Thursday, 2015-02-05

clarkbjogo: so maybe shelve is broken with >1 node even with one router?00:00
fungigreghaynes: great (and terrible) point00:00
clarkbgreghaynes: oh it uses git protocol? I guess if its just a set of pointers to the internet that would be a problem00:00
clarkbtl;dr learn to make mirrorable package repos00:00
clarkbmordred: r1chardj0n3s we could also do the terrible thing of setting up our own pointers that were all local then make a mirror of the pointer list00:00
greghaynesyep, and basically do a rewrite of the references to point to your own git repos00:01
mordredthats what I was basically thinking00:01
clarkbjogo: they all look like ssh timeouts so likely networking problem of some sort00:01
mordredgo grab the index - fetch all of the git repos, and then rewrite the index and serve it out00:02
clarkbjogo: it is possible that we need part of my multi host change where the public and flat interfaces are completely separate. ebtables got cranky with them together00:02
clarkbgreghaynes: you should create that project on openstack-infra the name is good00:02
dstufftr1chardj0n3s: obivously you should just write a totally new package manager and a new index website00:03
greghayneswell, if this is what the solution is then I think the name fits great :)00:03
dstufftmake it better than bower00:03
*** ociuhandu has quit IRC00:04
clarkbjogo: I am going to split the multi host change in two. first part with the separated interfaces and second turn on nova multihost00:04
fungigreghaynes: our member companies are already employing rube _and_ goldberg to work full-time on our projects anyway00:05
mordredso bower already knows how to talk to a local registry00:05
mordredwhich means that part is part of the standard toolchain - AND that can be configured in a config file, so we can drop the file on our servers and the repos and scripts could still work for normal people00:06
greghaynesmordred: as long we dont allow people to use absolute uri's00:06
greghayneswhich seems reasonable00:06
* dstufft considers writing a guide to not re-learning all of the problems in a new packaging ecosystem that everyone else already learned and solved (or not solved as the case may be)00:07
mordreddstufft: if only people in new language ecosystems would read it00:07
dstufftmordred: I never realized how bad an assumption it was that "the internet generally works" was until software started relying on that assumption in test runs across the world00:10
mordred <-- bower registry server in django00:10
fungidstufft: you're totally not talking about pypi there, i'm sure00:10
greghaynesor even generalize it a bit more to "the network generally works" :)00:10
dstufftfungi: all I'm gonna say is if your package manager doesn't support retries you're gonna be in for a bad time00:11
fungigreghaynes: or "people will just try again later"00:11
dstufftpip still doesn't handle retries during downloading the file itself, just during connection00:12
dstufftand I'm trying to fix that00:12
fungidstufft: i think we'll collectively be happier once you do, so thank you00:12
*** arosen has quit IRC00:13
dstufftit requires some changes in urllib3 to make it happen, I have a +1 on the idea from shazow (urllib3 maintainer) I just need to implement it :|00:13
*** dangers is now known as dangers_away00:14
dstufftAlso non obvious thing: indirection is so important, absolute urls (ala golang's importing from github) are the worst idea00:14
greghaynesalso, that makes dealing with major versioning HILLARIOUS00:14
mordredalso - any tool that ASSUMES github is fail00:15
greghaynes(using absolute urls that is)00:15
openstackgerritClark Boylan proposed openstack-infra/devstack-gate: Enable nova multi host on aiopcpu tests
openstackgerritClark Boylan proposed openstack-infra/devstack-gate: Separate public and flat interfaces for multinode
dstufftwell there are ways around like, that using absolute urls that are versioned00:15
clarkbjogo: ^ there I split them. Will run experimental jobs on them now but I think we may need that first one then the second can happen when shelve works00:15
dstufftmordred: I don't think it actually assumes github, but just that you can point to random vcs repos00:15
dstufft this is my favorite story00:15
mordredoh wow00:17
*** dmsimard is now known as dmsimard_away00:17
mordredI never should have read that00:17
*** dims_ has quit IRC00:17
mordred"So basically, the way Go handles dependencies like headers is to hardcode a path to a github repo that has the header you want then download a local copy of that header so you can edit it independently of the repo author because all Go programmers are enlightened perfect angels and this could never cause dependency problems ever."00:18
*** Sukhdev has quit IRC00:18
fungiin case i ever needed another reason to be even more allergic to go...00:19
mordredalso - we use puppet and not chef for infra because chef required git, fwiw00:20
*** annegent_ has quit IRC00:20
mordredI dont' know if that proves or disproves anything - but it makes me amused everytime I think about it00:20
clarkbmordred: puppet was chosen during the Bee Zed ARRR days?00:21
mordredI guess I shoudl say "one of the reasons we use puppet and not chef" - because that blocked me from playing with chef further which made puppet the de facto winner...00:21
mordredclarkb: yah man00:21
mordredclarkb: have you ever seen my script that emits puppet manifests?00:21
clarkbI don't think so00:21
clarkband now I am slightly afraid00:21
mordredoh - you should be very afraid00:22
mordredit was before jeblair figured out how to do virtual resources00:22
mordredbecause all I really wanted to do first was manage user accoutns00:22
mordredwhich turns out to be possibly the most complex thing to do in puppet00:22
dstufftianw: ping00:23
clarkbmordred: its funny too that the common example for puppet is to set up a user but then as you dig in more you realize its a lot more complicated than it seems00:23
dstufftmordred: I've used chef and salt now, I prefer salt over chef :| none of them make me particularly happy though00:24
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Added OAuth check for response_type 'code'
r1chardj0n3ssorry folks00:25
r1chardj0n3sso, if I understand correctly, infra is going to set up a bower registry with all the files we need in it and we don't need
jogoclarkb: cool00:26
sigmavirus24r1chardj0n3s: why not
jogoclarkb: about to look into the failures myself00:26
r1chardj0n3ssigmavirus24: you're not helping00:26
* sigmavirus24 isn't wearing his helping hat. It's meeting time not helping time00:27
*** stevemar has quit IRC00:27
*** craigbr has joined #openstack-infra00:28
r1chardj0n3sso django-bower-registry needs an Internet connection itself to fetch packages into itself...00:28
*** marun has quit IRC00:28
clarkbjogo: if 153093 works try rebasing on that00:28
*** VijayTripathi has quit IRC00:28
r1chardj0n3salso, the bower registry is configured in the bower.json file in the project, which means either the gate has a modified bower.json file, or users do, or users all hit the gate's registry00:29
mordredclarkb: very very early iterations on puppet00:29
*** jpeeler has joined #openstack-infra00:30
*** jpeeler has quit IRC00:30
*** jpeeler has joined #openstack-infra00:30
clarkbwith open("users.pp", "w") as user_pp: <- \o/00:30
mordredr1chardj0n3s: oh - we're certainly not decided on a course of action or actual technology00:30
dstufftr1chardj0n3s: bower doesn't let you have a site specific bower.json? :|00:30
clarkbmordred: that is awesome00:30
mordredr1chardj0n3s: mostly just brainstorming00:30
mordredclarkb: right?00:30
mordredclarkb: I believe that was the first thing jeblair fixed when he started00:30
*** amitgandhinz has joined #openstack-infra00:31
mordredI'm sure you can imagine the look on his virtual face00:31
sigmavirus24clarkb: congrats00:31
r1chardj0n3smordred: ok. this is something Horizon is bumping up against, hence my proposal of as a low-resistance solution (which is 70% implemented because I had an itch I wanted to scratch at LCA)00:31
clarkbsigmavirus24: wait what did I do?00:31
mordredr1chardj0n3s: yah - I think it's been very useful bumping and we've all learned a bunch more things00:31
sigmavirus24clarkb: you made me exhale rapidly00:32
*** mriedem has joined #openstack-infra00:32
sigmavirus24few people do that00:32
*** tnovacik has quit IRC00:33
r1chardj0n3sI guess the registry search order could work, assuming that the gate used a registry that no-one else could contact00:33
mordredr1chardj0n3s: yuck on no global registry config option00:33
*** annegent_ has joined #openstack-infra00:33
mordredr1chardj0n3s: it's entirely possible that we're goign to need BOTH AND a mirror of some sort :)00:33
* sigmavirus24 still thinks is a better name00:34
r1chardj0n3smordred: my solution included creating a mirror for to use - that's where the "cache" came in00:34
*** wenlock1 has quit IRC00:34
*** bhunter71 has joined #openstack-infra00:34
clarkbsigmavirus24: and now I have the super mario song stuck in my head00:34
clarkbalso I may play some mario tonight00:34
sigmavirus24you're welcome00:34
sigmavirus24win-win if you ask me00:34
*** gyee has quit IRC00:34
dstufftr1chardj0n3s: I think sychronizing a cache is a super hard problem compared to setting up a mirror server ala bandersnatch00:34
r1chardj0n3ssigmavirus24: if you hang around here much more I'm totally getting your boss to second you on to working on this :)00:34
*** amitgandhinz has quit IRC00:35
r1chardj0n3sdstufft: the problem is you're trying to mirror *github*00:35
r1chardj0n3s(and its APIs)00:35
dstufftdoes bower actually use github's API, I thought it just poked to find some git repos and then sucked them down00:35
r1chardj0n3sdstufft: no, it's not like that00:35
sigmavirus24r1chardj0n3s: fair enough00:35
sigmavirus24I'll stick around however long it takes00:35
*** Sukhdev has joined #openstack-infra00:36
r1chardj0n3sdstufft: the "cache" which I shall no longer refer to as a "cache" but rather a "mirror" would do precisely what you're talking about00:36
*** rverchikov has joined #openstack-infra00:36
dstufftr1chardj0n3s: ok I'm re-asserting my suggestion to write a whole new packaging mananger and repository instead of bower and convince all the front end devs to use it instead00:36
clarkbso I really wonder how terrible it would be to just host in npm then everyone already gets a mirror00:36
sigmavirus24dstufft: write it in rust00:36
r1chardj0n3sdstufft: undertstanding the API of github (like bandersnatch does for pypi) it copies all the stuffs for various packages (limited in scope, unlike bandersnatch, because copying everything would be insane)00:37
clarkbjust have not npm client suck them down and don't run code to do it00:37
*** achanda has quit IRC00:37
clarkbmordred: clearly we just need to mirror all of github00:37
r1chardj0n3sdstufft: I'm effectively doing that thing: writing a new packaging manager (but a very small one)00:37
sigmavirus24r1chardj0n3s: it'll grow00:37
sigmavirus24just like distutils00:38
dstufftclarkb: generalized to, clearly we need to mirror all of the internet00:38
r1chardj0n3ssigmavirus24: ah, but in this case I have confidence that it won't :P00:38
sigmavirus24== dstufft00:38
dstufftopenstack infra aka archive.org00:38
sigmavirus24r1chardj0n3s: that's what you said 20 years ago00:38
*** annegent_ has quit IRC00:38
* sigmavirus24 wasn't cognizant 20 years ago00:38
r1chardj0n3ssigmavirus24 oh, sure, throw pypi back in my face00:38
clarkbdstufft: exactly00:39
fungir1chardj0n3s: if anybody's going to get pypi thrown back in their face, then... yeagh00:39
mordredr1chardj0n3s: oh - right - because people can, and do, list things that are not in the registry00:39
r1chardj0n3smordred: yeah, but no-one does00:39
clarkbdstufft: that would actually be a fun thing to do. we just need some container datacenters we can expand as the internet gets bigger00:39
openstackgerritElizabeth K. Joseph proposed openstack-infra/puppet-zanata: Initial commit of Zanata-specific sections
jogoclarkb: the  tempest.scenario.test_minimum_basic.TestMinimumBasicScenario.test_minimum_basic_scenario is failing on the shelve fix as well00:39
jogoclarkb: so just maybe the shelve fix is actually working00:39
r1chardj0n3sit's like bower has the capability for refs to point *all over the place* but in practise everyone uses github00:39
jogoclarkb: and we have a regression somewhere00:39
clarkbjogo: could be00:39
mordredr1chardj0n3s: oh? so you only have to mirror the parts of github that are first referenced in bower.io00:39
r1chardj0n3smordred: yes00:39
clarkbjogo: I think should give us answers (and we should run it a few times)00:40
jogoclarkb: same very high level failure00:40
r1chardj0n3smordred: though I would limit it to just the stuff referenced in bower.json for the project00:40
mordredr1chardj0n3s: or is is possible someone's bower.json would refer to emonty/foo with that being a github reference but not registered in bower.io00:40
jogoclarkb: ahh yes00:40
jogoit should00:40
r1chardj0n3smordred: no, we control the bower.json, we can dictate what goes in it00:40
clarkbjogo: since that should have known good networking other than the nova networking stuff00:40
mordredr1chardj0n3s: right - but if we just limit it to things in the bower.json for the project (which we might have to do mind you) we're still stuck with needing to update a central bower.json to update a mirror or a cache and _then_ landing the change to use it00:41
clarkbjogo: er nova multihost00:41
r1chardj0n3smordred: hmm, yes00:41
mordredr1chardj0n3s: I'm not saying that's a non-starter per-se - just that we've had two different systems that worked like that in the past00:41
mordredand both got pushback from devs being confused00:41
*** cody-somerville has quit IRC00:41
mordredso - word of warning00:41
jogoclarkb: I think the other two failures in tempest-full are related to not running a second metadata server00:42
clarkbso hear me out on a slightly crazy idea00:42
* mordred listens00:42
clarkbwhat if we made git.o.o mirror more than just review.o.o repos, then used submodules00:42
clarkbthat is terrible for a couple reasons but might be simpler overall?00:42
clarkbjogo: thats what adam_g's change fixes? we can put that one low on the stack too00:42
mordredso ...00:42
mordredclarkb: ansible just recently moved to using submodules for its module library00:42
jogoclarkb: yeah ... its right above yours I think00:43
dstufftianw: whenever you're around again, please remind me what fix you needed in pip00:43
clarkbjogo: ya but maybe it should be under00:43
mordredclarkb: and I continually have no idea how to get my working tree in the right state00:43
jogobbiab time for a run00:43
mordredclarkb: and I, it turns out, am one of the most advanced git users on the planet00:43
clarkbmordred: I agree submodules complicate how to keep repo in good state00:43
mordredmuch to my chagrin00:43
jogoclarkb: possible, well once the results are in from your new patch we should know a lot more00:43
jogobut we are getting pretty there00:43
*** achanda has joined #openstack-infra00:43
mordredmost of our devs are not, it turns out, the most advanced git users on the planet- no offense towards them intended00:44
*** ChuckC_ has joined #openstack-infra00:44
clarkbmordred: ya I am there with you on not liking submodules00:44
clarkbbut I think this is he problem submodules actually exist to solve00:44
mordredyes. it is00:44
*** mbacchi has quit IRC00:44
Shrewsmordred: i think your ansible submodule problems come from github, not so much git00:44
mordredbut that still doesn't mean we should use it00:45
clarkbits when you need to include files from a specific version of a thing. its not a replacement for packaging but it sounds like we don't actually have real packaging here00:45
mordredclarkb: your system would be just as laggy as r1chardj0n3s's00:45
r1chardj0n3smordred: yep, I can understand that. so if your change uses a new bower component not previously seen, you have two options: 1. your change will fail for up to 24 hours, or 2. you have to submit a separate change to get the component in for testing00:45
mordredbecause if we wanted to add a new depend00:45
*** craigbr has quit IRC00:45
greghaynesIt just occured to me that even if we say that we will only use registered names for bower packages we need we cant guarantee that those packages dont have dependencies that are absolute urls00:45
mordredwe'd need to have INFRA add that depend to our git mirrors00:45
*** ChuckC has quit IRC00:45
mordredso that a submodule pointing to our mirror of the git repo could be added to the master repo00:46
clarkbmordred: yes except that it would be all history so at least versions within that mirror wouldn't have lag00:46
mordreddon't care00:46
clarkbbut yes adding completely new thing would have lag00:46
mordredI'm talking about dev workflow for adding a dep00:46
mordredand the overhead that carries00:46
clarkbjogo: tempest.scenario.test_shelve_instance.TestShelveInstance.test_shelve_instance [122.746042s] ... ok fwiw00:46
dstufftmordred: I think I'm pretty good at git and submodules ranks somewhat around shoving a fork into my thigh in how happy I am to do them00:46
mordredr1chardj0n3s: you only have 200:46
mordredr1chardj0n3s: 1 can't happen because we gate00:46
*** dims__ has joined #openstack-infra00:46
mordredr1chardj0n3s: so your change that adds the new dep would never land, so we'd never process it00:47
mordredr1chardj0n3s: your only choice would be making a separate change00:47
* r1chardj0n3s smacks forehead00:47
mordredr1chardj0n3s: welcome to our world of insane sets of interwoven requirements :)00:47
r1chardj0n3sso either we mirror *all* of, or we have that 2-step dance00:47
mordredI think those are the two choices00:48
r1chardj0n3sI'm not even sure you can list all of bower.io00:48
r1chardj0n3slemme look at the API "spec" for the site00:48
mordredregardless of how we do the 2-step dance00:48
dstufftmordred: obviously what you need is the ablility to link two CRs in different repos together with a circular dependency so they merge together00:48
jogoclarkb: that one will pass if unshelve to same node00:48
mordreddstufft: funny you should say that ...00:48
clarkbjogo: ya definitely need to run it a fe times00:48
jogoyeah, although if it goes all green at least once that is a very good sign00:49
jogoit if doesn't fail at all in 10 runs then something is wrong too though ;)00:49
dstufftmordred: oh god there's a half done branch that does that isn't there00:49
mordreddstufft: not circular - we explicitly trap for those00:49
clarkbjogo: nope basic scneario and snapshot fail so It hink it must be a regresssion00:49
clarkbmtreinish: ^ fyi00:49
mordreddstufft: but jeblair is about to land declarative depends - so you can indicate when a patch depends on a patch in another repo landing first00:50
r1chardj0n3soh, yep, we can do a 100% mirror if we want to. no idea how long it'd take to build, or how grumpy github might get tho ;)00:50
fungirandom aside: git submodules still strike me as a solution in search of a problem00:51
fungirelated: having that hammer makes a lot of problems look like nails00:51
*** rmcall has quit IRC00:51
sigmavirus24dstufft: your worst nightmares come true?00:51
*** annegent_ has joined #openstack-infra00:52
*** annegent_ has quit IRC00:52
mordredr1chardj0n3s: ok. so we can then in theory build one, should we decide that we want to do such a thing00:53
r1chardj0n3sthere are 26915 components listed in at present00:53
*** achanda has quit IRC00:53
r1chardj0n3sand at present we desire to have access to about 12 of those00:53
mordredr1chardj0n3s: it does look like one can use a .bowerrc00:54
mordredr1chardj0n3s: and you can put a registry in that00:54
fungisounds like roughly the same % of our pypi mirrors we actually use ;)00:54
openstackgerritElizabeth K. Joseph proposed openstack-infra/puppet-zanata: Initial commit of Zanata-specific sections
fungiwithin one or two insignificant orders of magnitude00:55
r1chardj0n3sok, but if you have the local registry there's still the question of how to populate it00:55
r1chardj0n3sI *guess* you could build the django app somewhere and copy it and its db to your imagew00:56
mordredyah. something it gonna have to go all bandersnatch on the situation00:56
mordredIm starting to be more and more in favor of the cache idea, tbh - one-day-lag for new cache notwithstanding00:57
*** zz_dimtruck is now known as dimtruck00:57
r1chardj0n3sI think the one day lag has been accepted (ish) in the past and can be in the future00:58
*** oomichi has joined #openstack-infra00:58
r1chardj0n3swell, where getting a change through to the packages list is done in a day00:58
clarkbjogo: interestingly only the shelve test fails on the multihost change00:58
clarkbjogo: I wonder if the regression is specific to multiple nodes with a single router00:58
clarkbjogo: I am going to grab a test env for the next time 153093 runs multinode tests00:59
fungir1chardj0n3s: is continuing along the existing xstatic path looking any more appealing yet?01:00
r1chardj0n3sfungi: gods no01:00
fungijust checking!01:01
*** sigmavirus24 is now known as sigmavirus24_awa01:03
*** melwitt has quit IRC01:04
*** chlong has joined #openstack-infra01:05
*** mohitsha_ has joined #openstack-infra01:06
openstackgerritDavid Shrewsbury proposed openstack-infra/shade: Add support for creating/deleting volumes
*** tmcpeak has quit IRC01:07
mordredr1chardj0n3s, clarkb: "terrible" idea - what if we had a post-job that ran on horizon that did r1chardj0n3s local cache building thing01:08
*** craigbr has joined #openstack-infra01:08
mordredand then we copied that somewhere - say swift for sake of argument01:08
*** shashankhegde has quit IRC01:09
mordredand then jobs that do something with horizon could have a "download cache dir" macro01:09
r1chardj0n3smordred, clarkb: oh, so outside of the gate, running the job on code that's committed but not tested?01:09
mordredso that there would still be a 2-step01:09
mordredbut it wouldn't be tied to image builds01:09
mordredr1chardj0n3s: no - I mean only running when a change to horizon is merged01:10
mordredand then maybe also nightly just cause01:10
mordred(to make sure we didn't fail on a post-job)01:10
r1chardj0n3smordred: ... but what about running it on commit? with appropriate levels of error handling, of course ;)01:10
mordredr1chardj0n3s: right - that's on-merge for us01:10
r1chardj0n3sthat way new packages would most likely be present when the tests are run01:10
