack-infra09:24
*** rakhmerov has quit IRC09:25
*** dteselkin has joined #openstack-infra09:26
*** tsufiev has joined #openstack-infra09:26
*** HeOS has joined #openstack-infra09:27
*** katyafervent_awa has joined #openstack-infra09:27
*** e0ne has quit IRC09:27
*** e0ne has joined #openstack-infra09:28
*** mattymo has joined #openstack-infra09:28
*** mattymo has quit IRC09:28
*** mattymo has joined #openstack-infra09:30
*** mattymo has quit IRC09:30
*** rakhmerov has joined #openstack-infra09:30
*** marun has joined #openstack-infra09:31
*** mattymo has joined #openstack-infra09:32
*** mattymo has quit IRC09:32
*** pelix has joined #openstack-infra09:32
*** cdent has joined #openstack-infra09:34
*** mattymo has joined #openstack-infra09:34
*** mattymo has quit IRC09:34
*** jgallard has quit IRC09:34
*** mattymo has joined #openstack-infra09:36
*** mattymo has quit IRC09:36
*** amotoki_ has joined #openstack-infra09:36
*** jgallard has joined #openstack-infra09:37
*** yamamoto_ has quit IRC09:37
*** mattymo has joined #openstack-infra09:38
*** mattymo has quit IRC09:38
*** mattymo has joined #openstack-infra09:40
*** mattymo has quit IRC09:40
*** amotoki_ has quit IRC09:41
*** jgallard has quit IRC09:41
openstackgerritJeffrey Zhang proposed a change to openstack-infra/config: Add keystone-formula project
*** jgallard has joined #openstack-infra09:41
*** nosnos has quit IRC09:41
*** nosnos has joined #openstack-infra09:42
*** mattymo has joined #openstack-infra09:42
*** mattymo has quit IRC09:42
*** mattymo has joined #openstack-infra09:44
*** mattymo has quit IRC09:44
*** mattymo has joined #openstack-infra09:46
*** mattymo has quit IRC09:46
*** _buzz has quit IRC09:46
*** nosnos has quit IRC09:47
*** andreykurilin has joined #openstack-infra09:47
*** _buzz has joined #openstack-infra09:48
*** maxbit has joined #openstack-infra09:48
*** mattymo has joined #openstack-infra09:48
*** mattymo has quit IRC09:48
*** mattymo has joined #openstack-infra09:50
*** mattymo has quit IRC09:50
*** mattymo has joined #openstack-infra09:52
*** mattymo has quit IRC09:52
*** unicell1 has joined #openstack-infra09:54
*** michchap_ has quit IRC09:54
*** mattymo has joined #openstack-infra09:55
*** mattymo has quit IRC09:55
*** michchap has joined #openstack-infra09:55
*** unicell has quit IRC09:56
*** mattymo has joined #openstack-infra09:57
*** mattymo has quit IRC09:57
*** mattymo has joined #openstack-infra09:59
*** mattymo has quit IRC09:59
*** Longgeek has joined #openstack-infra09:59
*** gokrokve has joined #openstack-infra10:00
*** dtantsur is now known as dtantsur|brb10:00
*** mattymo has joined #openstack-infra10:01
*** mattymo has quit IRC10:01
*** gokrokve has quit IRC10:01
*** gokrokve has joined #openstack-infra10:02
*** mattymo has joined #openstack-infra10:03
*** mattymo has quit IRC10:03
openstackgerritDavid Pursehouse proposed a change to stackforge/gertty: Allow to authenticate to Gerrit with HTTP basic auth
*** yamamoto_ has joined #openstack-infra10:03
*** jgallard has quit IRC10:04
*** mattymo has joined #openstack-infra10:05
*** mattymo has quit IRC10:05
*** marun has quit IRC10:05
*** praneshp has quit IRC10:06
*** marun has joined #openstack-infra10:06
*** gokrokve has quit IRC10:07
*** mattymo has joined #openstack-infra10:07
*** mattymo has quit IRC10:07
*** john-davidge has joined #openstack-infra10:07
*** yamamoto_ has quit IRC10:07
*** yamamoto_ has joined #openstack-infra10:08
*** mattymo has joined #openstack-infra10:09
*** mattymo has quit IRC10:09
openstackgerritDavid Pursehouse proposed a change to stackforge/gertty: Fix exception in change list when change owner has no name
*** pcm_ has joined #openstack-infra10:11
*** mattymo has joined #openstack-infra10:11
*** mattymo has quit IRC10:11
*** yamamoto_ has quit IRC10:12
*** ociuhandu has joined #openstack-infra10:12
*** mattymo has joined #openstack-infra10:13
*** mattymo has quit IRC10:13
*** yfried has joined #openstack-infra10:13
*** pcm__ has joined #openstack-infra10:14
*** mattymo has joined #openstack-infra10:15
*** mattymo has quit IRC10:15
*** marun has quit IRC10:16
*** mattymo has joined #openstack-infra10:17
*** mattymo has quit IRC10:17
*** pcm_ has quit IRC10:17
*** cnesa10 has quit IRC10:18
*** jerryz has quit IRC10:19
*** mattymo has joined #openstack-infra10:19
*** mattymo has quit IRC10:19
*** yamamoto_ has joined #openstack-infra10:21
*** mattymo has joined #openstack-infra10:21
*** mattymo has quit IRC10:21
*** mattymo has joined #openstack-infra10:23
*** mattymo has quit IRC10:23
*** yamamoto_ has quit IRC10:23
*** yamamoto_ has joined #openstack-infra10:23
*** chenglch|3 has quit IRC10:24
*** cnesa9 has joined #openstack-infra10:24
*** mattymo has joined #openstack-infra10:25
*** mattymo has quit IRC10:25
*** mattymo has joined #openstack-infra10:27
*** mattymo has quit IRC10:27
*** yamamoto_ has quit IRC10:28
*** mattymo has joined #openstack-infra10:29
*** mattymo has quit IRC10:29
*** mattymo has joined #openstack-infra10:31
*** mattymo has quit IRC10:31
*** mattymo has joined #openstack-infra10:33
*** mattymo has quit IRC10:33
*** penguinRaider has joined #openstack-infra10:33
*** mattymo has joined #openstack-infra10:35
*** mattymo has quit IRC10:35
*** mattymo has joined #openstack-infra10:37
*** mattymo has quit IRC10:37
*** daya_k has quit IRC10:39
*** mattymo has joined #openstack-infra10:39
*** mattymo has quit IRC10:39
*** koolhead17 has quit IRC10:39
*** berendt has joined #openstack-infra10:41
*** mattymo has joined #openstack-infra10:41
*** mattymo has quit IRC10:41
*** koolhead17 has joined #openstack-infra10:42
*** koolhead17 has joined #openstack-infra10:42
*** mattymo has joined #openstack-infra10:43
*** mattymo has quit IRC10:43
*** mattymo has joined #openstack-infra10:46
*** mattymo has quit IRC10:46
openstackgerritDavid Pursehouse proposed a change to stackforge/gertty: Allow to authenticate to Gerrit with HTTP basic auth
*** cnesa9 has quit IRC10:46
*** asselin has quit IRC10:46
sdaguelifeless: ok, so is it getting better?10:47
*** mattymo has joined #openstack-infra10:48
*** mattymo has quit IRC10:48
openstackgerritDavid Pursehouse proposed a change to stackforge/gertty: Add gertty-env to .gitignore
*** mattymo has joined #openstack-infra10:50
*** mattymo has quit IRC10:50
*** cnesa9 has joined #openstack-infra10:51
*** cnesa9 has quit IRC10:52
*** mattymo has joined #openstack-infra10:52
*** mattymo has quit IRC10:52
*** cnesa9 has joined #openstack-infra10:53
AJaegerflaper87, ping10:53
flaper87AJaeger: pong10:54
*** mattymo has joined #openstack-infra10:54
*** mattymo has quit IRC10:54
AJaegerflaper87, will not merge as is.10:54
AJaegerflaper87, you need to remove the approval and approve again - you previously approved while there was no +2.10:54
flaper87AJaeger: done :)10:55
AJaegergerrit does not handle this10:55
AJaegerflaper87, now gerrit is happy ;)10:55
flaper87As long as it's happy, we're all happy :D10:55
*** mattymo has joined #openstack-infra10:56
*** mattymo has quit IRC10:56
*** dims_ has joined #openstack-infra10:57
*** mattymo has joined #openstack-infra10:58
*** mattymo has quit IRC10:58
*** miqui has quit IRC11:00
*** mattymo has joined #openstack-infra11:00
*** mattymo has quit IRC11:00
*** gokrokve has joined #openstack-infra11:00
*** mattymo has joined #openstack-infra11:02
*** mattymo has quit IRC11:02
*** mattymo has joined #openstack-infra11:04
*** mattymo has quit IRC11:04
*** gokrokve has quit IRC11:05
*** ppai has quit IRC11:05
*** mattymo has joined #openstack-infra11:06
*** mattymo has quit IRC11:06
*** koolhead17 has quit IRC11:07
*** koolhead17 has joined #openstack-infra11:07
*** mattymo has joined #openstack-infra11:08
*** mattymo has quit IRC11:08
ekarlso-hmm, doesn't curl or httpie like certificiates with SAN names ?11:09
*** mattymo has joined #openstack-infra11:10
*** mattymo has quit IRC11:10
*** yamamoto has joined #openstack-infra11:11
*** koolhead_ has joined #openstack-infra11:11
*** koolhead17 has quit IRC11:12
*** mattymo has joined #openstack-infra11:12
*** mattymo has quit IRC11:12
*** mattymo has joined #openstack-infra11:14
*** mattymo has quit IRC11:14
*** Longgeek_ has joined #openstack-infra11:14
openstackgerritDavanum Srinivas (dims) proposed a change to openstack/requirements: Update oslo.messaging to latest version
*** sld has joined #openstack-infra11:15
*** mattymo has joined #openstack-infra11:16
*** mattymo has quit IRC11:16
*** Longgeek has quit IRC11:18
*** mattymo has joined #openstack-infra11:18
*** mattymo has quit IRC11:18
openstackgerritDavanum Srinivas (dims) proposed a change to openstack/requirements: Update oslo.serialization to latest version
*** ppai has joined #openstack-infra11:19
*** zz_dimtruck is now known as dimtruck11:19
*** eharney has quit IRC11:20
*** mattymo has joined #openstack-infra11:20
*** mattymo has quit IRC11:20
*** dims_ has quit IRC11:21
*** dims_ has joined #openstack-infra11:21
*** katyafervent_awa is now known as katyafervent11:22
*** daya_k has joined #openstack-infra11:22
*** dims__ has joined #openstack-infra11:22
*** mattymo has joined #openstack-infra11:24
*** dims_ has quit IRC11:25
*** baoli has joined #openstack-infra11:26
*** loki184 has quit IRC11:30
openstackgerritJeffrey Zhang proposed a change to openstack-infra/config: Add keystone-formula project
*** dmsimard_away is now known as dmsimard11:32
*** ashaeron has quit IRC11:34
openstackgerritRicardo Carrillo Cruz proposed a change to openstack-infra/config: Add version-properties builder to macros.yaml
*** ashaeron has joined #openstack-infra11:35
*** dimtruck is now known as zz_dimtruck11:37
*** ekarlso- has quit IRC11:37
*** ekarlso- has joined #openstack-infra11:37
*** amotoki_ has joined #openstack-infra11:37
*** mwagner_lap has quit IRC11:38
*** adalbas has joined #openstack-infra11:39
*** zz_dimtruck is now known as dimtruck11:40
*** amotoki_ has quit IRC11:41
*** isviridov_away is now known as isviridov11:42
*** KanagarajM has quit IRC11:46
*** luqas has quit IRC11:49
*** koolhead_ has quit IRC11:49
jd__fungi: could you add oslo-core to pylockfile-core? it seems like it has been forgotten11:51
*** yaguang has quit IRC11:53
*** marcusvrn has joined #openstack-infra11:53
*** koolhead17 has joined #openstack-infra11:57
*** daya_k has quit IRC11:57
*** koolhead17 has quit IRC11:59
*** gokrokve has joined #openstack-infra12:00
*** dprince has joined #openstack-infra12:02
*** rfolco has joined #openstack-infra12:03
*** gokrokve has quit IRC12:05
flaper87guys, I'd like to create a job for zaqarclient that runs the client's functional tests against a running zaqar instance. AFAIU, I should go with something similar to this patch:,cm12:07
*** sld has left #openstack-infra12:07
flaper87what I don't know is where the client tests should be run12:08
flaper87I mean, should I clone the client there and run the tests?12:08
flaper87is there a better way to do that ?12:08
ekarlso-phschwartz: you around  :)12:12
*** mmedvede has joined #openstack-infra12:12
*** jamespage_ has quit IRC12:17
*** bswartz has quit IRC12:18
openstackgerritRadomir Dopieralski proposed a change to openstack/requirements: Update xstatic-jquery-ui for Horizon
*** jgallard has joined #openstack-infra12:21
*** dkranz has joined #openstack-infra12:21
*** michchap has quit IRC12:25
*** michchap has joined #openstack-infra12:25
*** michchap has quit IRC12:27
*** weshay has joined #openstack-infra12:27
*** michchap has joined #openstack-infra12:27
*** aysyd has joined #openstack-infra12:27
*** dimtruck is now known as zz_dimtruck12:29
*** miqui has joined #openstack-infra12:29
*** mbacchi has joined #openstack-infra12:33
*** dims__ has quit IRC12:34
EmilienManteaya: hey, I'm reading your review, let me know if you are around12:35
*** hdd has joined #openstack-infra12:35
*** dims_ has joined #openstack-infra12:35
*** jcoufal has quit IRC12:35
*** jcoufal has joined #openstack-infra12:36
*** mwagner_lap has joined #openstack-infra12:37
*** jcoufal_ has joined #openstack-infra12:38
*** dims_ has quit IRC12:38
*** jcoufal has quit IRC12:38
*** dims_ has joined #openstack-infra12:38
*** rainya has joined #openstack-infra12:39
*** rainya has quit IRC12:39
*** jcoufal_ has quit IRC12:39
*** jcoufal has joined #openstack-infra12:40
*** rainya has joined #openstack-infra12:40
*** Clabbe has joined #openstack-infra12:40
*** zz_dimtruck is now known as dimtruck12:40
*** mwagner_lap has quit IRC12:41
*** mwagner_lap has joined #openstack-infra12:41
*** rainya has quit IRC12:42
*** rainya has joined #openstack-infra12:42
*** radez_g0n3 is now known as radez12:43
*** rainya has quit IRC12:43
*** Daisy has joined #openstack-infra12:44
*** jcoufal has quit IRC12:45
*** jcoufal has joined #openstack-infra12:45
*** hdd has quit IRC12:48
*** KanagarajM has joined #openstack-infra12:49
*** nelsnelson has joined #openstack-infra12:52
*** oomichi has quit IRC12:54
*** hdd has joined #openstack-infra12:54
*** eharney has joined #openstack-infra12:55
*** yfried has quit IRC12:57
*** sdake_ has joined #openstack-infra12:58
*** gokrokve has joined #openstack-infra13:00
*** afazekas_ has quit IRC13:01
*** nelsnelson has quit IRC13:01
*** dtantsur|brb is now known as dtantsur13:02
*** nelsnelson has joined #openstack-infra13:02
KiallWeird.. I've not seen a gate job go to SKIPPED before like this? The dependant patchset is passing etc
*** gokrokve has quit IRC13:05
*** luqas has joined #openstack-infra13:08
*** ppai has quit IRC13:09
*** dane_leblanc has joined #openstack-infra13:13
*** dustins has joined #openstack-infra13:13
fungiKiall: if you hover the cursor over the black circle, it should tell you why it's skipped13:14
Kiallfungi: oh, too late... never noticed that mouseover before!13:14
fungi"This change was unable to be automatically merged with the current state of the repository. Please rebase your change and upload a new patchset."13:15
fungiso it was skipping it because it had a merge conflict elsewhere, even though the change ahead of it did not13:15
fungipossibly it merge-conflicted with the patch ahead of it?13:15
KiallYea - I just saw that, makes perfect sense now :) Just never seen the review stay on the Zuul status page when it doesn't merge cleanly before!13:16
KiallThanks :)13:16
*** mriedem has joined #openstack-infra13:16
*** afazekas_ has joined #openstack-infra13:16
fungiKiall: it does that in case the change ahead of it fails for some other reason, then it can try to merge automatically13:16
KiallOh, right.. the next patch could fail gate, so the conflict would be gone.. nice.13:17
fungiwhen you're being integration-tested in a long dependent pipeline with other changes which have a non-insignificant chance of failing ahead, it's a bit of a time-saver13:17
KiallYea, I can imagine :)13:18
fungijd__: i've included oslo-core in pylockfile-core now13:18
jd__thanks fungi13:19
*** craigbr has joined #openstack-infra13:19
*** amotoki has quit IRC13:21
*** ildikov has quit IRC13:21
*** flaviof is now known as flaviof_zzz13:22
*** dkranz has quit IRC13:24
*** sdake_ has quit IRC13:24
*** pballand has quit IRC13:24
*** hdd has quit IRC13:25
*** sabeen1 has quit IRC13:26
fungiclarkb: jamielennox: -python34 jobs may not work until merges (we don't have the headers on those slaves by default)13:26
fungialso my plan is to just replace -python33 with -python34 in the python3-jobs template when the time comes13:27
*** julim has joined #openstack-infra13:28
fungiso adding arbitrary -python34 jobs now is premature and eventually unnecessary (though i saw tempest-lib already did for some reason)13:28
*** _nadya_ has quit IRC13:28
*** esker has joined #openstack-infra13:29
fungiarosen: pep8 *can* actually fail on a change in combination with another change ahead of it in the gate pipeline, which would not have been tested in that way in the check pipeline just being merged to the branch tip... though we've discussed a distinct probability of assuming that will be a rare wedge, ceasing to retest it in the gate and just letting it break projects if and when it does happen13:29
*** kgiusti has joined #openstack-infra13:32
fungimika: it looks like maybe you're using gertty with a gerrit which is not i think so far it's only been heavily tested with gerrit 2.8 so if the gerrit you're hitting is a different version, some additional research may be needed13:32
*** ashaeron has quit IRC13:33
*** bknudson has joined #openstack-infra13:34
*** dkranz has joined #openstack-infra13:37
AJaegerfungi, I broke the propose-upstream-translation jobs with my last patch - sorry. Good thing is it gave some translations time to merge in with the long queue before a new patch was proposed ;)13:37
AJaegerfungi, patch to fix it at
*** torandu has joined #openstack-infra13:38
*** amotoki has joined #openstack-infra13:38
*** rainya has joined #openstack-infra13:38
*** markmcclain has joined #openstack-infra13:40
*** markmcclain has quit IRC13:40
AJaegerfungi, you didn't want to give translation another day to catch up? ;) Thanks!13:41
fungiAJaeger: yeah, that was a trivial and obviously correct fix, so no reason to delay it further13:41
*** markmcclain has joined #openstack-infra13:41
* AJaeger is ashamed that my testing was too limited and didn't catch this ;(13:42
fungiAJaeger: well, i reviewed the previous patch and didn't catch it either13:42
*** amotoki has quit IRC13:43
AJaegerfungi, is a missing merge-check in layouts.yaml critical? I fixed one case in
fungiAJaeger: it won't break anything, it will just cause that project to miss out on more rapid identification of merge conflicts13:44
*** tomoe has joined #openstack-infra13:44
AJaegerfungi, so no need to push that patch ;) Thanks13:45
fungiand thanks... it was on my to do list to reprocess layout.yaml and look for any stragglers after the initial massive patch went through13:45
*** tomoe has quit IRC13:45
*** tomoe has joined #openstack-infra13:46
*** rainya has quit IRC13:47
*** rcarrillocruz has joined #openstack-infra13:48
rcarrillocruzhelo helo13:48
rcarrillocruzwho are the stackalytics core reviewers? are the same as infra?13:49
*** dimtruck is now known as zz_dimtruck13:49
*** mfer has joined #openstack-infra13:49
*** tomoe has quit IRC13:50
*** hdd has joined #openstack-infra13:51
*** nelsnelson has quit IRC13:51
*** tomoe has joined #openstack-infra13:52
flaper87fungi: did you, by any chance, read the questions I sent before?13:53
*** yamahata has joined #openstack-infra13:54
AJaegerfungi, you're welcome - I'm sure another review in a week won't hurt ;)13:54
flaper87fungi: I'd like to add a job for zaqarclient that runs the client functional tests against a running zaqar instance. I was wondering what the steps for doing that are.13:54
flaper87fungi: I guess I need something like:,cm13:54
flaper87but I'm not sure where the tests are actually called13:55
*** Longgeek_ has quit IRC13:56
*** nelsnelson has joined #openstack-infra13:56
fungircarrillocruz:,members (not the same as infra-core)13:56
openstackgerritBarnaby Court proposed a change to openstack-infra/jenkins-job-builder: Adds credentials-binding plugin support to wrappers
openstackgerritA change was merged to openstack-infra/config: Fix pattern match for translations
fungiflaper87: without digging into the details i'd say it sounds like you want something similar to the swift and neutron functional tests?13:57
flaper87fungi: mmh, lemme take a look at those jobs13:58
flaper87fungi: I want every patch proposed to zaqarclient to be tested live against a running zaqar instance.13:58
* flaper87 checks swift tests13:58
