Wednesday, 2014-05-28

jeblairjhesketh, mordred: ^00:00
dimsjeblair, where are the logs you were looking at? (just curious)00:00
mordredjeblair: +2 +A00:01
*** matsuhashi has joined #openstack-infra00:01
*** msabramo has quit IRC00:01
openstackgerritA change was merged to openstack-infra/config: Reduce gearman logging level
jeblairi pushed it through ^00:02
*** gokrokve has quit IRC00:03
*** lcheng_ has joined #openstack-infra00:03
fungipizza made and eaten, scrollback consumed as desert00:03
*** gokrokve has joined #openstack-infra00:04
mattoliveraufungi: sounds like a delicious dessert :P00:04
jeblairfungi: short version: logs here:  nodepool status command sent to gearman taking > 30 seconds seems to be the problem00:04
jeblairfungi: cause unknown00:04
openstackgerritMark Sturdevant proposed a change to openstack/requirements: Remove hplefthandclient from global-requirements
fungiand additional logging seems to have exacerbated it00:04
*** msabramo has joined #openstack-infra00:05
jeblairfungi: i don't really have any reason to suspect that other than that's the main thing we've changed and it seems to have been getting worse00:05
jeblairfungi: so that one's just a shot in the dark00:05
jeblairfungi: i'm hoping that plus changing the timeout to 300 seconds will at least make things go00:06
jeblairfungi, mordred, clarkb, SergeyLukjanov: zuul is using a manually installed "gear" library with the timeout and logging changes00:06
*** W00dy__ has joined #openstack-infra00:07
jeblairfungi, mattoliverau, dims: i added in the extra logging entries between the status command start/stop to the etherpad00:07
mattoliverauThanks jeblair00:07
mordredjeblair: there's a lot of us now00:08
jeblairfungi, mattoliverau, dims: the main context missing from that is that normally the log is awash in send/receive packet log entries, but there aren't any during that time because it's busy00:08
mordredit's hard to just ping infra-core00:08
*** liyuezho has joined #openstack-infra00:08
*** gokrokve has quit IRC00:08
jeblair#status log zuul is using a manually installed "gear" library with the timeout and logging changes00:08
openstackstatusjeblair: finished logging00:08
jeblairmordred: thanks for reminding me00:08
jeblairmordred: well i was really pinging infra-root there00:09
*** msabramo has quit IRC00:09
mordredjeblair: nod00:09
jeblairrestarting zuul00:09
*** mrodden has quit IRC00:09
jogoso random idea / question:00:10
jogofor unit tests, we have a problem where we skip some tests for optional dependencies00:10
jogoand those aren't listed in the local requirements file.00:10
*** otherwiseguy has quit IRC00:10
*** otherwiseguy has joined #openstack-infra00:11
jogolong term it would be nice to do a requirments directory00:11
jogonot just a single list00:11
jogobut until then, can we just run the unit tests with all of global-requirements installed?00:11
fungijogo: in the venv?00:11
jogomaybe as an additional job00:12
jogofungi: presumably in a venv, but I don't think that is required00:12
jeblairzuul is running00:12
fungijogo: worth noting, that won't solve version-dependent tests00:13
anteayalet's hope it stays that way, running that is00:13
jogofungi: right. but if we have those (I am sure we do) that is pretty ugly00:13
fungijogo: you could test it by replacing requirements.txt with the current global reqs and emptying out test-requires.txt, just to see what blows up00:14
*** lcheng_ has quit IRC00:15
jogofungi: good idea00:15
*** melwitt has quit IRC00:17
*** otherwiseguy has quit IRC00:17
fungijogo: as far as an additional job, that would be pretty easy (and i think necessarily separate, since otherwise we won't test that the listed test requirements are sufficient to run tests)00:18
mordredjogo: we've also discussed running unittests in a devstack00:19
devanandahemna: hi! around? i'm curious why you're removing hplefthandclient from global req's00:21
jogomordred: that seems like overkill or at least as as second phase00:22
*** SumitNaiksatam_ has joined #openstack-infra00:22
fungion the other hand, it may be nice to test that devstack provides a suitable environment for running unit tests (a long-standing request)00:23
*** SumitNaiksatam has quit IRC00:23
*** SumitNaiksatam_ is now known as SumitNaiksatam00:23
*** zzelle has quit IRC00:24
*** chuckC has quit IRC00:24
*** bhuvan_ has quit IRC00:27
*** gabriel-bezerra has left #openstack-infra00:27
*** SumitNaiksatam has quit IRC00:28
jogofungi: yeah, although the overhead of spinning up another devstack sounds concerning00:29
*** praneshp has quit IRC00:29
jogobut not unreadable00:29
mordredjogo: well, we alraedy spin up throwaway nodes00:30
mordredjogo: thing to be aware of also - pip dependency lists are order dependent, and global-requirements is not intended to be installed, so it's alphabetized00:31
mordredso you'll get some potential weirdness00:31
mordredjogo: I actually think doing the first thing you mentioned is actually less hard than you might think - since it's already a supported construct in setuptools00:32
*** gokrokve has joined #openstack-infra00:32
mordredjogo: so we'd need to add some support to pbr to deal with inserting stuff into extra_requires00:33
mgagneis gerrit replication plugin a system wide option or can it be enabled per project?00:34
*** dstanek_zzz is now known as dstanek00:34
*** zhiyan_ is now known as zhiyan00:34
mordredmgagne: the replication plugin is system wide - but it has config to control what gets replicated00:34
mordredmgagne: one of the controls is that you can tell it to only replicate things that a particular user or group can access00:34
mgagnemordred: ok so it's only a matter of creating a group with limited access?00:35
mordredmgagne: so you could, if you needed really specific controls, "enable" replication for a repo by having a group that by default could see nothign and than explicitly granting that group view permissions on each repo you wanted00:35
mgagnemordred: cool thanks00:35
*** zhiyan is now known as zhiyan_00:36
*** Sukhdev has quit IRC00:38
*** zhiyan_ is now known as zhiyan00:38
* jogo is still waiting for nova tests to finish00:38
*** yamahata has joined #openstack-infra00:39
*** zhiyan is now known as zhiyan_00:41
openstackgerritMorgan Fainberg proposed a change to openstack-infra/gitdm: Update company affiliation for morganfainberg
*** dstanek is now known as dstanek_zzz00:44
*** UtahDave has quit IRC00:45
*** SumitNaiksatam has joined #openstack-infra00:46
*** gokrokve_ has joined #openstack-infra00:56
devanandajogo: i vaguely recall vim encoding lines in nova files at one point. do you know if/why they were removed?00:57
*** BadCub01_ has quit IRC00:57
*** wenlock has joined #openstack-infra00:57
* jogo can't tell if devananda is trolling or not00:58
*** liyuezho has quit IRC00:58
devanandajogo: sort of -- but i am actually looking for the reference00:58
mordreddevananda: we removed them because they were both incorrect and also because we didn't want to support all the editors in teh world00:59
devanandathat's what i thought00:59
jogodevananda: there was a big ML thread but
jogoand what mordred said00:59
jogoturns out they aren't very secure, and are disabled by default etc00:59
*** gokrokve has quit IRC01:00
devanandaremember what the ML thread month / title was?01:00
devanandai just realized there's a bunch still in ironic/client, and someone proposed converting them all to emacs lines01:00
*** dstanek_zzz is now known as dstanek01:01
devanandajogo: perfect, thanks01:01
devanandai was searching 2014 :(01:01
jogodevananda: another reason they are a termi-ism01:01
*** ArxCruz has joined #openstack-infra01:01
jogoamazingly enough I just clicked the first link when I googled 'vim modelines openstack'01:02
* devananda fails01:02
jogoI didn't say that to shame you, I said it because I am impressed with google01:02
jogothis time01:02
jogoimpressed with google this time*01:03
* devananda doesn't feel shame for not using google01:03
mordreddevananda: you're a google01:03
* devananda googles mordred01:04
* mordred returns a wet cat01:04
jogo"known as a notorious traitor"01:06
openstackgerritA change was merged to openstack-infra/storyboard-webclient: Fix StoryBoard spelling
*** wenlock has quit IRC01:07
*** chuckC has joined #openstack-infra01:10
jogomordred: thats what wikipedia says about you01:10
fungiwikipedia is never wrong01:14
* mordred throws wet cats at everyone01:15
* jesusaurus "corrects" whichever wiki page claims that wikipedia is never wrong01:16
morganfainbergmordred, i .. great, do you know how hard it is to get wet cat smell out of carpet?01:16
*** wenlock has joined #openstack-infra01:18
*** camunoz has quit IRC01:20
*** saschpe has quit IRC01:23
*** dstanek is now known as dstanek_zzz01:24
*** dstanek_zzz is now known as dstanek01:24
*** saschpe has joined #openstack-infra01:24
*** zhiyan_ is now known as zhiyan01:26
*** gokrokve_ has quit IRC01:27
*** gokrokve has joined #openstack-infra01:28
jogofungi: running nova unit tests with all of global deps installed worked01:28
jogoI am very surprised01:28
*** wenlock has quit IRC01:29
*** weshay has quit IRC01:29
jogoso assuming folks want to run the unit tests with all of global reqs installed, are there any infra side concerns?01:29
jogowhen do we EOL havana?01:31
*** gokrokve has quit IRC01:32
jeblairfungi, mordred, mattoliverau, jhesketh: i updated with an strace01:34
fungia while--i think we extended it past k assuming interested parties step up to keep it working01:35
fungijogo: ^01:35
*** melwitt has joined #openstack-infra01:35
jeblairfungi, mordred, mattoliverau, jhesketh: i ran the 'status' command repeatedly until it looked like it was very slow, and then started an strace01:35
jeblairfungi, mordred, mattoliverau, jhesketh: it's looking like the thread handling the status command (15898) was swapped out in favor of the connection handling thread (15899) which stayed running for a VERY long time01:36
fungidoes the connection handling thread need to yield more often?01:37
jogofungi: thanks01:37
*** weshay has joined #openstack-infra01:37
fungijust curious what the usual approach to that is01:37
jeblairfungi: it should spend most of its time in poll(), so i wouldn't expect it to be a problem01:39
fungimaybe there are connection setup/teardown operations which are blocking for longer periods?01:39
devanandajeblair: regarding making ironic vote on nova as a third-party driver, when'd be a good time to chat about how to make that happen?01:40
devanandajeblair: or is there a magical wiki that will answer all my questions on that?01:40
jeblairdevananda: not today :)01:40
devanandajeblair: tomorrow?01:41
jeblairdevananda: i hope so01:42
jeblairfungi: take a look at the timestamps in the strace; i don't fully understand it yet...01:42
jeblairbut it's well past dinner time here, so i have to run01:42
fungigoes silent after poll for 4-5 seconds a couple times, yeah01:43
fungilooks like it accepts a new connection after each of those?01:45
fungidifferent initiating ip addresses01:46
jeblairfungi: yeah, i think that's the case; if you recall from the earlier logs, the only thing geard was duing during the slow time was accepting a handful of new connections01:46
fungiahh, right. still not sure what the implication of that is, but maybe it'll come to me01:47
*** nosnos has joined #openstack-infra01:48
openstackgerritA change was merged to openstack-infra/jenkins-job-builder: Fail if specified set of defaults is not found.
mattoliveraujeblair, fungi yeah the silence is strange, it that the time for the system to accepts the previous conncetion, and then what polls and waits for a second connection. Or is it working with the connection (without issuing system calls, which sounds doubtful) during the silent times. Unless there is a context switch to a different process (that the connection thread needs to wait for), which wont be01:52
mattoliveraulogged as it isn't actually run by zuul.01:52
clarkbreading the sb now01:53
mattoliverauMaybe I'll look at zuul io load, or number on context switches. Though I would think on a multi core/threaded system if it was in wait, it could continue running the status thread.01:53
*** ArxCruz has quit IRC01:54
jeblairmattoliverau, fungi: nothing substantial happens between the poll and accept calls, so i suspect the time really is spent in poll01:56
jeblairmattoliverau, fungi: here's the loop in question:
*** changbl has joined #openstack-infra01:57
jeblair(and you can tell when it gets to the bottom with the write('1')01:57
clarkbjeblair are things "working" with the longer timeout?01:57
clarkband it isnt possible to use C gearmand at this point right?01:58
jeblairclarkb: yeah, hasn't hit again yet01:58
*** Ivo has quit IRC01:58
jeblairclarkb: no one has added the needed bits to it01:58
mattoliveraulots of the memeory is used up in caches, so although most of it is free, maybe the gap is time in kernl space, as there seems to be many many gigs of caches being used. But yeah, or it's just idle, by why not run the other thread again then.. hmm. I'll look at the loop now.01:58
fungiseems to be working so far, and the queue backlogs are once again being whittled away01:58
jeblairclarkb: i'm not at all convinced we've hit some kind of scalability limit here...02:00
fungiit does feel like we're exposing some sort of infrequent race/deadlock02:01
jeblairclarkb: we have a thread that is spending _seconds_ inside of a system call while another thread is starving.  i'm really interested in understanding why that's the case.02:01
*** wenlock has joined #openstack-infra02:02
jeblairfungi: i suspect "rate of client connects/disconnects" is related, the nodepool config changes that made hpcloud 1.1 work better may have been the trigger02:02
*** Ivo has joined #openstack-infra02:02
*** alugovoi has quit IRC02:03
fungitiming would be about right, given that today was the first heavy use day we've seen for a while02:03
*** yaguang has joined #openstack-infra02:06
clarkbjeblair wow fun (re seconds in syscall)02:06
clarkbpoll is where that happens?02:07
jeblairyeah, there's an instance where it's 4 seconds from poll to the next call02:07
fungiaccording to the systrace anyway02:07
jeblairpoll realeases the GIL, so if another thread gets scheduled, it should be able to run02:08
clarkbcan you s/poll/epoll/? curious if that changes anything02:08
*** zz_gondoi is now known as gondoi02:09
jeblairclarkb: we're not actually dealing with reading/writing here, we're just accepting connections02:09
jeblairclarkb: (the poll is to determine if there's a connection to accept)02:10
*** gondoi is now known as zz_gondoi02:10
jeblairclarkb: and we're not blocking on the accept, it's the poll that's sitting there02:11
clarkbright but isnt the syscall different?02:11
*** homeless has quit IRC02:11
*** zz_gondoi is now known as gondoi02:11
jeblairclarkb: yeah, but why would this make a difference?02:11
clarkbbug in libc or kernel maybe02:12
mattoliverauThe first poll was interuppted, and continued about 6 seconds later, the second poll wasn't, but still didn't continue for about 5 secionds.. So I don't know why it didn't get interupted but looks like it normally takes 5 or so seconds to poll, and that is the lost time.02:12
jeblairclarkb: the only way we have to test this is in production; i've been unable to produce it locally, even in a setup that's twice the size we're running02:12
jeblairclarkb: so i don't want to try _every_ idea :)02:12
clarkbah ok02:12
*** camunoz has joined #openstack-infra02:13
jeblairokay, really dinner now.  good luck, anyone who wants to work on this.  ;)02:13
*** nati_uen_ has quit IRC02:13
clarkbthe other thing we can do is put a timeout on the poll02:23
clarkbbut without a reproduceable case this is hard to test02:23
*** weshay has quit IRC02:24
mattoliverauclarkb: yeah, lookingingat more of the polls in the strace, the first one only takes a second, and then it started to take longer. jeblair said we ran status a few times before running the strace, so maybe it has something to do with the number of connections. Maybe a red herring? or number of connections is the trigger?02:27
mattoliverau*he ran02:28
*** rfolco has quit IRC02:29
*** wenlock_ has joined #openstack-infra02:30
*** gondoi is now known as zz_gondoi02:30
clarkbmattoliverau: these poll() calls don't specify a timeout so should block02:31
clarkbbut as I understand it they shouldn't block because the FDs polled on have data?02:31
*** asselin_ has joined #openstack-infra02:32
*** mkerrin1 has joined #openstack-infra02:32
jeblairso the problem could be related to the thread that's doing the connection handling (poll); or it could equally be related to the other thread, the one handling the status command02:33
jeblairperhaps it's not being scheduled because it isn't ready; maybe the sendto is blocking or something02:33
*** mkerrin has quit IRC02:33
*** asselin has quit IRC02:33
*** Ivo has quit IRC02:33
*** YorikSar has quit IRC02:33
jeblairclarkb: no, they should block.02:34
*** YorikSar has joined #openstack-infra02:34
jeblairclarkb: which is fine.  the problem is that the other _thread_ isn't running.02:34
jeblairclarkb: (which it should be able to do since the poll is blocking)02:34
clarkbjeblair: the handleStatus() thread is not running02:35
mattoliveraujeblair: yeah, interestingly, the sendto (status) is swapped out for a poll, but on line 88, a poll is swapped out for the sendto.. so yeah, the poll could be a red herring and sendto could be being blocked on something else, or isn't ready.02:35
jeblairclarkb: yeah it's the one doing all the sendto calls ^02:36
*** zhiyan is now known as zhiyan_02:36
clarkbjeblair: I think they are the same thread02:37
clarkbhandleAdminRequest() is called in the main _pollLoop()02:38
clarkber wait maybe that is a different class02:38
jeblairclarkb: no, that thread is 15898; the connection handling thread is 15899.02:39
clarkbnope just overridden02:39
jeblairclarkb: the packet handling loop and the connection handling loop are constructed the same way, so it's easy to confuse them02:39
clarkbah I see, the poll are part of connection handling not packet handling02:41
*** alugovoi has joined #openstack-infra02:43
*** unicell has joined #openstack-infra02:44
*** olaph has quit IRC02:47
jeblairclarkb: yup02:48
*** gyee has quit IRC02:49
*** zz_gondoi is now known as gondoi02:49
*** otherwiseguy has joined #openstack-infra02:49
*** zhiyan_ is now known as zhiyan02:49
*** gondoi is now known as zz_gondoi02:49
*** zz_gondoi is now known as gondoi02:50
*** alugovoi has quit IRC02:52
*** gondoi is now known as zz_gondoi02:54
*** otherwiseguy has quit IRC02:54
*** msabramo has joined #openstack-infra02:54
mattoliverauI'm off to grab some lunch, bbs02:54
*** rcarrill` has joined #openstack-infra02:55
*** dims has quit IRC02:57
*** rcarrillocruz has quit IRC02:58
*** Ivo has joined #openstack-infra03:00
lifelessmordred: any thoughts on whats up in ?03:04
*** mmaglana has joined #openstack-infra03:08
jeblairfungi, clarkb, mattoliverau: i got a full packet capture and i see duplicate acks, retransmissions, window updates and zero window packets03:11
mgagnemordred: although more cumbersome to manage, it's possible to list projects you want to replicate (with the gerrit plugin) using the remote.NAME.projects config
jeblairfungi, clarkb, mattoliverau: i'm leaning toward the idea that we're hitting a condition where nodepool can't accept any more data, and the send is blocking03:13
mgagnemordred: might be useful for people with inconsistent naming scheme where the gerrit project name does not match the one on the destination03:14
*** harlowja_ is now known as harlowja_away03:16
*** zhiyan is now known as zhiyan_03:16
*** reed has joined #openstack-infra03:20
*** dstanek is now known as dstanek_zzz03:21
*** zhiyan_ is now known as zhiyan03:22
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for bug 1308715
uvirtbotLaunchpad bug 1308715 in nova/icehouse "Deadlock on quota_usages" [High,Fix committed]
*** plars has quit IRC03:30
*** plars has joined #openstack-infra03:31
*** wenlock_ has quit IRC03:35
openstackgerritA change was merged to openstack-infra/gear: Log admin requests
*** zhiyan is now known as zhiyan_03:41
mmaglanahi, i'm trying to setup the jenkins job for aviator ( and was wondering if i can define somewhere in the config a specific ruby version (1.9.3 and 2.1.0) to be used for running tests?03:42
*** arnaud has joined #openstack-infra03:44
*** nosnos has quit IRC03:47
*** yfried has joined #openstack-infra03:50
*** pcrews has quit IRC03:55
mattoliveraujeblair: nice work. That sounds very plausible. It seems to fit what we are seeing in the strace and the symtoms.03:56
*** jamespage has quit IRC04:00
*** gnuoy has quit IRC04:00
*** gnuoy has joined #openstack-infra04:00
*** jamespage has joined #openstack-infra04:00
*** david-lyle has joined #openstack-infra04:01
mgagnemmaglana: afaik, you can't precisely select the ruby version. rvm/rbenv aren't installed on slaves. You can however (at least) use ruby1.9 to run your tests.04:08
mgagnemmaglana: example of a job/macro using ruby1.9:
mmaglanathanks mgagne. unfortunately, activesupport 3.2.8 is incompatible with ruby 1.9.1. it needs 1.9.3 or higher. i will try to downgrade activesupport though to see if it works.04:11
mgagnemmaglana: although the binary is named "ruby1.9.1", it is in fact a weird choice made by Debian. 1.9.1 represents the ABI version, no ruby itself. 1.9.3 could very well be installed on the slaves.04:11
mmaglanamgagne: oh!04:11
mgagnemmaglana: indeed04:11
*** afazekas has quit IRC04:12
mmaglanamgagne: nice. thanks for that!04:12
mmaglanaslaves are running precise, right?04:13
*** zhiyan_ is now known as zhiyan04:13
mgagnemmaglana: by default, yes, not sure if they setup trusty slaves yet04:13
mmaglanamgagne: as long as it's not Lucid, I'm a happy camper! :)04:13
mgagnemmaglana: they usually aim for LTS releases04:14
mmaglanamgagne: cool. thanks a lot for your help!04:14
mgagnemmaglana: yw, can't wait for aviator to be properly setup at stackforge =)04:14
*** matsuhashi has quit IRC04:15
mgagnemmaglana: no trusty in view:
mmaglanamgagne: same here! :)04:15
*** yfried has quit IRC04:16
*** reed has quit IRC04:23
*** marcoemorais has joined #openstack-infra04:25
*** matsuhashi has joined #openstack-infra04:26
*** nosnos has joined #openstack-infra04:27
*** praneshp has joined #openstack-infra04:27
*** praneshp_ has joined #openstack-infra04:29
*** praneshp has quit IRC04:32
*** praneshp_ is now known as praneshp04:32
*** marcoemorais1 has joined #openstack-infra04:32
*** marcoemorais has quit IRC04:34
*** otherwiseguy has joined #openstack-infra04:35
*** lcheng_ has joined #openstack-infra04:36
*** afazekas has joined #openstack-infra04:40
*** wenlock_ has joined #openstack-infra04:44
openstackgerritValeriy Ponomaryov proposed a change to openstack-infra/config: Enabled pylint job for manila project
*** thomasbiege has joined #openstack-infra04:51
*** yfried has joined #openstack-infra04:51
*** yfried has quit IRC04:53
*** zhiyan is now known as zhiyan_04:53
*** wenlock_ has quit IRC04:56
*** david-lyle has quit IRC04:57
*** david-lyle has joined #openstack-infra04:57
*** david-lyle has quit IRC05:02
*** wenlock_ has joined #openstack-infra05:06
openstackgerritA change was merged to openstack-infra/pypi-mirror: Add a pointer to the DIB pypi element's README
*** wenlock has quit IRC05:06
openstackgerritMark Maglana proposed a change to openstack-infra/config: Define Jenkins Job for Aviator
*** thomasbiege has quit IRC05:09
*** thomasbiege has joined #openstack-infra05:12
*** zhiyan_ is now known as zhiyan05:14
openstackgerritMark Maglana proposed a change to openstack-infra/config: Define Jenkins Job for Aviator
