Wednesday, 2015-07-08

*** dannywilson has quit IRC00:00
greghaynesAlso, using the power of expect, one can grab serials for all our servers in hp2 if anyone was wondering ;)00:00
greghayneswithout having to boot them00:00
* fungi expected as much00:00
nibalizerfungi: *slow clap*00:01
*** salv-orlando has joined #openstack-infra00:01
*** bdemers has quit IRC00:01
fungii'll be here all night folks (seriously, i'm wired into the circuitry!)00:01
fungidon't forget to tip the wait staff00:01
*** claudiub has quit IRC00:02
* fungi should haev said "wait(2) staff"00:02
*** kdas_ has joined #openstack-infra00:02
openstackgerritMike Perez proposed openstack/requirements: Bumping Cinder client to 1.3.0
*** Guest7393 has quit IRC00:04
*** sdake_ is now known as sdake00:04
mordredI clearly walked away at the WRONG time and missed that00:06
clarkbhow did image builds go?00:06
*** heyongli has joined #openstack-infra00:07
mordredclarkb: I really want to make a joke response, but fungi has won for the evening already00:07
*** kdas_ has quit IRC00:07
fungiclarkb: for devstack-trusty-2-node?00:07
*** achanda has quit IRC00:08
fungilooks like we got new images00:08
fungiin all 8 locations00:08
*** gyee has joined #openstack-infra00:09
fungii haven't observed any multinode jobs in the ~4 hours since the last of those images built00:09
fungiso not sure if it solved the issue00:09
clarkbI can kick some off00:10
*** ddieterly has joined #openstack-infra00:10
fungino need00:10
fungiclarkb: just completed00:11
*** heyongli has quit IRC00:11
*** chlong has joined #openstack-infra00:12
fungii think the hostnames look sane now00:12
fungioh, wait, that's rax. were we exceeding there or just in hpcloud?00:12
clarkbin rax too looks like00:13
clarkbit failed on subnode devstacking00:14
fungisudo: unable to resolve host devstack-trusty-2-node-rax-iad-3535243-13519.slave.openstack.or00:14
*** salv-orlando has quit IRC00:14
fungimordred: greghaynes:
openstackgerritThai Tran proposed openstack/requirements: Create a repository for stackforge/xstatic-angular-gettext
mordredfungi: I blame greghaynes00:18
*** puranamr has joined #openstack-infra00:18
fungiclarkb: jhesketh: lifeless: looks like we've got /usr/zuul-env/bin/zuul-cloner on now (and on too)00:18
openstackgerritThai Tran proposed openstack/requirements: Adding angular-gettext to global requirements
lifelessfungi: can we tickle the job ?00:19
fungilifeless: i'm working on a manual run now. i'll |tee it to a log i can stick on paste.o.o00:20
freerunnerHi folks! I have a question about upper-constraints in requirements. Murano-dsvm-job tests failed due installation of ply==3.6 which listed in upper-constraints but not listed in global-requirements instead of 3.4. What we should to do? We need to remove ply from upper-constraints at all or fix version of ply in upper-constraints?00:20
mordredfreerunner: lifeless is your man00:21
*** hdd has quit IRC00:21
freerunnermordred: ty00:21
lifelessfreerunner: I don't see ply in global-requirements at all00:21
freerunnerlifeless: Yep.00:22
openstackgerritlifeless proposed openstack/requirements: Don't permit embedded #'s in URLs.
openstackgerritlifeless proposed openstack/requirements: Manual run of generate-constraints.
lifelessfreerunner: so whats the problem ?00:22
*** tsekiyam_ has joined #openstack-infra00:22
freerunnerlifeless: ply in upper-constraints =(00:23
lifelessfreerunner: ply is a transitive dependency of something in global-requirements.txt00:23
lifelessfreerunner: 3.6 is the version resulting from installing all of global-requirements.txt00:23
*** mayurig has quit IRC00:23
*** dims has joined #openstack-infra00:23
lifelessfreerunner: if you want to change that version, there are two options: a) put ply in global-requirements.txt or b) figure out the transitive dependency bringing that version in and change *that*00:23
lifelessfreerunner: upper-constraints.txt just gives us what pip does, but in advance so rather than find out in gate jobs, we find out in a controlled way00:24
*** jamielennox is now known as jamielennox|away00:24
*** doug-fish1 has left #openstack-infra00:25
*** heyongli has joined #openstack-infra00:25
freerunnerlifeless: Thanks for explanation ;)00:26
*** tsekiyama has quit IRC00:26
*** tsekiyam_ has quit IRC00:27
*** sigmavirus24 is now known as sigmavirus24_awa00:27
lifelessfreerunner: given what you've said, I'd say ply should be in global-requirements.txt00:27
lifelessand murano should have a representative job as a gate job for openstack/requirements00:28
fungignar. we still need to set zuul_ref in this job to {branch-override} i think. i'll have an update here shortly assuming i get this to work00:29
freerunnerlifeless: We working on this.00:29
*** heyongli has quit IRC00:30
*** mtanino has quit IRC00:30
jrollfreerunner: lifeless: I'm really curious what depends on ply, ply is a tool to manage a set of patches on top of upstream :/00:31
jrollunless this is a different ply I don't know about00:32
lifelessjroll: its a parser library00:32
jrolloh, plypatch is the patch tool, ignore me :)00:32
fungijroll: it's a parser00:32
jrollplypatch uses ply as the entrypoint00:32
jrollso I get confused00:32
jroll'ply restore' etc00:32
lifelessbecause terrible names are our trade00:33
jrollit is the number 1 problem in software dev00:33
fungiwe pave the internet with terrible names00:33
lifelesswe built this 'net on terr.... I can't do it00:33
*** fallenpegasus has quit IRC00:33
*** jamielennox|away is now known as jamielennox00:35
clarkbI am curious how if 3.4 vs 3.6 matters this ever worked otherwise00:35
clarkbyou got 3.6 before so it should work right? so why not just use it00:35
lifelessclarkb: I'm guessing a local pin of somesort affecting murano00:36
lifelessclarkb: e.g. lib/murano00:36
fungilifeless: out of curiosity, how long do you expect this to take at the generate-constraints stage? it's been going for some minutes already00:36
*** rlandy has joined #openstack-infra00:36
lifelessfungi: cold cache? 5-10 tops00:37
fungiokay, thanks. i'm still well within that range i think00:37
lifelessfungi: most of the time is the first build of numpy etc00:37
freerunnerclarkb: We got 3.4 all time. And in manual devstack installation we still got 3.4. Looks like there is bug in yaql + ply 3.6 =(00:37
*** vahidh has quit IRC00:37
*** Swami has quit IRC00:38
*** heyongli has joined #openstack-infra00:38
fungilifeless: okay it picked back up again though i'm a little worried...00:38
freerunnerIt looks really strange, because of compatibility with 3.6 ply was fixed a three months ago..00:39
*** ddieterl_ has joined #openstack-infra00:39
fungilifeless: Ignoring pyngus: markers "python_version=='2.7'" don't match your environment00:39
*** shardy_ has joined #openstack-infra00:39
fungilifeless: this is on an ubuntu 14.04 lts server00:39
fungilifeless: is that normal?00:39
*** shardy has quit IRC00:40
*** esmute__ has quit IRC00:40
lifelessfungi: yes00:40
lifelessfungi: its because its building 3.4 now00:40
*** ddieterly has quit IRC00:41
*** esmute__ has joined #openstack-infra00:41
fungiright, that's why i replaced the 12.04-based server with a new one last week ;)00:41
fungii suppose we could stand to add some additional informational output from the generate call, but not critical00:42
*** jasondotstar has quit IRC00:43
*** shardy_ has quit IRC00:44
*** Sukhdev has quit IRC00:45
lifelessclarkb: another possibility00:45
lifelessclarkb: two deps on ply00:45
*** shardy has joined #openstack-infra00:45
lifelessclarkb: and ordering, one is versioned lower, one not00:45
*** sdake has quit IRC00:46
*** heyongli has quit IRC00:48
openstackgerritOpenStack Proposal Bot proposed openstack/requirements: Updated from generate-constraints
lifelessfungi: what file are you using for bindep reqs ?00:50
fungilifeless: if the project doesn't have a other-requirements.txt in it00:50
* fungi is in the middle of updating the experimental job to do that bit now00:50
*** yamada-h has joined #openstack-infra00:51
asselinzaro, you around?00:51
lifelessahha other-requirements.txt00:51
fungilifeless: yep, the bindep default you originally picked00:51
*** esmute__ has quit IRC00:52
*** heyongli has joined #openstack-infra00:52
lifelessfungi: yah, I saw it come through here.00:55
lifelesswe're in business00:55
fungilifeless: we only proposed a constraints update to the requirements repo... suspecting we've short-circuited the loop somewhere00:55
lifelessfungi: no, thats all this does00:55
lifelessfungi: its what stops us resting on our laurels forever00:55
fungilifeless: oh, it's only supposed to propose to requirements, not to every project? i think i forgot that ;)00:55
fungianyway, one-liner fix for the job definition on the way in a few minutes00:55
lifeless'all'. Key bit of loop closing and its 'all' :)00:56
*** shashank_hegde has quit IRC00:57
openstackgerritlifeless proposed openstack-infra/release-tools: Document repo format for list_unreleased_changes
openstackgerritlifeless proposed openstack-infra/release-tools: Permit using local repos in release_postversion
openstackgerritlifeless proposed openstack-infra/release-tools: Document binary package requirements.
openstackgerritlifeless proposed openstack-infra/release-tools: Ignore setuptool's .eggs dir.
*** heyongli has quit IRC00:57
*** yamada-h has quit IRC00:57
*** shashank_hegde has joined #openstack-infra00:57
*** gildub has quit IRC00:57
*** madong_ has quit IRC01:02
*** shashank_hegde has quit IRC01:04
*** signed8bit has quit IRC01:07
openstackgerritMerged openstack-infra/release-tools: Document repo format for list_unreleased_changes
*** btully has quit IRC01:09
*** gmann has quit IRC01:12
*** salv-orlando has joined #openstack-infra01:12
*** Longgeek has joined #openstack-infra01:16
*** sdake has joined #openstack-infra01:16
*** baoli has joined #openstack-infra01:16
*** yamada-h has joined #openstack-infra01:17
*** rlandy has quit IRC01:17
openstackgerritlifeless proposed openstack-infra/release-tools: Document binary package requirements.
openstackgerritlifeless proposed openstack-infra/release-tools: Permit using local repos in release_postversion
openstackgerritlifeless proposed openstack-infra/release-tools: Ignore setuptool's .eggs dir.
openstackgerritlifeless proposed openstack-infra/release-tools: Fixup requirements
*** Longgeek has quit IRC01:20
*** heyongli has joined #openstack-infra01:20
*** gildub has joined #openstack-infra01:20
*** sdake_ has joined #openstack-infra01:20
*** yamada-h has quit IRC01:20
*** davideagnello has quit IRC01:21
*** spzala has joined #openstack-infra01:22
openstackgerritJeremy Stanley proposed openstack-infra/project-config: Use bindep in thin worker experimental job
*** annegentle has joined #openstack-infra01:23
*** sdake has quit IRC01:23
*** yamada-h has joined #openstack-infra01:24
*** dims_ has joined #openstack-infra01:24
*** dims_ has quit IRC01:24
*** heyongli has quit IRC01:25
*** yamada-h has quit IRC01:25
*** salv-orlando has quit IRC01:26
openstackgerritJeremy Stanley proposed openstack-infra/project-config: Provide ZUUL_BRANCH in constraints generation job
fungilifeless: ^ that was necessary to get the constraints job to run01:27
*** dims has quit IRC01:27
*** annegentle has quit IRC01:27
*** hogepodge has quit IRC01:29
lifelessjhesketh: clarkb:
lifeless(I'd really like to get this squared away)01:30
*** heyongli has joined #openstack-infra01:30
fungialso feel free to take it over if it's not as belts-and-braces as you'd like. i'm about to head to bed01:31
*** woodster_ has quit IRC01:31
jheskethlifeless: lgtm01:32
*** sdake_ is now known as sdake01:34
clarkbdoesnt cloner work without the zuul params?01:34
clarkbmaybe the job is passing in enough that it wants branch?01:35
lifelessit needs to run off of a branch, either master or kilo are currently defined in project-config01:36
lifelessnot that we expect kilo to have constraints backported to it - but certainly liberty's branch will01:36
fungiclarkb: it does not01:37
*** gmann has joined #openstack-infra01:37
*** heyongli has quit IRC01:38
fungiclarkb: at least in periodic pipeline jobs, it needs either ZUUL_BRANCH or ZUUL_REF otherwise you get an error01:38
fungibut also, we want to run it for stable branches as well as master, so we really need to tell zuul-cloner what branch we expect too01:40
*** hogepodge has joined #openstack-infra01:41
*** sabeen has joined #openstack-infra01:42
*** puranamr has quit IRC01:44
openstackgerritOm Kumar proposed openstack/diskimage-builder: Adds Ubuntu and Debian to ironic-python-agent Support-list
openstackgerritOm Kumar proposed openstack/diskimage-builder: Removes hardcoded refrences for ethernet interface
*** thorst has joined #openstack-infra01:45
*** heyongli has joined #openstack-infra01:45
pleia2StevenK: thanks re: manuals, I was a bit confused on that point and your email helped01:45
*** armax has quit IRC01:46
StevenKpleia2: And I realized about 30 seconds after I sent it that it also impacts Horizon -- django_openstack_auth is seperate for the same reason.01:47
* pleia2 nods01:48
*** thorst has quit IRC01:50
StevenKpleia2: I will fiddle with the etherpad today, trying to come up with wording for Manuals and Horizon01:50
*** laughterwym_ has joined #openstack-infra01:50
pleia2StevenK: awesome, thanks01:50
*** laughterwym has quit IRC01:53
*** unicell has quit IRC01:53
clarkbfungi it is supposed to default to master and not fail01:54
clarkbI need to test that and fix it I guess01:55
*** asselin_ has joined #openstack-infra01:55
fungiclarkb: then it's the combination of parameters you get in a periodic pipeline which throw it off and make it think it needs something more01:55
fungiclarkb: regardless, that job needed to export a branch name for zuul-cloner to consume because we're running it on more than just master01:56
*** sridhar_ram has quit IRC01:57
openstackgerritMerged openstack-infra/project-config: Provide ZUUL_BRANCH in constraints generation job
fungiwhich means maybe we want it to continue to be "broken" that way, to catch when we try to use it in periodic jobs without explicitly setting a branch01:57
*** heyongli has quit IRC01:57
*** heyongli has joined #openstack-infra01:58
clarkbor have a glag to be strict02:00
jheskethclarkb, fungi: either of you good people know much about the apache puppet module?02:04
*** gyee has quit IRC02:04
fungii know next to nothing about modern versions of it, just our ancient fork02:05
*** dprince has joined #openstack-infra02:06
jheskethfungi: according to the modules.env file we're using the puppetlabs version?02:07
*** btully has joined #openstack-infra02:07
jheskethmind you, it is set to 0.0.4 which looks ancient02:07
*** asettle has joined #openstack-infra02:08
jheskethbasically I'm just looking for a way to notify a restart02:08
jheskethor should I just exec 'service apache2 restart'?02:09
jheskethwhich seems terrible and won't work on centos02:09
*** heyongli has quit IRC02:10
fungijhesketh: what do you need to restart apache on? i think it already triggers restarts if you change modules02:10
jheskethfungi: when os-loganalyze is updated02:10
fungiso presumably has the necessary logic exposed somewhere02:10
jheskethfungi: as in when apache mods are changed?02:11
fungibut off the top of my head i don't know of anywhere specifically we're doing arbitrary restarts through the apache puppet module02:11
fungiyeah, apache mod_whatever changes02:11
*** heyongli has joined #openstack-infra02:12
fungianyway, i'm out for the night02:13
jheskethhmm okay02:14
jheskeththanks, have a good evening :-)02:14
clarkbjhesketh you canhave the apache service subscribe to your os lganalyze stuff02:15
jheskethclarkb: right, but I can't see where for logserver the apache service is defined02:15
jheskethwe include the apache module, but don't instantiate the class anywhere obvious02:16
clarkbI think include apache does it02:16
jheskethclarkb: right, so I'd have to overwrite what that does02:16
jheskethclarkb: do you konw what our puppet-httpd does? it looks like it might overwrite some of the puppetlabs-apache classes02:16
jheskethI'm trying to figure out if it's in a different namespace or not02:16
*** heyongli has quit IRC02:17
pleia2I seem to recall we called it puppet-httpd specifically to avoid the name collision02:17
*** oomichi has joined #openstack-infra02:17
nibalizerjhesketh: it shouldn't be instantiated along with a host that has puppet-apache on it02:18
pleia2so we should be able to ease into using the new module02:18
nibalizerone or teh other02:18
clarkbits just there to allow us to use old apache module without name colliding02:18
nibalizerbut they can exist on the puppetmaster together02:18
clarkbbecause new module is bad02:18
clarkbbut things want to use new module02:18
nibalizerand pabelanger has been doing the work of migrating us so \o/02:18
jheskethright, but puppet-httpd declairs an 'apache' class02:19
nibalizerwell thats a bug02:19
nibalizerlet me look02:19
nibalizeri think its a bug anyways02:19
jheskethnibalizer: heh, it's more likely I don't understand what's going on02:19
*** dprince has quit IRC02:20
openstackgerritMerged openstack/requirements: Don't permit embedded #'s in URLs.
jheskethlooking at how apache is used in puppet-cgit and elsewhere, ^ looks like the thing to do02:21
openstackgerritJoshua Hesketh proposed openstack-infra/puppet-openstackci: Reload apache on os-loganalyze update
nibalizerjhesketh: so the httpd module defines a class 'httpd'02:21
jheskethheh, ^^ looks like the thing to do :p02:21
nibalizeroh probably not02:21
nibalizerso we /want/ to move to puppet-httpd, haven't acutally done it in most places yet02:22
nibalizeror any places, maybe02:22
*** jasondotstar has joined #openstack-infra02:22
jheskethnibalizer: sigh, my puppet-httpd branch is so out of date it still says apache :-(02:22
nibalizerhah, glad we cleaned that up02:22
jheskethnibalizer: so do you know an easy way to reload the old apache puppet module without moving everything across to the httpd module02:22
*** heyongli has joined #openstack-infra02:24
clarkbalso you cant reload iirc its always a restart because puppet02:24
nibalizerya reviewing your thing02:24
*** heyongli has quit IRC02:28
*** heyongli has joined #openstack-infra02:28
*** patrickeast has quit IRC02:29
*** yamamoto has joined #openstack-infra02:32
*** annegentle has joined #openstack-infra02:34
*** bhunter71 has quit IRC02:35
*** rossella_s has quit IRC02:37
*** rossella_s has joined #openstack-infra02:37
*** annegentle has quit IRC02:37
*** shardy_ has joined #openstack-infra02:38
*** hdd has joined #openstack-infra02:38
*** shardy has quit IRC02:39
*** davideagnello has joined #openstack-infra02:41
*** shardy_ has quit IRC02:43
*** shardy has joined #openstack-infra02:43
*** heyongli has quit IRC02:43
*** heyongli has joined #openstack-infra02:45
*** BharatK has joined #openstack-infra02:45
*** rkukura has joined #openstack-infra02:45
*** changbl has joined #openstack-infra02:48
*** fawadkhaliq has joined #openstack-infra02:51
*** Kennan2 is now known as Kennan02:53
*** jasondotstar has quit IRC03:02
*** ddieterl_ has quit IRC03:02
jheskethnibalizer: thanks for your review, I've replied:
jheskethclarkb: ^ if you're interested03:06
*** yamahata has quit IRC03:06
jheskethclarkb: btw, I /think/ this might have been why osloganalze failed during that patch series a while ago (but again having not seen the logs makes it difficult to tell)03:07
clarkboh interesting03:08
clarkbI have to do this feeding but will try to take a look03:08
jheskethno pressure :-)03:09
nibalizerjhesketh: ya i think you're right03:09
jheskethalso, holy crap, that module is outdated ;p03:10
clarkbbut it supports things like mod rewrite :)03:11
jheskeththere's a reason I'm not attempting to upgrade it ;-)03:12
*** puranamr has joined #openstack-infra03:13
*** salv-orlando has joined #openstack-infra03:13
*** Longgeek has joined #openstack-infra03:16
*** binbincong has quit IRC03:19
*** Longgeek has quit IRC03:22
*** Sukhdev has joined #openstack-infra03:24
*** xyang1 has quit IRC03:28
*** salv-orlando has quit IRC03:28
*** puranamr has quit IRC03:30
*** puranamr has joined #openstack-infra03:30
*** Sukhdev has quit IRC03:31
*** binbincong has joined #openstack-infra03:33
*** annegentle has joined #openstack-infra03:37
*** kiran-r has joined #openstack-infra03:38
*** BharatK has quit IRC03:40
*** annegentle has quit IRC03:41
*** sdake has quit IRC03:42
*** puranamr has quit IRC03:43
*** patrickeast has joined #openstack-infra03:45
openstackgerritYAMAMOTO Takashi proposed openstack-infra/project-config: Move Midonet v2 job to the experimental queue
openstackgerritYAMAMOTO Takashi proposed openstack-infra/project-config: midonet-v2: Update after recent plugin changes
*** koolhead17 has joined #openstack-infra03:45
*** patrickeast has quit IRC03:49
*** patrickeast has joined #openstack-infra03:49
*** patrickeast has quit IRC03:50
*** zz_dimtruck is now known as dimtruck03:51
*** patrickeast has joined #openstack-infra03:51
*** BharatK has joined #openstack-infra03:51
*** koolhead17 has quit IRC03:51
*** boris-42 has quit IRC03:52
*** tlian has quit IRC03:53
*** baoli has quit IRC03:53
*** ayoung has quit IRC03:56
lifelessjhesketh: +2A please03:56
lifelessclarkb: ^ answered the point about routes in that for you03:56
openstackgerritSteve Martinelli proposed openstack/requirements: Bump oslo.cache to 0.2.0
*** yamahata has joined #openstack-infra03:57
jheskethlifeless: +2, but I've got an appointment in a bit, so I'm cautious of merging it03:58
lifelessjhesketh: k03:58
*** armax has joined #openstack-infra04:02
*** ddieterly has joined #openstack-infra04:03
*** jamespage has quit IRC04:05
*** gnuoy has quit IRC04:05
*** gnuoy has joined #openstack-infra04:06
*** jamespage has joined #openstack-infra04:06
*** mtreinish has quit IRC04:07
*** ddieterly has quit IRC04:08
*** fawadkhaliq has quit IRC04:11
*** r-daneel has joined #openstack-infra04:13
*** kiran-r has quit IRC04:15
*** mtreinish has joined #openstack-infra04:16
openstackgerritSteve Kowalik proposed openstack-infra/project-config: Add openstack-manuals to populate-zanata-project
*** btully has quit IRC04:19
*** david-ly_ has joined #openstack-infra04:19
*** david-lyle has quit IRC04:22
*** heyongli has quit IRC04:22
*** maishsk has quit IRC04:22
*** fallenpegasus has joined #openstack-infra04:29
*** salv-orlando has joined #openstack-infra04:29
*** maishsk has joined #openstack-infra04:30
*** Sukhdev has joined #openstack-infra04:31
*** maishsk has quit IRC04:31
*** maishsk has joined #openstack-infra04:31
*** heyongli has joined #openstack-infra04:32
*** chlong has quit IRC04:35
*** maishsk has quit IRC04:36
openstackgerritJarod Du proposed openstack-infra/project-config: Add networking-lenovo to stackforge
*** AzherKhan has joined #openstack-infra04:41
*** salv-orlando has quit IRC04:42
*** laughterwym__ has joined #openstack-infra04:43
*** harlowja_at_home has joined #openstack-infra04:43
*** chlong has joined #openstack-infra04:44
*** unicell has joined #openstack-infra04:46
*** laughterwym_ has quit IRC04:46
*** mgedmin has quit IRC04:47
*** davideagnello has quit IRC04:47
*** amotoki has joined #openstack-infra04:49
*** mgedmin has joined #openstack-infra04:50
*** chlong has quit IRC04:51
*** btully has joined #openstack-infra04:51
*** yamada-h has joined #openstack-infra04:52
*** changbl has quit IRC04:52
*** fawadkhaliq has joined #openstack-infra04:55
*** fawadk has joined #openstack-infra04:59
*** AK has joined #openstack-infra05:00
*** HeOS has quit IRC05:00
*** fawadkhaliq has quit IRC05:02
*** AzherKhan has quit IRC05:03
*** ddieterly has joined #openstack-infra05:04
*** ddieterly has quit IRC05:09
*** yamada-h_ has joined #openstack-infra05:09
*** dims has joined #openstack-infra05:11
*** AK has quit IRC05:11
*** AzherKhan has joined #openstack-infra05:11
*** yamada-h has quit IRC05:12
openstackgerritKhai Do proposed openstack-infra/puppet-openstackci: Add Jenkins Job Builder manifest
*** abregman has joined #openstack-infra05:16
*** Longgeek has joined #openstack-infra05:16
*** hdd has quit IRC05:17
*** deepakcs has joined #openstack-infra05:17
*** unicell has quit IRC05:18
*** unicell has joined #openstack-infra05:18
*** unicell has quit IRC05:18
*** unicell has joined #openstack-infra05:18
*** asettle has quit IRC05:18
*** achanda has joined #openstack-infra05:20
*** chlong has joined #openstack-infra05:20
*** boris-42 has joined #openstack-infra05:21
*** camunoz_mtg has quit IRC05:23
openstackgerritKhai Do proposed openstack-infra/system-config: Use third party ci manifest to setup jenkins job builder.