fungiflaper87: maybe something like you have there, though i think the other functional tests generally reduce the set of enabled services in devstack significantly13:58
*** _nadya_ has joined #openstack-infra13:59
rcarrillocruzjust added my company affiliation (something i should've done months ago, but oh well...)13:59
flaper87I was thinking that I could disable all services and just enable Zaqar, since it's the only one I care about for this job13:59
*** flaviof_zzz is now known as flaviof13:59
fungiflaper87: yeah, i think that's what the others are doing... though this is a relatively new frontier for projects, so there may be some lessons learned. i think marun was working on this sort of stuff most recently for neutron, so he might have suggestions14:00
*** gokrokve has joined #openstack-infra14:00
fungii believe he to a great extent started with the swift functional test model and evolved it to the neutron use case14:00
flaper87fungi: cool, thanks for the info. I'll dig into this14:01
fungialso mtreinish might be able to project what the tempest-lib breakout implies for the future design of functional tests14:01
*** adalbas has quit IRC14:01
fungii gather there's interest in using that to simplify them somewhat14:01
*** gokrokve has quit IRC14:02
fungibut not sure what the timeline is for having it usable to that end14:02
*** gokrokve has joined #openstack-infra14:02
mtreinishfungi: well, the hope is it'll provide some useful testing patterns and interfaces (like the rest client) for spinning up functional testing14:03
*** dustins_ has joined #openstack-infra14:03
mtreinishbut they should be more coupled to the project than just what's in the lib which will just provide some nice building blocks14:03
*** dangers_away is now known as dangers14:03
*** dustins has quit IRC14:03
openstackgerritBarnaby Court proposed a change to openstack-infra/jenkins-job-builder: Adds credentials-binding plugin support to wrappers
mtreinishas for timeline I'm planning to have the first version on pypi within the next week or 214:03
mtreinishbut that will just cover the cli testing framework from tempest14:04
fungisounds like awesome progress14:04
*** yamamoto has quit IRC14:04
*** stevemar has joined #openstack-infra14:05
*** armax_ has joined #openstack-infra14:06
*** torandu has quit IRC14:08
*** afazekas_ has quit IRC14:08
*** torandu has joined #openstack-infra14:09
*** craigbr has quit IRC14:10
*** wenlock has joined #openstack-infra14:10
*** Longgeek has joined #openstack-infra14:11
*** sandywalsh_ has joined #openstack-infra14:11
*** koolhead17 has joined #openstack-infra14:12
*** koolhead17 has joined #openstack-infra14:12
*** atiwari has joined #openstack-infra14:12
*** pcrews has joined #openstack-infra14:15
*** sandywalsh has quit IRC14:18
zuldhellmann: also six is used everywhere in oslo.i18n but its not listed in the requirements.txt14:23
zuldhellmann: LP: #136776114:24
fungizul: starting to apply r1chardj0n3s's to projects as a static check will help ferret out things like that14:24
zulfungi: cool14:24
fungizul: he has work underway to get it enforced for nova first as a proof-of-concept and then fan out to all official projects]=14:25
fungi(it is the result of a discussion we had at the infra/qa mid-cycle in germany)14:25
zulfungi: wouldnt it be better to get started with the oslo projects and fan out from there?14:26
*** tonytan4ever has joined #openstack-infra14:27
zulfungi: just staying14:27
dhellmannzul: on it14:27
fungizul: well, we found obvious examples in nova, so it was easy to use as a demonstration of its effectiveness14:27
*** pkoniszewski has quit IRC14:28
*** zz_jgrimm is now known as jgrimm14:28
fungiwhile his manager is the nova ptl, i think the choice actually had more to do with it being the oldest and largest (loc-wise) project and so more likely to have that sort of cruft14:29
*** rwsu has joined #openstack-infra14:30
dhellmannfungi: is there a patch up for review for that work? if it's in-project work, I'd like to add it to the oslo cookiecutter template. If it's a separate job, then I guess I won't have anything to do?14:30
*** mpaolino has joined #openstack-infra14:30
*** atiwari has quit IRC14:30
fungithe conversation actually started when one project, i forget which, spontaneously broke the first day of the mid-cycle because it was importing a module from a transitive dependency which released a new version no longer depending on that package... *boom*14:30
fungidhellmann: i believe the hope is to spread it as a separate job first and then once it's enforced everywhere combine it with another existing job/tox env. might make sense in the cookiecutter but talk to r1ch4rdj0n3s when he's around (au daytime) on the details14:31
dhellmannfungi: ok14:32
openstackgerritFlavio Percoco proposed a change to openstack-infra/config: Add a job for functional tests in zaqarclient
flaper87fungi: does this make sense? how can I test that job?14:33
* flaper87 is sure as hell that won't work :P14:33
fungiflaper87: i'd suggest as the currently documented way to test out devstack-gate jobs manually14:34
fungiflaper87: though sdague may have suggestions based on his container work14:34
*** Longgeek has quit IRC14:34
flaper87sdague: any hint? :)14:35
flaper87fungi: thanks14:35
*** yamamoto has joined #openstack-infra14:35
*** garyh has joined #openstack-infra14:35
sdaguehonestly, I haven't gotten that far on that yet14:35
flaper87mmh, so I should probably set it as non-voting until it's working as expected14:36
flaper87I don't think I've the time/resources to test it manually14:36
fungiflaper87: if marun's experiences with the neutron functional test are any indication, you *definitely* want to do that, yes14:36
*** maxbit has quit IRC14:36
flaper87fungi: LOL, ok!14:37
fungiflaper87: only because he managed to wedge neutron's testing at least a couple times while he was still tweaking the job14:37
flaper87fungi: yeah, that was exactly my fear14:37
fungiflaper87: but that still shouldn't stop you from testing it out manually to reduce the number of changes you need to propose to infra. as long as you have access to a cloud or have a spare machine sitting around on which you can run devstack, i highly recommend taking it for a spin. that's why i wrote that howto section14:39
openstackgerritFlavio Percoco proposed a change to openstack-infra/config: Add a job for functional tests in zaqarclient
*** dhellmann is now known as dhellmann_14:39
*** yamamoto has quit IRC14:39
fungiflaper87: or team up with one of the other zaqar devs to help test it with you if you're short on time/systems14:39
*** _nadya_ has quit IRC14:39
flaper87fungi: yeah, I think we can find something to do a manual test14:40
*** dhellmann_ is now known as dhellmann14:40
*** _nadya_ has joined #openstack-infra14:40
flaper87fungi: would landing the first version and start manual testing once it has landed make sense?14:40
flaper87at least to get an idea of how wrong it is based on the output :P14:40
flaper87I guess not, I'll try to find someone that would help testing14:40
*** habib has joined #openstack-infra14:41
fungiflaper87: i took advantage of rackspace's generous offer. $500 services credit a month for 2 years just for having a commit somewhere in an official openstack project is plenty to run devstack many, many, many times over14:41
*** Longgeek has joined #openstack-infra14:42
*** _nadya_ has quit IRC14:42
*** ashaeron has joined #openstack-infra14:43
*** mpaolino has quit IRC14:43
*** pblaho_ has joined #openstack-infra14:43
*** pblaho_ has quit IRC14:44
*** pblaho_ has joined #openstack-infra14:44
jrollfungi: wow, that page is really really broken :(14:44
fungijroll: yeah, it is14:44
*** john-davidge has quit IRC14:44
* jroll makes noise14:45
fungijroll: i brought it to rackspace's attention, and their answer was to send me a pdf version to print, fill out, scan and e-mail back to them14:45
* jroll yells14:45
jrollthanks for that tidbit, fungi14:45
*** pblaho_ is now known as pblaho14:46
jrollso embarassing14:46
*** pballand has joined #openstack-infra14:46
fungijroll: just doing my bit! ;)14:46
fungijroll: i didn't complain further once they sent me the pdf, as it's an awesomely generous offer and i'm not one to look a gift horse in the mouth14:47
*** jgrimm is now known as zz_jgrimm14:47
*** rwsu has quit IRC14:48
*** maxbit has joined #openstack-infra14:48
*** Longgeek has quit IRC14:49
fungijroll: if you have access to the bug tracker, i can give you the bug number i opened, for extra ammunition14:49
sandywalsh_hey y'all, so we're about to add some new stacktach projects to stackforge ... but I have a question about the security groups:14:49
jrollfungi: sure!14:49
*** hemna_ is now known as hemna14:49
sandywalsh_1. we want our stacktach-core group to be a member of the monasca-core group ... is that possible?14:49
jrollthat would be great14:49
*** zz_gondoi is now known as gondoi14:49
sandywalsh_2. and vice-versa :)14:50
fungisandywalsh_: yeah, gerrit groups can include other groups, but group management is a manual process in the webui14:50
sandywalsh_fungi: great ... I'm actually glad it's manual. Easier for me to fix my mistakes :) thanks14:51
*** david-lyle has joined #openstack-infra14:51
openstackgerritJoakim Löfgren proposed a change to openstack-infra/jenkins-job-builder: Add PMD publisher
fungijroll: tenant id 901056 ticket id 140403-ord-0001009 (opened and closed in april)14:52
jrollfungi: beautiful, thank you :)14:52
*** annegentle has joined #openstack-infra14:52
*** marun has joined #openstack-infra14:54
*** rpodolyaka has quit IRC14:54
sdaguesandywalsh_: I would be concerned about the circular definition there14:55
*** zz_dimtruck is now known as dimtruck14:55
*** reed has joined #openstack-infra14:55
sdaguewhy not just create a new group, put them both in it, and make that the review team for both projects14:56
sdagueor just flatten all out into one group14:56
*** craigbr has joined #openstack-infra14:58
*** datsun180b has joined #openstack-infra14:58
*** rmcall has joined #openstack-infra14:59
*** craigbr has quit IRC15:00
*** Daisy has quit IRC15:00
jeblairmika: can you run the following command and capture the output?15:01
jeblairmika: curl --anyauth -u <username>:<password> "https://<hostname>/a/changes/ngcpcfg-api~master~I45d11990aa71855d7be84c47d9f6fc583039dae4?o=DETAILED_LABELS&o=ALL_REVISIONS&o=ALL_COMMITS&o=MESSAGES&o=DETAILED_ACCOUNTS"15:01
*** dane_leblanc has quit IRC15:01
jeblairmika: there's probably too much for you to sanitize to paste, so just let me know when you have it and i can walk you through what i need15:01
jrollfungi: noises have been made, thanks again15:02
*** annegentle has quit IRC15:03
fungijroll: funny story about *why* that's bad... since the workaround was manual and i couldn't cut/paste from the dashboard, i ended up mis-entering my account number (mea culpa!) which they didn't catch and so never credited me. i've sent in a corrected one a couple months ago after talking to the sales manager but they don't seem to have applied it yet (i'm still being billed) so i need to follow up15:03
fungiwith him again15:03
*** habib has quit IRC15:03
jeblairfungi: that's a really funny story :(15:04
*** mattymo has quit IRC15:04
fungiwell, i feel responsible for the initial transcription error15:04
* jroll appends that15:04
jrollbecause we apparently can't run a simple form on the internet15:04
*** andreykurilin has quit IRC15:04
jrollthat's absurd15:04
*** rainya has joined #openstack-infra15:05
*** rainya has quit IRC15:05
fungijroll: the paperless future has not _yet_ arrived, it seems15:05
* jroll looks around for a table15:05
jeblairi think we should start an openstack project for that so all openstack clouds can have a consistent data entry form experience15:05
*** yamamoto has joined #openstack-infra15:05
jrollmeh, we'll just mess it up downstream15:06
fungijeblair: web forms as a service!15:06
fungijeblair: can we make it send the content in an e-mail message, just like the old days?15:07
*** dkranz has quit IRC15:07
jrollas long as it relies heavily on really complicated javascript, I'm in15:07
jroll<form> and <input> are way too easy15:07
*** bswartz has joined #openstack-infra15:07
fungijroll: well, what it needs is a hidden form element so that we can specify the recipient address in the http post15:09
fungijroll: that makes it nice and reusable!15:09
*** j^2 has joined #openstack-infra15:09
* fungi stops with the 20/20 hindsight and making fun of poor php devs who were doing the best they could to write software in something which wasn't a programming language15:10
*** yamamoto has quit IRC15:11
jrollfairly certain all websites are just a giant hack15:11
fungijroll: all computers are a giant hack, so by extension... yes15:11
*** dane_leblanc has joined #openstack-infra15:12
*** morganfainberg is now known as morgan15:12
*** _nadya_ has joined #openstack-infra15:12
*** morgan is now known as Morgan15:13
*** Sukhdev has joined #openstack-infra15:14
*** habib has joined #openstack-infra15:14
*** rlandy has quit IRC15:15
HeOSHow do you change project.config for gerrit default project "All-Projects" on Using jeepyb or through gerrit web-interface?15:15
*** Morgan is now known as morgan15:15
fungiHeOS: we manually change it through the webui or manually push it to the internal jgit15:17
fungiHeOS: we've never externalized management of that acl because we can't bootstrap it (it's self-enforcing, so you need to modify it to give access to modify it)15:17
*** lcheng_ has joined #openstack-infra15:17
fungiHeOS: though i believe yolanda once wrote a bootstrapper for it while working on the gerrit charm for canonical's cloud service, so she probably has some ideas on how to go about it15:18
flaper87guys, could any of you delete from pypi? I don't have rights on this package and it was renamed to `glance_store`15:18
HeOSfungi, how you do it in jgit?15:19
*** sandywalsh_ has quit IRC15:19
*** stevemar has quit IRC15:20
*** bdpayne has joined #openstack-infra15:21
HeOSfungi, do you clone All-Project repository on local computer for manual editing?15:21
*** stevemar has joined #openstack-infra15:21
*** dkranz has joined #openstack-infra15:21
fungiHeOS: basically, yes, you clone All-Projects.git, then checkout origin/refs/meta/config and edit project.config, then push that back to refs/meta/config through the ssh interface15:22
*** mmedvede has quit IRC15:22
HeOSfungi, thanks.15:22
*** dtantsur is now known as dtantsur|brb15:22
fungiHeOS: but the account you do that with needs additional privileges granted in the all-projects acl, thus the chicken-and-egg bootstrapping issue15:23
HeOSfungi, do you have procedure of review for this changes?15:23
openstackgerritJJ Asghar proposed a change to openstack-infra/config: Fix ruby-191-prep to uninstall 1.8 and install bundler
*** rainya has joined #openstack-infra15:23
*** yamamoto has joined #openstack-infra15:23
*** jerryz_ has joined #openstack-infra15:23
*** Sukhdev has quit IRC15:23
HeOSfungi, "then push that back to refs/meta/config through the ssh interface". Understand.15:23
*** Sukhdev has joined #openstack-infra15:24
fungiHeOS: we normally document the alteration with a proposed change to and then one of the admins manually alters the acl via whatever method once that's approved15:24
HeOSfungi, thanks.15:26
*** marun has quit IRC15:27
openstackgerritMark Vanderwiel proposed a change to openstack-infra/config: Fix ruby-191-prep to use rvm
*** yamamoto has quit IRC15:28
HeOSfungi, which workflow you use for creating new groups in gerrit? Is new group creating through project.config when anyone from administration team modified config for any project?15:28
*** jerryz_ has quit IRC15:30
*** Hal has joined #openstack-infra15:31
*** Hal is now known as Guest124515:31
*** gyee has joined #openstack-infra15:32
*** k4n0 has quit IRC15:34
*** Sukhdev has quit IRC15:34
*** rkukura has quit IRC15:35
*** ashaeron has quit IRC15:35
*** rkukura has joined #openstack-infra15:36
openstackgerritA change was merged to stackforge/gertty: Allow to authenticate to Gerrit with HTTP basic auth
openstackgerritA change was merged to stackforge/gertty: Fix exception in change list when change owner has no name
*** loki184 has joined #openstack-infra15:36
openstackgerritA change was merged to stackforge/gertty: Add gertty-env to .gitignore
*** Sukhdev has joined #openstack-infra15:37
*** jerryz_ has joined #openstack-infra15:38
*** patrickeast has joined #openstack-infra15:38
fungiHeOS: it's being handled through jeepyb except in unusual circumstances where we need a group which is not directly referenced in an acl (usually one which is intended to only be included in another group)15:39
*** virmitio has joined #openstack-infra15:40
*** bdpayne has quit IRC15:40
*** andreykurilin has joined #openstack-infra15:40
*** amotoki has joined #openstack-infra15:41
openstackgerritMatt Riedemann proposed a change to openstack-infra/elastic-recheck: Add query for gate-oslo.db-python33 setup failure
*** garyh has quit IRC15:47
*** mmaglana has joined #openstack-infra15:47
*** garyh has joined #openstack-infra15:50
*** HeOS has quit IRC15:52
*** ZZelle has quit IRC15:52
*** ZZelle has joined #openstack-infra15:53
*** daya_k has joined #openstack-infra15:53
*** arxcruz has quit IRC15:55
*** dkranz has quit IRC15:56
*** zz_jgrimm is now known as jgrimm15:56
*** habib has quit IRC15:58
*** IvanBerezovskiy has left #openstack-infra15:59
*** gokrokve has quit IRC15:59
*** kgiusti has quit IRC16:01
*** asselin has joined #openstack-infra16:01
*** yamamoto has joined #openstack-infra16:01
*** dustins_ is now known as dustins16:01
*** pblaho has quit IRC16:01
*** arnaud has joined #openstack-infra16:02
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for gate-oslo.db-python33 setup failure
*** arnaud has quit IRC16:03
*** loki184 has quit IRC16:04
nibalizerjeblair: clarkb fungi good morning!16:05
jeblairnibalizer: good morning; i just aprvd your patches, so we should be good to go soon16:06
nibalizer:D :D16:07
*** jistr has quit IRC16:07
openstackgerritA change was merged to openstack-infra/config: Formatting updates in hiera.yaml
*** dkranz has joined #openstack-infra16:08
*** sandywalsh has joined #openstack-infra16:09
*** mwagner_lap has quit IRC16:09
*** marun has joined #openstack-infra16:10
*** mestery has quit IRC16:10
*** mestery has joined #openstack-infra16:11
*** tomoe has quit IRC16:11
*** tomoe has joined #openstack-infra16:11
openstackgerritA change was merged to openstack-infra/config: Enable puppet3 master bootstrap
*** rwsu has joined #openstack-infra16:12
*** markmcclain has quit IRC16:12
*** dkranz has quit IRC16:13
*** daya_k has quit IRC16:13
*** Guest1245 has quit IRC16:15
*** _shaps_ has quit IRC16:15
*** _shaps_ has joined #openstack-infra16:16
*** tomoe has quit IRC16:16
*** ndipanov has quit IRC16:17
*** loki184 has joined #openstack-infra16:19
*** _nadya_ has quit IRC16:19
*** _nadya_ has joined #openstack-infra16:19
openstackgerritA change was merged to openstack-infra/config: Switch to using single projects yaml job
openstackgerritAndrew Bramley proposed a change to openstack/requirements: Update python-troveclient for Horizon
