Tuesday, 2013-09-03

*** michchap has quit IRC00:01
*** michchap has joined #openstack-infra00:01
*** michchap has quit IRC00:13
*** jhesketh__ has joined #openstack-infra00:14
*** primeministerp has joined #openstack-infra00:14
*** michchap has joined #openstack-infra00:14
*** sgviking has joined #openstack-infra00:23
*** CaptTofu has quit IRC00:27
*** CaptTofu has joined #openstack-infra00:27
*** derekh has quit IRC00:39
*** pcm__ has quit IRC00:42
*** stevebaker has left #openstack-infra00:44
*** dina_belova has joined #openstack-infra00:44
*** dina_belova has quit IRC00:49
*** stevebaker has joined #openstack-infra00:53
*** zul has quit IRC00:57
*** zul has joined #openstack-infra00:57
*** markmcclain has quit IRC01:01
*** CaptTofu has quit IRC01:11
*** CaptTofu has joined #openstack-infra01:11
*** CaptTofu has quit IRC01:13
*** mikal has quit IRC01:13
*** CaptTofu has joined #openstack-infra01:14
*** mikal has joined #openstack-infra01:14
*** yaguang has joined #openstack-infra01:26
*** dina_belova has joined #openstack-infra01:45
*** dina_belova has quit IRC01:49
*** marun has quit IRC01:55
*** nati_ueno has quit IRC02:09
*** nati_ueno has joined #openstack-infra02:09
*** nati_ueno has joined #openstack-infra02:10
*** pabelanger has joined #openstack-infra02:13
*** rfolco has joined #openstack-infra02:20
*** xchu has joined #openstack-infra02:21
*** CaptTofu has quit IRC02:29
*** nati_ueno has quit IRC02:30
*** CaptTofu has joined #openstack-infra02:30
*** nati_ueno has joined #openstack-infra02:30
*** vogxn has joined #openstack-infra02:33
*** nati_ueno has quit IRC02:35
*** dina_belova has joined #openstack-infra02:45
*** anteaya has quit IRC02:47
*** rfolco has quit IRC02:48
*** dina_belova has quit IRC02:50
*** amotoki has joined #openstack-infra02:50
*** vogxn has quit IRC02:52
*** oubiwann has quit IRC02:59
*** garyk1 has joined #openstack-infra02:59
*** nati_ueno has joined #openstack-infra03:00
*** garyk has quit IRC03:00
*** oubiwann has joined #openstack-infra03:02
*** oubiwann is now known as Guest6801503:02
*** nati_uen_ has joined #openstack-infra03:06
*** nati_uen_ has joined #openstack-infra03:06
*** garyk1 has quit IRC03:08
*** nati_ueno has quit IRC03:09
*** garyk has joined #openstack-infra03:18
*** rfolco has joined #openstack-infra03:26
*** nati_ueno has joined #openstack-infra03:29
*** dims has quit IRC03:36
lifelessclarkb: / jeblair: where do you get the rax dns client from?03:40
*** SergeyLukjanov has joined #openstack-infra03:40
lifelessclarkb: / jeblair: also, what size instance is gerrit? And does everyone need that size?03:41
*** pabelanger has quit IRC03:43
*** dina_belova has joined #openstack-infra03:46
*** UtahDave has joined #openstack-infra03:49
*** marun has joined #openstack-infra03:50
*** dina_belova has quit IRC03:50
*** nati_ue__ has joined #openstack-infra03:54
*** nati_uen_ has quit IRC03:58
*** pabelanger has joined #openstack-infra04:00
*** yaguang has quit IRC04:03
*** Ryan_Lane has quit IRC04:14
*** Ryan_Lane has joined #openstack-infra04:14
*** yaguang has joined #openstack-infra04:16
*** jhesketh__ has quit IRC04:27
*** garyk has quit IRC04:38
*** vogxn has joined #openstack-infra04:46
*** dina_belova has joined #openstack-infra04:47
*** dina_belova has quit IRC04:51
*** pabelanger has quit IRC04:53
*** boris-42 has joined #openstack-infra04:54
*** jhesketh__ has joined #openstack-infra04:57
morganfainbergclarkb: ping05:08
morganfainbergah, probably a bit late, i'll try and hit ya up tomorrow.05:08
*** nicedice_ has quit IRC05:19
*** SergeyLukjanov has quit IRC05:19
*** pblaho has joined #openstack-infra05:43
* mordred is back in the world of internet05:50
mordredlifeless: instance size of gerrit is the largest instance that rax has - I think 64G05:51