*** annegentle has joined #openstack-infra05:26
*** flepied has quit IRC05:29
*** camunoz has joined #openstack-infra05:29
*** annegentle has quit IRC05:30
*** flepied has joined #openstack-infra05:30
*** mpaolino has joined #openstack-infra05:32
*** mpaolino has quit IRC05:33
*** dims has quit IRC05:40
openstackgerritKhai Do proposed openstack-infra/jenkins-job-builder: Improve delete all job performance
*** maishsk has joined #openstack-infra05:51
*** ig0r__ has joined #openstack-infra05:51
*** flepied has quit IRC05:51
*** yamada-h_ has quit IRC05:52
*** ig0r_ has quit IRC05:52
*** heyongli has quit IRC05:54
*** heyongli has joined #openstack-infra05:55
*** claudiub has joined #openstack-infra05:55
*** scheuran has joined #openstack-infra06:05
*** achanda has quit IRC06:05
*** ParsectiX has joined #openstack-infra06:05
*** ddieterly has joined #openstack-infra06:05
*** dguerri` is now known as dguerri06:07
*** rfolco has quit IRC06:09
*** ddieterly has quit IRC06:10
*** smyers has quit IRC06:11
*** kiran-r has joined #openstack-infra06:12
*** dguerri is now known as dguerri`06:14
*** smyers has joined #openstack-infra06:16
*** dimtruck is now known as zz_dimtruck06:17
*** patrickeast has quit IRC06:19
*** rfolco has joined #openstack-infra06:21
*** kushal has joined #openstack-infra06:23
*** yamada-h has joined #openstack-infra06:24
*** flepied has joined #openstack-infra06:30
*** claudiub has quit IRC06:31
*** salv-orlando has joined #openstack-infra06:31
*** Sukhdev has quit IRC06:31
*** coolsvap|away is now known as coolsvap06:32
*** ildikov has quit IRC06:35
*** fawadk has quit IRC06:36
*** rossella_s has quit IRC06:36
*** rossella_s has joined #openstack-infra06:37
*** otter768 has quit IRC06:38
*** mrunge has joined #openstack-infra06:41
*** salv-orlando has quit IRC06:42
*** fhubik has joined #openstack-infra06:45
*** jtomasek has joined #openstack-infra06:45
*** fhubik is now known as fhubik_afk06:45
*** markus_z has joined #openstack-infra06:46
*** juggler_ has joined #openstack-infra06:49
*** yamamoto has quit IRC06:49
*** eantyshev has quit IRC06:50
*** harlowja_at_home has quit IRC06:52
*** juggler has quit IRC06:52
*** isviridov_away is now known as isviridov06:52
*** camunoz is now known as camunoz_gone06:54
*** binbincong has quit IRC06:56
*** markus_z has quit IRC06:58
*** k4n0 has joined #openstack-infra06:59
*** yamamoto has joined #openstack-infra07:00
*** fawadkhaliq has joined #openstack-infra07:00
*** fhubik_afk is now known as fhubik07:01
*** Sukhdev has joined #openstack-infra07:02
*** salv-orlando has joined #openstack-infra07:03
*** yamada-h has quit IRC07:05
*** ddieterly has joined #openstack-infra07:06
*** fhubik is now known as fhubik_afk07:08
*** yamada-h has joined #openstack-infra07:10
*** fhubik_afk is now known as fhubik07:11
*** ddieterly has quit IRC07:11
*** soren has joined #openstack-infra07:11
*** isviridov is now known as isviridov_away07:14
*** luqas has joined #openstack-infra07:14
*** dtantsur|afk is now known as dtantsur07:14
*** annegentle has joined #openstack-infra07:14
*** binbincong has joined #openstack-infra07:15
*** heyongli has quit IRC07:18
*** annegentle has quit IRC07:19
*** yamada-h has quit IRC07:19
*** yamahata has quit IRC07:20
*** asettle has joined #openstack-infra07:20
*** luqas has quit IRC07:22
*** heyongli has joined #openstack-infra07:23
*** fhubik is now known as fhubik_afk07:23
*** e0ne has joined #openstack-infra07:25
*** fhubik_afk is now known as fhubik07:26
*** yamada-h has joined #openstack-infra07:26
openstackgerritSergey Vilgelm proposed openstack/requirements: Bump oslo.log to 1.6.0 version
*** fhubik has quit IRC07:32
openstackgerritSteve Kowalik proposed openstack-infra/project-config: Add a script and job
*** arxcruz has joined #openstack-infra07:33
*** scheuran_ has joined #openstack-infra07:33
*** fhubik has joined #openstack-infra07:33
*** dizquierdo has joined #openstack-infra07:34
*** scheuran has quit IRC07:34
*** bexelbie has joined #openstack-infra07:34
*** afazekas has joined #openstack-infra07:34
*** sshnaidm has joined #openstack-infra07:36
*** fhubik is now known as fhubik_afk07:39
*** ildikov has joined #openstack-infra07:39
mrungehello, I would like to propose a patch to django_openstack_auth in stable/kilo branch07:40
mrungegit review tells me: ! [remote rejected] HEAD -> refs/publish/stable/kilo/bug/145223207:40
*** vivcheri has quit IRC07:40
mrungeis there something else to configure?07:40
*** Sukhdev has quit IRC07:41
mrungeit seems, it was the first backport for stable/kilo for django_openstack_auth07:41
*** dims has joined #openstack-infra07:41
*** hashar has joined #openstack-infra07:41
openstackgerritMerged openstack-infra/puppet-openstackci: Reload apache on os-loganalyze update
*** dims_ has joined #openstack-infra07:42
*** fhubik_afk is now known as fhubik07:43
*** jistr has joined #openstack-infra07:43
*** dizquierdo has quit IRC07:45
*** dims has quit IRC07:45
*** btully has quit IRC07:46
*** dims_ has quit IRC07:46
*** binbincong has quit IRC07:46
*** markvoelker has quit IRC07:47
clarkbit should tell you why it rejected it07:48
clarkbcan you paste the full git review output?07:48
*** ihrachyshka has joined #openstack-infra07:48
*** e0ne is now known as e0ne_07:49
mrungeclarkb, the full output is
mrungereproducer is easy07:50
*** e0ne_ is now known as e0ne07:51
mrungegit checkout origin/stable/kilo07:51
mrungegit cherry-pick -x 8aa029307:51
mrungegit review07:51
*** _nadya_ has joined #openstack-infra07:51
clarkbthe error is in the paste, no changes07:51
clarkbso gerrit says its a noop and refuses the commit07:52
mrungeunfortunately, there is a change07:52
clarkbnot according to git07:52
mrungethat should mean, the patch is already in07:52
clarkbwhat does git diff HEAD^ say?07:53
mrungedisplays the patch from cherry-pick07:53
clarkboh is the change already proposed?07:54
mrungeto be really sure, everything was clean, I did a clean checkout, just now07:54
mrungeclarkb, good catch, I missed that07:54
clarkbya I bet there us already a change up fir it07:54
mrungeit is up07:54
clarkb(this is why cherry picking can be bad)07:55
mrungethanks a lot for your help, and sorry for the noise07:55
clarkbit breaks any association between commits and gerrit rejects noop work07:55
*** arxcruz has quit IRC07:55
*** arxcruz has joined #openstack-infra07:56
*** afazekas has quit IRC07:56
*** jlanoux has joined #openstack-infra07:57
*** e0ne is now known as e0ne_07:57
*** shashank_hegde has joined #openstack-infra07:58
*** e0ne_ is now known as e0ne07:59
*** binbincong has joined #openstack-infra08:00
*** e0ne has quit IRC08:00
*** asettle has quit IRC08:01
*** ociuhandu has joined #openstack-infra08:01
*** armax has quit IRC08:02
*** HeOS has joined #openstack-infra08:04
*** ddieterly has joined #openstack-infra08:06
*** Ala has joined #openstack-infra08:07
*** afazekas has joined #openstack-infra08:09
*** ddieterly has quit IRC08:11
*** ociuhandu has quit IRC08:13
*** HeOS has quit IRC08:15
*** binbincong has quit IRC08:16
*** shashank_hegde has quit IRC08:17
*** binbincong has joined #openstack-infra08:18
*** jcoufal has joined #openstack-infra08:19
*** HeOS has joined #openstack-infra08:20
*** hichihara has quit IRC08:20
*** ildikov has quit IRC08:21
*** chlong has quit IRC08:21
*** heyongli has quit IRC08:22
*** heyongli has joined #openstack-infra08:23
*** fallenpegasus has quit IRC08:23
*** gildub has quit IRC08:25
*** alexsyip has joined #openstack-infra08:25
*** binbincong has quit IRC08:29
*** yamada-h has quit IRC08:29
*** pelix has joined #openstack-infra08:31
*** boris-42 has quit IRC08:32
*** ildikov has joined #openstack-infra08:35
*** eantyshev has joined #openstack-infra08:36
*** oomichi has quit IRC08:37
*** weshay has quit IRC08:37
*** weshay has joined #openstack-infra08:38
*** nilasae has joined #openstack-infra08:38
*** dguitarbite has joined #openstack-infra08:39
*** dims has joined #openstack-infra08:43
*** dims_ has joined #openstack-infra08:44
*** mpavlase has joined #openstack-infra08:44
*** laughterwym__ has quit IRC08:44
*** dims__ has joined #openstack-infra08:45
*** dims___ has joined #openstack-infra08:46
*** mpavlase has left #openstack-infra08:47
*** dims has quit IRC08:47
*** markvoelker has joined #openstack-infra08:48
*** dims_ has quit IRC08:49
*** dims__ has quit IRC08:49
*** dims___ has quit IRC08:50
openstackgerritIgor Belikov proposed openstack-infra/project-config: Fix acls for stable/6.1 branch of fuel projects
*** markvoelker has quit IRC08:53
*** mfmcdonagh has joined #openstack-infra08:55
*** mcdonagm has joined #openstack-infra08:55
*** devvesa has joined #openstack-infra08:59
*** maishsk has quit IRC09:01
*** nijaba has quit IRC09:02
*** dizquierdo has joined #openstack-infra09:02
*** e0ne has joined #openstack-infra09:03
*** annegentle has joined #openstack-infra09:03
openstackgerritIgor Belikov proposed openstack-infra/project-config: Fix acls for stable/6.1 branch of fuel projects
*** maishsk has joined #openstack-infra09:04
*** flepied has quit IRC09:05
*** scheuran has joined #openstack-infra09:05
*** nijaba has joined #openstack-infra09:06
*** scheuran_ has quit IRC09:07
*** ddieterly has joined #openstack-infra09:07
*** annegentle has quit IRC09:08
*** r-daneel has quit IRC09:08
zigolifeless: Hey, can I take over the python-testtools package and push it to the OpenStack PKG group in Alioth, rather than keeping the stupid SVN?09:10
openstackgerritAndreas Jaeger proposed openstack-infra/project-config: Add missing projects to specs list
zigo(ie: moving it out of the DPMT team until they switch to Git...)09:11
*** ddieterly has quit IRC09:11
*** afazekas has quit IRC09:12
*** hashar_ has joined #openstack-infra09:13
*** hashar has quit IRC09:14
*** alexsyip has quit IRC09:15
*** madong has joined #openstack-infra09:15
*** markus_z has joined #openstack-infra09:16
*** laughterwym has joined #openstack-infra09:19
*** john-davidge has joined #openstack-infra09:19
SergeyLukjanovlifeless, hey, approved09:19
*** devvesa has quit IRC09:20
lifelessSergeyLukjanov: thanks09:21
lifelesszigo: ask the maintainers09:21
*** e0ne is now known as e0ne_09:21
*** afazekas has joined #openstack-infra09:21
zigolifeless: Which is you, me and Jelmer Vernooij <>09:21
*** flepied has joined #openstack-infra09:22
lifelesshuh, I didn't realise I was still listed ;)09:22
zigolifeless: Is it ok for *you* ?09:22
zigoWell, we still have a super-old testtools in Sid.09:22
zigoNow I need to update it because Nova Liberty b1 needs >= 1.4.009:22
zigoIt needs unittest2 >= 1.0.0, experimental doesn't even have that... :(09:22
*** john-davidge_ has joined #openstack-infra09:22
* zigo goes to lunch09:23
*** ajo is now known as ajo|pocassemblin09:24
*** ajo|pocassemblin is now known as ajo|QoSPOC09:24
*** fhubik is now known as fhubik_afk09:24
*** john-davidge has quit IRC09:25
*** john-davidge_ is now known as john-davidge09:25
*** fawadkhaliq has quit IRC09:25
*** e0ne_ is now known as e0ne09:26
*** hashar_ is now known as hashar09:26
*** fawadkhaliq has joined #openstack-infra09:30
*** fhubik_afk is now known as fhubik09:31
*** ParsectiX has quit IRC09:31
openstackgerritMerged openstack-infra/project-config: Fix acls for stable/6.1 branch of fuel projects
*** ParsectiX has joined #openstack-infra09:32
Nggreghaynes: yeah those machines are no longer required. They were for upgrade testing09:33
lifelesszigo: I don't particularly care. I'd rather not splitout from a decent sized team to a team of ?2 ?, simply due to the bus factor thing.09:37
*** e0ne is now known as e0ne_09:37
lifelesszigo: can't you use svn for an upload or two until the DPMT migration occurs?09:37
openstackgerritVladimir Kozhukalov proposed openstack-infra/project-config: Add project fuel-agent repo
*** devvesa has joined #openstack-infra09:40
openstackgerritVladimir Kozhukalov proposed openstack-infra/project-config: Add project fuel-agent repo
openstackgerritHao Liu proposed openstack-infra/jenkins-job-builder: Replace templates by traversing component
*** ociuhandu has joined #openstack-infra09:44
*** jasondotstar has joined #openstack-infra09:45
*** maishsk has quit IRC09:47
*** maishsk has joined #openstack-infra09:47
openstackgerritDave Walker proposed openstack/requirements: [WIP] Added stackforge/anchor to projects.txt
*** _nadya_ has quit IRC09:55
*** fhubik is now known as fhubik_afk09:59
*** e0ne_ is now known as e0ne10:00
*** fhubik_afk is now known as fhubik10:02
*** _nadya_ has joined #openstack-infra10:02
*** maishsk has quit IRC10:06
*** claudiub has joined #openstack-infra10:07
*** ddieterly has joined #openstack-infra10:08
*** e0ne is now known as e0ne_10:11
*** bradjones has quit IRC10:13
*** ddieterly has quit IRC10:13
*** e0ne_ is now known as e0ne10:13
*** bradjones has joined #openstack-infra10:15
*** bradjones has quit IRC10:15
*** bradjones has joined #openstack-infra10:15
*** fawadkhaliq has quit IRC10:19
*** mpaolino has joined #openstack-infra10:19
*** fawadkhaliq has joined #openstack-infra10:19
*** maishsk has joined #openstack-infra10:19
*** pc_m has joined #openstack-infra10:21
*** mpaolino has quit IRC10:22
*** vikram has left #openstack-infra10:23
DavieyAJaeger: Are you sure the openstack/requirements project.txt change should Depend-On project-config zuul layout.yaml change and not the other way around?10:24
Daviey^ Or anyone else, when adding a project to global-requirements10:24
AJaegerDaviey, I would stage them this way: The rename, on top of that the layout.yaml and after that the requirements one.10:24
DavieyAJaeger: So layout.yaml can have a check-requirements job without it being in project.txt first?10:25
AJaegerDaviey both are independent. check-requirements gates that you only have lines in there that are part of global-requirements10:25
AJaegerThe projects.txt pushes changes from global-requirements to your project.10:25
*** maishsk has quit IRC10:25
DavieyAJaeger: So when this is first added, should it be marked as non-voting?10:26
AJaegerYou could make them independent as well but with the rename, it's better to depend - and running check-requirement is part of the contract10:26
AJaegerDaviey: Check your local requirements file now and if it's fine, no need for non-voting10:26
DavieyAJaeger: Yeah, it currently fails.10:27
AJaegerAnd check-requirement only gates on *changes*.10:27
DavieyAh, perfect10:27
DavieyAJaeger: Okay, that is helpful.. thanks :)10:27
openstackgerritMerged openstack/requirements: Updated from generate-constraints
AJaegerDaviey: you're welcome!10:27
openstackgerritMerged openstack-infra/releasestatus: Loop to retrieve all reviews
zigolifeless: Clearly, I've been so far the only person taking care of these test tools (testr, subunit, etc.), and right now, James Page and Corey Bryan are working together with me on these, so we're 3 instead of just 1...10:27
zigoSo moving to OpenStack PKG group in Alioth is for sure an improvement.10:28
zigoAs for SVN, I just hate it. :(10:28
zigolifeless: There's a lot of progress in the switch to Git though.10:28
*** afazekas has quit IRC10:29
rcarrillocruzare you around10:30
rcarrillocruzhave a couple questions about glean/simple-init10:30
rcarrillocruzping me when around pls10:30
*** maishsk has joined #openstack-infra10:33
*** coolsvap is now known as coolsvap|away10:33
openstackgerritDave Walker proposed openstack-infra/project-config: [WIP]Add check-requirement job to openstack/anchor
zigolifeless: BTW, is there unit tests for testtools? I haven't seen this "documented" in a tox.ini...10:35
*** vikram has joined #openstack-infra10:36
*** binbincong has joined #openstack-infra10:36
*** afazekas has joined #openstack-infra10:40
*** e0ne is now known as e0ne_10:40
*** e0ne_ is now known as e0ne10:42
openstackgerritMerged openstack-infra/project-config: Graduate Linux Bridge job from experimental status
*** yamamoto has quit IRC10:42
*** shardy_ has joined #openstack-infra10:44
openstackgerritSomay Jain proposed openstack-infra/jenkins-job-builder: Improvement in the Credentials Binding Plugin
*** shardy has quit IRC10:45
*** hashar has quit IRC10:46
openstackgerritvikram.choudhary proposed openstack-infra/project-config: Adding networking-onos as a part of neutron project.
*** markvoelker has joined #openstack-infra10:49
*** shardy_ has quit IRC10:49
*** claudiub has quit IRC10:50
*** hamzy has quit IRC10:50
*** shardy has joined #openstack-infra10:50
*** annegentle has joined #openstack-infra10:52
TheJuliarcarrillocruz: I might be able to answer/help with glean/simple-init :)10:53
rcarrillocruzTheJulia: morning!10:53
rcarrillocruzso, as chatted yesterday, i have a non-reproduceable issue10:54
*** markvoelker has quit IRC10:54
rcarrillocruzwhere servers being booted with bifrost/ironic/simple-init10:54
*** fawadkhaliq has quit IRC10:54
rcarrillocruzwill have the eth0 configuration applied set10:54
rcarrillocruzsometimes not10:54
rcarrillocruz/etc/network/interfaces.d/eth0.cfg is always set tho10:54
rcarrillocruzi've looked at the glean code10:54
rcarrillocruzand it just sets ssh/hostname and network interfaces configuration files10:55
TheJuliaOh... its a race then10:55
rcarrillocruzhowever, it doesn't bring up the configuration on the interfaces after putting those files10:55
rcarrillocruzsomething like 'ifup eth0'10:55
rcarrillocruzor similar10:55
TheJuliawhat OS is this?10:55
rcarrillocruzi'm puzzled why 'ifconfig' shows ip address set sometimes10:55
rcarrillocruzand sometimes not10:55
rcarrillocruzcos glean clearly doesn't bring up interfaces10:56
rcarrillocruzit just pushes /etc/network/interfaces files10:56
rcarrillocruzcan you give me a hint?10:56
TheJuliaso glean does not bring up interfaces, but simple-init should be in such a place for the system initialization that it should start up the interface10:56
*** annegentle has quit IRC10:56
TheJuliarcarrillocruz: let me finish a big review I'm in the middle of, ~5 minutes10:56
rcarrillocruzsure, thx10:57
*** cdent has joined #openstack-infra10:58
*** AzherKhan has quit IRC10:59
*** ParsectiX has quit IRC10:59
TheJuliaYeah, looking at that, I'm kind of suspecting that it might be configuring a little too late11:02
*** baoli has joined #openstack-infra11:08
*** ddieterly has joined #openstack-infra11:09
*** baoli has quit IRC11:10
rcarrillocruzinterestingly enough11:10
rcarrillocruzi have one server which got network configured after provisioning11:10
rcarrillocruzand another failed11:10
rcarrillocruzand look at the differences on dmesg:11:11
*** baoli has joined #openstack-infra11:11
rcarrillocruzdmesg |grep simple-init|grep eth011:12
rcarrillocruz(no results on ok server)11:12
rcarrillocruzdmesg|grep simple-init|grep eth011:12
rcarrillocruzmain process terminated with status 111:12
rcarrillocruzso, simple-init failed on eth0 on the server that i don't have ip configured for eth011:12
*** hashar has joined #openstack-infra11:12
rcarrillocruzsomething's fishy with simple-init definitely11:12
TheJuliainteresting... what happens if you just run glean from the command line on the econd server?11:13
*** afazekas has quit IRC11:13
*** ddieterly has quit IRC11:13
openstackgerritLucas Alvares Gomes proposed openstack-infra/project-config: Make the tempest-dsvm-ironic-pxe_ipa job voting
rcarrillocruzruns ok, doesn't give any output back11:15
rcarrillocruzi traced glean on the failing server11:15
rcarrillocruzand it works ok11:15
rcarrillocruzit writes the network config etc11:15
rcarrillocruzbut as you said, glean just sets up the network config files11:15
rcarrillocruzwon't configure the ip for the interface on its own11:16
rcarrillocruzif i reboot the server, then i have everything set up, cos the simple-init seems to run correctly after the network-interface script kicks in11:16
rcarrillocruzso it's a race condition11:16
rcarrillocruzi was thinking of putting a patch to bring up interfaces on glean, but not sure if the scope is right11:17
*** fhubik_afk has joined #openstack-infra11:17
*** fhubik_afk is now known as fhubik_11:17
*** hashar has quit IRC11:17
*** _kiran_ has joined #openstack-infra11:17
*** kiran-r has quit IRC11:17
*** fhubik has quit IRC11:18
*** shardy_ has joined #openstack-infra11:19
TheJuliarcarrillocruz: I've pondered that myself, although I also wonder if it is not the right thing for glean to do, espesically with the way it can be executed on a per-interface level11:20
*** shardy has quit IRC11:21
*** ParsectiX has joined #openstack-infra11:23
clarkbglean basically does metadata -> config. it doesnt do services or device setup I dont rrally see that as gleans responsibility11:24
*** shardy_ has quit IRC11:25
*** shardy has joined #openstack-infra11:25
*** yamamoto has joined #openstack-infra11:27
*** _kiran_ has quit IRC11:27
*** fhubik_ is now known as fhubik_afk11:27
*** ParsectiX has quit IRC11:28
*** ParsectiX has joined #openstack-infra11:29
*** yamamoto has quit IRC11:29
*** Somay has joined #openstack-infra11:32
*** fawadkhaliq has joined #openstack-infra11:32
*** e0ne is now known as e0ne_11:34
openstackgerritSergey Vilgelm proposed openstack/requirements: Bump 1.8.0 to oslo.util
*** e0ne_ is now known as e0ne11:35
*** _nadya_ has quit IRC11:36
openstackgerritSergey Vilgelm proposed openstack/requirements: Bump 1.8.0 to oslo.utils
*** e0ne has quit IRC11:38
sdagueclarkb: man, what is your sleep schedule now?11:39
*** ldnunes has joined #openstack-infra11:40
*** Longgeek has quit IRC11:41
*** Longgeek has joined #openstack-infra11:42
*** doug-fish has joined #openstack-infra11:42
*** Longgeek has quit IRC11:42
*** Longgeek has joined #openstack-infra11:42
clarkbthere is no schedule :) >1 baby is basically the end to all scheduling11:42
*** Longgeek has quit IRC11:43
clarkbso nights have been worst but I have to be awake during day to do things otherwise I would have completely inverted my sleep schedule11:43
clarkbthough yesterday was pretty inverted11:45
sdagueso... is my proposed way to get to grenade multinode with d-g. It feels pretty isolated. I know it's not the approach you wanted, but I'd like to see what you think about it.11:45
*** dims has joined #openstack-infra11:47
*** zul has quit IRC11:48
*** dims_ has joined #openstack-infra11:48
*** dims__ has joined #openstack-infra11:49
*** thorst has joined #openstack-infra11:50
*** markvoelker has joined #openstack-infra11:50
clarkbsdague wfm if thats how you want grenade to do it, I might split that new function in two though, first half to do the localrc stuff and second half for networking11:51
*** asselin_ has quit IRC11:51
*** asselin_ has joined #openstack-infra11:52
*** claudiub has joined #openstack-infra11:52
*** dims has quit IRC11:52
sdagueclarkb: sure, that's fair. Honestly, I kind of need - landed first so I can see how any of this works, and if I forgot something11:52
*** dims_ has quit IRC11:53
sdagueI expect much like the multinode job work, using experimental queues for a bit is going to be pretty important to see what hasn't yet been considered11:53
*** _nadya_ has joined #openstack-infra11:53
clarkbsdague maybe start with neutron since thats where you want to do partial stuff?11:54
*** markvoelker has quit IRC11:54
*** dims__ has quit IRC11:54
clarkbor is nvoa net just easier to bootstrap?11:54
*** hashar has joined #openstack-infra11:56
*** amotoki has quit IRC11:57
KiallAnyone about able to +2/+A a requirements change? - adding dnspython311:57
*** salv-orlando has quit IRC11:58
*** fhubik_afk is now known as fhubik_11:58
*** dprince has joined #openstack-infra11:59
clarkbshouldnt you also be updating upper constraints? I suppose it will evebtually happen for you but during the interim that is an unconstrained dep12:00
*** john-davidge has left #openstack-infra12:00
Kiallclarkb: Not quite sure what you mean? If i understand correctly, no no upper should be set (at least not as part of this change). DNSPython plans to continue having 2 different packages on pypi for py2 vs py312:01
clarkbKiall see upper-constraints.txt we effectively pin everything for gate testing12:02
clarkbso by not pinning dnspython3 any jobs using it can break12:02
KiallOh - That's new, never seen it before.12:02
KiallI think I see what needs doing there.. I think12:03
*** fhubik_ has quit IRC12:04
*** fhubik_ has joined #openstack-infra12:05
*** prad has joined #openstack-infra12:06
*** gordc has joined #openstack-infra12:06
openstackgerritKiall Mac Innes proposed openstack/requirements: Add dnspython3 to global-requirements.txt
Kiallclarkb: is ^ what you meant? or am I wayy off base and need to go read up on what that file does ;)12:08
AJaegerclarkb, Kiall: I thought upper-constraints is autogenerated, so no need to touch it manually. Ask lifeless and let's check his emails12:09
*** ddieterly has joined #openstack-infra12:09
KiallIt looks like it's auto-updated, but it *seems* like adding a new entry for new requirements makes sense12:09
AJaegerKiall: I was wrong ;( Let me find a link quickly...12:10
sdagueclarkb: nova is easier to bootstrap, and we already have a nova partial job that we can replicate here12:10
KiallOtherwise, it's missing an entry for the new one - and the edited one (addion of a env marker for dnspython) makes it inconsistent .. So... Humm!12:10
*** arxcruz has quit IRC12:10
sdaguethe multinode work has been running on nova longer, so I'm a little more confident that all the bugs will be of my making12:10
AJaegerlifeless: could you review, please?12:11
sdagueAJaeger: it might be nice if upper-contraints.txt had a comment at the top that it was fully autogenerated, don't touch12:11
*** afazekas has joined #openstack-infra12:11
KiallAJaeger: ah, so updating the file was correct. clarkb could you re-review the updated review? thanks :)12:11
Kiallsdague: according to that email, you are meant to touch it12:12
AJaegersdague: Looking at the referenced mail from lifeless, it should be updated.12:12
Kiall"* If you are adding a new requirement you should also add it to upper-constraints.txt with an exact pin."12:12
sdagueoh, really?12:12
AJaegerbut it's autogenerated as well ;/12:12
sdagueok, that's confusing  :)12:12
AJaegersdague: Indeed, very confusing.12:12
*** bswartz has quit IRC12:12
*** markvoelker has joined #openstack-infra12:13
*** ddieterly has quit IRC12:14
*** arxcruz has joined #openstack-infra12:14
*** shardy_ has joined #openstack-infra12:14
*** kiran-r has joined #openstack-infra12:14
*** ParsectiX has quit IRC12:14
*** arxcruz has quit IRC12:15
openstackgerritDarragh Bailey proposed stackforge/python-jenkins: Remove unneeded wrapping of imports
openstackgerritDarragh Bailey proposed stackforge/python-jenkins: Remove duplicate unittest2 for python 2.6 test env
*** arxcruz has joined #openstack-infra12:15
*** shardy has quit IRC12:16
openstackgerritDarragh Bailey proposed stackforge/python-jenkins: Add socket timeout tests and helpers
openstackgerritSomay Jain proposed openstack-infra/jenkins-job-builder: Support for more conditions type in conditions-buildstep plugin
*** masayukig_ has quit IRC12:16
*** masayukig_ has joined #openstack-infra12:17
*** yamamoto has joined #openstack-infra12:17
*** shardy_ has quit IRC12:20
sdagueKiall: while you happen to be around, does the designate team have an install guide chapter somewhere? Like what's in the rest of the install guide for older openstack projects12:20
*** shardy has joined #openstack-infra12:20
*** salv-orlando has joined #openstack-infra12:21
*** yamamoto has quit IRC12:22
Kiallsdague: - it's likely a tad out of date though :(12:22
*** kgiusti has joined #openstack-infra12:22
sdagueKiall: ok, yeh, that's quite a bit different than for the other projects, which assume package installation12:23
KiallYou looking for devstack addition?