*** MaxV has quit IRC16:21
*** MaxV has joined #openstack-infra16:22
*** todd_dsm has joined #openstack-infra16:24
*** dkranz has joined #openstack-infra16:25
*** marcoemorais has joined #openstack-infra16:25
*** SumitNaiksatam has quit IRC16:25
*** jlibosva has quit IRC16:26
*** jlibosva has joined #openstack-infra16:26
*** MaxV has quit IRC16:26
*** cdent_ has joined #openstack-infra16:27
*** jlibosva has quit IRC16:28
*** cdent has quit IRC16:29
*** cdent_ is now known as cdent16:29
markvanclarkb: fungi: when ya get a chance...we switched to use rvm for ruby setup.   looks cleaner and should be easier to deal with down the road. thx.16:30
*** melwitt has joined #openstack-infra16:30
*** jgallard has quit IRC16:31
*** emagana has joined #openstack-infra16:31
*** gokrokve has joined #openstack-infra16:32
*** dane_leblanc has quit IRC16:32
*** dkranz has quit IRC16:34
*** marcoemorais has quit IRC16:35
*** bdpayne has joined #openstack-infra16:35
*** todd_dsm has quit IRC16:35
*** marcoemorais has joined #openstack-infra16:36
*** marcoemorais has quit IRC16:36
*** marcoemorais has joined #openstack-infra16:36
*** todd_dsm has joined #openstack-infra16:36
*** bhuvan has joined #openstack-infra16:38
*** SumitNaiksatam has joined #openstack-infra16:39
*** Ryan_Lane has quit IRC16:43
*** dane_leblanc has joined #openstack-infra16:43
*** adalbas has joined #openstack-infra16:43
openstackgerritA change was merged to openstack-infra/config: Add integration apply test
*** jerryz_ has quit IRC16:44
*** praneshp has joined #openstack-infra16:45
openstackgerritSpencer Krum proposed a change to openstack-infra/config: Updating node defn for puppet 3 master
*** jcoufal has quit IRC16:45
*** esker has quit IRC16:45
*** todd_dsm has quit IRC16:45
jeblairwow.  so many puppet certs from random people on the internet waiting to be signed by our puppetmaster :(16:46
nibalizerjeblair: so i think we want in as well, sorry it wasn't rebased last night16:46
nibalizerjeblair: ya, to be honest im pretty sure i contribute to that problem16:46
nibalizerif we template out the server = parameter, then put it hiera, we can fix that16:46
nibalizerjeblair: is there anything from * ? or that's all likely me16:47
*** dkranz has joined #openstack-infra16:47
jeblairnibalizer: it looks like is just a whitespace change, but the commit msg says something about removing all testing configuration from the script16:48
*** rmcall has quit IRC16:48
nibalizerwill fix16:48
*** bswartz1 has joined #openstack-infra16:49
openstackgerritBarnaby Court proposed a change to openstack-infra/jenkins-job-builder: Adds credentials-binding plugin support to wrappers
openstackgerritSpencer Krum proposed a change to openstack-infra/config: Updating node defn for puppet 3 master
nibalizerthe 'removing testing configuration' was folded down16:49
openstackgerritDerek Higgins proposed a change to openstack-infra/config: Install libxslt-devel on Fedora nodepool templates
jeblairclarkb, fungi, mordred:
openstackgerritA change was merged to openstack-infra/elastic-recheck: Remove fingerprint for bug 1291605
uvirtbotLaunchpad bug 1291605 in nova "unit test test_create_instance_with_networks_disabled race fail" [Medium,Invalid]
*** rmcall has joined #openstack-infra16:50
*** markmcclain has joined #openstack-infra16:51
*** bswartz has quit IRC16:53
*** emagana has quit IRC16:54
*** marcoemorais has quit IRC16:54
*** emagana has joined #openstack-infra16:55
*** marcoemorais has joined #openstack-infra16:55
*** derekh has quit IRC16:55
*** rmcall has quit IRC16:56
*** rmcall has joined #openstack-infra16:57
*** harlowja_away is now known as harlowja16:58
*** emagana has quit IRC16:59
*** loki184 has quit IRC16:59
*** mmedvede has joined #openstack-infra17:01
*** garyh has quit IRC17:01
*** chuckC has left #openstack-infra17:01
*** pballand has quit IRC17:01
*** pballand has joined #openstack-infra17:02
*** emagana has joined #openstack-infra17:03
fungijeblair: nibalizer: is 118485 safe to approve now?17:04
*** garyh has joined #openstack-infra17:05
jeblairfungi: yep17:05
*** amcrn has joined #openstack-infra17:05
*** loki184 has joined #openstack-infra17:06
*** arnaud has joined #openstack-infra17:07
*** arnaud__ has joined #openstack-infra17:07
*** kgiusti has joined #openstack-infra17:08
*** denis_makogon has quit IRC17:10
*** loki184 has quit IRC17:12
*** rushiagr is now known as rushiagr_away17:13
*** andreykurilin has quit IRC17:13
*** _nadya_ has quit IRC17:15
*** rmcall has quit IRC17:18
clarkbfungi: did you catch my notes on gdb'ing python?17:19
clarkbfungi: I wasn't able to load symbols properly or something because I couldn't get line numbers. but it looks like a python bug17:19
clarkbwhich is cool17:19
*** rmcall has joined #openstack-infra17:20
*** lcheng_ has quit IRC17:20
fungiclarkb: saw but i haven't gotten back to it yet--sidetracked by catching up on e-mail17:21
*** emagana has quit IRC17:21
clarkbfungi: tl;dr is in Py_Exit garbage collection segfaults17:21
clarkbit doesn't happen 100% of the time17:21
fungiyeah, that's very, very weird17:21
clarkbdhellmann: are you better with python + gdb than I am?17:21
clarkbdhellmann: trying to make the backtrace in useable17:22
*** cnesa9 has quit IRC17:22
fungii have to assume if it were this easy to trigger, it would be fixed... maybe there's a patch in later point releases and we just need to pester ubuntu to backport17:22
clarkbits possible. I hadn't looked at the python bug tracker yet. was trying to get line numbers and things out of gdb17:22
jeblairclarkb: did you install python3.4-dbg ?17:23
clarkbjeblair: I did. I wasn't running the dbg interpreter though just loading the symbols. The etherpad has the command I ran17:23
*** HeOS has joined #openstack-infra17:23
*** craigbr has joined #openstack-infra17:24
*** emagana has joined #openstack-infra17:25
openstackgerritA change was merged to openstack-infra/config: Updating node defn for puppet 3 master
jeblairnibalizer: okay, here goes ^17:26
nibalizerjeblair: wooo17:26
*** craigbr has quit IRC17:26
jogofungi: ping
jogofungi: do you think setting the download cache is worth it even if we don't prepopulate the cache?17:27
clarkb there is that but that is for 2.7.417:27
jogofungi: I am thinking it can't hurt but may not help very much17:28
clarkbjogo: it probably is, just because we have so many common deps17:28
clarkband C compilation will be sped up too right?17:28
jogoclarkb: yeah, well in devstack since we don't use venvs we only install things once17:28
clarkboh right17:28
clarkbso I don't think it will hurt but it won't actually help anything?17:28
jogoclarkb: this actually could help with unit test mirrors17:28
jogounit test nodes *17:29
clarkbjogo: except we don't reuse those nodes17:29
clarkbjogo: so same situation17:29
jogoclarkb: oh we throw them out between runs?17:29
jogoclarkb: with all of our !python deps that makes sense17:29
*** jp_at_hp has quit IRC17:29
fungijogo: maybe prepopulating would help, but we might be limited on space for that. knowing about how much room we'd need on the snapshot might help us figure out whether that's viable at all17:30
*** dtantsur|brb is now known as dtantsur17:30
jogofungi: that should be easy to sort out, I can do a pip install -r global-reqs17:30
*** marcoemorais has quit IRC17:30
jogoand see how big that is17:30
*** marcoemorais has joined #openstack-infra17:31
* clarkb gives gdb python another go. do ping me if there is any puppet3 related things I can help with17:31
fungijogo: if it's ~100bb then that's likely something we could manage without too much trouble17:31
*** dizquierdo has quit IRC17:32
*** MaxV has joined #openstack-infra17:32
*** marcoemorais has quit IRC17:32
fungiclarkb: sure thing. i'm about to take a swing through all the voting pypy jobs too. if those all run fine then there's a quick win by just changing its node label override once 118859 merges17:33
*** marcoemorais has joined #openstack-infra17:33
*** e0ne has quit IRC17:33
clarkbfungi: sounds good17:33
fungiwell, once that merges and we get new nodepool bare-trusty images anyway17:33
*** marcoemorais has quit IRC17:33
*** mwagner_lap has joined #openstack-infra17:33
*** marcoemorais has joined #openstack-infra17:33
jeblairclarkb: don't go far :)17:34
clarkbfungi: does pypy need a dev package of some sort?17:34
fungiclarkb: not sure but i'll know soon17:34
clarkbfungi: should probably install pypy-dev as well?17:34
clarkboh right ok :)17:34
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Fix variable ref error in launch node
jeblairclarkb: ^17:35
clarkbjeblair: approved17:36
jeblairi applied that manually in my checkout and am running it now; puppetmaster is building17:36
*** rushiagr_away is now known as rushiagr17:36
*** MaxV has quit IRC17:36
* nibalizer does a happy dance17:36
*** garyh has quit IRC17:37
*** jab has quit IRC17:37
*** Sukhdev has quit IRC17:37
*** Sukhdev has joined #openstack-infra17:38
*** rhsu has joined #openstack-infra17:38
*** emagana has quit IRC17:38
*** Ryan_Lane has joined #openstack-infra17:39
*** emagana has joined #openstack-infra17:39
clarkbon the jshint front. we did remove jshint from horizon once before17:39
dhellmannclarkb: I wonder if it would work better if you weren't using a virtualenv wrapped python?17:40
clarkbdhellmann: ya I think that may be part of it17:40
clarkbdhellmann: maybe I should sacrifice a node to the cause17:40
clarkbit is worth a shot /me holds a trusty node17:41
dhellmannclarkb: times like this make me wish there was an easy way to turn a random vm into a test node. I guess there's a way to do that with puppet, but I don't know it well enough.17:41
dhellmannsomething equivalent to chef-solo, that doesn't need a server17:41
jeblairnibalizer, clarkb:
jeblairall of that was red error output from the end of the initial puppet run17:42
jeblairhowever, it did not appear to cause the server to be deleted17:42
clarkbdhellmann: ya, it is one of the reasons we want to push for more dib (and in the case of dox docker) in our env. To make get test node simple17:42
*** david-lyle has quit IRC17:42
dhellmannclarkb: that makes sense, too17:43
*** emagana has quit IRC17:43
clarkbjeblair: puppet can return 0 even when it fails...17:43
openstackgerritEdward Raigosa (wenlock) proposed a change to openstack-infra/config: lsb_release is missing for debian setup
clarkbjeblair: in this case it may help the debugging so probably not a terrible thing17:44
jogofungi: hmm is there a good way to populate the download cache without installing things?17:44
nibalizerjeblair: :/ that doesn't look good17:44
jeblairclarkb: yeah, i'm tired of doing this, and i don't intend to spin up another node17:44
jeblairif someone wants puppetmaster bootstrap to work, they can fix it on their own time; we need to keep moving.17:44
nibalizercan't negotiate its environment17:44
rhsuanteaya: vmware nsx ci seems to have lost the ability to vote on patches as of 2 days ago. any way you can help us with authentication issues?17:45
fungijogo: i believe there's a no-install option to pip, but it will cause it to miss setup_requires and other things where packages shell out to setuptools/easy_install to get their deps in some non-pip-oriented way17:45
jeblairwell, i say that, but apparently i can't sudo on that node.17:45
*** _nadya_ has joined #openstack-infra17:45
fungijogo: though those also tend to not benefit from the pip cache anyway, so might not matter17:45
*** masayukig has quit IRC17:45
clarkbjeblair: we almost need a launch-node option that says give me node and do everything but tie it to puppet. also splat this pubkey in it for me17:46
clarkbjeblair: I guess that is just nova boot17:46
nibalizerclarkb: what about just running util/ ?17:46
clarkbthe biggest thing we lose is the ipv6 check17:46
fungijogo: it also might make sense, once devpi can cache from bandersnatch mirrors, to use that to create the local cache17:46
nibalizeri guess with no sudo can't do that17:46
nibalizerand hacking for a one off sounds gross too17:46
jeblairnibalizer: any chance the 3.6 change broke the ability to talk to the 2.7 puppetmaster?17:47
nibalizerjeblair: absolutely17:47
nibalizeryou're not supposed to be able to have 3x nodes talk to 2.7 masters at all17:47
jeblairokay, well, looking at the early errors, that seems to be what's happening.17:47
nibalizerthats why i put a bunch of effort into, so it could bootstrap itself17:47
jeblairokay, well, that's not how we launch nodes17:48
nibalizerya, i learned that17:48
nibalizerso idunno what to do, what do you guys think?17:48
*** david-lyle has joined #openstack-infra17:48
*** amotoki has quit IRC17:48
clarkbpuppet masters being the bootstrappers may need to be a special class of thing that we nova boot directly?17:48
clarkbor have a puppet apply vs puppet agent thing?17:49
openstackgerritEdward Raigosa (wenlock) proposed a change to openstack-infra/config: give install_modules options for loading requirements externally
*** rushiagr is now known as rushiagr_away17:51
*** masayukig has joined #openstack-infra17:51
mordredI've always bootstrapped puppetmasters by hand before17:52
mordredI'm not saying it's a great plan17:52
clarkbso for all of the trouble tox gives us, trying to run python3 things directly is a pain17:53
openstackgerritEdward Raigosa (wenlock) proposed a change to openstack-infra/config: give install_modules options for loading requirements externally
clarkbmordred: don't forget to address that in dox17:53
mordredwhat if we spin up the puppet3 master as a puppet2 master pointed at the current puppet2 master17:53
mordredthen land a change that tells the puppet3 master to look at itself as its puppetmaster17:53
jeblairmordred: i don't think we can do that on a trusty vm17:54
mordredthen land a change to upgrade it to puppet317:54
mordredoh. right. damn17:54
jogofungi: doing a 'pip install --download -r requirements.txt' tries to compile things17:54
clarkbjogo: you want it to17:54
mordredjeblair: I'm pretty sure there's going to be some by-hand then17:54
jogofungi: doing a 'pip install --download /cachedir -r requirements.txt'* tries to compile things17:54
clarkbjogo: so that things are built and cached17:54
jogoclarkb: that means the non-python dependencies need to be installed in the system17:54
fungijogo: yeah, stuff gets built, just not installed17:55
openstackgerritEdward Raigosa (wenlock) proposed a change to openstack-infra/config: provide an option in plugin to pre-update a plugin with puppet
nibalizerwe could hand roll a puppet 3 master real fast, then tell the real puppetmaster to bootstrap it off the puppet 3 master, kindof the tripleo approach17:55
nibalizerbut thats kinda silly17:55
clarkbjogo: yes, that is the case17:56
clarkbjogo: so it may interfere with devstack17:56
jogoclarkb: yeah, although it would just mean the packages are installed before hand17:57
*** mrmartin has joined #openstack-infra17:57
fungijogo: that moves a lot of the devstack package-install logic out of devstack though...17:57
jeblairnibalizer, mordred, clarkb, fungi: i am going to build a puppet3 master by hand.  it will be completely unreproducible, but we're out of time on this project.17:58
mordredjeblair: ++17:58
fungijogo: if the proposal is preinstalling system packages devstack wants so that we can compile the pip packages17:58
clarkbjeblair: ok17:58
fungijeblair: unfortunate, but i agree on the urgency17:59
*** pballand has quit IRC17:59
*** rmcall has quit IRC17:59
nibalizerjeblair: ok17:59
*** Sincler has joined #openstack-infra18:00
*** pballand has joined #openstack-infra18:00
jogofungi: yes, at least they way I am trying this it looks like we would need to preinstall system packaages that devstack wants to build the pip download cache. Is that a problem? If so there may be other options18:00
*** koolhead17 has quit IRC18:01
*** rmcall has joined #openstack-infra18:01
jeblairjogo: what is the end goal of the project you're working on?18:01
clarkbI cannot import oslo.messaging after pip3 install . in oslo.messaging18:01
fungijogo: well, it causes us to no longer notice in testing if devstack ceases to actually try to install one of the packages it needs because now we've already installed that where we're testing18:01
jogojeblair: following through with
jogofungi: after running 'pip install -d /pipcache/ -r global-requirements.txt18:02
jogopipcache is 193MB18:02
*** ociuhandu has quit IRC18:03
jogojeblair: exploring what the implications of pre populating a pip download cache are18:03
fungijogo: that's probably a reasonable amount if we decide that pre-caching those is warranted18:03
fungijogo: though note that you probably only tested master18:03
clarkbhrm oslo.messaging isn't ending up in /usr/local/lib/python3.4/dist-packages18:03
jogofungi: true, but doing master is a good start at least18:03
fungijogo: yeah, i wouldn't expect that adding the stable branches would bump it up by more than another 10-20%18:04
clarkbneeded a --upgrade?18:04
jeblairjogo: nothing in that change says anything about pre-populating18:04
clarkbok reproducing without tox/virtualenvs now18:04
*** ramashri has joined #openstack-infra18:05
jeblairjogo: also, we used to pre-populate but stopped; why do you want to start again?18:05
jogojeblair: yes that patch doesn't say anyhting about prepopulating but since we don't us virtualenvs it wouldn't be useful without pre-populating18:06
jogojeblair: not saying we should, just exploring the idea18:06
jogojeblair: why did we stop? because we have the mirrors. The thinking is if we pre-populate we have less network traffic18:07
jogojeblair: but once again this is just exploring the implications18:07
*** reed has quit IRC18:07
openstackgerritA change was merged to openstack-infra/config: Fix variable ref error in launch node
*** emagana has joined #openstack-infra18:08
jeblairLast login: Thu Jan  1 00:00:10 197018:09
jeblair^ from root's account on the puppetmaster18:09
clarkbdhellmann: woot that worked thanks.18:10
clarkbfungi: dhellmann etherpad now has a potentially useful bt18:10
jogofungi: so I don't see a way to pre-populate via 'pip install --download' that doesn't require system packages that devstack usually installs18:10
*** koolhead17 has joined #openstack-infra18:11
*** atiwari has joined #openstack-infra18:11
clarkbnow to poke around in this stack18:11
*** _nadya_ has quit IRC18:13
fungijogo: yeah, probably a dead-end then, unless we do something like maintain a devpi cache which we rsync into new images as they're being built, or maybe the "put our mirror on afs and attach to it from clients on workers" suggestion brought up at the mid-cycle18:13
*** pballand has quit IRC18:13
*** yamahata has quit IRC18:13
jeblairi think it's okay to install something like build-essential on the node, but we're trying to simplify node types, so i don't think we'd want any specific library dev packages installed18:14
dhellmannclarkb: nice18:14
jeblairis "/etc/kernel/postinst.d/apt-auto-removal" what i think it is?  if so, i'm very excited18:15
clarkbdhellmann: I think testtools is throwing an exception18:15
jogojeblair: so this would need things like 'libffi-dev' so I think that is a no-go18:15
clarkbdhellmann: and that is getting into exit under duress code18:16
*** jamespage_ has joined #openstack-infra18:16
jeblairjogo: yeah :/  does it by any chance keep going after failing to build that, or does it blow up?18:16
jogofungi: I wonder if there is a way to tell devpi what to pre-populate18:16
jogojeblair: blow up18:16
jogojeblair: let me see if I can make it continue18:16
jogojeblair: the internets say do cat requirements.txt | while read PACKAGE; do pip install "$PACKAGE"; done18:18
jeblairjogo: we could iterate over all the packages individually and ignore errors18:18
jeblairsomething like that :)18:18
jeblairthat might even be what we used to do18:19