mordredthere is no on commit01:10
* mordred thinking...01:11
clarkbjogo: ok I have env held. will poke around in it once the test fails01:11
*** tmcpeak has joined #openstack-infra01:11
mordredI think that will work nicer with AFS - and I'm not just saying that to troll01:12
mordredbut since this works well with local dirs and has a "don't talk to the internet" mode01:12
mordredthe post-job could do the caching and could have write permissions to the bower cache volume in AFS01:12
mordredAFS would handle making sure that we had local copies in each region01:13
mordredjob slaves would be configured to look in /afs/ or something for their local bower dir01:13
mordredand we could make that read access to system:anyuser01:13
mordredbut it means also we'd get differential writing out of the box01:14
mordredso that we arent' downloading and uploading tarballs of tarballs all the time01:14
mordredand as long as the process doesn't remove things that arent' the version we care about ...01:14
mordredwe could even have the same caching process triggered by anything that has a bower anything01:14
r1chardj0n3sI don't think anything would be removed01:14
mordredclarkb, fungi: ^^ am I smoking too much crack here?01:15
clarkbmordred: maybe? I actually can't speak much to afs as a mirror backend simply because it hasn't done much for me there01:15
r1chardj0n3sI still like the idea of on-merge instead of post-job, but I'm not close enough to infra to know why that's a horrid idea ;)01:15
fungimordred: afs does fix everything01:15
mordredclarkb: I don't mean as a backend - I mean as the 100% of the solution01:15
mordredr1chardj0n3s: the main thing there is the theory of bad actors01:16
mordredif we ran on patch submission01:16
*** doug-fish has left #openstack-infra01:16
mordredthena nyone could submit a patcha nd we'd run a thing on that patch which had write access to a location in our system01:16
fungiclarkb: we still need to put it into place. i got sidetracked on the git mirror backend part and need to pick that back up soon01:16
mordredthat's a security hole big enough to drive a truck through01:16
r1chardj0n3smordred: hmmm, yep01:16
clarkbfungi: right but I connected to it from my desktop and promptly gave up on it01:17
r1chardj0n3sok, so we cop a delay01:17
*** bradjones has quit IRC01:17
mordredr1chardj0n3s: once it's post-merge, the patch has at least been reviewed by someone, so the chances that it's a rootkit are lower :)01:17
clarkbfungi: performance isn't good enough to use it over a thing on the internet over http01:17
fungiclarkb: oh, i meant for our workers01:17
clarkbfungi: I imagine they will have similar issues01:17
clarkbfungi: at least until we have read only replicas in all regions which we don't yet iirc01:17
fungiclarkb: with a warmed cache on the images?01:17
mordredwe do ...01:18
clarkbmordred: oh I thought hpcloud was missing it01:18
mordredbut yes - read only replicas in all regions is the only way it helps or makes sense01:18
*** david-lyle is now known as david-lyle_afk01:18
*** bradjones has joined #openstack-infra01:18
mordredwithout that, it will not be helpful - it's smart-cached-data-locality which is the key in this case01:18
fungiwe do still need more afs servers for that bit, agreed01:18
*** andreykurilin_ has quit IRC01:19
*** jamielennox|away is now known as jamielennox01:19
*** otter768 has joined #openstack-infra01:19
*** Ryan_Lane has quit IRC01:19
*** chlong has quit IRC01:22
*** mtanino has quit IRC01:22
*** ChuckC has joined #openstack-infra01:23
*** annegent_ has joined #openstack-infra01:23
*** rverchikov has quit IRC01:23
*** otter768 has quit IRC01:24
r1chardj0n3smordred: so, where to now?01:25
*** ChuckC_ has quit IRC01:26
*** zhiwei has joined #openstack-infra01:28
*** rfolco has quit IRC01:28
fungir1chardj0n3s: the bottle01:28
fungi(getting late on my side of the rock)01:29
*** mohitsha_ has quit IRC01:29
fungiclarkb: specifically only two rax regions have fileservers at the moment, you're right01:31
clarkbyay I am not more crazy than I know01:31
* fungi would be willing to bet we're all at least marginally more crazy than we know01:32
mordredr1chardj0n3s: honestly, I've started drinking01:32
mordredr1chardj0n3s: I think we should talk about this at the next infra meeting01:32
jogoclarkb: it may have not been a regression, it may be your patch is actually needed01:32
mordredbecause we have some good ideas here, but we should sort through them before we ask people to work on a thing or promise to have a thing01:33
clarkbjogo: no my patch is breaking the same way01:33
jogoclarkb: link?01:33
clarkbjogo: so something somewhere has regressed01:33
*** shashankhegde has joined #openstack-infra01:34
*** dmsimard_away is now known as dmsimard01:34
jogoclarkb: yeah, :(01:34
clarkbjogo: and I am just waiting for the next run to fail so that I can manipulate the test env directly01:34
jogoclarkb: well metadata service failure failed 9/20: up 104.07. request failed01:34
clarkbjogo: hopefully will have some answers soon. I can give you access too if it beomes necessary01:35
*** annegent_ has quit IRC01:35
clarkbjogo: I don't know what that means. Is that from the logs?01:35
jogoclarkb: yes01:35
jogoand search for checking
jogoin one of the expanded failures01:36
clarkbjogo: want to rebase adam_g's change so that it is 1) mergeable and 2) testable without all of the other changes?01:36
jogoclarkb: yes, doing that now01:36
r1chardj0n3smordred: ok, thanks01:36
clarkbjogo: I see that in the logs now so metadata service is failing01:37
clarkbjogo: potentially because it isn't runnin on the subnode01:37
jogoclarkb: my hunch is that went to the second server01:37
jogolet me confirm that before rebasing01:37
fungir1chardj0n3s: if 1900 utc on tuesday is hard for you, just remember than jhesketh manages... or write something up for us ahead of time and sleep in--don't be a sucker01:38
clarkbya it could be. I should be able to confirm that things work on main node and not on sbubnode manually too01:38
r1chardj0n3sfungi: it's 6am for me which is doable just fine01:38
*** rverchikov has joined #openstack-infra01:38
fungir1chardj0n3s: 6am sounds unpleasant if you ask me01:39
r1chardj0n3sfungi: I'm usually up at 7, so 6 isn't a stretch01:39
jogoclarkb: yup its on the subnode 73e54df1-acfb-4707-ad5b-d6fb4a3f86501:39
*** dannywilson has quit IRC01:41
r1chardj0n3sand now: lunch01:41
*** tiswanso has joined #openstack-infra01:41
*** SumitNaiksatam has quit IRC01:41
*** SumitNaiksatam has joined #openstack-infra01:42
*** rwsu is now known as rwsu-afk01:43
openstackgerritJoe Gordon proposed openstack-infra/devstack-gate: Enable nova-api-metadata for multi host nova-network
*** shashankhegde has quit IRC01:44
jogoclarkb: ^ there is is, I pulled it out of the big stack01:44
ianwdstufft: see and and the issue in there
clarkbjogo: cool I have a node up that I can ping via its flat network and public network ip addrs01:45
fungislogan for next release: "walk softly and carry a big stack"01:45
clarkbjogo: but ssh works in neither case01:45
clarkbjogo: I am going to boot a second in a similar manner and do that until I get one on the main node01:45
jogoclarkb: ssh into the node fails if you can't hit the metadata service01:45
*** tiswanso has quit IRC01:46
ianwdstufft: with install -U, it seems pbr pulls in the latest pip; so it's hard to pin devstack down to a specific version of pip01:46
jogoclarkb: because ssh key is added via metadata service and not config drive (since its off for live migration in this test)01:46
*** ddieterly has joined #openstack-infra01:47
jogoso I expect ssh to work on instances on the main node01:47
dstufftianw: ok01:47
clarkbjogo: ok confirmed that I can ssh into nodes on the main node but not the subnode01:47
*** craigbr has quit IRC01:47
clarkbjogo: going to try confirming that metadata server is the problem01:47
jogoclarkb: awesome01:48
clarkbit seems to be trying to do diffie hellman then going nowhere01:48
*** salv-orlando has quit IRC01:48
clarkbso the service is up and running and I appear to have packets going to and from it01:48
clarkbjogo: I can ask qemu for the console log directly ya?01:49
jogoI think so, you can just ask nova though it should work too01:49
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth Functional Tests
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth check for response_type 'code'
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth Storage Abstraction Removed
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth check for no response_type
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth check for no client_id
jogoclarkb: that fits with a missing key01:49
dstufftonce tests pass I'll merge and release01:50
clarkbjogo: and I confirmed that metadata failed01:50
jogoclarkb: cool, so we have the fix it looks like01:50
clarkbjogo: and the reason this regressed was we turned off configdrive because of migrations?01:50
jogoonce the tests pass for that patch we can get it in and go from there01:50
clarkbjogo: +101:51
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: REMAININGfunctional tests for the OAuth Endpoint
*** vilobhmm has quit IRC01:51
jogoclarkb: not a regression per se. we forgot that patch was needed01:51
*** bdpayne has quit IRC01:51
jogoI turned of configdrive a while back01:51
jogofor live migratoin01:51
jogoconfig drive doesn't support it01:51
jogoclarkb: this was more of a human memory regression then a software one01:51
clarkbjogo: its odd that config drive doesn't support it, couldn't we copy the iso from old hypervisor to new and mount it as before?01:52
*** mohitsharma has joined #openstack-infra01:52
clarkbjogo: I suppose some details in the image may need to be updated?01:52
jogoclarkb: yeah, ask mikal about it01:53
*** achanda has joined #openstack-infra01:53
jogomikal: config drive and live migrate01:53
clarkbjogo: also since I split my change the change to enable multihost is much smaller and I think we can just wait on enabling that for later. But the other change should likely get in as it makes the networking much cleaner from the ebtables stance01:54
*** markvoelker has joined #openstack-infra01:54
*** vilobhmm has joined #openstack-infra01:55
*** EmilienM is now known as EmilienM|afk01:55
*** tmcpeak has quit IRC01:56
jogoclarkb: ++01:56
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Explicit Hook Priorities
jogoclarkb: thankfully I think we are at the point where we can work on some of this in parallel now that the basic thing is mostly working01:56
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Pass the old entity values in the NotificationHook
*** tmcpeak has joined #openstack-infra01:57
clarkbjogo: yup01:57
clarkbjogo: I think the only true dependency we have right now is the multihost dependency on network refactor01:58
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: API events now include resource before/after state.
*** craigbr has joined #openstack-infra01:59
clarkblol every 4 weeks sdague abandons the fix :)01:59
*** achanda has quit IRC01:59
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email base plugin.
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email Templating Engine
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email SMTP Sender
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email configuration.
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: [WIP] Email Send Cron Worker
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email Working Directory Utility Method
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Email Outbox
*** hdd has quit IRC02:01
*** vilobhmm has quit IRC02:02
clarkball the emails02:03
*** dmsimard is now known as dmsimard_away02:03
clarkbjogo: so why does neutron work without metadata service on all nodes?02:10
clarkbjogo: does it forward the metadata requests to the main node?02:10
*** tqtran has quit IRC02:10
jogoclarkb: AFAIK yes02:11
*** MarkAtwood has joined #openstack-infra02:13
openstackgerritBen Swartzlander proposed openstack-infra/project-config: Make manila tempest jobs votable
clarkbjogo: woo shelve instance failed02:14
*** dims__ has quit IRC02:14
*** atan8 has joined #openstack-infra02:14
clarkbjogo: so I am not entirely convinced that shelve instance works in any capacity02:15
clarkbya see
clarkbjogo: but maybe put my change that checks the nova fix on top of adam_g's change02:15
clarkbI will do that rebase now02:15
jogoyeah, that is a good test02:16
openstackgerritClark Boylan proposed openstack-infra/devstack-gate: DO NOT MERGE testing nova multinode fix
clarkbjogo: also failed tempest.scenario.test_snapshot_pattern.TestSnapshotPattern.test_snapshot_pattern02:18
jogometadata service still 2015-02-05 02:17:27.024 |     failed 8/20: up 92.50. request failed02:18
clarkbalso metadata server requests failed still02:18
clarkbI need to call it a night02:19
jogome too02:19
clarkbbut it appeas that metadata is more broken that it appears02:19
jogoto be continued02:19
jogowell its fickle heh02:19
clarkbhonestly I would be inclined to kill metadata with fire02:19
clarkbrequire config drive for your migration testing02:19
clarkbif only bceause metadata server seems to never be reliable02:19
jogohmm nothing got routed to it02:20
clarkbya its probably an underlying issue with the networking02:20
jogoI wonder if its because the devstack bits to set up routes arein the n-net bits02:20
clarkbthat shouldn't matter02:20
clarkbespecially since metadata service doesn't listen on that ip range02:20
*** ashleighfarnham has joined #openstack-infra02:20
clarkball we have done is allow each compute node to ping public floating ips of VMs on either compute node02:21
clarkbits all on the public network side and shouldn't touch the private networking02:21
jogoits the magic ip02:21
clarkbwhich is completely independent of
*** signed8bit has joined #openstack-infra02:22
*** craigbr has quit IRC02:23
jogohmm anyway we can dig in further tomorrow02:23
clarkbjogo: any idea what process that runs as?02:23
jogomaybe adam_g remembers02:23
clarkbthe metadata service02:23
jogoclarkb: yes
nibalizerianw: unfortunately i am in europe ttill the end of the week02:24
nibalizermakes collaborating with american timezones hard02:24
clarkbjogo: is the process name nova-api-meta?02:24
clarkblooks like it sorry metadata02:24
ianwnibalizer: no worries02:25
jogoyeah on the second node02:25
jogofirst node its part of n-api I think02:25
clarkboh what ok02:25
*** ashleighfarnham has quit IRC02:25
jogoI think whats going on is nova-network sets of the routes to
jogobut its not running on the second node ...02:25
jogoanyway really have to go, dinner time02:26
*** chlong has joined #openstack-infra02:26
*** tmcpeak has quit IRC02:31
clarkbjogo: ok progress for whenever you get back. The main node has the NAT rule for metadata service but the subnode does not02:32
*** SumitNaiksatam has quit IRC02:32
clarkbjogo: but maybe thats not a problem since everything private is on the network bridged on the main node02:34
dstufftianw: 6.0.8 is released, a new virtualenv will be shortly too02:34
ianwdstufft: awesome thanks ... that should get centos7 working in the gate again02:35
*** Sukhdev has quit IRC02:36
*** koolhead17 has joined #openstack-infra02:36
*** SumitNaiksatam has joined #openstack-infra02:36
*** thedodd has joined #openstack-infra02:37
*** thedodd has quit IRC02:37
*** ivar-lazzaro has quit IRC02:38
*** melwitt has joined #openstack-infra02:39
clarkbjogo: also according to docs we only need the second metadata service if we have two routers02:40
*** chlong has quit IRC02:40
*** tiswanso has joined #openstack-infra02:43
*** mohitsharma has quit IRC02:44
*** dims__ has joined #openstack-infra02:45
*** tiswanso_ has joined #openstack-infra02:46
*** dims_ has joined #openstack-infra02:46
*** tiswanso has quit IRC02:48
*** salv-orlando has joined #openstack-infra02:49
*** dims__ has quit IRC02:50
*** achanda has joined #openstack-infra02:51
*** mohitsharma has joined #openstack-infra02:51
*** yamahata has quit IRC02:55
*** yamamoto_ has joined #openstack-infra02:57
*** Longgeek has joined #openstack-infra02:57
*** rverchikov has quit IRC03:02
*** salv-orlando has quit IRC03:06
*** woodster_ has quit IRC03:06
*** dstufft|laptop has joined #openstack-infra03:06
*** marun has joined #openstack-infra03:11
*** stevemar has joined #openstack-infra03:12
*** dims_ has quit IRC03:12
*** rmcall has joined #openstack-infra03:13
*** rmcall has quit IRC03:14
*** rmcall has joined #openstack-infra03:15
*** marun has quit IRC03:15
*** rmcall_ has joined #openstack-infra03:19
*** otter768 has joined #openstack-infra03:20
*** rmcall has quit IRC03:21
*** rmcall_ has quit IRC03:23
*** r1chardj0n3s is now known as r1chardj0n3s_afk03:24
*** rmcall has joined #openstack-infra03:25
*** otter768 has quit IRC03:25
*** amotoki has joined #openstack-infra03:25
*** rmcall has quit IRC03:29
*** rmcall has joined #openstack-infra03:30
*** chlong has joined #openstack-infra03:30
*** woodster_ has joined #openstack-infra03:34
*** chlong has quit IRC03:37
*** stevemar has quit IRC03:38
*** mriedem has quit IRC03:39
*** stevemar has joined #openstack-infra03:40
*** craigbr has joined #openstack-infra03:45
*** sdake_ has quit IRC03:46
*** rmcall has quit IRC03:46
*** rmcall has joined #openstack-infra03:47
openstackgerritVenu Murthy proposed openstack-infra/project-config: In project ec2-driver updating to not run any tests and setting noop-jobs, moto is not ready, and sharing secrete keys is risky
*** MarkAtwood has quit IRC03:49
*** Longgeek has quit IRC03:53
jogoclarkb: nova-network sets up the NAT rule03:54
jogoI think it needs to be done manually on the second node03:54
jogofor not multihost03:54
*** harlowja is now known as harlowja_away03:56
*** rmcall_ has joined #openstack-infra03:56
openstackgerritVenu Murthy proposed openstack-infra/project-config: For project ec2-driver setting to noop-job in zuul layout
*** rmcall has quit IRC03:57
*** rmcall_ is now known as rmcall03:57
*** jerryz has joined #openstack-infra03:57
*** r1chardj0n3s_afk is now known as r1chardj0n3s03:59
*** tiswanso_ has quit IRC03:59
*** rmcall has quit IRC04:00
*** sdake_ has joined #openstack-infra04:01
*** hdd has joined #openstack-infra04:02
*** patrickeast has quit IRC04:03
*** rmcall has joined #openstack-infra04:03
clarkbjogo I am not sure I see the vm to 169 address packets on router node04:03
clarkbjogo but dont see return packets04:03
mikalclarkb: hello, I was at lunch04:05
mikalclarkb: do you still want to talk about config drive and live migrate?04:05
*** BharatK has joined #openstack-infra04:06
*** rmcall has quit IRC04:08
*** signed8bit has quit IRC04:08
clarkbI mean we are using metadata service to make live migrate work but metadata service doesnt work04:10
*** Longgeek has joined #openstack-infra04:10
clarkbmy consumption of public clouds has taught me this is normal04:10
clarkbso I pretty much think we should config drive everything04:10
openstackgerritBharat Kumar Kobagana proposed openstack-infra/project-config: GlusterFS: passthrough devstack config
openstackgerritBharat Kumar Kobagana proposed openstack-infra/project-config: GlusterFS: passthrough devstack config
clarkbmikal for even more lol metadata works with nova multihost but then shelve doesnt work04:17
clarkbso its like a trifecta of bugs preventing us from having multinode tests pass04:17
*** hughhalf is now known as hugh_afk04:18
*** mohitsharma has quit IRC04:19
*** mohitsharma has joined #openstack-infra04:20
jogoclarkb: AFAIK this is the missing bit
jogoiptables -t nat -A PREROUTING -d -p tcp -m tcp --dport 80 -j DNAT --to-destination $NOVA_API_IP:877304:22
clarkbyou have to manually add that on all nodes that arent the router?04:22
clarkbits in place on the router node04:22
jogoclarkb: yeah, that is usually done by nova-network04:23
jogohmm hold on04:23
jogoit may be a nova.conf setting now04:23
jogoahh CONF.metadata_host04:24
jogodefaults to $my_ip04:24
clarkbonly on the router04:24
jogoif you do a iptables-save on the subnode what does it say?04:24
clarkbit doesnt have that rule04:25
clarkbbut main node does04:25
clarkbpackets from subnode vm to 169 addr make it to main node04:25
clarkbI have a hunch its the return traffic that gets lost04:25
jogoclarkb: oh that is very possible04:27
*** salv-orlando has joined #openstack-infra04:27
jogoguh networking is  such a pain04:27
* jogo wonders off04:27
* jogo wanders off*04:28
clarkbwell it would help if we didnt support crazy nat rules like this04:28
clarkbits dnat only too no snat04:28
*** harlowja_at_home has joined #openstack-infra04:29
mikalclarkb: sounds like you're feeling the love04:29
clarkbalso pretty sure it works with multihost because then each nodd is its own metadata server04:30
*** xyang1 has joined #openstack-infra04:31
*** lhcheng has joined #openstack-infra04:32
*** julim has joined #openstack-infra04:32
openstackgerritKhai Do proposed openstack-infra/puppet-gerrit: Make sure Gerrit sites are using the proper libraries.
*** chlong has joined #openstack-infra04:35
*** dannywilson has joined #openstack-infra04:38
*** hdd has quit IRC04:41
*** dannywilson has quit IRC04:42
openstackgerritMerged openstack-infra/jenkins-job-builder: Fix formatting for inclusion of README.rst
openstackgerritMerged openstack-infra/jenkins-job-builder: Add plugins_info to module registry object.
*** salv-orlando has quit IRC04:47
*** achanda has quit IRC04:52
*** miqui has joined #openstack-infra04:53
*** chlong has quit IRC04:54
openstackgerritMerged openstack-infra/jenkins-job-builder: Add support for JDK axis
*** mohitsharma has quit IRC05:01
*** mohitsharma has joined #openstack-infra05:01
*** achanda has joined #openstack-infra05:01
*** chlong has joined #openstack-infra05:03
*** atan8 has quit IRC05:03
*** mohitsharma has quit IRC05:06
openstackgerritKhai Do proposed openstack-infra/jenkins-job-builder: Allow JJB to set connection timeout to jenkins server
*** hdd has joined #openstack-infra05:11
openstackgerritAnusha Ramineni proposed openstack-infra/project-config: Add ACL to create branch/tags
openstackgerritAnusha Ramineni proposed openstack-infra/project-config: Add ACL to create branch/tags
*** craigbr has quit IRC05:20
*** garyh has quit IRC05:20
*** harlowja_at_home has quit IRC05:21
*** otter768 has joined #openstack-infra05:21
*** ivar-lazzaro has joined #openstack-infra05:25
*** otter768 has quit IRC05:25
*** jerryz has quit IRC05:26
openstackgerritAnusha Ramineni proposed openstack-infra/project-config: Add ACL to create branch/tags in proliantutils.