Kiall(converting to a real devstack plugin is on the 3 km long TODO list)12:24
sdagueKiall: no, actually, I've been building a manually installed environment from docs, mostly to experience what we ask of our users12:24
*** Longgeek has joined #openstack-infra12:25
*** abregman has quit IRC12:25
*** jasondotstar has quit IRC12:25
Kialli.e. from git? The install guides do that rather than Ubuntu packages for designate -
Kiall(there's deb installs for powerdns/bind/mysql etc - but not for Designate itself)12:26
*** _nadya_ has quit IRC12:26
*** _nadya_ has joined #openstack-infra12:27
KiallAll that said - I've not walked through that guide is like a year.. I probably should go do that again soon12:27
Kiallin like*12:27
*** annegentle has joined #openstack-infra12:30
*** zul has joined #openstack-infra12:30
*** chlong has joined #openstack-infra12:31
*** afazekas has quit IRC12:31
*** _nadya_ has quit IRC12:31
*** binbincong has quit IRC12:32
openstackgerritSomay Jain proposed openstack-infra/jenkins-job-builder: Support for health-report-amplification-factor in junit plugin
*** amrith is now known as _amrith_12:33
openstackgerritSomay Jain proposed openstack-infra/jenkins-job-builder: Support for health-scale-factor in junit plugin
*** yamamoto has joined #openstack-infra12:36
*** abregman has joined #openstack-infra12:37
*** yamamoto has quit IRC12:38
*** aysyd has joined #openstack-infra12:38
*** erlon has joined #openstack-infra12:39
*** jlanoux_ has joined #openstack-infra12:40
*** thorst has quit IRC12:42
*** afazekas has joined #openstack-infra12:42
sdagueKiall: there are deb packages in Ubuntu for kilo12:42
freerunnerlifeless: Hi! Are u here?12:42
Kiallsdague: Yes, there are.. But that guide doesn't use them..12:42
*** jlanoux has quit IRC12:43
sdagueKiall: right, I know. But the install guide for all the other projects do12:43
*** ParsectiX has joined #openstack-infra12:43
KiallOkay - Now I get what you want ;)12:43
KiallI'll find some time tomorrow to write that up :)12:43
sdagueKiall: if you start here
sdaguethen go to the install guide for your platform12:44
freerunnerlifeless: Is there a way to trigger Proposal Bot for upper-constraints update?12:44
sdagueyou get something like -
*** binbincong has joined #openstack-infra12:45
*** tmorin has joined #openstack-infra12:46
AJaegerfreerunner: it is run once a day12:46
*** tmorin has quit IRC12:46
openstackgerritArnold Bechtoldt proposed openstack-infra/jenkins-job-builder: support com.cloudbees.plugins.credentials properties in authorization matrix
openstackgerritMerged openstack-infra/shade: Add CRUD methods for Keystone domains
*** tmorin has joined #openstack-infra12:47
Kiallsdague: Yep, I'll write up content following that pattern tomorrow at some point - will ping you :)12:47
openstackgerritMerged openstack-infra/shade: Support project/tenant and domain vs. None
sdagueKiall: cool12:48
fungiAJaeger: freerunner: looks like the most recent update was merged just a couple hours ago
*** ninag has joined #openstack-infra12:49
AJaegerfreerunner: lifeless is resting. Why do you want to trigger it?12:50
*** dims has joined #openstack-infra12:50
AJaegerfungi, yes, indeed - lifeless was asking kindly to get this in earlier today12:50
sdagueKiall: so I have an entirely odd question, for home network setup, is it possible to have designate generate dynamic dns update requests to a dns server that it doesn't directly control?12:51
fungifreerunner: i wonder if you're under the misconception that constraints files will be proposed to every repo rather than simply being hosted in the requirements repo?12:51
freerunnerAJaeger: fungi: We just released a new version of yaql, which have compatibility with ply==3.6, and we need to change version of yaql from 0.2.6 o 0.2.7 in upper-constraints to unblock our dsvm-gate...12:51
fungifreerunner: okay, so has global-requirements.txt in openstack/requirements been updated to cap or blacklist ply so that version won't be selected?12:52
*** coolsvap|away is now known as coolsvap12:52
AJaegerfreerunner: looking at, you need to send a review request for requirements repo and update both global-requiremetns and upper-constraints12:53
fungioh, yep, i read your problem statement backwards12:53
Kiallsdague: lol - that's an odd request ;) Dynamic DNS updates? no.. You could have the 3rd party slave from us.. but that's as close as it gets12:53
*** ZZelle_ has joined #openstack-infra12:53
AJaegerfreerunner: I see - global-requirements has "yaql>=0.2.6,!=0.3.0  # Apache 2.0 License", so it's only upper-constraints.12:54
KiallDesignate really isn't built for the home-network use case :P12:54
fungifreerunner: you need to raise the yaql cap in global-requirements.txt and optionally also in upper-constraints.txt (or wait for it to be auto-updated the day after your global-reqs change lands)12:54
AJaegerfreerunner: Looks like you best wait for the Australian sun to wake up lifeless ;(12:54
*** dims has quit IRC12:54
fungifreerunner: assuming there's any cap of yaql in global requirements currently12:55
*** ZZelle_ is now known as ZZelle12:55
AJaegerfungi see above, there's no cap12:55
*** afazekas has quit IRC12:55
fungionce i'm at a computer where i can check easily, i'll refine my suggestion12:55
*** rlandy has joined #openstack-infra12:55
mordredSpamapS, Shrews: I learned two things yesterday we should reflect in shade12:55
*** tmorin has quit IRC12:55
sdagueKiall: gotcha, well, it was a thought. Would be nice to propogate up dns updates to existing dns server here. But such it is.12:56
fungiAJaeger: freerunner: got it, so this is entirely about upper-constraints.txt noticing new releases showing up on pypi12:56
mordredShrews: one - /v3/auth/domains still doesn't return "domains scoped to this token" - because that's apparently not a thing - it returns domains the user has a role in12:56
*** dboik has joined #openstack-infra12:56
mordredShrews: which means that, for instance in devstack, it's the empty list for both admin and demo12:56
*** tmorin has joined #openstack-infra12:56
mordredShrews: the workaround12:56
mordredShrews: is that we can get /v3/auth/projects and pull the list of domain ids from that - and we will get the list of relevant domains for the user12:57
fungiAJaeger: freerunner: agreed, i have no idea if lifeless had any ideas around shortening the "we released a new version to solve a problem, why do we have to wait a day to use it in devstack" dilemma12:57
*** doug-fish has quit IRC12:57
*** thorst has joined #openstack-infra12:57
mordredShrews: NOW ... this leads to the question "why are we listing domains at all?" - and I think the answer is "to support figuring out if a domain exists before creating it"12:57
*** nilasae has quit IRC12:57
*** tmorin has quit IRC12:57
*** doug-fish has joined #openstack-infra12:57
*** mnaser has quit IRC12:58
freerunnerfungi: AJaeger: I think I can bump yaql version to >=0.2.7, !=0.3.0. And fix constraints as well. This should fix the problem ;)12:58
openstackgerritThomas Morin proposed openstack-infra/project-config: Create branch permission for networking-bagpipe-l2
mordredShrews: which is more relevant for admin users anyway12:58
Shrewsmordred: is there a "duplicate" exception thrown that we can catch instead at creation time?12:58
*** fhubik_ is now known as fhubik_afk12:58
mordredthat's where I was going12:59
*** tmorin has joined #openstack-infra12:59
Shrewsor does that not help us12:59
fungiAJaeger: freerunner: though proposing a change to the upper constraints file and convince requirements reviewers to approve it is probably the fastest route12:59
*** eharney has joined #openstack-infra12:59
mordredI'm thinking we should investigate just trying to create teh domain when someone asks for it12:59
*** msuriar has quit IRC12:59
*** dguitarbite has quit IRC12:59
*** AJaeger has quit IRC12:59
*** tmorin has quit IRC12:59
*** ZZelle has quit IRC13:00
*** bradjones has quit IRC13:00
*** rcarrillocruz has quit IRC13:00
mordredShrews: and/or - I believe we can also do "domains.get(domain_id"13:00
mordredShrews: at the keystoneclient level13:00
*** ninag has quit IRC13:00
*** mfisch has quit IRC13:00
*** mrodden has quit IRC13:00
mordredShrews: which might let us avoid this altogether13:00
*** jsavak has joined #openstack-infra13:00
*** ninag has joined #openstack-infra13:00
*** msuriar has joined #openstack-infra13:00
*** bradjones has joined #openstack-infra13:00
*** bradjones has quit IRC13:00
*** bradjones has joined #openstack-infra13:00
*** dguitarbite has joined #openstack-infra13:00
mordredShrews: since a user is going to have one and only one domain13:01
*** tmorin has joined #openstack-infra13:01
*** AJaeger has joined #openstack-infra13:01
*** AJaeger has joined #openstack-infra13:01
*** julim has joined #openstack-infra13:01
*** dboik has quit IRC13:01
mordredShrews: and for an admin, "list domains" is apparently something they never want to do because it's super expensive13:01
*** fawadkhaliq has quit IRC13:01
mordredor something13:01
*** mrodden has joined #openstack-infra13:01
*** mpaolino has joined #openstack-infra13:01
mordredor rather - if they want to - it's a thing they want to very explicitly request - they don't want something listing domains helpfully for them to do something else13:01
*** mnaser has joined #openstack-infra13:01
*** tmorin has quit IRC13:01
*** ZZelle has joined #openstack-infra13:02
AJaegerfreerunner: looking at the referenced email, it says " New releases of *any* library will not be used by devstack-gate jobs until upper-constraints.txt has been edited to enable them."13:02
AJaegerSo, how comes that your dsvm job is failing? And which one exactly?13:02
*** ricky1 has joined #openstack-infra13:02
*** hamzy has joined #openstack-infra13:02
AJaegerfreerunner: But go with fungi's suggestion and update upper-constraints.txt manually and ask for quick reviews.13:02
zigolifeless: Setting up python-linecache2 (1.0.0-1) ...13:03
zigo  File "/usr/lib/python2.7/dist-packages/linecache2/tests/", line 10213:03
zigo    def keyworded(*arg1, arg2=1):13:03
zigo                            ^13:03
zigoSyntaxError: invalid syntax13:03
Shrewsmordred: So, unless infra has a need to do such a thing, I really don't see admin types using shade for such adminy things. They'd likely use the official clients, no?13:03
EmilienMmordred: hey, just wanted to introduce you to mhu, the author of
mordredShrews: ansible13:04
Shrewsmordred: ah13:04
openstackgerritVictor Ryzhenkin proposed openstack/requirements: Bump YAQL version to 0.2.7
EmilienMmordred: I remember you were interested by a plugin, openid was it?13:04
mordredmhu: hi! so ... I think we want to use cauth in infra13:04
*** tmorin has joined #openstack-infra13:04
*** tmorin has quit IRC13:04
*** jasondotstar has joined #openstack-infra13:05
AJaegerfreerunner: please add a proper comment explaining why this is needed.13:05
mordredmhu: starting as a thing we want to run in front of phabriactor to help us do SSO with launchpad13:05
*** peristeri has joined #openstack-infra13:05
freerunnerAJaeger: Sure. One moment.13:05
mordredmhu, EmilienM: we have pecan/wsme based openid code in openstack/storyboard that could be leveraged I'm sure13:05
mhumordred: hi ! good timing, we're switching to a plugin architecture for auth methods and supported services13:05
Shrewsmordred: i can't immediately anticipate any problems with just catching the "already exists" exception and returning the existing domain13:06
mhumordred, great, an openid plugin is something I want to implement during our next sprint (starting tomorrow)13:06
*** tmorin has joined #openstack-infra13:06
*** mestery_ has joined #openstack-infra13:06
*** masayukig_ has quit IRC13:06
mordredmhu: so - please feel free to use as much of the code from storyboard as you like13:06
* EmilienM connects people13:07
mhumordred, will do !13:07
mhuEmilienM, you're like vodka, or Nokia - can't remember which is which13:07
mordredmhu: the one feature we always need with openid plugins (which the storyboard code supports) is the ability to configure it to be SSO to a pre-configured openid url13:07
mordredI like vodka13:07
*** xyang1 has joined #openstack-infra13:07
mhumordred, I'll check how it's done in storyboard13:08
openstackgerritVictor Ryzhenkin proposed openstack/requirements: Bump YAQL version to 0.2.7
*** tlian has joined #openstack-infra13:09
*** mestery has quit IRC13:09
freerunnerAJaeger: done13:09
*** tmorin has quit IRC13:09
*** shashank_hegde has joined #openstack-infra13:09
mordredmhu: let me know where to send bribes ... I mean presents ...13:11
*** tmorin has joined #openstack-infra13:11
*** tmorin has quit IRC13:11
*** dustins has joined #openstack-infra13:12
mhumordred: and by presents you mean vodka ?13:12
mordredmhu: :)13:12
*** tmorin has joined #openstack-infra13:13
*** masayukig_ has joined #openstack-infra13:13
Shrewsmordred: so, the domains stuff that just merged.... broken?13:14
*** amotoki has joined #openstack-infra13:14
*** deepakcs has quit IRC13:14
mordredShrews: not broken - just not as optimal as it could be - patch coming13:14
*** miqui has joined #openstack-infra13:15
*** afazekas has joined #openstack-infra13:15
mordredShrews: all of the things that merged will do the things they say they will do... just maybe not at scale13:15
*** tlian has quit IRC13:15
*** asselin__ has joined #openstack-infra13:16
*** shashank_hegde has quit IRC13:17
Shrewsmordred: plz to make them webscale13:17
openstackgerritSomay Jain proposed openstack-infra/jenkins-job-builder: Adding support for javadoc plugin
mordredShrews: it's going to make the domain interface inconsistent with the others - but it's basically an admin interface, so I think it's ok13:18
KiallRandom Q - What's the current story with pep8/flake8/hacking version bumps? I've not noticed one come along in quite some time..13:18
*** tmorin has quit IRC13:19
*** asselin_ has quit IRC13:19
*** changbl has joined #openstack-infra13:19
*** bknudson has joined #openstack-infra13:19
*** ayoung has joined #openstack-infra13:20
*** puranamr has joined #openstack-infra13:22
*** dims has joined #openstack-infra13:22
yolandamordred, can i get some help? i'd like to have a new version of gerritlib cut, i have some changes for manage-projects and description, blocked for months and months13:23
openstackgerritArnold Bechtoldt proposed openstack-infra/jenkins-job-builder: support com.cloudbees.plugins.credentials properties in authorization matrix
mordredyolanda: any outstanding changes? or just need a new release?13:24
*** ari has joined #openstack-infra13:24
yolandanew release, i need a change i pushed last year, to manage project descriptions13:24
*** mriedem_away is now known as mriedem13:24
*** shardy_ has joined #openstack-infra13:25
Shrewsmordred: that should be fine13:25
*** tlian has joined #openstack-infra13:25
mordredfungi: ^^ any concerns with me cutting a new gerritlib?13:25
*** shardy has quit IRC13:26
fungimordred: none. i saw yolanda asking for one yesterday (and weeks prior) but have unfortunately not gotten around to it13:27
*** hdd has joined #openstack-infra13:28
openstackgerritMonty Taylor proposed openstack-infra/gerritlib: Bump pbr version
yolandayep, i asked several times, i had some change to add project descriptions into gerrit, and wanted to move that forward13:28
mordredfungi: we shoudl merge that first ^^13:28
mordredfungi: because gerritlib is currently pinning to an old pbr13:28
*** yamamoto has joined #openstack-infra13:28
*** yamamoto has quit IRC13:29
*** woodster_ has joined #openstack-infra13:29
*** yamamoto has joined #openstack-infra13:29
mordredfungi: also, while we're at it:
mordredit's a docs change13:30
*** shardy_ has quit IRC13:30
mordredgah. and it's going to merge conflict on pbr version ... :)13:30
* zigo uploaded linecache2 & traceback2 to Sid.13:30
zigoI'll proceed with updating testtools to 1.8.013:30
yolandamordred, fungi, if there is anything i can help to sort that, i'm glad to collaborate13:31
*** akshai has joined #openstack-infra13:31
*** shardy has joined #openstack-infra13:31
mordredyolanda: could you rebase on top of real quick (keep the pbr versions from 199568) ?13:31
*** masayukig_ has quit IRC13:32
fungimordred: on a similar note, will get good sphinxified docs for bindep, and the next point release should put them on the docs site (i have a separate wip change to add it to our infra docs index)13:32
mordredyolanda: (while you're in that patch - also please delete that Makefile)13:32
*** whitenoise has quit IRC13:33
*** salv-orlando has quit IRC13:33
*** puranamr has quit IRC13:34
yolandamm, fails with pep813:34
*** eharney has quit IRC13:34
mordredyolanda: ok. if it needs work, let's cut your release and come back to it13:35
*** tmorin has joined #openstack-infra13:35
jklarehi, anybody around who knows how the infra jenkins instances where set up?13:36
yolandamordred, thx13:36
mordredjklare: many people13:36
fungisdague: your tags e-mail reminds me that i (or someone in the vmt at any rate) should propose a tag or tag set to replace
*** whitenoise has joined #openstack-infra13:36
*** whitenoise has joined #openstack-infra13:36
jklarei am wondering, since the manifests tell that the jenkins is installed from package, from the official repo as far as i can tell13:36
mordredjklare: yes, that's right. we use packages from jenkins-ci.org13:37
sdaguefungi: cool, I was hoping it would kick off some thoughts13:37
jklarethe currently used version on the infra servers is 1.565.313:37
*** radez_g0n3 is now known as radez13:37
*** bswartz has joined #openstack-infra13:37
jklarewhich is not included in the repo anymore13:37
mordredjklare: we try to never update them because every jenkins release so far has broken something13:37
jklareas far as i can tell13:37
jklareso they were set up a long time ago and never updated13:37
mordredjklare: and we have not made a new jenkins master in quite some time13:37
fungijklare: right, that puppet manifest just installs whatever the latest version is, doesn't update13:37
jklarei see13:37
mordredjklare: that's right. if a security update comes out, we'll make a plan to update13:37
*** tmorin has quit IRC13:38
mordredalthough we're hoping we'll be off of jenkins before we have to deal with the pain of another security update13:38
sdaguefungi / mordred would one of you be willing to approve - to get some experimental jobs on this grenade partial idea13:38
jklarei guess we will also go with this version then13:38
mordredjklare: sure thing!13:38
*** masayukig_ has joined #openstack-infra13:38
fungijklare: looks like we upgraded our jenkins masters last in october. so not _that_ long ago depending on how you calculate time13:38
*** tmorin has joined #openstack-infra13:38
jklarefungi october is fine i guess, the issue for me is that you could not set up a new one with the current manifests13:39
jklarefungi which is a little broken13:39
fungijklare: yes, i agree that's unfortunate. i wonder why they delete releases13:39
fungithat seems anti-social13:40
mordredyolanda: you just made a new jenkins master in gozer ... what did you do?13:40
fungii mean, we don't ever delete any release we publish for our software13:40
*** flepied has quit IRC13:40
jklarefungi i also asked myself the same question, especially since they are still hosting the debs on their archive site13:40
jklarefungi so why not put them in the repo...13:40
yolandamordred, well, Tim was the one doing that. It was a lot of manual effort13:40
fungisdague: on a related note, input on this experimental job change would be appreciated
mordredyolanda: yah - it's not very automated13:40
*** tmorin has quit IRC13:40
yolandaas jenkins versions are not correctly installed by manifest, and plugins also are not automated13:40
mordredyolanda: I look forward to the day where we no longer have to deal with that :)13:41
*** tmorin has joined #openstack-infra13:41
yolandamordred, yes, that jenkins manifest needs a bit of love. Even security is not automated, every config step is manual13:41
fricklerso you intend to replace jenkins? with what?13:41
yolandamordred, is there any expected ETA for jenkins drop? i think it will be a long way, right?13:42
fungifrickler: we've been gradually replacing jenkins for 2+ years13:42
fungifrickler: (zuul, nodepool, gearman, et cetera)13:42
mordredfrickler: ansible.
mordredfrickler: jenkins currently performs one task for us - remote shell execution13:42
fungirelatively poorly and inefficiently i might add13:43
mordredit's a bit heavy weight for just that :)13:43
mordredit breaks all the time13:43
* fungi had to restart 4 jenkins masters this week alone13:43
yolandawe don't normally need restarts, but an issue we are having, is that it drops connections with slaves quite frequently13:43
*** sdake has joined #openstack-infra13:44
fungimaybe the thread leaks have been fixed in newer jenkins, but the fact that nobody has touched clarkb's bug report on it in months after they requested a thread dump speaks volumes13:44
yolandaeven now, with 2 jenkins, and less load, it keeps happening13:44
mordredyolanda: I am hoping we can get zuulv3 finished before tokyo13:44
yolandamordred, really?! wow13:44
yolandathat spec is huge13:44
fungisdague: lgtm, though mordred has already approved13:45
mordredsure. but it should not take ages to implement, largely because we're not going to try to support both systems in parallel (which is what usually makes things take forever)13:45
*** annegentle has quit IRC13:45
*** masayukig_ has quit IRC13:45
*** tmorin has quit IRC13:45
* yolanda is afraid of the backport to our systems13:45
mordredyolanda: I mean, I could very well be wrong - but it's what I hope to achieve - because the current jenkins thing is very crap13:45
openstackgerritMerged openstack-infra/project-config: add experimental grenade multinode jobs
*** annegentle has joined #openstack-infra13:45
mordredyolanda: biggest issue will be anyone on your side who has successfully managed to get an additional jenkins plugin added13:46
*** tmorin has joined #openstack-infra13:46
*** fawadkhaliq has joined #openstack-infra13:46
mordredyolanda: we'll hopefully have a script to translate jjb templates13:46
yolandamordred, we have several plugins that are not used upstream, yes13:46
*** tmorin has quit IRC13:46
mordredyolanda: so you'll need to top using those plugins - and you'll need to get log publishing to swift rolled out13:46
yolandamordred, i wanted to propose that swift stuff for next sprint13:47
yolandanow i have more reasons for that :)13:47
mordredyolanda: OR - you'll need to figure out ansible equivs for the plugins they're using - so that your jjb template translation can migrate from those plugins then13:47
*** tmorin has joined #openstack-infra13:47
yolandayou'll have it ready for Tokyo, we can have it ported for Barcelona :)13:48
*** mestery_ has quit IRC13:48
*** hdd has quit IRC13:48
*** sdake has quit IRC13:49
*** Somay has quit IRC13:49
*** fawadkhaliq has quit IRC13:49
*** sdake has joined #openstack-infra13:49
*** tmorin has left #openstack-infra13:51
openstackgerritMerged openstack-infra/gerritlib: Update project docs
*** masayukig_ has joined #openstack-infra13:52
openstackgerritDarragh Bailey proposed openstack-infra/jenkins-job-builder: Removed unnecessary parameter from delete_old_jobs
openstackgerritDarragh Bailey proposed openstack-infra/jenkins-job-builder: Added parallelization options
*** _nadya_ has joined #openstack-infra13:57
*** armax has joined #openstack-infra13:57
openstackgerritThomas Morin proposed openstack/requirements: Add networking-bagpipe-l2 to projects.txt
*** asilenkov__ has joined #openstack-infra13:59
*** dustins has quit IRC13:59
asilenkov__Hello! I have a question about gate-fuel-specs-python2713:59
openstackgerritJeremy Stanley proposed openstack-infra/system-config: Add a server
asilenkov__ - error inside14:00
openstackgerritMonty Taylor proposed openstack-infra/project-config: Use zuul-cloner for bifrost jobs
*** ajo|QoSPOC is now known as ajo14:01
asilenkov__http:get:: /dump/(snapshot_name)14:01
asilenkov__it seems that is missing in test env.14:01
*** jlanoux has joined #openstack-infra14:01
*** dulek has joined #openstack-infra14:01
*** dulek has left #openstack-infra14:01
asilenkov__Should I fill the bug for that ?14:01
fungiasilenkov__: is it in the requirements or test-requirements list for that project?14:01
*** eharney has joined #openstack-infra14:02
fungiasilenkov__: your repo should declare the python packages it needs to install and perform its tests14:02
*** annegentle has quit IRC14:02
asilenkov__fuel-specs/requirements.txt contains sphinxcontrib-httpdomain14:02
*** dboik has joined #openstack-infra14:02
*** dustins has joined #openstack-infra14:03
fungii see that14:03
openstackgerritMonty Taylor proposed openstack-infra/project-config: Use zuul-cloner for bifrost jobs
*** btully has joined #openstack-infra14:03
asilenkov__job is green and this error seems non destructive )14:03
fungi"py27 installed: ...,sphinxcontrib-httpdomain==1.3.0,..."14:04
*** ParsectiX has quit IRC14:04
fungiaccording to the job log14:04
*** jlanoux_ has quit IRC14:04
pabelangerAny core my a +1 workflow?  It will help me troubleshoot some puppet changes for the puppet-apply job14:04
*** mtanino has joined #openstack-infra14:04
asilenkov__yep (( can't understand the reason.14:04
*** baoli has quit IRC14:04
fungiasilenkov__: if you 'tox -e py27' locally do you get the same error?14:04
AJaegerasilenkov__: "sphinxcontrib-httpdomain==1.3.0" is installed14:05
pabelangerman, I buggered that request14:05
pabelangerneed coffee14:05
mordredpabelanger: done14:05
*** baoli has joined #openstack-infra14:05
mordredfungi: double check me on ?14:05
pabelangermordred, thank you kind sir14:05
*** dboik_ has joined #openstack-infra14:05
fungipabelanger: lgtm14:05
fungioh, mordred approved too14:06
pabelangermore the better :)14:06
*** jsavak has quit IRC14:07
*** _amrith_ is now known as amrith14:07
*** masayukig_ has quit IRC14:07
*** dboik has quit IRC14:07
*** bhunter71 has joined #openstack-infra14:07
asilenkov__^^fungi locally it's ok )14:08
*** kiran-r has quit IRC14:08
eantyshevHi! Anyone who has core rights in zuul, please, review
fungiasilenkov__: how about `git clean -dfx; tox -e py27`? asilenkov__ also what version of sphinxcontrib-httpdomain does it tell you is being installed when you run it locally?14:09
*** f1ller is now known as filler14:10
asilenkov__^^fungi yes -dfx shows error <string>:52: (ERROR/3) Unknown directive type "http:get"14:11
KiallAnyone happen to know if Florian Weimer of RedHat is on IRC / what his nick is?14:11
*** baoli has quit IRC14:12
*** fhubik_afk is now known as fhubik_14:12
fungiasilenkov__: okay, so you probably had a cached venv with an older sphinxcontrib-httpdomain is my guess14:12
*** jsavak has joined #openstack-infra14:12
*** baoli has joined #openstack-infra14:12
fungiasilenkov__: which would be odd if true, because 1.3.0 is a year old this month14:13
*** masayukig_ has joined #openstack-infra14:13
asilenkov__yep sphinxcontrib-httpdomain==1.3.014:13
openstackgerritMerged openstack-infra/project-config: Upload applytest folder gate-infra-puppet-apply-*
fungiasilenkov__: was this job previously passing for fuel-specs, or has it only recently been running14:13
*** fhubik_ is now known as fhubik14:13
*** fhubik has quit IRC14:14
asilenkov__this job is passing but error still in logs (14:14
fungiasilenkov__: oh, got it14:14
*** ildikov has quit IRC14:14
*** fhubik has joined #openstack-infra14:15
AJaegerasilenkov__: You're running the test suite with tox -e py27, and use docutils, not sphinx14:15
fungiasilenkov__: aha, that dependency is only a week old14:15
*** zz_dimtruck is now known as dimtruck14:16
AJaegerasilenkov__: could that be the problem?14:16
asilenkov__not sure )14:17
fungiasilenkov__: so anyway, looks like that error first appears on the test runs for the change which added it last week
*** esker has joined #openstack-infra14:17
fungiasilenkov__: e.g. it's been there "all along" (at least for as long as sphinxcontrib-httpdomain has been used on that repo)14:18
freerunnerAJaeger: Gate jobs passed! Can u review this?14:18
fungier, s/e.g./i.e./ (darn latin abbreviations)14:18
*** BharatK has quit IRC14:18
*** yamamoto has quit IRC14:18
AJaegerfreerunner: I'm not core on that repo, but happily give you my +1 ;)14:19
*** wenlock has joined #openstack-infra14:19
*** amitgandhinz has joined #openstack-infra14:19
freerunnerAJaeger ;)14:19
AJaegerasilenkov__: the error message comes from .tox/py27/lib/python2.7/site-packages/docutils/parsers/rst/states.py14:20
AJaegerSo, it'S docutils that does not parse the RST extension you use14:20
*** puranamr has joined #openstack-infra14:23
Davieyttx: I used to be a member of requirements-core and it seems at some point i was silently dropped. Do you know when this was?14:26
*** armax has quit IRC14:26
*** annegentle has joined #openstack-infra14:27
*** hdd has joined #openstack-infra14:27
*** dimtruck is now known as zz_dimtruck14:30
*** tonytan4ever has joined #openstack-infra14:30
*** sodo has joined #openstack-infra14:32
openstackgerritVladimir Kozhukalov proposed openstack-infra/project-config: Add project fuel-agent repo
*** armax has joined #openstack-infra14:34
*** ildikov has joined #openstack-infra14:34
jeblairmordred: regarding  -- it looks like that's step 1 to joining bifrost, ironic, and shade into a shared change queue, which would pull infra and openstack together14:35
*** salv-orlando has joined #openstack-infra14:35
openstackgerritRussell Bryant proposed openstack/requirements: Stop adding extra newline after unparseable lines.