*** krtaylor has quit IRC05:15
*** thomasbiege has quit IRC05:17
*** W00dy__ has quit IRC05:18
*** otherwiseguy has quit IRC05:21
*** Longgeek has joined #openstack-infra05:23
*** zhiyan is now known as zhiyan_05:25
*** alexpilotti has quit IRC05:26
*** arnaud has quit IRC05:32
*** pcrews has joined #openstack-infra05:34
*** wenlock has joined #openstack-infra05:35
*** ildikov has quit IRC05:43
*** renlt has joined #openstack-infra05:43
*** starmer has joined #openstack-infra05:46
*** rlandy has joined #openstack-infra05:47
*** rcarrillocruz has joined #openstack-infra05:49
*** rcarrill` has quit IRC05:51
*** markmcclain has joined #openstack-infra05:52
*** resker has joined #openstack-infra05:54
*** talluri has joined #openstack-infra05:54
*** markmcclain1 has joined #openstack-infra05:54
*** alexpilotti has joined #openstack-infra05:54
*** wenlock has quit IRC05:55
*** markmcclain has quit IRC05:57
renltHi, I'm now maintaining a project under stackforge, I want to create a new stable branch, but it seems that it does not work as a normal github repository does, which just to type a branch name to create..Would anyone shed some light on this, please? Thanks in advance!05:57
*** esker has quit IRC05:57
*** alexpilotti has quit IRC05:59
*** markmcclain1 has quit IRC06:01
*** andreykurilin has joined #openstack-infra06:05
*** _nadya_ has joined #openstack-infra06:06
*** alexpilotti has joined #openstack-infra06:08
*** ildikov has joined #openstack-infra06:13
*** dkehn has quit IRC06:20
*** dkehn has joined #openstack-infra06:22
*** _nadya_ has quit IRC06:28
*** _nadya_ has joined #openstack-infra06:30
*** _nadya_ has quit IRC06:38
openstackgerritA change was merged to openstack-infra/config: Fix rally-cli tests
rcarrillocruzoops, wrong window06:42
*** e0ne has joined #openstack-infra06:44
*** boris-42 has quit IRC06:45
*** wenlock_ has quit IRC06:46
*** doude has joined #openstack-infra06:48
*** e0ne has quit IRC06:48
*** boris-42 has joined #openstack-infra06:48
*** yfried has joined #openstack-infra06:48
*** e0ne has joined #openstack-infra06:49
*** e0ne has quit IRC06:49
*** yfried has quit IRC06:49
*** yfried has joined #openstack-infra06:49
*** e0ne has joined #openstack-infra06:50
*** rdopiera has joined #openstack-infra06:50
*** yfried_ has joined #openstack-infra06:50
*** renlt has quit IRC06:51
*** e0ne has quit IRC06:51
*** e0ne has joined #openstack-infra06:51
*** lcheng_ has quit IRC06:52
*** yfried has quit IRC06:54
*** e0ne has quit IRC06:56
*** yfried_ is now known as yfried06:57
*** rcarrill` has joined #openstack-infra06:59
*** praneshp has quit IRC06:59
*** rcarrillocruz has quit IRC07:01
*** skolekonov has joined #openstack-infra07:02
*** e0ne has joined #openstack-infra07:05
*** _nadya_ has joined #openstack-infra07:07
*** derekh_ has joined #openstack-infra07:08
*** jhesketh has quit IRC07:10
*** e0ne has quit IRC07:10
*** flaper87|afk is now known as flaper8707:10
*** e0ne has joined #openstack-infra07:11
*** e0ne has quit IRC07:13
*** pblaho has joined #openstack-infra07:13
*** rcarrill` is now known as rcarrillocruz07:15
openstackgerritSergey Skripnick proposed a change to openstack-infra/config: Make rally-install jobs voting
*** zhiyan_ is now known as zhiyan07:18
*** e0ne has joined #openstack-infra07:20
*** jlibosva has joined #openstack-infra07:25
*** _nadya_ has quit IRC07:26
*** zhiyan is now known as zhiyan_07:27
*** jcoufal has joined #openstack-infra07:28
*** e0ne has quit IRC07:36
*** e0ne has joined #openstack-infra07:36
*** e0ne has quit IRC07:39
mattoliverauNight all, have a great day/night07:39
*** mrda is now known as mrda-away07:43
openstackgerritSpencer Krum proposed a change to openstack-infra/config: Logstash: Modifying rewrite rules to allow kibana 3
nibalizerjesusaurus: clarkb at long last you have your one character change -_- (sorry, ive been busy doing stuff)07:46
openstackgerritSpencer Krum proposed a change to openstack-infra/config: Logstash: Modifying rewrite rules to allow kibana 3
*** ihrachyshka has joined #openstack-infra07:51
*** Hal_ has joined #openstack-infra07:52
*** ihrachyshka has quit IRC07:53
*** yfried has quit IRC07:53
*** markmcclain has joined #openstack-infra07:55
*** jgallard has joined #openstack-infra07:55
*** ihrachyshka has joined #openstack-infra07:56
*** andreykurilin has quit IRC07:56
openstackgerritafazekas proposed a change to openstack-infra/elastic-recheck: Add query for 1320617
*** jpich has joined #openstack-infra07:58
openstackgerritStephan Renatus proposed a change to openstack-infra/config: Use strainer for chef checks when possible
*** markmcclain has quit IRC08:00
openstackgerritChristian Berendt proposed a change to openstack-infra/config: cleaning up index.rst file
*** yfried has joined #openstack-infra08:04
*** Hal_ has quit IRC08:07
*** Hal_ has joined #openstack-infra08:08
*** derekh_ has quit IRC08:09
*** yfried has quit IRC08:09
*** derekh_ has joined #openstack-infra08:11
*** habib has joined #openstack-infra08:12
*** matsuhashi has quit IRC08:14
*** matsuhashi has joined #openstack-infra08:15
*** matsuhashi has quit IRC08:17
*** matsuhashi has joined #openstack-infra08:17
*** andreaf has joined #openstack-infra08:18
*** doude has quit IRC08:20
*** yfried has joined #openstack-infra08:22
*** doude has joined #openstack-infra08:26
*** IvanBerezovskiy has joined #openstack-infra08:27
*** mmaglana has quit IRC08:27
openstackgerritNikita Konovalov proposed a change to openstack-infra/storyboard: Project Groups API
*** amotoki has joined #openstack-infra08:30
*** habib has quit IRC08:35
*** habib has joined #openstack-infra08:39
*** markmcclain has joined #openstack-infra08:50
*** isviridov|away is now known as isviridov08:53
*** doude has quit IRC08:56
*** e0ne has joined #openstack-infra08:58
*** ominakov has joined #openstack-infra08:59
*** e0ne_ has joined #openstack-infra09:03
*** pelix has joined #openstack-infra09:03
*** ildikov_ has joined #openstack-infra09:03
*** ildikov has quit IRC09:04
*** e0ne has quit IRC09:06
*** marcoemorais1 has quit IRC09:07
*** jp_at_hp has joined #openstack-infra09:12
*** melwitt has quit IRC09:18
*** zhiyan_ is now known as zhiyan09:19
*** rlandy has quit IRC09:21
*** yamahata has quit IRC09:22
*** thomasbiege has joined #openstack-infra09:24
*** markmcclain has quit IRC09:24
*** zhiyan is now known as zhiyan_09:28
*** _nadya_ has joined #openstack-infra09:32
*** Hal_ has quit IRC09:33
*** zhiyan_ is now known as zhiyan09:37
*** zzelle has joined #openstack-infra09:37
*** thomasbiege has quit IRC09:37
*** Alexei_987 has joined #openstack-infra09:39
*** Hal_ has joined #openstack-infra09:41
*** Ajaeger has joined #openstack-infra09:43
*** unicell has quit IRC09:45
*** zhiyan is now known as zhiyan_09:46
*** unicell has joined #openstack-infra09:47
*** Hal_ has quit IRC09:47
*** habib has joined #openstack-infra09:47
*** jooools has joined #openstack-infra09:48
*** boris-42 has quit IRC09:52
*** habib has quit IRC09:53
*** habib has joined #openstack-infra09:53
*** markmcclain has joined #openstack-infra09:54
*** thomasbiege has joined #openstack-infra09:55
*** habib has quit IRC09:56
*** habib has joined #openstack-infra09:56
*** thomasbiege has quit IRC09:57
*** maxbit has quit IRC10:02
*** unicell has quit IRC10:04
openstackgerritDaniel Berrange proposed a change to openstack-infra/git-review: Fix typo in manpage s/gireview/gitreview/
openstackgerritA change was merged to openstack-infra/git-review: Correct git review -l over http(s)
*** boris-42 has joined #openstack-infra10:15
*** ihrachyshka has quit IRC10:18
openstackgerrityolanda.robla proposed a change to openstack-infra/storyboard-webclient: Fixes misaligned comments on very long lines Using <span> container also for main text, and switching orders of it Closes-Bug: #93
*** e0ne_ has quit IRC10:21
*** e0ne has joined #openstack-infra10:22
openstackgerrityolanda.robla proposed a change to openstack-infra/storyboard-webclient: Fixes misaligned comments on very long lines Using <span> container also for main text, and switching orders of it Closes-Bug: #93
*** e0ne has quit IRC10:26
*** wenlock has joined #openstack-infra10:31
*** jgallard has quit IRC10:33
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Fix django_openstack_auth translation setup
*** wenlock has quit IRC10:40
*** ihrachyshka has joined #openstack-infra10:41
*** thomasbiege has joined #openstack-infra10:41
openstackgerritmark mcclain proposed a change to openstack-infra/reviewstats: Update list to reflect recent changes to core team
*** matsuhashi has quit IRC10:48
*** thomasbiege has quit IRC10:49
*** markmcclain has quit IRC10:50
openstackgerrityolanda.robla proposed a change to openstack-infra/config: Checks if requirements overlap between head and branch reqs
*** markmc has joined #openstack-infra10:56
*** e0ne has joined #openstack-infra10:59
*** ildikov has joined #openstack-infra11:00
*** ildikov_ has quit IRC11:01
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: propose_translation_update: Handle .tx/config in git
*** dizquierdo has joined #openstack-infra11:02
*** e0ne has quit IRC11:04
*** _nadya_ has quit IRC11:07
*** lcostantino has joined #openstack-infra11:08
*** _nadya_ has joined #openstack-infra11:12
*** saschpe has left #openstack-infra11:20
*** thomasbiege has joined #openstack-infra11:22
*** andreykurilin has joined #openstack-infra11:28
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Remove duplication in translation scripts
*** e0ne has joined #openstack-infra11:32
*** habib has quit IRC11:32
YorikSarzzelle: Hello. Around?11:36
zzelleYorikSar, yes11:36
YorikSarzzelle: I've found out that the only external dependency (except argparse) in git-review is requests added by you.11:37
YorikSarzzelle: Is it really needed?11:37
AjaegerThe requirements proposal job currently fails, fixed via - any requirements core around to approve this, please?11:37
zzelleYorikSar, requests is used for http queries, hooks download11:38
YorikSarzzelle: It's just that the argument "git-review should have minimal set dependencies" is kind of wrong already...11:38
zzelleYorikSar, because urllib (or urllib2) does not check https certificates11:38
*** andreykurilin_ has joined #openstack-infra11:39
YorikSarzzelle: Ah that's why it's there.11:39
*** andreykurilin has quit IRC11:39
zzelleYorikSar, i wasn't sure, i remember it was a trouble for fungi or clarkb to introduce requests11:39
zzellethe security reason was more important than the miminal dependencies set11:40
Ajaegeranteaya: ok, will add to it ;)11:41
anteayaGheRivero might be up11:41
Ajaegerjust fixing another fallout...11:41
anteayajust showing you who the cores are11:41
Ajaegeranteaya: thanks!11:41
anteayamarkmc: might be around as well11:41
YorikSarzzelle: Yeah... I wonder if switching to using six is not that bad then.11:42
zzelleYorikSar, if we switch to six, we should target to switch everything11:43
GheRiverohi anteaya11:43
openstackgerritAndreas Jaeger proposed a change to openstack/requirements: Rename climate -> blazar
zzelleYorikSar, that's why i am not really "fan" of the proposed change11:43
YorikSarzzelle: Sure. I wonder if Christian is up to it.11:43
AjaegerAlex_Gaynor, GheRivero, dhellmann, markmc, sdague: and are needed to unbreak the requirements proposal job, could you review, please?11:44
GheRiveroAjaeger: on my way11:44
AjaegerThanks, GheRivero !11:44
anteayahey GheRivero, thanks11:45
YorikSarzzelle: By the way. I wonder why there was no release of git-review for a long time.11:45
sdagueAjaeger: +211:45
*** e0ne_ has joined #openstack-infra11:45
*** dims has joined #openstack-infra11:45
zzelleYorikSar, if the change solves a bug, it's fine, i am ok with the change base. if it is cosmetic i prefer we avoid to add a new requirement11:46
YorikSarzzelle: It looks like there's a lot of sweet changes waiting in master but never used by broad audience.11:46
AjaegerThanks a lot everybody!11:46
zzelleYorikSar, fungi is assigned a bug on releasing git-review11:46
YorikSarzzelle: My guess is that it's just a cosmetic change.11:47
*** _nadya_ has quit IRC11:47
zzelleYorikSar, which is a good idea with the upgrade of gerrit11:47
*** dstanek_zzz is now known as dstanek11:47
Ajaegerwow, seldom had a patch that quickly in - thanks, GheRivero, sdague, markmc!11:47
anteayaSergeyLukjanov: so this patch merged:
YorikSarzzelle: Oh, I didn't know there's a bug on this already.11:49
anteayaany objection to me submitting a patch to sync openstack-infra with the governance/programs.yaml?11:49
anteayaor did you want to do that?11:49
*** e0ne has quit IRC11:49
uvirtbotLaunchpad bug 1322365 in git-review "git-review 1.24 (or newer) needs to be released" [High,In progress]11:49
zzelleYorikSar, perhaps we should wait for the release before switching to six ?11:50
*** miqui has quit IRC11:51
YorikSarzzelle: As long as it doesn't fix anything, yes.11:51
*** thomasbiege has quit IRC11:51
YorikSarzzelle: Otherwise I think a new dependency should be introduced as early as possible.11:52
*** _nadya_ has joined #openstack-infra11:52
zzelleYorikSar, exactly11:52
zzelleYorikSar, i will ask Christian if the change is cosmetic and after discuss with fungi the release target11:53
YorikSarzzelle: We have another change waiting for new patchset:
YorikSarzzelle: btw, Christian's change somehow didn't make it to Zuul.11:55
YorikSarzzelle: There's no Jenkins comments at all.11:55
zzelleYorikSar, it seems to be linked to zuul troubles (yesterday)11:56
*** _nadya_ has quit IRC11:56
*** rlandy has joined #openstack-infra11:57
zzelleYorikSar, could be interesting to add to the release ?11:57
zzelleYorikSar, i am also asking myself if we should provide support for WIP before release ?11:58
YorikSarzzelle: I guess we can either wait for both these changes to land or just release more often :)11:59
YorikSarzzelle: I still don't understand why there was no releases for 9 months.12:00
zzelleYorikSar, don't know ... perhaps for normal git-review usage over ssh git-review 1.23 is enough12:01
*** markmcclain has joined #openstack-infra12:02
YorikSarzzelle: Well... There were some bugs like '(detached' topic.12:02
YorikSarzzelle: But it looks like they were fixed only recently.12:03
zzelleYorikSar, indeed12:03
*** markmcclain1 has joined #openstack-infra12:04
*** _nadya_ has joined #openstack-infra12:04
vponomaryovGuys from infra core, could you please review commit:
*** markmcclain1 has quit IRC12:04
zzelleYorikSar, i should leave, we can discuss with fungi tomorrow afternoon ?12:04
YorikSarzzelle: Sure12:04
*** markmcclain1 has joined #openstack-infra12:04
YorikSarzzelle: I think I'll be still around.12:04
zzelleYorikSar, bye12:04
*** habib has joined #openstack-infra12:06
*** markmcclain has quit IRC12:06
*** yaguang has quit IRC12:09
*** dprince has joined #openstack-infra12:10
ildikovanteaya: clarkb: are you around?12:11
*** habib has quit IRC12:12
*** habib has joined #openstack-infra12:13
anteayaildikov: I am12:13
anteayatoo early for clarkb yet12:13
anteayahow are you making out?12:13
ildikovanteaya: hi12:13
ildikovanteaya: ah, ok, I'm still not 100% aware of the timezones :)12:13
anteayait is okay12:14
anteayaonce you learn everyone's they travel somewhere12:14
anteayayou never win on the timezone question12:14
ildikovanteaya: so it seems that there is no problem with tox, I found the log with the Error message yesterday for the patch I showed12:14
ildikovanteaya: yeap, that's true :)12:14
anteayahave you identified where the issue lies?12:14
ildikovanteaya: in this log file:
ildikovanteaya: sorry: it's better:
ildikovanteaya: so in the linked lines you can see some error messages12:15
ildikovanteaya: any time I saw these errors, the build was always finished in the way that Sphinx generates all the doc it can12:16
anteayayes I can see some error messages12:16
ildikovanteaya: and leaves out those, that have formatting issues12:16
anteayamostly for me that the log has error messages but the build finishes success12:16
ildikovanteaya: so the doc is generated, but some parts are missing, for instance here the Cinder table, as I showed yesterday12:17
ildikovanteaya: yes, so I do not know if I can analyze the logs in that gate job or what could be the best solution for this issue12:17
anteayaokay so at a certain point there was action taken about logs12:18
anteayasince every project has a different interpretation about what WARNING means and so on12:18
ildikovanteaya: I didn't find a stop in case of errors config option, maybe it would worth some more time, but if there is no option like this, is there any other way to identify a partially failed build?12:18
*** aysyd has joined #openstack-infra12:18
*** pdmars has joined #openstack-infra12:18
anteayasome projects were setting everything to a certain level and the logs were getting way to big12:18
*** pblaho has quit IRC12:18
anteayaI wonder if the setting for the log has been adjusted and needs to be evaluated12:19
*** krtaylor has joined #openstack-infra12:19
*** pdmars has quit IRC12:19
anteayaildikov I would suggest asking sdague to help you understand the issue with logging levels in the hopes that will help you understand how logging is handled for ceilometer12:19
*** mbacchi has joined #openstack-infra12:20
ildikovanteaya: but it's Sphinx log and not Ceilometer12:20
*** thomasbiege has joined #openstack-infra12:20
*** julim has joined #openstack-infra12:21
ildikovanteaya: I mean I cannot modify the log levels easily and also, in the script, the doc build script is ran and then the result of it is success in most of the cases and then the deals with the generated docs, even if it is generated only partially12:21
anteayawell I would have to dig into sphinx logging then since I know nothing about it12:22
anteayaI need to eat something first12:22
ildikovanteaya: I guess the made it work this way because of a reason, I mean that the build does not fail in case of a formatting error12:23
*** weshay has joined #openstack-infra12:23
ildikovanteaya: *they* made12:23
*** smarcet has joined #openstack-infra12:24
ildikovanteaya: sure, but if sdague knows more about this maybe, then I will annoy him a bit :)12:25
ildikovanteaya: do you know when he will be available?12:25
openstackgerritA change was merged to openstack-dev/hacking: Updated from global requirements
*** jgallard has joined #openstack-infra12:29
openstackgerrityolanda.robla proposed a change to openstack-infra/storyboard: Added new priority changed event for tasks Defined new constants and created new methods to generate event
*** pdmars has joined #openstack-infra12:32
*** ArxCruz has joined #openstack-infra12:32
anteayaildikov: he is online now, he is probably busy with something else atm12:35
anteayayeah, so my question is why does the build succeed when error messages are logged12:35
anteayaI am thinking that is a quality of one of our jobs, not necessarily sphinx, but I don't know12:36
fungizzelle: YorikSar: i'm going to try and get git-review polished up for a 1.24 release this week12:36
*** markmcclain1 has quit IRC12:36
anteayaildikov: for instance I am going to start in here:
ildikovanteaya: if you run the sphinx-build command it succeeds in case of this formatting error12:38
*** dims has quit IRC12:39
ildikovanteaya: so it is nothing to do with the gate job, I mean it receives the 'build succeeded' verdict from the sphinx command12:39
anteayaokay so then the issue may reside with sphinx-build logging12:39
*** eharney has joined #openstack-infra12:39
anteayahave you chatted with any of the sphinx devs at all?12:40
SergeyLukjanovanteaya, hey, mornging12:40
*** andreaf has quit IRC12:41
anteayaSergeyLukjanov: morning/afternoon12:41
anteayaSergeyLukjanov: what do you want to do about sync'ing infra with governance/programs.yaml?12:41
ildikovanteaya: not yet, that's the next step on the list, if analyzing of the logs in the gate job is not an option12:41
anteayaildikov: I think you have demonstrated to me that this issue resides with sphinx12:41
anteayathe logs in the gate job are accessible to you with every test run12:42
anteayayou get all we have12:42
fungiit could reside within pbr. the build_sphinx target is provided by pbr, which then in turn calls out to the sphinx tools12:42
ildikovfungi: hi12:42
anteayaI didn't know pbr might have an effect on how build_sphinx interprets errors in the build log12:43
openstackgerrityolanda.robla proposed a change to openstack-infra/storyboard: Added new priority changed event for tasks
anteayawould pbr still be playing a role if ildikov gets the same result by running build-sphinx?12:43
fungiildikov: hi. i'm looking at pbr right now to see whether maybe it's had some recent changes to its sphinx hooks... though i'm not particularly familiar with that part of the code so it's unlikely i'll find an obvious cause12:44
ildikovfungi: you mean that this issue could be solved in pbr maybe?12:44
*** thomasbiege has quit IRC12:44
fungianteaya: well, the result is that whatever's invoking the sphinx build is existing with a o return code afterward, which tox interprets as a success and so the job passes12:44
fungier, a 0 return code12:45
*** talluri has quit IRC12:45
*** W00dy__ has joined #openstack-infra12:45
jd__fungi: does an import problem with lockfile.linklockfile on the py33 nodes rings a bell to you?12:46
*** lcostantino has quit IRC12:46
jd__something like that
fungijd__: i asked a search engine friend of mine, and he responded with
jd__fungi: uh, I'm lost as the developer says it's fixed in 0.9.1 which is what is installed12:50
*** nosnos has quit IRC12:50
SergeyLukjanovanteaya, I'll sync infra projects today to gov12:50
ildikovfungi: hmm, I do not know how can Sphinx configured in that way to return something else in case of an error, if even it exists12:50
anteayaSergeyLukjanov: okay12:50
anteayaanyone have anotherjesse's email address12:51
SergeyLukjanovanteaya, /me was thinking about automating it, but it's not always could be determind12:51
fungijd__: yeah, i too find that confusing... could be we're looking at a regression somewhere12:51
anteayaSergeyLukjanov: yeah, going to be tough to get the regex right12:51
ildikovfungi: is there a chance maybe to analyze the logs and searching for error and mark the gate job as 'Failed' in this case12:51
ildikovfungi: ?12:52
*** prad has joined #openstack-infra12:52
jd__fungi: oh it's something more pervert I think12:52
fungiildikov: it's supposed to fail build_sphinx on errors, i believe (and we also do something to make sure it catches warnings too)12:52
StevenKI'm guessing that is populated by openstack-manuals, but I can't work out which bit of openstack-manuals does it.12:52
*** miqui has joined #openstack-infra12:53
fungiStevenK: it's api jobs for the individual projects, not the manuals12:53
StevenKfungi: Ah ha, I shall paw through -infra/config then12:54
*** zehicle_at_dell has quit IRC12:54
fungiStevenK: yeah, you should see plenty of api jobs12:54
ildikovfungi: I use the pure sphinx-build command locally, when I update some docs and it succeedes, even if there are some formatting errors in the logs12:54
ildikovfungi: or did you mean pbr or some other parts of the gate job, that should fail?12:55
fungiildikov: it's also possible this is due to the sphinx configuration in ceilometer12:55
fungiildikov: the process should return nonzero on error12:55
jd__fungi: it's simpler to explain on a paste so here's my current thoughts
*** jcoufal has quit IRC12:56
fungiildikov: i would start by comparing to the same file in other projects to see if maybe there's something missing for that12:57
*** jcoufal has joined #openstack-infra12:57
StevenKfungi: Strangely, not one for ironic12:58
fungiStevenK: "baremetal" maybe?12:58
anteayaanotherjesse doesn't have an email in launchpad, hmmmm12:58
StevenKfungi: No dice for that either12:58
ildikovfungi: ok, I will give it a try, tnx :)12:59
anteayaStevenK: maybe ask in the ironic channel?12:59
anteayaStevenK: they might know where their docs are12:59
*** gibi has joined #openstack-infra12:59
StevenKI know where they are, I'm trying to work out how they get to
*** skolekonov has quit IRC13:00
anteayalook at the post jobs in the zuul/layout.yaml file?13:01
*** pafuent has joined #openstack-infra13:01
*** fifieldt has quit IRC13:02
*** _nadya_ has quit IRC13:02
fungiStevenK: aha, there are separate docs projects for those...