*** jamespage has joined #openstack-infra05:28
*** jyuso has quit IRC05:30
*** naggappan has joined #openstack-infra05:33
*** salv-orlando has joined #openstack-infra05:33
*** salv-orlando has quit IRC05:34
*** yamahata has joined #openstack-infra05:36
jheskethclarkb, mordred: are either of you guys still around per chance?05:37
*** abhirc has quit IRC05:40
*** yfried__ has quit IRC05:41
r1chardj0n3sjhesketh: the previous discussion of bower turned them to drinking a few hours ago ;)05:43
* jhesketh will read backlog then05:44
*** dguerri has quit IRC05:44
*** salv-orlando has joined #openstack-infra05:44
*** lhcheng has quit IRC05:44
r1chardj0n3sstarts around 10:20AM05:45
*** dguerri has joined #openstack-infra05:45
*** mohitsharma has joined #openstack-infra05:48
*** coolsvap_ is now known as coolsvap05:49
*** ivar-laz_ has joined #openstack-infra05:50
*** craigbr has joined #openstack-infra05:51
*** salv-orlando has quit IRC05:51
*** dstufft|laptop has quit IRC05:52
*** ivar-la__ has joined #openstack-infra05:52
*** ivar-lazzaro has quit IRC05:52
*** ivar-laz_ has quit IRC05:55
naggappanhi may i know is this a unit test job for neutron?
naggappanAlso please let me know where this particular job yamal file is presetn. Basically i am looking to run Neutron unit test job each time in our internal CI set up05:59
*** achanda has quit IRC06:00
*** salv-orlando has joined #openstack-infra06:02
*** salv-orlando has quit IRC06:03
*** rmcall has joined #openstack-infra06:06
*** jerryz has joined #openstack-infra06:10
*** hdd has quit IRC06:12
*** dims__ has joined #openstack-infra06:13
*** jyuso has joined #openstack-infra06:15
*** salv-orlando has joined #openstack-infra06:16
*** salv-orlando has quit IRC06:16
*** sputnik13 has joined #openstack-infra06:16
*** imcsk8 has quit IRC06:17
*** dims__ has quit IRC06:18
*** garyh has joined #openstack-infra06:21
*** craigbr has quit IRC06:24
*** rverchikov has joined #openstack-infra06:24
*** achanda has joined #openstack-infra06:24
*** sputnik13 has quit IRC06:25
*** rmcall has quit IRC06:26
*** salv-orlando has joined #openstack-infra06:26
*** mohitsharma has quit IRC06:28
*** heyongli has joined #openstack-infra06:29
*** salv-orlando has quit IRC06:31
*** r1chardj0n3s is now known as r1chardj0n3s_afk06:33
*** vigneshvar has joined #openstack-infra06:35
*** sputnik13 has joined #openstack-infra06:40
*** mohitsharma has joined #openstack-infra06:40
*** achanda has quit IRC06:41
*** denis_makogon has joined #openstack-infra06:42
*** rverchikov has quit IRC06:44
*** vilobhmm has joined #openstack-infra06:44
*** mohitsharma has quit IRC06:44
*** mohitsharma has joined #openstack-infra06:45
*** shashankhegde has joined #openstack-infra06:46
*** MarkAtwood has joined #openstack-infra06:52
*** MarkAtwood has quit IRC06:57
*** markvoelker has quit IRC07:06
*** ildikov has quit IRC07:06
*** mohitsharma has quit IRC07:06
*** wuhg has joined #openstack-infra07:09
*** jlibosva has joined #openstack-infra07:10
*** yfried__ has joined #openstack-infra07:10
*** doude has joined #openstack-infra07:10
*** tnovacik has joined #openstack-infra07:12
openstackgerritMerged openstack-infra/project-config: Stop running the non-voting neutron-full-pg job in the gate
*** yfried__ has quit IRC07:14
*** yfried has joined #openstack-infra07:14
*** yfried is now known as yfried|afk07:15
*** mohitsharma has joined #openstack-infra07:16
*** mohitsharma has quit IRC07:18
*** yfried|afk is now known as yfried07:21
*** otter768 has joined #openstack-infra07:22
openstackgerritMarton Kiss proposed openstack-infra/system-config: Move Drupal settings to conf variable
*** harlowja_away has quit IRC07:23
*** shashankhegde has quit IRC07:26
*** otter768 has quit IRC07:27
*** scheuran has joined #openstack-infra07:27
*** sdake_ has quit IRC07:29
*** SumitNaiksatam has quit IRC07:29
*** woodster_ has quit IRC07:36
*** che-arne has quit IRC07:37
*** salv-orlando has joined #openstack-infra07:38
*** jcoufal has joined #openstack-infra07:38
*** ildikov has joined #openstack-infra07:38
*** chlong has quit IRC07:39
*** achanda has joined #openstack-infra07:41
*** achanda has quit IRC07:47
*** e0ne has joined #openstack-infra07:51
*** salv-orlando has quit IRC07:51
*** achanda has joined #openstack-infra07:54
*** SumitNaiksatam has joined #openstack-infra07:54
*** afazekas has joined #openstack-infra07:55
*** luqas has joined #openstack-infra07:57
*** e0ne has quit IRC07:58
*** ivar-lazzaro has joined #openstack-infra07:59
*** ivar-laz_ has joined #openstack-infra08:02
*** belmoreira has joined #openstack-infra08:02
*** ivar-la__ has quit IRC08:02
*** ivar-lazzaro has quit IRC08:04
*** mrmartin has joined #openstack-infra08:05
*** vilobhmm has quit IRC08:07
*** teran has quit IRC08:10
*** jgjl has joined #openstack-infra08:12
*** stevemar has quit IRC08:12
*** Ala has joined #openstack-infra08:12
*** Guest82498 has joined #openstack-infra08:19
*** Guest82498 has quit IRC08:19
*** vigneshvar has quit IRC08:20
*** ivar-laz_ has quit IRC08:20
*** teran has joined #openstack-infra08:26
openstackgerritPavel Sedlák proposed openstack-infra/jenkins-job-builder: Fix NullPtr issue in ownership props
*** Hal has joined #openstack-infra08:27
*** Hal is now known as Guest7504208:27
*** zz_avozza is now known as avozza08:28
*** thingee has joined #openstack-infra08:29
*** markus_z has joined #openstack-infra08:29
*** achanda has quit IRC08:29
*** teran has quit IRC08:30
*** vigneshvar has joined #openstack-infra08:31
*** jgallard_ has joined #openstack-infra08:32
*** salv-orl_ has joined #openstack-infra08:32
*** achanda has joined #openstack-infra08:34
*** yamamoto_ has quit IRC08:35
*** yamahata has quit IRC08:36
*** mpaolino has joined #openstack-infra08:39
*** MaxV has joined #openstack-infra08:44
*** otherwiseguy has quit IRC08:49
*** achanda has quit IRC08:53
*** armax has quit IRC08:54
*** jpich has joined #openstack-infra08:55
*** achanda has joined #openstack-infra08:55
*** denis_makogon has quit IRC08:59
*** _shaps__ has joined #openstack-infra09:00
*** jgjl has quit IRC09:03
*** Longgeek has quit IRC09:05
*** dimtruck is now known as zz_dimtruck09:05
*** jistr has joined #openstack-infra09:07
*** _shaps__ has quit IRC09:08
*** teran has joined #openstack-infra09:09
*** achanda has quit IRC09:09
*** _shaps__ has joined #openstack-infra09:11
*** achanda has joined #openstack-infra09:11
*** _shaps__ has quit IRC09:15
*** otherwiseguy has joined #openstack-infra09:15
*** achanda has quit IRC09:19
*** salv-orl_ has quit IRC09:19
*** dizquierdo has joined #openstack-infra09:19
*** andreykurilin_ has joined #openstack-infra09:21
*** dtantsur|afk is now known as dtantsur09:22
*** otter768 has joined #openstack-infra09:23
*** derekh has joined #openstack-infra09:24
*** zz_dimtruck is now known as dimtruck09:26
*** otter768 has quit IRC09:27
*** mpaolino has quit IRC09:29
*** mpaolino has joined #openstack-infra09:30
*** dimtruck is now known as zz_dimtruck09:38
*** marun has joined #openstack-infra09:41
*** salv-orlando has joined #openstack-infra09:41
*** ihrachyshka has joined #openstack-infra09:43
*** marun has quit IRC09:45
*** zz_johnthetubagu is now known as johnthetubaguy09:54
openstackgerritMateusz Matuszkowiak proposed openstack-infra/project-config: Add Fuel Contrail Plugin to Stackforge
*** hichihara has quit IRC09:58
*** dtantsur is now known as dtantsur|bbl10:03
openstackgerrityolanda.robla proposed openstack-infra/storyboard-webclient: Fix to show errors again
*** marcusvrn has joined #openstack-infra10:04
openstackgerritMateusz Matuszkowiak proposed openstack-infra/project-config: Add Fuel Contrail Plugin to Stackforge
*** jp_at_hp has joined #openstack-infra10:15
*** cdent has joined #openstack-infra10:19
*** skolekonov has joined #openstack-infra10:19
openstackgerritAdam Gandelman proposed openstack-infra/shade: Adds some more swift operations
*** _shaps__ has joined #openstack-infra10:24
*** coolsvap is now known as coolsvap_10:29
*** BharatK has quit IRC10:31
*** zhiwei has quit IRC10:33
*** sputnik13 has quit IRC10:33
*** sputnik13 has joined #openstack-infra10:34
openstackgerritIhar Hrachyshka proposed openstack-infra/release-tools: Be more informative when putting -2 for stable freeze
*** salv-orlando has quit IRC10:35
*** pblaho has joined #openstack-infra10:36
*** ghostpl has joined #openstack-infra10:39
*** ldnunes has joined #openstack-infra10:42
*** sputnik13 has quit IRC10:46
*** chlong has joined #openstack-infra10:50
*** mase_x200 has joined #openstack-infra10:51
*** vdo has joined #openstack-infra10:52
*** boris-42 has quit IRC10:52
*** swamireddy has joined #openstack-infra10:52
*** ab1892 has joined #openstack-infra10:53
swamireddyHi Core reviewers - can you please look in the patch
ab1892in my ci setup i want to run only volume test, can anyone tell me how to do that using the devstack-gate script only??10:54
*** alexpilotti has joined #openstack-infra10:56
*** rfolco has joined #openstack-infra10:58
*** pelix has joined #openstack-infra11:02
*** yfried is now known as yfried|afk11:02
*** dtantsur|bbl is now known as dtantsur11:03
*** yfried|afk is now known as yfried11:03
*** pcrews has quit IRC11:08
*** pcrews has joined #openstack-infra11:11
ab1892hi all, in my ci setup i want to run only volume test, can anyone tell me how to do that using the devstack-gate script only??11:12
*** andreykurilin_ has quit IRC11:12
*** andreaf has joined #openstack-infra11:12
*** e0ne has joined #openstack-infra11:12
*** denis_makogon has joined #openstack-infra11:14
*** rlandy has joined #openstack-infra11:16
*** ghostpl has quit IRC11:16
*** pc_m has joined #openstack-infra11:19
*** yfried is now known as yfried|afk11:19
*** yfried|afk is now known as yfried11:23
*** mpaolino has quit IRC11:25
*** _shaps__ has quit IRC11:35
*** salv-orlando has joined #openstack-infra11:35
*** ociuhandu has joined #openstack-infra11:36
*** yamamoto has joined #openstack-infra11:36
*** e0ne is now known as e0ne_11:37
*** salv-orlando has quit IRC11:38
*** mpaolino has joined #openstack-infra11:39
*** markvoelker has joined #openstack-infra11:40
*** adalbas has joined #openstack-infra11:42
openstackgerritAdam Gandelman proposed openstack-infra/devstack-gate: Bump ironic build/callback timeouts
*** coolsvap_ is now known as coolsvap11:46
*** markvoelker has quit IRC11:47
*** dims__ has joined #openstack-infra11:47
swamireddymarkmcclain: is it possible check this patch -
*** e0ne_ is now known as e0ne11:55
*** e0ne has quit IRC12:00
*** dims__ has quit IRC12:00
*** coolsvap is now known as coolsvap_12:01
*** EmilienM|afk is now known as EmilienM12:02
*** dims__ has joined #openstack-infra12:03
*** nfedotov has joined #openstack-infra12:05
*** reed has joined #openstack-infra12:06
*** salv-orlando has joined #openstack-infra12:07
*** wuhg has quit IRC12:09
*** e0ne has joined #openstack-infra12:09
*** jlibosva has quit IRC12:10
*** salv-orlando has quit IRC12:11
*** jlibosva has joined #openstack-infra12:12
*** yfried is now known as yfried|afk12:12
openstackgerritIhar Hrachyshka proposed openstack-infra/release-tools: Be more informative when putting -2 for stable freeze
*** ghostpl has joined #openstack-infra12:17
*** ab1892 has quit IRC12:20
*** ghostpl has quit IRC12:22
*** hashar has joined #openstack-infra12:24
*** xyang1 has quit IRC12:25
*** yfried|afk is now known as yfried12:25
*** aysyd has joined #openstack-infra12:25
*** salv-orlando has joined #openstack-infra12:28
*** weshay has joined #openstack-infra12:33
*** salv-orlando has quit IRC12:33
*** mwagner_lap has quit IRC12:35
*** chlong has quit IRC12:38
*** salv-orlando has joined #openstack-infra12:39
*** oomichi has quit IRC12:39
*** oomichi has joined #openstack-infra12:42
*** markvoelker has joined #openstack-infra12:43
*** salv-orlando has quit IRC12:43
*** oomichi has quit IRC12:44
*** hashar is now known as hasharAway12:47
*** markvoelker has quit IRC12:47
*** marcusvrn1 has joined #openstack-infra12:47
*** marcusvrn1 has quit IRC12:48
*** marcusvrn1 has joined #openstack-infra12:48
*** woodster_ has joined #openstack-infra12:49
rcarrillocruzhey folks, i get errors when installing reviewstats from latest ...12:51
rcarrillocruzanyone else?12:51
*** marcusvrn has quit IRC12:51
rcarrillocruzbtw, if reviewstats has its own channel, pls let me know...12:51
mordredrcarrillocruz: nope. this is the closest we've got12:54
*** mbacchi has joined #openstack-infra12:55
rcarrillocruzk, thx mordred12:55
*** yfried_ has joined #openstack-infra13:00
openstackgerritNikita Konovalov proposed openstack-infra/python-storyboardclient: Improved exception handling
*** markvoelker has joined #openstack-infra13:01
*** mase_x200 has quit IRC13:01
*** yfried has quit IRC13:03
*** oomichi has joined #openstack-infra13:04
*** yfried__ has joined #openstack-infra13:05
*** ghostpl has joined #openstack-infra13:05
*** amitgandhinz has joined #openstack-infra13:06
*** afazekas has quit IRC13:06
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Allow to wait for jenkins to start up
*** yamamoto has quit IRC13:07
*** oomichi has quit IRC13:08
*** yfried_ has quit IRC13:08
*** kgiusti has joined #openstack-infra13:13
*** mpaolino has quit IRC13:16
*** enikanorov_ has quit IRC13:16
*** dkliban_afk is now known as dkliban13:16
*** mpaolino has joined #openstack-infra13:17
*** jerryz has quit IRC13:18
*** dprince has joined #openstack-infra13:18
mordredclarkb: btw - related to our conversation about generating puppet from yesterday - all of that code is actually in commit 113:19
mordredin system-config13:19
*** bswartz has quit IRC13:20
sdaguemordred / fungi - could I get you folks to push this through - ?13:20
nibalizermordred: what, generting puppet what?13:21
*** baoli has joined #openstack-infra13:22
pc_mHi! Can I get a second core to review for me?13:22
*** amitgandhinz has quit IRC13:23
*** amitgandhinz has joined #openstack-infra13:24
*** mwagner_lap has joined #openstack-infra13:27
openstackgerritDenis M. proposed openstack-infra/project-config: Enable inner requirements parsing during requirements check job
*** zhiwei has joined #openstack-infra13:29
*** yfried__ has quit IRC13:30
openstackgerritAleksey Ripinen proposed openstack-infra/storyboard: Added new way to filter stories by tags
*** jerryz has joined #openstack-infra13:31
*** mjturek1 has joined #openstack-infra13:32
*** amitgandhinz has quit IRC13:34
openstackgerritDenis M. proposed openstack-infra/project-config: Enable inner requirements parsing during requirements check job
*** ihrachyshka has quit IRC13:35
*** jerryz has quit IRC13:36
*** abhirc has joined #openstack-infra13:36
*** _nadya_ has joined #openstack-infra13:41
*** abhirc has quit IRC13:41
openstackgerritRadomir Dopieralski proposed openstack-infra/project-config: Automatically trigger building of docs for stackforge/tuskar-ui-extras
mordrednibalizer: enjoy this in your morning:
mordrednibalizer: it's commit 1 from openstack-infra ... and it's a script to look at a group on launchpad and use that as a the basis of generating a puppet manifest that creates local users with everyone's ssh keys as imported from launchpad13:43
nibalizerno i get it13:43
mordrednibalizer: :)13:43
nibalizerwe dont still have this?13:43
*** ghostpl has quit IRC13:43
*** vigneshvar has quit IRC13:43
mordredI dunno man - we got lazy :)13:44
nibalizerwell its super trivial to feed modern puppet json or yaml or whatever13:45
mordredit was probably super trivial to do it then too13:45
mordredbut I still hadn't gotten to grokking puppet classes at that point ... so let's not kid ourselves about the basis for decision making13:45
nibalizerok well you get points for being hillarious13:46
nibalizerthis is totes going in my next talk13:46
openstackgerritAdam Gandelman proposed openstack-infra/shade: Fix broken object hashing
openstackgerritAdam Gandelman proposed openstack-infra/shade: Adds some more swift operations
koolhead17hello world13:47
openstackgerritSpencer Krum proposed openstack-infra/puppet-httpd: Rename apache2 to httpd globally
*** dims__ has quit IRC13:49
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Allow to wait for jenkins to start up
reedmrmartin, what's up with the solr patch? isn't there a puppet module that works?13:52
openstackgerritDenis M. proposed openstack/requirements: Update global requirements to reflect requirements of AMQP1 driver
mrmartinthe vamsee-solr is broken13:53
*** jedimike has joined #openstack-infra13:53
mrmartinI was talking with a guy, who wrote a patch for that, but finally forked the repo, but he is not sure how long will he support that13:53
*** afazekas has joined #openstack-infra13:55
mrmartinso we have several options there: a, work with the fork b, fork it to ourselfs (not much difference, we need to maintain the module), or use the one I wrote13:55
*** _shaps__ has joined #openstack-infra13:56
reedare there other modules we can use?13:57
reedlike this one
* anteaya is home again and online in time for, you guessed it, a meeting!13:58
mrmartinI've checked those modules during the implementation, and for example this don't have a solr core support. (in solr terminology they call separate indexes to cores)13:58
openstackgerritAleksey Ripinen proposed openstack-infra/storyboard: Added new paginate query
openstackgerritDenis M. proposed openstack-infra/project-config: Adding required jobs for supported datastores
mrmartinand we need three different cores deployed, due to chinese translation13:59
swamireddyHi Core reviewers - can you please look in the patch
mrmartinand of course custom configuration for chinese core, because it is not trivial to setup, they are using different "letters", so it requires a custom analyzer13:59
openstackgerritAleksey Ripinen proposed openstack-infra/storyboard: Added new way to filter stories by tags
swamireddyThere is ceph ceilometers integration patch is on hold due to this review pending14:00
reedmrmartin, sounds like you have enough details then14:00
*** amotoki has quit IRC14:00
mrmartinI've really checked it when wrote the askbot modules.14:01
*** vdo has quit IRC14:01
*** dims_ has joined #openstack-infra14:02
reedgot it, mrmartin ... maybe next time it would be good to add such details to the commit message, to avoid being questioned by reviewers14:02
*** dims___ has joined #openstack-infra14:02
*** jlibosva has quit IRC14:02
*** abhirc has joined #openstack-infra14:03
*** salv-orlando has joined #openstack-infra14:04
*** dims____ has joined #openstack-infra14:04
*** scheuran has quit IRC14:04
*** dims_ has quit IRC14:06
*** dims__ has joined #openstack-infra14:06
*** dims__ has quit IRC14:06
*** dims___ has quit IRC14:06
*** gema has left #openstack-infra14:07
*** andymaier has joined #openstack-infra14:07
*** dims__ has joined #openstack-infra14:07
*** salv-orlando has quit IRC14:08
*** dims__ has quit IRC14:08
*** dims____ has quit IRC14:09
*** alazarev has quit IRC14:09
*** tiswanso has joined #openstack-infra14:10
*** therve has quit IRC14:10
*** therve has joined #openstack-infra14:10
*** yolanda has quit IRC14:11
mordrednibalizer: in terms of talks - you've seen my slide on what it takes to run puppet agent successfully, right?14:11
*** alazarev has joined #openstack-infra14:11
*** yolanda has joined #openstack-infra14:11
*** dims_ has joined #openstack-infra14:11
mordrednibalizer: because if you could explain to people that a return code of 2 for a successful run is just really not ok, that would be greaaaat14:12
*** e0ne is now known as e0ne_14:14
*** omrim has joined #openstack-infra14:16
omrimIs someone encounter an access problem to:ubuntu webrepoo in his ci?14:17
omrim16:15:12 2015-02-05 14:15:11.964 | W: Failed to fetch http://webrepo/ubuntu-mirror/dists/trusty-backports/universe/binary-amd64/Packages  500  Internal Server ErrorFor example:14:17
*** amitgandhinz has joined #openstack-infra14:17
*** dustins has joined #openstack-infra14:17
openstackgerritAnne Gentle proposed openstack/requirements: Adds Sphinx theme, openstackdocstheme, for
adam_ganyone able to push a devstack-gate quickie through?