*** rossella_s has quit IRC14:37
*** sodo has quit IRC14:37
*** rossella_s has joined #openstack-infra14:37
*** sabeen has quit IRC14:41
*** otter768 has joined #openstack-infra14:42
*** cindyo has joined #openstack-infra14:44
mordredjeblair: yeah - that started as "bifrost has a set of functional tests as a consumer of shade via ansible that might be nice to run" - and then got there. I'm not sure we want to participate in that symmetric gate ... but regardless bifrost should be using zuul-cloner to clone the repos its using and not code in its test script14:45
*** puranamr has quit IRC14:46
jeblairmordred: ++14:46
*** otter768 has quit IRC14:46
*** salv-orlando has quit IRC14:47
*** akraminsky has quit IRC14:47
*** maishsk has quit IRC14:48
*** rfolco has quit IRC14:51
asilenkov__^^fungi ^^AJaegar thank you for pointing this out. I'm not an expert here, so just say if you need a bug or something to fix it14:51
*** rfolco has joined #openstack-infra14:51
fungiasilenkov__: it's up to the fuel-specs developers to fix. it's not an infrastructure problem14:51
*** fawadkhaliq has joined #openstack-infra14:51
asilenkov__alright ty14:52
fungiDaviey: ttx: i've started digging through the mailing list archives for the announcement. i vaguely recall a year or two back there was a flushing of requirements core reviewers with few or no reviews on the repo over some lengthy period of time, but i'm not finding the discussion so i may have imagined it14:53
openstackgerritSergey Vilgelm proposed openstack/requirements: Bump 1.8.0 to oslo.utils
openstackgerritMonty Taylor proposed openstack-infra/shade: Allow use of admin tokens in keystone
openstackgerritMonty Taylor proposed openstack-infra/shade: Rework how we get domains
mordredShrews: ^^ I have tested that manually locally against my devstack and the get and search things work as expected14:54
mordredShrews: I'll follow up with real tests in a sec14:55
ttxmordred: just had an interesting discussion on #phabricator14:55
mordredttx: awesome14:55
ttxI filed our "requests" there14:55
*** austin81 has joined #openstack-infra14:55
ttxgot an openstack "project" craeted there14:55
*** austin81 has left #openstack-infra14:55
*** wenlock_ has joined #openstack-infra14:56
fungithat seems like a promising development14:56
*** dpose has joined #openstack-infra14:56
anteayattx does #phabricator have a url for its irc logs?14:56
*** rkukura has quit IRC14:56
Davieyfungi: Yeah, that would have been pretty accurate.. but I missed the announcement.  I'm not saying it was wrong, i just found it odd that i wasn't informed.14:56
ttxanteaya: not that I can find14:56
anteayattx okay thanks14:56
anteayawasn't questioning your accuracy just thought I might save you some typing14:57
*** yamahata has joined #openstack-infra14:57
ricky1SpamapS: hiya14:57
*** ricky1 is now known as rcarrillocruz14:57
ttxmordred, jeblair: summary is.. they have a tool in the pipe to track products and branches14:58
ttxcaled releeph14:58
ttxcalled even14:58
openstackgerritSergey Vilgelm proposed openstack/requirements: Bump 1.8.0 to oslo.utils
ttxbut it's extremely far from being ready for anyone14:58
jeblairttx: oh, interesting... i'm reading the tasks now14:58
*** _nadya_ has quit IRC14:59
ttxthere are some they already dismissed/merged14:59
ttxsee links on the wiki page14:59
*** _nadya_ has joined #openstack-infra14:59
ttxmordred, jeblair: having this Releeph on the long-term roadmap makes them extremely unlikely to support features that would help us emulate it in Maniphest basically15:00
*** ninag has quit IRC15:00
*** ari has quit IRC15:00
jeblairttx: i'm assuming releeph will integrate with maniphest to do the things you need?15:01
*** ninag has joined #openstack-infra15:01
ttxjeblair: that is epriestley's assumption yes15:01
*** yamahata has quit IRC15:01
ttx<epriestley> On change tracking in general we have some tooling in the pipeline, but it's currently not anywhere near the top of the priority list. T2714 is the top-level task for Releeph, the release-management / where-was-stuff-merged tool.15:01
*** mrunge has quit IRC15:02
*** yamahata has joined #openstack-infra15:02
jeblairttx: part of me says "cool they are working on something to support this case", but part of me is wary, because there are a lot of tools that we either do not want or are not ready to use in phabricator (prime examples code review, git hosting), so we also need to be wary of whether releeph in turn requires something we don't want to use.15:02
ttx"in the long term our intent is to provide purpose-built release tracking tools via Releeph (which has formal objects like "Products" and "Branches" which can, I think, map to things like "Cinder -> Kilo")."15:02
ttxyeah, puts us in a sort of tricky situation15:02
ttxReleeph might well require us to use Diffusion, for example15:03
*** sabeen1 has joined #openstack-infra15:03
fungialso makes me leery of the possibility that releeph might impose some specific release tracking workflows15:03
sc68calanteaya: linux bridge CI job is passing 80% - 27 passes 7 failures.15:03
sc68calanteaya: in the last 48 hours15:03
anteayasc68cal: woot!15:03
anteayawell done!15:04
ttxfungi: I think that's the key issue with it today, it's pretty tied to Facebook's needs and thy need to clean that up15:04
*** dims has quit IRC15:04
*** sabeen3 has joined #openstack-infra15:04
*** coolsvap is now known as coolsvap|away15:04
*** _nadya_ has quit IRC15:04
sc68calanteaya: I haven't checked if the failures are patches that failed other jobs too15:04
anteayado they want to clean that up?15:04
fungittx: got it. so it's less that they're creating from whole cloth, and more that this already is in use in facebook but not suitably generalized15:04
anteayaor is that a feature15:04
anteayasc68cal: yay15:05
ttxOne thing that he suggested was to develop a huge CustomField that would hold that project/branch information, the same way the task list in Launchpad does15:05
*** dims has joined #openstack-infra15:05
ttxand display it in tabular form15:05
jeblairhow would that look when you edited it?15:06
*** arxcruz has quit IRC15:06
ttxCustomFields can override display methods and be displayed however you want15:06
ttxsame for their edit control15:06
sc68calanteaya: at least one - 191707 failed all the tempest jobs15:06
sc68calanteaya: so the patch was probably bad15:07
jeblairoh ok15:07
*** sabeen1 has quit IRC15:07
*** ninag has quit IRC15:07
anteayasc68cal: let's move to -neutron?15:07
ttxjeblair: it's just a pretty massive development I think. We'd store all that data in a single customfield and have nifty display/edit controls/search hooks in it15:07
sc68calanteaya: sure15:08
*** jsavak has quit IRC15:08
ttxa light version might be a textfield that says "nova/kilo: DONE; cinder/kilo: INPROGRESS"15:08
*** ivar-lazzaro has quit IRC15:09
ttxbut that would be ugly and errorprone15:09
jeblairyeah :/15:10
*** zz_dimtruck is now known as dimtruck15:10
*** dims has quit IRC15:10
ttxbut you get the idea15:10
ttxI guess we could spec it and then see if we find someone to get it done, one way or another15:11
openstackgerritMatt Riedemann proposed openstack-infra/elastic-recheck: Fix the module index link in the docs
*** coolsvap|away is now known as coolsvap15:12
ttxjeblair: on the long run, it's one drawback of using only a part of an integrated suite -- we'll always be stuck in corners15:12
*** rfolco has quit IRC15:12
openstackgerritJeremy Stanley proposed openstack-infra/system-config: Add a server
ttxI fear they will always prioritize for the case of people that will use the full suite15:13
openstackgerritMatt Riedemann proposed openstack-dev/pbr: Add more documetnation around building docs
ttxwikimedia apparently decided to adopt them all15:13
ttxand will migrate from gerrit15:14
fungiunderstandable. we do the same with the tools we develop15:14
*** rfolco has joined #openstack-infra15:14
openstackgerritMatt Riedemann proposed openstack-dev/pbr: Add more documentation around building docs
fungi(for example, the discussion yesterday about being able to use nodepool without zuul)15:15
fungi"yes you can probably do it, no we don't recommend it, be prepared for some painful workarounds and/or custom patches"15:15
ttxthey also resist adding options, so whatever they define as "the one true way" is likely to be what we have to use15:16
*** jlanoux has quit IRC15:16
mordredI do tend to agree with the theory of not adding tons of options15:17
*** unicell1 has joined #openstack-infra15:18
mordredmriedem: aw. I like "documetnation" :)15:18
*** unicell has quit IRC15:19
*** yamamoto has joined #openstack-infra15:19
ttxmordred: they also don't mind humans updating tasks. So triaging bugs, and closing tasks when things merge are all human tasks15:19
*** _nadya_ has joined #openstack-infra15:19
ttxI guess that's a by-product of having staff paid to do that15:19
*** austin81 has joined #openstack-infra15:20
mordredoh weird15:20
austin81join #openstack15:20
*** jgriffith has quit IRC15:20
austin81oops haha15:20
ttxin openstack we kinda like not having to update things manually15:20
*** boris-42 has joined #openstack-infra15:21
ttxmordred: I'm trying to picture in my mind a world where we'd use Maniphest purely for task tracking (organization of work, backlogs), completely separately from commit/change tracking, and see if that could work15:22
ttxI just fear that would be a regression for horizontal teams like the VMT15:22
openstackgerritgreghaynes proposed openstack-infra/nodepool: Trigger image delete using gearman
*** dims has joined #openstack-infra15:23
*** dizquierdo has quit IRC15:23
mordredttx: I'm having a hard time picturing that world in my mind15:23
ttxI mean, it's diable. Use Maniphest the way some teams use Trello.15:23
*** jsavak has joined #openstack-infra15:23
mordredttx: in your imagination there, what do we use for the other thing?15:24
*** yamamoto has quit IRC15:24
jeblairi don't think we want to be updating things manually; istr mordred saying that the api was a low priority for them, but if people were interested in using it, priority may increase15:24
fungidyeable too. we can make it any color we want15:24
mriedemmordred: yeah, i like not having to read source code :)15:24
ttxI'm trying to see if we could do without it.15:24
mriedemmordred: and not having typos15:24
openstackgerritMerged openstack/requirements: Bump YAQL version to 0.2.7
ttxbecause having the two (change/release tracking and persnal/team task tracking as separate tools is a bit insane)15:25
ttxthe question becomes, do we need change/release tracking15:26
*** asilenkov__ has quit IRC15:26
ttxI see three areas15:26
ttxone is complex task coordination, the "vulnerability" use case15:26
* anteaya claps hands, here comes a ttx list15:26
ttxone is post-release tracking, "what's in a release", and we already said we should generate that from commit messages15:27
*** tsekiyama has joined #openstack-infra15:27
ttxlast one is release goals tracking. Ability to say "RC2 is blocked on landing this set of bugfixes"15:27
ttxanteaya: I'll admit I seriously abuse the list construct15:28
jeblairwell, that's what this whole conversation is about.  :)15:28
ttxI guess the last one is simple enough and could be emulated using tasks alright15:29
ttxI'm mostly concerned about the first one. And now that I abandoned the VMT, I'm mostly concerned by proxy15:29
anteayattx is how you think15:30
anteayattx I look forward to your lists15:30
ttxbeing able to say "this task is done" using commit data to back it up15:30
*** spzala has quit IRC15:30
ttxrather than humans processing it and forgetting something15:30
anteayaor having something move while you were asleep and you lose track15:31
fungiright, it's generally helpful in vulnerability reports which involve more than one stable branch and more than one project to be able to know when fixes are proposed and when they're merged and track which ones are still missing15:31
ttxbasically Launchpad did both (task/release), Phabricator only does one (task). Storyboard currently only does one too (task). The release tracking stuff was still on the pipe15:31
*** bexelbie has quit IRC15:32
*** asselin__ has quit IRC15:32
ttxI thought Maniphest tasks mapped to Launchpad bugtasks. That was a mistake. They map to Launchpad bugs (and Storyboard stories)15:32
mordredyah - they are bugs - except with dependency relationships15:33
mordredso - they're a hybrid of bugs and blueprints from lp15:33
fungiwe've had plenty where a bug was introduced in oslo-incubator, copied into two projects in one release, then split out into a separate library which also carried that vulberability, and one of those two projects switched to consuming the library so no longer has the bug on a later branch wher eit should instead be fixed in the library, but the other project is still syncing the bug from the15:33
*** dpose has quit IRC15:33
fungiincubator... it becomes a mess to untangle15:33
ttxsubtasks are really made to track compeltely separate tasks15:33
openstackgerritPaul Belanger proposed openstack-infra/system-config: WIP: Migrate to puppet-httpd module
ttxand the long-term prospect of using Releeph will block any short-term improvement there15:33
jeblair(the dependencies will be useful -- we're writing specs with multiple "work items")15:34
*** ddieterly has joined #openstack-infra15:34
ttxand we are not even sure releeph will be usable for us and not depend on some other *ph* thing that we don't want to use15:34
*** jgriffith has joined #openstack-infra15:34
mordredfungi, ttx: agree that the VMT tasks is a mess to solve - otoh, it's also the most specialized use case that gets driven by a tight cadre15:34
ttxjeblair: sure. They are useful the same way relationships between blueprints were useful15:34
ttxnot the same way tasks in bugs were useful.15:34
jeblairttx: i got the impression from epriestly that releeph might be mostly there; at least enough maybe we can look at it and see what it's integration (with maniphest and possibly other phtools) is like15:35
fungiwaiting for them to name a project "phuket" (also seems like it would be a fun place to visit)15:35
mordredfungi: I want to go to there15:35
ttxjeblair: <epriestley> Releeph is totally unusable today though and probably nowhere near close enough to working to be made workable.15:36
ttxnot "mostly there" in my book15:36
mordredjeblair: yah. also - depending on what the diffusion integration story looks like, replicating from gerrit to a diffusion wouldn't be much worse than gerrit having a gitweb that we mostly never look at15:36
ttxyeah, diffusion can be made to mirror external repos15:37
BobBallAre there known issues with jenkins job builder compatability with newer jenkins?  It's generating a job with a publisher just fine, however when it's uploaded to jenkins the publisher step is missing from the job's config15:37
mordredwhich is to say - having an additional place that visualizes git commits is basically adding a fourth to a list that's already got 3 thigns in it15:37
ttxrequires to create a separate prefix for each repo though, which wil be fun with our hundreds of repos :)15:37
mordredyah. that's my least favorite part about what they do with diffusion15:37
mordredttx: if we could get a way out of that prefix stuff, I think that woudl be more palatable if it's the mechanism we need to use to trigger something15:38
pabelangernibalizer, so, not sure why but I am starting to see more puppet-apply-trusty failures15:38
fungimordred: i wouldn't really say that the vmt use case is unrepresentative of openstack in general, just that we have some examples of particularly complicated bug proliferation between projects and branches. any bug could exhibit similar proliferation, it doesn't have to be a security vulnerability... those are just ones we want to be a lot more diligent about not letting fall through the cracks in15:39
mordredfungi: totally15:39
ttxmordred: think one of those bugs where you tried to land a pbr fix in dozens of projects15:39
*** bswartz has quit IRC15:39
*** Sukhdev has joined #openstack-infra15:39
*** cindyo has quit IRC15:39
ttxthe list of projects to strike though came handy there too15:39
ttxWe can definitely track it manually in a comment / description checklist15:40
ttxbut that is a bit full of unfun15:40
*** e0ne has joined #openstack-infra15:40
rcarrillocruzTheJulia: i believe i found a solution to the race condition15:40
fungiin fact, we _would_ have a bunch more non-vmt examples of this _if_ launchpad weren't so totally broken that it makes them a pain to deal with. the vulnerabilities are simply a situation where enduring the pain is better than giving up15:41
rcarrillocruzas I commented earlier, whenever I got a provisioned server without eth0 IP config set up, i saw on dmesg a corresponding 'simple-init main process terminated with exit 1' yadayada15:41
rcarrillocruzi commented out the dib-install include15:41
mordredttx: indeed - but I actually tend to track those in gerrit rather than a bug tracker15:41
*** zul has quit IRC15:41
mordredttx: largely because tracking them in the bug tracker was always more work than tracking their status with gerrit15:41
rcarrillocruzmodified simple-init.conf from /opt/stack on the ironic host so it had a respawn stanza15:42
rcarrillocruzand now15:42
*** david-ly_ is now known as david-lyle15:42
rcarrillocruzi get all the interfaces set up15:42
sdaguemordred: well, that's good if one person is generating all the patches. But we had things like using python urllib with https, which all projects did15:42
jeblairttx: 14:58 <@epriestley> Releeph is hypothetically not all that far from being usable, but there's nowhere near enough room on the roadmap to sneak it in anytime soon.15:42
sdagueand the per project patches were different15:42
ttxmordred: the trick is... you can only track them when they are in. Not the need to create them.15:42
jeblairttx: i'll agree with you that the status as described by epriestly is unclear :)15:42
mordredsdague: totally. not discounting the use case15:43
*** dustins has quit IRC15:43
jeblairsdague: change metadata 'tags' in gerrit will help with that (go sdague!)15:44
rcarrillocruzit's as if when we have multiple interfaces the concurrent simple-init/glean executions step each other, if one of them corresponding to eth0 gets terminated then you are out of luck and you get a provisioned server without the management NIC configured15:44
ttxReleeph is certainly something we can enable to see. It's in "beta" mode15:44
*** rkukura has joined #openstack-infra15:44
rcarrillocruzmordred: ^, i will push  a change to the simple-init element to add a respawn, we can hit race issues on servers when multiple NICs are available15:44
TheJuliarcarrillocruz: that kind of makes sense if upstart doesn't realize that multiple ones are firing up intentionally.15:44
rcarrillocruz(my servers have up to 6 NICs!)15:45
ttxmordred: moving off LP partly because it could track more than 12 tasks and switching to seomthing that only supports 1 is kind of weird :)15:45
mordredrcarrillocruz: nice!15:45
ttx"couldn't track more", I mean15:45
TheJuliarcarrillocruz: That is just wrong... but in solidarity I'll look for a few cards to add to my servers locally :)15:45
*** annegentle has quit IRC15:45
*** laughterwym has quit IRC15:45
ttxthat said, there are a lot of things I like in Phab. The Herald stuff is pretty cool.15:45
*** annegentle has joined #openstack-infra15:46
*** kiran-r has joined #openstack-infra15:46
rcarrillocruzi just gave you a justification to get a beefy server for your testing ;-)15:46
fungislap a couple of quad-port nics in and you're set15:46
ttxlets you define personal rules that apply automatically on specific events. Like if the new bug mentions rootwrap, add me to cc. Or add "rootwrap" project15:47
*** enikanorov2 has joined #openstack-infra15:47
*** coolsvap is now known as coolsvap|away15:47
*** salv-orlando has joined #openstack-infra15:47
mordredttx: that is pretty cool15:47
ttxWill help with keeping triaging under control15:47
ttxIf I didn't like it, I wouldn't spend so much head-wrapping time trying to figure out how we could use it :)15:48
*** coolsvap|away is now known as coolsvap15:48
*** fallenpegasus has joined #openstack-infra15:48
*** jasondotstar has quit IRC15:48
*** abregman has quit IRC15:48
mordredhehe. excellent point15:48
*** patrickeast has joined #openstack-infra15:49
*** BharatK has joined #openstack-infra15:50
*** devvesa has quit IRC15:51
TheJuliarcarrillocruz: already had one, except I keep promising to actually fire up a deploy on it... guess I should stop rebuilding it with ironic every day :)15:51
openstackgerritDavid Shrewsbury proposed openstack-infra/shade: Add a testing framework for the Ansible modules
*** yamada-h has joined #openstack-infra15:51
Shrewsmordred: ^^^^ an idea15:51
ttxmordred: but that was a useful exercise. I think we need to rethink our initial idea of using subtasks to track things landing in a set of needed repositories. Maybe we should use a gerrit topic instead15:52
Shrewsmordred: i wanted something flexible. i think that works (in my head, at least)15:52
mordredShrews: looks great15:53
*** patrickeast has quit IRC15:53
ttxmordred: maybe we could file placeholder gerrit changes to describe needed work, and track all the related things under a single topic.15:53
ttxthinking out loud15:53
ttxstill a bit of a pain for the VMT15:54
*** unicell1 has quit IRC15:55
ttxsince their patches can't be public15:55
*** yamamoto has joined #openstack-infra15:55
*** spzala__ has joined #openstack-infra15:55
ttxlike have automation that will autoclose tasks if all the changes in the linked topic are done15:56
*** jistr has quit IRC15:56
*** salv-orlando has quit IRC15:56
ttxand maybe have that automation update a table at the top of the description based on the linked topic.15:56
*** davideagnello has joined #openstack-infra15:57
ttxI think I'm onto something;15:57
*** yamahata has quit IRC15:57
openstackgerritDavid Shrewsbury proposed openstack-infra/shade: Add a testing framework for the Ansible modules
fungiproblem is that review topics can grow additional changes later. not sure if you need to deal with the possibility of a new change added to a review topic after the master task has been closed?15:57
*** yamahata has joined #openstack-infra15:57
*** jlanoux has joined #openstack-infra15:58
ttxfungi: the review topic could be strongly tied to the task. Like be called "T34523"15:58
*** jasondotstar has joined #openstack-infra15:58
ttxif you add more you could reopen it15:58
fungijust wondering if that should trigger reopening15:58
fungiyeah, that15:58
sdaguejeblair: well I have to get time to start looking at that :)15:59
*** spzala__ is now known as spzala15:59
ttxfungi: tables in phab descriptions can be made to look nice:
*** patrickeast has joined #openstack-infra15:59
funginot bad. though i'm less worried about the aesthetics of it and more about the accuracy/utility of it16:00
*** ddieterly has quit IRC16:00
ttxthat could be a solution while waiting for releeph to happen16:00
ttx"while waiting for relief" heh16:00
*** bswartz has joined #openstack-infra16:00
fungisweet release16:00
*** sridhar_ram has joined #openstack-infra16:00
*** salv-orl_ has joined #openstack-infra16:00
ttxfungi: gerrit would update it on changes16:01
*** geoffarnold has quit IRC16:01
ttxnot worse than te current gerrit->LP situation16:01
fungiwe still have problems with that now, but maybe a better/less failure-prone api will make it more robust16:01
ttxthe main issue is that there is no convenient way to say "a change will be needed for nova/kilo"16:02
fungiother than to literally type it out in a comment, right16:02
*** amrith is now known as _amrith_16:02
ttxshort from pushing a placeholder change on that topic16:02
fungior use a kanban16:02
*** scheuran has quit IRC16:02
ttxno I mean to track that in gerrit itself16:02
ttxsince Maniphest would just mirror the status of the Gerrit topic16:03
*** akraminsky has joined #openstack-infra16:03
ttxI'm just pushing mordred's "I track things in Gerrit" metaphor to the logic conclusion and see where that puts us16:04
fungii see16:05
*** jsavak has quit IRC16:06
ttxfungi: maybe some git-review --placeholder nova/kilo16:06
*** jsavak has joined #openstack-infra16:06
ttxgit-review --placeholder T2345 nova/kilo16:07
*** Ala has quit IRC16:07
* ttx stops dumping his brains in the middle of the channel and will soon call it a day16:07
fungithat could be fairly easy to add. make it create a dummy file in the repo so that gerrit won't reject a diffless change, and then know to delete that file when the actual change is added. though that makes whoever creates the placeholder also the change owner16:07
*** zul has joined #openstack-infra16:08
*** kiran-r has quit IRC16:08
fungiunless the placeholder change is abandoned and replaced by a completely new change-id16:09
ttxyeah, the owner bit is slightly problematic16:09
*** patrickeast has quit IRC16:09
ttxalthough it could be seen as a feature.16:10
*** patrickeast has joined #openstack-infra16:10
*** dontalton has joined #openstack-infra16:10
*** ivar-lazzaro has joined #openstack-infra16:10
mordredhi nibalizer16:11
fungittx: perhaps, except we also use that (currently) to track contributions for elections and summit passes16:11
*** coolsvap is now known as coolsvap|away16:11
*** dtantsur is now known as dtantsur|afk16:11
*** jsavak has quit IRC16:12
*** jsavak has joined #openstack-infra16:12
*** fallenpegasus has quit IRC16:14
sdagueclarkb / fungi so subnode seems to still not be working due to dns truncation issues16:14
sdaguethis is a run from this morning16:14
fungisdague: yep, saw that last night. pointed it out to clarkb/greghaynes/mordred16:14
openstackgerritKhai Do proposed openstack-infra/puppet-openstackci: Add Jenkins Job Builder manifest
fungiapparently the glean fix was insufficient16:15
*** markus_z has quit IRC16:15
*** dims has quit IRC16:15
mordredwhen greghaynes awakens I expect him to be unhappy16:15
fungisubnodes have extra information in the hostname which puts the fqdn well over the limit16:15
openstackgerritMatt Riedemann proposed openstack-infra/project-config: Remove non-voting gate-horizon-dsvm-integration from gate queue
mriedemdavid-lyle: sdague: mtreinish: ^16:16
*** ihrachyshka has quit IRC16:16
fungiwe may want to consider sticking less information in the hostnames themselves, or abbreviating some things16:16
*** dims has joined #openstack-infra16:16
mordredwell, I mean- the glean change was supposed to not put the domain name in16:18
fungiyeah, i need to jump on one and see what exactly it ended up configuring16:19
fungii'll do that real quick16:19
*** ivar-lazzaro has quit IRC16:19
sdagueso, the 64 byte limit is truncating from where?16:19
*** ChanServ changes topic to "Common CI Sprint happening now in #openstack-sprint"16:19
fungisdague: HOST_NAME_MAX16:19
*** ivar-lazzaro has joined #openstack-infra16:20
* mordred blames C16:20
*** dims has quit IRC16:21
*** ivar-laz_ has joined #openstack-infra16:21
fungisdague: mordred: greghaynes: clarkb:
*** jbrc has joined #openstack-infra16:21
sdagueit also seems odd to encode all that into the hostname16:21
zaroasselin: made the change to zuul_scheduler.pp16:22
asselinzaro, want to join us in #openstack-sprint?16:22
fungiso anyway, i think glean isn't actually doing what we expected there?16:22
mordredfungi: agree16:22
mordredfungi: can you check glean version real quick?16:22
fungimordred: where is it installed?16:23
*** whoops has joined #openstack-infra16:23
mordredpbr freeze | grep glean shoudl do it16:23
mordredare we sure these are glean-based?16:23
fungimaybe we're not installing it?16:23
mordreddpkg -l cloud-init ?16:23
mordredI didn't think we were using any glean-based servers in any gate jobs yet, tbh16:23
fungimordred: yep, cloud-init is installed16:24
sdagueyeh, doesn't look like clean is there16:24
mordredfungi: so - how about we move the multi-node nodes to use ubuntu-trusty instead of devstack-trusty16:24
fungimordred: these are just our typical devstack-trusty snapshot images in rax, i think, unless we switched to glance'ing those everywhere16:24
*** ivar-lazzaro has quit IRC16:24
mordredwe've tested and devstack works on the ubuntu-minimal nodes16:24
sdagueso, how is that compute node dns name generated?16:25
fungimordred: seems reasonable. i'll write a quick patch16:25
SpamapSmordred: what are the two things you learned that we should reflect in shade?16:25
*** jasondotstar has quit IRC16:25
openstackgerritMatt Riedemann proposed openstack-infra/elastic-recheck: Add query for neutron functional test race bug 1472692
openstackbug 1472692 in neutron "test_port_creation_and_deletion randomly fails with: eventlet.timeout.Timeout: 60 seconds" [Undecided,New]
mordredSpamapS: ah - so ... domains dont' work quite like we expected16:26
*** mayurig has joined #openstack-infra16:26
mordredSpamapS: (there's a bunch more scrollback between Shrews and I) - I'm about to add tests to that16:26
*** unicell has joined #openstack-infra16:27
mordredfungi: I also made a patch to move the rest of the gate to ubuntu-trusty - but it's already in merge conflict :)16:27
*** hashar has quit IRC16:28
*** changbl has quit IRC16:29
*** jaypipes has quit IRC16:29
fungimordred: greghaynes: i'm also unconvinced glean is doing what we expect with hostnames where it is in use...