fungiet cetera13:02
*** rfolco has joined #openstack-infra13:03
StevenKBut not for ironic, but it seems to be done via the openstack-client-publish-jobs template13:03
anteayaStevenK: ironic doesn't have any post docs jobs:
StevenKanteaya: But python-ironicclient does13:04
*** dkliban_afk is now known as dkliban13:04
*** pafuent has quit IRC13:05
*** andreykurilin_ has quit IRC13:05
fungiStevenK: okay, so the project entry for openstack/ironic in zuul's layout includes the openstack-server-publish-jobs project template which has {name}-docs in the post pipeline, which ends up expanding to ironic-docs as a job name13:06
*** dstanek is now known as dstanek_zzz13:07
fungiStevenK: in the jjb config tree there's a openstack-publish-jobs.yaml which has the job template for {name}-docs13:09
*** yfried has quit IRC13:10
fungiStevenK: it in turn calls out to the "docs" builder in macros.yaml13:10
*** mriedem has joined #openstack-infra13:10
fungiStevenK: which runs '/usr/local/jenkins/slave_scripts/ {github-org} {project}'13:10
StevenKfungi: So where I'm really going with this is trying to get things like os-apply-config and friends up there, which doesn't fit into either -server or -client. However, it's >2300 and I need beauty sleep, so I shall dig more into all the crumbs you've fed me tomorrow.13:11
fungiStevenK: that should be enough to get you going after you sleep on it13:11
openstackgerrityolanda.robla proposed a change to openstack-infra/storyboard-webclient: Added priority on task row
*** bknudson has quit IRC13:11
fungijd__: i'm wondering if we have a system site packages negative interaction going on... looking to see if the tox.ini for gnocchi has any obvious clues13:12
jd__fungi: I set sitepackages=False, but yeah, I'm surprised. I've sent a patch with a requirement specifying lockfile>=0.9.1 instead of just lockfile, to see13:12
fungijd__: yeah, agreed. so it shouldn't be touching the system installed lockfile anyway13:13
*** _nadya_ has joined #openstack-infra13:13
*** _nadya_ has quit IRC13:13
jd__yes, there's something weird or so big I miss it :(13:13
fungijd__: i think i see the problem13:15
*** mburned is now known as mburned_out13:16
fungi-rw-rw-r-- 1 jenkins jenkins 16949 Mar 25  2013 /srv/static/pypi/openstack/lockfile/lockfile-0.9.1.tar.gz13:16
fungi4e4c7ea4c4301500da5e7f3b51e01cfe  /srv/static/pypi/openstack/lockfile/lockfile-0.9.1.tar.gz13:16
jd__HAN HAN13:16
fungijd__: then look at
jd__released on 23 Aug yeah13:17
jd__do we have a time traveling machine?13:17
fungiseems the lockfile authors think it's perfectly okay to fix bugs and reuse the same version numbers when they rerelease13:17
jd__sounds like that indeed :(13:17
jd__sounds like we solved a mystery /nick scoobidoo13:18
annegentleruh roh raggy13:18
fungiit was old man weathers underneath that monster mask all along13:18
annegentlefungi: lol13:18
fungiand he would have gotten away with it too if it weren't for us nosy developers13:18
*** signed8bit has joined #openstack-infra13:19
*** jaypipes has quit IRC13:20
fungijd__: so anyway, the proper thing to do here is to "encourage" the lockfile authors to bump their release number, then adjust your requirements to insist on >= that new release13:22
*** dstanek_zzz is now known as dstanek13:23
*** pafuent has joined #openstack-infra13:23
jd__fungi: I can do that, but can we reset the mirror in the meantime?13:23
fungijd__: since there's really no telling who has already cached/shipped/baked in the previous 0.9.1, so just papering over it in our mirror is probably a bad idea13:23
openstackgerritRadomir Dopieralski proposed a change to openstack-infra/config: Add gerritbot to the #tuskar channel
*** jaypipes has joined #openstack-infra13:23
jd__fungi: fair enough13:23
openstackgerritAnita Kuno proposed a change to openstack/requirements: Add barbican and python-barbicanclient to global-requirements
fungijd__: better, i think, if we know that certain things are broken for people with the "old" 0.9.113:24
fungiwhich we won't know any longer if we replace it with the "new" 0.9.113:24
*** signed8bit has quit IRC13:25
*** signed8bit has joined #openstack-infra13:25
*** rgerganov has joined #openstack-infra13:25
jd__fungi: FYI
*** lcostantino has joined #openstack-infra13:26
fungijd__: aha! i was looking in their google code issue tracker linked from pypi, which is apparently old13:27
jd__yeah they forgot to change the page from pypi…13:27
jd__suckerz :)13:27
*** W00die has quit IRC13:32
*** mfer has joined #openstack-infra13:32
*** bknudson has joined #openstack-infra13:34
*** prad_ has joined #openstack-infra13:34
anteayaRyan_Lane seems to have adjusted the timeout for wiki signin13:35
anteayaI have more than 5 minutes at a time now13:35
*** prad has quit IRC13:37
*** prad_ is now known as prad13:37
*** alexpilotti has quit IRC13:39
*** homeless has joined #openstack-infra13:39
*** zz_gondoi is now known as gondoi13:40
*** gokrokve has joined #openstack-infra13:40
*** pafuent has quit IRC13:40
*** gokrokve has quit IRC13:40
*** gokrokve has joined #openstack-infra13:41
*** W00die has joined #openstack-infra13:42
fungiildikov: i've confirmed that the build_sphinx target provided by pbr does pass warningiserror=True into application.Sphinx() and then tries to build() and reraises any exceptions from that except for the docutils.utils.SystemMessage class13:44
*** gabriel-bezerra has joined #openstack-infra13:44
gabriel-bezerraGuys, doesn't check-grenade-dsvm update the entire devstack repository?13:45
fungiildikov: but if it was from the ignored class there should have been a message starting with "reST markup error:" in the log (dumped to stdout by pbr)13:45
ildikovfungi: wow :)13:45
fungier, to stderr13:45
fungii'm starting to wonder if this job isn't using pbr for sphinx doc generation13:46
fungiexcept it must be, because ceilometer's doesn't have a build_sphinx target of its own13:47
gabriel-bezerraMy patch changes the line 'enable_apache_site horizon.conf' to 'enable_apache_site horizon' but the tests fail with the old line in the logs13:47
ildikovfungi: yes, I've checked in Ceilo13:47
fungigabriel-bezerra: it updates code but keeps old configuration13:47
*** rgerganov has quit IRC13:48
fungigabriel-bezerra: so i think any attempt to modify the configuration may get ignored, if that's what you were trying to accomplish there (hard to know without additional context)13:48
gabriel-bezerraIt also introduces a new function to devstack/lib/apache, and there is another breakage with "apache_site_config_for not found"13:48
ildikovfungi: but I haven't seen "reST markup error"13:49
gabriel-bezerrasearch apache_site in that log13:49
*** mburned_out is now known as mburned13:49
gabriel-bezerrafungi: ^13:49
fungiildikov: agreed, so any exceptions from sphinx should have been reraised and that call should have returned a nonzero exit code13:49
fungigabriel-bezerra: oh, a change *to* devstack. i think grenade should be starting with your proposed change in place in that case13:50
ildikovfungi: where is it set that what is reraised and what not?13:52
gabriel-bezerraIs there anything I can do to fix that?13:52
*** otherwiseguy has joined #openstack-infra13:52
*** andreaf has joined #openstack-infra13:52
*** pafuent has joined #openstack-infra13:53
*** habib has quit IRC13:53
*** jcoufal has quit IRC13:54
*** jcoufal has joined #openstack-infra13:55
*** jcoufal has quit IRC13:56
*** jcoufal has joined #openstack-infra13:56
*** habib has joined #openstack-infra13:56
*** habib has quit IRC13:57
*** habib has joined #openstack-infra13:57
*** gokrokve has quit IRC13:58
*** gokrokve has joined #openstack-infra13:58
fungigabriel-bezerra: is this the issue?
fungigabriel-bezerra: /usr/sbin/apachectl: 87: ulimit: error setting limit (Operation not permitted)13:59
ildikovfungi: tnx, looking14:00
*** yfried has joined #openstack-infra14:00
*** amotoki has quit IRC14:00
*** BadCub has joined #openstack-infra14:01
fungiildikov: also, dhellmann and/or mordred may have some insight into sphinx+pbr failure modes and where we should be looking14:01
*** gokrokve has quit IRC14:02
ildikovfungi: I checked this run() function14:03
fungiildikov: oops, sorry, not run() but _sphinx_run()14:04
ildikovfungi: so as a short summary, did you mean that for the utils.SystemMessage class the exceptions are not reraised and they should appear as reST markup error?14:04
dhellmannildikov: I've lost some backtrace, do you have a link to the error log?14:04
*** andreykurilin has joined #openstack-infra14:04
ildikovfungi: yes, sorry I was lazy to write here the correct one :)14:04
fungidhellmann: quick summary, trying to figure out why that ^ job didn't fail14:04
ildikovdhellmann: in this case the docs gate job is green as the sphinx build is marked as succeeded, but the malformed table is of course missing from the generated doc14:05
dhellmannfungi: ah, got it14:05
gabriel-bezerrafungi: /opt/stack/new/devstack/lib/keystone: line 91: apache_site_config_for: command not found14:05
fungi(which has devolved into a dissecting of pbr's sphinx hooks)14:05
*** kaisero has joined #openstack-infra14:05
ildikovfungi: so if the not reraised exceptions would be the problem then they should be marked with the "reST markup error" I guess, or I misunderstood what you wrote above14:06
fungigabriel-bezerra: ahh, yep that would also be a problem. looks like lib/apache may not be sourced yet at that point?14:06
gabriel-bezerrafungi: I think that what really breaks is this line
gabriel-bezerrafungi: It works on 2 vms I used to test14:06 worked there14:07
gabriel-bezerrathat line with the ulimit error disappeared when I used sudo in a later patch14:07
gabriel-bezerrathe line with enable_apache_site is also wrong, my patch removed the .conf suffix and it shows up in the log14:08
gabriel-bezerrait seems that my patch was not applied before those tests run14:09
gabriel-bezerraor partially applied?14:09
ildikovfungi: so it seems that 'warnerrors' is not in 'option_dict' in run(), but maybe it's a wrong direction14:10
fungiildikov: yeah, the warnerrors point was somewhat tangential... you also have actual errors in that log, so i would have expected it to fail regardless14:11
gabriel-bezerraor it may have been applied on the wrong moment14:11
W00dieHi guys, I'm working with gerrit provided by openstack on a server and I'm kind of new, Is there a way to run the "manage-projects" command from jeepyb manually?14:12
*** mengxd has joined #openstack-infra14:12
*** gokrokve has joined #openstack-infra14:12
dhellmannildikov, fungi : the clean_exit function in looks suspicious14:13
*** markmcclain has joined #openstack-infra14:13
dhellmannalso the fact that the venv environment is spinning up mongo to build docs is likely a problem14:13
fungigabriel-bezerra: oh! somehow i forgot that devstack does actually have stable branches... so the initial devstack detup does not have your patch applied. you need to be able to upgrade from the tip of stable/icehouse to your patch on master (which means the initial devstack setup lacks your patch, but then it's applied by grenade during the upgrade)14:14
ildikovdhellmann: it was a next step on my list to fix it14:14
pafuentHi. I would like to change the topics of Blazar IRC channels, but I get "You're not a channel operator". How I should change the topics? I need to became a channel operator? The change could be achieved through a patch in infra?14:14
ildikovdhellmann: I asked about my options here yesterday and clarkb mentioned this issue, that it shouldn't be needed there14:15
dhellmannildikov: changing tox.ini to not run that script doesn't fix the issue locally; I still see warnings that are not converted to errors14:15
fungiW00die: we do sometimes run manage-projects manually. have you tried, and what isn't working for you?14:15
ildikovdhellmann: I do not know if we use this venv to anything else, I did not have the time to check it14:15
*** maxbit has joined #openstack-infra14:16
dhellmannildikov: that env is meant to be used by infra jobs, so if the ceilometer team needs a generic env for running commands with the test environment configured we should add one with a different name14:16
ildikovdhellmann: sphinx-build itself reports that the build succeeded even if there were some formatting errors during the build process14:16
dhellmannildikov: yeah, I'm looking at how that works now14:16
anteayapafuent: SergeyLukjanov is the person you need to talk to14:17
gabriel-bezerrafungi: what should I do, then?14:17
ildikovdhellmann: so the simple sphinx-build behaves this way14:17
pafuentanteaya: Thanks, I'll contact him14:17
gabriel-bezerrafungi: how am I going to possibly get that merged, if my patch doesn't come on the right moment?14:18
*** resker has quit IRC14:18
*** esker has joined #openstack-infra14:19
ildikovdhellmann: as for the venv, you mean that venv should remain for doc and another generic one should be created for any other case?14:19
*** dstanek is now known as dstanek_zzz14:19
dhellmannildikov: that venv is used for several different CI jobs (docs, packaging, etc.), so yes, we should add a separate one for running commands that needs mongo and the other test environment stuff set up14:19
*** yamahata has joined #openstack-infra14:20
*** msabramo has quit IRC14:20
ildikovdhellmann: ok, great, I just wanted to be sure that we think about the same thing before modifying anything in tox.ini14:20
ildikovdhellmann: as for sphinx, fungi pointed me to pbr:
*** e0ne_ has quit IRC14:21
dhellmannildikov: yes, and I noticed that ceilometer is using an old version of sphinx so I'm updating now to see if that fixes the problem14:21
*** e0ne has joined #openstack-infra14:21
dhellmannthe warnings I am seeing are not going through the sphinx function that raises the exception when warningiserror is true14:22
ildikovdhellmann: well, it worth giving it a try, I mean the Sphinx update14:22
fungigabriel-bezerra: what is it that you want to do? there's a good chance that you'll need to fix this in all supported versions, or provide some means to upgrade the configuration between devstack versions14:23
ildikovdhellmann: you mean the warnings of logs of the sphinx job?14:23
fungigabriel-bezerra: you might want to bring it up with the devstack developers in #openstack-qa for better guidance on what they want14:23
gabriel-bezerrafungi: I want to fix the way apache sites are configured, because it doesn't work on Ubuntu 14.04 the way it is now.14:24
*** david-lyle has joined #openstack-infra14:24
anteayahmmmm, I offered a patch to add barbican and python-barbicanclient to requirements/global-requirements but the tests failed since it can't find a barbican14:24
dhellmannildikov: when I run "tox -e venv -- python build_sphinx" I see warnings for values missing from the glossary and for some formatting errors in measurements.rst but none of those cause the doc build to exit with an error code14:24
anteayabarbican isn't on pypi, what needs to happen to get it there?14:25
anteayado we know?14:25
ildikovdhellmann: a-ha, ok, I see those also14:25
gabriel-bezerrafungi: So I put all the distribution dependent code in a single place: lib/apache, and created a new function that returns the full path of the site configuration file.14:25
dhellmannildikov: in fact, I see an "ERROR: Unexpected indentation" which should trigger the failure, too14:26
rdopieraanteaya: I saw a pypi job defined in some projects, I would guess you need a correct and that job14:26
rdopieraanteaya: or you can do 'python register' maually14:26
anteayardopiera: cool thanks14:26
rdopieraanteaya: and add openstackci to the owners14:26
fungigabriel-bezerra: right, i mostly understand the intent of the patch, just not how it fits with the life cycle of the devstack stable branches and what the upgrade story is for it14:26
dhellmannildikov: if I run ".tox/py27/bin/sphinx-build doc/source doc/build" I get an error that it can't import oslo.messaging14:27
anteayardopiera: to the owners of the pypi package?14:27
rdopieraanteaya: yeah14:27
ildikovdhellmann: yes, that is what I noticed earlier and why I reported a bug about it, because usually no one checks the output of the docs gate job, if it succeeds14:27
rdopieraanteaya: in the 'roles' tab at pypi14:27
ildikovdhellmann: do you run it inside the virtual env?14:28
anteayardopiera: cool, thanks not sure if that is infra's job or barbican's14:28
mriedemsdague: figured out why we weren't hitting that cinder db_sync issue with rhel 6.5 and python-six 1.5.214:28
gabriel-bezerrafungi: Neither do I. What do you mean by "life cycle of the devstack stable branches" and the " upgrade story"?14:28
rdopieraanteaya: I suppose that pypi jenkins/gerrit job actually does register and upload14:28
fungianteaya: does barbican need to be on pypi? is it a server (we don't release servers to pypi currently) or a client/library?14:28
mriedemwe're shipping python-oslo-messaging 1.3.0a9 which didn't have this change:
dhellmannildikov: yes14:29
*** markmcclain has quit IRC14:29
fungigabriel-bezerra: saying i'm not sure how the devstack developers usually handle upgrades between branches for changes to the way service configuration is handled, so you should ask them14:29
dhellmannildikov: oslo.messaging wasn't installed in the virtualenv for some reason14:29
anteayafungi: the server14:29
dhellmannildikov: let's move to #openstack-ceilometer to continue this14:29
gabriel-bezerrafungi: ok, I'll check that14:29
ildikovdhellmann: ok14:30
fungianteaya: you'll be shocked to discover that we don't even release nova to pypi ;)14:30
anteayathe test failed because it couldn't find a barbican to download14:30
anteayaI'm shocked14:30
fungianteaya: which test?14:30
anteayaonly cause you told be to be though14:30
anteayaNo distributions at all found for barbican (from -r /tmp/tmpddm9LL (line 7))14:31
anteayawhere does it have to be to be found?14:31
fungianteaya: out of curiosity, why? your commit message just says "the projects should be included in the requirements repo" without any explanation for what other package on pypi is depending on them14:32
fungianteaya: or did you mean to add them to projects.txt rather than global-requirements.txt?14:32
W00diethanks fungi, I was waiting for "manage-proyects" to run automatically but it never created my project, now that i ran it manually it did created it :D14:32
anteayathe fact they are incubated is insufficient for them to be included in requrirements?14:32
*** jcoufal has quit IRC14:33
fungiW00die: right, we trigger runs of manage-projects from our puppet master when configuration is updated14:33
*** markmcclain has joined #openstack-infra14:33
anteayaI looked at other patches, projects were added to global-requirements14:33
anteayaso that is what I did14:33
anteayaplease do explain the difference?14:33
*** dstanek_zzz is now known as dstanek14:34
mriedemafazekas: added more information here:
uvirtbotLaunchpad bug 1316328 in devstack "Stack trace when hits "cinder-manage db sync"" [Undecided,Fix released]14:34
mriedemafazekas: is RDO building olso.messaging packages from trunk?14:34
fungianteaya: projects.txt is the file which defines which projects should be forced to use requirements which match our global-requirements.txt and which will get requirements updates proposed when the global requirements list changes14:35
fungianteaya: global-requirements.txt is an aggregate list of the projects used by other projects in openstack as python dependencies14:35
anteayain that light then, I had intended them to be in global-requirements14:35
W00diefungui, is there a sugested documentation for connecting gerrit with github?, I see on the openstack doc that it is possible but i dont see much info on how it should be, I'm trying this:14:36
W00diebut I'm not sure if its outdated14:36
anteayaif they are incubated, they should be available for use by other projects, should they not?14:36
fungianteaya: which brings me back to the question, what other openstack projects are using those as python dependencies and listing them in their individual requirements.txt or test-requirements.txt files?14:36
anteayanone that I know of14:37
anteayaisn't the fact they are incubated sufficient?14:37
fungisufficient for what?14:37
anteayafor inclusion in global-requirements?14:37
anteayaI can abandon my patch, if you prefer14:37
*** rbowen has joined #openstack-infra14:37
*** wenlock has joined #openstack-infra14:38
fungito drive home the point, please take note of which of the following projects are listed in global-requirements.txt: nova, swift, glance, cinder, keystone...14:38
*** wenlock has quit IRC14:38
rbowenI wonder if someone can point me to a contact for the OpenStack youtube channel. There's a video - - that has the audio track wrong, and I was hoping we could get that fixed.14:38
anteayaI did14:39
*** rgerganov has joined #openstack-infra14:39
anteayait sounds like you want me to abandon the patch, I will14:39
*** wenlock has joined #openstack-infra14:39
fungianteaya: yes, please abandon the patch. packages get added to global-requirements.txt when they are needed to install other openstack projects. it is not intended to act as a registry of official projects14:40
*** andreykurilin has quit IRC14:41
fungirbowen: try contacting and he hopefully can get it corrected14:41
rbowenThank you.14:41
*** mrodden has joined #openstack-infra14:41
gabriel-bezerrafungi: how does it work? It starts with the stable branch, then run,, then applies the branch, then,
*** doude has joined #openstack-infra14:43
*** msabramo has joined #openstack-infra14:45
fungigabriel-bezerra: it seems to be described at
fungigabriel-bezerra: so it doesn't unstack... looks like it stops services individually and runs upgrade scripts for each of them, then starts them again14:46
fungibased on the description there14:46
fungigabriel-bezerra: though again, it may be more fruitful to talk to the grenade developers in #openstack-qa if you have specific questions about it. #openstack-infra is more about the servers all this stuff runs on, not so much what gets run within a job14:47