nibalizermordred: lol yea14:18
nibalizerwe should be greatful we get a return code14:19
nibalizerfor a while we were just gonna get a yaml blob on stdout with a report of what happened14:19
mordrednibalizer: we should be sholdn't we?14:19
*** ghostpl has joined #openstack-infra14:19
mordrednibalizer: well, a return code of 0 or 1 with a yaml blob would be nicer, tbh14:19
mordrednibalizer: but I hear you14:19
*** doug-fish has joined #openstack-infra14:20
nibalizermordred: so uh, in my 10 seconds of testing just now14:21
nibalizerit looks like if you /dont/ specify --detailed-exitcodes then it works right?14:22
*** bswartz has joined #openstack-infra14:23
*** joesavak has joined #openstack-infra14:23
*** e0ne_ has quit IRC14:24
*** hdd has joined #openstack-infra14:24
*** boris-42 has joined #openstack-infra14:25
*** mriedem has joined #openstack-infra14:25
*** mtanino has joined #openstack-infra14:26
*** e0ne has joined #openstack-infra14:26
*** salv-orlando has joined #openstack-infra14:26
*** gordc has joined #openstack-infra14:26
*** zhiwei has quit IRC14:29
*** jcoufal_ has joined #openstack-infra14:30
funginibalizer: not reallly14:32
*** achuprin_ has quit IRC14:32
funginibalizer: if you don't specify --detailed-exitcodes you get a return code of 0 for lots of things which any sane person would consider a failure (anyone outside of the crowd who feels "eventually consistent" is okay for config management)14:33
*** jcoufal has quit IRC14:33
fungi(where eventually might be "never" or perhaps just "until an admin notices stuff is terribly broken")14:34
openstackgerritMerged openstack-infra/project-config: encapsulate updown testing in devstack tree
*** ZZelle has quit IRC14:34
*** ZZelle has joined #openstack-infra14:35
*** enikanorov_ has joined #openstack-infra14:35
*** dangers_away is now known as dangers14:35
*** abhirc_ has joined #openstack-infra14:38
zuli just tagged nova 2014.2.2 for stable/juno should the tarball be created now or is backed up because of kilo-214:38
fungizul: might be a slight delay... looking now14:38
*** melwitt has quit IRC14:39
zuloh its still building nova-docks looks like it14:39
fungizul: says it finished already14:39
*** _shaps__ has quit IRC14:39
fungizul: and i see it at
zulfungi:  ok im just slow this morning apparently14:40
*** abhirc has quit IRC14:40
funginova-2014.2.2-py2.py3-none-any.whl and nova-2014.2.2.tar.gz are showing there with today's date14:40
*** mtanino has quit IRC14:40
zulfungi: coolio thanks14:42
*** signed8bit has joined #openstack-infra14:42
*** signed8b_ has joined #openstack-infra14:44
*** ddieterly has quit IRC14:46
*** zz_gondoi is now known as gondoi14:47
*** laurelm has joined #openstack-infra14:47
*** signed8bit has quit IRC14:47
*** kashyap has joined #openstack-infra14:48
*** achuprin_ has joined #openstack-infra14:48
*** prad has joined #openstack-infra14:50
*** jlibosva has joined #openstack-infra14:52
*** hasharAway has quit IRC14:53
*** xyang1 has joined #openstack-infra14:56
bhunter71good morning, infra. I have a Windows user for whom the SCP of the commit-hooks is failing when they git-review.   I've checked "everything", but at am a loss for how SCP is failing [in git bash on windows].   Any ideas?    see for more details.  Thanks!14:58
*** _shaps__ has joined #openstack-infra14:58
*** dizquierdo has quit IRC14:58
*** bswartz1 has joined #openstack-infra14:59
*** carl_baldwin has joined #openstack-infra14:59
*** prad has quit IRC14:59
*** bswartz has quit IRC14:59
*** bswartz1 is now known as bswartz15:00
*** prad has joined #openstack-infra15:00
*** sigmavirus24_awa is now known as sigmavirus2415:00
*** hdd has quit IRC15:00
sdagueso... citrix ci15:01
sdagueanyone know what's up with that?15:01
BobBallTrying to set up a 3rd party CI starting Zuul I get "ImportError: Could not import any of: daemon.pidlockfile, daemon.pidfile"... python-daemon==2.0.5 which should include daemon.pidfile as far as I can tell15:01
BobBallsure sdague15:01
BobBallWhat's the question?15:01
sdagueit's all the failing15:02
*** armax has joined #openstack-infra15:02
sdaguean example15:02
BobBallFailure rate isn't good ATM I'll admit... 'Passed': 214, 'Failed': 38 over 24 hours15:02
*** dalgaaf has joined #openstack-infra15:02
anteayaBobBall: care to shut it down?15:03
fungibhunter71: are you certain the private key in ~user/.ssh/id_rsa corresponds to the public key that user added at
bhunter71fungi, yeah, pretty certain.   Double checked it, and can also ssh directly to gerrit with it.15:03
*** mfink_ has joined #openstack-infra15:03
fungibhunter71: and the username showing in the scp command details matches the username for that user as displayed at
BobBallI would very much prefer not to anteaya - most of the tests are giving correct results - passes or fails - but the fails failure rate is indeed higher than I would ideally like.15:04
*** jlibosva has quit IRC15:04
anteayaBobBall: give me a sense of what to tell folks should they ask like sdague did15:04
BobBallsdague: The TestShelveInstance failure there is a known race that we're trying to investigate15:04
anteayakeep in mind someone else from infra might shut it down15:05
anteayacan you investigate without the ci voting?15:05
anteayathat might be a good start15:05
fungibhunter71: "Permission denied (publickey)." generally means that either the key or the username didn't match. is it possible scp is finding some different key than ssh? that seems unlikely, but...15:05
BobBallit's affecting around 5% of jobs15:05
anteayaBobBall: I think we are not agreeing to the same point15:05
*** craigbr has joined #openstack-infra15:05
*** marun has joined #openstack-infra15:05
fungibhunter71: if you can get me the ip address they're coming from, i can double-check the gerrit ssh api logs for some clue15:06
anteayaa ci that is voiting with a known failure is a candidate for being disabled15:06
BobBall(12 jobs failed with the Shelve Instance failure compared with 200 passes in the same period)15:06
*** rmcall has joined #openstack-infra15:06
anteayaBobBall: are you willing to diable voting?15:06
BobBallanteaya: No CI system, even the official one, has 100% pass rate15:06
anteayaBobBall: agreed15:06
anteayahowever in the weighing of question from dev and ci system the ci system looses15:07
BobBallSo this race that affects 5% of jobs shouldn't mean it's disabled15:07
bhunter71thanks fungi, I'll get back to you.   Just knowing that you don't recognize this, helps.15:07
anteayacan you see this from my point of view?15:07
anteayait is easier to just disable and move on15:08
anteayaI'm asking if you are willing to disable voting, you can still comment15:08
BobBallOf course - and we're trying to fix it - but non-voting is useless.  People don't always respect the -1 vote as it is (e.g. the devstack merge a few days ago that broke the Citrix CI)15:08
openstackgerritThanh Ha proposed openstack-infra/jenkins-job-builder: Add support for new patchset-created exclude fields
anteayaBobBall: well you are welcome to spin the roulette wheel, another query about the ci, or should someone else from infra choose and your system may be disabled15:09
anteayaBobBall: and yes, I hear you, it is because folks get so much noise from ci systems that they don't pay any attention to them15:09
anteayait is an unfortnate manifestation of having too many15:10
BobBallQueries about the CI I can cope with - please redirect them to me.  I will happily disable it if there is a serious issue (and indeed I have constant monitoring of the pass rate and did disable it when the devstack merge broke it before anyone mentioned it to me)15:10
anteayathen devs don't pay attention to them15:10
anteayaBobBall: well stay close, and put an alert on citrix15:10
fungianteaya: BobBall: it's completely within the purview of the members of nova-release (in this case mikal) to remove citrix ci from the nova-ci group if it is deemed problematic15:11
fungithough he is just one person, and so not awake 24 hours a day15:11
BobBallOf course.  What I want to avoid is a question "What's going on with the Citrix CI, it failed XYZ" to mean it gets blanket disabled.  In sdague's case it was a 5% failure but the vast majority of changes are passing successfully.15:12
BobBallI'll update the status page to point out this one known race that we're investigating15:13
BobBallI haven't done that yet.15:13
anteayaBobBall: if you were one of five ci systems we could agree to an exception for your system15:13
anteayathat time has long passed15:13
BobBallI don't want an exception :)15:13
anteayaeven if you could convince me to behave in a certain way15:13
anteayaI can't speak for all folks with the ability to affect permissions on a system15:13
anteayawell what you are asking for is an exception15:14
anteayasince our current workflow is to disabled errant systems15:14
*** _shaps__ has quit IRC15:14
anteayawe have over 10015:14
BobBall5% failure rate shouldn't be considered errant15:14
anteayait is for sdague15:14
clarkbBobBall does your CI run on multiple node test envs? shelve is broken on >1 node15:14
BobBallunfortunately not clarkb - single node, but it's a race I think in the xenserver plugins15:15
*** dustins has quit IRC15:15
clarkbgotcha just checking as I am aware of a multinode bug but its fail rate is closer to 100%15:15
BobBallit all passes fine in serial - I suspect it's when multiple jobs are causing garbage collection on the xenserver host15:16
BobBalland one job times out because it waits too long15:16
sdagueBobBall: yeh, I may have just been noticing it more recently (like it happened to hit a bunch of my patches)15:16
BobBallbut trying to track it down is being painful :/15:16
*** ddieterly has joined #openstack-infra15:16
*** e0ne is now known as e0ne_15:17
mrmartinpleia2: I've deployed your zanata into vagrant, and have some questions if you are available.15:17
BobBallI will, of course, raise the priority of trying to get this addressed internally15:17
*** rmcall has quit IRC15:17
*** e0ne_ is now known as e0ne15:17
*** rm_work|away is now known as rm_work15:18
fungiin my mind, the whole reason we implemented project-specific ci groups in gerrit is to let representatives of those projects self-organize and control voting permissions for third parties which they do or don't find a benefit to the review process, so that the infra team can stop having to care about it15:19
BobBallWould it be helpful if the CI systems were asked to publish failure rates and/or known racy jobs?  I could easily produce statistics showing the current failure rate of individual jobs.  Perhaps I could even update the message to say "This job is failing with XYZ rate"15:19
*** dkranz has joined #openstack-infra15:20
* BobBall likes that idea actually... I think I'll update the message. I can't triage it automatically like elastic recheck but just having a job-level failure could be useful at this point.15:20
anteayafungi: true15:20
anteayawas just hoping BobBall would change things on his side since it is faster to change rather than leting folks change things on the gerrit side15:21
*** swamireddy has quit IRC15:21
anteayawhich is much slower to change15:21
anteayaBobBall: would it be helpful to ask ci systems to publish failure rates? well I am not asking ci systems to do anything more than we are already asking since many of them can't figure out what we currently require15:22
fungisure, but ultimately that's between the citrix ci operators and the nova reviewers, so as long as they're on the same page i'm not particularly worried15:22
*** zhiwei has joined #openstack-infra15:22
anteayafungi: fair enough, was hoping to save BobBall some time15:22
fungiunderstood, and certainly worth raising the point15:22
*** timcline has joined #openstack-infra15:22
*** david-lyle_afk is now known as david-lyle15:23
fungiBobBall: as for your original zuul question, did just merge on monday15:24
*** isviridov is now known as isviridov_away15:25
BobBallfungi: unfortunately I've got 2.0.5 installed15:25
fungiBobBall: so that may have prompted your system to update to a newer python-daemon if yuo're deploying zuul automatically from the master branch15:25
BobBalloh, yes I am15:25
*** tonytan4ever has joined #openstack-infra15:25
fungiBobBall: is is possible you also have a system package of python-daemon installed and import is finding the old one?15:25
*** dizquierdo has joined #openstack-infra15:26
clarkbyou have to update lockfile too15:26
*** craigbr has quit IRC15:26
clarkbI think zuul needs a new lower bound on it too15:27
fungiyeah, in production we have python-daemon==2.0.5 and lockfile==0.9.1 right now and if i can import daemon.pidfile but not daemon.pidlockfile15:28
fungier, s/if //15:28
fungiBobBall: so see if lockfile==0.9.1 helps15:29
BobBallLockfile was 0.815:29
BobBallLet me update that and remove the python-daemon.  A few other things to try too.  Thanks.15:29
*** dkranz has quit IRC15:30
fungiclarkb: i wonder whether will dtrt there?15:31
clarkbuhm it is used15:32
clarkbso at the very least the commit message is wrong. previously python-daemon didnt depend on it properly15:32
*** wenlock has joined #openstack-infra15:33
*** dkranz has joined #openstack-infra15:33
BobBallthanks fungi - upgrading lockfile was the issue15:33
fungiclarkb: python-daemon declares an install_requires on lockfile >= 0.1015:33
openstackgerritMatt Riedemann proposed openstack-infra/elastic-recheck: Add query for neutron bug 1321207
*** jcoufal_ has quit IRC15:33
fungithe bzr repo for it is at though you probably need an alioth account to browse it15:34
*** jcoufal has joined #openstack-infra15:34
*** nfedotov has quit IRC15:34
fungidsa seriously needs to get anon access fixed for alioth15:34
jeblairit is browsable anonymously15:34
clarkboh I know. zuul manipulates lockfile directly15:34
fungiokay, so it was just the bug tracker i had to have an account to get to15:35
clarkbthat is why we depend on it in zuul15:35
fungiclarkb: in that case maybe we just make sure that python-daemon is listed before lockfile since it has a declared minimum15:35
fungiotherwise epip15:35
fungii'll update 148252 to that end15:36
fungijd__: ^15:36
jeblairclarkb, fungi: is what jd__ is saying perhaps that it is not needed because we can rely on python-daemon now?15:36
jeblairand the transitive dependency15:36
jeblairsince we now require daemon >2 or whatever15:37
clarkbmaybe but we import lockfile in zuul it should be a dep of zuul15:37
fungiwe shouldn't rely on a transitive dependency if we call lockfile directly. though if we only call it via daemon then it's probably fine to drop it from the reqs list15:37
jeblairclarkb: where do we import it?15:37
*** gondoi is now known as zz_gondoi15:37
*** mtanino has joined #openstack-infra15:37
fungii only see us importing daemon.pidlockfile15:38
fungiand daemon.pidfile15:38
fungibut no lockfile15:38
*** tmcpeak has joined #openstack-infra15:38
fungiso i think jd__'s 148252 should be fine as-is15:38
clarkbah I see15:38
fungii've +2'd but will let one of you approve if you agree15:39
*** rmcall has joined #openstack-infra15:39
clarkbya removing is fine if its imported via daemon15:39
jeblairjhesketh has +2d the depends-on stack if anyone else wants to review/approve; ends at 14455615:40
*** dustins has joined #openstack-infra15:41
fungithanks, revisiting that now15:42
*** sdake_ has joined #openstack-infra15:42
*** dkranz has quit IRC15:43
*** jaypipes has joined #openstack-infra15:43
*** kashyap_bbiab has joined #openstack-infra15:43
*** rmcall has quit IRC15:44
*** zhiwei has quit IRC15:44
*** dkranz has joined #openstack-infra15:44
*** hdd has joined #openstack-infra15:45
openstackgerritMatt Riedemann proposed openstack-infra/elastic-recheck: Add query for tempest bug 1415143
*** jcoufal has quit IRC15:45
*** kashyap has quit IRC15:45
*** jcoufal has joined #openstack-infra15:46
openstackgerritMerged openstack-infra/zuul: Make needs_change plural
openstackgerritMerged openstack-infra/zuul: Add a test for enqueuing complex dependencies
*** jcoufal_ has joined #openstack-infra15:47
*** yamahata has joined #openstack-infra15:47
*** zz_dimtruck is now known as dimtruck15:49
*** stevemar has joined #openstack-infra15:50
*** jcoufal has quit IRC15:50
*** signed8b_ is now known as signed8bit_ZZZzz15:51
*** signed8bit_ZZZzz is now known as signed8b_15:53
*** achanda has joined #openstack-infra15:53
*** kashyap_bbiab is now known as kashyap15:54
*** rm_work is now known as rm_work|away15:54
*** rverchikov has joined #openstack-infra15:57
openstackgerritMateusz Matuszkowiak proposed openstack-infra/project-config: Add Fuel Contrail Plugin to Stackforge
*** zz_gondoi is now known as gondoi15:57
*** pblaho has quit IRC15:58
clarkbjeblair: so you just list multiple depends on if they exist?15:58
jeblairclarkb: that's the idea15:58
clarkband that creates an order I assume? /me is reviewing the crd changes15:58
openstackgerritMerged openstack-infra/elastic-recheck: Add query for neutron bug 1321207
clarkbbut still in commit message and test stuff15:59
jeblairclarkb: yeah, i think the order is git-depends then commit-depends in order15:59
*** kgiusti has quit IRC16:00
*** craigbr has joined #openstack-infra16:00
openstackgerritMateusz Matuszkowiak proposed openstack-infra/project-config: Add Fuel Contrail Plugin to Stackforge
*** craigbr has quit IRC16:00
pc_mclarkb: ping16:01
*** skolekonov has quit IRC16:01
*** pblaho has joined #openstack-infra16:01
*** swamireddy has joined #openstack-infra16:02
*** baoli has quit IRC16:02
silehtfungi, hi, perhaps you have a bit of bandwith for ?16:02
*** baoli has joined #openstack-infra16:03
openstackgerritDavanum Srinivas (dims) proposed openstack/requirements: WIP: trying to see if we can bump pycadf
pelixzaro: wondering if I should go ahead and approve as we seem to be lacking enough JJB core reviewers at the moment when it comes to code from either of us :p16:03
openstackgerritDenis M. proposed openstack-infra/project-config: Enable inner requirements parsing during requirements check job
*** erikmwil_ has joined #openstack-infra16:04
*** erikmwilson is now known as Guest7083716:04
*** erikmwil_ is now known as erikmwilson16:04
openstackgerritDavanum Srinivas (dims) proposed openstack/requirements: WIP: trying to see if we can bump pycadf
*** erikmwil- has joined #openstack-infra16:04
clarkbpc_m pong16:04
pc_mclarkb: Can you review for me please? Would like this running in check queue.16:04
*** achanda has quit IRC16:05
*** CIbuilds has quit IRC16:05
*** pblaho has quit IRC16:06
*** kashyap_bbiab has joined #openstack-infra16:06
*** avozza is now known as zz_avozza16:07
*** achuprin_ has quit IRC16:09
openstackgerritMatt Riedemann proposed openstack-infra/elastic-recheck: Add query for tempest bug 1415143
*** kashyap has quit IRC16:10
*** ghostpl has quit IRC16:10
*** doobi-sham-95717 has joined #openstack-infra16:11
*** vishy has quit IRC16:11
*** sdake has quit IRC16:11
*** dustins has quit IRC16:12
*** sdake has joined #openstack-infra16:12
*** sdake has quit IRC16:12
*** sdake has joined #openstack-infra16:12
nibalizerfungi: ah yea I didn't test it well16:12
*** dustins has joined #openstack-infra16:12
*** vishy has joined #openstack-infra16:12
funginibalizer: i've lost context... to what were you replying there?16:13
*** kashyap_bbiab is now known as kashyap16:13
openstackgerritMerged openstack-infra/project-config: oslo.messaging: uses devstack-f21 instead f20
openstackgerritMerged openstack-infra/project-config: oslo.messaging: add zeromq functional job
openstackgerritMerged openstack-infra/project-config: oslo.messaging: enable functional jobs
*** jcoufal_ has quit IRC16:17
*** mjturek1 has quit IRC16:17
*** jcoufal has joined #openstack-infra16:17
*** kashyap has quit IRC16:18
*** naggappan has quit IRC16:18
*** thedodd has joined #openstack-infra16:19
*** dtantsur is now known as dtantsur|afk16:19
*** rm_work|away is now known as rm_work16:22
*** dkranz has quit IRC16:23
openstackgerritMerged openstack-infra/zuul: Support cross-repo-dependencies in dependent pipelines
*** dkranz has joined #openstack-infra16:23
*** kashyap_bbiab has joined #openstack-infra16:23
*** kashyap_bbiab is now known as kashyap16:23
*** ildikov has quit IRC16:24
*** achuprin_ has joined #openstack-infra16:25
*** imcsk8 has joined #openstack-infra16:26
*** imcsk8 has quit IRC16:27
*** jlibosva has joined #openstack-infra16:27
*** imcsk8 has joined #openstack-infra16:27
*** rwsu-afk is now known as rwsu16:28
*** dizquierdo has quit IRC16:29
*** dannywilson has joined #openstack-infra16:29
*** ihrachyshka has joined #openstack-infra16:30
clarkbjeblair commented on 144555 I dont think it wouldve been a -1 so not a big deal that it just merged16:31
clarkbjeblair but curious to hear your response(s)16:31
fungiclarkb: oops, sorry didn't realize you were reviewing that one or i'd have waited16:32
clarkbno problem I would've likely +2'd anyways16:33
nibalizerfungi: sorry, puppet is bad at return codes, yep yep :(16:33
funginibalizer: yupyupyup</muppet_alien>16:33
clarkbring ring16:33
*** craigbr has joined #openstack-infra16:33
jeblairclarkb: responded16:34
*** ddieterly has quit IRC16:34
fungiseriously wishing the remaining licensing issues would get worked out so that they can release later muppet show seasons on dvd16:34
*** hashar has joined #openstack-infra16:35
*** ddieterly has joined #openstack-infra16:35
jeblairoh is that the holdup?!16:35
fungias it was they had to do quite a bit of negotiating on the first few seasons. mostly over the music performance rights16:35
jeblairdid disney buy the old shows?  if so, i'm sure they can strong-arm whatever they need now.16:36
fungiartists had licensed a lot of that for broadcast but not for other forms of distribution, and in some cases only for broadcast in specific countries (which is part of why the musical numbers differed somewhat between usa and uk broadcasts)16:36
*** hashar has quit IRC16:36
fungiclarkb: i'm also good with 144556 but didn't approve now that i know you're currently reading through that stack16:38
clarkbjeblair oh is multiline required for findall I guess that makes sense.16:38
clarkbfungi thanks16:39
*** ihrachyshka has quit IRC16:40
nibalizerspeaking of releasing puppets, anyone want to try to release an openstackci puppet module?16:41
nibalizerI think all the parts are there, just hasn't been tested16:41
clarkbI was semi hoping we would clean each one up first16:42
clarkbso that we dont have to support things like building node.js from scratch (should use system node on trusty) for example16:43
clarkbbut maybe we have a module that is nice and ready to go?16:44
*** _nadya_ has quit IRC16:44
*** _nadya_ has joined #openstack-infra16:45
funginibalizer: in light of i'm curious as to your opinion on whether we can retire that module entirely16:45