SpamapSmordred: gets direct to the API seems like something we should almost always do if we're not aggressively caching. :)16:29
SpamapSmordred: but I digress, that makes sense.16:29
*** davideagnello has quit IRC16:29
*** davideagnello has joined #openstack-infra16:30
*** sshnaidm has quit IRC16:31
mordredSpamapS: so - it gets tricky16:33
mordredSpamapS, Shrews: because it's hard to do name_or_id if we don't filter the list client side :)16:34
SpamapSmordred: I know that, but I also think there should be an 'id' optional arg that short circuits listing.16:35
SpamapSmordred: because this isn't the only 'list' operation that is inefficient.16:35
SpamapSBUT, in most cases, aggressive caching is enough to mitigate that and keep things simpler.16:36
SpamapSdomain listing is an example where it's not performance, but policy, that stands in the way16:36
*** tonytan4ever has quit IRC16:36
mordredwell, it's both16:36
mordredthe policy is there because of the extreme expense of lissting domains16:36
mordredsince the listing is not user scoped but cloud global16:37
mordredwhich could potentially mean a global listing search on an LDAP backend16:37
*** Swami has joined #openstack-infra16:38
*** ihrachyshka has joined #openstack-infra16:38
mordredSpamapS: so - in this case, I'm going to suggest we tie the domain API to use explcit IDs, rather than the typical name_or_id pattern16:39
mordredbecause it seems to be an area where we need people to be specific about what they mean16:39
openstackgerritKhai Do proposed openstack-infra/puppet-openstackci: Add Jenkins Job Builder manifest
SpamapSmordred: could this also just be one of those pieces that is only useful in admin context, and consistency is still a win even if listing domains is inefficient?16:40
*** dims has joined #openstack-infra16:40
mordredSpamapS: I think yes and no16:40
mordredI think it's a piece that is definitely only useful in an admin context16:41
mordredbut haneef from the HP keystone team basically told me that _ever_ running a domain list is just a monumentally terrible thing to do16:41
mordredso I think we should expose it16:41
*** achanda has joined #openstack-infra16:41
*** dontalton is now known as bitblt16:41
mordredbut not use it when someone did not ask us to16:41
*** bitblt has quit IRC16:41
*** bitblt has joined #openstack-infra16:41
*** fitoduarte has joined #openstack-infra16:41
*** puranamr has joined #openstack-infra16:42
SpamapSmordred: having something in the keystone API that is a terrible thing to do... is perhaps something we should address upstream.16:42
*** bitblt has quit IRC16:42
mordredwell - it's possible you might need to do it - but as an admin you'll know that16:42
*** _nadya_ has quit IRC16:42
*** afazekas has quit IRC16:42
*** otter768 has joined #openstack-infra16:42
mordredlike, 'sudo rm -rf /' is a thing that's possible to do after all16:43
*** _nadya_ has joined #openstack-infra16:43
mordredbut the cases in which you _want_ to are pretty low16:43
*** liam_ has joined #openstack-infra16:43
*** fallenpegasus has joined #openstack-infra16:43
SpamapSYeah, k. So another way to address this is to allow overriding the list you might get from that API call with a static list, possibly in the cloud config.16:44
*** liam_ is now known as Guest7196316:44
SpamapSmordred: still. I'd like to see Keystone push that list into a cache rather than zomg realtime queries for everything.16:44
mordredalthough I have yelled enough at morganfainberg recently16:45
SpamapSsort of like domain transfers.. "close enough"16:45
SpamapSmordred: I've landed a number of scalability fixes in Keystone.. I think I speak their language. :)16:45
*** ashleighfarnham has joined #openstack-infra16:46
openstackgerritMikhail S Medvedev proposed openstack-infra/puppet-project_config: Allow specifying repo ref as a class parameter
*** _nadya_ has quit IRC16:47
jeblairmordred/SpamapS tag team yelling at morganfainberg16:47
*** otter768 has quit IRC16:47
openstackgerritMonty Taylor proposed openstack-infra/shade: Rework how we get domains
mordredSpamapS: there ^^ that's what I'm thinking16:47
mordredoh - wait - one more change16:48
morganfainbergWhat did i do?16:48
mordredget _does_ work for normal users16:48
mordredmorganfainberg: we're still on domains16:48
SpamapSjeblair: no, its good cop bad cop. mordred yells, I console. ;)16:48
*** ihrachyshka has quit IRC16:48
mordredmorganfainberg: and I still think they should be named realm - but we weren't really talking about that ;)16:48
morganfainbergReasonable to make that list a cache aware thing.16:48
morganfainbergmordred: I don't disagree re: realm.16:48
SpamapSmorganfainberg: I'm thinking it needs to be cached more aggressively than clients can reasonably be expected to do.16:49
mordredmorganfainberg: haneef gave the feedback that using GET /v3/domains and then filtering client side to find the id given a domain name was exceptionally evil for things like ldap backend16:50
SpamapSmorganfainberg: if the backends really are fragile enough not to be able to spit a list back in realtime, then keystone should probably do that query in a way friendly to the backend. LDIF anyone? ;-)16:50
*** fallenpegasus has quit IRC16:50
mordredSpamapS: well, I think "get" and list with a filter are friendly enough ... you know ... we COULD just run list twice - once with id=name_or_id and if that's empty once with name=name_or_id16:51
mordrednow I've become unconvinced of whether or not a non-admin user can do get16:53
mordredmorganfainberg: ^^16:53
SpamapSmordred: yeah, that of course assumes ids will never collide with names (safe enough where we aren't likely to care about the one guy whose domain is deadbeefdeadbeefdeadbeefdeadbeef and gets hacked....)16:53
mordredmorganfainberg: the devstack config allows the demo user to do /v3/domains ... but I know that's intended to only be admin16:53
mordredmorganfainberg: is a thing an end user can expect to run?16:54
SpamapSseems like domain-admins wouild be able to get16:54
*** enikanorov2 has quit IRC16:54
SpamapSend-users that would possibly be a site specific policy decision16:54
* mordred wants site-specific policy decisions to diaf16:54
*** maishsk has joined #openstack-infra16:54
mordredthey make absolutely everything terrible16:55
SpamapSThough this is where I start to really really really want a rich ACL akin to IAM. ;)16:55
SpamapSmordred: you are so right, but the answer is per-domain policies that put the power to make your app homogenous in the users' hands, not homogenous policy for all the clouds which will drive some users away.16:56
SpamapSand IIRC that discussion is still in its infancy.16:56
SpamapSbecause I think that baby has been drowned a few times16:56
mordredSpamapS: right. BUT16:56
mordredthe things outside of your domain admin control16:56
mordredsuch as access to the global list of domains16:56
*** kutija has joined #openstack-infra16:56
mordredneed a consistent policy across clouds16:56
mordred"can I ask the cloud for the details of the domain I have access to" should be consistent16:57
mordred"what can the users I created do in the project I created in the domain I control" ... yeah, that's all me16:57
*** bexelbie has joined #openstack-infra16:58
openstackgerritMonty Taylor proposed openstack-infra/shade: Rework how we get domains
mordredShrews, SpamapS: ^^ k. that works now16:58
*** ayoung has quit IRC16:59
*** _amrith_ is now known as amrith17:00
*** sridhar_ram1 has joined #openstack-infra17:01
*** dustins has joined #openstack-infra17:02
*** maishsk has quit IRC17:02
*** maishsk has joined #openstack-infra17:03
*** sridhar_ram has quit IRC17:03
openstackgerrityolanda.robla proposed openstack-infra/puppet-cgit: Stop including policycoreutils-python
*** bswartz has quit IRC17:05
*** shashank_hegde has joined #openstack-infra17:06
*** enikanorov2 has joined #openstack-infra17:06
*** vikram_ has joined #openstack-infra17:07
*** jlanoux has quit IRC17:08
*** bswartz has joined #openstack-infra17:08
vikram_how we can resolve "Build timed out (after 30 minutes). Marking the build as failed" issue for a patch17:10
vikram_any idea?17:10
*** achanda has quit IRC17:12
openstackgerritSpencer Krum proposed openstack-infra/puppet-os_client_config: Module boilerplate
*** achanda has joined #openstack-infra17:12
*** Guest71963 has quit IRC17:13
*** puranamr has quit IRC17:14
fungivikram_: make the job run faster or set the timeout longer17:14
*** jsavak has quit IRC17:15
fungivikram_: if it's just affecting one patch, that patch may be making the software or the tests slower17:15
*** claudiub has quit IRC17:15
vikram_fungi: can we rerun it?17:15
vikram_fungi: how we can set the timeout longer17:16
fungivikram_: you can, if you think it was an isolated incident. just leave a review comment starting with the word "recheck" on the change where you got a -117:16
vikram_fungi: thanks.. will do a recheck17:17
fungivikram_: oh, you should have said you were talking about the gate-infra-puppet-apply-centos6 job. it's non-voting because we have a git performance issue we're trying to track down on centos 6.x17:17
fungivikram_: it's fine for that job to report a timeout right now. it won't prevent your patch from being reviewed/approved17:17
vikram_fungi: thanks for the help!17:18
fungivikram_: you're welcome17:18
*** vikram_ has left #openstack-infra17:18
*** ari has joined #openstack-infra17:18
*** abitha has joined #openstack-infra17:18
*** crc32 has joined #openstack-infra17:19
*** fhubik has quit IRC17:20
*** jsavak has joined #openstack-infra17:20
*** alexsyip has joined #openstack-infra17:21
*** ivar-laz_ has quit IRC17:24
*** ivar-lazzaro has joined #openstack-infra17:24
*** puranamr has joined #openstack-infra17:24
openstackgerritMerged openstack-infra/shade: Add a testing framework for the Ansible modules
*** ociuhandu has quit IRC17:25
openstackgerritVipul Sabhaya proposed openstack-infra/project-config: update cue project list in rally gate job
*** e0ne has quit IRC17:28
*** maishsk has quit IRC17:29
*** yamahata has quit IRC17:30
*** BharatK has quit IRC17:32
*** pelix has quit IRC17:32
*** spzala_ has joined #openstack-infra17:33
*** crc32 has quit IRC17:34
*** spzala has quit IRC17:35
*** spzala_ is now known as spzala17:35
openstackgerritMonty Taylor proposed openstack-infra/shade: Rework how we get domains
mordredShrews, SpamapS: now with testing17:36
*** dannywilson has joined #openstack-infra17:37
*** maishsk has joined #openstack-infra17:37
*** bknudson has quit IRC17:38
*** dims has quit IRC17:39
*** Sukhdev has quit IRC17:39
*** tonytan4ever has joined #openstack-infra17:39
*** dims has joined #openstack-infra17:39
*** yamamoto has quit IRC17:41
ShrewsSpamapS: also, we are starting to add, to the shade repo, playbooks to exercise ansible modules. feel free to add any you may have17:42
*** fallenpegasus has joined #openstack-infra17:42
SpamapSoh crap I have some hiding somewhere17:42
* SpamapS lost the thread on that one.17:43
*** Sukhdev has joined #openstack-infra17:43
SpamapSoh well now this is silly..17:44
SpamapSthe puppet vswitch modules _always_ install openvswitch-datapath-dkms ... even though it is not needed with kernels > 3.1117:44
fungithat seems rather a shotgun approach to kernel module handling17:44
openstackgerritCody Littley proposed openstack-infra/jenkins-job-builder: According to the message on line 443:
SpamapSyeah, well it breaks apt on trusty17:45
fungialso at this point i have limited interest in platforms with a kernel <=3.1117:45
SpamapSbecause openvswitch-datapath-dkms refuses to build when kernel is > 3.11 since it is already in-kernel.17:45
SpamapSfungi: like.. trusty? ;)17:45
openstackgerritCody Littley proposed openstack-infra/jenkins-job-builder: Add deprication warning in documentation
clarkbwhy does the package exist at all in that case?17:45
SpamapSclarkb: because trusty is o_O on kernel17:46
SpamapSTrusty has 3.11 as it's default kernel17:46
*** yamamoto has joined #openstack-infra17:46
SpamapSbut you can get 3.1317:46
clarkbso it should work there17:46
SpamapSand you SHOULD get 3.1317:46
SpamapSactually wait no!17:46
fungiahh, that's why i thought it had 3.13 as the default17:46
clarkbno you get 3.11 by default in the cloud17:46
SpamapStrusty has 3.13 ...17:46
SpamapSso yeah, why the hell does this exist?17:46
clarkbbecause default is 3.1117:47
mordredSpamapS: sounds like a good patch to puppet vswitch17:47
SpamapSclarkb: I think thats only on 14.04.017:47
clarkbSpamapS and all cloud images17:47
mordredSpamapS: there is a facter kernelversion fact taht should be easy to put a conditional on17:48
SpamapSclarkb: uggggh17:48
clarkbyou have to opt in to new kernel17:48
clarkband upgrade then reboot17:48
mordredclarkb: do we opt in to new kernel on the dib images we build?17:48
SpamapSapparently typing 'sudo apt-get install linux-image-generic' == opting in17:48
mordredclarkb: maybe we should?17:48
*** mayurig has quit IRC17:48
mordredor do we not on purpose?17:48
*** amotoki has quit IRC17:48
*** puranamr has quit IRC17:48
clarkbon purpose of we test default trusty by way of not caring ?17:49
SpamapSyeah really, the 3.11 they included in that early Ubuntu 14.04 is just a 3.12 without some of the bug fixes. It's weeeeirrrd17:49
*** marzif_ has joined #openstack-infra17:49
*** mayurig has joined #openstack-infra17:49
SpamapSclarkb: the Ubuntu images on HP Cloud are 3.1317:49
*** mayurig has quit IRC17:49
clarkbSpamapS we dont use those17:50
fungiat this point at least, i only see 3.11 anywhere for precise
*** mayurig has joined #openstack-infra17:50
mordredfungi, clarkb: I mean, yeah - except anyone deploying openstack on trusty is almost certainly goign to apt-get upgrade first, no?17:50
SpamapSclarkb: sure we do, they're the same one that diskimage-builder builds from17:50
clarkbwe use what dib makes and Kiall ran into 3.1117:50
clarkbSpamapS arw they?17:50
SpamapSunless we use ubuntu-minimal17:50
mordredSpamapS: not yet17:50
clarkball i know is kiall got 3.1117:50
mordredalthoguh SOON17:50
clarkbso pretty sure we 3.1117:50
SpamapSclarkb: Kiall likely had an older image. They upload new ones and rename them all the time.17:50
fungicould Kiall have been using precise?17:50
SpamapSclarkb: and diskimage-builder gets 'current'17:51
mordreddid Kiall get 3.11 in nodepool from us recently?17:51
mordredor elsewhere?17:51
SpamapSPrecise is 3.1117:51
SpamapSin fact17:51
SpamapSI am actually sorry, Trusty never had 3.1117:51
mordredSpamapS: oh yeah - we know all about teh upload/rename17:51
clarkbfrom nodepool recently enough17:51
clarkbmayve a month ago17:51
mordredclarkb: sounds like precise according to SpamapS ^^17:51
* fungi is now far less confused by the statements contradicting evidence ;)17:51
clarkbok maybe he beedes >3.1317:52
SpamapSTrusty has 3.13 + a metric ton of patches from 3.1417:52
clarkbwhat I know is we dont use hardware compat kernel17:52
mordredSpamapS: the upload/rename is one of the things driving us to dib :)17:52
SpamapSmordred: because we use ids?17:52
mordredSpamapS: no. because we do not ids17:52
SpamapSbecause.. I think it's a fair thing to do.17:52
mordredSpamapS: they rename17:52
clarkbwe use the .0 lts release kernel base version whatever that number is17:52
fungiit seems that precise is 3.2 with available opt-in updates to 3.1117:52
mordredSpamapS: we use named17:52
mordredSpamapS: we use names17:52
mordredSpamapS: but they change the names17:52
mordredSpamapS: so our configs break17:52
clarkbuuids change too17:52
clarkbso both would break17:52
*** puranamr has joined #openstack-infra17:53
*** yamahata has joined #openstack-infra17:53
mordredwell, the old uuid would work - but they actually add the string "deprecated" to the name of the old image17:53
clarkbno they change the old uuid too17:53
mordredthey do? jeez17:53
clarkbbecause its a delete + add17:53
mordredit's a completely madness17:53
SpamapSoohh they don't rename _to the thing that works all the time_17:53
SpamapSdohhh wtf!17:53
mordredit's insane pants17:53
mordredso - dib is a big win for us17:53
SpamapSthats such a..17:54
mordredbecause we can manage logical constructs with it17:54
SpamapSlike who is making this decision?17:54
mordredSpamapS: nobody17:54
SpamapSbah, onward17:54
fungibasically there is no commonality in their image uuids nor names. when they replace an image they rename and create a new uuid for the old image so you can't boot it any longer, but the new image doesn't assume the name nor uuid of the old image either17:54
SpamapSthe real question is in fact "why does that package exist" regarding openvswitch-datapath-dkms in Trusty17:54
*** mpaolino has quit IRC17:54
mordredSpamapS: ++17:54
*** puranamr has quit IRC17:55
fungiSpamapS: i wager it's an oversight17:55
fungicoupled with poor testing17:55
openstackgerritK Jonathan Harker proposed openstack-infra/puppet-log_processor: Make the subunit-files config section optional
fungior rather likely incomplete testing because they probably can't test kernel modules sanely17:55
fungiso maybe they skip that17:55
SpamapSoh hah17:56
SpamapSthe upstream vswitch puppet module is already fixed17:56
*** Longgeek has quit IRC17:57
mordredcrinkle wins17:57
KiallDid I what now? Only time I've mentioned kernel versions in here was related to gate slaves etc ;)17:57
*** Longgeek has joined #openstack-infra17:57
*** Longgeek has quit IRC17:57
SpamapScrinkle: how do we get them to tag that 0.2.0?17:58
*** madorn has joined #openstack-infra17:58
openstackgerritArmando Migliaccio proposed openstack-infra/project-config: Drop unstable configuration for tempest-dsvm-neutron job
crinkleSpamapS: you should be looking at
fungiKiall: we just like to blame everything on you. except for git/gerrit bugs which we still blame on SpamapS17:59
crinkleit's been tagged 1.0.017:59
*** Longgeek has joined #openstack-infra17:59
openstackgerritCody Littley proposed openstack-infra/jenkins-job-builder: Add deprication warning in documentation
madorni am trying to resolve an issue with alternate emails for one user showing up at when attempting to select a review. can someone point me in the right direction?  there is only one launchpad user and one email associated with this user.18:00
madornreview = reviewer18:00
crinkleo.0 i didn't even know puppetlabs was maintaining a different one18:00
SpamapScrinkle: ooh thats just great18:00
anteayaBobBall: you had mentioned interest in the common ci sprint, it is currently happening in #openstack-sprint18:00
SpamapScrinkle: why is that openstack specific?18:01
anteayaBobBall: as a reminder18:01
*** puranamr has joined #openstack-infra18:01
crinkleSpamapS: the puppet-openstack team maintains it since it's used a lot with puppet-neutron so we pretend it's openstack18:01
*** mestery has joined #openstack-infra18:01
SpamapScrinkle: I'm good with pretending.18:02
*** dannywilson has quit IRC18:02
* crinkle wonders if the issue is also fixed in openstack/puppet-vswitch18:03
SpamapScrinkle: we'll see. :)18:03
Kiallfungi: right, I'll take the blame so ;)18:03
*** bknudson has joined #openstack-infra18:03
*** dannywilson has joined #openstack-infra18:03
fungimadorn: sounds like someone has duplicate accounts. what's the user you're seeing this with?18:03
madornfungi: thanks. its mine, madorn.18:04
crinkleSpamapS: looks like it is
openstackgerritMerged openstack-infra/project-config: update cue project list in rally gate job
madornfungi:  i have one launchpad account, one gerrit account. one email. but for some reason i can see: and  If i add madorn@gmail, it won't work.18:05
*** ivar-laz_ has joined #openstack-infra18:05
madornfungi: if i add, it works - and sees that user as
*** ivar-laz_ has quit IRC18:05
*** spzala_ has joined #openstack-infra18:06
fungimadorn: yep. i'm poking at your account in the backend database. bear with me18:06
madornfungi: thank you.18:06
openstackgerritArmando Migliaccio proposed openstack-infra/devstack-gate: Drop these config switches as they are no longer being used
*** ivar-laz_ has joined #openstack-infra18:06
openstackgerritMerged openstack/requirements: Ironicclient minimum is too low for nova virt
*** yamamoto has quit IRC18:08
openstackgerritMerged openstack-infra/project-config: Remove non-voting gate-horizon-dsvm-integration from gate queue
j^2hey everyone, so i’m trying to build an openstack cluster with chef on infra: and i assume that eth1 is there, it seems that some of the experimental queue don’t always have eth1, is there anyway to make sure it’s always there?18:08
*** ivar-lazzaro has quit IRC18:08
*** spzala has quit IRC18:08
*** spzala_ is now known as spzala18:08
openstackgerritMerged openstack-infra/project-config: Experiment with an increase time for gate-trove-functional job
openstackgerritMerged openstack-infra/project-config: Create branch permission for networking-bagpipe-l2
openstackgerritMerged openstack-infra/project-config: Add tacker-horizon to gerritt IRC notifications
openstackgerritMerged openstack-infra/project-config: Add openstack-manuals to populate-zanata-project
openstackgerritMerged openstack-infra/project-config: Move Midonet v2 job to the experimental queue
mordredj^2: eth1 is only there on half of our clouds18:10
openstackgerritMerged openstack-infra/project-config: Update ACLs for Neutron sub-projects
mordredj^2: it's safer to assume it does not exist18:10
j^2mordred: so it seems, is there a NIC i can use for OVS testing?18:11
fungimadorn: yep, there's a second account id (17218)18:11
*** sridhar_ram has joined #openstack-infra18:11
openstackgerritMerged openstack-infra/project-config: Add stable ACLs for Neutron sub-projects
madornfungi: where did it come from?18:11
openstackgerritMerged openstack-infra/project-config: Add a create ACL for networking-bgpvpn
openstackgerritMerged openstack-infra/project-config: Update merge commit gerrit ACL for Neutron repositories
*** ayoung has joined #openstack-infra18:11
fungimadorn: the second account has username mdorn rather than madorn and a different http password18:11
openstackgerritMerged openstack-infra/project-config: puppet & tripleo: use devstack-logs publisher for logs
mordredj^2: eth0 is the only nic that exists everywhere we have ... clarkb may know more though since he did the multi-node setup stuff18:11
krotschecksdague: Thjanks for the +2a18:11
sdaguekrotscheck: no prob18:12
*** jsavak has quit IRC18:12
mordredsdague: speaking of +2s...18:12
*** sridhar_ram1 has quit IRC18:12
krotschecksdague: Of course, now I need people to look at this Infra Spec :)
*** resker has joined #openstack-infra18:12
j^2mordred: yeah multinode is the eventual testing we’re doing but i’m trying to get AIO to be consistant first, clarkb can you help here?18:12
*** puranamr has quit IRC18:12
mordredj^2: totally - and the right choice for sure!18:12
openstackgerritSpencer Krum proposed openstack-infra/puppet-os_client_config: Manage OSCC config file with Puppet
madornfungi:  one sec checking something, thanks.18:12
openstackgerritMerged openstack-infra/project-config: specs: fix puppet specs URL
openstackgerritMerged openstack-infra/project-config: Add tests for bifrost
mordredsdague: on the starter-kit patch - I looked through for a place to add in mention to provider-network-linuxbridge but didnt' see an obvious place18:13
mordredsdague: any suggestions?18:13
*** jsavak has joined #openstack-infra18:13
j^2mordred: it’s not far, i can taste it, it’s just figuring out the correct incantation :P18:13
fungimadorn: it looks like the later account (created yesterday) has filed contact information but the older account has not, so chances are you've not been submitting patches under the older account18:13
sdaguemordred: so... some of the context and additional direction from the base patch was lost in jogo's fixups18:13
*** puranamr has joined #openstack-infra18:13
mordredthat may be why I don't see a place like what you mentioned18:13
openstackgerritMerged openstack-infra/project-config: Changed the gerrit acl of stackforge/rack
openstackgerritMerged openstack-infra/project-config: Publish docs for networking-ovn.