jogojeblair: I'll try that idea18:19
fungijogo: though also, using devpi for that needs, as a prerequisite, the current work in progress to make it possible to devpi to a bandersnatch mirror (currently doesn't work)18:19
jeblairjogo: i'm still not really sure what we're trying to accomplish.18:19
jogojeblair: reduce the amount of network traffic during the jobs18:20
mordredyeah. pip download caches dont' avoid network traffic - pip still pings the mirror index for each package18:20
jogojeblair: as we have seen networks fail on us18:20
jeblairjogo: sure, but we JUST created per-region mirrors, which is the opposite of what you're doing18:20
jogomordred: yes, it doesn't avoid traffic but reduces it18:20
jeblairjogo: the failure modes we're seeing with networks will likely still occur as long as there is any network traffic18:21
*** johnthetubaguy is now known as zz_johnthetubagu18:21
*** pballand has joined #openstack-infra18:21
mordredI think honestly fixing the network inside of our cloud regions is gonna get us further18:21
jogomordred: ++18:22
jogojeblair mordred: I am just exporing the idea not saying we should do it18:22
mordredthat said, I would love to see our old pypi-mirror script run on our requirements file in --create-wheel-only mode to make partial mirrors with pre-built wheels in it18:22
jogomordred: yes fixing the network is way better18:22
jogomordred: you mean 'pip wheel'?18:22
mordredif we build wheels and add them to our mirrors, then we can spend less time building and rebuilding lxml18:23
*** garyh has joined #openstack-infra18:23
mordredwhich would likely add up in crunch time18:23
jeblairjogo: okay cool.  i'm trying to strike a balance between helping you out while not misleading you about whether we've decided to actually do this :)18:23
*** morgan is now known as morganfainberg18:23
jogomordred: I run 'pip wheel -r ' on my box on a regular basis18:23
*** morganfainberg is now known as morgan18:24
mordredjogo: yah. I have a patch up to pypi-mirror to be able to splat a wheelhouse into a pypi mirror format18:24
jogomordred: nice18:24
*** koolhead17 has quit IRC18:24
mordredbut I havent' gotten aroudn to doing anything with it18:24
*** arnaud has quit IRC18:25
*** arnaud__ has quit IRC18:25
jogojeblair: at this point my take is pre-poluating the download cache isn't the best choice18:25
*** e0ne has joined #openstack-infra18:26
jeblairclarkb, fungi, mordred: rick says don't bother trying to match instance az to neutron virtual router host az; apparently that latency is not a big deal18:26
*** jamespage_ has quit IRC18:27
*** yamamoto has quit IRC18:28
*** e0ne has quit IRC18:28
mordredjeblair: but what about failure domains18:28
mordredshould we match them to make sure that if all of az2 fails we don't also lose the network on some of our other nodes?18:29
jeblairgood question18:29
mordredsince we treat each router as a "cloud" anyway18:29
jeblairmordred: you want to ask (you're still on the thread), or shall i?18:29
mordredit would be fairly trivial to include a strict az config option in each cloud as well18:29
jeblairmordred: yeah, we already have that and in fact use it due to capacity issues in az118:30
mordredlet me ask18:30
*** todd_dsm has joined #openstack-infra18:30
*** koolhead17 has joined #openstack-infra18:31
*** Morgan_ is now known as morgan_remote18:31
mordredjeblair: actually, why don't we just match them for the hell of it?18:32
jogoif we skip failed installs, the pip download cache is 100MB18:32
jogohopefully we can fix the underlying network issues and not have to do this18:32
*** ociuhandu has joined #openstack-infra18:33
*** emagana has quit IRC18:33
*** rushiagr_away is now known as rushiagr18:33
*** emagana has joined #openstack-infra18:34
*** todd_dsm has quit IRC18:34
jeblairnibalizer: okay, the puppetmaster is bootstrapped; what do we need to do with the certs?18:34
mordredjeblair: which certs?18:34
jeblairi'm trying to recall the part of the migration plan regarding the puppet ca18:35
*** todd_dsm has joined #openstack-infra18:35
jeblairalso, we need to figure out what to do to have the puppetmaster be able to run puppet on itself18:35
jeblair(regularly; it's completed the first run)18:36
mordredoh. you know what?18:36
mordredwe should update the ansible script18:36
mordredit should not run just puppet agent --test18:36
mordredit should run puppet agent --test $hostname_of_locahost18:36
mordredso that if you run it on ci-puppetmaster, it will tell nodes to connect to ci-puppetmaster18:37
mordredand if you run it on puppetmaster, it will tell nodes to connect to puppetmaster18:37
mordredit should run puppet agent --test --puppetmaster $hostname_of_locahost18:37
mordredlemme make a patch18:37
jeblairmordred: wait18:37
jeblair/etc/puppet.conf has ca_server = ci-puppetmaster.openstack.org18:38
nibalizermordred: i think you want --server18:38
*** vhoward has left #openstack-infra18:38
jeblairthat's probably what needs updating18:38
*** emagana has quit IRC18:38
*** koolhead17 has quit IRC18:38
nibalizerjeblair: we dont have a migration plan regarding the puppet_ca18:38
nibalizerwe can make one though18:38
jeblairanyway, nobody go and do anything yet :)  i want everyone here to talk about this18:38
jeblaircause there's like 50 interlocking things we need to work out18:39
mordredso - the puppet.conf value has to come from somewhere18:39
nibalizerfor the certs on puppetmaster.o.o (the p3 master) you'll need to puppet cert generate on the ci-puppetmaster, the copy over the 4 important files18:39
mordredthat sounds too complex18:39
jeblairlets take notes here:
mordredwhy would we do that?18:39
nibalizerthose being, /var/lib/puppet/ssl/certs/{ca.pem,puppetmaster.o.o.pem} /var/lib/puppet/ssl/private_keys/puppetmaster.pem and wherever the crl is kept18:40
mordredI think there is a very simple path taht does not involve the puppetmasters knowing about each other18:40
*** e0ne has joined #openstack-infra18:40
mordrednibalizer, jeblair: can we talk here fora  sec before we do etherpad?18:41
mordredthere is a fundamental thing I think we need to sync on before we make plans for doing it18:42
jeblairsure, i'm not trying to make plans18:42
jeblairi'm trying to figure out what all the things are that need changing18:42
jeblairlike where does ca_server get set18:42
*** e0ne has quit IRC18:43
nibalizermordred: sure18:43
mordredI don't know why we're setting ca_server18:43
nibalizercan i take a stab at explaining that?18:43
jeblairah, that's only set on the new puppetmaster18:44
*** rainya has quit IRC18:44
nibalizerthe puppet master brings its own CA, each master, by default, assumes it is its own CA, we want to set the new puppetmaster to use the old puppetmaster for CA activities18:44
mordredright. I think that's a mistake18:44
nibalizerwe want to do that, instaed of two separate CA's, because then all clinets can check in to both masters18:44
mordredI think we _want_ each puppetmaster to be a completely seperate CA18:44
mordredwe dont' want that18:44
jeblairmordred: part of the plan was to move one client at a time18:44
mordredwe _Want_ each master to have a completely separate cert list18:45
mordredbut we drive puppet via ansible via puppet cert list18:45
mordredso there is a very simple mechanism, if we don't try to make both talk to both18:45
mordredwhere we revoke the cert on teh old master18:45
mordredsign it on the new18:45
*** rainya has joined #openstack-infra18:45
mordredand each node will only ever talk to one and only one master at any time18:45
mordredand we can do it node at a time18:46
nibalizerthe question is when is a node ready to move to puppet 318:46
nibalizeri think the iteration loop should be for a rooter to go to a node, run puppet with --server=puppetmaster.o.o --noop and see if there are changes18:47
jeblairmordred: we have to be _very_ careful with that; we will need to log into each host and verify cert fingerprints, etc.18:47
nibalizerif there are changes we need to improve the puppet codebase until there are not changes18:47
*** maxbit has quit IRC18:47
nibalizeronce there are no changes, the node can be set to the new puppetmaster and the client software upgraded18:48
mordredjeblair: so you're saying use the old puppetmaster for the ca until everythign is moved, then move the whole CA over?18:48
*** bknudson has quit IRC18:48
jeblairmordred: i did not say that.18:48
jeblairmordred: i think nibalizer is saying that?18:49
*** Sukhdev has quit IRC18:49
nibalizeri think we use the old puppetmaster durring the transistion to puppet 318:49
nibalizerthen the last thing that goes over is the CA18:49
*** pelix has quit IRC18:49
nibalizerbut i haven't done too much thought on what the final state of the CA should be, so i'm kinda punting on that18:50
nibalizerwe can have that discussion now if you like18:50
nibalizerbut i have nothing prepared18:50
jeblairnibalizer: when you say we run puppet --noop, aren't we still running puppet2?  is that going to tell us what we want to know?18:50
fungithe plan nibalizer outlines makes sense. otherwise we're talking about hacking certs back and forth while we iterate on patches to get puppet3 working on a given machine, or finding some way to test applying from the puppet3 master without actually being trusted by it18:50
nibalizerjeblair: yes, because the providers and such( the client code) dont change at all really, the big changes is in catalog parsing and generation (server side)18:51
*** julim has quit IRC18:51
*** garyh has quit IRC18:51
jeblairright, so we would need to address mordred's point about the ansible iteration18:51
jeblairwhich currently comes from puppet cert list18:51
*** todd_dsm has quit IRC18:51
mordredthis is an absolute mess and puppetlabs should be VERY ashamed of themselves18:52
jeblairnibalizer: what determines which puppetmaster a node contacts?18:52
nibalizerthis is a longer version of the plan:
jeblairthat line in puppet.conf18:52
clarkbjeblair: its puppet.conf. you can override with puppet agent --server18:53
*** emagana has joined #openstack-infra18:53
nibalizerjeblair: yea server = <server> is set in puppet.conf, and can be ovevrided18:53
*** julim has joined #openstack-infra18:53
nibalizeri *think* we have puppet machinery in place to set that in site.pp durring the transition, i could be wrong, its an easy fix if not18:53
mordredyah. that's what I was saying earlier about having the ansible script pass in --server in its invocation, so that if it's ansible on the new puppetmaster driving the run, it will point it at itself18:53
jeblairso the act of actually switching a node means we either update site.pp for each node to override the master, or use mordred's ansible idea to drive the selection from the puppetmaster18:54
jeblairi kind of like the second one because it's less config/site.pp churn18:54
*** vigneshvar has joined #openstack-infra18:54
*** ZZelle_ has joined #openstack-infra18:54
mordredI'm still not sure when we get puppet3 on the new node18:54
mordredon the node18:54
nibalizermordred: on the client?18:54
mordredI _know_ that puppet3 talking to a puppet2 master breaks18:54
mordredsince that's broken for us18:55
*** bknudson has joined #openstack-infra18:55
mordredis puppet2 talking to a puppet3 master supposed to work?18:55
nibalizerso once you run puppet agent --test --server=old and puppet agent --test --server=new and they do the same thing (which is hopefully nothing) then you're clear to move the client to connect only to the p3 master and upgrade the clinet software18:55
vigneshvarhi guys. Kindly review my patch
nibalizermordred: yes p2 client to p3 server is guaranteed to work18:55
wenlockyou don't plan to rebuild the new node for trusty to run with the 3 option?18:55
nibalizerfor the next 20 some-odd days18:55
mordredwenlock: no, we do not18:55
nibalizerthats another option, is the wipe-out, rebuild18:55
mordredno. it's not18:55
mordredit's not an option18:55
mordredI will replace puppet wholesale before I accept a plan that isa  wipe and rebuild of all of our nodes18:56
jeblairmordred: ++18:56
nibalizerokay, so we gotta be pretty sure then that the puppet code/ ansible script that does the puppet client upgrade it good18:56
fungivigneshvar: discussion of devstack changes usually happens in #openstack-qa18:56
*** doug-fish has joined #openstack-infra18:56
nibalizerwe dont want a bunch of squiched nodes18:56
vigneshvaranybody who can help me with "MySQL is gone away error", I keep getting this sometimes and it works other times18:56
vigneshvarfungi: Thanks18:57
mordredjeblair: my ansible idea was predicated on the idea of moving the certs one at a time too18:57
mordredjeblair: if we're going to keep one CA and move it at the end, the ansible set the server trick won't work I don't think18:57
nibalizerwell hold on a second18:57
fungimordred: what's your proposal for testing/reversing it before committing to leaving it migrated?18:57
jeblairmordred: no, we'd need another way of listing.  we could switch to using a static file for now18:57
mordredjeblair: we could for sure18:58
mordredthat would be fairly easy18:58
nibalizerif we set sever = <server> in puppet.conf ; and then ansible runs puppet agent --test; then ansible only needs to run on one machine18:58
mordrednibalizer: yes, but then we have 76 patches to site.pp18:58
nibalizerthe nodes will check into the correct machine because the correct server is written in their puppet.conf18:58
*** yamamoto has joined #openstack-infra18:58
nibalizermordred: we'll need those anyways because we have to set puppet_version => 3.6 on every client node anyways18:58
fungito be fair, it's probably nowhere near 76... we could batch related systems. it's still a fair number of changes though yes18:58
mordrednibalizer: not at the same time18:58
mordrednibalizer: if puppet2 client can talk to puppet3 master18:59
*** rmcall has quit IRC18:59
mordredwe can migrate to puppet3 master without upgrading puppet client anywhere18:59
mordredand then go back and start migrating puppet3 clients one at a time18:59
nibalizeryes, but only for 20 days, the puppet client software is EOL sept 30 too18:59
*** radez is now known as radez_g0n318:59
jeblairthat's slightly less concerning as its not listening on the network.19:00
mordredif we're sure about the p2client->p3master ... then transitioning to the new puppetmaster is probably only a few hours work19:00
mordredbecause we don't need to test that new puppet can run the old code, yeah?19:00
nibalizermordred: oh hold on let me be very clear very very clear19:00
nibalizerthe puppet code and how its interpreted from 2.7 to 3 is a change, that happens on the master19:01
*** e0ne has joined #openstack-infra19:01
nibalizerso there could be changes that break our infrastructure if we just switch all the clinets to the new master19:01
mordredoh holy jesus19:01
nibalizerwhat is guaranteed, is that there is not some line-protocol bug like you saw earlier between a 2.7 client and a 3 master19:01
mordredso we actually have three different transitions19:01
nibalizerexplain that19:02
*** dustins has quit IRC19:02
*** rushiagr is now known as rushiagr_away19:02
*** dtantsur is now known as dtantsur|afk19:02
mordredwe need to verify that puppet3 master parses the old code, this is done by trying to run puppet on a node against the new master19:02
mordredoh - only two, my bad19:02
*** yamamoto has quit IRC19:02
mordredthen we need to verify that puppet3 _client_ can actually execute what's passed to it19:02
mordredonce we upgrade the client19:03
*** derekh has joined #openstack-infra19:03
nibalizermordred: i don't think we need to sweat that, they didn't change provider behavior19:03
nibalizerlet me get the release notes19:03
*** dustins has joined #openstack-infra19:03
mordrednibalizer: no offense, but I think at this point the track record is that we need to empirically test and not trust19:03
mordredthat this is a breaking change at all proves that the people sitting in that office do not know anything about people running large infrastructures19:04
mordredand I'm very angry with them19:04
nibalizer--noop testing sufficient for you?19:04
nibalizerrelease notes:
mordredno idea. I suppose it has to be perhaps?19:04
*** e0ne has quit IRC19:04
nibalizerwell actually, i guess we're taking every change in the 3.x series19:05
mordredfungi: so - I guess the testing plan is what nibalizer is saying with --noop19:05
mordredjeblair: I just  had a terrible idea19:05
nibalizerso step 1: puppet 2.7 client on puppet 3 master with --noop; step2: puppet 3 client on puppet 3 master with --noop ; if both cause no changes, then we're ok to upgrade and switch that node?19:05
mordrednibalizer: we don't really have a way to test step 2 without just upgrading I think19:06
fungimordred: yeah19:06
mordredso we may have to trust the statement that the danger is in step 119:06
mordredjeblair: what if ...19:07
mordredwe expand the ansible playbook for this19:07
*** marun has quit IRC19:07
mordredand have it, for every node, run puppet --noop --server puppetmaster.openstack.org19:07
nibalizermordred: we could play a game where we do rvm in /tmp/ and put puppet 3 in it and try19:07
mordredthen, if that _fails_ ... it runs puppet --server ci-puppetmaster.openstack.org19:07
mordredand if it succeeds, it runs puppet --server puppetmaster.openstack.org19:07
jeblairmordred: i think we're getting too clever19:08
nibalizermordred: you're not catching all the failure modes19:08
jeblairmordred: i think that's going to hide errors and stuff19:08
jeblairwe need to actually see this ourselves, and probably run it a few times to converge even...19:08
jeblairso i believe i've written out the tentative plan in the etherpad...19:08
nibalizerthe commoon failure mode to expect is a variable that was in scope for 2.7 is out of scope for 3, so your template has whitespace instead of confiugration dircetives19:08
jeblairbut i think the small detail of "actually upgrade from 2 to 3 on the client" is not small19:09
*** markmcclain has quit IRC19:09
nibalizeryou made it sound like you were flexible on running 2.7 clients after sept 30th though, just not the master19:09
jeblairnibalizer: yes, but i'm not eager to19:10
mordredjeblair: I think that is a great plan for step 119:10
*** amuller has quit IRC19:10
mordredand I think we should execute and finish it and then circle back to work on upgrading teh clients19:10
*** masayukig has quit IRC19:10
mordredmaybe we can install puppet3 into an rvm for the purposes of testing without fully uprading?19:10
jeblairmordred: i'm not entirely sure what you're saying?19:11
jeblairmordred: what's a great plan?  and what's step1?19:11
mordredjeblair: your plan in the etherpad is great19:11
nibalizerim confused about line #919:11
mordredjeblair: step one is "do the puppetmaster transition with clients on puppet2"19:11
nibalizercopying the CA over is not what we discussed, do you mean copy only the puppetmaster.o.o certificates?19:12
jeblairnibalizer: no i meant the whole thing.  but you're saying we don't need to do that because of ca_server19:13
jeblairnibalizer: right?19:13
fungiwell, eventually the new puppetmaster will need the ca too, but it can be copied over at any time before we start asking it to sign certs for new servers19:13
nibalizerjeblair: yes19:13
jeblairnibalizer: that will cause puppet3master to do what exactly?19:13
nibalizerwe should be setting ca_server = ci-puppetmaster.o.o on puppetmaster.o.o19:13
openstackgerritpatrick-crews proposed a change to stackforge/dox: Document multiple images in dox.