*** thedodd has joined #openstack-infra14:48
*** rgerganov has quit IRC14:49
*** alexpilotti has joined #openstack-infra14:50
gabriel-bezerraoh, ok. thanks fungi.14:50
*** jgrimm has joined #openstack-infra14:52
*** thedodd has quit IRC14:53
*** thedodd has joined #openstack-infra14:53
*** markmcclain has quit IRC14:55
*** habib has quit IRC14:55
*** gokrokve has quit IRC14:56
*** lcheng_ has joined #openstack-infra14:57
*** gokrokve has joined #openstack-infra14:57
*** radez_g0n3 is now known as radez14:58
phschwartzmorning infra15:02
*** gokrokve has quit IRC15:02
*** mengxd has quit IRC15:02
*** rdopiera has quit IRC15:03
afazekasmriedem: I do not know, what is packaged. I need to check the spec files. But probably one of the release tar balls is packaged.15:03
* afazekas I need to hurry, I got on urgent SMS..15:03
*** otter768 has joined #openstack-infra15:04
* afazekas will be back in ~1h15:05
*** otter768_ has quit IRC15:05
krtayloranteaya, from yesterday's meeting - did we decide on a third party string? Or did we just decide to agree on one15:07
anteayawe decided to agree on one15:08
anteayaI am working on that over here:
anteayathere are some issues15:09
anteayafor starters you will see starting on line 23 that gerrit confirms the following account names15:09
*** andreykurilin has joined #openstack-infra15:09
anteayaon line 3 from an email by HenryG there are some account names that don't match up with what gerrit has in the third party grouping15:10
*** afazekas has quit IRC15:10
anteayaalso on line 8 you will see from something called gerrymander a list of accounts which I can match to gerrit names15:10
anteayafor instance rocktown15:10
anteayawhere is that coming from?15:10
*** alexpilotti has quit IRC15:11
anteayaI was just about to find out what Daniel P. Berrange's irc nick was if I could15:11
anteayaso I could ask him15:11
anteayahi phschwartz15:11
krtayloranteaya, berrange maybe?15:12
phschwartzanteaya: I will be trying to get the wip patch I have done today for you to review. I got held up due to a back injury :(15:13
anteayaI don't see that nick in here15:13
*** signed8bit has quit IRC15:13
anteayaphschwartz: oh no, I hope you are okay15:13
anteayaphschwartz: how are you feeling?15:13
*** signed8bit has joined #openstack-infra15:13
krtayloranteaya, danpb15:13
krtayloranteaya, so, this is account names prefix?15:14
phschwartzanteaya: In pain, need to make a dr appointment. I hurt it jumping shoulder first into rocks in the waves at the beach saving my friends 2 year old that got rolled up and pulled under.15:14
anteayakrtaylor: what do you mean account names prefix?15:14
*** matrohon has quit IRC15:14
anteayathis is what I get when I go to gerrit groups to look up account names15:15
fungikrtaylor: i think the idea is that we have some particular pattern we'll use to rename them and for future names (whether that includes a common prefix, suffix or both)15:15
anteayakrtaylor: note there is a subgroup in that group15:15
anteayaI am stumped on rocktown, bsn, pattabi-ayyasami-ci15:16
krtayloranteaya, account names organization and standardization is good, but we were talking ML filtering, I was just looking for a tag15:16
fungikrtaylor: since there will also be other criteria we'll potentially want to place on the names for these (for example, length?) the actual visible names will be chosen by the infra team when the accounts are created based on input from the operators requesting them15:16
anteayaCi for one and Bot for another15:16
mikahm, is pointing to openstack with a 404 - anyone of you aware where the most recent version is supposed to be?15:17
krtaylorfungi, yes, that sounds great15:17
fungianteaya: in previous conversations (this has been coming up for about a year, initially suggested by HenryG) we may want things like "OpenStack Zuul Bot" for official systems and "IBM PowerVC Bot" for unofficial/third-party systems15:18
krtaylorwell, if all third party testing is CI then CI works15:18
*** ihrachyshka has quit IRC15:18
* krtaylor is looking at cinder15:18
anteayamika: looks like 0.2.1 was released last August:
mikaanteaya: i see, thanks!15:20
fungiso that people can filter on a from header where the name starts with "OpenStack" for official/upstream messages and anything ending with "Bot" for all automated systems15:20
*** UtahDave has joined #openstack-infra15:20
fungisomething along those lines anyway15:20
anteayaI agree that our systems should be using OpenStack in the account name15:20
anteayagot the sense that accounts would either use CI or Bot depended on how the definition split,15:21
mikaanteaya: hm, would be nice to have a new release of python-jenkins, do you know whom we could ping about that? :)15:21
anteayanot having all third party ci named CI Bot15:21
anteayamika: lets look at the log to see who contributes to it15:22
mikaanteaya: Marc Abramowitz mainly15:22
anteayajames page did the last release15:22
fungianteaya: yeah, i think we should factor out all but one common term just to keep things short, but probably keep a common prefix which will allow a simple filter to still catch all (official and unofficial) messages from automated systems in one go15:22
anteayayes either one of those two15:22
fungior suffix, or something15:22
anteayamsabramo: you are marc abramowitz if i remember correctly?15:23
fungiand this will work great until we get a contributor whose last name is actually "Bot" ;)15:23
anteayawe will convince that person to change their name15:23
*** Ajaeger has quit IRC15:23
anteayait will be easier15:24
krtaylorthats a good start, I'd also like to see [Third Party xyz] in the subject like [Nova]15:24
msabramoanteaya: yes, that is correct15:24
*** wenlock_ has joined #openstack-infra15:24
anteayamsabramo: great, mika has some python-jenkins questions for you15:24
fungikrtaylor: subject mangling will be harder, i think, since gerrit decides that for us15:24
msabramoanteaya: OK, thanks15:25
msabramomika: you have some questions about python-jenkins?15:25
mikamsabramo: no questions, just a new (tagged) release would be awesome to see it show up in distributions :)15:25
fungikrtaylor: chances are that would require a new gerrit feature upstream in an upcoming release, or some sort of mail proxy which transforms messages emanating from the gerrit server15:25
krtaylorfungi, true for gerrit, but for general discussion, just getting everyone to agree to use Third Party would help15:26
msabramomika: Ah yes, I agree. I don't have the power to do that. I'm not sure if that would be zaro or James E. Blair or who15:26
fungikrtaylor: sure15:26
mikamsabramo: buddy of mine noticed that 0.2.1-0.1 as available in debian has problems e.g. with spaces in job names and this seems to be fixed in git already15:26
mikamsabramo: i see, any chance you could talk to them?15:26
msabramomika: I have proposed a lot of changes that have been merged, but I don't have the ability to approve changes, create release tags, etc.15:26
mikaah :)15:26
fungimika: msabramo: anteaya: right now (since it was recently imported) the python-jenkins-ptl group in gerrit with permission to push tags to the project is empty. probably should add jeblair at a minimum15:27
*** signed8bit has quit IRC15:27
msabramofungi: Yeah python-jenkins was recently brought into stackforge by James E. Blair. Before that it lived for a while on Launchpad with some different authors and maintainers15:27
*** signed8bit has joined #openstack-infra15:28
fungimsabramo: i'm unsure what the project structure is for stackforge/python-jenkins, but i'm fairly sure we can work out here who needs to push tags and add them. input from jeblair would be helpful at any rate15:28
*** wenlock_ has quit IRC15:28
msabramo+1 to jeblair being able to push python-jenkins, or if he doesn't want it, he can suggest someone else15:29
mikajey, thanks folks :)15:29
msabramopython-jenkins git master could probably use a bit more testing before we push it15:29
derekh_nodepool seems to be creating instances on tripleo-test-cloud over and above quota, any idea if there are loads of instance running it doesn't know about?15:30
msabramomika: have you been testing the version from git master?15:30
mikamsabramo: not yet15:30
anteayakrtaylor: I don't have any hope these folks will use the [Third Party][3rd party][third-party] email tag even if there is any consenseous from any group on what the tag should be15:30
anteayakrtaylor: these people genuinely believe that they are unique, not part of a group that already exists15:31
fungiderekh_: it's entirely possible that nodepool thinks things got deleted when they actually stuck around (during the ups and downs you've had). i'll see what the alien list looks like for those regions15:31
*** gokrokve has joined #openstack-infra15:31
anteayait doesn't occur to them that there may already be a structure to address their concerns15:31
msabramojenkins-job-builder uses python-jenkins; is it possible in OpenStack CI infrastructure to make JJB use the python-jenkins from git so that it can be tested? Or do we have to push a new release of python-jenkins in order to pick it up and test it with JJB?15:31
anteayain the row of walls to bang my head on, I have given up on that one15:32
anteayajust as an fyi15:32
krtayloranteaya, sigh, yeah, but we can start with a cross project wiki page, requirements and make baby steps15:32
anteayaif you want to carry any battle standard, I support you15:32
anteayaoh yes, by all means15:32
anteayaI'm working on the wiki pages15:32
fungimsabramo: what you probably need is a new gate job for python-jenkins which exercises it with jenkins-job-builder, to confirm changes to the former work with the latter15:33
krtayloranteaya, I'll volunteer to be the first CI account page to work out any problems15:33
anteayaso I have listed all the accounts I have heard anything about on this etherpad:
derekh_fungi: cool, thanks, R2 seems to be running fins at about the rate I'd expect, R1 isnt being used at all, if we can get it back running for a couple of hours we'll now if the work lifeless did helped things15:33
anteayawe will need a seed account for the wikipages15:33
anteayaso thank you15:33
krtayloranteaya, it was so hard to participate on my mobile client yesterday, I would have volunteered then15:34
msabramozaro: fungi said "msabramo: what you probably need is a new gate job for python-jenkins which exercises it with jenkins-job-builder, to confirm changes to the former work with the latter". Is that something you'd like to do?15:34
anteayathanks for tuning in15:34
*** pblaho has joined #openstack-infra15:35
*** pblaho has quit IRC15:35
krtayloranteaya, are you ready for prime time on the wiki template?15:35
*** msabramo has quit IRC15:35
fungianteaya: it's worth noting that the lists of accounts people coughed up on the ml were based on historical data, but there are definitely accounts which have been renamed upon request over time15:35
*** pblaho has joined #openstack-infra15:35
*** andreykurilin has quit IRC15:36
fungianteaya: so not being able to find some of them is an expected outcome15:37
anteayakrtaylor: no, not yet, I got some good feedback from yesterday's meeting that I need to incorporate, plus I have the template for openstack programs I need to put together15:37
krtaylorcan't we just use the current gerrit account list?15:37
*** jgallard has quit IRC15:37
anteayafungi: hmmm, okay, is rocktown ringing any bells for you?15:37
anteayakrtaylor: use the current gerrit account list for the name format discussion?15:38
krtayloranteaya, ok, I'll wait, let me know if there is anything I can help with15:38
fungianteaya: yeah, i believe it was renamed. you may need to beat the infra ml and openstack-ci bug lists with a search engine to figure out why/when though15:38
krtayloranteaya, enumeration15:38
anteayayes but if you look at line 5 there isn't even agreement between the gerrit group and the name that shows up in comments15:38
fungikrtaylor: yes, i think the list of third-party ci and voting third-party ci group members, plus any official accounts we happen to recognize which we didn't think of originally ought to be sufficient15:38
anteayaah I was hoping to talk to danpb15:39
anteayaI don't think I care that much to go chasing the infra ml and openstack-ci bug lists15:39
anteayaI also think part of the exercise is account cleanup and removing stale accounts if they are stale15:40
*** Sukhdev has joined #openstack-infra15:40
anteayabut that is a sub goal not a prime goal15:40
anteayaif they are still active and are overlooked hopefully they will make themselves known15:40
fungianteaya: right. suffice to say we can use the lists people provided on the ml as reminders of things to check into, but we already mostly have what we need to build a correct list ourselves anyway15:40
anteayaso right now I would go with line 19 and down on
*** msabramo has joined #openstack-infra15:42
anteayamorning zaro15:45
SergeyLukjanovzaro, morning15:45
*** gyee has joined #openstack-infra15:46
*** e0ne has quit IRC15:47
*** e0ne has joined #openstack-infra15:48
*** e0ne_ has joined #openstack-infra15:49
*** HenryG has quit IRC15:49
fungiderekh_: found 57 alien nodes in that region aside from the te-broker, and the names of all of them suggest they were at some point created by nodepool. i'm removing them directly via novaclient now15:49
derekh_fungi: ok, let me know if you need any reset, or want me to delete any, thanks15:50
fungiderekh_: well, novaclient is being very slow to respond to me on the deletes, so it's possible it's going to time out getting a response for them15:51
*** e0ne has quit IRC15:52
*** HenryG has joined #openstack-infra15:52
zaromsabramo: setting up a job to test both directions  python-jenkins <-> jjb is probably a good idea.  i'll see if i can propose a change for that.15:53
derekh_fungi: ok, if it does, will see if I can find out why15:54
anteayajeblair: when you are about:
*** afazekas has joined #openstack-infra15:54
anteayalet me know if you want additional expansion on any of the items in the table15:55
fungiderekh_: the list of instances i'm trying to delete is
*** HenryG has quit IRC15:55
*** HenryG has joined #openstack-infra15:55
fungiderekh_: all my nova delete calls went through without error, so now i'm just waiting for them to actually disappear from the nodepool alien-list output (11 remaining)15:56
derekh_fungi: k, thanks, will check in a few minutes to see whats left15:56
msabramozaro: cool, thanks!15:57
msabramozaro: also note that mika was asking about a new release of python-jenkins15:58
*** gokrokve has quit IRC15:58
*** gokrokve has joined #openstack-infra15:58
*** derekh_ has quit IRC16:00
*** marcoemorais has joined #openstack-infra16:00
*** otherwiseguy has quit IRC16:00
*** derekh_ has joined #openstack-infra16:01
openstackgerritcraigtracey proposed a change to openstack-infra/config: Add nova-docker to the gerritbot config
*** doude has quit IRC16:02
notmynamejeblair: ttx: FYI storage polices patches have been proposed to master in swift. soft freeze in effect16:03
*** gokrokve has quit IRC16:03
* anteaya is afk for a bit16:03
*** marcoemorais has quit IRC16:04
*** marcoemorais has joined #openstack-infra16:04
pleia2good morning16:05
*** marcoemorais has quit IRC16:05
*** marcoemorais has joined #openstack-infra16:05
*** atiwari has joined #openstack-infra16:06
derekh_fungi: thanks, the region is back in use now16:06
*** arunkant has joined #openstack-infra16:07
*** marcoemorais has quit IRC16:07
*** marcoemorais has joined #openstack-infra16:07
*** hogepodge has quit IRC16:08
*** hogepodge has joined #openstack-infra16:09
*** jooools has quit IRC16:10
*** crc32 has joined #openstack-infra16:11
fungiderekh_: excellent16:11
ttxnotmyname: ack16:12
mordredmorning all16:12
*** crc32 has quit IRC16:12
arunkantHi..I am trying to add new package dependency for addressing a keystone issue. This is my first time for requesting change in global requirements so want to see if following the process correctly. llink:
zaromsabramo, mika : i'm not privileged to make a new release.  best one to ask for that would be jeblair16:14
*** IvanBerezovskiy has left #openstack-infra16:15
pleia2the jboss licensing/version world is definitely not my forte, anyone have thoughts on JBossAS vs JBossERB re: for our needs?16:16
fungiarunkant: at a quick look, seems like a reasonable change request though you've placed your additional requirement in the section which is commented as being for test-time-only dependencies (while your keystone review implies it's a run-time dependency) and you don't have it inserted in alphabetical order (which most of the rest of the contents seem to be in)16:16
pleia2naturally I'm inclined toward the one we don't seem to need a license for, but if it's going to make zanata unmanagable...16:16
fungipleia2: i think we ought to push for non-commercial infrastructure. if zanata doesn't work without the proprietary jboss distribution, then we should help or encourage them to fix that problem16:17
clarkbpleia2: whatever is available in ubuntu16:17
arunkantfungi: I will address those concerns in next patch..did not knew about the order and section..thanks for looking into this.16:18
pleia2fungi: ok, open source high road it is :)16:18
fungiarunkant: sure, glad to help (keep in mind that i don't really exercise my requirements core rights for anything which isn't infrastructure related or a gate emergency, so you'll still need to solicit actual reviews from the real reqs core reviewers)16:19
*** e0ne_ has quit IRC16:19
pleia2clarkb: yeah, JBossAS is the one in the Ubuntu repos (JBossERB is installable though, but it's a tarball kind of deal)16:19
bodepd_jesusaurus: I'm getting started on the splits today. Can you create the empty projects that I need?16:19
*** e0ne has joined #openstack-infra16:19
*** Hal_ has joined #openstack-infra16:20
mordredlook, it's a bodepd_16:20
*** otherwiseguy has joined #openstack-infra16:20
bodepd_I know. I exist. It's crazy.16:20
mordredI mean, I'm not sure I believe you EXIST16:21
fungipleia2: a prime tenet of our infrastructure work is that we build our systems on top of components which can be reused by anyone, and basing some of it on a proprietary platform gets in the way of that16:21
* mordred agrees with fungi16:22
pleia2fungi: so I don't think their enterprise version is actually *proprietary* it's just subscription-based and you need a license (like redhat)16:22
bodepd_mordred: I don't even think I saw you at the summit, which is evidence that you might not exist as well16:22
fungipleia2: so "high road" is perhaps not entirely accurate, but i do think we should help other projects be more free ;)16:22
pleia2 says GPL16:22
*** dprince has quit IRC16:22
mordredbodepd_: that's fair. I think I heard that you were there though...16:23
pleia2and we did look into getting redhat licenses before going with centos, so I wasn't sure where our line was16:23
*** starmer has quit IRC16:23
fungipleia2: weird. how can they enforce subscription fees to obtain it? or is it just that it's freely obtainable from other sources but to get it directly from them you have to pay?16:23
*** derekh_ has quit IRC16:23
mordredpleia2: well, we were rather uncomfortable with that, but it was really just to test that openstack worked _on_ rhel16:24
*** e0ne has quit IRC16:24
afazekasmriedem: pong16:24
pleia2fungi: I don't really know how that works16:24
mordredI don't think we should run any pieces of the actual infrastructure on things that requires licenses or subscriptions16:24
mroddenmordred: not sure if it still matters but i put out a 0.1.1 bash8 with that ignore (-i) fix16:24
mordredmrodden: woot! thanks!16:24
fungipleia2: because if it's really all released under the gpl, then someone else ought to be able to publish copies of it for free if they want (perhaps we could offer to mirror it for them for free?)16:25
mordredmrodden: I was starting to run it on some of our things to check it out - and we have a _bunch_ of bad indentation :)16:25
openstackgerritA change was merged to openstack/requirements: Remove hp3parclient from global-requirements
mroddenmordred: i totally understand :)16:25
openstackgerritA change was merged to openstack/requirements: Rename climate -> blazar
mroddeni haven't had time to put it to work against our internal stuff yet16:25
mroddeni'm not prepared for that yet16:26
mroddenalso, bash8 still needs some work apparently16:26
*** Ivo has quit IRC16:26
pleia2mordred: so that's AS - the community version they said zanata may not do so great on16:27
mordredthat's EAP16:27
clarkbas is packaged for ubuntu (also maybe we make this a trusty server to see what breaks? and agent master will break)16:27
pleia2heh, EAP is at the top16:27
pleia2oh, that's a confusing page16:27
mordredyeah it is16:27
fungipleia2: and asp is listed as lgpl!16:28
fungipleia2: perhaps zanata was suggesting that they could get us a technical support license for it?16:28
mordredI _think_ maybe we can mvn install it from there?16:28
*** arnaud has joined #openstack-infra16:28
clarkbmordred speaking of trusty, I think I am ready to start pushing your fixes through16:28
pleia2fungi: well their email said it "doesn't have all the required libraries and versions" so I think it was more than support16:29
clarkbwell review and so on16:29
clarkband babysit16:29
mordredclarkb: oh. wait - you're goign to review them? that means I'll have to fix things ...16:29
clarkbmordred: yes :)16:29
clarkbactually iirc they have had a round of fixes and really just need babysitting now16:29
*** arnaud has quit IRC16:30
mordredthis is not helpful:16:30
fungipleia2: mordred: ewww... you have to create an account at before you can download the release archives or view source code16:30
openstackgerritArun Kant proposed a change to openstack/requirements: Adding ldappool module dependency as needed by keystone bug #1320997.