mordredjogo: causing trouble still I see18:14
madornfungi:  i've only submitted one bug thus far and one test commit.  based on that, how do you suggest i proceed if i'd like to keep the madorn user account18:14
openstackgerritMerged openstack-infra/project-config: Make gate-os-brick-python34 voting and add gate
fungimadorn: i can set the old account inactive to stop it showing up as an available reviewer or group member in typeahead completion18:14
mordredsdague: hrm. I've got an idea ...18:15
openstackgerritSpencer Krum proposed openstack-infra/puppet-os_client_config: Manage OSCC config file with Puppet
openstackgerritSpencer Krum proposed openstack-infra/puppet-os_client_config: Module boilerplate
fungimadorn: oh, got it. so you want the madorn account and not the mdorn account. if i switch your current openid url to that older one, you'll need to go back through cla/contact info steps i suspect18:15
sdagueso I'm trying to think how to naturally put some additional context back in there18:15
fungimadorn: or i can move the madorn username to the newer account18:16
mordredsdague: lemme send you a patch idea - 2 minutes18:16
madornfungi:  whatever you think is most appropriate.  my only concern is keeping the madorn username so its consistent with irc, email, etc.18:16
*** dprince has quit IRC18:16
*** esker has quit IRC18:16
fungimadorn: though things like review history won't move. so if you had open changes or group memberships or anything those will be orphaned18:17
fungimadorn: well, the only one who should normally see that username anywhere is you. it's just used in your ssh/http configuration as authentication credentials. but since i have to get rid of one anyway it's easiest enough to keep the one you want18:18
openstackgerritK Jonathan Harker proposed openstack-infra/puppet-kibana: Add kibana3 support
nibalizermordred: can you see the p-oscc patches up there ^18:18
*** dannywilson has quit IRC18:18
*** xyang1 has quit IRC18:19
madornfungi: is there any activity associated with the old account?18:20
fungimadorn: hard to say. activity is scattered all over a bunch of different tables18:20
madornfungi: i see. can you remove the old account?18:21
*** abregman has joined #openstack-infra18:21
*** ZZelle_ has joined #openstack-infra18:22
nibalizergerrit breaking?18:22
*** boris-42 has quit IRC18:22
fungimadorn: just finished doing that, yes (well, removing/moving the external id rows for it and setting the account to inactive status). you may need to log out and back into gerrit to be sure you're authenticated as the active account18:22
madornfungi: thank you. for the new account created yesterday..did you keep the same username? mdorn?18:23
funginibalizer: no, it just gets slow to respond when the flush-caches api call is issued, which i needed to do after making account adjustments in the database18:23
nibalizerah okay18:23
*** pc_m is now known as pc_m_AWAY18:23
fungimadorn: you said you wanted to keep the madorn username rather than mdorn18:23
*** bexelbie has quit IRC18:23
madornfungi: correct, thans.18:24
fungimadorn: you should be able to confirm it's correct at
madornchecking now18:24
fungiwhen you're logged back in18:24
*** jtomasek has quit IRC18:25
*** resker has quit IRC18:25
madornfungi: it works - thanks a lot for your help.  i appreciate it.18:27
fungimadorn: you're welcome. as for activity on the old account, it looks like you only had two test reviews on which you left comments,n,z and no reviews for which you were an owner,n,z18:28
jogomordred sdague: what did I do?18:29
*** doug-fish has quit IRC18:29
clarkbmordred j^2 rax has eth1 and eth0 but you completely ignore that for multinode testing18:29
jogoahh the governance patch thing18:29
clarkbinstead we make a l2 overlay using ovs and gre18:29
fungimadorn: based on,n,z i think you were using account 11442 (madorn) in the web interface and 17218 (mdorn) for ssh to upload your test change18:29
mordredjogo: I think sdague and I just figured it all out18:30
clarkband use those interfaces18:30
j^2clarkb: ah damn, ok, is there any chance i can get what i would like done: IE have a eth0 for network access and eth1 for OVS?18:30
clarkbbecause hpcloud has no eth118:30
clarkbso your only option is n overlay18:30
madornfungi: cool, good to know18:30
mordredjogo: and the one before it - let me know what you think of the language18:30
j^2clarkb: cool, i’ll look into that then, do you have any examples to point in me the direction?18:31
j^2rock on thanks!18:32
SpamapSgrub issues now18:32
morganfainbergmordred: I'll need to look more closely at domains.get but that iirc is meant to be admin.18:33
mordredmorganfainberg: ok. so - basically, the entire domain api is intended to be admin18:33
morganfainbergYeah. Afaict18:34
morganfainbergMostly for security reasons.18:34
mordredmorganfainberg: raisins are tasty18:34
*** bswartz has quit IRC18:34
openstackgerritMonty Taylor proposed openstack-infra/shade: Rework how we get domains
mordredmorganfainberg: there you go then ^^ ( Shrews, SpamapS ^^ )18:35
*** kushal has quit IRC18:35
*** jbrc has quit IRC18:35
jogomordred: looks fine to me18:35
*** dustins has quit IRC18:36
clarkbj^2 the nice thing about this is yoi can test the network you want to test regardless of underlying env18:36
j^2i think i get that18:36
*** rossella_s has quit IRC18:37
j^2i’m still trying to grok it18:37
*** rossella_s has joined #openstack-infra18:37
reedmordred, I'm reading legal, I don't understand how shade uses ansible-modules-core18:37
SpamapSmordred: you have raisins?18:37
*** Sukhdev has quit IRC18:38
clarkbj^2 it sets up a hub and spoke switch where you do spoke - ovs - gre - ovs hub - gre - ovs - spoke18:38
openstackgerritArmando Migliaccio proposed openstack-infra/project-config: Drop unstable configuration for tempest-dsvm-neutron job
clarkbthats the linear representation18:38
j^2clarkb: ahh nice18:38
openstackgerritDavid Shrewsbury proposed openstack-infra/shade: Add Ansible module test for networks
mordredreed: it does not18:38
mordredreed: ansible-modules-core uses shade18:38
j^2so if i add this to my bootstrap script, before anytihng i do this’ll give me the device i want18:39
mordredreed: all of the new ansible openstack modules are based on shade18:39
reedohhhh, that's neat18:39
openstackgerritArmando Migliaccio proposed openstack-infra/devstack-gate: Drop these config switches as they are no longer being used
*** esker has joined #openstack-infra18:39
mordredreed: we're hoping to have the full set of them landed before the upcoming ansible 2.0 release18:39
*** dustins has joined #openstack-infra18:39
fungimordred: reed: so gplv3 ansible modules consuming aslv2 shade library18:40
clarkbj^2: ya you have to run it on all the nodes and X times for the number of interfaces you want18:40
mordredfungi: yes18:40
reedok, now I have a better picture18:40
clarkbj^2: but once you do that you can have arbitrary numbers of interfaces until the soft swith falls over due to load18:40
reedthe GPLv3 is so misunderstood, it makes me sad18:40
j^2clarkb: rock on that is exactaly what i’m looking for18:41
fungimordred: reed: and i guess kolla wants to use (bundle?) some of those gplv3 ansible modules?18:41
sdakefungi we decided we just rolled our own module18:41
*** dprince has joined #openstack-infra18:41
sdakefungi but originally that was the plan yes18:41
openstackgerritMerged openstack-infra/project-config: Make python34 whitespace consistent with others
mordredfungi: the issues is that they are not released yet18:42
reedthe message by sdake says "We use a set of libraries only available in Ansible v2.0.  Shade [2]  ASL2.0 uses these Ansible python libraries [3] licensed under a  GPLv3 license.  We use Shade in Kolla directly"18:42
mordredfungi: so there is no good way for them to install them18:42
sdakereed perhaps I got the order wrong18:42
sdakeI don't actually write this code, so I dont know precisely how it works - so take with a grain of sand - ask the experts ;)18:42
reedsdake, looks like it :) I got the correction18:42
sdakei copied the appropriate people (mordred) specificially on teh mail to make corrections whwere I said something incorrect18:43
sdakeso again, apologies mordred if I mistated something18:43
*** otter768 has joined #openstack-infra18:43
fungisort of unfortunate for a probably-not-legally-problematic license difference to cause you to reinvent that wheel18:43
pabelangerfungi, how often does JJB run these days?18:43
reedsdake, just do me a favor and don't use the word 'contamination' when referring to GPL :) 10 hairs fall off my head every time18:43
*** doug-fish has joined #openstack-infra18:44
sdakewhich word would you prefer ;)18:44
fungipabelanger: it's run by puppet, so as often as puppet applies. generally every 30-60 minutes18:44
*** HeOS has quit IRC18:44
reedsdake, persistence, for example, or just 'provisions'. The GPL doesn't contaminate anything18:44
sdakereed ack noted18:44
openstackgerritSpencer Krum proposed openstack-infra/system-config: Test the apply test
* reed goes combing the few hair left :)18:45
pabelangerfungi, how do we insure all the jenkin master say in sync for the JJB jobs? I assume each master runs JJB locally, I can only assume there is some puppet drift on executing18:45
sdakereed i'd loan you some but i'm low on hair myself :)18:45
fungipabelanger: they're eventually consistent"18:45
sdakefungi yes we had to reinevent there - its unfortunate but hopefully we can sort it out once the 2.0 mdoulez are actually out18:46
pabelangerfungi, /me nods18:46
sdakedamn keyboard18:46
fungipabelanger: we also have a pretty wide race where job name changes/additions wind up taking effect in zuul before they take effect on the jenkins masters because jjb takes a while to sift through our body of (now rougly 6500) jobs18:46
*** jamielennox is now known as jamielennox|away18:47
clarkbmordred: could/should ansible make a prerelease on pypi if people want to consume things from them this way?18:48
*** otter768 has quit IRC18:48
mordredclarkb: I can ask18:48
reedsdake, so Kolla needs to import shade or use the ansible modules?18:49
*** mrunge has joined #openstack-infra18:49
openstackgerritMerged openstack-infra/project-config: Make gate-sqlalchemy-migrate-python34 voting
sdakereed i am not sure and the perso nthat nows, samyaple, is alseep atm18:50
reedno problem :)18:50
sdakereed whatever it is, it is the same model as osad was using according to sam - which was his big annoyance with not using shade directly because he didn't want to take design shortcuts18:50
openstackgerritMerged openstack-infra/project-config: Add jhesketh to accessbot
sdakebut it is what it is18:50
openstackgerritMerged openstack-infra/project-config: Add python3 jobs for neutron-lbaas
openstackgerritMerged openstack-infra/project-config: networking-ovn: Drop custom services list
openstackgerritMerged openstack-infra/project-config: Add python3-jobs for networking-midonet
openstackgerritMerged openstack-infra/project-config: Added eslint-config-openstack
reedsdake, what's osad?18:51
*** sridhar_ram has quit IRC18:51
openstackgerritMerged openstack-infra/project-config: Make gate-networking-ofagent-python34 non-voting
openstackgerritMerged openstack-infra/project-config: OpenStack Foundation Transparency Policy project
*** patrickeast_ has joined #openstack-infra18:51
openstackgerritMerged openstack-infra/project-config: Switch jobs from old pypi to new release worker
clarkbopenstack something soemthing seomthing18:51
openstackgerritMerged openstack-infra/project-config: Restrict stable branches approval on ironic-inspector
fungireed: apparently "openstack ansible deployment"18:51
reedthanks fungi18:52
TheJuliarcarrillocruz: So I think I just ran into the same issue you ran into with simple-init only two nics in an ubuntu :)18:52
sdakereed osad is openstack ansible deploy18:52
sdakethey use a similar model where they use modules to control ansible iirc18:53
*** baoli has quit IRC18:53
*** sridhar_ram has joined #openstack-infra18:54
*** puranamr has quit IRC18:54
*** jamielennox|away is now known as jamielennox18:54
jeblairmordred: can you join #openstack-sprint ?18:55
clarkboh right the spring is happening18:56
*** sigmavirus24_awa is now known as sigmavirus2418:56
*** heyongli has quit IRC18:57
*** baoli has joined #openstack-infra18:58
*** heyongli has joined #openstack-infra18:59
*** gordc has quit IRC18:59
*** pc_m has joined #openstack-infra19:01
pabelangerokay, I was hoping to get some help with the zuul_swift_upload script.  I added a chance recently to puppet-apply-jobs to also include the applytest folder however, I am not seeing the folder under apache.19:03
pabelangerHere is the jenkins jobs calling the script:
pabelangerhowever, only lists the console file19:04
pabelangeram I missing something?19:04
*** Sukhdev has joined #openstack-infra19:04
*** Sherif has joined #openstack-infra19:04
*** heyongli has quit IRC19:06
*** maishsk has quit IRC19:06
*** heyongli has joined #openstack-infra19:06
openstackgerritMichael Krotscheck proposed openstack-infra/system-config: Added NPM Publishing credentials to pypi slave.
*** maishsk has joined #openstack-infra19:08
*** yamamoto has joined #openstack-infra19:08
*** Sherif is now known as sherif19:08
*** e0ne has joined #openstack-infra19:09
*** odyssey4me_ has joined #openstack-infra19:10
openstackgerritMichael Krotscheck proposed openstack-infra/system-config: Added NPM Publishing credentials to pypi slave.
*** odyssey4me has quit IRC19:10
*** bhunter71 has quit IRC19:11
*** heyongli has quit IRC19:11
*** achanda has quit IRC19:11
krotscheckCan someone point me at the code that checks to see whether openstackci is allowed to publish a particular pypi package?19:12
* krotscheck needs to implement something similar for npm19:12
krotscheckIt's somewhere in the layout job, I think :/19:12
*** heyongli has joined #openstack-infra19:13
*** jsavak has quit IRC19:13
*** bitblt has joined #openstack-infra19:13
*** yamamoto has quit IRC19:13
*** coolsvap|away is now known as coolsvap19:14
*** odyssey4me_ is now known as odyssey4me19:14
anteayakrotscheck: I'm not seeing it publish anything19:14
*** jsavak has joined #openstack-infra19:14
anteayaand I don't know about allowed to publish19:15
*** Rockyg has joined #openstack-infra19:15
*** puranamr has joined #openstack-infra19:15
*** fawadkhaliq has quit IRC19:15
anteayaI check if a repo has name reserved on pypi before I +2 a patch with a publish to pypi job on it19:15
*** puranamr has quit IRC19:15
anteayaa new publish to pypi job that is19:15
*** puranamr has joined #openstack-infra19:15
anteayanot sure if that gives you what you want or not19:16
*** amitgandhinz has quit IRC19:16
krotscheckanteaya: I was under the impression that something checks to make sure the openstackci user has the permissions to publish a given package.19:16
*** bswartz has joined #openstack-infra19:16
anteayaI don't think I fully understand what you are thinking19:16
anteayaif there is I don't know what it would be19:16
*** fawadkhaliq has joined #openstack-infra19:16
krotscheckanteaya: Ok, so if I add an IRC channel, but don't add the openstack bot to the admin list first, the build fails.19:16
anteayaso I'm uncertain how you managed to get that impression19:17
anteayaoh that would be a test19:17
anteayawe don't have a test that checks pypi to see if the name of the repo is registered on pypi19:17
krotscheckanteaya: I understand that pypi tests do something different.19:17
krotscheckBut similar.19:17
anteayawe have pypi tests?19:17
clarkbno all the pypi jobs do is grab tarball and upload19:17
clarkbif the perms are not correct it will fail19:17
*** puranamr has quit IRC19:17
krotscheckclarkb: Allllrighty.19:18
krotscheckclarkb: But wouldn't it be nice if we didn't allow new projects to push to pypi unless we can detect that the user has perms?19:18
clarkbI mean we already don't allow it by virtue of pypi rejecting it19:19
*** jsavak has quit IRC19:19
*** sdake_ has joined #openstack-infra19:19
*** heyongli has quit IRC19:19
*** jsavak has joined #openstack-infra19:20
anteayamorning lifeless19:21
lifelesszigo: yes. 'make'. (testtools tests)19:21
openstackgerritDavid Shrewsbury proposed openstack-infra/shade: Add Ansible module test for subnet
rcarrillocruzTheJulia: hah19:21
krotscheckclarkb: Alright then, less work for me.19:21
rcarrillocruzi need to dig a bit still19:21
rcarrillocruzi have all my nics configured  with the respawn19:21
rcarrillocruzbut for some reason the keys are being injects19:22
TheJuliarcarrillocruz: also, on ubuntu, I see what you meant where cloudinit was running completely19:22
rcarrillocruzeven though I see from dmesg output that is indeed executed19:22
lifelessAJaeger: / sdague: I would expect folk to run it by hand. This is all going into README in the requirements repo19:22
*** sdake has quit IRC19:23
*** r-daneel has joined #openstack-infra19:23
*** heyongli has joined #openstack-infra19:25
*** sdake has joined #openstack-infra19:26
openstackgerritPaul Belanger proposed openstack-infra/system-config: WIP: Migrate to puppet-httpd module
lifelessfungi: AJaeger: freerunner: yes - I want to run that job more frequently than daily; and autoapprove it when it works. We're not there yet.19:26
lifelesszigo: linecache2 - yes, thats cosmetic, doesn't affect anything. Its test data that can't compile is all.19:27
*** fallenpegasus has quit IRC19:27
*** jcoufal_ has joined #openstack-infra19:28
*** _nadya_ has joined #openstack-infra19:29
*** esker has quit IRC19:29
*** nfedotov has joined #openstack-infra19:29
*** esker has joined #openstack-infra19:29
*** reazem has joined #openstack-infra19:30
*** sdake_ has quit IRC19:30
openstackgerritMerged openstack-infra/project-config: Cleanup proposal slave workspaces once a day after builds
*** woodster_ has quit IRC19:31
openstackgerritMerged openstack-infra/project-config: Use in-tree devstack plugin to fix rally gates
*** heyongli has quit IRC19:33
*** ddieterly has joined #openstack-infra19:33
*** puranamr has joined #openstack-infra19:33
*** patrickeast_ has quit IRC19:33
*** sherif has quit IRC19:34
*** zul has quit IRC19:34
*** amitgandhinz has joined #openstack-infra19:34
openstackgerritPaul Belanger proposed openstack-infra/system-config: Display puppet apply results right after running
*** heyongli has joined #openstack-infra19:37
*** changbl has joined #openstack-infra19:37
pabelangernibalizer, ^ there is a case where the cat statement is not called properly19:38
pabelangerthat should fix it19:38
pabelanger for example19:38
pabelangernot sure why, but something is killing the build step19:39
*** cdent has quit IRC19:39
*** linkedinyou has joined #openstack-infra19:40
*** _nadya_ has quit IRC19:41
openstackgerritMerged openstack-infra/devstack-gate: Add openvswitch logs to output
openstackgerritMichael Krotscheck proposed openstack-infra/project-config: Added npm publish job.
*** sherif has joined #openstack-infra19:41
nibalizerpabelanger: commented19:41
*** _nadya_ has joined #openstack-infra19:41
openstackgerritMerged openstack-infra/project-config: Move gate-keystone-python34 to check
*** flepied has joined #openstack-infra19:42
openstackgerritMichael Krotscheck proposed openstack-infra/system-config: Added NPM Publishing credentials to pypi slave.
*** unicell1 has joined #openstack-infra19:44
*** unicell has quit IRC19:45
nibalizerpabelanger: so whats the issue? I don't understand19:45
pabelangernibalizer, look at the logs.o.o URL above19:45
nibalizeroh is it because of set -x?19:45
nibalizerer set -e19:45
pabelangerit is because we are calling exit19:46
nibalizerso it detects that something went wrong then fails?19:46
*** sherif has quit IRC19:46
pabelangerwhich I think is killing the shell19:46
pabelangerand for some reason, your cat call is not executing19:46
*** Volundr has joined #openstack-infra19:46
*** _nadya_ has quit IRC19:46
*** whoops has quit IRC19:46
openstackgerritMichael Krotscheck proposed openstack-infra/project-config: Added npm publish job.
*** Volundr is now known as whoops19:46
*** e0ne has quit IRC19:47
nibalizerso that sounds like the expected behavior of set -e to me19:48
nibalizerso maybe turn of set -e then turn it back on?19:48
clarkbor do || true19:48
*** sherif has joined #openstack-infra19:50
lifelessttx: around ?19:50
*** rbrooker has joined #openstack-infra19:50
*** fawadkhaliq has quit IRC19:51
openstackgerritMichael Krotscheck proposed openstack-infra/project-config: Added --silent flag to npm run template.