nibalizerit will reject certificate signing requests, and it won't build a CA in its own /var/lib/ssl and it won't sign its own master certificate19:13
jeblairnibalizer: right, how will it do validation?19:14
*** ramashri has quit IRC19:14
nibalizerwe will copy over the ca.pem and the crl.pem which it uses to do validation and crl verifying19:14
nibalizerits a bit suqirly about refreshing its crl.pem, if you like we can add a cronjob to be explicit about refreshing it daily or something19:14
fungica_server = ci-puppetmaster.o.o tells it to use the ci-puppetmaster's self-cert to confirm the validity of server certs19:15
fungiif i'm understanding19:15
jeblairfungi: that's a statement, not a question, right?19:15
openstackgerritEmilien Macchi proposed a change to openstack-infra/config: Introduce tripleo-puppet-elements into Stackforge
*** Ryan_Lane has quit IRC19:15
jeblairokay, that sounds like a question now19:15
nibalizerfungi: eh not really, i mean yes19:16
nibalizerreally verification of the clietn certificate is done in the apache configuration file19:16
nibalizerpuppet never touches it19:16
fungibut the ca (the actual private key on ci-puppetmaster used to sign new certs) can just be copied over at some point before we complete the transition and start creatign new servers which the new master needs to be able to sign19:16
*** masayukig has joined #openstack-infra19:16
funginibalizer: am i understanding correctly?19:17
nibalizerfungi: im an ops person19:17
jeblairnibalizer: okay, so validation happens in apache with the ca public key.  what does ca_server actually get us then?19:17
nibalizerso 'you CAN' do all kinds of things19:17
jeblairnibalizer: we don't actually do any signing, so why not just copy the whole thing over?19:17
nibalizerif you run the puppet master application, without any special configuration, it will look for its CA, if it doesn't find one, it will generate one19:17
nibalizerca = false; ca-server = <something else> turns off that behaviour19:18
fungithis is a good point. if we either freeze creation of new servers while we're doing this, or only create them from the new puppetmaster and only with puppet3 then the trust referral to the old puppetmaster is irrelevant i think19:18
fungisince signing is a manual process for us19:18
*** mgagne has quit IRC19:18
*** dtroyer has quit IRC19:19
fungiwe create the signature (certificate) when we generate the server key at launch time19:19
*** mgagne has joined #openstack-infra19:19
*** mgagne is now known as Guest7525019:19
jeblairfungi: it sounds like it's really just a convenient way to disable the ca generation on the new master19:19
jeblairbut copying the whole ca would do that as well19:20
nibalizerokay but copying the whole CA is the thing i absolutely reccomend against19:20
funginibalizer: can you formalize the risk statement for that?19:20
nibalizerunless you make the guarantee that you wont sign certs on one of the locations19:20
*** dtroyer has joined #openstack-infra19:21
nibalizerwhen certs are signed their fingerprint and index number is written down in a file in there somewhere19:21
fungiit sounds like it's fine to sign certs in the new location if we freeze the copy on the old puppet 2 master19:21
nibalizerand its just sequential19:21
nibalizerand its used to revoke19:21
nibalizerso if you sign in two places you can't really revoke reliably any more19:21
wenlocknibalizer, thats the inventory.txt19:21
fungiyeah, "in there somewhere" is the serial file19:21
nibalizerya that^^19:21
wenlockyou would have to consider maintaining the file in a copy i supose19:22
nibalizerno i dont fully understand why we dont look at fingerprints, but i've had the fear of god put into me on this from jeff mccune19:22
*** _nadya_ has joined #openstack-infra19:22
*** KanagarajM has quit IRC19:23
nibalizersome people do stuff like put the CA on two servers, but use nfs to keep only one copy, or inotify/rsync to keep both directories fully in sync19:23
fungiif we can guarantee (upon pain of additional certificate regeneration) that we will stop signing server certs on the old puppetmaster *and* only use the new puppet master to launch new servers, then it shouldn't be a problem, right?19:23
jeblairfungi: i think you've probably hit the issue -- if we want to launch new servers under puppet2, we should use ca_server; if we want to launch them under puppet3, we should copy the ca over19:24
mordredI vote for copy - we shoudl not be launching anything new on puppet 2 at this point, right/19:26
mordredI mean, that would just be crazy new work19:26
jeblairmordred: well, if we needed to spin up a logstash or zuul worker immediately, it might be awkward.  but hopefully within a few days of starting this we'd be in that position.19:26
jeblairi think i'd be okay with that.19:26
fungiyeah, i think as long as the 5 of us with ssh access to the puppet masters agree not to sign new certs/launch new servers from the puppet 2 master starting now, we should be fine19:27
mordredand if we do need to spin up logstash or zuul workers - that will motivate us to fix puppet3 problems in those modules with haste19:27
jeblairnibalizer: think that can work? ^19:28
*** garyh has joined #openstack-infra19:28
*** yamamoto has joined #openstack-infra19:29
fungiand if so, that simplifies things a bit. no need for a transitive trust, no piecemeal syncing of the certificate authority and server certificate files, et cetera19:29
*** markmcclain has joined #openstack-infra19:30
*** emagana has quit IRC19:30
openstackgerritDoug Hellmann proposed a change to openstack-infra/infra-specs: Add RSS feed
openstackgerritDoug Hellmann proposed a change to openstack-infra/infra-specs: Remove templates from toctrees
openstackgerritDoug Hellmann proposed a change to openstack-infra/infra-specs: Use the current date for the copyright statement
*** emagana has joined #openstack-infra19:30
nibalizerI think ca_server = ci-puppetmaster.o.o is a much better solution, i don't see the advantage of trying to 'hurry-up' move everything off of the p27 master19:31
openstackgerritpatrick-crews proposed a change to openstack-infra/elastic-recheck: Adding query namespace to elastic-recheck
JayFanteaya: I wanted to make sure you knew, because I saw all the scrollback hulabaloo about elections... I got an email from OpenStack Foundation telling me b/c I hadn't voted in two years, I had to click a link to keep my account enabled. I clicked it and got an option to "Continue as a Member" or "Delete Account" ... made me suspicious if this method of account invalidation / deletion existed before19:32
JayFthe explicit "resign" button did19:32
*** yamamoto has quit IRC19:34
openstackgerritDoug Hellmann proposed a change to openstack-dev/specs-cookiecutter: Add RSS feed
clarkbzul: ^ I think that is a bugfix ubuntu may want to pick up19:34
*** emagana has quit IRC19:34
*** mriedem has quit IRC19:34
*** emagana has joined #openstack-infra19:34
clarkbzul: I need to do more testing and check that that actually fixes this thing. Which is exciting because building python? oh boy19:34
nibalizeri guess i think of a puppet run against 3 right now being the exception19:35
nibalizeras we gain confidence with our codebase on puppet 3, then it makes sense to default new nodes to puppet 3, but that doesn't make sense because we haven't even done initial testing19:35
*** berendt has quit IRC19:35
clarkbfungi: so, do we have to wait on a backport bugfix from ubuntu? not sure hwat you want to do next19:37
jeblairnibalizer: i think the issue is that we're all a bit more comfortable with having the ca be on one host.  it's easier to understand than the ca_server option.19:38
jeblairwe can live with the idea that it will be hard to spin up a new node of a type that has not been ported to puppet3 for the next two weeks19:38
jeblairand at the end of those two weeks, it would be hard regardless19:38
jeblairso unless there's a reason why it won't work, i think can go with the copy-the-ca option19:39
arosenHi, I was working on setting up 3rd party CI following jay's blog about a month ago. I ended up giving up and I want to start again. I was wondering if there is a newer document that talks about how to set this up that is being mantained now?19:39
arosenI believe asselin was working on something?19:39
jeblairclarkb, nibalizer, fungi, mordred: so we still have the problem on line 1819:40
openstackgerritDavid Shrewsbury proposed a change to stackforge/dox: Better support for multiple commands
nibalizerjeblair: im struggling here19:40
jeblairhow do we actually upgrade the client19:40
nibalizeri want the ca on one host19:40
jeblairnibalizer: okay, tell me why that's required19:40
*** dimtruck is now known as zz_dimtruck19:40
nibalizerthe group seems to be interested in copying it to another host, so that we ahve two19:40
nibalizeryou said "we're all a bit more comfortable with having the ca be on one host"19:41
nibalizerits already on one host, that host is working well19:41
nibalizerca_server = <server>/ca = false is the standard way to spin up a second puppet master19:42
nibalizerpeople do it for load scaling or for stuff like this where you're testing an upgrade19:42
*** derekh has quit IRC19:42
nibalizeroften, people will move the CA to a lightweight vm all by itself so CA services and puppetmaster services are totally separated19:42
nibalizerbut yes, I dont have a hard reason why we can't copy the cA19:43
nibalizerwe can do it, and we can get it to work19:43
*** rmcall has joined #openstack-infra19:43
nibalizerand good is great, but done is better19:43
fungiJayF: the resign button was added specifically for this expiration announcement. the foundation has only now existed long enough to have to follow its planned expiration processes19:44
JayFfungi: aha! That makes a lot of sense then. I just wanted to make sure the knowledge was sufficiently disseminated... I would assume most others are less-lax than I am about voting in foundation elections19:44
fungiclarkb: we could adjust the oslo.messaging unit tests to avoid tickling that bug, or stop testing it with python 3 while we wait, or delay the transition to 3.419:45
clarkbfungi: I am not sure how we would massage the tests. this is so subtle and appears to possibly depend on the garbage collectors unreachable list19:45
greghaynesI dont thinkg we know where in oslo.messageing that bug creeps in...19:45
clarkbgreghaynes: right19:45
greghaynesyoud prolly have to bisect to find it19:45
greghayneswhich might not be a bad idea19:45
clarkbgreghaynes: well except that that will only tell you what code introduced it but not how to fix it19:46
clarkbor at least to not tickle it19:46
greghaynessorry, I mean on tests19:46
greghaynesnot on commits19:46
jeblairnibalizer: we don't _do_ any ca services, so none of that actually matters to us.  let's just copy the ca and be done with it.19:46
nibalizerjeblair: okay19:47
nibalizercopy /var/lib/puppet/ssl/ca on one to the other19:47
nibalizerthen we'll need a patch to infra to unset ca = false19:47
greghaynesactually, its totally possible that a large number of the tests cause that bug19:47
*** dprince has quit IRC19:48
clarkbgreghaynes: yup19:48
fungigreghaynes: clarkb: which lends me to want to wait until ubuntu has a working python3.4 interpreter19:48
clarkbanyways /me figure out how to file a bug against ubuntu's py34 package19:48
openstackgerritSpencer Krum proposed a change to openstack-infra/config: puppetmaster is a CA
nibalizerjeblair: ^^19:48
jeblairnibalizer, clarkb, fungi, mordred: so any ideas about line 18?19:48
clarkbpleia2: any idea how I can do that?19:48
openstackgerritpatrick-crews proposed a change to stackforge/dox: Document multiple images in dox
greghaynesfungi: +119:49
JayFdhellmann: I'm assuming for (add rss feed, ironic-specs), you're using the same/similar configs other projects are using?19:49
clarkbjeblair: what happens if we update it by hand? will apt downgrade it for us against the pin?19:49
nibalizerjeblair: i dont think thats a problem19:49
clarkbjeblair: that maight be a simple workaround19:50
nibalizersince you can pass in puppet_version to openstack_project::server19:50
clarkboh or we can ^ do that19:50
jeblairhow do we pass that in, exactly?19:50
nibalizerwhich will give it to template, then to base, then to the tempalte19:50
jeblairwith 76 patches to infra/config?19:50
clarkbjeblair: ya :(19:50
dhellmannJayF: yes, I'm making the same changes in all of the projects19:50
JayFdhellmann: So dtantsur|afk raised a concern about putting the whole body in the RSS feed... Is it safe to assume you know it'll be huge and are OK with that?19:51
jeblairclarkb, nibalizer: we could temporarily stop managing that file, and perform the pin change manually on each host19:52
nibalizerjeblair: probably more actually19:52
dhellmannJayF: I think that's OK myself. But we can add an option to yasfb to prune it.19:52
clarkbjeblair: thats not a bad idea19:52
JayFdhellmann: I'm OK with it too :)19:52
*** sballe has joined #openstack-infra19:52
nibalizersince openstack_project::server is wrapped by a lot of classes, we'd have to make openstack_project::review take a puppet_version, soley to pass it on to ::server19:53
nibalizerwhich is terribad sounding19:53
JayFdhellmann: just one small thing then and you'll have my +2 ... I'm also willing to just fix that for you if you're tied up atm :)19:53
dhellmannJayF: I'll go through and fix that trailing slash for all of the projects, then19:54
openstackgerritAndreas Jaeger proposed a change to openstack-dev/specs-cookiecutter: Remove docutils pin
clarkbgreghaynes: fungi after reading the bug I think it is if objects refer to each other19:54
JayFdhellmann: sweet, thanks. I'll look for the update and when it hits you have my vote :)19:54
clarkbgreghaynes: fungi so clearly we should remove all referecnes and not garbage collect :)19:54
nibalizerclarkb: jeblair im okay with not managing the pin for a while19:55
nibalizeri expect we will hit a few nodes that need refactors in the puppet code, then it will be downhill19:55
jeblairclarkb, fungi, mordred, nibalizer: okay i think i have the full list of steps in the etherpad;  can you all double check that please?  steps that need changes in gerrit are in bold; we should get all of those except the last one merged before we start.19:56
*** unicell1 has quit IRC19:56
nibalizerand probably there will be a couple days of just upgrading everything then the default will be 3.6 and any node not on 3.6 will be looked at funny19:56
fungijeblair: are we pinning under 3 as well? if not, then i think that's a swell solution. we stop puppet-managing the apt preferences file for it and delete it to upgrade, then put it back and downgrade again if there's an issue19:56
openstackgerritKen Giusti proposed a change to openstack-infra/config: Add a test to verify oslo.messaging's AMQP 1.0 messaging protocol support
*** Ryan_Lane has joined #openstack-infra19:57
*** Ryan_Lane has quit IRC19:58
*** david-lyle has quit IRC19:58
*** _nadya_ has quit IRC19:59
jeblairfungi: i think currently we are set up to pin under 3, so i was expecting that step to end up being a standard 'sed' command we apply to the pin file20:00
*** weshay has quit IRC20:00
jeblairfungi, nibalizer: maybe we don't need to pin under 3?  worth considering?20:01
*** Sukhdev has joined #openstack-infra20:01
nibalizerjeblair: just sorta 3.lastest?20:01
*** Ryan_Lane has joined #openstack-infra20:01
nibalizer4 is in flight so i'd like to pin to 3. but I think I'm okay with pinning to 3. as opposed to 3.620:01
*** MarkAtwood has joined #openstack-infra20:01
nibalizerincidentally 3.7 came out like monday or something20:01
fungihaving not paid close enough attention, i assumed we were just going to take it as it came. i thought our original pin was there to keep us from prematurely upgrading to puppet 320:02
jeblairi'm guessing that 4 will break everything again?20:02
fungibut yeah, not insta-upgrading to puppet 4 might be good to insure20:02
jeblairso pinning to 3.* will at least keep that from happening, but we can continue to track the hopefully not completely breaking 3.x changes?20:02
wenlockjeblair, if i remember correctly there was a flag to test 4 lang with 3 client20:03
fungisounds good to me20:03
nibalizerwenlock: yea you can set parser = future20:03
nibalizeror future_parser = true or somtehing20:03
*** sdake has joined #openstack-infra20:03
*** sdake has quit IRC20:03
*** sdake has joined #openstack-infra20:03
*** markmcclain has quit IRC20:03
jeblaircan you set parser=stable_and_non_breaking? ;)20:03
nibalizerjeblair: lols hopefully after 2ish years of 3.x the 3.x parser is pretty stable20:04
*** sdake has quit IRC20:04
*** sdake has joined #openstack-infra20:05
jeblairi'm deleting the answered inline questions in the etherpad to make things more clear20:06
nibalizerwhat do you think of lines 18+19 ?20:06
nibalizerdoes it sound right to do that?20:06