fungiit'll be pretty gutted after that lands16:45
*** _amrith_ is now known as amrith16:45
*** kashyap has quit IRC16:47
nibalizerfungi: oo nice16:48
clarkbjeblair new question there doesnt seem to be any checking againsts depends on in the same project. I guess that is desired to express multiple deps without merge commits or serializing the git tree?16:48
openstackgerritJames E. Blair proposed openstack-infra/zuul: Add aditional CRD tests
*** vilobhmm has joined #openstack-infra16:48
nibalizerfungi: i always had a fantasy that we should ship that provider into the stankevich-python module16:48
jeblairclarkb: you have inspired additional (passing! yay!) tests ^ :)16:48
nibalizersince that is a natural home for it16:49
*** ayoung has quit IRC16:49
*** tkelsey has joined #openstack-infra16:49
*** _nadya_ has quit IRC16:49
nibalizerclarkb: i also want to clean  up the modules before releasing, but I think testing the release workflow is a separate task16:50
jeblairclarkb: i'm not sure it's a pattern i want to encourage, but i didn't see a reason to make a rule against it at the moment (other than it will fail it it causes a cycle)16:50
*** hdd has quit IRC16:50
*** e0ne has quit IRC16:51
*** kgiusti has joined #openstack-infra16:51
*** ddieterly has quit IRC16:51
*** kgiusti has left #openstack-infra16:51
*** e0ne has joined #openstack-infra16:51
*** ddieterly has joined #openstack-infra16:51
clarkbjeblair ya I think my only concern is circumventing the checks gerrit already makes. Not a problem for us with zuul enforcing but possibly for downstreams?16:52
funginibalizer: in the case of the pip module, we already preinstall pip on all our systems and don't manage it with puppet anyway16:52
*** mjturek1 has joined #openstack-infra16:52
anteayareed: so I was looking for rossella_s's talk from paris summit and found the schedule with her talk but couldn't find the video, which is here:
*** hdd has joined #openstack-infra16:53
anteayareed: any way that the video can be linked back to the schedule for easier discovery?16:53
*** vilobhmm has quit IRC16:53
nibalizerfungi: so even the package{ $pip: } is redundant?16:54
reedanteaya, I don't know much about that site, let me check though16:54
nibalizerbecause get-pip runs as first as soon as the computer starts?16:54
funginibalizer: a number of classes in openstack_project actually does an include pip but i suspect that could be switched to an existing module16:54
*** ildikov has joined #openstack-infra16:54
funginibalizer: right we pull down and run before we even install puppet16:54
pleia2mrmartin: hey, I'm around now16:54
jeblairclarkb: circumventing which check?16:54
mrmartinhi pleia216:55
mrmartinso I've two commetns on this zanata patch16:55
nibalizerfungi: yea so all this is doing is telling apt/yum to install pip too, which is probably the opposite of what we want16:55
funginibalizer: likely16:55
pleia2mrmartin: ok16:55
mrmartinpleia2: I've a working widlfly, then how can I access zanata?16:55
funginibalizer: is the pip package provider part of stdlib?16:55
fungithe normal one i mean16:56
fungii have to assume so16:56
clarkbjeblair the your git parent must be mergeable check16:56
*** ddieterly has quit IRC16:56
pleia2mrmartin: go to port 808016:56
nibalizerfungi: puppet core actually, but yes16:56
clarkbjeblair you could push deps that dont rely on git to establish those deps so gerrit never knows about it16:56
jeblairclarkb: sure, but you can do that today :)16:57
mrmartinpleia2: it brings up the welcome to wildfly screen16:57
nibalizerfungi: in case you care
*** ddieterly has joined #openstack-infra16:57
clarkbjeblair but today there is no implicit dep, but point taken16:57
nibalizerpleia2: wildfly still sounds like a katy perry song16:57
*** ddieterly has quit IRC16:58
mrmartinpleai2: how do you test those puppet modules?16:58
pleia2mrmartin: then it didn't work :) if the standalone.xml did get installed, you can try restarting the wildfly service16:58
*** ddieterly has joined #openstack-infra16:58
funginibalizer: thanks. i'll take a stab at removing references to that module and see where that leaves us16:58
*** sarob has joined #openstack-infra16:59
pleia2nibalizer: hehe16:59
openstackgerritSpencer Krum proposed openstack-infra/puppet-apparmor: Update metadata.json and License
nibalizerfungi: ++16:59
funginibalizer: worth mentioning, we don't use puppet-apparmor17:00
fungire 15329417:00
pleia2mrmartin: and note that wildfly+zanata is slooow, so it takes a couple minutes to bring up the service with zanata17:00
mrmartinpleia2: I've tried that, but got a wunderschön java exception (a lot)17:00
*** ashleighfarnham has joined #openstack-infra17:00
*** luqas has quit IRC17:00
* fungi loves the term wunderschön17:00
*** mattfarina has joined #openstack-infra17:01
fungione of those german words which just doesn't work as well when translated into english17:01
mrmartinpleia2: this is the reason I asked how are you doing the tests, maybe we can share some vagrant scripts17:01
*** abhirc_ has quit IRC17:02
mrmartinfungi: that was the first word in my mind when I saw the huge amount of java exception17:02
pleia2mrmartin: I don't use vagrant, I've got a prepped VM (snapshotted to where I want it with puppet installed) in hpcloud17:02
fungimrmartin: it describes giant java backtraces rather well, if somewhat sarcastically17:02
*** SumitNaiksatam has quit IRC17:02
*** SumitNaiksatam has joined #openstack-infra17:03
nibalizerfungi: all the better to use it as a gunnia pig for release pipeline testing17:03
mrmartinpleia2: ok, and what you have deployed inside that vm? do you have a script for that?17:03
pleia2mrmartin: so to do it manually, the steps are here:
*** markmcclain has quit IRC17:04
*** amrith is now known as _amrith_17:04
*** denis_makogon has quit IRC17:05
*** sarob has quit IRC17:05
*** sarob has joined #openstack-infra17:05
*** bdpayne has joined #openstack-infra17:05
*** sarob has quit IRC17:05
*** sarob has joined #openstack-infra17:06
*** tqtran has joined #openstack-infra17:06
mrmartinpleia2: I'm trying to figure out what I'm doing wrong if it works for you17:07
pleia2mrmartin: it doesn't work for me17:07
mrmartinoh ok17:07
pleia2that's why the patch is WIP17:07
reedanteaya, which page would you expect the recordings to be in?17:08
pleia2mrmartin: if you look at patch set 2, I explain the order in which we *want* things to happen in, but it isn't quite there yet17:08
reedanteaya, do you mean you'd like to see the recording linked from the site?17:08
mrmartinpleia2: ok, and how can I help you to move forward? What I see that I can't even login with the admin account to wildfly17:08
clarkbI have been nerd sniped by the doreconfigureevent function17:08
mrmartinand this two dependency things must be resolved17:09
clarkbjeblair ^ where does it re enqueue it seems to only remove builds/items17:09
*** SumitNaiksatam has quit IRC17:09
*** markmcclain has joined #openstack-infra17:09
pleia2mrmartin: mostly identifying dependencies that are failing, fixing them in the way that makes the most sense (ie - depending on Services rather than Files) and testing until we get it right17:09
*** mriedem has quit IRC17:09
*** mpaolino has quit IRC17:10
*** sarob has quit IRC17:10
*** pblaho has joined #openstack-infra17:11
*** ashleighfarnham has quit IRC17:11
*** ashleighfarnham has joined #openstack-infra17:11
mrmartinpleia2: I'll review this again tomorrow, and at least reach a point where I can access the admin console in wildlfy and somehow remove the zanata java exceptions17:11
openstackgerritElizabeth K. Joseph proposed openstack-infra/puppet-zanata: Initial commit of Zanata-specific sections
pleia2mrmartin: thanks, this should fix the comments you had ^^17:12
openstackgerritJames E. Blair proposed openstack-infra/zuul: Add aditional CRD tests
clarkboh derp I see it now17:13
pleia2mrmartin: I'm mostly away Friday and Monday, but I'll try to pop in to have a look at any comments you might have17:13
*** SumitNaiksatam has joined #openstack-infra17:13
clarkbnow to finish this review17:13
jeblairclarkb: whew, excellent timing.  i just read yor question.  glad i could help!17:13
mrmartinI'll have some meetings tomorrow afternoon too, but I guess it will require some time to deep dive in the java17:14
*** gondoi is now known as zz_gondoi17:14
openstackgerritChris Dent proposed openstack/requirements: Update gabbi to 0.6.0
*** rverchikov has quit IRC17:16
pleia2mrmartin: I'm hoping it doesn't have to go that far, the manual instructions on the etherpad do actually work17:16
pleia2mrmartin: so it's just a matter of getting things in the right order for the puppet module17:16
anteayareed: yes, so when I search I can find her talk in the old schedule with the description, would be great to have the link there too17:17
*** zz_gondoi is now known as gondoi17:17
mrmartinpleia2: yeah it is a basic rule in IT, you just need to push buttons and click with the mouse in proper order, and voilà everything starts to work magically :)17:17
*** sarob has joined #openstack-infra17:17
*** _amrith_ is now known as amrith17:18
*** swamireddy has quit IRC17:18
*** ldnunes has quit IRC17:18
reedanteaya, since is proprietary stuff, I'm not sure I want to give them more valuable content but I do agree with you that the videos on don't have enough metadata17:18
mrmartini'm going to find the proper order. sometimes it is not trivial to find the dependency between steps and find the notification chain between services.17:19
reedanteaya, for example if videos on contained also the information about the track, they would be easier to consume and to group17:20
pleia2mrmartin: yeah, I figured that out pretty quickly :\17:20
pleia2mrmartin: thanks for your help17:20
*** ddieterly has quit IRC17:20
*** ldnunes has joined #openstack-infra17:20
*** andymaier has quit IRC17:20
*** ddieterly has joined #openstack-infra17:20
reedanteaya, anyway, I think your suggestion is worth making. Best thing is filing a bug against openstack-www on launchpad17:20
*** rmcall has joined #openstack-infra17:21
jeblairor, perhaps confusingly, against 'openstackweb' in storyboard.  :/!/project/72917:21
*** koolhead17 has quit IRC17:22
anteayaah how about the launchpad one with a storyboard story with the launchpad bug link?17:23
anteayareed jeblair do we like that or no?17:23
*** signed8b_ has quit IRC17:23
*** jpich has quit IRC17:23
reedjeblair, i'm not sure the relevant actors are overseeing Storyboard, but I may be wrong17:24
*** yamahata has quit IRC17:24
fungiyeah, i think until we have them committing directly into the project on gerrit rather than continuing in github, filing stories in storyboard is premature anyway17:24
reedanteaya, I'd suggest to file on launchpad so that the developers of www can chime in (I don't think they're using SB yet)17:25
openstackgerritMerged openstack-infra/puppet-jenkins: Drop unused python3 classvar from jenkins::slave
*** pradk has joined #openstack-infra17:25
*** cdent has quit IRC17:25
openstackgerritJeremy Stanley proposed openstack-infra/system-config: Stop using the puppet-pip module
funginibalizer: ^17:25
*** MaxV has quit IRC17:26
*** jlibosva1 has joined #openstack-infra17:26
*** abhirc has joined #openstack-infra17:26
*** rverchikov has joined #openstack-infra17:26
*** ayoung has joined #openstack-infra17:26
*** Ryan_Lane has joined #openstack-infra17:27
*** timcline has quit IRC17:28
*** jlibosva has quit IRC17:28
*** BobBall is now known as BobBall_AWOL17:28
*** timcline has joined #openstack-infra17:29
anteayareed: okay I'll have some lunch and do so17:30
*** rverchikov has quit IRC17:31
*** abhirc has quit IRC17:33
clarkbjeblair: comments on 144556, trying to grok a couple pieces of behavior there17:33
*** timcline has quit IRC17:33
clarkbjeblair: I think at least one of them may be a bug but I am still wrapping my head around it17:33
clarkbjeblair: the bug is second comment on The other may be a bug too but I am less confident in that17:33
funginibalizer: actually seems to pass all tests, so maybe we can just abandon that module/repo now17:34
*** yolanda has quit IRC17:34
*** abhirc has joined #openstack-infra17:36
*** rverchikov has joined #openstack-infra17:36
*** che-arne has joined #openstack-infra17:36
*** abhirc has quit IRC17:36
*** tiswanso has quit IRC17:37
*** tiswanso_ has joined #openstack-infra17:38
*** koolhead17 has joined #openstack-infra17:39
*** koolhead17 has joined #openstack-infra17:39
*** e0ne is now known as e0ne_17:39
*** e0ne_ is now known as e0ne17:41
*** SumitNaiksatam has quit IRC17:41
*** belmoreira has quit IRC17:41
*** rverchikov has quit IRC17:42
*** kashyap has joined #openstack-infra17:43
*** johnthetubaguy is now known as zz_johnthetubagu17:43
*** tsufiev is now known as tsufiev_17:44
*** jgallard__ has joined #openstack-infra17:44
*** jgallard_ has quit IRC17:45
pleia2nibalizer: re: 151098 does stuff in the spec/ directory matter?17:46
*** jgallard__ has quit IRC17:52
*** shashankhegde has joined #openstack-infra17:54
*** baoli has quit IRC17:55
*** jp_at_hp has quit IRC17:55
*** ddieterly has quit IRC17:56
*** tiswanso_ has quit IRC17:56
*** ddieterly has joined #openstack-infra17:56
*** yolanda has joined #openstack-infra17:57
*** armax has quit IRC17:57
*** ayoung has quit IRC17:57
*** baoli has joined #openstack-infra17:57
*** yamahata has joined #openstack-infra17:59
jeblairclarkb: i think all your points are good.  i will exercise them with tests18:00
*** dustins_ has joined #openstack-infra18:00
*** bswartz has quit IRC18:01
*** ddieterly has quit IRC18:01
*** jlibosva1 has quit IRC18:01
*** armax has joined #openstack-infra18:01
fungitteggel: gilliard: i have an interesting failure when trying to delete an active but unused instance from usa east/region b... "Instance 9a89245c-6604-45af-a39d-0b20f0c8e9c7 is locked (HTTP 409) (Request-ID: req-19e591c0-43d6-4428-9ec5-b78854bdea1f)"18:02
*** dustins has quit IRC18:03
fungitteggel: gilliard: as far as i know nothing should have an outstanding request against that instance unless something's hung at your end18:03
*** derekh has quit IRC18:03
*** vponomaryov has left #openstack-infra18:03
*** vigneshvar has joined #openstack-infra18:06
*** ivar-lazzaro has joined #openstack-infra18:06
*** shashankhegde has quit IRC18:06
*** SumitNaiksatam has joined #openstack-infra18:07
reedanteaya, ACK18:07
*** ivar-lazzaro has quit IRC18:08
anteayareed: feel free to add if I didn't capture what you think needs to be conveyed18:09
*** ivar-lazzaro has joined #openstack-infra18:09
*** jistr has quit IRC18:09
*** ddieterly has joined #openstack-infra18:09
*** armax has quit IRC18:09
*** che-arne has quit IRC18:10
*** jistr has joined #openstack-infra18:10
*** patrickeast has joined #openstack-infra18:11
*** tkelsey has quit IRC18:11
*** sputnik13 has joined #openstack-infra18:12
*** mwagner_lap has quit IRC18:12
*** tiswanso has joined #openstack-infra18:13
*** gondoi is now known as zz_gondoi18:13
*** carl_baldwin has quit IRC18:14
*** carl_baldwin has joined #openstack-infra18:14
afazekasianw: ping18:15
*** bdpayne has quit IRC18:15
*** zz_gondoi is now known as gondoi18:15
nibalizerpleia2: the stuff in the spec/ directory is rspec-puppet tests18:17
*** armax has joined #openstack-infra18:17
nibalizerprobably they wouldn't pass even if we ran them renamed because our apache is so old and rspec has moved on18:17
*** ddieterly has quit IRC18:17
nibalizerfungi: i think i agree, if any resource used the 'pip3' provider, puppet would fail because that provider isn't available, which our testing should find18:18
funginibalizer: right, i already stripped that out in an earlier patchset which is already merged and causing no blowback afaik18:18
*** signed8bit has joined #openstack-infra18:18
pleia2nibalizer: ok, thanks18:19
*** mattfarina has quit IRC18:19
*** abhirc has joined #openstack-infra18:19
*** signed8bit is now known as signed8bit_ZZZzz18:19
funginibalizer: it was removed in which merged a couple days ago18:19
nibalizerfungi: i'm convinced18:21
fungiwow, we have a ton of (208!) disassociated floating ips in hpcloud. cleaning those up now18:21
*** e0ne has quit IRC18:21
zulcan someone let my email go through for openstack-announce?18:22
*** signed8bit_ZZZzz is now known as signed8bit18:22
*** jcoufal has quit IRC18:22
*** jistr has quit IRC18:22
fungittx: reed: ^18:22
openstackgerritAnne Gentle proposed openstack/requirements: Adds Sphinx theme, openstackdocstheme, for
*** timcline has joined #openstack-infra18:23
reedzul, on it18:23
*** abhirc has quit IRC18:23
zulreed:  thanks18:23
*** radez_g0n3 is now known as radez18:23
*** craigbr has quit IRC18:24
reedis HTML kosher on that list? :)18:24
*** ddieterly has joined #openstack-infra18:24
reedit's a neat html, approved18:25
*** ChuckC_ has joined #openstack-infra18:25
fungireed: the text/plain attachment is perfectly readable anyway18:25
*** achanda has joined #openstack-infra18:25
*** annegent_ has joined #openstack-infra18:26
fungizul: consider openpgp-signing release announcements in the future (can't remember if the other stable release managers have done traditionally but it's a good idea regardless)18:26
zulfungi:  k18:27
reedanteaya, did you google "openstack summit talks rosella" or "openstack summit talks rossella" with two S?18:27
*** melwitt has joined #openstack-infra18:27
*** timcline has quit IRC18:27
fungizul: also thanks for your hard work on 2014.2.2!18:27
reedanteaya, just out of curiosity, can you try again using 'rossella', two ss?18:27
zulfungi:  no worries18:27
*** mattfarina has joined #openstack-infra18:28
*** hdd has quit IRC18:28
*** ChuckC has quit IRC18:29
*** harlowja has joined #openstack-infra18:30
*** abhirc has joined #openstack-infra18:32
*** abhirc has quit IRC18:32
jeblairclarkb: i think the parts of the code you were looking at are correct.  i have provided explanations in follow-up comments.18:34
clarkbjeblair thanks looking in a moment18:34
jeblairclarkb: i'll add "and not live" to address the scheduler one in a followup change (since that one is not technically broken and has +2s)18:34
*** shashankhegde has joined #openstack-infra18:36
fungithe answers there make sense to me too18:36
*** markus_z has quit IRC18:37
*** gondoi is now known as zz_gondoi18:38
*** bdpayne has joined #openstack-infra18:38
*** unicell has quit IRC18:38
*** tonytan4ever has quit IRC18:39
openstackgerritJoe Gordon proposed openstack-infra/project-config: Make novaclient-dsvm-functional gating for novaclient
*** fandi has joined #openstack-infra18:40
*** afazekas has quit IRC18:40
*** zz_gondoi is now known as gondoi18:40
*** tkelsey has joined #openstack-infra18:41
openstackgerritBen Nemec proposed openstack-infra/reviewstats: Update tripleo core list
clarkbjeblair: oh I see the all jobs complete check actually iterates over jobs objects that never ran18:41
clarkbjeblair: rather than just being an empty list and promptly returning True18:42
clarkbjeblair: and your comment covers the case where ^ may be possible18:42
mordredjeblair, clarkb: design/style question ...18:42
clarkbok I am approving that change now. Thanks for looking into those items18:42
*** mriedem has joined #openstack-infra18:43
mordredbecause keystone auth is completely pluggable and unknowable upfront, we're currently treating auth params as an opaque dict in os-client-config and shade current, that gets validated against the declared auth_plugin18:43
mordredthat's the fully general thing and works across the board18:43
mordredHOWEVER - it requires the common case of username/password to be folded up into an auth dict too18:43
clarkbmordred: right I think that was the genesis of my complaint about keystone auth the other day18:44
mordredwhat do you think about special casing the original auth_url/username/password params to be providable at the top level18:44
mordredsince, even though it's not discoverable, it's the default and common enough18:44
*** Guest75042 has quit IRC18:44
clarkbmordred: I think it is likely best to not do that to better reflect what keystone expects18:44
mordredso you can say "if the user gave username/password/etc in the params, pop them and add them to the auth dict" - but not really make any attempts to do that for other forms18:44
clarkbmordred: will make everything a bit more readable when compared to reality18:45
*** tkelsey has quit IRC18:45
mordredclarkb: well, keystone expects a third thing18:45
clarkbthe plugin right?18:45
mordredclarkb: which is just that the keyword params are all at one level18:45
openstackgerritJames E. Blair proposed openstack-infra/zuul: Add aditional CRD tests
openstackgerritJames E. Blair proposed openstack-infra/zuul: Add extra liveness check to queue processor
mordredand if you didn't happen to provide the right ones, things just bomb18:45
*** Shrews has quit IRC18:46
*** teran has quit IRC18:46
jeblairclarkb: thank you!  i added an extra test in 153289 to cover re-enqueuing and the other comment is addressed in 15333418:46
clarkbmordred: ya I guess we have ot balance useability against debuggability18:47
clarkbmordred: I am a fan of consistency18:47
*** gondoi is now known as zz_gondoi18:47
mordredjust to be clear - the question is whether both of those forms should be allowed - or only one18:47
clarkbmordred: right I think just the one so it is consistent with other auth types18:48
jeblairmordred: and i assume the first form is what is required for anything other than userpass?18:48
mordredjeblair: yes18:48
*** Shrews has joined #openstack-infra18:48
jeblairmordred: i think i agree with clarkb, the consistency is not that hard on usability, and it makes it much easier to transition between that and other auth forms18:48
mordredadam_g, Shrews: have we missed anything ?18:49
*** rverchikov has joined #openstack-infra18:49
jeblairmordred: (eg, if you have the second form in the wild, then you have to have a long conversation about how using severed toes for authentication is completely different)18:49
mordredjeblair: ugh. I don't like talking about severed toes18:49
*** zz_gondoi is now known as gondoi18:49
lifelessbiometrics man18:49
jeblairmordred: yeah, me neither, which is why i'm highly negative about biometrics18:49
fungitoeprint scanner!18:49
* mordred clutches his toes18:49
Shrewsmordred: uhh.. i was absent from irc temporarily so i probably missed much18:50
mordredjeblair: how does one pass a severed toe as an arugment to a python function?18:50
*** koolhead17 has quit IRC18:50
*** thedodd has quit IRC18:50
fungiShrews: might be for the best anyway18:50
jeblairbut i think my work here is done; we can totally talk about "plucked hairs" if this bothers people18:50
*** signed8bit has quit IRC18:50
dhellmannfungi, sdague, clarkb: can we get some help unblocking oslo.messaging with this icehouse requirements cap?