krotscheckThere's some manual acl things that need to happen for , anyone here able to help out now that it's merged?19:53
pabelangernibalizer, you want to disable the -x for the puppet apply command?19:53
nibalizerthe -x isn't the problem19:53
nibalizerthat just tells the shell to print out what it is doing19:54
pabelangersorry, not following the logic of toggling it19:54
pabelangervs adding a build job to simple render the logs post execute, along side the upload of them19:54
krotscheck(There's comments from mtreinish there)19:55
nibalizerwell hrm19:55
*** HeOS has joined #openstack-infra19:56
nibalizerya so if we set +e that will turn of the failures and allow the cat to run, but then the test wont fail where it should19:56
nibalizerya where are you at with the publisher to just grab the right things19:56
*** shashank_hegde has quit IRC19:56
nibalizerthat is the future really anyways19:56
clarkbuse || true19:56
nibalizerclarkb: will that let the test fail correctly?19:56
clarkbif cat is expected to fail then yes19:56
pabelangerclarkb, do you know why a swift upload would not render a folder upload correctly under apache?19:56
nibalizerclarkb: so the problem is this line fails:
*** baoli has quit IRC19:57
nibalizerwhich causes the shell to exit because of set -e which prevents the cat from running19:57
*** coolsvap is now known as coolsvap|away19:58
*** crc32 has joined #openstack-infra19:58
clarkband that fail is acceptavle?19:58
clarkbif so use || true19:58
nibalizerclarkb: I dont think that failure is acceptable, because thats how the test knows if the puppet code is good or not19:59
clarkbthen dont change anything19:59
pabelangerYa, I don't think we want to manually trap errors by toggling +e -e. Seems like it could get tricky fast20:00
pabelangereven moving the cat $filename into test_puppet_apply isn't going to work in all cases.  Since if puppet apply directly does the exit, then everthing after is going to bomb20:01
pabelangerbasically that is what is happening now20:01
openstackgerritMerged openstack-infra/shade: Add Ansible module test for networks
*** kutija_ has joined #openstack-infra20:02
*** shashank_hegde has joined #openstack-infra20:03
*** baoli has joined #openstack-infra20:04
openstackgerritPaul Belanger proposed openstack-infra/system-config: Display puppet apply results right after running
pabelangernibalizer, removing the -P on xargs for now20:04
clarkboh dont do that20:04
pabelangerwonder if puppet is freaking out of the parallel process or something20:04
clarkbthe job will likely timeout20:04
pabelangerclarkb, k20:04
nibalizerya the parallism isnt the problem20:05
pabelangerk, never notice we did parallel before20:05
*** kutija has quit IRC20:05
nibalizerpabelanger: so your patch to add a publisher to the apply-test, where is that?20:06
nibalizerI can't find it20:06
pabelangerit is actually merged20:06
pabelangerbut logs are not rendered properly under apache20:06
pabelangersee scroll back for me asking for some help20:06
nibalizerah thats whats going on20:07
clarkboh did you add a seci d publisher?20:07
clarkbyou can only do one with swift20:07
clarkbotherwise the index generation doesnt have a full picture20:07
pabelangerseci d publisher?20:07
*** gyee has joined #openstack-infra20:07
clarkbfor whatever publishibg you did20:07
*** bhunter71 has joined #openstack-infra20:08
clarkbis it a second publisher in jjb?20:08
pabelangerclarkb, I simply change zuul-swift-upload-console-log to zuul-swift-upload-logs-with-console20:08
pabelangerbut, I guess I missed somethign?20:09
*** coolsvap|away is now known as coolsvap20:09
*** claudiub has joined #openstack-infra20:09
pabelangerbut don't have logs into the swift publish script to see what is happening20:10
*** openstackgerrit has quit IRC20:10
*** amrith is now known as _amrith_20:10
clarkbyou should20:10
*** pc_m has quit IRC20:10
clarkbvia the jenkins master20:10
*** openstackgerrit has joined #openstack-infra20:10
*** pc_m has joined #openstack-infra20:10
*** shashank_hegde has quit IRC20:10
*** pc_m has quit IRC20:11
pabelangerk let me check that20:11
*** dahoo has joined #openstack-infra20:11
pabelangerI see:
clarkbya that20:11
pabelangerwhich calls zuul_swift_upload20:11
*** rbrooker has quit IRC20:11
pabelangerbut, no logs from that20:11
*** e0ne has joined #openstack-infra20:11
pabelanger is folder20:11
pabelangerbut like you mention, index is not generated correctly20:11
*** ivar-laz_ has quit IRC20:11
*** e0ne has quit IRC20:12
*** ivar-lazzaro has joined #openstack-infra20:12
clarkbwe can infer from the warbings it only uplaoded two files20:13
clarkbibdex and console20:13
clarkbso your glob likely does not match20:13
pabelangerokay, let me check on that20:13
pabelangerand retest locally20:13
rcarrillocruzTheJulia: can you confirm if you get your ssh key injected on root's authorized_keys with simple_init?20:14
rcarrillocruznot sure what's going on20:14
rcarrillocruzi've set 'set -eux' on and confirms that on boot it's run 'glean --ssh --hostname --skip-network'20:15
rcarrillocruzhowever, auhtorized_keys on /root is void20:15
rcarrillocruzif i run that very same command on root's shell, the pub key is injected on root's authorized_keys20:15
openstackgerritMichael Krotscheck proposed openstack-infra/project-config: Added npm publish to eslint-config-openstack
mordredrcarrillocruz: wow. you get all the fun!20:16
rcarrillocruzfirst nics20:16
rcarrillocruznow ssh keys20:16
openstackgerritSpencer Krum proposed openstack-infra/project-config: Dump the output of the apply test
mordredrcarrillocruz: is there something _else_ running that's deleting the key maybe?20:16
rcarrillocruzcomputers hate me20:16
mordredrcarrillocruz: computers are terrible20:16
nibalizerpabelanger: clarkb thats a simple workaround ^^20:16
pabelangernibalizer, right, thats what I was considering too20:17
pabelangerand just remove the dump of the logs from the apply.sh20:17
*** kutija has joined #openstack-infra20:17
rcarrillocruzmordred: no that i'm aware :/20:18
pabelangernibalizer, path issue20:18
nibalizerah yep, fixing20:18
*** ari has quit IRC20:18
clarkbmordred cloud-init20:19
clarkbI am half serious20:19
openstackgerritSpencer Krum proposed openstack-infra/project-config: Dump the output of the apply test
TheJuliarcarrillocruz: I'll need to check tomorrow, i need to call it a day shortly20:19
mordredclarkb: yah. me too20:19
*** maishsk has quit IRC20:19
*** puranamr has quit IRC20:19
mordredrcarrillocruz: you using the ubuntu-minimal element?20:19
openstackgerritPaul Belanger proposed openstack-infra/project-config: Use full path to applytest folder
rcarrillocruzsure, enjoy20:19
rcarrillocruzmordred: vm simple-init devuser20:19
mordredrcarrillocruz: what is before vm ?20:20
pabelangernibalizer, clarkb that should fix swift upload ^20:20
dstanekhow can i unstick ?20:20
pabelangermight you, not a fan of full path now20:20
*** kutija_ has quit IRC20:20
*** heyongli has quit IRC20:20
rcarrillocruzah k20:21
*** baoli has quit IRC20:21
rcarrillocruzit uses a playbook default20:21
fungidstanek: talk to the guy who proposed and get it working first?20:21
rcarrillocruz'ubuntu' element20:21
rcarrillocruzmordred: ^20:21
pabelangernibalizer, hopefully we can get some +1 workflow action on that :)20:21
*** hashar has joined #openstack-infra20:21
*** maishsk has joined #openstack-infra20:21
*** baoli has joined #openstack-infra20:21
fungidstanek: or rebase 176434 to no longer depend on 17455620:21
mordredrcarrillocruz: yup. that's your problem20:21
mordredrcarrillocruz: dont' use that20:22
mordredrcarrillocruz: use ubuntu-minimal20:22
nibalizerpabelanger: haha did fixing my bug show you your bug? thats awesome20:22
clarkbpabelanger is openstack-infra in the path?20:22
dstanekfungi: opps duh, ok thx20:22
*** dustins has quit IRC20:23
*** tonytan4ever has quit IRC20:23
rcarrillocruzk, let me override the element then, thx20:23
*** woodster_ has joined #openstack-infra20:23
rcarrillocruzmordred: out of curiosity, what is wrong with the ubuntu element that messes with keys, do you know?20:23
clarkbit rubs cloud init20:24
openstackgerritMerged openstack-infra/shade: Add Ansible module test for subnet
*** heyongli has joined #openstack-infra20:24
pabelangerclarkb, I don't think so. We explicitly cd into openstack-infra/system-config to run apply-tst20:24
mordredrcarrillocruz: what clarkb said20:24
mordredrcarrillocruz: glean/simple-init is designed to work with ubuntu-minimal - not with ubuntu20:25
*** dizquierdo has joined #openstack-infra20:25
pabelangerclarkb, to be honest, I would have to see what the jenkins workarea looks like again20:25
rcarrillocruzi asked that same question in ironic channel20:25
rcarrillocruzwas wondering if getting cloud-init running was normal20:25
rcarrillocruzsince simple-init is supposedly replacing it20:25
mordredrcarrillocruz, TheJulia: I strongly recommend updating bifrost to reference ubuntu-minimal and not ubuntu if it's making use of glean's support for IPs in config drive20:26
mordredrcarrillocruz, TheJulia: strongly as in "using the ubuntu element is 100% unsupported and WILL break"20:26
rcarrillocruzmordred: to be fair, it doesn't use simple-init by default20:26
*** geoffarnold has joined #openstack-infra20:26
*** jcoufal_ has quit IRC20:27
rcarrillocruzi think it would be good to put a comment saying 'hey, you wanna use simple-init? make sure you switch to ubuntu-minimal'20:27
rcarrillocruzi can do that change20:27
mordredrcarrillocruz: ah! cool20:27
rcarrillocruzotoh, i think it would be good for bifrost to switch to simple-init altogether20:27
rcarrillocruzbut i leave that to TheJulia20:27
rcarrillocruzi love its simplicity20:28
rcarrillocruzthat it20:28
rcarrillocruzall the stuff you need you just use ssh/ansible to bootstrap20:28
rcarrillocruzno need for cloud-config stuff20:28
TheJuliarcarrillocruz: thats on my list for very soon... like on the list sitting in front of me :)20:28
openstackgerritSpencer Krum proposed openstack-infra/system-config: Revert "Print the node definition before apply test"
rcarrillocruzok then , i'll hold on and await for new release :-)20:29
*** crc32 has quit IRC20:29
mordredrcarrillocruz: that's excatly right20:30
mordredrcarrillocruz: if you're going to ansible - it's all you need20:30
*** baoli has quit IRC20:32
*** baoli has joined #openstack-infra20:33
*** baoli has quit IRC20:33
*** bswartz has quit IRC20:33
*** marzif_ has quit IRC20:34
*** baoli has joined #openstack-infra20:35
openstackgerritPaul Belanger proposed openstack-infra/system-config: Display puppet apply results right after running
nfedotovHello zuul operators!20:36
openstackgerritMerged openstack-infra/system-config: Script to install modules during integration
*** _amrith_ is now known as amrith20:38
*** rbrooker has joined #openstack-infra20:38
*** crc32 has joined #openstack-infra20:38
*** kgiusti has left #openstack-infra20:39
nfedotovDid somebody face memory leaks in zuul ?20:39
greghaynesmordred: I thought we disabled cobfigdrive datasource for cloud init whe  you use glean?20:39
greghaynesagreed that users should be pointed at ubuntu-minimal reguardless20:39
clarkbno our memory use is pretty constant
*** jasondotstar has joined #openstack-infra20:40
greghayneslooks like I missed some glean fun while I was in flight?20:40
nfedotovclarkb: Oh. +/- 10 Gb is ok20:40
clarkbit is only using about 520:41
clarkband thats constant20:41
*** bdemers has joined #openstack-infra20:41
*** mrunge has quit IRC20:42
*** heyongli has quit IRC20:42
nfedotovOn my machine the Zuul takes 2 Gb in about 24 hours .20:43
*** shashank_hegde has joined #openstack-infra20:43
*** heyongli has joined #openstack-infra20:43
*** sherif has quit IRC20:43
*** claudiub|2 has joined #openstack-infra20:44
*** otter768 has joined #openstack-infra20:44
*** jgriffith has quit IRC20:45
*** claudiub has quit IRC20:45
clarkbgear can leak if you arent processing jobs20:46
clarkbbyt thats the only place I would look off the rop of my head20:46
clarkbzuul does have a profiler feature maybe you can get the info from that?20:47
nfedotovclarkb: Gear is installed on "Jenkins" machine only.20:47
clarkboh we run gear on the zuul node20:48
clarkbeither way should work fine20:48
nfedotovclarkb: Profiler? Could you point me to it? Thanks.20:48
*** otter768 has quit IRC20:49
*** fallenpegasus has joined #openstack-infra20:49
*** shardy has quit IRC20:50
nfedotovThank you! I will take look :-)20:50
*** spzala has quit IRC20:50
greghaynesfungi: you all get the glean thing from earlier sorted?20:52
openstackgerritSpencer Krum proposed openstack-infra/puppet-openstackci: Check resources in the openstackci::loghost class
openstackgerritSpencer Krum proposed openstack-infra/puppet-openstackci: Update spec helper for zuul-cloner
*** fallenpegasus has quit IRC20:54
openstackgerritJames E. Blair proposed openstack-infra/project-config: Use '-' in macro names
*** annegentle has quit IRC20:55
*** radez is now known as radez_g0n320:55
*** tqtran has joined #openstack-infra20:56
tqtranlifeless: question regarding generate-constraints for patch
tqtranwhere is the generate-constraints cmd coming from?20:57
openstackgerritMerged openstack-infra/puppet-openstackci: Add beaker boilerplate to openstackci
*** vnod has joined #openstack-infra20:58
ekarlso /buffer 2420:58
lifelesstqtran: pip install git+
openstackgerritScott Moser proposed openstack-infra/git-review: support urls as change-id input
openstackgerritlifeless proposed openstack-infra/release-tools: Be safer with git refs.
*** jsavak has quit IRC21:04
openstackgerritSpencer Krum proposed openstack-infra/puppet-os_client_config: Manage OSCC config file with Puppet
openstackgerritK Jonathan Harker proposed openstack-infra/system-config: Add jenkins masters credentials to secure.conf
openstackgerritScott Moser proposed openstack-infra/git-review: support urls as change-id input
*** Sukhdev has quit IRC21:04
tqtranlifeless: Collecting Django<1.8,>=1.4.2 (from -r global-requirements.txt (line 29))21:05
tqtran  Using cached Django-1.7.9-py2.py3-none-any.whl21:05
tqtran Command '['/tmp/tmpwTFIem/bin/pip', 'install', '-r', 'global-requirements.txt']' returned non-zero exit status 121:05
tqtranany idea what that could mean?21:05
lifelesstqtran: it failed to install all of global-requirements21:05
fungigreghaynes: not really, no. i've not been usefully around today21:05
*** rfolco has quit IRC21:05
lifelesstqtran: that may mean your change is incompatible, or there may be more output that you have not pasted21:05
*** katyafervent is now known as katyafervent_awa21:06
fungigreghaynes: though it doesn't look (to me) like glean is setting short hostnames on current ubuntu-trusty workers, only fqdns21:06
tqtranim running this on master21:06
tqtranlet me pastebin it21:06
*** abregman has quit IRC21:06
*** achanda has joined #openstack-infra21:07
*** esker has quit IRC21:07
*** dizquierdo has quit IRC21:08
*** jasondotstar has quit IRC21:08
openstackgerritKhai Do proposed openstack-infra/puppet-openstackci: Add Jenkins Job Builder manifest
*** maishsk has quit IRC21:08
*** maishsk has joined #openstack-infra21:09
sdaguefungi: so, out of curiosity, do we know what changed to create these new node names that are so giant?21:10
openstackgerritMerged openstack-infra/system-config: Display puppet apply results right after running
clarkbcould just be node id added another digit21:11
tqtranlooks like is really slow atm21:11
sdagueclarkb: could be, we seem to be overflowing only by 121:11
sdagueclarkb: where is the hostname generation code?21:11
fungisdague: clarkb: yeah, chances are we were already skirting the limit. and the subnode id adds a good chunk of additional data21:11
clarkbsdague in nodepool21:11
*** baoli has quit IRC21:11
clarkbit does follow a configurable template we can change21:12
sdagueok, so short term, can we change that?21:12
clarkbprobably or just make basename shorter21:12
sdaguebecause all multinode is broken right now because of these 65 character hostnames21:12
sdagueoh, well N+1 hostnames :)21:12
fungiright now it's (label-name)-(provider-location)-(node-id)[-(subnode-id)].slave.openstack.org21:12
sdagueok, where is said template?21:13
clarkbwe can trim subnode id maybe?21:13
sdaguehonestly, is label-name really useful?21:13
*** ddieterly has quit IRC21:13
lifelesstqtran: this is the problem:21:13
lifeless  Hash of the package (from (52358ac5f48fd514d6c4eb9e08830985) doesn't match the expected hash 2df76c90453477a9d0bb6a1608f0dfa5!21:13
lifelessBad md5 hash for package (from
clarkbsdague yes?21:14
mordredsdague: it lets us verify the node - we're about to be swapping out the base image and we'll want to make sure we know that it ran on old or new21:14
clarkbit allows you to glob in logstash for example21:14
mordredalso logstash21:14
greghaynesfungi: hrm, did new uploads work?21:14
fungisdague: it lets us know what kind of worker the job ran on. in the past we had some jobs which could run on more than one node type. for example branch-dependent jobs that might run on precise or trusty21:14
mordredgreghaynes: yes21:14
fungigreghaynes: the nodes have glean 1.1.0 on them21:14
sdagueso there is no way to register that information except to treat the hostname as a database?21:15
*** ldnunes has quit IRC21:15
lifelesstqtran: you could try deleting ~/.cache/pip/http21:15
lifelesstqtran: in case its a bad cached copy21:15
tqtranlifeless: ok let me try21:15
lifelesstqtran: (if that fixes it file a bug on pip - pip should be able to self-recover from that)21:15
greghayneswierd. I even wrote tests! ;)21:15
mordredsdague: not quickly. we could implement something different ... but I think that would be a longer process than the other thigns21:15
fungisdague: well, it's sort of being marshalled through jenkins to get into logstash at the moment and hostname is something jenkins actually knows. it might be possible to work around it21:15
mordredyah. I mean - it's code - anything is possible - this just happens to be how it's designed right now21:16
mordredso I think a template hack to get us past today, combined with fixing glean is likely the least amount of work to be solid for the future21:17
sdagueok, so anyway, this issue means that live migration no longer is being tested, and hasn't been for a while21:17
sdagueso some sort of short term, if imperfect fix, would be really nice21:17
clarkbit wad t testd for other reasons befire that21:17
fungileast amount of work would be to stop nodepool, delete all nodes, and reset its counters ;)21:17
clarkbamd no one cared until jogo drove by21:17
sdagueclarkb: it has been working for a while, and people were looking at it21:18
fungibut that doesn't solve the problem, merely buys us some time21:18
clarkbno it hadnt21:18
clarkbit was broken for about a month21:18
*** rlandy has quit IRC21:18
mordredthe real fix is the glean thing ...21:18
mordredgreghaynes: so - earlier when fungi looked in to it21:18
sdagueok, so how far away is the real fix?21:19
mordredgreghaynes: he ran glean by hand ona  held node and it set the hostname right21:19
*** rlandy has joined #openstack-infra21:19
mordredfungi: that's right, right?21:19
greghaynesyeaa, the code for it is really simple so im surprised21:19
mordredoh. wait.21:19
mordredI forgot21:19
mordredit's not glean21:19
greghaynesoh, well there you go :p21:20
mordredmulti-node is running on cloud-init hosts21:20
mordredand we were going to make a patch real quick to move it to ubuntu-trusty21:20
mordredfungi: did you make that patch? or you want me to?21:20
mordredsorry - I totally forgot we walked through this this morning21:20
tqtranlifeless: in progress, its taking a while.....21:21
*** coolsvap is now known as coolsvap|away21:22
*** jgriffith has joined #openstack-infra21:24
openstackgerritMonty Taylor proposed openstack-infra/system-config: Use ubuntu-trusty for multi-node testing
sdaguewell, whatever could be done to kick things forward there would be nice. Because it's currently blocking my grenade work here. And it would be kind of nice to be able to see the state of live migration again.21:24
mordredsdague: working on it21:25
sdaguemordred: thanks21:25
*** aysyd has quit IRC21:25
*** sdake_ has joined #openstack-infra21:25
clarkbmordred does rax provide that image yet?21:25
sdagueok, got to run for the evening to run the local lug meeting21:25
*** jsavak has joined #openstack-infra21:25
clarkbif not need to remove it21:25
*** annegentle has joined #openstack-infra21:25
mordredclarkb: wait - what?21:25
clarkbmordred in that change you assert rax is providing that label21:26
lifelesstqtran: first run will be ~10m21:26
lifelesstqtran: after that its a minute or so21:26
clarkbthis requires swift image uploads abd glean in rax to work and be working21:26
tqtranlifeless: ok didn't work, where do i submit the bug? and is there anything i can do to sidestep this atm?21:26
mordredclarkb: yes.21:26
fungimordred: no, earlier when i looked at it, after you suggested moving all devstack jobs to ubuntu-trusty to fix it, i logged into an ubuntu-trusty node and confirmed that glean hadn't fixed the hostname on them either, so it's not (yet) going to solve this particular issue21:27
tqtranall we're looking for is the diff right? so if i remove that from both master and branch, i can still find out the constraints?21:27
mordredfungi: OH21:27
mordredgreghaynes: ^^21:27
mordredfungi, greghaynes: so both paths need work currently21:27
*** jgriffith has quit IRC21:27
fungilet me dig up my pastes from earlier21:27
mordredwell- amusingly -21:28
mordredchanging it to ubuntu-trusty will fix it21:28
lifelesstqtran: Django 1.7.9 was uploaded today to pypi - 2015-07-0821:28
mordredbecause ubuntu is shorter than devstack :)21:28
mordredby 2 chars21:28
lifelesstqtran: I think its bad data there. Try adding a !=1.7.9 to global-requirements.txt on Django21:28
clarkbits a good switch in general too so we xan move forward21:28
*** sdake has quit IRC21:28
clarkbthen fix glean21:28
openstackgerritMonty Taylor proposed openstack-infra/project-config: Use ubuntu-trusty for multi-node tests
mordredso there's the other bit21:29
fungimordred: greghaynes: compare and
*** e0ne has joined #openstack-infra21:29
tqtranlifeless: running it now21:29
*** DericHorn-HP has joined #openstack-infra21:29
fungimordred: very good point! it squeaks us in under the limit21:29
fungi(for now)21:29
*** reazem has quit IRC21:29
lifelesstqtran: (you don't need to include the Dango change to global-requirements in what you submit, since the periodic job will also fail to pull it in until pypi is fixed21:30
*** fifieldt has quit IRC21:30
tqtranunderstood, but i can't get pass the generate-constraints cmd21:30
tqtranto even get the baseline file to compare to21:31
greghayneslater tonight I can look at the details of whats happening with glea21:31
lifelessyou should be able to with the django 1.7.9 exclude in g-r21:31
openstackgerritJames E. Blair proposed openstack-infra/project-config: Add openstackci beaker jobs to several repos
lifelessI'm going to try a couple runs now21:31
greghayneswhen I finally escape ATL21:31
*** rlandy has quit IRC21:31
fungigreghaynes: there is no escape. if you walk all the way to terminal a it loops back to terminal f again21:31
ClintATL will always be with you21:31
fungigreghaynes: not many know this, but atl is actually a hypertoroidal surface21:32
*** maishsk has quit IRC21:32
krotscheckWhen pushing a release tag, I'm guessing the format matters? For instance, checking oslo,middleware, I see a tag called 2.3.0. Is there anything that would prevent me from using a tag format more like 'v2.3.0'? It'd make javascript things a bit easier.21:32
openstackgerritKhai Do proposed openstack-infra/system-config: Use third party ci manifest to setup jenkins job builder.
lifelesskrotscheck: yes. There's a regex config on tags in gerrit,a nd in the zuul pipelines that run after IIRC21:32
krotschecklifeless: Hrm. Ok, let me see if there's a way for me to override the npm default.21:33
fungikrotscheck: to trigger jobs in the release pipeline the tag name must match the regex at
mordredthe zuul pipelines are the important ones21:33
*** miqui has quit IRC21:33
lifelessif each repo is only one project21:33
lifelessand thus only releasing artifacts in one language21:33
lifelesswe can probably do separate regexs21:33
mordredkrotscheck: now - if there are repos that are javascript and not python and there is a reaon those need to have a leading v - I see no reason we couldn't have a second pipeline ...21:33
mordredwhat lifeless just said21:33
*** heyongli has quit IRC21:34
mordredhowever - it would be neat if we didn't have to diverge21:34
fungikrotscheck: basically (at the moment) one or more digits separated by one or more . and not starting or ending with . and no multiple adjacent ..21:34
*** puranamr has joined #openstack-infra21:34
krotscheckfungi, mordred: we could just add a v? at the beginning.21:34
fungiwe do have a "tag" pipeline which can trigger jobs on any tag regardless of the tag name's format21:34
*** puranamr has quit IRC21:35
mordredkrotscheck: physically? yes. otoh - it adds more "well, you coudl do this, or you could do this" to the mix21:35
fungikrotscheck: well, at the moment this is our only protection against the foot-canon of people tagging python projects with something pbr won't render into a pep-440 compliant version21:35
*** tlian has quit IRC21:35
krotscheckfungi: Yeah, but using that would make the release pipeline a very institutional knowledge thing.21:35
fungikrotscheck: sure. probably a safer solution is to have tooling-specific release pipelines21:36
fungiso an npm-release pipeline and a pypi-release pipeline21:36
fungiand they can have separate regexes21:36
*** pc_m_AWAY is now known as pc_m21:36
mordredfungi: I do like the theory of things working kinda the same21:36
clarkbtge v is annoying21:36
lifelesstqtran: ok, Django 1.7.9 came down ok for me21:36
clarkbmeans you have to parse and remove it for everything else21:36
mordredI mean, requireing a v is as arbitrary as not allowing one ...21:36
lifelesstqtran: it might be your local fastly pop21:37
clarkbexcept that no distro wants v in there21:37
fungiand it's generally easier to see if someone has added a pypi/python-specific job into the npm-release pipeline or vice versa21:37
clarkband its redundant21:37
*** e0ne has quit IRC21:37
mordredit is a redundant character for sure21:37
clarkbbut yes arbitrary21:37
lifelesstqtran: dstufft has fixed it. You may need to delete your http cache again.21:37
tqtranok let me try again21:37
openstackgerritK Jonathan Harker proposed openstack-infra/system-config: Use new api for jenkins masters credentials
*** puranamr has joined #openstack-infra21:37
lifelesstqtran: but the exclude of 1.7.9 should have worked too21:37
fungiand we can't really cast stones of arbitrariness at other communities' choices of acceptable version formats, since we've done plenty of that ourselves (albeit with some good justification for our choices)21:38
*** heyongli has joined #openstack-infra21:38
krotscheckclarkb, mordred, fungi: Let me doublecheck to make sure that format isn't customizable.21:38
*** jsavak has quit IRC21:38
mordredfungi: nope. can't cast stones at all21:39
*** pc_m has quit IRC21:41
Shrewsmordred: if work while at fullsteam can I expense my beers? fungi says yes21:41
clarkbtge real reason I get mad is longer typing to tab complete with git >_>21:41
openstackgerritKhai Do proposed openstack-infra/system-config: Use third party ci manifest to setup jenkins job builder.
mordredShrews: sure21:42
*** Rockyg has quit IRC21:42
*** thorst has quit IRC21:42
*** jsavak has joined #openstack-infra21:43
*** Sukhdev has joined #openstack-infra21:43
*** fifieldt has joined #openstack-infra21:43
jogoclarkb: what did I do?21:44
clarkbnoticed multinode wasnt working at all21:45
jogoclarkb: oh yeah, well I have invested a lot of time into trying to get it to work21:45
*** bknudson has quit IRC21:45
*** rlandy has joined #openstack-infra21:45
openstackgerritSpencer Krum proposed openstack-infra/puppet-openstackci: Check resources in the openstackci::loghost class
clarkbme too just highly distracted lately21:45
*** achanda has quit IRC21:45
lifelessok, time to get back to backporting21:46
*** my24thautumn has joined #openstack-infra21:46
mordredShrews: you're trying to recruit fungi while there, right?21:46
*** amitgandhinz has quit IRC21:47
Shrewsumm, let's go with yes, even though he's not here21:47
*** jasondotstar has joined #openstack-infra21:48
mordredShrews: ssh. you talk too much21:48
*** annegentle has quit IRC21:49
fungiShrews: if you work at the outer banks brewing station, _i'll_ pay for your beers21:49
EmilienMnibalizer: I'm probably doing the wrong way but can you look at when you got time ?21:49
fungiShrews: also as of tomorrow i have several available guest bedrooms21:49
Shrewsfungi: sold21:49
krotscheckmordred, clarkb, fungi: It appears there's a configuration option we can set that alters that prefix. The next question I have to answer is whether npm will publish this version. I _believe_ it will, but I'm about to find out.21:49
*** annegentle has joined #openstack-infra21:49
clarkbwait I can get paid to brew beer?21:50
*** heyongli has quit IRC21:51
tqtranlifeless: looks like it works, but getting different error now lol21:51
*** xyang1 has joined #openstack-infra21:51
lifelesstqtran: ...21:51
*** heyongli has joined #openstack-infra21:51
mordredclarkb: yes. there are jobs called "brewer" that pay people to "brew" "beer"21:52
nibalizerEmilienM: sorry im working on the infra sprint today21:52
EmilienMdon't worry21:52
jogoclarkb: yeah, we are trying to use multinode for a lot so it got important (grenade partial etc)21:52
tqtranlifeless: something to do with x86_64-linux-gnu-gcc21:53
* tqtran facepalms, this is sooo beyond me.21:53
*** my24thautumn has quit IRC21:53
openstackgerritlifeless proposed openstack-infra/release-tools: Be safer with git refs.