nibalizerwe want to have some confidence that the master is functioning20:07
jeblairnibalizer: yeah, pretty sensible, but i think i want to move them, 1 sec20:07
openstackgerritDoug Hellmann proposed a change to openstack-infra/infra-specs: Use the current date for the copyright statement
openstackgerritDoug Hellmann proposed a change to openstack-infra/infra-specs: Remove templates from toctrees
openstackgerritDoug Hellmann proposed a change to openstack-infra/infra-specs: Add RSS feed
openstackgerritDoug Hellmann proposed a change to openstack-dev/specs-cookiecutter: Add RSS feed
jeblairnibalizer: moved them to lines 23/24, so that as the puppetmaster starts working, it's not trying to run ansible on all hosts20:08
clarkbdhellmann: fungi zul I am sure I did something wrong but meh20:08
uvirtbotLaunchpad bug 1367907 in python3.4 "Segfault in gc with cyclic trash" [Undecided,New]20:08
dhellmannclarkb: "tox -repy34"?20:09
clarkbdhellmann: yes20:09
clarkbdhellmann: oh wait I forgot something20:10
jeblairmordred: i've assigned some ansible related tasks to you in the etherpad, think you can do those real quick?20:10
dhellmannoh, heh, I read that as "-e repy34"20:10
clarkbdhellmann: you need to edit tox.ini to have a py34 target that is a copy of py33 target20:10
* clarkb updates bug20:10
*** dkehn has quit IRC20:10
dhellmannclarkb: let's add the py34 thing to oslo.messaging20:10
*** penguinRaider has quit IRC20:10
clarkbdhellmann: ok updated20:11
dhellmannclarkb: approved with a link to the bug20:11
*** david-lyle has joined #openstack-infra20:11
clarkband I added oslo.messaging to the lp bug20:11
*** dkehn has joined #openstack-infra20:12
dhellmannclarkb: I assigned that to you so I remember you're the one to talk to about it. :-)20:12
jeblairclarkb, fungi, nibalizer: i guess mordred is not around; anyone else interested in doing those tasks?20:12
dhellmannclarkb: if we need any changes, we can take it back if you like20:12
clarkbjeblair: looking now20:12
*** craigbr has joined #openstack-infra20:13
clarkbjeblair: I can certainly give it a shot20:13
jeblairmordred: nevermind ^20:13
*** annegentle has joined #openstack-infra20:13
fungijeblair: i'll volunteer for whatever is needed20:14
clarkbfungi: I can ping you if ansible makes my brain melt20:14
fungii've never touched it, but no time like the present to learn ;)20:14
jeblairyeah, i put it on the list as two things, though that might end up being one change20:14
jeblairit's really annoying that we're using both python and puppet 2.720:16
fungip words20:16
fungiclearly we should be using python and puppet 3 instead20:16
*** craigbr has quit IRC20:17
*** markmcclain has joined #openstack-infra20:18
nibalizerjeblair: im about to lunch20:18
nibalizerbut will start poking after lunch20:19
vishymordred: is dox part of your evil plan to eventually have containerized versions of all the services with no dependencies20:19
vishycutting the distros out of the loop altogether?20:19
lifelesssdague: it should be completely rectified20:19
sdaguelifeless: cool20:19
mordredvishy: that comes after I rewrite OpenStack in go20:19
fungicontainerized go would be like having a handlebar mustache on your handlebar mustache20:20
openstackgerritA change was merged to stackforge/dox: Document multiple images in dox
vishyunfortunately go dependencies are a complete mess20:20
vishybut at least they are only needed for compilation :)20:21
fungiyep, so far their answer has mostly been "statically compile, forget dynamic linking"20:21
vishybut there is not a good standard way to version packages20:21
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Stop managing puppet apt pins
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Revert "Stop managing puppet apt pins"
vishybasically you do go get and hope things haven’t broken20:22
mesteryHowdy infra folks! I think I've found an issue with openstack's neutron git. The commit found here ( is not in the openstack neutron git, but is in the github neutron repo.20:23
*** Ryan_Lane has quit IRC20:23
mesteryThis is for changeid Ie8e97a6d4fde0bda7cf049696e1384586492f676.20:23
mesteryAt first, I thought I was going nuts for not finding it after updating my local tree, but it's not upstream either:20:23
mesteryBut is on github. Can anyone help me understand this?20:23
*** krotscheck has quit IRC20:24
fungimestery: sounds like replication to one of the git backend servers in the farm may have broken for that repo. i'll give them a quick checkup20:25
jeblairfungi: oh thank goodness, there's a wide option to show-queue now: ssh review gerrit show-queue -w20:25
*** marcoemorais has quit IRC20:25
*** krotscheck has joined #openstack-infra20:25
jeblairfungi: run that command and check out task 6a1d12fa20:25
fungijeblair: oh! i missed the -w20:25
mesteryfungi: Thank you so much sir! Mostly I wanted to confirm I'm not going crazy, though this isn't the best litmus test I realize. ;)20:25
*** marcoemorais has joined #openstack-infra20:26
fungiha! it's been hung pushing neutron on git03 for a couple days, it seems20:26
jeblairfungi: i reckon we should kill the task?20:27
*** marcoemorais has quit IRC20:27
fungijeblair: yeah, was looking at the help to remind myself how20:27
jeblairfungi: afa453a8 after it should come along and clean things up20:27
jeblairfungi: you have the controls20:28
*** marcoemorais has joined #openstack-infra20:28
*** marun has joined #openstack-infra20:28
fungiahh, it's just 'kill 6a1d12fa' (or cancel)20:29
jogo"Attackers might be trying to steal your information from (for example, passwords, messages, or credit cards)."20:29
*** yamamoto has joined #openstack-infra20:29
jeblairjogo: that is an accurate description20:29
jogonew chrome warning20:30
jogojeblair: ^_^20:30
*** marcoemorais has quit IRC20:30
fungijogo: it's a nefarious plot. cloudbees is behind it. tinfoil around your wallet will help20:30
*** kgiusti has left #openstack-infra20:30
*** marcoemorais has joined #openstack-infra20:30
fungif589e249 20:30:18.475 20:29:18.475      (retry 1) push
*** Ryan_Lane has joined #openstack-infra20:31
clarkbhrm I haven't written puppet in a while :)20:31
fungilooks like it finished... any better now mestery?20:31
sdaguethat will pull the jenkins ssl certs and mark them as private trusted certs for the host in question20:31
mesteryfungi: Works like a champ now, thanks so much!20:31
fungimestery: thanks for bringing it to our attention. thankfully nothing else seemed to be hung20:32
mesteryfungi: Anytime, glad it was a simple kick to get things rolling again. :)20:32
jogosdague: thanks, although I am not sure if I should trust jenkins now20:32
*** yamamoto has quit IRC20:33
jeblairi don't20:34
*** amcrn has quit IRC20:35
*** yfried has joined #openstack-infra20:35
clarkbmordred: are you going to be around in just a moment? will have ansible question but want to push code so you can see20:37
*** craigbr has joined #openstack-infra20:37
sdaguejogo: well he does have a mustache, that makes him untrustworthy20:39
*** TravT has joined #openstack-infra20:40
*** dustins has quit IRC20:40
*** JayF has quit IRC20:40
*** baoli has quit IRC20:40
*** gondoi has quit IRC20:41
*** freyes has quit IRC20:41
*** baoli has joined #openstack-infra20:41
openstackgerritNathan Kinder proposed a change to openstack-infra/config: Add keystoneclient-kerberos repository
*** freyes has joined #openstack-infra20:41
*** krtaylor has quit IRC20:42
*** gondoi has joined #openstack-infra20:42
*** JayF has joined #openstack-infra20:42
fungisdague: are you saying jenkins is a hipster?20:42
jogosdague: haha, what does that mean for
sdaguejogo: oh, logstash was always a hipster20:43
*** weshay has joined #openstack-infra20:43
nibalizerjeblair: clarkb okay im on line20:44
mtreinishjogo: heh, shouldn't it be logstache then?20:44
fungii mustache you a question, but i'm shaving it for later20:44
*** lcheng_ has joined #openstack-infra20:45
*** Ryan_Lane has quit IRC20:46
openstackgerritClark Boylan proposed a change to openstack-infra/config: Make ansible speak to appropriate puppetmaster
openstackgerritSpencer Krum proposed a change to openstack-infra/infra-specs: HP owns everything. I own nothing.
clarkbmordred: ^ something like that you will see that the hostlists are mepty and I am not sure how to pass args to playbooks yet20:48
clarkband with that I am going to go find food20:48
sdaguefungi: where in the tree is the script that updates launchpad on gerrit changes?20:56
fungisdague: openstack-infra/jeepyb:jeepyb/cmd/update_{blueprint,bug}.py20:57
*** marun has quit IRC20:57
fungisdague: at least that's what gets called... if you want the shell script which serves as the gerrit hook to fire that on new patchsets, that's in the config repo but not particularly interesting20:58
*** pblaho has joined #openstack-infra21:00
jeblairclarkb: what's "#need to pass puppetmasterargument here." about?21:00
sdaguecool, I think I might need to care about the shell script as well, as we've got the issue where people don't always get the bug tag in the first commit21:00
clarkbjeblair I added a required arg that needs passing in not sure how yet21:01
fungisdague: openstack-infra/config:modules/openstack_project/templates/gerrit_patchset-created.erb21:02
sdaguefungi: cool, thanks21:02
ekarlso-anyone at rax know if phschwartz is awake ? ;)21:02
*** marcoemorais has quit IRC21:04
*** rfolco has quit IRC21:04
*** marcoemorais has joined #openstack-infra21:04
*** pcm__ has quit IRC21:05
*** marcoemorais has quit IRC21:05
*** marcoemorais has joined #openstack-infra21:06
*** jpeeler has quit IRC21:06
*** annegentle has quit IRC21:06
*** marcoemorais has quit IRC21:07
*** lcheng__ has joined #openstack-infra21:08
*** marcoemorais has joined #openstack-infra21:08
*** mjturek has quit IRC21:08
*** aysyd has quit IRC21:08
sdaguelaunchpad changed their typography now? seriously?21:08
*** marun has joined #openstack-infra21:08
openstackgerritSandy Walsh proposed a change to openstack-infra/config: StackTach.v3 Quincy REST API
*** jpeeler has joined #openstack-infra21:09
jeblairfungi: 120577 has an ansible puzzle if you're interested (see "need to pass puppetmasteragurement here")21:09
jeblairfungi, clarkb, mordred, nibalizer: i'm afk, back in an hour21:10
*** lcheng_ has quit IRC21:10
*** sabeen has joined #openstack-infra21:10
*** bswartz1 has quit IRC21:10
fungilooking, thanks!21:10
nibalizerjeblair: ok21:11
*** mfer has quit IRC21:11
*** tonytan4ever has quit IRC21:11
fungiclarkb: design question before i dive too deep... did you consider letting ansible run against all servers and check the contents of the local config to see if it matches the master invoking it?21:12
clarkbfungi: I did not21:12
clarkbbut that means updating all servers?21:12
*** miqui has quit IRC21:12
fungiclarkb: i mean the puppet config on each server which tells it what master it's reporting into21:12
fungiclarkb: since we have that information already, it just happens to be scattered across all the servers21:13
clarkbwe are not updatiung that according to etherpad21:13
* fungi rereads21:13
nibalizerfungi: that file is controlled by puppet21:13
*** andreykurilin has joined #openstack-infra21:14
jeblairyeah, there was a suggestion to remove that from puppet.conf; i think it makes sense though i'm kind of ambivalent about it...21:14
funginibalizer: the puppet.conf server line on each server?21:14
*** ramashri has joined #openstack-infra21:14
nibalizerfungi: yes21:14
fungihrm... yeah so if we're updating it that way instead of by hand, then i guess it is a catch-2221:15
nibalizerwe can manage it to be absent, which means when someone runs 'puppet agent --test' it will fail instead of just checking in to the p27 master21:15
nibalizermeans a manual run has to be explicit21:15
fungiokay, i'm cool with the design in 120577 in that case. reviewing more closely21:15
*** dkranz has quit IRC21:17
*** rainya has quit IRC21:18
*** krtaylor has joined #openstack-infra21:19
*** todd_dsm has joined #openstack-infra21:20
fungiclarkb: so...
fungiclarkb: the bit on parameterized roles21:20
fungiclarkb: that seems to be the syntax you're seeking/>21:20
fungii'm burrowing through now to see whether i can find a clear example of this pattern21:24
*** mestery has quit IRC21:25
* nibalizer total ansible newb21:26
*** mestery has joined #openstack-infra21:26
funginibalizer: this was the first time i looked at ansible configuration, so welcome to the club21:26
fungino, i take that back... michael dehaan visited rtp and presented to our local osug on managing openstack deployments with ansible21:27
fungiso i definitely saw some in his slides21:27
*** emagana has quit IRC21:27
*** rainya has joined #openstack-infra21:28
nibalizerwe're getting a talk at pdxdevops on ansible on monday21:29
*** yamamoto has joined #openstack-infra21:29
nibalizercouldn't come soon enough :)21:29
*** annegentle has joined #openstack-infra21:29
clarkbfungi: yup I think thats it. sorry was finishing lunch21:29
clarkbwill push a patch shortly21:30
*** Ryan_Lane has joined #openstack-infra21:30
*** armax_ has quit IRC21:30
sdakewhich ansible work is planned for infra21:32
*** ramashri has quit IRC21:32
clarkbsdake: we use ansible to trigger puppet. This allows us to do some orchestration of our puppet runs21:32
arosenany idea why nodepool image-delete isn't work for me here?
clarkbsdake: I am currently updating it to handle multiple puppet masters21:32
sdakecool thanks clarkb21:32
fungisdake: we've, so far, been using it for complex orchestration tasks where puppet's staticness gets in the way. heat would have been our preference i think, but having working lightweight standalone heat was a prerequisite for that21:33
clarkbarosen: make sure you that provider still exists?21:33
clarkbarosen: oh actually I think maybe because there is no image id yet? that may be a real bug21:33
sdakefungi standalone heat is possible, but with the changes planned around convergence, I think it is unlikely the standalone model will work well21:34
arosenclarkb:  i deleted the instance manually myself but it's still stuck. Should i drop the nodepool db and restart nodepool?21:34
*** yamamoto has quit IRC21:34
sdakejust a guess tho21:34
fungiarosen: since that image never got into a working state, you could just delete that row21:34
sdakei like ansible, it isn't super competitive to heat21:34
openstackgerritClark Boylan proposed a change to openstack-infra/config: Make ansible speak to appropriate puppetmaster
clarkbfungi: ^ something like that?21:34
sdakeheat orchestrates infrastructure really well which no other project does that I'm aware of :)21:34
sdakesoftware config isn't a super interesting problem to me :)21:35
*** annegentle has quit IRC21:35
sdakeits a must have, but there is so much competition hard to be a #1 solution there21:35
clarkbarosen: I think you might have to. looking at the traceback it appears it is complaining that the snap_image is None whihc means it isn't reading it out of the db properly21:36
arosenk, doing that seemed to fix it. thanks.21:36
*** mrmartin has quit IRC21:36
clarkbfungi: note I Have not tested that. I mostly wanted to hack something out first and make sure everything looked good21:37
fungiclarkb: i think so (you could also make that a tree instead of a one-liner association) but also you have failing ci jobs on it21:37
clarkbnot quite sure how I would test it actually. oh right I need lists of hosts and need to fix tests21:37
clarkbworking on making tests happy now21:37
nibalizerclarkb: so,  i think one of the things that was expressed was that we dont want to hit gerrit to switch a node21:37
clarkbfungi: maybe you can get a list of nodes out of puppet by running the puppet-inventory script21:38
nibalizerso maybe managing the hostlist files in puppet isn't the right idea21:38
clarkbnibalizer: hrm21:38
fungiclarkb: sure, you want it pasted into the etherpad?21:38
*** oanufriev has quit IRC21:38
clarkbI can remove puppet's management of that file21:38
clarkbfungi: sure21:38
clarkbnibalizer: ya I think I will just go ahead and remove management of those files from puppet21:40
clarkbthen we can edit /etc/ansible/hostfile by hand21:40
clarkbon the appropriate nodes21:40
*** oanufriev has joined #openstack-infra21:40
clarkbnibalizer: if I ensure => present but have no source is that equivalent to touch file if it doesn't exist but leave it alone otherwise?21:41
openstackgerritClark Boylan proposed a change to openstack-infra/config: Make ansible speak to appropriate puppetmaster
clarkbnibalizer: fungi ^ ok that is sort of an in between compromise21:42
clarkbbut gives reviewers enough context that we should be able to decide on one, update the change, then merge it21:42
nibalizeralso unix pop quiz, echo > /tmp/foo; chmod +x /tmp/foo; ./tmp/foo;  echo $?21:42
* nibalizer goes to review21:43
clarkbnibalizer: thats a shell thing right?21:43
fungiclarkb: it was 80 in total...