uvirtbotLaunchpad bug 1320997 in keystone "Common Ldap handler connection pooling" [Medium,In progress]
mordredmarkmc: get your people to make better pages ^^16:30
clarkbok afk to finish morning things. back soon. also I have dentist appointment at 2 I will afk for16:30
mordredfungi: is it a captive thing or a fake thing like mysql's?16:30
*** hashar has joined #openstack-infra16:31
*** arnaud has joined #openstack-infra16:31
mordredfungi: I supposed I can find that out...16:31
pleia2yeah, so I went the `apt-get install jboss` route when I got started, then went to the zanata install docs and it was all kinds of scary (written for redhat, assumed enterprise jboss)16:31
pleia2so right now is just my sanity check before I dive in there16:31
*** bearhands is now known as comstud16:32
*** otherwiseguy has quit IRC16:32
fungiweird that the jbossenterprisemiddleware/community-enterprise page is being served to me half in chinese16:32
mordredit's captive16:32
afazekasmriedem: I haven't seen six related issue with rdo icehouse, I had issue stable/icehouse devstack. The code part which referencing to six>=1.6 is exists in almost all openstack stable/icehouse component, but it is usually not called anywhere. The global requirements  and all project requirements should have been changed when the added the call for six>=1.6 feature.16:33
mordredI just logged in, got to the download link, and it includes an auth token16:33
mordredpleia2: I think we should soldier on with apt-get install jboss16:33
mordredand if there are 12 other java libraries we have to install - well then, that's what puppet is for :)16:33
mordredand we can make a nice self-contained zanata puppet module for folks16:34
pleia2ok, will do :)16:34
*** jpich has quit IRC16:34
fungiagreed. anything we can do to make zanata more consumable for others without needing to get enterprise subscriptions for things is a benefit to the community16:34
openstackgerritKhai Do proposed a change to openstack-infra/config: add python coverage job to check pipeline.
fungizaro: the commit message for that ^ could stand to mention what project it's for16:36
mordredbtw - I read the article W00die linked in above:
mordredit's rather strange - but I went ahead and poked at gerrithub so that I could play with the things16:37
mordredand I'm 100% confused as to what benefits any of it has16:37
jeblairfungi, pleia2, mordred: last time i looked, jboss is basically not installable on rhel without extra (extremely expensive) licenses.  afaict that's intentional due to business models.16:37
jeblairlike, there just plain aren't any packages that do it16:38
mordredthere is a plugin which will connect your gerrit to github - and there is a page you can go to to "import pull requests" - but it's manual - and it keeps the pull request open, and comments don't sync16:38
openstackgerritKhai Do proposed a change to openstack-infra/config: add python coverage job to check pipeline for JJB.
fungimordred: well, they seem to provide gerrit-as-a-service at least, so there's presumably some benefit to that part (the github integration is less interesting as far as i'm concerned)16:39
mordredfungi: oh - sure - that part is neat16:39
mordredI was more commenting on the point of the github plugin16:39
jeblairso yeah, i support the 'use ubuntu' route.  and failing that, asking them 'how do we do this on centos without proprietary licenses?', and failing that, working on pootle.16:39
mordredjeblair: ++16:39
mordredjeblair: given that apt-get install jboss is a thing, everything past that really should just be java, right?16:40
morganfainbergthere is a github plugin for gerrit that does... more than the replication stuff?16:40
morganfainberg... why?16:40
fungimorganfainberg: because of stockholm syndrome16:40
jeblairmordred: seems reasonable!  can't wait to find out why we're wrong about that.  ;)16:40
morganfainbergfungi, good answer.16:40
jeblairanteaya: the wiki page looks great, thanks16:41
mordredjeblair: well, we'll be wrong because of java16:41
fungimordred: it's enterprise-grade *and* web-scale!16:41
mordredfungi: it can only be web-scale if it uses mongodb. does it use mongodb?16:42
fungimordred: it might! it might...16:42
W00diehehe, doesn't openstack has it's repositories linked to github? I heard that what openstack does is close pull request and redirect them to gerrit16:42
mordredwow. that would be the best of both worlds - an "enterprise" grade java app that stores its data in mongo16:42
mordredW00die: we replicate to github using normal gerrit replication for dev convenience16:43
mordredW00die: and then we have a bot which closes the pull requests - that's right16:43
*** bhuvan has joined #openstack-infra16:43
fungiW00die: yep. there are some people who like to be able to "fork us over on github" so we replicate there. the pull-request-closer is more of a reaction to deal with people who think that because we mirror on github means anyone is actually watching their pull requests (one of the only bits you can't disable directly)16:44
*** pblaho has quit IRC16:44
fungiW00die: we disable issues, wiki, et cetera but there's no feature to disable pull requests or we would just do that instead16:45
*** jlibosva has quit IRC16:46
mordredwow. I'm learning all sorts of things today16:47
mordrednow - the real question is - why would that operate on the contents of the commit message and not on a tag?16:48
*** whoops has joined #openstack-infra16:48
mordredI mean, otherwise that's a great little service to run for folks16:48
openstackgerritDan Bode proposed a change to openstack-infra/config: Remove pip module from config
*** bodepd_ has quit IRC16:50
clarkbmordred gerrithub scares me. iirc the great jenkins github fubar was a result of gerrithub16:50
mordredclarkb: lovely16:50
*** gokrokve has joined #openstack-infra16:51
*** bodepd has joined #openstack-infra16:51
fungilooks like the reqs update proposer is broken since our project renames...
clarkbmordred I think your trusty stuff needs a rebase now that jesusaurus' refactor merged16:51
*** ociuhandu has joined #openstack-infra16:51
fungifatal: Project not found: openstack/database-api16:51
clarkbthar was what prevented merging earlier16:52
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Remove duplication in translation scripts
clarkbfungi needs a projects.txt update?16:52
fungiclarkb: looks like it--about to propose a patch now16:53
*** talluri has joined #openstack-infra16:53
*** Ajaeger has joined #openstack-infra16:54
*** otherwiseguy has joined #openstack-infra16:55
Ajaegerinfra team, I would appreciate a review of - the jobs site for translation of log messages. Thanks!16:55
*** _nadya_ has joined #openstack-infra16:56
openstackgerritA change was merged to openstack/requirements: Paramiko: bump to 1.13.0
*** talluri has quit IRC16:57
*** nati_ueno has joined #openstack-infra16:57
*** talluri has joined #openstack-infra16:57
openstackgerritJeremy Stanley proposed a change to openstack/requirements: Update projects list for May 23 project renaming
fungibackports of that ^ to stable branches on the way too16:58
mordredclarkb: okie16:58
mordredclarkb: the nodepool or the install_puppet or both?16:59
clarkbmordred: the change that touches slave.pp16:59
clarkbbecause it was moved16:59
*** rgerganov has joined #openstack-infra16:59
*** BadCub has quit IRC17:00
*** rdopiera has joined #openstack-infra17:02
*** talluri has quit IRC17:02
*** yamahata has quit IRC17:03
*** praneshp has joined #openstack-infra17:06
*** andreaf has quit IRC17:06
*** ominakov has quit IRC17:06
jesusaurusbodepd: which splits are you starting with?17:06
*** ildikov has quit IRC17:09
*** andreaf has joined #openstack-infra17:10
*** andreaf has quit IRC17:10
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Changes to get nodepool working with trusty
*** otherwiseguy has quit IRC17:11
mordredclarkb: ^^17:11
fungiSergeyLukjanov: if that lands, we still need to deal with the oslotest rename there and also get that backported to the stable/icehouse branch17:11
mordredbodepd: btw - the pip provider in puppet does not honor proxy settings17:11
SergeyLukjanovfungi, yup, I mean it's already on the go :)17:12
*** hashar has quit IRC17:12
mordredbodepd: there is a place inside where it makes a direct ruby xmlrpc query to check if something needs to be upgraded, and that does not pass through configured proxy settings17:12
*** andreaf has joined #openstack-infra17:12
mordredbodepd: I've been meaning to write a patch for about a year now and simply haven't17:12
fungiSergeyLukjanov: SergeyLukjanov yep. SlickNik: are you backporting 94722 to stable/icehouse?17:13
krtayloranteaya, we will rename all the public names for third party  accounts, do we want to do that before requiring everyone to create their ThirdPartySystems contact page?17:13
*** Hal_ has quit IRC17:13
krtayloranteaya, or is the rename only for new accounts, grandfathering in all existing17:13
Ajaegerfungi: do you want a backport of and - or will you make one patch that combines your patch with these two?17:14
Ajaegerfungi: missed your last message ;)17:14
AjaegerLet me backport 96144 now - ok?17:14
fungiAjaeger: oh, good point. i missed 96144 wasn't applied on icehouse yet17:15
fungiAjaeger: separate backports are better if they landed separately in master17:15
*** habib has joined #openstack-infra17:15
fungiAjaeger: i'll fix up 96205 and my backport to just take care of oslotest17:16
*** harlowja_away is now known as harlowja_17:16
*** marcoemorais has quit IRC17:16
*** dizquierdo has quit IRC17:16
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Update get-pip URL to the new location
*** marcoemorais has joined #openstack-infra17:16
Ajaeger is the backport17:16
Ajaegerfungi: ok17:16
*** dstanek is now known as dstanek_zzz17:17
openstackgerritJeremy Stanley proposed a change to openstack/requirements: Update projects list for oslo.test renaming
*** habib has quit IRC17:18
*** habib has joined #openstack-infra17:19
zarofungi: i want to run jjb tests when there are changes to python-jenkins to verify that the upstream dependency doesn't break jjb.  could you please help?17:20
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add an entry for a new puppetmaster for puppet3
*** talluri has joined #openstack-infra17:21
*** Longgeek has quit IRC17:22
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard-webclient: Configure Timeline from Story details
*** otherwiseguy has joined #openstack-infra17:22
fungizaro: sure, do we already have jjb jobs which use python-jenkins in some way (presumably with a mock jenkins of some kind)?17:22
zarofungi: python-jobs runs the jjb tests that uses python-jenkins.  is that what you mean?17:23
fungizaro: i suppose that's a start. how much of python-jenkins is actually exercised by that?17:23
fungizaro: enough to be able to at least tell whether or not changes to it will break jjb?17:24
clarkbnone of it17:24
*** Ryan_Lane has joined #openstack-infra17:24
*** vponomaryov has quit IRC17:24
clarkbthe tests only test yaml to xml compilation17:24
zarofungi: ohh yeah, clarkb is correct.17:25
fungizaro: so given clarkb's assertion there, you probably need something akin to the integration test we have for git-review changes17:25
*** vponomaryov has joined #openstack-infra17:25
fungizaro: basically, spin up a jenkins master and then try to inject/delete some job definitions17:26
zarozaro: ok. i gotha.17:26
fungizaro: once that's a thing, you can test jjb and python-jenkins that way if you want17:26
fungizaro: alternatively, if python-jenkins grew some mechanism to mock jenkins interactions, you could use that instead of an actual jenkins backend for the job17:27
fungizaro: the latter is preferable, but i have no idea how much work that would be to implement17:27
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard: Remove unnecessary files
*** mika has quit IRC17:28
*** mika has joined #openstack-infra17:30
fungii think nodepool may have grown very unhappy with us. the node count graph has plateaued17:30
SlickNikSergeyLukjanov / clarkb: Regarding I'm torn, and would like some input.17:30
fungiand we're running almost no jobs. looking into it17:30
funginodepoold is not running17:31
zaromsabramo: above ^ about JJB testing17:31
SlickNikSergeyLukjanov/clarkb: I'm using DEST=/opt/stack, because that's where the image-build scripts pulls down the OpenStack repos (trove / diskimage-builder).17:31
fungii'm starting it back up and then looking into the logs to see why it died17:32
clarkbSlickNik: wfm17:32
*** praneshp has quit IRC17:32
*** esker has quit IRC17:33
SlickNikSergeyLukjanov: I could use a different dir as DEST to avoid confusion with a devstack-gate job, but that means that we won't be consistent as to where we're pulling OpenStack bits across jobs.17:33
*** Ivo has joined #openstack-infra17:34
*** esker has joined #openstack-infra17:34
*** lcostantino has quit IRC17:34
fungi[Wed May 28 16:40:43 2014] Out of memory: Kill process 7105 (nodepoold) score 896 or sacrifice child17:34
fungi[Wed May 28 16:40:43 2014] Killed process 7105 (nodepoold) total-vm:20821676kB, anon-rss:7164608kB, file-rss:0kB17:34
AjaegerSlickNik: should I backport for you?17:34
*** praneshp has joined #openstack-infra17:35
*** Ryan_Lane has quit IRC17:35
bodepdmordred: do you guys carry a patch for the pip provider? That isn't related specifically to the stuff I was working on, but it's not an unreasonable thing for me to patch upstream.17:35
mordredbodepd: we don't - we don't have an proxy problems17:35
*** Ryan_Lane has joined #openstack-infra17:35
bodepdmordred: oh, I see that you guys have a pip3 provider17:35
mordredbodepd: we carry an alternate pip provider internally at HP - which I have a not-quite-working patch for in flight here17:36
bodepdjesusaurus: I was starting with pip today17:36
mordredbodepd: but that will probably just make you cry17:36
fungibodepd: we were working around the challenges of having pip install python 2.7 and 3.3 versions of things on the same system, since the backported 3.3 debs to precise came with a separate python3-pip17:36
*** UtahDave has quit IRC17:37
* mordred really wants to patch the pip provider to be 2/3 aware in general - and also to install pip using get-pip instead of distro packages17:38
mordredbut it's really above my current level of ruby zen17:38
SlickNikAjaeger: Sure, go for it.17:38
SlickNikAjaeger: looks like it merged this morning.17:38
*** esker has quit IRC17:38
*** UtahDave has joined #openstack-infra17:39
bodepdmordred: and being 2/3 aware is not sufficient  just to select a provider?17:39
AjaegerSlickNik: will do - yeah, it merged nicely17:39
bodepdmordred: you need to be able to specify both 2 and 3 packages?17:39
clarkbmordred: zen?17:39
mordredbodepd: yeah - we need to be able to tell it to install a package into the python3 install using pip, or into the python2 install using pip17:40
mordredbodepd: selecting a pip3 provider is fine, actually17:40
*** arnaud has quit IRC17:40
*** Sukhdev has quit IRC17:40
*** lcostantino has joined #openstack-infra17:40
bodepdmordred: is that built into pip, or is that something else (in ruby-land, it would be called rvm)17:40
bodepdmordred: people have this same issue with rvm, I would start there and see if someone has an elegant solution17:40
AjaegerSlickNik, fungi: is the backport17:41
mordredbodepd: totally different - you can have different parallel installs of python17:41
bodepdmordred: this class of problem was intended to be solved with composite namevars, but last I checked, they were a giant POS17:41
bodepdmordred: what is managing that process? Is it pip?17:41
mordredso, the pythons are installed via distro packages17:41
nibalizerrvm does enable sane multiple installs of ruby/rubygems17:41
nibalizerbut the virtualenv way of doing it (python) is totally different17:42
mordredpip itself simply installs things into the python env it was executed from17:42
bodepdyeah, that is what I was thinking of is virtualenv.17:42
mordredso if you had a /usr/bin/python3 and that was the python that invoked pip17:42
mordredit will cause the packages to be installed into /usr/local/lib/python3.3/site-packages17:42
bodepdthe easiest way to do it might be to create two types17:42
AjaegerSlickNik:No, it did not merge yet - was approved but not through the gate yet17:42
bodepdpip2 and pip317:42
mordredbodepd: virtualenv is the way that a lot of people deal with this issue - but in our case, we're installing some system level things that need to be server-wide17:43
mgagnebodepd: and what if you want to test against 2 versions of python3...17:43
* mgagne runs away17:43
bodepdmgagne: each resource declaration would only be for one version17:43
nibalizerbodepd: so this is where its at now:
mordredmgagne: yeah...17:43
nibalizermordred: did your massive pip change land yet?17:44
*** ildikov has joined #openstack-infra17:44
mordredbodepd: actually, (mgagne thanks) that's the reason in my brain I was thinking adding a version to the provider would be nice17:44
bodepdmgagne: if you need to test against both:
mgagnenibalizer: still a WIP according to the link above17:44
clarkbI am beginning to think that needs to come in bite size changes17:44
*** rgerganov has quit IRC17:44
mordredbecause you might want to install a package into python 2.6, python 2.7, python 3.2 or python 3.317:44
bodepdmordred: adding a version to the provider is a bit complicated, b/c of uniqueness17:44
mordredclarkb: I agree17:44
mordredbodepd: yah17:45
mgagnemordred: what about adding venv support to the package resource :D17:45
bodepdmordred: it puts you down the composite namevar path17:45
mordredultimately, I think we could live without that - it's jsut a completeness thing17:45
bodepdunless... it accepted a list of versions17:45
nibalizeroh yes bodepd if you would comment on mordreds patch that would be much appreciated17:45
mordrednibalizer: I think it needs work too17:45
mordrednibalizer: like, last time we tried it, it broke17:46
nibalizerfor anyone who doesn't know, bodepd literally wrote the book on types and providers17:46
mgagnenibalizer: and I'm still struggling to write code for them =)17:46
nibalizermordred: :< i guess the soft-freeze on infra/config didn't hold ?17:46
*** praneshp has quit IRC17:46
mikai assume it's also jeblair who releases new version of jenkins-job-builder? (nice stuff in git which isn't yet in the 0.7.0 release)17:47
bodepdmordred: nibalizer still looking, but I would say that is the 'easy' way to do it17:47
fungibodepd: yeah that was the other challenge i faced. if we wanted, for example, to have puppet install a package named "tox" for both python 2.7 and python 3.3, we'd have to find some way to work around the fact that puppet doesn't let you install two packages of the same name with different providers, due to nonuniqueness17:47
mgagnemika: zaro can too17:48
clarkbnibalizer: iirc heartbleed happened17:48
jesusauruslifeless: SpamapS: im trying to find a slide from a presentation that has a large diagram of an ideal CI/CD system that has testing of code, creation of packages from the code, creation of images from the packages, testing of images and promotion of images to golden images. is this ringing a bell? do you know where I can find this slide?17:48
bodepdfungi: yeah, two ways: 1. create different types 2. make version a param that accepts an array17:48
mikazaro: any chance we could see a new jenkins-job-builder release? :)17:48
mikamgagne: thx!17:48
mordredjesusaurus: I think that's in the tripleo deck?17:48
bodepdfungi: there is something called composite namevars that definitely did not work after it was created17:48
jesusaurusmordred: i think you gave that presentation in seattle like a year ago17:48
nibalizerbut yea plzcanhaz compostite namevars17:48
bodepdI doubt its been fixed b/c the issues were *complicated*17:49
jesusaurusmordred: maybe? where can i find the tripleo deck? where are the various decks kept?17:49
mordredjesusaurus: yeah - lemme send you a copy17:49
fungibodepd: aha. so instead of using type "package" we'd have pip2package and pip3package types?17:49
nibalizerfungi: you get it in other places to, such as needing to install the system package sqlite3 and the rubygem sqlite3 -> kerboom!17:49
mordredjesusaurus: the tripleo deck just gets emailed around17:49
jesusaurusah, okay17:49
mgagnebodepd: like most bugs opened for years at puppet: too complicated and bug stays open in limbo state. =)17:49
clarkbmordred: 88477 lgtm17:50
*** _nadya_ has quit IRC17:50
anteayajeblair: wonderful thank you, I'll work on the template for openstack programs now17:50
clarkbjeblair: fungi SergeyLukjanov ^ I am happy to babysit that change if someone else wants to review and approve it17:50
mgagnefungi: you could still use the package resource but implement your own provider which would make clever use of the resource name17:50
anteayakrtaylor: we haven't exactly discussed timing and scope for third party naming17:51
anteayakrtaylor: I think agreeing on a format is our first step17:51
Ajaegerclarkb: could I get some comments from you on, please?17:51
*** habib has quit IRC17:51
mgagnefungi: package { 'tox@py27': ensure => present, provider => 'pyninja' }17:51
bodepdfungi: yeah, that woudl work. mgagne 's suggestion would work as well, but you would have to embed the parsing logic in the current provider17:52
mgagnebodepd: isn't it what we do with our ini provider? =)17:52
jeblairanteaya: template for openstack programs?17:52
mgagnebodepd: what about: venv { 'py27': ensure => present } and package { 'tox@py27': ensure => present, provider => 'venv' } :D17:53
mordredjeblair: sent17:53
mordredjesusaurus: sent17:53
anteayajeblair: yes a template for openstack programs17:54
jeblairanteaya: i don't understand what you're saying17:54
anteayaline 2017:54
*** otherwiseguy has quit IRC17:54
fungimgagne: well, we're talking about system-wide installs rather than virtualenvs (parallel installations within the same system context)17:54
jeblairanteaya: i understand now, thanks!17:54
anteayajeblair: np sorry it was opaque17:55
mgagnefungi: alright, lets not go down the path of venv hell in puppet then :P17:55
anteayaAjaeger: how do I add a link to
nibalizermgagne: as an aside i find this module to be really good at dealing with python and virtualenv
Ajaegeranteaya: [[Category:ThirdPartySystems]]17:56
anteayaAjaeger: when I click the edit button I can only edit: This page links to all pages tagged as relevant for 3rd party systems.17:56
fungimgagne: in fact, we're mainly using it to get the things we need to be able to create virtualenvs (that is, newer versions of pip, virtualenv, tox, et et cetera than are distributed by the operating system)17:56
anteayaAjaeger: I don't understand the context for how to use that bit of code17:56
clarkbmordred: I note that the thread on 1.1 problems seems to have gone quiet. We do still see lots of error'd nodes though17:56
mgagnenibalizer: now that someone did it already, I lost interest :D17:56
anteayaAjaeger: I need to create a new template17:57
anteayaAjaeger: I don't know how to create a new template17:57
Ajaegeranteaya: sorry, was confused....17:57
Ajaegeranteaya: What's the name of the new template?17:57
AjaegerProgramsThirdParty ?17:57
jesusaurusmordred: awesome thanks17:57
Ajaegeranteaya: go to https:/ and create the page17:58
AjaegerAnd fill it with content ;)17:58
anteayaAjaeger: ahh, okay that could work17:58
AjaegerCopy from the other template what you like ;)17:58
anteayajust navigating to the url creates the page?17:58
fungiclarkb: 88477 lgtm (slightly sketchy about adding things to profile.d but i suppose since this is just envvar manipulation that's okay). you can approve when you're ready to watch it17:59
clarkbfungi: weren't those additions always there?17:59
fungiclarkb: oh, indeed! that was a rename18:00
*** talluri has quit IRC18:00
Ajaegeranteaya: Navigate to it and then there's a button "Create the page"18:00
msabramozaro et al: I would lean towards mocking to test python-jenkins with JJB18:00
lifelessjesusaurus: we have one like that without the packages bit :)18:00
lifelessjesusaurus: in most of the tripleo presentations18:01
msabramomuch faster and less error-prone than spinning up a bunch of Jenkins stuff18:01
msabramoand I already mock Jenkins in the python-jenkins tests, which have 100% coverage18:01
Ajaegeranteaya: if you go to the page, a box opens, if you save that the page is created...18:01
AjaegerNo need t osay Create (just checked)18:01
msabramoI'm not sure I'm a fan of having python-jenkins run tests with an upstream component like JJB - that seems backwards to me18:02
clarkbmsabramo: I think its the other way around18:03
fungimsabramo: zaro: agreed that's the best possible solution. so really just need to figure out how to exercise jjb against python-jenkins using the mocks you already have to provide proper responses18:03
clarkbmsabramo: its gating just like we do with other related components18:03
*** melwitt has joined #openstack-infra18:03
mgagnemsabramo: for important projects such as pbr and JJB, we test against consumers of the projects18:03
clarkbbut I agree. There hasn't been a need for that so far18:03
fungiyeah, the oslotest solution is more of an extreme example of that18:04
msabramoclarkb: yeah, I've seen this "gating" idea before. It seems weird to me, but I am willing to entertain that it may make total sense for the way OpenStack development works18:04
clarkbmordred: fungi: approved the trusty slave fixes change18:05
msabramoclarkb: but it is weird when I submit simple changes for something like python-jenkins and then it fails because of some weird problem in heat or neutron, which are things I don't know anything about18:05
fungimsabramo: well, the idea being that you can know when you propose a change to python-jenkins whether it will cause jenkins-job-builder to break. you might simply run that as an advisory job rather than actually having it prevent you from being able to approve and merge that change18:05
mordredclarkb: ossum18:06
mgagnefungi, msabramo: just like how we compare the XML configs of our CI jobs before and after a change in JJB.18:06
msabramofungi: yeah, I see the value. It seems like maybe something that could be solved by inverting things. LIke when there is a new python-jenkins, it triggers a new build of JJB and that JJB test is where the problems are caught; not in python-jenkins18:06
fungimsabramo: but at least then you'll know that before you release a new python-jenkins version, jenkins-job-builder will likely need to receive a new version which will either conflict with that or support it18:06
msabramobut this is a very well-established paradigm in OpenStack so I don't think I should argue it18:07
msabramoI'm sure there is a lot of history behind it and you guys did this for good reasons18:07
SergeyLukjanovSlickNik, that's a point too18:07
clarkbmsabramo: illustrates it pretty well18:08
*** dstanek_zzz is now known as dstanek18:08
fungimsabramo: well, it's more than "these are tests you can run if there are particular downstream consumers of your library which you want to make sure can continue to use it"18:08
SergeyLukjanovclarkb, which CR I should review? :)18:08
*** lcheng_ has quit IRC18:08
fungimsabramo: we just happen to have a mechanism to run that for you on every proposec commit18:08
clarkbSergeyLukjanov: 88477 is the one I just approved18:08
openstackgerritArun Kant proposed a change to openstack/requirements: Adding ldappool module dependency as needed by keystone bug #1320997.