openstackgerritlifeless proposed openstack-infra/release-tools: Fixup requirements
krotscheckclarkb: Given where you live, I'm both surprised, and totally unsurprised, about how you don't know that.21:54
openstackgerritMerged openstack-infra/subunit2sql: Switch functools partials usage to kwargs
krotscheckOh look, a neat little config option that makes npm sign our rlease tags!21:55
*** amrith is now known as _amrith_21:55
*** jsavak has quit IRC21:55
krotscheckThat's helpful :)21:55
lifelesstqtran: you need the build-dependencies. We should put these in a bindep file21:55
lifelesstqtran: but for now, let me grab the list from pbr for you21:55
*** jsavak has joined #openstack-infra21:56
lifelesssudo apt-get update21:56
lifelesssudo apt-get install -y --force-yes libvirt-dev libxml2-dev libxslt-dev libmysqlclient-dev libpq-dev libnspr4-dev pkg-config libsqlite3-dev libzmq-dev libffi-dev libldap2-dev libsasl2-dev ccache libkrb5-dev21:56
mordredkrotscheck: yay!21:56
*** mriedem has quit IRC21:56
tqtranok cool, retrying now21:57
*** hichihara has joined #openstack-infra21:57
krotscheckmordred: Can we enforce only being allowed to push signed tags? Because it feels like our current flow assumes that step is still manual.21:58
openstackgerritK Jonathan Harker proposed openstack-infra/system-config: Use new api for jenkins masters credentials
openstackgerritK Jonathan Harker proposed openstack-infra/system-config: Stop passing unused nodepool parameters
*** jasondotstar has quit IRC22:00
*** enikanorov2 has quit IRC22:01
rcarrillocruzany of you folks getting ubuntu-minimal instances stuck at 'Stopping userspace bootsplash' ?22:02
rcarrillocruzwonder if i should add some kernel option on grub to disable splash22:02
fungikrotscheck: gerrit acls enforce that very well22:03
clarkbyou cant push unsigned tags22:03
krotscheckfungi: Perfect. And I was just able to release a thing that doesn't have the v prefix.22:03
fungikrotscheck: we normally allow only signed tags and have gerrit reject the rest22:03
krotscheck(On a test project I'm running on the side)22:03
*** boris-42 has joined #openstack-infra22:04
*** achanda has joined #openstack-infra22:04
*** spzala has joined #openstack-infra22:04
*** hichihara has quit IRC22:05
*** mayurig has quit IRC22:05
krotscheckfungi: Thanks for renaming that slave btw.22:06
krotscheckfungi: It's not upgraded by any chance, is it?22:06
rcarrillocruzgreghaynes: ^22:06
rcarrillocruzyou've done infra-cloud stuff , haven't you?22:06
*** claudiub|2 has quit IRC22:07
rcarrillocruzi'm giving a shot to this:
rcarrillocruzlet's see22:07
greghaynesI havent seen that, no22:08
greghaynesaside from the hostname thibg ububtu-minimal seems to be working fine for us22:08
*** mayurig has joined #openstack-infra22:09
*** jsavak has quit IRC22:09
*** peristeri has quit IRC22:10
*** dprince has quit IRC22:10
*** jayh has joined #openstack-infra22:11
jeblairfungi: in can we drop the zuul-cloner in 'install-distro-packages' (in favor of a file burned into the image)?  maybe later, but before we use this at scale?22:11
*** jayh has left #openstack-infra22:11
krotscheckSo, if I propose a change to the release slave, should I go ahead and install necessary bits as well? Forinstance, I notice that it's installing pip.22:14
openstackgerritMadhusudhan Kandadai proposed openstack-infra/project-config: New jobs for neutron-lbaasv2 scenario tests
fungijeblair: sure, we can bake the bindep-fallback.txt into images similar to what we do with slave scripts22:14
fungikrotscheck: no, i haven't built the new one yet. need to check and see if those changes have merged22:15
*** lhcheng has joined #openstack-infra22:16
lhchenghello, can anyone help how to look for the logs why the merge is failing in
fungijeblair: the thought with using zuul-cloner to get the file for now is that i can iterate a little faster on confirming additions to that file by using depends-on rather than waiting for patch approval and image update (and retried image update, and...)22:18
lhchengall the checks passed, not sure where to look at the cause of issue22:18
clarkblhcheng incompatible requirements22:18
jeblairfungi: agreed; i +2d with that note22:18
clarkbone test failed22:18
fungijeblair: awesome--thanks!22:19
lhchengclarkb: how do you figure that out? :)22:19
clarkbtoggle ci22:19
*** jgriffit1 has joined #openstack-infra22:20
* lhcheng looking for the toggle22:20
clarkbbottom of page22:20
fungilhcheng: button at the bottom22:20
jeblairso... apparently stripe uses jenkins job builder22:20
*** jgriffit1 has quit IRC22:20
jeblair_and_ the blame jenkins plugin22:20
nibalizeryou get some emails jeblair ?22:20
jeblairas, i think, all of the jjb contributors have just learned if they check their email22:20
lhchengfungi: ha! thanks!22:21
fungilhcheng: though looking at the error, it appears it's because we've got a discrepancy on stable/kilo for requirements checks, waiting on a backport of the requirements tooling in master to find its way into kilo22:21
*** annegentle has quit IRC22:21
clarkboh fun22:21
fungilhcheng: lifeless is working on that in
clarkbthat email is amazing22:22
fungiit looks very close to passing tests now22:22
openstackgerritMichael Krotscheck proposed openstack-infra/system-config: Added NPM Publishing credentials to pypi slave.
clarkband url does nit resolve22:22
lhchengfungi: cool22:22
*** heyongli has quit IRC22:22
*** achanda_ has joined #openstack-infra22:23 "web and mobile payments, built for developers"22:23
fungiwell okay then22:23
lhchengfungi: longest commit msg I've seen lol22:23
*** Guest69012 has joined #openstack-infra22:23 "built _by_ developers who don't check their ci configuration carefully"22:23
*** achanda has quit IRC22:24
lhchengfungi: thanks for the help!22:24
fungilhcheng: you're welcome22:25
openstackgerritMerged openstack-infra/project-config: Use '-' in macro names
*** heyongli has joined #openstack-infra22:25
clarkbI am not syre where our emails come from if its unintentional22:26
clarkbbecause we dont have them set up in any jobs22:26
*** shashank_hegde has quit IRC22:27
clarkbtgere isa #stripe we can talk to22:27
openstackgerritMerged openstack-infra/puppet-cgit: Updated config for CentOS 7 and Apache 2.4
*** nfedotov has quit IRC22:27
openstackgerritMerged openstack-infra/system-config: Pass exim flag to template rather than exim class
openstackgerritMerged openstack-infra/system-config: Move the server class settings out of the ask class
fungiclarkb: the blame plugin seems to be e-mailing the git committer e-mail addresses22:27
openstackgerritMerged openstack-infra/system-config: Moved the server class settings out of the eavesdrop class
*** DericHorn-HP has quit IRC22:28
openstackgerritMerged openstack-infra/system-config: Moved out the server class from the etherpad and etherpad-dev classes
fungiclarkb: their jenkins is testing the state of jjb, finds it wanting, and so git blame plugin is letting all the jjb authors know about that situation22:28
*** DericHorn-HP has joined #openstack-infra22:28
*** eharney has quit IRC22:28
lifelessoh happybase22:29
lifelessI'm not sure how to fix that. I'll need to page it in :/22:29
fungihappybase != happymaking22:29
openstackgerritMerged openstack-infra/system-config: Moved out the server class from the nodepool_prod class
*** austin81 has left #openstack-infra22:29
rcarrillocruzgreghaynes: what elements do you use for the infra-cloud servers?22:31
rcarrillocruzcan you link me pls22:31
*** Guest69012 has quit IRC22:32
*** akshai has quit IRC22:33
openstackgerritMerged openstack-infra/project-config: Add openstackci beaker jobs to several repos
mordredrcarrillocruz: we're still just experimenting ... but i think ubuntu-minimal vm should about do it22:33
openstackgerritlifeless proposed openstack/requirements: Replace coinor.pulp with pulp.
openstackgerritlifeless proposed openstack/requirements: Update Kilo requirements code
mordredrcarrillocruz: SpamapS ^^ you know the answer?22:33
rcarrillocruzbleh, that's what i use22:33
rcarrillocruzand i'm getting this freeze on the splash22:33
rcarrillocruzi think there's something going on the filesystem created22:34
openstackgerritMichael Krotscheck proposed openstack/requirements: Updated oslo.middleware to >=2.3.0
*** doug-fish has quit IRC22:35
*** sigmavirus24 is now known as sigmavirus24_awa22:35
lifelessfungi: mordred 198588 +2A would be nice22:37
SpamapSrcarrillocruz: ubuntu and dhcp-all-interfaces, that is it22:40
SpamapSrcarrillocruz: oh and vm since we used bifrost which does that kind of image22:40
rcarrillocruzk thx22:41
mordredSpamapS: oh - you're still using ubuntu and not ubuntu-minimal22:41
*** eharney has joined #openstack-infra22:42
*** _nadya_ has joined #openstack-infra22:42
rcarrillocruzred herring22:42
rcarrillocruzso, the boot was full of glean messasegs22:42
rcarrillocruzscrolling back22:42
rcarrillocruzit was showing the / partition had issues22:42
*** hashar has quit IRC22:43
openstackgerritMerged openstack-dev/cookiecutter: Remove executable bit and shebang from
rcarrillocruzi hate the iLO console, just saying...22:44
*** otter768 has joined #openstack-infra22:45
*** mriedem has joined #openstack-infra22:46
*** bknudson has joined #openstack-infra22:47
*** eharney has quit IRC22:47
*** _nadya_ has quit IRC22:47
*** otter768 has quit IRC22:50
*** achanda_ has quit IRC22:50
openstackgerritColleen Murphy proposed openstack-infra/puppet-ansible: Add Gemfile and puppet 4 checks + fix lint errors
openstackgerritSpencer Krum proposed openstack-infra/puppet-os_client_config: Manage OSCC config file with Puppet
rcarrillocruzcat: hosts: No such file or directory22:52
rcarrillocruzso glean fails22:52
rcarrillocruzcos in ubuntu-minimal it appears /etc/hosts is non-existant22:52
rcarrillocruzmordred: ^ you aware of this?22:52
fungircarrillocruz: i believe glean populates it22:53
* rcarrillocruz is shaving a gigantic yak today...22:53
mordredyes. glean populates it22:53
mordredoh. ZOMG22:53
fungircarrillocruz: at least when we boot ubuntu-minimal we have an /etc/hosts file provided by glean22:53
rcarrillocruzyeah, but it assumes /etc/hosts just exists22:53
mordredwell - that would certainly be a bug22:53
* mordred is bashful22:54
rcarrillocruzi mean, i see on boot.log a 'No such file or directory' error from glean22:54
rcarrillocruzon /etc/hosts22:54
*** juggler has joined #openstack-infra22:54
rcarrillocruzi'll see if i can push a path tomorrow morning prior to leave on short vaca22:55
rcarrillocruzalso, kind of annoying vim-tiny is not even available on ubuntu-minimal22:55
rcarrillocruznor less22:55
rcarrillocruzi mean22:55
rcarrillocruzusing more is awful, lol22:55
*** achanda has joined #openstack-infra22:55
*** mattoliverau has quit IRC22:56
*** boris-42 has quit IRC22:56
*** matt_____ has joined #openstack-infra22:56
*** crc32 has quit IRC22:56
*** juggler_ has quit IRC22:56
*** diego__ has quit IRC22:56
jeblairi've used systems without more :)22:56
jeblairyou just pick the paper up off of the floor and read as far back as you need to22:57
*** diego__ has joined #openstack-infra22:57
fungiindeed. cat and paging the virtual console (or scrolling the serial terminal)22:57
*** heyongli has quit IRC22:57
fungioh, right, your teletype22:57
*** boris-42 has joined #openstack-infra22:57
rcarrillocruzi think my first text processor  was edlin?22:57
rcarrillocruzms-dos thingy22:57
rcarrillocruzthen wordstar22:58
rcarrillocruzthat was like OMG the future22:58
* fungi remembers edlin very well22:58
*** heyongli has joined #openstack-infra22:58
*** DericHorn-HP has quit IRC22:58
jeblairoh, yeah and don't forget software flow control.... rcarrillocruz: you can use ^S and ^Q to pause/release your terminal :)22:58
fungiyou can _still_ do that too!22:58
rcarrillocruzmordred: i really doubt anyone has taken simple-init+ubuntu-minimal too far? i touched /etc/hosts to get past the issue, now it fails cos there's no openssh-server22:58
* fungi pauses his xterm for fun22:59
rcarrillocruzunless you have some other machinery or element that sets those deps in place22:59
jeblairfungi: you can't see me!22:59
mordredrcarrillocruz: no - we boot images made with simple-init in infra currently22:59
fungijeblair: toggled my scroll lock back off22:59
mordredrcarrillocruz: but we DO run a few other elements on those22:59
rcarrillocruzthat might explain22:59
fungijeblair: i can totally see you again22:59
mordredrcarrillocruz: so it's possible that ubunt-minimal simple-init vm would be too few22:59
jeblairfungi: we had fun while you were gone22:59
rcarrillocruzi'm uber noob on disk-image, is there an element to install a list of packages?23:00
rcarrillocruzi.e. you pass some dib vars 'pls install me these'23:00
jeblairmordred, rcarrillocruz: maybe our puppet run is doing that?  it should be installing openssh-server23:00
mordredrcarrillocruz: there is - I dont' remember what it is23:00
rcarrillocruztoo late for me23:00
mordredjeblair: yes, I believe our puppet run takes care of it23:00
rcarrillocruzi'll look at it tomorrow23:00
mordredrcarrillocruz: ok - we can help you tomorrow23:00
*** erlon has quit IRC23:00
rcarrillocruzthx for the help folks23:00
*** gildub has joined #openstack-infra23:00
rcarrillocruztalk to you next week , be back on Sunday from short vaca23:00
mordredrcarrillocruz: havea  good vaca!23:01
rcarrillocruzthere ya go23:01
krotscheckAnyone know how I can locally simulate the apply-precise build? These log messages are not really helpful.23:02
krotscheck(on system-config)23:03
*** heyongli has quit IRC23:03
mordredkrotscheck: I think nibalizer had a something?23:03
clarkbif you run the script with puppet installed that hould work but I think nibalizer and pabelanger were fixibg issues there23:03
*** heyongli has joined #openstack-infra23:03
krotscheckThat whole build is a special kind of juju magic.23:04
mordredjeblair: I can confirm that stripe use jjb :)23:04
krotscheckmordred, jeblair: Me too!23:04
*** yamahata has quit IRC23:05
openstackgerritColleen Murphy proposed openstack-infra/puppet-apparmor: Add Gemfile and puppet 4 checks
nibalizerkrotscheck: so ./; puppet apply --modulepath=/etc/puppet/modules:modules foo.pp23:05
nibalizerwhere foo.pp contains the contents of the node definiton23:05
nibalizeri should probably docment that23:05
krotschecknibalizer: I am _certain_ there was english in there somewhere.23:06
*** dimtruck is now known as zz_dimtruck23:06
clarkbor run the script...23:06
krotschecknibalizer: Ok, so I... do what. Spin up a precise vm, check out system-config, and run that?23:06
nibalizerya thats what I do23:06
*** sdake has joined #openstack-infra23:07
*** asettle has joined #openstack-infra23:09
*** achanda has quit IRC23:09
mordredmumble vagrant mumble docker mumble mumble cloud mumble ansible mumble puppet mumble cloudfoundry mumble23:10
anteayamorning jhesketh23:10
mordredlook it's a jhesketh23:11
*** sdake_ has quit IRC23:11
jheskethwe're all doomed?23:11
mordredso doomed23:11
jheskethso going back to bed is an option then right?23:12
*** fitoduarte has quit IRC23:13
fungispeaking of the obscurity of our puppet apply jobs, i can usually manage to get my divining rod to twitch over the right log entries but i'm having no luck finding the error in the job log for
nibalizerfungi: ya the errors were supressed23:16
nibalizerbug in the apply test23:16
openstackgerritColleen Murphy proposed openstack-infra/puppet-apps_site: Add Gemfile and puppet 4 checks
nibalizerrecheck it and more information will flow23:16
funginibalizer: thanks! thought i was going crazy :/23:16
jeblairnibalizer: should have been fixed by though, right?23:17
fungithat's what i get for not really being here today23:17
tqtranneed help with decrypting this log, anyone know why python-qpid-proton fails?
jeblairnibalizer: i see only 'exit 0's in that log23:18
*** jgriffith has joined #openstack-infra23:18
fungitqtran: i think you need to install the uuid-dev package?23:19
nibalizer198869 was testsed at 8:45 am so I think it didn't have our fixes23:19
nibalizeryea I have no information on why 199707 failed23:20
jeblairnibalizer: i'm looking at the log from krotscheck which was at 22:3123:20
krotscheckdoom doom doom, dooomy doom doom doom23:21
fungikrotscheck: no more singing the doom song!23:21
*** ZZelle_ has quit IRC23:22
openstackgerritMatt Riedemann proposed openstack/requirements: Raise minimum python-openstackclient to 1.4.0
dhellmannmordred: there is a very very old swift-private group in launchpad that is tripping some sort of permission issue on random bug notifications when we run release scripts. notmyname doesn't think it's used any more. Can we delete it, or fix its permissions or something to get rid of the notification errors?23:25
dhellmannttx: this is the thing that caused the milestone scripts to bomb for me, too ^^23:25
notmynamemordred: I think it's a list of RAX employees in july of 201023:26
openstackgerritJames E. Blair proposed openstack-infra/system-config: Remove -e from apply test applier
jeblairnibalizer, krotscheck, fungi: ^23:26
mordreddhellmann: I know nothing about it?23:26
notmynamemordred: you're an admin23:26
dhellmannmordred: it's not clear how the group is even related to the bugs -- apparently you're an administrator23:26
notmynamemordred: actually, the only admin still active in openstack :-)23:26
notmynamemordred: check that link! it's a fun walk down memory lane ;-)23:26
mordredI know what that is23:27
fungimordred: notmyname: in fact for those of us not members of the group, it's not visible at all in lp23:27
mordredyah - remember that chunk where we had private launchpad stuff because we hadn't announced at OsCon yet?23:27
notmynamefungi: right23:27
notmynamemordred: oh yeah :-)23:27
anteayapleia2: thinking of the rename on friday, have you had a chance to spin up an etherpad yet?23:27
mordrednotmyname: this is kindof an awesome back-in-the-day easter egg :)23:27
lifelessfungi: can bindep install things yet?23:28
notmynamemordred: I know!23:28
lifelessfungi: or just report?23:28
fungilifeless: it only reports, it doesn't do anything which would require root-level privileges23:28
mordredturns out - 5 years and 1 day ago, I set up the private launchpad projects in what would become openstack :)23:28
notmynamemordred: a new openstack shibboleth. "are you in the swift-private group?!"23:28
krotscheckjeblair: I don't understand enough about the underlying system to really have a clue, but sure, lgtm!23:28
mordrednotmyname: totally23:28
fungilifeless: however the -b/--brief option is suitable for feeding directly to your platform's package manager23:28
lifelessfungi: I don't recall23:28
lifelessfungi: ahha that what I wa going to ask23:28
anteayahappy birthday?23:28
lifelessfungi: is that in a release?23:29
mordreddhellmann: so what broke for you?23:29
fungilifeless: yep23:29
lifelessfungi: if not, please do so :)23:29
lifelessfungi: its not in the help on
*** zul has joined #openstack-infra23:29
mordrednotmyname: this is something you don't see often:23:29
lifelessfungi: but it is in the code.c ool23:29
mordredLatest members23:29
mordred    Todd Willey23:29
fungilifeless: the only things not yet in a release are bindep accepting - as a package list file path (to read from stdin) and the documentation change which needs to get a second review23:29
dhellmannmordred: trying to release python-cinderclient, which did get released and the bug did get updated but I've had to update the release tool to ignore this particular error23:29
dhellmannmordred: maybe just making the group not private any more would fix my issue?23:30
mordreddhellmann: nod23:30
mordreddhellmann: it is not possible23:30
mordreddhellmann: that was the first bug we uncovered in launchpad :)23:30
* dhellmann shakes his fist at launchpad23:30
dhellmanncan the group be deleted? or can "openstack-admins" be added as administrators (which would be a side-ways way to give me access)? or can we add the release-team (which would give our other release team members access)?23:31
mordreddhellmann: so - I'm sure I can delete it23:31
fungihuh. why can't i browse i'm getting connection resets23:31
pleia2anteaya: doing it tomorrow23:32
mordreddhellmann: but right now I don't see any objects anywhere that would indicate an assocition between swift-private and anything23:32
dhellmannmordred: we could also try to just break the subscription to swift bugs, but notmyname couldn't find that23:32
mordredoh - bug subscription23:32
fungianybody else having trouble with git.o.o?23:32
anteayapleia2: ah okay, was going to see if I could get the patches in some kind of shape while giving authors time for a respin where required23:32
anteayapleia2: so tomorrow then, mind if we do it first thing so I can pick away at it?23:33
pleia2anteaya: that would be great23:33
pleia2anteaya: sure23:33
nibalizerjeblair: lgtm23:33
openstackgerritColleen Murphy proposed openstack-infra/puppet-accessbot: Add Gemfile and puppet 4 checks
mordredlifeless: do you remember how teams get subscribed to bugs?23:33
openstackgerritColleen Murphy proposed openstack-infra/puppet-ansible: Add Gemfile and puppet 4 checks + fix lint errors
openstackgerritColleen Murphy proposed openstack-infra/puppet-apparmor: Add Gemfile and puppet 4 checks
openstackgerritColleen Murphy proposed openstack-infra/puppet-apps_site: Add Gemfile and puppet 4 checks
mordreddhellmann: sorry - I'm still poking23:35
lifelessmordred: yes23:35
*** rlandy has quit IRC23:35
lifelessmordred: IIRC its limited to owners not members23:35
dhellmannmordred: thanks23:35
Davieymordred: A team owner can subscribe a team to a bug23:35
mordredlifeless: more importantly -do you know how to unsubscribe a team?23:35
fungii've got also got a manual puppet apply --test going on release.slave.o.o and it's basically just hanging23:35
DavieyAh, lifeless is here23:35
mordredI am the team owner23:35
lifelessmordred: should be pointy clicky23:35
lifelessmordred: what bug ?23:35
openstackgerritMerged openstack-infra/bindep: Update documentation from cookiecutter
openstackLaunchpad bug 1455102 in python-novaclient "some test jobs broken by tox 2.0 not passing env variables" [High,In progress] - Assigned to Sean Dague (sdague)23:36
mordredlifeless: but there is no way it's a direct subscription23:36
mordredthat bug is way too new23:37
mordredlifeless: only thing i can think of is that there is a relationship between ~swift-private and the swift project?23:37
*** matt_____ is now known as mattoliverau23:38
*** blaedd has quit IRC23:38
*** heyongli has quit IRC23:38
lifelessfungi: the help also needs to mention that profiles can go on the command lind23:38
*** sdake_ has joined #openstack-infra23:38
lifelessfungi: if you're poking at it23:38
fungilifeless: i wasn't. currently trying to wrangle an upgraded worker for our release automation23:39
fungithough i should be knocking off for the night23:39
lifelessmordred: visit
openstackLaunchpad bug 1455102 in python-novaclient "some test jobs broken by tox 2.0 not passing env variables" [High,In progress] - Assigned to Sean Dague (sdague)23:39
fungibut for some reason i can't reach git.o.o23:39
openstackgerritKhai Do proposed openstack-infra/puppet-openstackci: Move jenkins job builder into jenkins_master
lifelessmordred: it should - IIRC - tell you how you're getting mail23:39
mordredlifeless: brilliant23:40
mordredlifeless, dhellmann: DONE23:40
dhellmannmordred: \o/23:40
*** heyongli has joined #openstack-infra23:40
dhellmannmordred, lifeless, notmyname : thanks for working through that23:41
*** thorst has joined #openstack-infra23:41
*** achanda has joined #openstack-infra23:41
mordred"You receive emails about this bug because you are a member and administrator of the team OpenStack CI Core, which is the owner of OpenStack-Gate, which has no bug supervisor."23:41
lifelessmordred: (that link btw is in the ui - upper right, near the 'edit bug mail' link23:41
*** shashank_hegde has joined #openstack-infra23:41
mordredlifeless: awesome. it never occured to me that that would do anyting other than edit my personal bug mail :)23:42
*** sdake has quit IRC23:42
lifelessmordred: there are two links in that widget.23:42
lifelessmordred: its a little subtle I suspect23:42
lifelessone is 'edit bug mail' (which also allows team subscriptions)23:42
lifelessand the other is that ^ link23:42
lifelesswhich is the diagnose-this-now-thanks23:42
lifelessmordred: now q for you23:43
mordredlifeless: aw. I wanted to diagnose-this-now23:43
openstackgerritKhai Do proposed openstack-infra/system-config: use updated jenkins_master.pp manifest
*** Sukhdev has quit IRC23:43
lifelessmordred: you did :)23:44
*** sabeen3 has quit IRC23:45
lifelessfungi: oh btw, is bindep available on workers now ?23:46
fungii think we have some sort of incident going on here. i seriously can't get to our git servers via http/https from home nor from the new server i just spun up in rackspace. it looks like something is trying to feed replacement content for it... i would have expected a misbehaving transparent proxy in my isp except that i'm also seeing puppet failing to vcsrepo things with "fatal: unable to access23:46
fungi'': gnutls_handshake() failed: A TLS packet with unexpected length was received."23:46
lifelessfungi: like[6~23:47
fungiSERIOUSLY is anyone else seeing this problem23:47
openstackgerritlifeless proposed openstack/requirements: Move integration test dependencies into bindep
fungimordred: jeblair: clarkb: ?23:47
lifelessfungi: ^ - ^ I'm going to try cgit now23:47
clarkbfungi where is that from?23:47
clarkbchecking from phonr23:47
lifelessfungi: reproduced, I have that issue too23:47
mordredyes. I have it23:48
clarkbyes its broken23:48
lifelessERR_CONNECTION_CLOSED from chromium23:48
fungiclarkb: spinning up a new server, trying to puppet apply23:48
*** Sukhdev has joined #openstack-infra23:48
fungithis has been going on for me for something like half an hour at least23:48
fungiif you manually connect to it and try to get / it feeds back a lot of interesting content which does not belong to us23:49
fungior at least i do23:49
asselinfungi, same for us23:50
*** mdenny has joined #openstack-infra23:50
*** puranamr has quit IRC23:51
fungihaving trouble browsing the content from individual git backend servers too23:52
jeblairfungi: hi, looking23:52
lifelessso its a network thing23:54
lifeless Next Gen Cloud Servers | ALL REGIONS | Starting on Monday, July 6th @ 10:00 PM CDT *In progress*23:54
lifelessRackspace will be performing upgrades on its Software Defined Networks. These upgrades will present a small disruption for customers (<2 seconds), however they will also provide an increase is overall performance.23:54
lifelesscould be related23:54
lifelesse.g. utterly failed23:54
fungilifeless: though not supposedly starting in dfw yet23:54
clarkbgit:// works o the backend23:55
fungibut yeah, possibly. i saw that too and considered the possibility23:55
jeblairthe haproxy is doing work...23:55
lifelessOur engineers continue to work to resolve an issue causing performance degradation for our Cloud Files product in the following regions:23:55
lifelessDFW, ORD, IAD23:55
clarkbso it only affects http?23:55
lifelessclarkb: https too23:55
fungilifeless: cloudfiles is swift, unrelated in this case23:55
fungiclarkb: yeah, looks like an overzealous proxy to me23:55
openstackgerritKhai Do proposed openstack-infra/puppet-openstackci: Update jenkins job builder parameters to be more sensible.
lifelessfungi: yes, but sometimes smoke -> fire23:55
lifelessclarkb: but the https impact culd be via http on the backend23:56
fungilike someone suddenly decided to transparently redirect all http and https traffic somewhere23:56
jeblairbut there is a substantial load reduction on git.o.o23:56
clarkbfungi: seems to affect the backends direclty too23:56
clarkbwhich are on non standard ports so its really funky23:56
fungiclarkb: yep, i confirmed that already23:56
fungiand i'm tcpdumping on git.o.o and don't see my web traffic arriving at all23:56
lifelessI'm taking out an 80% wager with myself that its SDN23:56
fungiyeah, it seems likely23:57
clarkbbut why would only http(s) be affected?23:57
clarkbor n that case why affect non standard port http(s)23:57
fungiclarkb: they're probably redirecting it even though it goes to broken23:57
clarkbwouldn't expect them to do l7 introspection with sdn23:57
jeblairis it breaking git:// ?23:57
fungithey might shunt 8080 and 4443 through their proxies just because those also look like alternate http/https ports23:58
lifelessis it breaking ssh ?23:58
fungissh is working fine too23:58
jeblairgrr.  telnet is not installed on git.o.o23:59
lifelessthat rules out sdn :>23:59
fungii noticed the same23:59
lifelessjeblair: nc ?23:59
funginor nc23:59
openstackgerritKhai Do proposed openstack-infra/system-config: use updated jenkins_master.pp manifest
*** wenlock_ has quit IRC23:59
jeblairtelnet is installed now23:59
fungiopenssl s_client -connect
jeblair(which also gives us another network datapoint)23:59

Generated by 2.14.0 by Marius Gedminas - find it at!