clarkbif all else fails run it as shell script21:43
clarkbfungi: thanks21:43
*** ramashri has joined #openstack-infra21:43
*** Sukhdev has quit IRC21:44
mordredclarkb: back online21:44
mordredclarkb: do you have all the ansibles? or do you need things?21:44
*** Sukhdev has joined #openstack-infra21:44
fungihe needs for to ansible with the ansible things21:44
fungiwe *hope* we figured out the right syntax21:45
fungimordred: is the current state of the patch21:45
fungier, no21:45
fungimordred: sorry,
*** sballe_ has joined #openstack-infra21:45
fungii picked the wrong \d*77\d* change21:46
clarkbmordred: please review 120577 I think that ansibles21:46
*** dane_leblanc has quit IRC21:48
*** sballe has quit IRC21:49
*** cipcosma has quit IRC21:50
*** alkari has joined #openstack-infra21:52
clarkbnew patchset pushed with more bug fixes21:52
*** gondoi is now known as zz_gondoi21:52
openstackgerritClark Boylan proposed a change to openstack-infra/config: Make ansible speak to appropriate puppetmaster
clarkbthere it is21:52
funginow with fixins21:52
*** amcrn has joined #openstack-infra21:53
fungiclarkb: lgtm other than hostfiles21:55
*** HeOS has quit IRC21:55
*** masayukig has quit IRC21:55
mordredclarkb: I think the remote_puppet.yaml could be done without splitting it into different files ... but I don't know how much I care21:55
*** HeOS has joined #openstack-infra21:55
clarkbmordred: how would we do that?21:56
clarkbwe need to pass arguments21:56
mordredclarkb: make a global variables file that gets .erb'd down that has $::fqdn in it, and then reference that variable in the one remote_puppet.yaml file21:57
clarkbfungi: ya so we need to decide if we want them and I can seed with the list you generated or if we should do them by hand21:57
jogorandom question:
jogoanyone understand that21:58
clarkbmordred: is that way preferable to you?21:58
mordredclarkb: but what you have will work - so meh21:58
clarkbmordred: ok I will leave it as is. The other thing to have an opinion on is whether or not we want to manage the host files with puppet21:58
clarkbmordred: I think jeblair doesn't want to do that so that we don't have 80 puppet changes to move nodes21:59
mordredclarkb: ... if you put a defaults/main.yaml in the puppet role and manage the content of that with puppet21:59
nibalizeri think we shoul dget jeblairs sense on that21:59
mordredclarkb: I think I agree with jeblair21:59
mordredclarkb: there's a bunch of manual here- I think we want to just manage those lists by hand21:59
clarkbok so I should probably just have puppet create the empty files for us and leave it at that21:59
clarkbjogo: django is trying to compile its css files via pyscss and failing because the output dir exists already22:00
clarkbthat is a silly failure mode to have but meh22:00
jogoclarkb: very22:01
*** masayukig has joined #openstack-infra22:01
* fungi is warning python projects not to be silly again22:01
mordredclarkb: that's the stupidest thing ever22:01
* mordred steps away again for a sec22:01
funginow, nobody loves a good joke as much as i do22:01
fungiexcept perhaps my wife and some of her friends22:02
fungioh, yes, and captain johnston22:02
openstackgerritClark Boylan proposed a change to openstack-infra/config: Make ansible speak to appropriate puppetmaster
fungicome to think of it, most people like a good laugh more than i do, but that's beside the point22:03
clarkbso there. once this goes in we will be puppetting no nodes and should edit /etc/ansible/hostfile to contain the current list22:03
*** ZZelle_ has quit IRC22:04
*** krtaylor has quit IRC22:04
clarkbmordred: do you hae a good way to test these types of changes?22:06
clarkbthat was going to be my next question for you22:06
*** armax has joined #openstack-infra22:07
*** craigbr has quit IRC22:08
*** jgrimm is now known as zz_jgrimm22:09
*** bhuvan has quit IRC22:10
*** bhuvan has joined #openstack-infra22:11
*** sdake has quit IRC22:11
*** dims__ has joined #openstack-infra22:13
arosenHrm, i'm getting this error when nodepool is creating an image for me. Does it assume we  have gcc installed on the host already?
arosenerr i mean libxml?22:13
arosenand xslt-config22:14
fungiarosen: hrm... our zuul module installs the python-lxml package to get around needing to compile it22:14
clarkbarosen: it shouldn't be. are you using vanilla nodepool scripts?22:15
clarkbfungi: oh is this the keystone thing?22:15
fungiclarkb: i think this is a regression we introduced today22:15
clarkbfungi: oh? I must've missed it while debugging python22:15
arosenclarkb:          setup: <-- i use this22:15
arosenand all the defaults.22:15
fungiclarkb: we just started installing zuul on all workers to make the merger tool available to jobs22:15
fungiclarkb: i bet the change to do that didn't consider its prereqs22:16
clarkbohhhhh yup and since zuul depends on swift/keystone clients...22:16
clarkbwe really need to make lxml not required22:16
fungiexcept those are dependencies of the zuul server and not of the zuul merger... so installing zuul just to get the merger ends up dragging in some seriously heavyweight dependencies22:16
*** dims_ has quit IRC22:17
arosenso this is a bug?22:17
clarkbarosen: yes I think so22:18
fungiarosen: yeah, i think it's a freshly created bug today22:18
jeblairi did test it; image builds worked at the time.22:18
arosenjeblair:  does it start from a fresh image or the last one it built?22:18
*** dims__ has quit IRC22:18
arosenmaybe this stuff was already installed/22:18
*** zehicle_at_dell has quit IRC22:19
*** lcheng__ has quit IRC22:19
*** dims_ has joined #openstack-infra22:19
openstackgerritJoe Gordon proposed a change to openstack-infra/elastic-recheck: Add query for bug 1345955
uvirtbotLaunchpad bug 1345955 in horizon "KeyError: <SassRule <RuleAncestry ()>, 0 props> Failure" [Critical,Confirmed]
*** david-lyle has quit IRC22:19
fungiarosen: yeah, looks like modules/openstack_project/manifests/thick_slave.pp installs python-lxml22:19
*** marcoemorais has quit IRC22:20
fungiarosen: jeblair: so it probably works on our thick slaves22:20
*** andreykurilin has quit IRC22:20
*** marcoemorais has joined #openstack-infra22:20
*** dmsimard is now known as dmsimard_away22:20
jeblairwhat's a thick slave again? :)22:20
fungium... bare-trusty et cetera22:20
clarkbjeblair: slave with all the things installed on it for unittesting22:20
jeblairthick/thin are meaningless to me :(22:20
*** marcoemorais has quit IRC22:20
* asselin is paying attention b/c he's about to setup a new 3rd party ci system....22:20
*** todd_dsm has quit IRC22:20
jeblairclarkb, fungi: but devstack slaves are "thin"?22:21
fungiarosen: what nodepool prep script is your entrypoint?22:21
clarkbjeblair: yes basically22:21
jeblairfwiw, if i had to guess, i would have guessed the opposite.22:21
*** marcoemorais has joined #openstack-infra22:21
jeblair(bare AND thick)?22:21
*** lttrl has quit IRC22:21
*** wenlock has quit IRC22:21
arosen./ "$HOSTNAME"22:21
clarkbjeblair: yes its an artifact of how we screwed up our puppet manifests22:21
fungijeblair: well, we preinstall lots more system packages on the "thick" slaves, and let devstack do that itself on the "thin" slaves?22:22
clarkbjeblair: which is why we introduced the new terminology to make it more explicit22:22
*** garyh has quit IRC22:22
*** arnaud has joined #openstack-infra22:22
clarkbjeblair: the next step is to stop using 'bare' completely22:22
*** lcheng_ has joined #openstack-infra22:22
clarkbbecause our bare slaves were never bare22:22
*** arnaud__ has joined #openstack-infra22:22
fungiour bare slaves are far, far, far from bare22:22
jeblairso yeah, this broke our devstack image builds too22:22
arosenfungi:  i believe it's prepare_node.sh22:22
*** flaper87 is now known as flaper87|afk22:22
fungiarosen: is that the one you list in your nodepool.yaml?22:23
arosen        # This is the script that will be used to prepare the image22:23
arosen        setup: prepare_node_devstack.sh22:23
jeblairso we should probably remove python-swiftclient from zuul reqs22:23
jeblairclarkb, fungi, jhesketh: we can move those to test-reqs and should be okay, right?  then add an explicit install for them in the zuul puppet module?22:24
clarkbjeblair: yeah that is the most immediate simple thing. then document it as an optional thing you can isntall22:24
clarkbjeblair: yup that will work22:24
*** dizquierdo has joined #openstack-infra22:24
fungijeblair: i think so, yeah22:24
jeblairi think it's already set up like that in the code (eg, imports are conditional)22:25
clarkbmorgan: jamielennox: keystoneclient really should stop depending on it alltogether22:25
*** lcheng__ has joined #openstack-infra22:25
arosenfungi: clarkb mind pinging me when there is a patch and I can update my setup and test it out for you guys?22:25
jeblairat hashar's request (he uses debs and ignores reqs.txt)22:25
fungijeblair: were you writing the patch or shall i?22:25
*** flaviof is now known as flaviof_zzz22:25
jeblairfungi: if you could that would be swell, i'm writing email :(22:26
*** todd_dsm has joined #openstack-infra22:26
morganclarkb, hrm?22:26
morganclarkb, lxml again?22:27
jeblairmorgan: rage lxml rage.  ;)22:27
clarkbmorgan: keystoneclient depends on python-lxml. This means you can't pip install keystoneclient anywhere22:27
*** morgan is now known as morganfainberg22:27
clarkbit really is detrimental to ease of use22:27
jeblairwe're about to pull it from zuul's deps basically just because of that :(22:27
morganfainbergclarkb, jeblair, we just started the work to split out the federation and kerberos stuff into separate repos22:27
morganfainbergso lxml will become optional22:27
clarkbwell we could just use the built in xml parser22:27
clarkbI think that is what bugs me most22:27
morganfainbergyeah the non-standard auth plugins for client should be used22:28
clarkbwe don't need to drop any features. we just have to be smart22:28
jeblairmorganfainberg: and nice -- ball is in our court to get around to reviewing the new repo change :)22:28
morganfainbergclarkb, i think this is a pysaml2 thing22:28
*** lcheng_ has quit IRC22:28
morganfainbergjeblair, i'll make sure we get the federated auth plugin moved as well22:28
openstackgerritJeremy Stanley proposed a change to openstack-infra/zuul: Make keystone and swift clients optional test reqs
clarkbmorganfainberg: iirc its a direct import and etree grep in keystoneclient22:28
fungiarosen: jeblair: ^22:28
clarkbmorganfainberg: we can write a grep function for the python xml.etree22:28
*** emagana has joined #openstack-infra22:29
morganfainbergclarkb, ok if we can't get the federation stuff moved we'll do that instead22:29
*** yamamoto has joined #openstack-infra22:29
*** jpich has joined #openstack-infra22:29
morganfainbergclarkb, in either case i'll make sure we evict lxml for the next release if at all possible22:29
*** Sincler has quit IRC22:29
arosenfungi:  okay so i've cd /opt/zuul; cherry-picked that commit; sudo python develop;22:30
arosennow i'll restart all the zuul services and nodepool image-delete everything that's building?22:30
*** emagana has quit IRC22:30
fungiarosen: that may not be sufficient since it's unlikely that your prep script is cloning from your local copies of projects22:30
jeblairarosen: i don't think that will help, as the script clones zuul; that change needs to land and then new image builds have to start22:31
*** emagana has joined #openstack-infra22:31
jeblairfungi, clarkb: we okay with the ordering there?  we're reversing the keystoneclient/swiftclient order.22:31
clarkbjeblair: it should be ok in this case I think22:32
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for bug 1345955
uvirtbotLaunchpad bug 1345955 in horizon "KeyError: <SassRule <RuleAncestry ()>, 0 props> Failure" [Critical,Confirmed]
arosenfungi:  TypeError: resolve() got an unexpected keyword argument 'replace_conflicting'22:32
fungijeblair: i figure if it's not, then hopefully zuul's tests will complain?22:32
arosenmaybe that's unrelated to your patch ?22:32
jeblairfungi: natch.  aprvd.22:32
*** rainya has quit IRC22:33
*** rainya has joined #openstack-infra22:33
fungiarosen: try 'pip install -e .' instead?22:33
*** bhuvan_ has joined #openstack-infra22:34
boris-42fungi clarkb  hi guys22:34
*** yamamoto has quit IRC22:34
boris-42fungi clarkb  one question probably I miss something but22:34
arosenfungi:  yup i had tried that too:
boris-42fungi clarkb
fungiarosen: i'm at this point a little fuzzy on the extent to which pbr breaks develop22:34
*** dims_ has quit IRC22:34
boris-42fungi clarkb
*** dims_ has joined #openstack-infra22:35
boris-42fungi clarkb  so oslo.config seems like be missing on pypi… or I missed something22:35
jeblairclarkb, nibalizer: what was the decision on removing server=?22:35
fungiboris-42: why do you think it's missing on pypi?22:36
*** arnaud__ has quit IRC22:36
clarkbjeblair: I think we can worry about that in a followup/cleanup change. I like the idea, but it isn't necessary yet22:36
*** bhuvan has quit IRC22:36
fungiboris-42: oslo.config- is right there in your list22:36
jeblairokay.  i'm going to start approving changes then.22:36
jeblairclarkb: can you aprv 120572 ?22:36
nibalizeri think thick_slave and thin_slave are puppet classes22:36
*** zz_dimtruck is now known as dimtruck22:36
nibalizeroops i was scrolled way back up22:37
funginibalizer: they are, yes22:37
*** marun is now known as marun_afk22:37
nibalizerjeblair: we can remove server = or not, up to you22:37
nibalizerprobably actually doesn't matter22:37
jeblairnibalizer: yeah, i'm just saying a can never remember how they map to our actual nodes.  since the "really big one" (devstack) is thin.  and the really small one "bare" is thick.22:37
clarkbjeblair: yup reviewing now22:37
*** rhsu has quit IRC22:38
jeblairclarkb: also 12056522:38
arosenfungi:  hrm i seem to be hitting this now :/
clarkband removing the pin from puppet is ok because the file will stay in place and we can modify it22:38
boris-42fungi ah sorry about that… some guys have issues22:38
* clarkb approves22:38
boris-42fungi with installing rally22:38
nibalizerclarkb: also remember it gets initalized to 2.7 on new nodes22:39
*** dims_ has quit IRC22:39
clarkbnibalizer: yup22:39
fungiboris-42: to install prereleases of openstack libraries you need to have python wheels support in pip, which means using pip 1.5.x or passing --use-wheel to 1.4.x22:39
boris-42fungi ok thanks22:40
*** datsun180b has quit IRC22:40
*** emagana has quit IRC22:41
BobBall_AWOLCan someone please re-eanble the xenserver CI account? citrix_xenserver_ci - it's running jobs fine but because it was disabled it can't post anything22:41
fungiboris-42: also be aware that if you're installing in a virtualenv, the virtualenv tool provides its own version of pip, so you need to be using a new enough version of virtualenv to have the version of pip you want22:41
*** lcheng__ has quit IRC22:41
clarkbjeblair: both of those changes have been approved22:41
*** emagana has joined #openstack-infra22:41
fungijhesketh: the ship has mostly sailed, but are you okay with ?22:42
*** daya_k has joined #openstack-infra22:42
fungiBobBall_AWOL: i take it you read the e-mail explaining why it was disabled and have corrected any deficiencies now?22:42
BobBall_AWOLit was disabled because all jobs were failing22:43
boris-42fungi ahh22:43
fungiBobBall_AWOL: cool, i'll set it back to active22:43
*** unicell has joined #openstack-infra22:43
boris-42fungi ya that makes sense thanks22:43
BobBall_AWOLperfect, thanks fungi22:43
fungiBobBall_AWOL: it should be active in gerrit again now22:44
jheskethfungi, jeblair: Yep, removing swiftclient looks good to me... I'll shoot through a patch to update the docs about installing i22:44
fungijhesketh: oh, awesome--thanks!22:45
fungijhesketh: for context, basically this came up because now we'd like to install zuul for the purpose of running its client utilities on some systems but not any of its servers22:45
openstackgerritJamie Lennox proposed a change to openstack-infra/config: Add keystoneclient-kerberos repository
*** emagana_ has joined #openstack-infra22:46
*** emagana has quit IRC22:46
jheskethfungi: Yep, makes sense... I assume you mean the zuul cloner?22:46
*** doug-fish has quit IRC22:46
clarkbuhm is the nodepool graph on status.o.o/zuul correct? it seems to indicate we have no nodes22:46
clarkbis this related to zuul?22:46
fungijhesketh: yeah22:46
fungiclarkb: i would agree with that assessment... zuul seems to be queuing a lot, so it may not be a nodepool issue22:47
fungior it may be nodepool22:47
funginodepool's still running, hasn't rebooted22:48
clarkband according to nodepool list we have ~140 nodes ready22:49
fungiclarkb: my guess is that zuul's choked up on processing merge-check?22:49
clarkbooohhh yup that could be it22:49
fungilots and lots of nova there, oldest ~12 minutes22:50
fungibut may have been at it for longer22:50
*** emagana_ has quit IRC22:50
*** jerryz has joined #openstack-infra22:51
jheskethfungi: btw, if you feel like reviewing zuul, I've got a heap of patches open,n,z22:51
*** bknudson has quit IRC22:53
*** marun_afk has quit IRC22:53
fungithey don't appear to have rebooted any of our other merger servers on us since zm0222:53
jeblairand zm02 is running22:53
clarkbso just nova having all the open changes22:54
clarkband being the biggest slowest repo22:54
jeblairalso, it takes ~30secs to test a nova merge22:54
fungimerger:merge    131     4       422:54
fungiaccording to 'echo status|nc -q 3 localhost 4730'22:54
*** alkari has quit IRC22:54
funginow 12022:54
*** bhuvan_ has quit IRC22:55
fungiso it's chewing through them surprisingly quickly22:55
*** virmitio has quit IRC22:56
jeblairi think we may not be applying the gearman precedence to the merge jobs; if we did, we could probably keep merge-check from interfering with the other pipelines22:56
openstackgerritA change was merged to openstack-infra/zuul: Make keystone and swift clients optional test reqs
clarkbjeblair: oh good idea. make the normal operation ones higher priority than the merge check pipeline jobs22:57
*** emagana has joined #openstack-infra22:57
jeblairif we didn't mind experimental/post fighting for mergers with merge-check, we can just use the same priority we use for the regular jobs.  i think i'm okay with that.22:58
jeblairwould mean check/gate have priority22:58
fungiclarkb: flaper87|afk: Alex_Gaynor: of all the voting pypy jobs which work on precise, the only one failing on trusty is zaqar's...