uvirtbotLaunchpad bug 1320997 in keystone "Common Ldap handler connection pooling" [Medium,In progress]
clarkbSergeyLukjanov: feel free to look at it though maybe you will see something18:08
SergeyLukjanovclarkb, okay:)18:09
mgagnemsabramo: JJB was written by openstack-infra to manage their own jobs. It wasn't widely used outside openstack before and made sense to do such tests so they don't break their own stuff.18:09
*** lcheng_ has joined #openstack-infra18:09
fungimsabramo: it's entirely optional. but it does help high-profile projects using your library be able to provide you real-time feedback on whether what you're doing will work for them18:09
*** pafuent has left #openstack-infra18:09
msabramomgagne, fungi: yeah I can see that18:10
openstackgerritA change was merged to openstack-infra/config: Add a doc8 stackforge project
msabramoperhaps with python-jenkins, it feel stranger, because it is a much more general module and it was recently imported to stackforge18:10
fungimsabramo: we already have plenty of libraries we use in various places which just instantly break us when they release new versions, so we've gotten pretty good at putting out those fires, but anywhere we can avoid more fire drills is nice18:11
SergeyLukjanovclarkb, it's lgtm, I've already reviewed it earlier today, but decided to postpone sending a mark to check ruby package names ;)18:11
openstackgerritVerónica Musso proposed a change to openstack/requirements: Add django-angular as requirement
bodepdmgagne: I would say those things are just not supported :)18:12
pleia2the jbossas4 package on Ubuntu is useless to us (and known "cannot be used as an application server yet" according to the README.Debian)18:13
openstackgerritA change was merged to openstack-infra/config: Changes to get nodepool working with trusty
clarkbpleia2: wow18:13
msabramofungi: yeah, alright I'll buy in. What was the original question, I forgot?18:13
pleia2at least one person has been working on it in a PPA, but that's kind of scary
mordredit exists in ubuntu, but doesn't actually work?18:13
pleia2elizabeth@zanata:/usr/share/doc/jbossas4$ cat README.Debian18:14
pleia2jbossas4 is currently in a very alpha stage of packaging. I can be used18:14
pleia2to build other libraries depending on JBoss like libhibernate3-java but18:14
pleia2it is not complete and cannot be used as an application server yet.18:14
fungimsabramo: mostly whether there's a way we can devise an integration test to exercise python-jenkins from jenkins-job-builder calls18:14
pleia2sad story18:14
mordredpleia2: that just makes me sadder than I can possibly describe18:14
fungimsabramo: leveraging your existing mocks so that we can emulate submitting job configs and deleting them again18:14
fungimsabramo: so that we test as many of the methods in the python-jenkins api as jenkins-job-builder currently relies on, ideally18:15
morganfainbergpleia2, wow18:15
pleia2it's pretty much just synced from debian, where not a whole lot of work is happening (last changelog entry in 2012)18:16
*** _nadya_ has joined #openstack-infra18:16
*** jamielennox is now known as jamielennox|away18:16
msabramofungi: I think basically we need more test coverage for JJB18:17
clarkbkrtaylor: was it you who mentioned that your jenkins server is capable of copying console logs to swift?18:17
clarkbkrtaylor: I am curious to know what your solution to that is18:17
msabramofungi: it just tests the YAML to XML conversion currently. We would test it more thoroughly, by mocking similarly to how we do that in python-jenkins18:17
*** dstanek is now known as dstanek_zzz18:17
msabramofungi: I have always thought that was something I might tackle at some point18:18
fungimsabramo: i don't doubt that. the original question was when someone (was it you?) said "will our next python-jenkins release break jenkins-job-builder" so i proposed a way to always know whether that's going to be the case. if we mock the python-jenkins api calls in jjb, then that won't really tell us whether python-jenkins itself works (integration testing)18:18
msabramofungi, zaro: Although I think probably be want to land first so that we know what our baseline coverage is and then can see it go up as we add tests18:18
bodepdnibalizer: mordred I reviewed and commented on the patch, I definitely did not process every single line of ruby code, but at least went through the code in general.18:19
fungimsabramo: yep. i was just waiting for test results to come back on that one18:19
*** esker has joined #openstack-infra18:19
msabramofungi: looks like it was verified last night by Jenkins at 7:10pm18:20
fungioh, that adds the tox coverage target. i was thinking about the one adding the job to the check pipeline18:20
fungireviewing that too18:20
msabramofungi: ah ok, different change sets18:20
msabramofungi = Jeremy Stanley?18:20
msabramoah ok, I forget all the names18:20
funginames are unimportant18:21
bodepdthe most important part for me is that I need to wait for tons of patches before I can split the bits out that I want18:21
*** praneshp has joined #openstack-infra18:21
msabramofungi: so yeah, I guess what I was thinking before but didn't articulate well is that we wouldn't mock out the entire python-jenkins API (because then yes, we're not testing python-jenkins at all). We would mock out the underlying HTTP stuff18:21
bodepdmordred: actually, when is the best time to try to get patches in to split things out18:21
*** gabriel-bezerra has left #openstack-infra18:21
fungimsabramo: that could work18:22
bodepdmordred: more specifically, once your pip provider patches get in, is that a good time to split pip out?18:22
msabramofungi: I haven't thought a ton about it but I might go for using a tool like CaptureMock or vcrpy to do said HTTP mocking18:22
bodepdmordred: could it happen before? Are there more design things that need to be discussed before splitting out modules? /cc jesusaurus18:22
msabramofungi: Those tools automatically record the requests and responses from a real server into a file and then can replay from that file when it exists and is deemed to be up to date18:22
fungimsabramo: yep, that's at least a good way to get a starting point, and then you can edit it from there as you need to extend it18:22
bodepdmordred jesusaurus for example: do we still want to install all external deps from the forge, or could we pin revisions and install deps from source18:23
nibalizerbodepd: awesome thanks18:23
mordredbodepd: I think we can split before that patch - we can re-parent that patch at the new repo if needed18:23
fungimsabramo: though openstack itself is full of rest apis, so it's entirely likely we already have a ton of mindshare behind ways to test and mock http interactions18:23
nibalizeri'm planning on adding a ruby-lint test of some kind to pass over our .rb18:23
nibalizerdo you have a prefered tool for linting? the two i found are both ruby2 only18:23
*** nati_ueno has quit IRC18:24
fungimsabramo: in which case we may have bits we can just hoist straight in to do that18:24
msabramofungi: yeah, if there is some other recommended approach I am all ears. Personally I have used vcrpy a little bit in my own projects and I have played with CaptureMock but not gotten quite as far18:24
mordredbodepd: and _I_ don't think we need to do forge if we can do a from-source method that works well - but things like librarian have been a non-starter for us so far18:24
clarkbnotmyname: are you around? any chance we can talk about the swift review sponsors thing again?18:24
clarkbnotmyname: want to make sure that we don't forget about you on that and figure out something that will work18:24
mordredclarkb, fungi, jeblair: ^^ thoughts on non-forge split-out modules?18:24
clarkbmordred: r10k18:24
msabramomgagne: Want to review Once we get coverage for JJB, then it will be more rewarding to add tests to JJB18:24
bodepdmordred: have you seen librarian-puppet-simple :)18:24
clarkblibrarian is poo18:25
notmynameclarkb: I'm here.18:25
bodepd(shameless plug :) )18:25
clarkbbut finch swaers r10k is good18:25
bodepdyeah, that is the real answer. I find it to be a little opinionated and complicated.18:25
jesusaurusclarkb: what dont you like about librarian? (ive never used it)18:25
mordredbodepd: I have not - but I think each time we try to use things that aren't our shell script, things go crazy18:25
mordredbodepd: however, our module install is run from a shell script - so conceivably one could do anythign in there18:25
clarkbnotmyname: thinking back to that paste that jeblair made. Is that something that won't work because individuals will need to sign up themselves?18:25
bodepdjesusaurus: he's referring to librarian-puppet18:25
fungimordred: seems fine, to me the bigger issue is separating the modules out so they can have their own life cycle separate from the config project. where they get published is more of an implementation detail18:26
bodepdjesusaurus: it's dep resolution code sucks and pretty much makes it impossible to use18:26
clarkbalso it depends on puppet318:26
clarkbwhich means you lose18:26
bodepdclarkb: are you guys not using Puppet 3?18:26
clarkbbodepd: not yet18:26
mordredwe're slowly starting to get there18:26
mordredit's a big task18:26
clarkbslowly moving towards that but there isn't much puppet 3 does that is compelling neough to move other than new distro has it18:26
clarkbsee python18:27
jesusaurusspeaking of dep resolution, how will we deal with nibalizer's nondep puppetboard stuff with something other than our shell script?18:27
notmynameclarkb: I don't know. I think people signing up could be very good, and perhaps the existing patch would work well. I was hoping for something slightly visually different, but the self-sign-up isn't bad18:27
bodepdintegrated hiera is killer. Also, the plugin system refactor stuff is pretty awesome18:27
fungiwell, and new puppet plays poorly with old puppet, which makes piecemeal upgrades with a puppet master a bit of a pain18:27
clarkbnotmyname: ok. as far as visuals go (or UX) what were you thinking?18:27
notmynameclarkb: suffice to say, it's just an idea right now, and one I'll be working on over the next couple of weeks. so I don't know what will work best :-)18:27
mordredjesusaurus: I think we should just keep using our script :)18:27
bodepdjesusaurus: I don;t have enough context. What is the issue?18:28
notmynameclarkb: I was hoping for another column on the main review page eg next to the author that would be the name of the core sponsor18:28
bodepdmordred: I'm ok with that. Perhaps have an install from github variable?18:28
notmynameclarkb: and also therefore able to sort and search on that18:28
mordredbodepd: yeah. I think that would be fine tbh18:28
clarkbnotmyname: I see. That is technically hard beacuse gerrit18:28
clarkbnotmyname: you can search using the proposed method though18:28
nibalizer is my patch to add a p3 master18:28
notmynameclarkb: right18:28
*** crc32 has joined #openstack-infra18:28
*** rbowen has quit IRC18:29
clarkbfungi: why doesn't docker work on precise nodes?18:29
jesusaurusbodepd: i dont recall why it was needed ( nibalizer can hopefully fill in that bit ) but we install nibalizer-puppetboard without dependency resolution because of varying versions of vcsrepo in different places18:29
nibalizeryea so i wrote a module that uses vcsrepo18:29
jesusaurusnibalizer: what was the deal there?18:29
notmynameclarkb: also, FYI the core sponsor thing will be getting secondary priority (as will everything else) since the storage polices patches have just been proposed and we're all focusing on those until they land18:29
nibalizerbut it uses puppetlabs-vcserpo and openstack-infra/config uses openstack-vcsrepo18:29
nibalizerwhich is just a fork of vcsrepo18:30
clarkbnotmyname: roger good to know. I will probably let this sit and ferment a bit then18:30
nibalizerbut the 'puppet module install <modname>' does try to do recursive dependency resolution18:30
fungiclarkb: i believe docker has some specific kernel version and cgroups support requirements not met on our current precise nodes?18:30
nibalizerso we have one pass where we install modules with dependencies18:30
bodepdI find that I generally don't want to use the built in dep resolution b/c I wind up changing things18:30
*** nati_ueno has joined #openstack-infra18:30
nibalizerand one pass where we install modules without dependencies18:30
fungiclarkb: could be that one of the newer kernels available for precise would solve it though18:30
bodepdlibrarian-puppet-simple is librarian-puppet without dependencies18:30
mordredheh. nice18:31
clarkbfungi: ah ok, I was pretty sure docker works fine on precise but it may require those newer kernels18:31
nibalizerits almost like we need to go by to the 'provides' or 'features' stuff thats already built into the type/provider system18:31
clarkbbodepd: without puppet3 too?18:31
mordrednibalizer: your patch need fixing, btw18:31
bodepdclarkb: it doesn't support forge installs (source only) so Puppet is not a dep18:31
nibalizerfungi: yea the party line on that is precise with a backported raring kernel18:32
nibalizer(for docker)18:32
nibalizerat least until trusty came out, not sure what the party line is now18:32
*** e0ne has joined #openstack-infra18:32
funginibalizer: right, when i talked to ewindisch about it (can't remember now if it was at the summit or before) consensus seemed to be that we'd be better off targeting trusty slaves for docker jobs18:33
nibalizermordred: you mean it has rebase changes? that doesn't suprise me, its like weeks and weeks old18:33
mordrednibalizer: well, also it failed jenkins18:33
nibalizermordred: hrm18:33
*** marcoemorais has quit IRC18:33
mordredoh - and merge conflicts18:33
nibalizerthe failure on the yaml looks like dns to me18:33
clarkbfungi: ya that is partially what prompted my question because there was a lot of interest in trusty for reasons I didn't quite grasp18:33
mordrednibalizer: actually, there is something I had in a different patch that needs to go in here...18:34
nibalizer2014-05-28 08:04:54.027 | fatal: unable to connect to
*** marcoemorais has joined #openstack-infra18:34
nibalizer2014-05-28 08:04:54.027 | Temporary failure in name resolution18:34
mordrednibalizer: mind if I edit it real quick?18:34
nibalizermordred: yea we need to sync up on the trusty stuff18:34
nibalizermordred: please do!18:34
nibalizerbut of course since it needs to rebase it might geet ugly18:34
anteayajeblair fungi clarkb mordred SergeyLukjanov pleia2 zaro Ajaeger here is the first draft of the template for the programs template:
anteayalet me know what I am missing and what I need to clarify18:35
clarkbAjaeger: looking at the log levels tx extraction now18:36
* Ajaeger hands on the mastery degree of wiki templates to anteaya !18:36
*** dstanek_zzz is now known as dstanek18:37
anteayaha ha ha18:37
Ajaegerclarkb: great, thanks!18:37
clarkbAjaeger: I am a bit worried that this appears to require people writing code to use different functions for each level18:37
anteayathanks Ajaeger :D18:37
clarkbAjaeger: do we expect them to do the right thing there? I suppose this is just going to be educational18:37
Ajaegerclarkb, ask dhellmann - he did that kind of work already18:37
clarkbAjaeger: ok18:37
krotschecknibalizer: You’re here! Im going to start working on puppet-storyboard, mind if I ping you for a review later?18:37
Ajaegerclarkb: yeah educational18:37
nibalizerkrotscheck: yea!18:38
Ajaegerclarkb: here's what's done already
krotschecknibalizer: Thanks!18:38
Ajaegeralready in icehouse18:38
hogepodgeCan zuul target different gates for different branches?18:38
Ajaegerclarkb: The only missing part is actual setting up the jobs18:38
krtayloranteaya, looks good18:39
Ajaegerclarkb: you discussed this in December with dhellmann ;)18:39
hogepodgei.e. have stackforge/project-name target master and stackforge/project-name/branchname target a branch?18:39
Ajaegerand proposed some of this18:39
Ajaegerclarkb: see also
clarkbAjaeger: ya I remember discussing it I just don't think we got into specifics18:40
fungianteaya: is the "program mission statement" supposed to be the mission statement for the program cut and pasted, or should it just be a link to where the mission statement for the program is actually published already? (or transcluded from another wiki page?) if it's duplicated, it may fall out of date over time and diverge18:40
Ajaegerclarkb: might be ;)18:41
Ajaegeranteaya: Quantity - do you want an exact number or something like "> 50" ?18:41
clarkbAjaeger: in any case +2 we can see how it turns out18:42
krtaylorAjaeger, my question too, and entities are?18:42
krtayloranteaya, ^^^18:42
krtaylorclarkb, yes, we have a script that takes care of swift log upload for PowerKVM18:42
fungihogepodge: i'm not sure what you mean by "different gates" but you can definitely give each job name a pattern of branches where it should be run or not run18:42
Ajaegerclarkb: thanks! We can change the msgmerge as followup if you like...18:43
anteayakrtaylor: thanks18:43
anteayafungi: good point18:43
Ajaegerclarkb: there're some more cleanup patches in the queue for those scripts ;)18:43
mgagnehogepodge: pipelines look to apply to all branches but as fungi pointed out, you can restrict branches against which the jobs will be run18:43
anteayafungi: yeah, so a link would probably be best, I just hope they click it18:43
hogepodgefungi: We’re talking about how to drop support for Puppet 2.7 in the coming year, so at some point we want to drop that gate from master, but leave it in place for the stable branches.18:44
fungihogepodge: for example shows the "gate-swift-tox-func" job running on any branch which is not stable/havana or stable/icehouse18:44
anteayaAjaeger: a guesstimate is fine, more than 5 would work for me18:44
hogepodgeOk, I see. Thanks.18:44
anteayathe point is that every third party person thinks they are the only one18:44
Ajaegeranteaya: then say it ;)18:44
anteayasay what?18:44
fungihogepodge: or runs "openstack-misc-guides-ja" only on branches named master18:44
bodepdjesusaurus: yeah, I found the code that related to it. I'm going to add a SOURCE_MODULES variable18:45
fungihogepodge: tons of examples in there18:45
bodepdnibalizer: if I add a variable for installing from source and without deps, can I remove the NODEP variable and just have it install from source (based on a revision?)18:45
anteayaI am using entites to mean collection of folks representing a third party interest18:45
*** arnaud has joined #openstack-infra18:45
anteayaI can't say account, because they might not have an account yet18:45
anteayaI'm open to a better word than entities18:46
Ajaegeranteaya: change "Quantity: How many Third Party entities interact with your Program?" to "Quantity: How many Third Party entities interact with your Program (order of magnitude number is fine)?18:46
jesusaurusbodepd: do you plan on installing from a sha, or from a git tag, or what?18:46
nibalizerbodepd: whats the context? the ?18:46
Ajaegeranteaya: or something like that18:46
jesusaurusnibalizer: yeah, install_modules.sh18:46
anteayaAjaeger: go for it18:46
anteayaalso while you are at it, can you show me how to get multi line in a value?18:46
anteayafor instance the meeting links aren't all going to have space on the same line18:47
fungianteaya: we've used the term "organizations" to this purpose in the past as well18:47
anteayathey each should have a line of their own18:47
anteayaorganizations, hmmmm18:47
nibalizerbodepd: i think thats fine18:47
bodepdjesusaurus: I was going to support all. They just work with git checkout18:47
anteayawell that gives them more credit than I have observed, but okay organizations might be a better fit18:48
nibalizeri'm assuming you want to install a module from source and don't want three loops so just change how puppetboard get installs?18:48
fungianteaya: it's an optimistic term ;)18:48
bodepdnibalizer: yep, that is exactly the context I did not provide :)18:48
krtayloranteaya, but I am still wondering what interacting means in this context of quantity - what is this trying to document?18:48
anteayasomething to hope for18:48
nibalizerbodepd: do et, post review, i will +118:48
nibalizerbodepd: also drinks soon is a moral imperative18:48
anteayawe expect you to test your code18:48
anteayaand when we interact with you we will tell you that18:49
Ajaegerkrtaylor: It's a different story if you think you're the first 3rd party driver to work with the project - or if you one of 60 - and those others work just fine ;)18:49
anteayaover and over an over18:49
anteayaso heads up18:49
*** _nadya_ has quit IRC18:49
*** _nadya_ has joined #openstack-infra18:49
anteayafor some reason these organizations feel they can do an end run around the requirements18:50
anteayaif they can just find the right person18:50
AjaegerI'm leaving now. If somebody wants to give a review today, I would be gratefull.18:50
openstackgerritA change was merged to openstack-infra/jenkins-job-builder: Add tox "coverage" target
*** Ajaeger has quit IRC18:50
anteayaso every program needs to tell them, we expect you to test your code18:50
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add node def for puppet3 master
mordrednibalizer: ^^18:51
krtayloranteaya, are you addressing my question about quantity?18:51
*** _nadya_ has quit IRC18:51
anteayayes interacting means in this context of quantity18:51
anteayathat was the question I was addressing18:52
anteayahave I missed the mark?18:52
*** hashar has joined #openstack-infra18:52
clarkbfungi: change 95935, its actually a bit harder to setup in log_processor because the log processors using logging.basicConfig() which does not appear to have a way to change the handler to WatchedFileHandler18:52
krtayloranteaya, hehheh, I still don't understand what interacting means18:52
anteayabecause you are doing it18:52
fungiclarkb: oh, ugh. +2 then18:52
anteayait is like explaining swimming to a fish18:53
anteayathink of the folks in the cinder meeting18:53
clarkbfungi: I think we should make the change, its just not a simple config change though. Will require MOARCODE18:53
anteayathe guy who when told that nightly builds was insufficient, changed his mind and talked about weekly builds18:53
anteayaremember that guy?18:54
anteayathat guy is going to read this page that cinder fills out18:54
krtayloranteaya, so you mean number of drivers in your project that are doing 3rd party testing?18:54
Alex_Gaynorclarkb: FYI, I've pinged our racskpace internal folks about the service catalog issue we discussed last week18:54
clarkbAlex_Gaynor: great!18:54
anteayanumber of accounts doing thrid party testing18:55
anteayanumber of organizations that are18:55
anteayaor should be18:55
anteayait doesn't have to be exact18:55
anteayait just has to empireically state18:55
anteayayou are not the only person dealing with this project18:55
fungiAlex_Gaynor: any indication it might get worked on?18:55
krtayloranteaya, so this template is meant to be used at a project level - nova, cinder, neutron, etc18:55
anteayathey still won't get the message18:55
anteayathis template is18:55
anteayafor organizations to consume18:56
* krtaylor has a lightbulb moment18:56
anteayathey don't read anything else anyway18:56
anteayaI can only hope they will read this18:56
krtaylorhehheh, so pessimistic18:56
krtaylorsad, but true18:57
anteayaspent the last release in -neutron18:57
anteayaI have no illusions18:57
nibalizermordred: whats #NB mean when its in a file with a little note?18:57
*** shardy is now known as shardy_afk18:57
mgagnenibalizer: Nota bene? :P18:59
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Remove usage of
nibalizermgagne: thanks19:00
*** jp_at_hp has quit IRC19:00
openstackgerritK Jonathan Harker proposed a change to openstack-infra/gitdm: Add Jonathan Harker
*** atiwari has quit IRC19:02
*** esker has quit IRC19:03
openstackgerritA change was merged to openstack-infra/config: Remove duplication in translation scripts
*** Sukhdev has joined #openstack-infra19:04
openstackgerritA change was merged to openstack-infra/config: cleaning up index.rst file
*** dims has joined #openstack-infra19:07
openstackgerritA change was merged to openstack-infra/config: Enabled pylint job for manila project
* mordred is baffled about people continuing to use pylint19:08
fungimordred: apparently there are still things it's able to catch which pyflakes does not due to the kind of analysis it performs19:09
*** e0ne has quit IRC19:10
openstackgerritA change was merged to openstack-infra/config: add python coverage job to check pipeline for JJB.