mordredShrews: I was asking about special casing username/pasword18:50
mordredShrews: and clarkb and jeblair both think we should not18:50
*** carl_baldwin has quit IRC18:50
mordredfor consistency18:50
dhellmannI don't have +2 in the stable branches in requirements, so if we can change that, too, I'd appreciate it18:50
*** signed8bit has joined #openstack-infra18:51
Shrewsmordred: for the record, i agree with clarkb and jeblair. i just don't see the need yet18:52
clarkbdhellmann: you aren't in stable-maint-core? maybe what we need is to make reqs core have +2 on stable branches? (I don't think ttx is generally a fan of that for projects but in this case maybe it makes sense?)18:52
*** zz_avozza is now known as avozza18:52
patrickeasthey all, i know its not specifically an infra project, but i have a review up for devstack to fix my nodepool image caching problems i was wondering if some of you guys who are more knowledgable with the nodepool scripts could take a look and make sure i’m not totally wandering off into the weeds, or if you have any better suggestions18:52
dhellmannclarkb: I think we just changed the way stable maint works so that each project has their own core group. Maybe I wasn't added for requirements?18:53
dhellmannor maybe that split didn't happen there18:53
clarkbdhellmann: doesn't look like the split happened there18:54
*** tonytan4ever has joined #openstack-infra18:56
mordredShrews: cool - I am too - but I also sometimes suffer from make-it-too-consistent-itis18:56
*** sarob has quit IRC18:57
*** mattfarina has quit IRC18:57
*** sarob has joined #openstack-infra18:57
*** mwagner_lap has joined #openstack-infra18:57
sdake_hey folks, magnum core dev here, we want to use a package that is not in the gate18:57
sdake_is there a dependency management process that we must follow to add a dependency?18:58
sdake_prad is working on the blueprint in question:18:58
mordredsdake_: when you say "use a package" - can you expand?18:58
SpamapSsdake_: magnum isn't in the integrated gate, so you should be able to use anything.18:58
mordreddo you mean a distro package or a python package?18:58
*** thedodd has joined #openstack-infra18:58
openstackgerritRoman Verchikov proposed openstack-infra/project-config: Add cloud validation (cloudv) project to stackforge
sdake_mordred a python package18:58
mordredsdake_: yah - just install it in your job before you revoke sudo access18:59
sdake_SpamapS so we need the file in requirements.txt - then how does it get into the gate?18:59
*** yfried__ has joined #openstack-infra18:59
mordredOR - wait - don't listen to me18:59
mordredjust add it to your requirements.txt file18:59
mordredwe don't restrict what you can install19:00
jogosdake_: infra uses a full pypi mirror now so you can install anything from pypi19:00
sdake_I thought there was a local mirror of the gate19:00
sdake_oh got it jogo cool tahnks19:00
mordredit's a full mirror now19:00
sdake_well there you go prad - is it in pypi?19:00
mordredif it's not in pypi - you'll want to get it there19:00
pradsdague, no its not19:00
pradsdake_, ^^19:00
*** pelix has quit IRC19:01
sdake_prad so that is the next logical step then19:01
sdake_if we want to actually use that as a dependency19:01
openstackgerritJoe Gordon proposed openstack-infra/project-config: Make novaclient-dsvm-functional gating for novaclient
jeblairmordred: seen this:  dib->docker
*** signed8bit is now known as signed8bit_ZZZzz19:01
mordredjeblair: yes I have!19:01
*** mattfarina has joined #openstack-infra19:01
sdake_ya that dib->docker is pretty sweet huh :)19:02
sdake_that guy does alot of good work19:02
mordredI'm mildly confused as to why it's not a patch to dib ... but I'm looking forward to poking at it19:02
sdake_mordred I plan on getting him to merge it upstream into dib at some point19:02
sdake_since it really belongs in dib19:02
sdake_I think he just didn't want to mes with the special cases possibly19:02
*** dizquierdo has joined #openstack-infra19:03
dhellmannclarkb: any thoughts on the best way to handle the stable branch permissions?19:03
jeblairoh, er, that commit wants submodules19:03
*** avozza is now known as zz_avozza19:03
clarkbdhellmann: we should probably bug ttx about splitting that stable maint core group too19:03
jeblairthat's not going to work19:03
clarkbdhellmann: then we can throw you in it19:03
dhellmannclarkb: makes sense, I'll talk to him about it19:04
*** signed8bit_ZZZzz is now known as signed8bit19:04
clarkbdhellmann: unless ttx is happy to use normal stable maint core thats an option too19:04
mordredjeblair: nope. it's not. I figured I'd let them discover that - but that's my passive aggressive side surfacing19:04
mordredsdake_: ^^19:04
clarkbsdake_: so read your blog post, it doesn't address iscsi. Is that something you just haven't tackled et or does that get fixed as part of using host net namespace?19:04
openstackgerritDan Prince proposed openstack-infra/project-config: Add TripleO f20 puppet job
sdake_clarkb haven't tackled yet but plan to test19:05
sdake_clarkb my speculation is it should work as is with the host net namespace19:05
*** dustins_ has quit IRC19:05
sdake_and sharing some devices using the --device flag to docker19:05
sdake_tbh I left rht adn had to ship my eqiupment back, so I essentially ran out of time19:05
clarkbsdake_: my understanding is that iscsi isn't namespaced at all so not sure if pid/net host namespace puts you in a place where that becomes a non issue19:06
clarkbsdake_: but am very curious to ehar if that does fix it19:06
sdake_it should be a non-issue - the container running with the super-rpived features is basically a super-process19:06
sdake_runnign on a host19:06
sdake_iscsi is not namespaced - but bindmounting /dev should get the job done19:07
sdake_also there is a doc pull request out from dan walsh to add --device=* to import the entire devfs into the container19:07
clarkbat what point do we stop calling it a container? :)19:07
sdake_my take is docker maintainers act pretty quick on these sorts of features because theya re mandatory to use containers for upgrade19:07
openstackgerritKhai Do proposed openstack-infra/jenkins-job-builder: Allow JJB to set connection timeout to jenkins server
sdake_clarkb granted, its not much of a container ;-)19:08
mordredit's plenty of a container19:08
fungiman, having tools which parse reference documents in the governance repo is a lot more effort than you might think. we seem to change file schemas in there at least once a cycle19:08
sdake_but it does allow atomic upgrade-  which is the part i'm after19:08
clarkbmordred: its basically a chroot with process management19:08
mordredif a container is a replacement for deb/rpm19:08
clarkbmordred: which is not quite what I think of when am told "container"19:08
mordredclarkb: well, and namespaced normal networking19:08
clarkbmordred: no they turn that off19:08
mordredclarkb: you're falling into the trap of wanting a word to mean something19:08
mordredclarkb: this is clouds19:09
clarkbmordred: maybe19:09
sdake_i dont think container is a replacement for deb/rpm, actually deb/rpm are still used - to build the container fs19:09
mordredclarkb: no words mean anything anymore19:09
clarkbmordred: ha19:09
mordredsdake_: I think it'sa  replacement for app deliery19:09
sdake_fluffy white things ;)19:09
*** thedodd has quit IRC19:09
sdake_mordred yup19:09
mordredsdake_: or, not a replacement -but sits in teh same space-ish19:09
mordredsdake_: "how do I run a copy of X without screwing my whole computer"19:09
jeblairmordred: please cloud me some cloud.19:09
clarkbmordred: the blog was basically make a super privileged container running in host process and network space with bind mount permissions19:09
sdake_one of the maintainers of  docker said "wow libvirt in a container, now I can recompile libvirt without busting my machine ;)"19:10
mordredclarkb: btw - that does not work with AFS19:10
mordredclarkb: so there is still_something_ the container is doing19:10
clarkbmordred: this allows you to make compute hosts in "containers" fluffy things that don't mean much anymore supposedly19:10
mordredthey're really easy to throw around and spin up and down19:10
fungimy take so far has been that application authors getting fed up with distro-specific toolchain and dependency issues and starting to distribute their software as container images are effectively turning themselves into (in many cases very inexperienced) derivative distro authors19:10
sdake_super prived containers have no isolation19:11
mordredfungi: yup - albeit most of them with a very single specific usecase19:11
mordredsdake_: yup19:11
clarkbmordred: ya that was sdake_'s use case. basically upgrade nova by turning container on and off19:11
*** ghostpl has joined #openstack-infra19:11
mordredfor instance - I have 2 docker containers running on my IRC host19:11
clarkbI mean I get it but to me that isn't a container19:11
mordredeach with a different version of bitlbee19:11
clarkbbut I will accept that "container" isn't supposed to mean anything19:11
mordredbecause $fail19:11
mordredbut it's really easy to do19:11
jeblairfungi: there's a security policy: they will totally upgrade dependencies if they see the press release on hackernews19:12
fungiand linux containers already don't provide great isolation from a security standpoint since there have been plenty of ways even recently to parlay uid=0 in the container to uid=0 outside the container, way more often than equivalent hypervisor breakout bugs anyway19:12
clarkbjogo: I have a new test env held to continue debugging the metadata stuff19:12
*** bradm has quit IRC19:12
clarkbjogo: I am pretty tempted to get you on the hosts at this point too if you are interested and can point me at an ssh pubkey19:12
*** dustins has joined #openstack-infra19:13
*** ddieterly has quit IRC19:13
fungiso basically people are running, in many cases very outdated and vulnerable, software inside containers which in turn can be leveraged by an attacker to gain control over the machine itself19:13
clarkbsdake_: but you only need this privilege for compute nodes right? everything else can run pretty isolated?19:13
*** ddieterly has joined #openstack-infra19:13
*** ChuckC has joined #openstack-infra19:13
mordredfungi: BUT - they were doing this before in virtualenvs and in alternate root paths with code vendored into source trees19:13
fungiand just yesterday one of my friends wanted to install freeipa to do centralized authentication. they seem to be encouraging people to download a docker image and run it that way19:14
sdake_I think storage may need super prived containers as well19:14
sdake_neutron will definately need --net=host (eg no network namesapce for the container)19:14
*** baoli has quit IRC19:14
sdake_so ya, the whole schebang needs some super privileges to do the job19:14
*** vilobhmm has joined #openstack-infra19:14
sdake_but the pid namespace sharing is only needed for compute19:14
openstackgerritKhai Do proposed openstack-infra/jenkins-job-builder: Change test folder name to 'fixture' for consistency
openstackgerritKhai Do proposed openstack-infra/jenkins-job-builder: Implement recursive search for test fixtures.
fungisdake_: yes, storage control from within a container needs extra care in fact because if you grant the container control over the system's root block device then it can achieve breakout trivially19:15
openstackgerritMerged openstack-infra/zuul: Support cross-repo-dependencies in independent pipelines
sdake_i'm planning a summit talk for the next 2 mo of r&d I do in this area19:15
*** ghostpl has quit IRC19:15
sdake_to present what I've found out :)19:15
mrmartinnibalizer re19:15
*** dizquierdo has quit IRC19:16
jeblairclarkb, fungi: if you have a min, could you look at 153289?  it has some fixes for racy tests  in the changes that just merged (which i recheck-bashed in, so... :)19:16
nibalizermrmartin: ?19:17
jeblairsdake_: hopefully it will be on the non-design-summit day and i can see it :)19:17
mrmartinnibalizer: here you commented an old patchset, I have newer one there, where the problem got solved finally19:17
*** ChuckC_ has quit IRC19:17
sdake_jeblair ya - i suspect that will be a busy day :)19:17
*** dustins has quit IRC19:17
mrmartinsame here:
nibalizermrmartin: oh yea i got gertty'd19:18
*** ddieterly has quit IRC19:18
clarkbjeblair: yup can do19:18
mrmartinok no prob19:18
SpamapSfungi: just responding to your anti-vendoring arguments late. nearly 20 years of Debian yelling at library authors to stop breaking API's hasn't worked.. so... people vendor.19:18
mrmartinawesome, it was out of snyc19:18
jogoclarkb: give me 10 minutes, and I can be ready19:18
*** ayoung has joined #openstack-infra19:18
mrmartinthanks anyway19:18
clarkbjogo: ok I am going to review zuul code while I wait19:18
SpamapSfungi: but you are right, you become your own mini-distro. and that is effective.. so it continues.19:19
clarkbSpamapS: did you see the thing about Go that dstufft linked esterday?19:19
clarkbSpamapS: was amazing19:19
SpamapSclarkb: no, canhaz?19:19
SpamapSclarkb: kthxbai19:19
*** ashleighfarnham has quit IRC19:20
fungiSpamapS: yep. for the freeipa example, i mostly trust the freeipa authors to design and maintain an identity and auth management platform. i don't so much trust them to maintain their own custom linux distro to go with it19:20
*** Shrews has quit IRC19:20
*** koolhead17 has joined #openstack-infra19:20
pleia2mmm free ipa19:20
pleia2(wrong ipa)19:20
fungipleia2: it's not as hoppy as the sort you pay for, but i won't turn it down19:21
*** ddieterly has joined #openstack-infra19:21
dstufftgetting yelled at for vendoring is fun19:21
*** ddieterly has quit IRC19:22
SpamapS"we can't use the source code that we have to recreate that game"19:22
SpamapSthats the pain I had the first time I tried to build juju's go rewrite like 3 years ago :-P19:22
*** ddieterly has joined #openstack-infra19:22
SpamapS(mostly remedied by the juju team hiring all the go people they could to maintain the libraries and keep things working)19:22
sdake_juju had a team? ;-)19:23
dstufftthat does not seem like a strategy that can scale effectively19:23
SpamapSsdake_: has one. Quite a large one filled with _excellent_ programmers.19:23
SpamapSdstufft: well eventually you figure out how to lock down these deps and ship them as a distro package. I'm not sure how they did it, but I assume they're just managing git trees and merging as needed.19:24
SpamapSaka "vendoring"19:24
*** timcline has joined #openstack-infra19:24
sdake_SpamapS interesting19:24
dstufftit was "fun" to watch Debian reinvent vendoring for python-pip except a jankier more error prone version of it19:24
nibalizerbuilding a go program and putting it in a container does solve the problem of 'maintaining a distribution' though19:25
mordreddstufft: ++19:25
*** gondoi is now known as zz_gondoi19:25
*** abhirc has joined #openstack-infra19:25
SpamapSnibalizer: go produces single binaries. There is one file. No container needed.19:25
mordredit's the idea that the distros are some how better equipped to handle the vendoring problem - when instead what they frequently do is backport python3.4 features into python2.619:26
SpamapSAnd it's a statically linked binary.19:26
mordredor don't release updates of completely broken things19:26
SpamapSSo as long as you aren't running crazy-crack versions of Linux or glibc, it will work.19:26
mordredwhile waxing poetic about the process they follow19:26
SpamapSmordred: and refuse to ship python3.4 bug fixes.... and...19:26
mordredthe truth is that it'sa  best-effort game all the way around19:26
mordredsome teams are good and release good stuff19:26
mordredsome are not and do not19:26
SpamapSYeah, do what you need to make it work for you and your users. Occasionally, optimize things where you can.19:27
mordredsome of them work in the context of distros, some do not19:27
*** spzala has joined #openstack-infra19:27
SpamapSDebian trades social harmony for process optimization. ;)19:27
*** signed8b_ has joined #openstack-infra19:27
*** zz_gondoi is now known as gondoi19:27
SpamapSSpeaking of Debian, we should use it. ;)19:28
nibalizerSpamapS: wait, are you saying debian processes are optimized?19:28
mordredSpamapS: fungi has a patch19:28
mordrednibalizer: HIGHLY19:28
mordrednibalizer: just not for the outcome you're wanting them to be optimized for19:28
*** timcline has quit IRC19:28
openstackgerritMerged openstack-infra/zuul: Add aditional CRD tests
fungimordred: still underway. i have to finish solving the move-mysql-setup-out-of-image-building problem as a next step19:29
nibalizerSpamapS: the container still provides much utility imho19:29
dstufftmordred: nibalizer they seem optimized for the constraints of the early 90's19:29
SpamapSnibalizer: optimized to their detriment. ;)19:30
mordreddstufft: listen - it's important to make sure the base os can fit on a floppy disk ...19:30
waynrzaro: it's not clear to me how to pull the changes you posted to my gerrit review so i can address pelix's concern19:30
*** ashleighfarnham has joined #openstack-infra19:30
mordredsnark aside- I do appreciate that about debian19:30
SpamapSnibalizer: oh yes, just saying in this case, the container isn't needed to ship separate libraries for Go things.19:30
mordredbecause the tiny output of debootstrap is totally usable and small19:30
mordredwhereas the fedora equiv is HUGE19:31
*** signed8bit has quit IRC19:31
fungimordred: SpamapS: the short story is that debian jessie has a python 3.4 which will work for oslo.messaging and oslo.rootwrap, but since jessie isn't released yet hpcloud has no base image for it (rax does), so we'll need to dib bare-jessie on hpcloud. except we're not dib'ing bare-.* workers yet because mysql database setup in a chroot is a no-go19:31
*** andreykurilin_ has joined #openstack-infra19:31
waynri ran the command listed in the email but it doesn't work and i don't have time to play around with gerrit19:31
*** sarob_ has joined #openstack-infra19:31
mordredfungi: yup19:31
* fungi gets back to that19:31
SpamapSoh interesting. If only we had a mysql-like database you could run in single process mode and pipe things into.19:32
* SpamapS hears trumpeting and splashing and smells something burning19:32
fungiSpamapS: you would drizzle things into it i think?19:32
dstufftmordred: I appreciate how the Fedora folks seem to care a lot more about remaining close to upstream, though I don't actually run FEdora because the release cycle is too fast for servers and RHEL is too slow for me19:32
nibalizerfungi: is mysql setup in a chroot a no go because of puppet? or because mysql likes to actually start or what is the reason?19:32
mordreddstufft: ++19:32
*** sarob has quit IRC19:32
mordrednibalizer: we need to start it and run some sql to set up the right things19:33
mordredand this fails quite spectacularly when we try it in our dib run19:33
funginibalizer: that's part of it i think, but more generally we'd like to not have separate bare-.* and devstack-.* workers anyway and so we should do things like db preconfig for unit tests in the unit test jobs instead19:33
*** timcline has joined #openstack-infra19:34
*** bswartz has joined #openstack-infra19:34
*** teran has joined #openstack-infra19:34
*** teran has quit IRC19:35
clarkbthough I think generally baking as much stuff as you can into an image is important19:35
fungidstufft: conversely, debian folks seem to care a lot more about being able to run on a broader variety of hardware and being able to coinstall pretty much anything with pretty much anything else without breaking (and also carry orders of magnitude more packaged applications)19:35
clarkband wish that dib had a better story for that beacuse if you exclude our needs saving a few sseconds across thousands of node boots is $$19:35
*** shashankhegde has quit IRC19:36
SpamapSWhat causes mysql not to start btw?19:36
dstufftfungi: quite possibly, which is probably why I don't like their policies much since there's pretty much only amd64 that matters to me and I prefer to get my applications from upstream not the wierdo debian version of them19:37
mordredSpamapS: dib, last we checked19:37
SpamapSI'd expect it would start just fine as long as you only have one running.19:37
clarkbSpamapS: dib does a thing to make all services not start19:37
SpamapSdib doesn't let it start because it was installed as a package.19:37
SpamapSYou can still /etc/init.d/mysql start19:37
clarkbSpamapS: iirc no19:37
clarkbSpamapS: puppet blows up doing that19:37
mordredSpamapS: it's possible there is something in what we're doing that you might have suggestions on how to do better19:37
SpamapSIf not that is totally fixable.19:37
fungithough that does also bring up the concurrency issue19:37
clarkbSpamapS: instead you would have to explicitly start a daemon19:37
mordredclarkb: I believe SpamapS just volunteered to fix this! :)19:37
fungibut since we're building dib diskimages serially in nodepool we can ignore that part19:38
* mordred hands SpamapS a fluffy goat19:38
clarkbfungi: ya I think we have decided concurrent builds are a problem for lots of reasons and shouldn't count on it19:38
clarkbalso I just confirmed that jeblairs fix appears to have fixed our double upload problem19:39
fungiclarkb: i also think that if we actually want build concurrency, that's when we pull the trigger on the image build worker spec19:39
clarkbfungi: yup19:39
SpamapSI'm quite happy to fix19:39
mordredclarkb: do we have an easy-to-point-SpamapS at thing to try? I can dig one up but if you have one handy ...19:40
*** Sukhdev has joined #openstack-infra19:40
*** gondoi is now known as zz_gondoi19:41
SpamapSYeah like a set of elements and a command line are all I need.19:41
mordredSpamapS: well, skin might also help19:41
clarkbmordred: just run the bare-* thick_slave.pp module in dib19:42
clarkbmordred: I hthink there is already an element that does it, it just doesn't work19:43
*** sarob_ is now known as sarob19:43
SpamapSclarkb: I don't have enough context in infra's puppet to know where any of that would be.19:43
mordredSpamapS: hang on - working on it19:43