fungi(not a lot to go on there either)23:00
openstackgerritA change was merged to openstack/requirements: restrict requests version with !=2.4.0
*** _nadya_ has joined #openstack-infra23:00
fungiarosen: the fix merged about 5 minutes ago, so you should be safe to retry your image updates whenever23:00
clarkbfungi: ya not sure what I would do next. hopefully Alex_Gaynor knows whats up23:00
arosenfungi:  k23:01
jeblairlooks like a dfw->dfw network error :(23:01
*** emagana has quit IRC23:02
*** dangers is now known as dangers_away23:02
openstackgerritArmando Migliaccio proposed a change to openstack-infra/config: Move check-tempest-dsvm-neutron-dvr to check queue
jeblairclarkb, fungi: and the np graph is spiking; i think things are normal now.23:03
*** hdd has quit IRC23:04
*** emagana has joined #openstack-infra23:04
*** pblaho has quit IRC23:04
*** todd_dsm has quit IRC23:04
*** _nadya_ has quit IRC23:04
*** armax has quit IRC23:04
jeblairclarkb, nibalizer:
*** todd_dsm has joined #openstack-infra23:05
jeblairoh i see.  1 sec.23:06
*** todd_dsm has quit IRC23:06
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Make ansible speak to appropriate puppetmaster
openstackgerritA change was merged to openstack-infra/config: Stop managing puppet apt pins
jeblairclarkb, nibalizer, fungi:  had a typo23:06
jeblairyay apply test23:06
clarkbwhoops I failed thank you for fixing23:07
openstackgerritA change was merged to openstack-infra/config: puppetmaster is a CA
*** eharney has quit IRC23:07
*** vigneshvar has quit IRC23:09
nibalizeryay apply test23:09
*** melwitt has quit IRC23:09
*** dizquierdo has quit IRC23:10
*** dprince has joined #openstack-infra23:10
*** pballand has quit IRC23:12
fungiAlex_Gaynor: is it possible the current pypy package in ubuntu 14.04 lts suffers from ?23:13
* fungi is starting to get the impression that you should give any lts release 6 months to stabilize before switching from the previous one ;)23:14
*** todd_dsm has joined #openstack-infra23:14
jeblairfungi: they actually don't recommend desktop users switch until the .1 release (the updater doesn't even suggest it until then)23:15
jeblairof course, that just came out :)23:16
*** MarkAtwood has quit IRC23:16
*** melwitt has joined #openstack-infra23:16
fungimaybe they should start recommending server users wait until the .2 point release23:16
mordredor we should just switch to gentoo like clarkb wants to do23:16
*** melwitt has quit IRC23:16
*** melwitt has joined #openstack-infra23:17
fungii thought clarkb was now fronting for lfs?23:17
*** melwitt has quit IRC23:17
clarkbno no arch23:17
mordredfungi: so what, I start to warm up to the idea of gentoo and he just goes crazier?23:17
clarkbor that gentoo hipster thing sabayon23:17
clarkbmordred: :)23:17
*** dprince has quit IRC23:17
*** melwitt has joined #openstack-infra23:17
*** armax has joined #openstack-infra23:18
clarkbmordred: fwiw gentoo would probably fix this python issue for us23:18
clarkband arch would out of ports at the very least23:18
*** melwitt has quit IRC23:18
clarkbthe SRU process is a bit confusing23:18
mordredclarkb: arch has dropped sysvinit23:18
jeblairit's pretty clear that the ideal system runs apps in docker on gentoo out of afs23:18
mordredclarkb: gentoo remains the only linux distro for people who like linux23:18
mordredjeblair: ++23:18
fungijeblair: or freebsd. it has perfectly fine afs support23:19
clarkbI just tested unicorn's python on trusty and it fixes the problem23:19
mordredfungi: but does freebsd have docker?23:19
fungimordred: freebsd has jails. who needs docker?23:19
JayFmordred: I run gentoo with systemd :)23:19
clarkbso now I have to update my bug and do all kinds of formatting to make it happy for SRUing23:19
pleia2clarkb: did you figure out how to file the bug against ubuntu's py34 package?23:19
*** emagana has quit IRC23:19
mordredJayF: why in god's name would you do something like that?23:19
pleia2aha, seems like23:19
mordredfungi: jails == lxc ... not docker23:19
JayFmordred: I'm one of those people who like systemd, very much :)23:19
*** melwitt has joined #openstack-infra23:19
mordredJayF: please get out of our channel then23:19
mordredJayF: :)23:20
jeblairso, funny upstart story from when i upgraded to trusty:23:20
clarkbpleia2: I filed currently working on making it SRU format compatible23:20
uvirtbotLaunchpad bug 1367907 in python3.4 "Segfault in gc with cyclic trash" [Undecided,New]23:20
*** emagana has joined #openstack-infra23:20
*** todd_dsm has quit IRC23:20
pleia2clarkb: ok, great23:20
fungimordred: freebsd has shell scripts--who needs docker?23:20
JayFmordred: you're too late, I've already infected systemd into openstack (IPA CoreOS image runs in a container under systemd-nspawn) :P23:20
mordredJayF: that's fine - I don't mind that systemd exists - I just don't want it to be running on a production system I actually have to care about23:21
mordredbecause I want my system to work23:21
fungiclarkb: i'll crib off your sru compatibility if it turns out i need to ask for newer pypy23:21
mordredand I've used pulseaudio before23:21
clarkbfungi: I have no idea what I am doing so your results may vary :)23:21
fungiclarkb: prolonged use may result in limb loss?23:22
*** yfried has quit IRC23:22
clarkband gum disease23:22
clarkbmordred: want to hear my funny pulseaudio story?23:22
mordredclarkb: yes please23:23
* fungi is still waiting for jeblair's funny upgrade story23:23
*** nelsnels_ has joined #openstack-infra23:23
fungithere can be only one23:23
clarkbmordred: so I upgraded from precise to trusty on my desktop and sound stops working. I login as guest user and its fine. So now I am grumpy. I fiddle with all the alsamixer knobs, try all soundcard and built in device nothing works23:23
*** rainya has quit IRC23:23
clarkbmordred: finally I install pavucontrol and clear as mud pulseaudio has decided to use my nvidia hdmi output which it marked as "unplugged" for its sound output23:24
*** emagana has quit IRC23:24
clarkbmordred: I almost rage quit computers right then and there23:24
jeblairfungi: oh, were you having trouble reading my story?  did you only see a blank screen after my lead-in?  i wonder where the output could have gone...23:24
clarkbmordred: it knows there is nothing attached, it knows that ther are things attached to this other port over here23:25
mordredjeblair wins23:25
clarkbmordred: it chose the unplugged decice23:25
mordredclarkb: AWESOME23:25
jeblairfungi: because that's what i got when i upgraded to trusty.  blank screens and no way to convince upstart to tell me why.23:25
clarkbanyways back to SRU stuff23:25
mordredclarkb: well, but but but - it has hipsters for extra awesome!23:25
*** nelsnelson has quit IRC23:25
clarkbfwiw when I ran arch I didn't pulse23:25
mordredclarkb: maybe using the working sound card is too obvious for pulseaudio23:25
clarkbI used alsa's poor mixing capabilities and lived with it23:26
* jeblair uses a mixer23:26
mordredI've mentioned I'm running gnome2 right?23:26
fungijeblair: clearly if it were systemd-tng and simply embedded directly into the kernel along with your window manager and dog-flosser, that would have been a non-issue23:26
jeblairit has excellent mixing capabilities23:26
*** melwitt has quit IRC23:26
nibalizeri use stock unity and blehhhhh, but sound does generally work23:26
jeblairfungi: i'm assuming i could have at least diagnosed it with led pulse codes23:26
mordredall I really want in life is ubuntu maverick23:27
mordred+ AFS23:27
*** emagana has joined #openstack-infra23:27
fungiclarkb: i still use alsa directly with an alsaconf which does surround reencoding from stereo and a variety of other useful things. works great23:27
*** bswartz has joined #openstack-infra23:28
fungieven adds a slight wavering rtl/ftb panning effect to make the surround a little more obvious23:28
*** yamamoto has joined #openstack-infra23:29
mordredoh. while we're being these people - I've mentioned I actually like X11 right?23:29
*** emagana has quit IRC23:30
fungimordred: why not. x11 is terribly bloated and inherits tons of design problems from its predecessors, but it also added a bunch of useful features no other display subsystem has yet to match23:31
*** yamamoto has quit IRC23:31
*** emagana has joined #openstack-infra23:31
*** yamamoto has joined #openstack-infra23:31
mordredfungi xterm & FTW23:32
mordredfungi: that's even better with AFS23:32
*** hdd has joined #openstack-infra23:33
fungiand kerberized remote x11 so you don't need tunneling/forwarding23:33
morganfainbergclarkb, jeblair, so next release of ksc will not have lxml in requires just test-requires23:33
morganfainbergclarkb, jeblair, that might be sufficient23:33
mordredfungi: I swear, one of these days my laptop will be as awesome as Athena was in 9323:34
clarkbfungi: you want tunneling for encryption right?23:34
fungimorganfainberg: it won't stop us from needing its system prereqs installed where we run keystoneclient's tests, but should solve the general "foo uses keystoneclient, welcome lxml trouble" case23:34
clarkbfungi: kerberos will only deal with the auth aspect?23:34
morganfainbergfungi, yes.23:34
clarkbmordred: if it makes you feel better your laptop's hardware is worse today too23:35
mordredclarkb: well, my current laptop is DEFINITELY worse23:35
clarkbmordred: you can have poor battery life, bad keyboard, bad touchpad/pointer, amazing display, and not much else23:35
fungiclarkb: back then we didn't care about network eavesdroppers23:35
*** emagana has quit IRC23:35
morganfainbergfungi, yeah we need better xpath support than built-in module has so its a bit more writing of code (the other xpath impls don't support namespace in the same way, lxml is sadly the *best* of the options)23:35
morganfainbergfungi, for now23:35
mordredclarkb: well, thank god it has an amazing display - that means that this-months-X-replacement can show me high def movies of cats while I try to figure out why systemd has decided to just start and stop mongodb in a loop23:36
*** yamamoto has quit IRC23:36
fungiclarkb: my new netbook is cobbled together out of a radxa rock. i'll hopefully have it fit to travel come paristime23:36
mordredbecause heaven forbid anybody actually do work that makes my laptop better for any of the things I use it for23:36
fungiclarkb: just got debian/testing working on it earlier this week, so far so good. need to decide what lvds panel to get for it now23:38
morganfainbergmordred, i like sysv init. :( can't we just go back in time...23:38
mordredmorganfainberg: I keep threatening to find a snapshot of ubuntu maverick, resurrect it and release it as a new distro23:38
mordredmorganfainberg: which has as it's mission "Remain Linux"23:38
mordredin fact23:38
morganfainbergmordred, hehe.23:38
mordredmaybe that's its name23:39
mordredor "Actual Linux"23:39
morganfainbergmordred, i kinda like ReMain Linux (RML)23:39
mordredor "I actually like Linux and don't want it to be like Windows or OSX"23:39
morganfainbergmordred, hahah23:39
mordredmaybe we can make the default window manager olvwm23:39
mordredwith no menu bars except via cntrl middle-mouse button on the root window23:40
morganfainbergmordred, but... what about E17?  or uh whatever that number is when it was last good23:40
morganfainbergmordred, :P23:40
mordredmorganfainberg: heh23:40
*** pabelanger has quit IRC23:40
*** roaet_ has quit IRC23:40
morganfainbergmordred, lets just use CDE!23:40
*** todd_dsm has joined #openstack-infra23:40
morganfainbergoh wait.. no lets not23:40
nibalizerfluxbox is an option... ill just let myself out23:41
lifelessmordred: the LFS folk got there first23:41
morganfainbergbleh, i can't break my enjoyment of apple devices.23:42
morganfainbergi keep trying23:42
morganfainbergbut ... alas23:42
*** roaet_ has joined #openstack-infra23:43
clarkblifeless: pleia2: how does look?23:43
uvirtbotLaunchpad bug 1367907 in python3.4 "Segfault in gc with cyclic trash" [Undecided,New]23:43
*** pabelanger has joined #openstack-infra23:43
*** Sukhdev has quit IRC23:44
*** emagana has joined #openstack-infra23:44
clarkbfungi: wait you macgyvered a netbook?23:46
fungiclarkb: am macguyvering currently23:47
clarkbfungi: wow23:47
openstackgerritA change was merged to openstack-infra/config: Make ansible speak to appropriate puppetmaster
fungieleballing a 5.6" tft off which does raw lvds (the rock pro has an lvds header)23:47
*** alkari has joined #openstack-infra23:47
clarkbfungi: I have an old acer 11.6" single core core2 laptop thingy you can have if you need a shell23:47
clarkbfungi: it was basically a netbook with real guts in it23:48
fungialready have the battery (10kah anker) and mini keyboard nailed down23:48
mordredmorganfainberg: I wish I liked apple devices ...23:48
*** hdd has quit IRC23:48
mordredmorganfainberg: since everyone seems hell-bent on copying them poorly, I might like life better if I thought anything about them was nice23:48
morganfainbergmordred, hehe.23:48
*** harlowja has quit IRC23:48
*** harlowja_ has joined #openstack-infra23:48
fungiclarkb: i think minimal shell will look cooler anyway ;)23:48
mordredas it is, I'm just madder because I Don't even like the perfect version of the thing that people are trying to emulate23:48
clarkbfungi: indeed. Make sure you have duct tape and chewing gum somewhere on it too23:49
*** atiwari has quit IRC23:49
clarkbmordred: its so bad :(23:49
morganfainbergmordred, fair point.23:49
clarkbmordred: I think google was going somewhere with the pixle23:49
clarkb5:4 hi res display because that is what you want23:49
*** emagana has quit IRC23:49
clarkbbut then priced it at insane prices for what is basically a gimpy google machine23:49
*** emagana has joined #openstack-infra23:50
morganfainbergthough tbh one of my best linux experiences was linux installed on a Macbook pro23:50
clarkbif I could just have a 12-13" device with 9 hours of battery life, non glossy hi res display, and a keyboard and pointer that are good23:50
clarkbits not asking for much23:51
*** rainya has joined #openstack-infra23:51
morganfainbergclarkb, the keyboard pointers are a deal breaker for me for the most part. I *really really* dislike the eraser things.23:51
clarkbdo not care for chicklet keyboards or these weird touchpads everyone does now23:51
morganfainbergclarkb, and the glossy screens *still* make me sad23:51
clarkbmorganfainberg: how can you dislike them? they give you carpal tunnel but they are the best things efer23:51
mordredmorganfainberg: yeah, that's the thing, I HATE HATE HATE trackpads23:52
morganfainbergclarkb, i keep bumping them and causing all sorts of wierd stuff, but on linux i also like sloppy focus23:52
mordredmorganfainberg: I only like the nipple23:52
mordredmorganfainberg: and real mouse buttons23:52
clarkbmordred: you have seen the latest trackpad thing on thinkpads right?23:52
morganfainbergclarkb, so accidently bumping the mouse changes my window focus23:52
mordredbut everyone wants to have huge trackpads and fake buttons that are regions of the trackpad23:52
clarkbmorganfainberg: so I have that problem with these bad trackpads. you graze them and boom mouse over there23:52
mordredclarkb: like, recent recent?23:52
*** hemna is now known as hemna_23:53
mordredclarkb: ++23:53
clarkbmordred: yeah in the x240 and x1 carbon23:53
clarkbmordred: there are no more physical buttons23:53
mordredclarkb: yeah - I just twitter-bombed lenovo about that yesterday23:53
clarkbmordred: instead the whole trackpad is a button that depresses23:53
morganfainbergmordred, apple has done trackpads better, but if you don't like trackpads like that, it's not a good option23:53
mordredI was not polite23:53
mordredmorganfainberg: yah23:53
clarkbmordred: with sensor zones23:53
morganfainbergi hate "sensor zones" and "fake buttons"23:53
mordredthat's the problem. I like a compeltely different approach to life, and apparently noone is manufacturing that anymore23:53
morganfainbergif you're going to go there, make the whole bloody thing a button and do multi touch... it really makers it more usable23:54
clarkbmordred: I think a lot of people are in that boat with you going why why why?23:54
clarkbmorganfainberg: that is how they are making them23:54
clarkbmorganfainberg: and its crap23:54
mordredclarkb: if only I worked for a company that made laptops23:54
*** rhsu has joined #openstack-infra23:54
*** emagana has quit IRC23:54
mordredclarkb: it would be so easy to go and talk to someone and tell them "there is a market segment that would be undyingly loyal if only you'd make a half-assed replica of the x22023:54
morganfainbergmordred, wait... a company that makes... laptops?23:55
clarkbmordred: right you don't even need to do anything new. just copy the x22023:55
mordredclarkb: and then they'd immediately realize that for the addition of a $10 keyboard part, they could dominate a market segment23:55
clarkbmordred: the worst thing about all this is that intel basically ruined the market23:55
clarkbthen windows8 made it worse23:55
mordredclarkb: it was intel?23:55
clarkbmordred: yes intel pushed the ultrabook thing23:55
mordredwhy can't I have an ultrabook with a real keyboard?23:56
clarkband to go super thin like that you have to sacrifice keyboard quality, battery life, and pointers23:56
clarkbmordred: because real keyboard needs room23:56
morganfainbergclarkb, and most ultrabooks aren't *that* thin23:56
mordredhell - at THIS POINT I think I'd even eat the chicklet keyboard if I could have mouse buttons back23:56
mordredclarkb: nod23:56
morganfainbergclarkb, which is the real tragedy23:56
clarkband now with windows 8 everything has a touch screen23:56
morganfainbergclarkb, god. no. more. touch. screens. in. desktop. replacements.23:56
clarkbmordred: buttons need room too23:57
clarkbbut ya Intel ruined the market23:57
clarkbbut it was always a play at copying apple23:57
morganfainbergnot sure if it was intel or apple in that case.. apple released the macbook air and that *might* have helped intel push for ultrabooks because it had fairly high ratings/adoption23:57
clarkbmorganfainberg: apple did thing and intel told everyone else to do it too23:58
*** todd_dsm has quit IRC23:58
mordredclarkb: you used to work there - I blame you23:58
clarkbmordred: it is my fault23:59
clarkbI shouldn't told paul no no no23:59
clarkber should've23:59
clarkbhe is retired now too23:59

Generated by 2.14.0 by Marius Gedminas - find it at!