*** e0ne has joined #openstack-infra19:10
openstackgerritA change was merged to openstack/requirements: Remove openstack/database-api from list of projects
*** e0ne has quit IRC19:14
*** e0ne_ has joined #openstack-infra19:14
*** andreykurilin has joined #openstack-infra19:15
*** _nadya_ has joined #openstack-infra19:15
mordredclarkb, fungi: dhellmann fixed the upload wheels patch:
mordredSergeyLukjanov and I both +2'd it - but given the nature of the patch, a couple of more eyes might be nice19:17
clarkbmordred: ok19:17
SergeyLukjanovmordred, ack19:17
clarkbnote I will not approve as many things nowish because dentist at 219:18
clarkbbut I will review19:18
openstackgerritA change was merged to openstack-infra/config: Update get-pip URL to the new location
mordredI think this is a "make sure someone can babysit"19:18
*** chuckC has quit IRC19:19
*** e0ne_ has quit IRC19:21
*** e0ne has joined #openstack-infra19:22
clarkbmordred: wheels are zip files too?19:24
clarkbmordred: `file -b foo.whl` will say zip in there?19:24
dhellmannmordred: That patch has to +2; I think SergeyLukjanov didn't realize I reworked the latest patch it so your +2 was ok?19:24
*** arnaud has quit IRC19:24
mordredclarkb: yup19:25
dhellmannsigh: s/to/2/19:25
mordredmordred@camelot:~/src/stackforge/gertty$ file -b /home/mordred/src/openstack-dev/pbr/dist/pbr-
mordredZip archive data, at least v2.0 to extract19:25
SergeyLukjanovdhellmann /me trying to realize, it's re ?19:26
clarkbmordred: well it does file -b | grep zip19:27
dhellmannSergeyLukjanov: yes, mordred's name is still on the patch but the latest patchset was my rebase -- I don't know how you treat that here, so maybe your +2 without the A was normal :-)19:27
clarkbwhich may not match that output19:27
dhellmannSergeyLukjanov: and clarkb seems to have found a potential issue19:27
openstackgerritJeremy Stanley proposed a change to openstack-infra/jeepyb: Wait for ACL creation in manage-projects
SergeyLukjanovdhellmann, I've not approved it to see more reviews for it19:27
SergeyLukjanovdhellmann :)19:28
*** pelix has quit IRC19:28
dhellmannSergeyLukjanov: yep, that's ok19:28
*** W00die has quit IRC19:29
*** dangers_away is now known as dangers19:30
clarkbdhellmann: doing grep -i there is fine19:30
clarkbdhellmann: mordred: why are we adding pypi-jobs to glance, horizon, et al?19:31
clarkbthat seems like a noop beacuse they don't end up in the layout file19:31
mordredclarkb: because - part of the "start uploading wheels" discussion also was about starting to upload releases of the server projects too19:31
clarkbmordred: right but this change doesn't do that19:31
mordredclarkb: hrm. good point19:32
openstackgerritJeremy Stanley proposed a change to openstack-infra/jeepyb: Wait for ACL creation in manage-projects
mordredprobably should split that into a separate patch19:32
fungimordred: also, what's the plan for server projects where the name is actually something else entirely on pypi?19:32
*** nati_ueno has quit IRC19:32
dhellmanndo any of the validation tools in this repository tell me that I have jobs defined that do not appear in the layout?19:32
clarkbdhellmann: no, it only checks the inverse thing that jobs in layout are defined in jenkins19:33
fungimordred: for example,
clarkbthe diff could probably be made strict when we check19:33
dhellmannclarkb: ok, I just thought I'd check (see some of the changes in the patches after this one)19:33
dhellmannso I should remove the pypi jobs from the server projects?19:34
mordreddhellmann: yeah.19:34
mordredfungi: oh for the love of pete19:34
clarkbthey won't hurt anything but it is confusing19:34
clarkbmordred: hahaha19:34
fungimordred: noticed this the other day with as well19:35
clarkbuse an openstack prefix for everything?19:35
clarkbopenstack-keystone, openstack-tempest19:35
mordredyeah - but we'd need to change that in the setup.cfg too19:35
mordredthis may be more trouble than it's worth19:35
*** wenlock_ has joined #openstack-infra19:38
*** W00die has joined #openstack-infra19:40
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Create and upload wheels
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Add zuul template for rtfd jobs
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Make it possible to run zuul layout test locally
dhellmannoops, that's wrong, hang on19:43
clarkbdhellmann: looks like you removed pypi-jobs from things that have them? in any case hanging on19:44
openstackgerritA change was merged to openstack-infra/config: Remove usage of
dhellmannclarkb: I removed the pypi jobs from anything that looked like a server project, as I thought had been requested?19:45
clarkbdhellmann: no it had been requested that you remove it from the servers that you added them too19:45
clarkbdhellmann: anything with existing jobs should keep them. We just don't want to be adding pypi-jobs to servers in this change19:45
dhellmannclarkb: ok, there are a bunch of jobs for publishing things that aren't on pypi19:46
clarkbdhellmann: but if those things don't use pypi-jobs then I guess it doesn't hurt to clean it up19:47
mordredhacking needs to stay at least19:48
dhellmannclarkb: ok, I'm going to restore them and just remove the ones that were being added by this patch originally19:48
mordreddhellmann: ++19:48
dhellmannmordred: yeah, that was my original oops19:48
mordredthen I think you're right - we shoudl go back through and clean house19:48
*** wenlock has quit IRC19:49
*** otherwiseguy has joined #openstack-infra19:49
*** wenlock_ has quit IRC19:50
*** wenlock has joined #openstack-infra19:52
*** markmc has quit IRC19:52
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Create and upload wheels
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Add zuul template for rtfd jobs
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Make it possible to run zuul layout test locally
clarkbok I am +2 on 56760 now, but can't babysit until later today (maybe tomorrow is a good time to babysit?)19:57
msabramofungi: thanks for reviewing at merging — that lets us see our coverage for JJB19:58
*** shardy_afk is now known as shardy19:58
clarkbjeblair: does diffing the infra-specs repo initial commit in gertty crash gertty for you?19:59
openstackgerritAnita Kuno proposed a change to openstack-infra/config: The process for making a Third-Party CI voting needs docs
*** chuckC has joined #openstack-infra20:03
*** Ryan_Lane1 has joined #openstack-infra20:03
clarkbjeblair: also, anything new that would help debug the gear stuff? maybe debugging on the nodepool side?20:04
*** Ryan_Lane has quit IRC20:05
*** nati_ueno has joined #openstack-infra20:06
*** jaypipes has quit IRC20:08
*** rlandy has quit IRC20:09
*** arnaud has joined #openstack-infra20:10
dhellmannclarkb, mordred : here's a list of projects that are in projects.yaml but not on pypi using the same name:
*** nati_ueno has quit IRC20:11
clarkbdhellmann: thanks. Feel free to push a change to clean that up or I can get around to it post dentist20:11
dhellmannclarkb, mordred : some of those look like things we want to be publishing20:11
jeblairclarkb: not sure about gertty; i know there's at least a bug regarding changes with comments on the commit message; i think there's one other i haven't had time to track down20:12
*** radez is now known as radez_g0n320:12
jeblairclarkb: i'm pretty confident i understand what's going on with zuul/gear/nodepool.  did you see the last?20:12
openstackgerritRyan Petrello proposed a change to openstack-infra/config: Gate a few new projects that are using pecan.
clarkbjeblair: I don't think I saw the last unless it was nodepool being unable to read fast enough20:12
clarkbI skimmed sb looking but didn't see muhc20:13
jeblairclarkb: that's it.  :)20:13
jogothis file search support in gerrit is awesome20:13
jeblairclarkb: that blocks geard20:13
jeblairclarkb: which blocks all its clients20:13
clarkbjeblair: should each geard Connection maintain a write buffer that is pushed into the tcp stack asynchronously?20:13
jeblairclarkb: i've been heads down working on a solution20:13
clarkbin that case I will leave you to it20:13
*** _nadya_ has quit IRC20:14
openstackgerritA change was merged to openstack-infra/git-review: Fix typo in manpage s/gireview/gitreview/
openstackgerritMark Sturdevant proposed a change to openstack/requirements: Remove hplefthandclient from global-requirements
*** mriedem has quit IRC20:17
clarkbdhellmann: comment for you on your gertty change20:18
*** mriedem has joined #openstack-infra20:18
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Ignore .egg directories
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Remove unused pypi jobs from projects
dhellmannclarkb: ^^ cleanup change for you :-)20:19
mordreddhellmann: blast. when we did the renames the other day, we should have put oslo.version in the attic20:20
Alex_Gaynordstufft: ping20:21
dhellmannclarkb: replied to your gertty comment20:22
dhellmannmordred: I keep meaning to ask you about oslo.version20:22
dhellmannwe're going to need a place for versionutils, would it make sense to put it there?20:22
*** r-daneel has joined #openstack-infra20:22
dhellmannmordred: if nothing in oslo.version is useful right now, I could erase the contents and reuse the lib name20:23
clarkbdhellmann: ooohhhh so all of my projects are unreviewed and I didn't realize it was three states there20:24
clarkbdhellmann: what do you think about default for unreviewed, dark gray for subscribed, and can we do a different background + default for unsubscribed?20:25
clarkbor something like that20:25
dhellmannclarkb: I'm not 100% sure. I really went in and poked at things until there were no blank places on my terminal, so some of these states might not even be used.20:25
clarkbdhellmann: maybe I remove my -1 and let jeblair chime in20:25
clarkbI will do that20:25
*** maxbit has quit IRC20:25
dhellmannclarkb: I'm OK with a different background, but I have a pretty hard time with reading light on dark20:26
dhellmannclarkb: in a perfect world we would just read this from a configuration file20:26
clarkbdhellmann: ya, it is configurable now iirc20:26
dstufftAlex_Gaynor: pong20:26
dhellmannclarkb: oh, I wasn't able to figure that out20:26
clarkbbut I think by default it should support light and dark terminals20:26
clarkbdhellmann: it just merged20:26
Alex_Gaynordstufft: So, do you remember how when Django upladed pre-release wheels on pypi it broke stuff?20:27
dhellmannclarkb: ah! if that's the case, we can probably just abandon my change, unless I'm actually going in something like the right direction20:27
dstufftAlex_Gaynor: I remember that it broke stuff, I don't recall how20:27
Alex_Gaynordstufft: Do you remember what the symptom was?20:27
clarkbdhellmann: I think what you are doing is good. light terminals exist and we should support them by default :)20:27
Alex_Gaynordstufft: I seem to recall it was "pip install Django" broke, but I want to make sure20:27
dstufftAlex_Gaynor: no idea20:27
Alex_Gaynordstufft: Ok, thanks. I've suggested openstack put off doing that for now.20:28
dstufftAlex_Gaynor: I wasn't around when it broke, and all the info I got from it was that it broke20:28
dhellmannclarkb: ok, I'll wait to hear what jeblair says20:28
Alex_Gaynordstufft: and whatever broke definitely hasn't been fixed, I assume :-)20:28
dstufftAlex_Gaynor: not unless it was fixed inadvertently20:28
dstufftI have no idea what that was, or if Django did something wrong or what20:29
Alex_Gaynordstufft: this is packaging, who knows! :-)20:29
dstufftmore or less20:29
*** e0ne has quit IRC20:29
*** e0ne has joined #openstack-infra20:30
mordreddhellmann: what's versionutils?20:30
*** terryw has joined #openstack-infra20:30
dhellmannmordred: deprecation logging stuff:
*** salv-orlando has joined #openstack-infra20:30
*** otherwiseguy has quit IRC20:31
dhellmannmordred: I think we've been planning to call the lib oslo.versionutils because I am bad at names20:32
*** terryw is now known as otherwiseguy20:32
*** andreaf has quit IRC20:32
*** e0ne has quit IRC20:33
*** ociuhandu has quit IRC20:33
clarkbfungi: I just approved your pypi-mirror change to stop whitelisting external links20:33
fungiclarkb: yay!20:33
fungiopenstack is free of pypi-external dependencies!20:34
fungiwell, it already was, but this is clean-up20:34
*** maxbit has joined #openstack-infra20:36
lifelessmordred: ping (pbr python2.6 fail insight needed)20:36
mordredlifeless: ola20:36
*** aconrad has joined #openstack-infra20:37
* clarkb & back later20:38
lifelessmordred: help20:38
mordredlifeless: yah. so - I don' tknow why it's only failing on 2.6 - but looking at the log output - that pip command sure does have the effect of writing files to the tree and not to the temp dir20:39
*** mestery has quit IRC20:39
zaromika: did want a jjb release?20:39
*** mestery has joined #openstack-infra20:40
*** radez_g0n3 is now known as radez20:40
mordredlifeless: oh! no, that's not what that's doing -that's installing something into the virtualenv inside of which the test is running20:40
lifelessright, to avoid the easy_install20:41
mordredyeah. but it's installing into the tox venv20:41
mordredCopying pbr_testpackage.egg-info to /home/jenkins/workspace/gate-pbr-python26/.tox/py26/lib64/python2.6/site-packages/pbr_testpackage-0.0-py2.6.egg-info20:41
mordredI believe that is not what we want to do20:41
lifelessdoing that prevents easy_install downloading pbr from pypi20:42
lifelessbut thats not why we need to do it20:42
lifelesswe need to do it because the tests that run have a setup_requires of pbr20:42
lifelessand if we *don't* do it, we're not using the pbr that is being tested20:43
mordredbut this isn't installing pbr20:43
mordredthis is installing the testpackage20:43
openstackgerritA change was merged to openstack-infra/nodepool: Add cloud-user to non-root user list
lifelessoh f*20:43
*** smarcet has quit IRC20:44
lifelessmordred: it shold be installing the root package, not the test package20:45
fungiclarkb: !!! gerrit ls-members --recursive20:47
fungiclarkb: tested and it works20:47
*** lakshmiS has joined #openstack-infra20:47
*** esker has joined #openstack-infra20:48
clarkbnice. I didnt realize it had that option20:49
lifelessmordred: on 2.720:49
*** nati_ueno has joined #openstack-infra20:50
openstackgerritA change was merged to openstack/requirements: Update projects list for oslo.test renaming
mgagneI just figured out the real inner working on project ACLs with refs/meta/config in Gerrit. Could the API be used instead to sync the ACL? Is it even possible?20:54
fungimgagne: depends on what you mean by "the api" and "sync"20:55
mgagnefungi: instead of creating an acl config per project and using jeppyb to update it through git, couldn't the API be used to set those up?20:56
*** miqui has quit IRC20:56
fungimgagne: perhaps, though that seems like it would make it harder to keep the actual state of the acls collected in a single git repository20:57
mgagnefungi: true, would probably make it harder by having to convert from the Gerrit ACL config format to API calls/format20:58
fungimgagne: unless there's a bulk-replace option for a project acl, i worry they'd quickly get out of sync20:59
mgagnefungi: right, I can see the diff hell from here :-/20:59
openstackgerritlifeless proposed a change to openstack-dev/pbr: Allow examining parsing exceptions.
fungimgagne: and the current push via git into gerrit is effectively a bulk replace plus sanity checking20:59
openstackgerritlifeless proposed a change to openstack-dev/pbr: Teach pbr VersionInfo about debian versions.
openstackgerritlifeless proposed a change to openstack-dev/pbr: Teach pbr about post versioned dev versions.
openstackgerritlifeless proposed a change to openstack-dev/pbr: Use the current pbr for testpackage tests.
openstackgerritlifeless proposed a change to openstack-dev/pbr: Add a converter to version_tuples.
openstackgerritlifeless proposed a change to openstack-dev/pbr: Break out a common version object from VersionInfo
openstackgerritlifeless proposed a change to openstack-dev/pbr: Allow _run_cmd to run commands in any directory.