SpamapSI will gladly gain such context. :)19:43
*** dims__ has joined #openstack-infra19:44
*** r-daneel has joined #openstack-infra19:44
*** dimtruck is now known as zz_dimtruck19:45
*** teran has joined #openstack-infra19:45
tchaypojeblair: I’m very confused by 15221319:45
*** adalbas has quit IRC19:46
*** zz_gondoi is now known as gondoi19:46
mordredclarkb: nope. doesn't look like there is an element already19:47
mordredclarkb: there is just nodepool/elements/node-devstack19:47
clarkbSpamapS: elements are there. is the puppet runner which is run by and install.d script in node-devstack19:47
*** dims_ has quit IRC19:47
clarkbSpamapS: mordred you would need to swith thin => true to false there19:47
mordredSpamapS: so what is needed is an element like node-devstack where THIN=false19:48
*** dims__ has quit IRC19:48
*** gondoi is now known as zz_gondoi19:48
clarkband for convenience19:48
clarkbs/devstack/otherthing/ there too19:48
*** abhirc has quit IRC19:48
*** zz_gondoi is now known as gondoi19:49
*** dims__ has joined #openstack-infra19:50
*** dims__ has quit IRC19:50
openstackgerritMonty Taylor proposed openstack-infra/project-config: Add element for bare nodes
mordredSpamapS: ^^19:50
*** hdd has joined #openstack-infra19:51
openstackgerritMonty Taylor proposed openstack-infra/project-config: Add element for bare nodes
mordredSpamapS: NODE_TYPE=node-bare tools/ should have you covered19:51
*** dims_ has joined #openstack-infra19:54
*** dims___ has joined #openstack-infra19:54
*** sarob has quit IRC19:55
*** shashankhegde has joined #openstack-infra19:55
*** sarob has joined #openstack-infra19:57
*** _nadya_ has joined #openstack-infra19:58
*** dims_ has quit IRC19:58
*** tonytan4ever has quit IRC19:59
*** abhirc has joined #openstack-infra20:00
*** dalgaaf has quit IRC20:00
*** dustins has joined #openstack-infra20:01
jeblairtchaypo: i don't know how it happened (current bug?  old bug?  interrupted sync?), but i found data like that in my local db, and it rendered that change unworkable in gertty because the correct data were never displayed20:01
*** dims__ has joined #openstack-infra20:01
jeblairi mean it rendered the change associated with the bad data20:01
*** sarob has quit IRC20:02
tchaypoah, so it’s not the case that the server has multiple approvals, it’s just that we’ve ended up with multiple local approvals?20:02
*** sarob has joined #openstack-infra20:02
*** dims___ has quit IRC20:03
*** gondoi is now known as zz_gondoi20:03
*** zz_gondoi is now known as gondoi20:04
*** aysyd has quit IRC20:05
anteayareed: I updated the bug report20:05
*** rfolco has quit IRC20:06
*** bdpayne has quit IRC20:06
*** bdpayne has joined #openstack-infra20:08
*** thedodd has joined #openstack-infra20:08
*** koolhead17 has quit IRC20:08
*** zz_dimtruck is now known as dimtruck20:09
*** dims_ has joined #openstack-infra20:09
openstackgerritMerged openstack/requirements: Update zake to 0.1.6
*** syerrapragada has joined #openstack-infra20:11
*** dims__ has quit IRC20:12
openstackgerritJeremy Stanley proposed openstack-infra/system-config: Generate machine-readable tree of change approvers
anteayaI just saw Lauren's forward of the L naming poll to the operator's list20:15
anteayaI'm not sure I can bring myself to vote20:15
*** ldnunes has quit IRC20:15
*** dims__ has joined #openstack-infra20:15
*** e0ne has joined #openstack-infra20:17
*** e0ne is now known as e0ne_20:17
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Use open() instead of file()
*** dims_ has quit IRC20:17
*** e0ne_ is now known as e0ne20:19
*** ildikov has quit IRC20:19
*** dprince has quit IRC20:19
jrollis there a way to only have a job vote on master and not juno?20:20
jroll(juno backports, that is)20:20
anteayajroll: yes20:20
* anteaya gets an example20:20
clarkbanteaya: is the problem that ops were asked to vote too? I thought we were being inclusive?20:21
*** ashleighfarnham has quit IRC20:21
*** sigmavirus24 is now known as sigmavirus24_awa20:21
jrollanteaya: awesome, ty20:22
anteayaclarkb: the problem is the four choices20:22
anteayaI'm glad ops are included20:22
*** shashankhegde has quit IRC20:22
anteayathose four words are not denoting canadian anything20:22
anteayathat job is not testing stable or icehouse branches20:22
jrollaha, awesome20:22
* jroll makes patches20:22
anteayajroll: welcome20:23
notmynameif global_requirements were a CI test (does the project's requirements conflict with the global requirements, taking into account versions) instead of an automated copy/paste into every source tree, would that be something that would be supported or opposed?20:24
jrollanteaya: for a job not already in the jobs: section, I can just add it and override defaults as necessary, correct?20:24
openstackgerritClark Boylan proposed openstack-infra/devstack-gate: DO NOT MERGE testing nova multinode fix
openstackgerritClark Boylan proposed openstack-infra/devstack-gate: Enable nova-api-metadata for multi host nova-network
anteayajroll: that would be a start, offer the patch and we will review it into shape from there20:24
jrollok, ty20:24
tchaypoanteaya: I don’t think the K release reflected anything french, nor J anything Georgian20:25
anteayajroll: not sure we are talking the same language when you say override defaults so let me see what you mean20:25
clarkbanteaya: ah yes, neither is havana very portlandy...20:25
anteayaclarkb: true20:25
anteayatchaypo: well metric and french20:25
anteayathey did come up with the metric system20:25
tchaypothere is that20:25
*** che-arne has joined #openstack-infra20:28
*** kgiusti has joined #openstack-infra20:30
*** yfried has joined #openstack-infra20:30
*** tiswanso has quit IRC20:30
*** dims_ has joined #openstack-infra20:30
*** tiswanso has joined #openstack-infra20:31
openstackgerritJim Rollenhagen proposed openstack-infra/project-config: Make agent_ssh job not run on stable/juno Ironic patches
jrollanteaya: ^ I'd rather see land and fix the real problem, however I can't guarantee that patch fixes it20:31
*** dims__ has quit IRC20:31
*** rlandy has quit IRC20:32
*** yfried__ has quit IRC20:33
*** zz_avozza is now known as avozza20:33
rcarrillocruzhey folks, i'm doing some modifications to an openstack-infra puppet module20:35
*** dims__ has joined #openstack-infra20:35
*** vigneshvar has quit IRC20:35
rcarrillocruznow that these modules are split, what's a quick way to install the module i'm modifying in a test system?20:35
rcarrillocruzdo i need to do 'puppet module build foo && puppet module install foo'20:35
rcarrillocruznibalizer: ^20:35
clarkbrcarrillocruz: no, just clone the module to your module path20:35
rcarrillocruzhmm, so doing a link on /etc/puppet/modules to my home folder containing my code would work, right?20:36
nibalizerso you want to clone it20:36
nibalizerya symlink hax also works :)20:36
rcarrillocruzexcellent, thx20:36
*** dims_ has quit IRC20:37
*** jedimike has quit IRC20:37
clarkbor put your repo path in the module path20:37
clarkbthat is what I do since I tend to run puppet apply --modulepath anyways20:37
*** avozza is now known as zz_avozza20:38
nibalizeralso --modulepath=. is a thing i do a lot20:38
funginotmyname: i'm not sure what you're asking... right now there is a ci test where if you propose a change to your requirements list and it conflicts with global requirements then the job fails. we just have an additional job which runs when a change merges to global requirements to propose the difference to all covered projects so that they'll be able to stay in sync with each other20:38
dstaneki've been working on how we can add real functional testing in Keystone and would appreciate some eyes on
dstanekit talks about how i'm using devstack/devstack-gate and thought you guys would have some great input20:39
rcarrillocruznibalizer: --modulepath honours order I assume?20:39
nibalizerbut ... im not sure...20:39
rcarrillocruzi can try , it's a quick thing to test20:39
nibalizerdont put different versions of the same module in your modulepath20:40
notmynamefungi: right. what I'd prefer to see is a CI job that parses the local requirements and the global requirements and shows where the local is in conflict. eg if my repo wants foo>=3.0 and global wants foo>=4, that's ok20:40
fungianteaya: keep in mind that we also almost had kyoto, which would have suggested not-france about as much as london suggests not-canada20:40
nibalizerthats trusting puppet not to do something bonghitsy, and it doesn't deserve that much trust20:40
notmynamefungi: but it's not ok if I have things that aren't in global or require different versions than global20:40
rcarrillocruzhah, yah, i noticed this evening with the concurrent puppet module install/upgrades :-)20:40
pleia2anteaya: fwiw, I think "love" when I think of canada :)20:41
pleia2even if it is kind of cold20:41
pleia2more kitty snuggles20:41
clarkbnotmyname: no we haven't done that because pip doesn't quite work that way. Its resolver is simple so >=3.0 and >=4.0 could result in vastly different installs20:42
funginotmyname: ahh, i think i see. so if global requirements has foo>1, project A requires foo>1,<2 and project B requires foo>2 you want them to succeed even though project A and project B can't be installed together?20:42
clarkbnotmyname: until pip supports proper dependency resolution we kinda have to stick with being identical20:42
notmynamefungi: well, mostly I'd like to not have a patch with 121 changesets in swift ;-)20:42
clarkba very recent example of this is the issue zuul has/had with lockfile20:42
clarkbnotmyname: we can stop proposing updates to swift if that is your preference20:42
clarkbnotmyname: we have done it at a convenience (and most projects seem to like it)20:43
*** gondoi is now known as zz_gondoi20:43
notmynameclarkb: no, I actually don't want them to stop being proposed20:43
notmynamebut I'd prefer that since our requirements actually aren't in conflict (ie if you install global-requirements, that satisfies swift) it should pass. instead of look for exact version matches20:43
funginotmyname: it's an attempt to reflect the state of requirements if you want to be able to install a python package of swift along with some other openstack project (for example to run tests), but if you don't want swift tested with other software that's probably an option20:43
notmynamefungi: no, I want to pass if installing global requirements versions would satisfy requirements. and fail otherwise20:44
*** ashleighfarnham has joined #openstack-infra20:44
clarkbnotmyname: right but unfortunately that isn't how pip works20:44
notmynamewhy is that dependent on pip? it's version strings that also affect distro packages20:44
funginotmyname: we'd need to trim down the given requirement in global requirements to the smallest common intersection of versions across all coinstalled projects20:45
clarkbnotmyname: for example zuul said it depended on lockfile>=0.8, python-daemon said it required lockfile>=0.10. If you had lockfile 0.8 already installed then zuul broke because python-daemon would not upate lockfile to 0.1020:45
clarkbnotmyname: because pip is the tool being used here20:45
clarkbnotmyname: its the forward looking we are testing things that haven't arrived in distros situation20:45
*** r1chardj0n3s_afk is now known as r1chardj0n3s20:46
funginotmyname: distros are taking clues from the requirements lists but those files are primarily to service python package toolchains not distro package toolchains20:46
notmynameso installing python-daemon with pip didn't upgrade to the minimum required version of lockfile (ie 0.10)?20:46
clarkbnotmyname: correct20:46
clarkbbecause pips version resolver is not so good20:46
notmynameclarkb: sounds like a bug in pip, not a bug with parsing version strings20:46
clarkbit is a bug in pip, I filedit a couple yaers ago20:46
clarkbdstufft has done some work to correct it20:46
clarkbbut its not a simple fix20:47
clarkbnotmyname: but given that is the tool we have we are somewhat stuck supporting it20:47
notmynamegotta run20:47
fungigiven how long that's been the case for pip (pretty much for as long as it's had the ability to take a list of packages) it would be more fair to call it a design shortcoming20:47
*** teran has quit IRC20:48
fungior perhaps an incomplete implementation20:48
*** zz_gondoi is now known as gondoi20:49
*** dims_ has joined #openstack-infra20:49
* dstufft doesn't bother to try to categorize it20:49
dstufftother than crappy thing that needs fixed20:49
*** ildikov has joined #openstack-infra20:49
*** dims__ has quit IRC20:50
*** vigneshvar has joined #openstack-infra20:50
openstackgerritDoug Wiegley proposed openstack-infra/project-config: Add pylint jobs to service modules
dstufftmost SAT solvers tend to want the entire boolean expression up front, but given that we can't construct a dependency graph without downloading things that's impossible for us, so we need to either resolve the entire boolean expression multiple times for each dependency, or we need something that can iteratively include additional boolean constraints which makes the implemetnation of the SAT solver itself harder20:52
*** carl_baldwin has joined #openstack-infra20:52
*** dims__ has joined #openstack-infra20:52
dstufftthen there's the whole, pip isn't really designedin a way that makes it easy to do this, it's code is spaghetti and we need to tease it apart better before we can really do that :/20:52
*** gondoi is now known as zz_gondoi20:53
anteayafungi: true20:54
anteayapleia2: well that is heartening20:54
*** dims_ has quit IRC20:54
*** abhirc has quit IRC20:54
anteayacurrently when I think of canada I think of a tanking sole industry economy and a falling dollar20:54
anteayaso perspective counts I guess20:54
*** signed8b_ is now known as signed8bit_ZZZzz20:55
*** zz_gondoi is now known as gondoi20:55
*** annegent_ has quit IRC20:55
*** dims__ has quit IRC20:55
*** dims__ has joined #openstack-infra20:56
*** dims__ has quit IRC20:56
*** andreykurilin_ has quit IRC20:57
*** gondoi is now known as zz_gondoi20:57
*** signed8bit_ZZZzz is now known as signed8b_20:59
jogoclarkb: success :)21:00
jogowe should recheck a few times though21:00
clarkbjogo: cool and yes21:00
*** annegent_ has joined #openstack-infra21:02
*** pblaho has quit IRC21:04
*** dims_ has joined #openstack-infra21:05
*** zz_avozza is now known as avozza21:06
*** zz_gondoi is now known as gondoi21:07
*** mrmartin has quit IRC21:10
*** _nadya_ has quit IRC21:10
fungidstufft: would having something pypi-side parsing declared dependencies of each version of every package and making that indexed set of relationships available to pip for download help at all?21:10
*** erikmwilson has quit IRC21:11
*** erikmwil- is now known as erikmwilson21:11
fungidstufft: i see that as being similar to how linux distro package managers handle the problem. they have indexes of available versions of software including what versions of what other software each depends on21:11
*** hugh_afk is now known as hughhalf21:11
*** zul has quit IRC21:12
fungithat provides the necessary up-front metadata the sat solver would need to chew on, right?21:13
*** tiswanso has quit IRC21:17
*** ociuhandu has quit IRC21:17
*** Shrews has joined #openstack-infra21:17
*** dizquierdo has joined #openstack-infra21:18
*** shashankhegde has joined #openstack-infra21:19
*** gondoi is now known as zz_gondoi21:19
*** dangers is now known as dangers_away21:20
*** zz_gondoi is now known as gondoi21:20
*** baoli has joined #openstack-infra21:21
*** achanda has quit IRC21:22
*** dangers_away is now known as dangers21:23
*** vigneshvar has quit IRC21:23
*** baoli has quit IRC21:24
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Allow to wait for jenkins to start up
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Use open() instead of file()
*** baoli has joined #openstack-infra21:24
SpamapSdisk-image-create --no-tmpfs makes me sad (doubles speed on my laptop..). We should find better heuristics to make that automatic.21:25
*** gondoi is now known as zz_gondoi21:25
clarkbya dont tmpfs21:25
*** ddieterl_ has joined #openstack-infra21:25
*** ddieterly has quit IRC21:25
SpamapSclarkb: No, _using tmpfs_ doubles the speed.21:25
greghaynesheh, thought you had that inverted21:26
clarkboh I dont use it because I dont have enough ram21:26
greghaynesunless youre using eatmydata21:26
SpamapSBut I know it breaks sometimes and that is a net loss in productivity.21:26
clarkbbut I think we can now that we dont chattr21:26
SpamapSclarkb: 16G not enough?21:26
clarkbI have 2 on my build box21:26
SpamapSgreghaynes: eatmydata used to make many thinks 'splode21:26
SpamapShaven't tried it in almost 2 years21:26
openstackgerritJeremy Stanley proposed openstack-infra/infra-manual: Link to the anonymously available ICLA full text
greghaynesdid it, in fact, eat your data?21:27
SpamapSno but it ate my afternoon21:27
pleia2fungi: ooh, good idea21:27
clarkbbut it why the file cp instead of using the image mounted bugs me21:27
fungipleia2: it's in response to a thread on the legal-discuss ml21:27
SpamapSclarkb: there's a time when RAM saves time. I live by the fact that trading $$ for time is always a win.. if you have $$ and you spend your time making more that is. :)21:28
pleia2fungi: yep, I saw21:28
*** zz_gondoi is now known as gondoi21:28
*** tnovacik has quit IRC21:28
greghaynesSpamapS: oh, have you ever considered adding an lvm output format for dib?21:28
clarkbSpamapS sure it should just toggle behavior if using tmpfs21:29
greghaynesbasically mount an lvm and build on it, then dont do any conversion after21:29
greghaynesor maybe even more general 'blockdevice'21:30
SpamapSgreghaynes: No, but it would be quite easy to add an arbitrary-block-device output format.21:30
* greghaynes might poke at this21:30
greghaynesI think it would speed up my dev box a lot21:31
dstufftfungi: can't do it generically for sdist 1.0, for Wheels we can and for sdist 2.0 we will be able to21:31
SpamapSclarkb: I think the real problem is that the heuristic we have must have failed at some point, hence the --no-tmpfs.21:31
SpamapSgreghaynes: so you're thinking just use an lvm instead of loop?21:31
fungipleia2: i didn't want someone at fsf europe to think that we aren't responsive to concerns about that sort of thing21:31
SpamapSgreghaynes: thats what schroot does, very useful21:31
SpamapSI've often wondered why we didn't build diskimage-builder on top of schroot.21:32
SpamapSI'm sure there are reasons. :-P21:32
pleia2fungi: indeed!21:32
anteayajhesketh: morning21:33
*** denis_makogon has joined #openstack-infra21:34
zarowaynr: what change are you refering to?21:35
fungipleia2: particularly someone on the fsf-europe legal task force, apparently :/21:36
*** vigneshvar has joined #openstack-infra21:37
*** ddieterl_ has quit IRC21:40
*** krtaylor has quit IRC21:40
*** Sukhdev has quit IRC21:40
*** ddieterly has joined #openstack-infra21:40
*** hdd has quit IRC21:42
*** ddieterly has quit IRC21:42
*** ddieterly has joined #openstack-infra21:42
clarkbpleia2: fungi: I have no context21:43
clarkbis there a thread I should be reading?21:44
pleia2I included a link to the thread in my review21:44
clarkbgotcha, doesn't seem to cover the CCLA though? But we aren't involved in that best I can tell21:46
fungiclarkb: right, and the complaint in that particular e-mail (though not the older thread to which he was replying) was about the icla anyway21:47
fungiso while his subject says ccla, that wasn't apparently the concern21:47
*** hdd has joined #openstack-infra21:47
clarkbfwiw if you google "openstack individual contributor license agreement" the gerrit page is hit 2 and the wiki how to contribute is 121:48
clarkbso while I never think better docs are a bad thing, I am not entirely convinced this is a problem21:49
pleia2still, fungi's patch is probably The Right Thing To Do21:49
clarkbyup just about to approve it21:50
*** ddieterly has quit IRC21:50
*** achanda has joined #openstack-infra21:50
openstackgerritMerged openstack-infra/infra-manual: Link to the anonymously available ICLA full text
*** dhellmann_ has joined #openstack-infra21:51
fungiclarkb: yeah, i noticed the same, but would have felt like a boob pointing out to someone that they're "wrong on the internet"21:52
fungii'm hoping that these edits squeeze that url up from #2 to the #1 spot21:52
*** krtaylor has joined #openstack-infra21:52
*** Rockyg has joined #openstack-infra21:54
*** MaxV has joined #openstack-infra21:55
*** dhellmann_ has quit IRC21:55
*** bswartz has quit IRC21:56
waynrlooks like you uploaded a patchset rebased on another change but pelix's comment regarding the basename checking wasn't addressed so i wanted to work off your patch set but was not able to follow the instructions in the email i received from gerrit to pull the new patchset21:57
*** dhellmann_ has joined #openstack-infra21:57
waynri guess i could just rebase on gerrit master21:57
zarowaynr: are you using git-review by chance?21:57
*** thedodd has quit IRC21:57
waynrnaw i just tried the git pull command in the email21:58
*** dhellmann_ has quit IRC21:58
zarowill make your life easier21:58
waynrbut i do have git-review21:58
*** ZZelle_ has joined #openstack-infra21:58
zarowaynr: just run 'git-review -d 151898'21:58
*** rverchikov has left #openstack-infra21:58
*** gondoi is now known as zz_gondoi21:58
zarowaynr: then make your changes, do 'commit --amend', save then, do 'review'21:59
*** andreaf has quit IRC21:59
zaroopps that command is actually 'git review -d 151898'21:59
*** MaxV has quit IRC22:00
*** zz_gondoi is now known as gondoi22:00
waynroh i didn't even notice the dash, got yer change now22:00
*** andreaf has joined #openstack-infra22:00
*** thedodd has joined #openstack-infra22:00
*** abhirc has joined #openstack-infra22:00
*** e0ne has quit IRC22:00
zarowaynr: opps that last command is actuualy 'git review' instead of just 'review'22:01
notmynameclarkb: sorry I had to run suddenly earlier. thanks for the info. also fungi22:01
*** dannywil_ has joined #openstack-infra22:01
openstackgerritWayne Warren proposed openstack-infra/jenkins-job-builder: Implement recursive search for test fixtures.