fungimgagne: gerrit upstream has tried to improve acl management by making it more self-service now (there is a feature to allow users to propose acl changes for review and then project owners to approve/merge them) but it doesn't really match our intended management workflow21:01
dhellmannlifeless, mordred : at some point we should talk about using oslo-specs for pbr like we're starting to do for the other oslo libs21:01
mordreddhellmann: ++21:01
mgagnefungi: ok, I'll check out jeepyb and see what I can do with it21:01
mordreddhellmann: how about now ...21:01
mordreddhellmann: I think we shoudl use oslo-specs for pbr21:01
dhellmanncool. I don't want to derail what lifeless is already doing, but maybe for future changes?21:02
mordredyah. because the barrier to entry for pbr changes should be fairly high at this point21:02
lifelessdhellmann: is there an oslo-specs now? I'm very happy to write up a spec for this21:02
lifelessgetting broad buy-in is one of the action items on this anyway21:03
dhellmannlifeless: yes, we created it just before the summit (you may literally have been on a plane at the time)21:03
dhellmannlifeless: I would be very happy to have one or more spec on this to review :-)21:03
dhellmannlike I said, I didn't want to derail work that was well under way21:04
*** Sukhdev has quit IRC21:04
signed8bitCurious as a first timer (, what is needed to get to a merged state? clarkb gave a +2 code review but I assume it needs more to continue through?21:04
mordredsigned8bit: it needs 2 +2s - lemme look21:05
lifelessso up to and including are nonconentious bugfixes basically21:05
lifelessI will write a spec for the top of the stack21:05
signed8bitAh thanks mordred21:06
*** aysyd has quit IRC21:06
*** wenlock_ has joined #openstack-infra21:06
mordredsigned8bit: and there you go ...21:06
*** marcoemorais has quit IRC21:07
openstackgerritA change was merged to openstack-infra/gerritbot: Force SSL Configuration Parameter
*** marcoemorais has joined #openstack-infra21:07
fungisigned8bit: +2 from me as well. i'm not find of the thought of us authenticating over a plaintext fallback21:07
signed8bitmordred: Thanks!21:08
mordredlifeless: +2'd all of those21:08
openstackgerritKhai Do proposed a change to openstack-infra/jenkins-job-builder: reference tests as examples
*** mfer has quit IRC21:09
openstackgerritA change was merged to openstack/requirements: update pycadf req to 0.5.1
*** ociuhandu has joined #openstack-infra21:11
signed8bitfungi: Yeah personally I'd have preferred it to not assume SSL for only one port but I didn't want to require a config change for existing deployments21:12
anteayafungi: line 85:
fungisigned8bit: pragmatic21:12
anteaya--recursive is handy21:12
fungianteaya: indeed it is21:12
openstackgerritKhai Do proposed a change to openstack-infra/jenkins-job-builder: reference tests as examples
fungianteaya: that was, in fact, one of the first groups i tested it on ;)21:13
anteayatook me a minute to get the syntax correct21:13
anteayait looks like part of the problem is that username and full name for some accounts are two entirely different things21:14
anteayanot very helpful21:14
*** gyee has quit IRC21:14
*** Sukhdev has joined #openstack-infra21:15
anteayaI still don't know which of the 3 brocade accounts are "Brocade CI" though:
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Extract translations for log messages
*** BadCub has joined #openstack-infra21:18
*** marcoemorais1 has joined #openstack-infra21:19
*** marcoemorais has quit IRC21:20
*** marcoemorais has joined #openstack-infra21:21
*** esker has quit IRC21:23
*** esker has joined #openstack-infra21:24
*** marcoemorais1 has quit IRC21:25
*** gokrokve has quit IRC21:26
*** gokrokve has joined #openstack-infra21:27
*** gokrokve has quit IRC21:30
davidlenwellSo in the last week I've been getting a lot of these fials.. when it wasn't an issue before.. this commit was just a spec.. so I don't understand how it can fail.
anteayaZero tests were run. At least one test should have been run.21:35
anteayaadd a teset21:35
*** david-lyle has quit IRC21:35
davidlenwellSo there is a test.. that does nothing21:35
*** signed8b_ has joined #openstack-infra21:36
davidlenwellnevermind ..  I guess there isn't21:36
davidlenwellbut what changed in the last week.. that didn't used to block things21:37
*** signed8bit has quit IRC21:37
anteayawhat do you want to get from running the python27 tests if there are no tests in the repo?21:37
anteayaand I don't know what changed in the last week in the python27 tests21:38
jogoso I am looking into bumping up the large-ops number, 100 appears to work but only did a few runs. what do you recomend as the next step?21:38
jogoturn large-ops-testing on as silent?21:38
jogofor a week?21:38
mgagneWould there be a way to tell Gerrit to add the -x option when cherry-picking?21:39
*** hashar has quit IRC21:39
openstackgerritMichael Krotscheck proposed a change to openstack-infra/puppet-storyboard: [WIP] Added initial puppet module for StoryBoard
krotschecknibalizer: ^ First pass, still a lot of stuff from puppet-dashboard in the README, but it’s a start.21:41
davidlenwellanteaya: there will be tests in the repo.. but we haven't bothered yet because the requirments have been in flux until now.. right now we're just commiting all the specs .. which will lead to tests.21:42
anteayadavidlenwell: I wasn't going to say anything about all the whitespace errors in that file, since I figured you were planning on correcting them21:42
anteayabut I see that patch is approved for merging21:42
anteayadon't you want to clean up the whitespace errors?21:42
davidlenwellanteaya: I don't consider white space in an rst file an error.21:43
mgagneforgot what I said, it does already include -x, got misleading info on my end ^^'21:43
davidlenwellin a python file .. absolutely and that will fail pep8..21:43
anteayawell until you add tests, why don't you stop testing against python2721:43
*** dangers is now known as dangers_away21:43
jogofungi: ^21:43
openstackgerritJames E. Blair proposed a change to openstack-infra/gear: Use non-blocking IO
clarkbjogo: wasn't it higher than 10021:44
clarkbjogo: so 100 is effectively lowering it?21:44
jogoclarkb: large-ops-testing was 175 but that didn't work21:44
jogoclarkb: I want to bump large-ops itself21:44
jogolarge-ops testing was to find a good testing number21:45
anteayamgagne: wasn't ignoring you, I didn't know the answer, thanks for sharing it21:45
davidlenwellanteaya: I would do that.. but we've been testing against it for months and it wasn't an issue until last week.. I'll add a few tests now that the api spec is finished. but my curriosity and questions stemed from the change in behavior21:45
*** wenlock_ has quit IRC21:45
jogoerr good number that we could bump the value too21:45
clarkbjogo: ah gotcha21:45
anteayadavidlenwell: okay, I don't have the answer21:45
jogolarge-ops started at 150 at first but cloud variance forced us to drop it21:45
mgagnein fact, it looks like the line "cherry picked from" is not added if the change you are cherry picking from is not yet merged.21:45
jogoclarkb: so I am thinking a few runs isn't enough data points to confirm the job is stable enough21:46
jeblairclarkb, fungi, mordred: addresses the issue we saw with zuul/gear/nodepool21:46
jogoclarkb: which is why I was thinking silent queue and I can track it via graphite21:46
anteayamgagne: cool, I didn't know that21:46
clarkbjogo: wfm21:47
jeblairclarkb, fungi, mordred: it's a substantial change.  considering that gear primarily wants to be a client library, and the problem was predominantly with the server, i think we have a few options.21:47
*** mrda-away is now known as mrda21:47
jeblairclarkb, fungi, mordred: we should evaluate that change on whether it improves things for clients and at a reasonable complexity cost21:48
*** Kevin_Bishop has joined #openstack-infra21:48
jeblairclarkb, fungi, mordred: on balance, i think it might make things slightly better for clients in bad network situations or which are moving lots of data around.  though it does make client calls like "send work data" non-blocking now.  not sure if that's a good thing or not.21:48
*** nati_ueno has quit IRC21:49
*** jvrbanac has joined #openstack-infra21:49
*** paul_glass has joined #openstack-infra21:49
jeblairclarkb, fungi, mordred: i'm not sure whether it's more complicated or not.21:49
clarkbjeblair: client objects can still query the Connection object for its status right?21:50
clarkbso this isn't really fire and forget, it just becomes a bit more work to remember21:50
dhellmannlifeless: that sounds like a good plan for the specs21:51
jeblairclarkb: i don't think i made a reliable way to do that.  i think it's more fire and forget.21:51
jvrbanacHowdy all! Is this the right place to ask a question about an error message i'm seeing with OpenStack Gerrit?21:51
jeblairclarkb: so if you are a worker, and you sendWorkComplete and actually sending that fails, i don't think you'd know.21:52
jeblairwhereas at the moment, that is synchronous21:52
davidlenwellanteaya: thank you for finding that for me21:52
davidlenwellmakes sense now21:52
*** bcrochet is now known as bcrochet|g0ne21:53
anteayajvrbanac: what is the error you are seeing?21:53
fungijeblair: i wonder whether the calls could be made to optionally block... or would that almost certainly increase complexity?21:53
anteayajvrbanac: if it is more than two lines, please use paste.openstack.org21:54
fungijeblair: and yeah, i guess that is a fairly big change, looking at it now21:54
jeblairclarkb, fungi, mordred: so if we decide that non-blocking is an anti-feature for clients, then i think we have two options:21:54
jogoclarkb: thanks21:55
jeblairclarkb, fungi, mordred: a) give geard its own poll loop (so it can do epoll edge-triggered non-blocking io without touching the client bits)21:55
jeblairclarkb, fungi, mordred: b) start hacking on the C server21:55
*** maxbit has quit IRC21:55
* fungi likes how "b" is also "c"21:56
anteayaone big ball of string21:56
lifelessjeblair: FWIW I am of the long-held opinion that all clients should be nonblocking, with blocking a thin shim on top.21:56
jvrbanacanteaya, Two of my devs (paul_glass & Kevin_Bishop) are trying to get setup with Gerrit and git review is telling them they need to have "current contact information"; however, when they attempt to update their information in Gerrit, they get a "Server Error - Cannot store contact information"21:56
anteayathey haven't become foundation members yet21:57
anteayawhat instructions are you following?21:57
jeblairclarkb, fungi, mordred: now that i have more or less the whole problem space in my head, i think (a) may be a much smaller change than this one, and generally not insane.21:57
*** dhellmann is now known as dhellmann_21:58
*** mriedem has quit IRC21:58
fungianteaya: jvrbanac: yeah, with the recent gerrit upgrade we lost the more descriptive error message we had hacked in there, so it's a little less instructive than it used to be. the faq entry for that is probably21:58
*** dhellmann_ is now known as dhellmann21:58
*** dhellmann is now known as dhellmann_21:58
nibalizerkrotscheck: im a bit confused, is this the code from the existing opentack-infra/config module or a new one?21:58
*** wenlock_ has joined #openstack-infra21:59
fungijeblair: i think there's still merit in maintaining a pure python implementation of the server to use in tests anyway, so might as well make it usable for production21:59
jvrbanacanteaya, fungi Opps! I didn't see that when I did a quick google search. We'll give that a shot. Thanks!21:59
fungijvrbanac: the narrative instructions (which list things needing to be done in the proper sequence) are at
clarkbmordred: by the way 1.1 has a new failure mode22:01
clarkbmordred: we are getting 500 responses22:01
clarkbjvrbanac: sure22:01
clarkbjeblair: I think that gear should at least make you aware of that22:01
clarkbjeblair: maybe by togglign a flag on the connection object that can be checked22:01
*** gondoi is now known as zz_gondoi22:01
*** ekarlso has quit IRC22:01
clarkbI am doing a proper review of the change now22:01
jeblairclarkb: okay.  as you can see, i consider the entire approach up for debate -- so keep that in mind as you review.  :)22:02
*** arnaud has quit IRC22:03
fungii have skimmed the change, but am going to flee to dinner and try to have a look at it when i get back22:03
jvrbanacfungi, that fixed it. Thanks a ton! Sorry to bug you about that.22:04
fungijvrbanac: no worries. just glad it wasn't something severely broken we needed to fix!22:04
* fungi -> out22:05
*** atiwari has joined #openstack-infra22:05
openstackgerritA change was merged to openstack-infra/pypi-mirror: No longer whitelist externally-hosted requirements
*** yamahata has joined #openstack-infra22:06
*** nati_ueno has joined #openstack-infra22:07
*** greghaynes has quit IRC22:07
clarkbjeblair: if a is much smaller and simpler maybe there is a better place to start?22:07
clarkbjeblair: then we can cut a 1.0 and start working on the less backward compatible client changes?22:07
*** greghaynes has joined #openstack-infra22:09
jeblairclarkb: it's an idea, though we'd end up with client and server being nonblocking, as they are in this change.  basically, that's probably the route to go if we want both to be nonblocking and we want the server to be extra awesome.22:10
openstackgerritJoe Gordon proposed a change to openstack-infra/config: Move large-ops-testing to silent queue for further testing
jogothat commit message took 10 times longer to write then the patch ^22:11
jeblairjogo: and it's still longer than 50 chars :P22:12
jogojeblair: heh22:12
*** jgrimm has quit IRC22:13
jogoon an unrelated note:
jogois pypi.o.o having issues?22:15
*** zehicle_at_dell has joined #openstack-infra22:15
*** bknudson has quit IRC22:16
*** dstanek is now known as dstanek_zzz22:19
*** pdmars has quit IRC22:19
*** dstanek_zzz is now known as dstanek22:20
*** whoops has quit IRC22:21
*** kaisero has quit IRC22:21
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for 1320617
*** Kevin_Bishop has quit IRC22:24
*** zehicle_at_dell has quit IRC22:25
*** andreykurilin has quit IRC22:26
*** r-daneel has quit IRC22:26
*** paul_glass has left #openstack-infra22:26
*** ekarlso has joined #openstack-infra22:26
*** greghaynes has quit IRC22:27
*** greghaynes has joined #openstack-infra22:28
clarkbjeblair: commented on the change. pep8 failed it for a few reasons22:28
clarkbjeblair: I think I am ok with that approach if we address my inline comment about setting a flag on the connection22:29
clarkbor a flag somewhere so that clients/workers can check if things have gone crazy22:29
*** UtahDave has quit IRC22:29
*** mburned is now known as mburned_out22:29
clarkband maybe make that flag = errno22:29
*** weshay has quit IRC22:29
jeblairclarkb: this is so that clients/workers could see what's going on?22:31
*** jp_at_hp has joined #openstack-infra22:31
clarkbjeblair: ya22:31
clarkbthough I was mostly thinking about this from a debugging/logging perspective rather than in code error handling22:32
clarkband gear should do the logging for us right?22:32
*** atiwari has quit IRC22:32
*** lakshmiS has quit IRC22:32
jeblairclarkb: connection losses are always logged, yeah22:32
*** greghaynes has quit IRC22:32
clarkbso maybe that isn't as useful as I thought it might be22:33
*** Ryan_Lane1 is now known as Ryan_Lane22:33
*** Ryan_Lane has quit IRC22:33
*** Ryan_Lane has joined #openstack-infra22:33
*** greghaynes has joined #openstack-infra22:34
*** mrodden has quit IRC22:34
*** ArxCruz has quit IRC22:34
*** rfolco has quit IRC22:35
*** thedodd has quit IRC22:35
*** marcoemorais has quit IRC22:36
*** marcoemorais has joined #openstack-infra22:36
openstackgerritcraigtracey proposed a change to openstack-infra/config: Add nova-docker to the gerritbot config
*** dangers_away is now known as dangers22:37
*** greghaynes has quit IRC22:37
*** greghaynes has joined #openstack-infra22:39
*** melwitt has quit IRC22:39
*** signed8b_ is now known as signed8bit_ZZZzz22:44
*** signed8bit_ZZZzz is now known as signed8b_22:46
*** signed8b_ is now known as signed8bit22:46
*** dkliban is now known as dkliban_afk22:47
*** melwitt has joined #openstack-infra22:47
*** 14WAC1CKA has joined #openstack-infra22:49
*** 14WAC1CKA has quit IRC22:49
*** jhesketh has joined #openstack-infra22:49
jeblairdevananda: i posted a reply to your ironic ci thread22:51
*** mmaglana has joined #openstack-infra22:51
*** dims has quit IRC22:51
clarkbjhesketh: o/22:52
*** dims has joined #openstack-infra22:52
clarkbjhesketh: so zuul is speaking swift now and I think I reviewed your JJB changes to put that into jobs, but I didn't see a job that uses it. Am I just not seeing that change?22:52
jheskethclarkb: oh cool, thanks22:53
jheskethbut no, no job doing that yet22:53
jheskeththat's next I guess :-)22:53
clarkbah ok, so I think thats next22:53
clarkbjhesketh: maybe add it to the config jobs22:53
clarkbjhesketh: and leave the scp in place so that we get two copies of logs for now22:54
jheskethclarkb: btw, could you take a look at my reply to this please
jheskethyeah I was going to make up a fake job but that's a good idea22:54
*** homeless has quit IRC22:54
clarkbjhesketh: for 90883 what if we restrict the data available to regular node names22:55
clarkbjhesketh: rather then changing what subnodes know about22:55
jheskethI'm not sure I follow sorry22:55
clarkbjhesketh: in the comment you mention subnodes don't have the full data set22:56
clarkbjhesketh: but we could go the other way and only allow self.node_id in launchNode22:56
clarkbjhesketh: then it would be consistent with launchSubNode22:57
jheskethah right22:57
jheskethbut then the names won't be able to be the same as already exists22:57
*** signed8bit has quit IRC22:57
clarkboh, does that mean sub node names can't be the same as would exist today if we were using subnodes?22:57
*** signed8bit has joined #openstack-infra22:57
*** arnaud has joined #openstack-infra22:58
clarkbactually it looks like nodeid is all that we use22:58
jheskethactually I just checked and it could be22:58
clarkbI think that would be better and will prevent drift in subnode and node names'22:58
*** msabramo has quit IRC22:59
jheskethso we allow provider and label objects, it'd be a pity to miss out on accessing the node object, but I tend to agree22:59
clarkbjhesketh: ya23:03
*** flaper87 is now known as flaper87|afk23:03
*** BadCub has quit IRC23:04
*** msabramo has joined #openstack-infra23:07
*** W00dy__ has quit IRC23:07
*** wenlock_ has quit IRC23:08
*** yamahata has quit IRC23:09
*** prad has quit IRC23:14
davidlenwellcan someone please remind me which file in config indicates the voting things on my project?23:15
clarkbdavidlenwell: modules/openstack_project/files/zuul/layout.yaml23:15
*** Sukhdev has quit IRC23:15
openstackgerritKhai Do proposed a change to openstack-infra/jenkins-job-builder: remove misleading jjb example
davidlenwellthanks clarkb23:17
openstackgerritDavid Lenwell proposed a change to openstack-infra/config: removed py26 testing from refstack
clarkbjeblair: fungi mordred 93950 adds a fun test to the infra/config project23:19
openstackgerritKhai Do proposed a change to openstack-infra/jenkins-job-builder: remove misleading jjb example
fungidavidlenwell: keeping in mind that any project templates your project entry includes in that file are also going to be run23:20
davidlenwellfungi:  that statement confuses me23:20
*** amotoki has joined #openstack-infra23:20
clarkbso I notice the redhat and hp tripleo clouds have significantly less fail going on23:20
fungidavidlenwell: which project? i can elucidate23:20
clarkbwhich is really cool23:20
fungiand by restack i assume refstack23:21
davidlenwelltyping is hard23:21
*** mburned_out is now known as mburned23:22
fungidavidlenwell: so... easy. refstack has no additional project templates at
davidlenwellI didn't think so23:22
*** hogepodge has quit IRC23:23
fungidavidlenwell: so it's just a matter of looking in the jobs section of that file for patterns any of the job names mentioned in your check and gate pipelines23:23
davidlenwellsweet .. thankf for the info23:23
fungidavidlenwell: noting that at voting is disabled for the gate-refstack-python33 job23:23
*** aconrad has quit IRC23:24
fungiand also for gate-refstack-python26 just below it23:24
davidlenwellyes .. as intended23:24
davidlenwellI removed that23:24
*** salv-orlando has quit IRC23:24
davidlenwellin that last commit23:24
davidlenwellI want non voting py33 but not py2623:24
fungiso the rest are presumably voting, assuming no unduly complicated regular expressions which missed my simple keyword search23:24
davidlenwellthanks again fungi and clarkb23:25
openstackgerritA change was merged to openstack-infra/config: Add serveraliases to cgit vhost template
*** salv-orlando has joined #openstack-infra23:29
fungidavidlenwell: and the change you linked is definitely not a match for what it does. comments added inline23:29
davidlenwelloh .. okay.. looking at that now23:30
*** radez is now known as radez_g0n323:30
openstackgerritDavid Lenwell proposed a change to openstack-infra/config: removed py26 testing from refstack
davidlenwellfungi:  better ? ^^23:31
clarkbjeblair: for 94440 (infra-specs initial commit) I think there are a few things that need addressing then we should probably get that in so that we can start working on real spec review and writing23:33
clarkbjeblair: also, I am a bit confused as to why we need a coveragerc and so on, but I think that came from cookiecutter so meh23:33
fungidavidlenwell: yep! +223:34
davidlenwellthank you!23:34
*** dstanek is now known as dstanek_zzz23:37
*** jp_at_hp has quit IRC23:37
*** jp_at_hp has joined #openstack-infra23:38
*** zzelle has quit IRC23:38
clarkbmordred: can you address the comment in 90234?23:38
*** mmaglana has quit IRC23:40
clarkbpleia2: 89928 would it be reasonable to ask for a mirror config that handled multiple ubuntu versions?23:40
clarkbpleia2: you should be able to make the variable an array then iterate over that in the erb file23:40
clarkbpleia2: that way we can mirror precise and trusty23:40
openstackgerritA change was merged to openstack-infra/config: Add change-abandoned hook for gerrit in puppet
clarkbpleia2: nevermind I see it is a define now and we can have multiple defines23:41
*** aconrad has joined #openstack-infra23:42
openstackgerritA change was merged to openstack-infra/config: removed py26 testing from refstack
*** oomichi has joined #openstack-infra23:46
clarkbpleia2: I commented on it with what I could figure out. Let me know what you think23:46
*** dstanek_zzz is now known as dstanek23:47
*** salv-orlando has quit IRC23:52
*** gyee has joined #openstack-infra23:53
*** jaypipes has joined #openstack-infra23:54
fungijesusaurus: any objection to me rebasing for you?23:56
* clarkb runs home back in a bit23:56
clarkbmordred: if you get back to a keyboard, I think we should follow up with the 1.1 stuff more aggressively. The failure mode appears to have changed and the amount of error is even greater23:57
jesusaurusfungi: oh, here, i'll take care of that in a sec23:59

Generated by 2.14.0 by Marius Gedminas - find it at!