*** mattfarina has quit IRC22:02
waynrthe awesomeness of gerrit just incremented in my mind22:02
*** patrickeast_ has joined #openstack-infra22:03
*** patrickeast has quit IRC22:03
*** patrickeast_ is now known as patrickeast22:03
waynrare there any video tutorials that demonstrate gerrity workfows?22:03
*** dhellmann_ has joined #openstack-infra22:03
fungiwaynr: funny you should ask22:03
*** dannywilson has quit IRC22:04
zarowaynr: openstack infra folks have lots of them on youtube.22:04
*** jaypipes has quit IRC22:04
anteayazaro: we do?22:05
krtayloranteaya, want to discuss the patch comments? I'd like to understand your concerns for the distinction between driver and non-driver testing22:05
zaroalthough i don't believe have have run across any video evidence of fungi22:05
*** bradm has joined #openstack-infra22:05
anteayakrtaylor: my concern is about adding a new requirement22:05
*** avozza is now known as zz_avozza22:05
anteayaas I have said several times, we can't get third party cis to follow the requirements we have22:05
krtayloranteaya, ok, but that is a different patch22:05
*** vilobhmm has quit IRC22:05
anteayamy postion remains the same22:06
anteayaregardless of what patch you are looking at22:06
zaroanteaya: i think openstack has a very specific gerrit workflow.  think it's even documented in infra manual?22:06
krtayloranteaya, I don't understand your position, that is why I am asking22:06
anteayazaro: yes in the infra manual22:06
anteayazaro: I didn't know we had youtube videos22:06
anteayakrtaylor: that is a real shame22:06
fungizaro: there exists video evidence on the internet of me giving presentations about openstack things, but i will not link them because i am not very confident in my public speaking abilities (working to improve them though!)22:06
anteayasince that is the sole point I had hoped you would understand22:07
anteayafor the entire time I spent supporting you22:07
fungiif either of my proposals for oscon get accepted i'll be forced to improve ;)22:07
anteayafungi: yay22:07
anteayafungi: I hope your proposals get accepted22:07
*** dhellmann_ has quit IRC22:08
anteayafungi: happy to support your public speaking work22:08
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Lazy load plugin list
*** denis_makogon has quit IRC22:08
krtayloranteaya, I am fairly intelligent, try me again :)  Do you not support me anymore? Now I really don't understand...22:08
krtayloranteaya, you want less requirements?22:08
krtayloranteaya, or to increase the learning/involvement of third party systems?22:09
openstackgerritGuido Günther proposed stackforge/python-jenkins: Don't throw exception on empty response
openstackgerritGuido Günther proposed stackforge/python-jenkins: Don't assume we get a valid response
*** joesavak has quit IRC22:09
* zaro suggest waynr viralize a vid of fungi, he's good.22:10
*** _nadya_ has joined #openstack-infra22:10
anteayakrtaylor: take some time and think about it22:11
*** gondoi is now known as zz_gondoi22:11
*** zz_gondoi is now known as gondoi22:12
krtayloranteaya, well, I'm not sure thinking about it will change my desire to help third party systems get better22:14
*** kgiusti has left #openstack-infra22:15
krtayloranteaya,  we are not perfect, by any means, but I don't see how pointing that out changes my desire to make things better22:15
anteayahow are you improving things for devs22:16
*** tkelsey has joined #openstack-infra22:16
anteayaI don't see any improvement for devs in anything you are doing22:16
anteayasome folks are doing work that helps devs22:16
krtayloranteaya, we are devs22:16
anteayaI jsut don't see any from you22:16
anteayawho do you mean when you use 'we'?22:16
*** mbacchi has quit IRC22:17
krtayloranteaya, we, my team, other external CI teams, we are all devs22:18
krtaylorwe learned that in Paris22:18
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Allow to wait for jenkins to start up
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Lazy load plugin list
*** _nadya_ has quit IRC22:18
krtaylorbut I still don't see how this has anything to do with the patches to improved the third-party documentation22:18
anteayaand I'm tired of trying to help you understand22:19
anteayasince nothing I do seems to make a dent22:20
anteayayeah wow22:20
anteayaI've been saying that myself for some time22:20
krtaylora dent in what?22:20
*** tkelsey has quit IRC22:20
*** dhellmann_ has joined #openstack-infra22:20
*** gondoi is now known as zz_gondoi22:20
krtayloranteaya, whatever, if you want to discuss the patches, great, but I'm not going to discuss how I don't get it22:21
anteayagreat, I gave up a long time ago22:21
*** zz_gondoi is now known as gondoi22:21
*** mjturek1 has quit IRC22:22
openstackgerritMerged openstack-infra/jenkins-job-builder: Implement recursive search for test fixtures.
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth check for no response_type
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth check for no client_id
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth check for response_type 'code'
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth validates redirect_uri
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth Authorize Scope Validation
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth tests for authorize_return
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth Exception on invalid oauth return data
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: OAuth checks for remote data access permissions.
*** dizquierdo has quit IRC22:24
*** salv-orlando has quit IRC22:25
*** abhirc has quit IRC22:25
openstackgerritJoshua Harlow proposed openstack/requirements: Add newly released oslo debtcollector library
*** andreaf has quit IRC22:26
*** ddieterly has joined #openstack-infra22:27
NobodyCamanyone have a second to look at a patch to unblock the Ironic stable gate?
openstackgerritMichael Krotscheck proposed openstack-infra/storyboard: Remaining functional tests for the OAuth Endpoint
krtaylorjeblair, thanks for the comment, that makes sense, I'll update the patch22:27
*** abhirc has joined #openstack-infra22:27
jeblairmarun: oh, sorry i have no idea why i was so intellectually lazy when i said "normal scenario testing".  I believe i meant "using the testscenarios module".  i suppose i thought that's "normal" for the kind of multiple-scenario testing22:29
clarkbjogo: it passed again with multihost fix22:29
jeblairmarun: i don't think zuul has any of that right now, so i was wondering if you established a reason it wouldn't work, or just didn't do that because of the lack of an existing pattern22:29
jeblairmarun: regardless, it might just be a candidate for future improvement :)22:30
clarkbjogo: I have been looking at iptables and ebtables counters and nothing really stands out to me as to why the packets are going away22:30
jeblairkrtaylor: cool, thanks22:30
*** salv-orlando has joined #openstack-infra22:30
marunjeblair: I've used testscenarios in neutron but I tend not to add it to a project if it doesn't exist there already.22:30
*** andreaf has joined #openstack-infra22:31
krtaylorjeblair, I was just looking for something to call ourselves, but this document may not be the right place22:31
marunjeblair: Its style of generative testing complicates debugging and is not compatible with tools like nose, after all.22:31
jeblairmarun: ok, got it.  ftr i think it's not it zuul because i didn't know about it when i wrote most of the tests, and haven't written a test that might use it since i learned about it.22:31
marunjeblair: the original test - with all those sets of parameters - was certainly cost to the model that testscenarios encourages.22:32
marunjeblair: After some experience with testscenarios, I'm not sure it's preferable to individual test methods that pass parameters to a helper.22:32
maruncost -> close22:32
*** gondoi is now known as zz_gondoi22:34
jeblairmarun: oh interesting.  i do like to run one test in the foreground when developing new features, so inasumch as testscenarios might complicate that, it might be a reason to avoid it...  but i don't have strong feelings about it in zuul though and think it might be worth a try for something.22:34
jeblairmarun: but i see you've changed the test (i'm still reading the old patchset), so it's probably moot at this point :)22:34
*** mriedem is now known as mriedem_zz22:34
clarkbjeblair: you should be able to do that with testscenarios, just provider a test name filter and it should only run that one test22:34
marunjeblair: my take is that testscenarios is more focused on writing/running test suites than debugging individual tests.  I have a similar complaints about other parts of the testr toolset.22:35
jeblairclarkb: okay, well, maybe at some point we can refactor these, though it's only two scenarios now, so, meh.22:35
marunclarkb: it doesn't play nice with editor integration, though22:36
jeblairmarun: you'll find that zuul's TESTING file has a lot of long commands that remind me how to do things i want to do with testr22:36
marunclarkb: I have emacs helpers that automatically execute a test that my cursor is on22:36
clarkbmarun: I don't use editor integration beyond :!tox22:36
marunclarkb: your loss :)22:36
mordredI have an xterm with bash in that I use to run things22:36
clarkbmordred: ya thats my typical workflow22:37
mordredI tend to use editors for editing code ... but I'm probably strange22:37
jeblairmordred: that's what i have, but marun has an interesting point....22:37
mordredoh - not disregarding - sorry22:37
jeblairthe subunit stream is pretty cool22:37
mordredjust being me and saying words22:37
jeblairand actually elisp to parse that and do something intelligent would be nifty22:37
clarkbmy guess is you could get the editor to run all the tests with the scenario prefix pretty easily22:37
mordredjeblair: +10022:37
clarkbbut more difficult to do specific scenarios22:37
marunjeblair: it exists already in the form of nose.el22:37
clarkbmarun: does it run with the prefix? oh using nose22:38
marunjeblair: I have a version I wrote myself before I knew about it22:38
mordredwell, but then that's nose22:38
marunjeblair: but the trick is22:38
*** ddieterly has quit IRC22:38
marunjeblair: all you really need is its discovery of path/module/method22:38
marunjeblair: the commands you invoke can be overriden22:38
mordredmarun: does nose support the discovery protocol now?22:38
marunjeblair: so I use nose2 or depending22:38
marunmordred: nose2 does22:38
*** ddieterly has joined #openstack-infra22:39
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Allow to wait for jenkins to start up
openstackgerritGuido Günther proposed openstack-infra/jenkins-job-builder: Lazy load plugin list
mordredmarun: even for things that override the method like testscenarios?22:39
marunmordred: nosetests, as much as I love it, is basically abandoned now.  nose2 is not as good but frequently better than testtools22:39
*** ddieterly has quit IRC22:39
*** ddieterly has joined #openstack-infra22:40
mordredthat was one of the gotchas originally - it ignored the full unittest discovery protocol - mainly the custom discovery hook bits22:40
marunmordred: nose2 is compatible. I updated the neutron docs recently to replace mention of nosetests with nose2 to avoid having to explain to people how to hack around the nose incompatibility.22:40
mordredthat's great to know!22:40
marunmordred: but I'm still of the opinion that testscenarios is a step backwards.22:41
mordredI actually like the idea of people being able to run nose2 if they want to but still be able to use testr in the gate22:41
*** dannywil_ has quit IRC22:41
mordredit's the _right_ kind of freedom22:41
marunmordred: agreed22:41
clarkbhopefuly nose2 fixes the py26 issues though22:41
mordredmarun: I like testscenarios some times... there are times when it makes making a suite of combinations so much easier22:41
clarkbgranted py26 is almost gone so maybe a non issue22:41
jeblairclarkb: (though we won't care for much longer)22:41
marunclarkb: people still use py26? ;)22:41
mordredmarun: only us22:41
mordredbut then there are times when I agree using is actually more mental gymanstics22:42
marunmordred: testscenarios is good for trivial testing that won't need troubleshooting, but otherwise...22:42
mordredyah -for blowing out tons and tons of combinations of input and output params22:42
marunmordred: that said, it should be possible to hack something together in elisp that can ask for enough info to generate a test name to execute22:43
*** dannywilson has joined #openstack-infra22:43
marunmordred: I should probably do that, because as much as I have mixed feelings about testscenarios I have a feeling its use is only going to grow.22:43
*** signed8b_ has quit IRC22:44
marunjeblair: do you run tests in emacs?22:45
jeblairmarun: no, just in a terminal.  i edit in emacs though.22:45
jeblairand read mail22:45
SpamapSmordred: ok, so trying your suggestion from earlier..22:45
SpamapSdib-run-parts Thu Feb 5 13:57:59 PST 2015 Running /tmp/image.OFUVr2pi/hooks/extra-data.d/55-cache-devstack-repos22:45
SpamapSCan not determine RELEASE22:45
*** radez is now known as radez_g0n322:46
marunjeblair: I really like being able to run tests in emacs. Running under pdb mode traces execution through the actual files.22:47
SpamapSah need DIB_RELEASE=trusty22:47
*** dannywilson has quit IRC22:47
*** dannywil_ has joined #openstack-infra22:47
SpamapSthough I'd have thought dib would do that.22:47
marunjeblair: Also, when you have post-mortem debugging enabled (we have this as a configurable option in the neutron test suite), it will drop into the environment of the error automatically on failure.22:48
greghaynesYea, I thought that was one of our 4 ways of determining distro22:48
greghaynesSpamapS: Youre making devstack + dib?22:48
SpamapSgreghaynes: I"m looking at why doing mysql things inside dib causes issues.22:48
*** timcline has quit IRC22:49
*** zz_gondoi is now known as gondoi22:50
*** thedodd has quit IRC22:53
*** thedodd has joined #openstack-infra22:53
*** hdd has quit IRC22:53
*** hdd has joined #openstack-infra22:54
*** dimtruck is now known as zz_dimtruck22:55
*** annegent_ has quit IRC22:55
*** amrith is now known as _amrith_22:57
*** dims_ has quit IRC22:58
*** dims__ has joined #openstack-infra22:59
*** dims__ has quit IRC22:59
*** amitgandhinz has quit IRC22:59
*** andreaf has quit IRC23:00
*** sdake_ has quit IRC23:01
*** gondoi is now known as zz_gondoi23:01
*** spzala has quit IRC23:02
*** dims__ has joined #openstack-infra23:03
openstackgerritKurt Taylor proposed openstack-infra/system-config: Rewrite for Third-party CI documentation Requirements section
*** ddieterly has quit IRC23:05
openstackgerritKurt Taylor proposed openstack-infra/system-config: Add requirement to get approval to post comments
*** abhirc has quit IRC23:07
*** ddieterly has joined #openstack-infra23:07
*** andreaf has joined #openstack-infra23:09
*** hichihara has joined #openstack-infra23:10
*** ashleighfarnham has quit IRC23:11
*** sarob has quit IRC23:11
clarkbfungi: where did that d-g change to fix the branch and repo origin stuff go?23:11
clarkbhrm looks like gertty may have stopped syncing for me23:12
clarkb^R makes it go23:12
*** zul has joined #openstack-infra23:12
*** dims__ has quit IRC23:13
*** dims__ has joined #openstack-infra23:14
fungiclarkb: ?23:14
*** zz_dimtruck is now known as dimtruck23:14
*** abhirc has joined #openstack-infra23:15
fungiclarkb: andreaf_ hasn't followed up to my response there yet and is still -1 since it will cause 'splosions in hpcloud's test rig23:15
clarkbya that one. It is/wasn't showing up in gertty for me but I manually started a resync so hopefully it arrives soon23:15
clarkbmtreinish: in 153036, your use of subunit-1to2 isn't quite right unless subunit-1to2 has an undocumented feature of taking a inpute file as a command arg23:16
clarkbmtreinish: am I just crazy or does that work?23:16
*** Rockyg has quit IRC23:16
mtreinishclarkb: I though it worked when I tested it, but you're probably right23:18
mtreinishlet me double check23:18
openstackgerritEmilien Macchi proposed openstack-infra/project-config: Jenkins/Puppet: increase timeout to 60 minutes
clarkbmtreinish: I always do subunit-1to2 < input > output23:18
EmilienMcrinkle: ^23:18
*** dkliban is now known as dkliban_afk23:18
clarkbmtreinish: but that may just be me never checking if the undocumented input arg works :)23:18
clarkbfungi: yup gertty synced it. Odd that it stopped syncing through23:19
*** sdake_ has joined #openstack-infra23:20
mtreinishclarkb: it seems to work:
clarkbmtreinish: thanks for checking that'll teach me to trust documentation23:21
*** ashleighfarnham has joined #openstack-infra23:21
clarkbmtreinish: also your terminal prompt looks scary23:21
mtreinishclarkb: it's all colorful too :)23:21
mtreinishI was bored one day freshman year and looked up all the fancy things you can do with zsh prompts and came up with that23:22
openstackgerritEmilien Macchi proposed openstack-infra/project-config: Jenkins/Puppet: increase timeout to 60 minutes
*** chlong has joined #openstack-infra23:23
*** zul has quit IRC23:24
*** dimtruck is now known as zz_dimtruck23:24
*** abhirc has quit IRC23:24
*** alexpilotti has quit IRC23:25
*** pc_m has quit IRC23:29
openstackgerritMerged openstack-infra/puppet-meetbot: Add Bugtracker plugin to meetbot supybot
*** dustins has quit IRC23:30
fungii'll restart the meetbot if necessary when ^ makes it onto eavesdrop in a few minutes23:30
*** ZZelle_ has quit IRC23:31
jeblairmorganfainberg: ^23:32
fungii already checked and there are no meetings running, so i have a little time to troubleshoot and roll back if it's devastating to the bot somehow23:33
clarkbI tried to get morganfainberg to confirm the config for the plugin was tested but never hread back fwiw23:33
*** EmilienM is now known as EmilienM|afk23:33
morganfainbergi thought i commented on that23:34
fungiyeah, that guy is never around23:34
fungioh, HI morganfainberg23:34
morganfainbergi tested those config options on my local install that is running for the keystone channel and triple-o channel atm23:34
morganfainbergseems to work correctly23:34
morganfainbergi mean i copy/pasted in the config options to the version from the bzr repo23:34
morganfainbergi can't speak to the puppet side, but nibalizer seemed to be happy with it [and the bzr provider looked to work like that afaict]23:35
*** annegent_ has joined #openstack-infra23:35
clarkbmorganfainberg: ya I was mostly interested in the bot config itself23:35
morganfainbergclarkb, yeah that config looked to work as expected23:35
fungiyeah, the change looked fine and the config seemed to look like what i found in the meetbot source23:36
morganfainbergif you want to poke the one in keystone channel its the same thing with one extra true for heat / assignment enabled23:36
morganfainbergwhich i felt was kinda silly for the generic bot to do23:36
morganfainbergs/heat/bug heat from LP/23:36
morganfainbergso the only question is... when does the meetbot restart so we can see the fun everywehre again? :)23:37
*** zul has joined #openstack-infra23:37
*** zul has quit IRC23:37
*** pradk has quit IRC23:38
*** prad is now known as prad-afk23:38
*** esmute has quit IRC23:39
openstackgerritMerged openstack-infra/jenkins-job-builder: Add test for name defaults overridden by template
*** fabo has quit IRC23:40
openstackgerritMerged openstack-infra/jenkins-job-builder: Interpolate params inside of params.
*** annegent_ has quit IRC23:41
*** esmute has joined #openstack-infra23:41
*** bswartz has joined #openstack-infra23:41
*** david-lyle is now known as david-lyle_afk23:41
*** zz_dimtruck is now known as dimtruck23:43
*** shashankhegde has quit IRC23:44
*** gordc has quit IRC23:45
*** baoli has quit IRC23:46
*** shashankhegde has joined #openstack-infra23:46
*** Sukhdev has joined #openstack-infra23:47
*** fabo has joined #openstack-infra23:47
*** fabo has joined #openstack-infra23:47
*** rm_work is now known as rm_work|away23:47
*** zul has joined #openstack-infra23:47
*** shashankhegde has quit IRC23:49
*** carl_baldwin has quit IRC23:51
*** bdpayne has quit IRC23:54
*** zz_gondoi is now known as gondoi23:54
*** vigneshvar has quit IRC23:54
*** mfink_ has quit IRC23:55
*** mikeit has joined #openstack-infra23:55
anteayafor the record I voted23:55
anteayaCanada probably has the lowest population of lizards of any country I know23:56
openstackgerritK Jonathan Harker proposed openstack-infra/zuul: Add a remote url override location
*** ddieterly has quit IRC23:56
*** openstack has joined #openstack-infra23:58
*** carl_baldwin has joined #openstack-infra23:58
clarkbfungi: looks like puppte just started and I had to ssh over ipv4 not 623:58
fungiError: /Vcsrepo[/opt/ubuntu_supybot_plugins]: Could not evaluate: undefined method `latest?' for #<Puppet::Type::Vcsrepo::ProviderBzr:0x7f2bdd676d48>23:58
*** mikeit has quit IRC23:58
SpamapSwow my devstack-cache element has been going for an hour :-P23:58
fungithat was me running it noop23:58
SpamapSI remember when I thought 30Mbit was fast enough :-P23:58
SpamapSthat was about 65 minutes ago23:58
*** Sukhdev has quit IRC23:58
*** dannywil_ has quit IRC23:58
fungiso i guess our first foray into the vcsrepo bzr provider is an abject failure23:59

Generated by 2.14.0 by Marius Gedminas - find it at!