mordredlifeless: it is not required to be that size - but if you wanted to run a gerrit that sees our kind of traffic you'd probably want one05:51
lifelessmordred: welcome back05:51
mordredlifeless: thanks!05:51
lifelessmordred: I have many questions if/when you have a time slice05:52
lifelessmordred: (after C goes to bed)05:52
Alex_Gaynormordred: welcome back05:52
mordredlifeless: I'm about to run down for a beer - I'm not realy here05:52
mordredthanks Alex_Gaynor !05:52
lifelessmordred: thats cool; I wouldn't want to interfere with beer!05:52
mordredlifeless: beer is important05:52
mordredlifeless: there will be a stewart and a devananda there too :)05:52
*** nati_ue__ has quit IRC05:56
*** garyk has joined #openstack-infra06:08
*** nati_ueno has quit IRC06:53
*** nati_ueno has joined #openstack-infra06:53
*** tian has quit IRC06:56
*** nati_ueno has quit IRC06:57
*** nati_ueno has joined #openstack-infra07:17
*** fbo_away is now known as fbo07:17
*** Bada has joined #openstack-infra07:19
*** boris-42 has quit IRC07:23
*** odyssey4me has joined #openstack-infra07:25
*** vogxn has quit IRC07:27
*** shardy_afk is now known as shardy07:30
*** pblaho has quit IRC07:33
*** dina_belova has joined #openstack-infra07:33
*** SergeyLukjanov has joined #openstack-infra07:33
*** dizquierdo has joined #openstack-infra07:37
*** jpich has joined #openstack-infra07:42
*** hashar has joined #openstack-infra07:47
*** vogxn has joined #openstack-infra07:57
*** sgviking has quit IRC08:01
*** nati_ueno has quit IRC08:05
*** nati_ueno has joined #openstack-infra08:05
*** nati_ueno has quit IRC08:09
*** jhesketh__ has quit IRC08:15
*** sgviking has joined #openstack-infra08:16
*** ruhe has joined #openstack-infra08:19
*** derekh has joined #openstack-infra08:21
*** nati_ueno has joined #openstack-infra08:22
*** pblaho has joined #openstack-infra08:34
*** fifieldt has quit IRC08:35
*** SergeyLukjanov has quit IRC08:36
*** vogxn has quit IRC08:40
*** vogxn has joined #openstack-infra08:41
*** SergeyLukjanov has joined #openstack-infra08:41
*** boris-42 has joined #openstack-infra08:50
*** mkerrin has joined #openstack-infra08:51
*** dina_belova has quit IRC09:06
*** nati_ueno has quit IRC09:07
*** nati_ueno has joined #openstack-infra09:08
*** nati_ueno has quit IRC09:12
*** nati_ueno has joined #openstack-infra09:19
*** nati_ueno has quit IRC09:34
*** nati_ueno has joined #openstack-infra09:35
*** alexpilotti has joined #openstack-infra09:39
*** vogxn has quit IRC09:43
*** blamar has quit IRC09:45
*** dina_belova has joined #openstack-infra09:45
*** mancdaz_ has quit IRC09:46
*** mancdaz has joined #openstack-infra09:47
*** jkt has left #openstack-infra09:48
*** thomasbiege has joined #openstack-infra09:49
*** hashar_ has joined #openstack-infra09:52
*** hashar has quit IRC09:55
*** hashar_ is now known as hashar09:55
*** thomasbiege has quit IRC09:56
*** ruhe has quit IRC09:58
*** vogxn has joined #openstack-infra09:59
*** ruhe has joined #openstack-infra10:06
*** xchu has quit IRC10:13
*** pcm_ has joined #openstack-infra10:18
*** ruhe has quit IRC10:18
*** pcm_ has quit IRC10:18
*** pcm_ has joined #openstack-infra10:19
*** SergeyLukjanov has quit IRC10:21
*** michchap has quit IRC10:27
*** sdague has joined #openstack-infra10:32
*** yaguang has quit IRC10:33
*** UtahDave has quit IRC10:33
openstackgerritA change was merged to openstack-dev/hacking: Use the name for this attribute that exists on Python3  https://review.openstack.org/4451210:34
*** SergeyLukjanov has joined #openstack-infra10:34
*** hashar has quit IRC10:35
*** SergeyLukjanov has quit IRC10:54
*** ruhe has joined #openstack-infra10:54
*** michchap has joined #openstack-infra10:58
*** SergeyLukjanov has joined #openstack-infra11:02
*** boris-42 has quit IRC11:04
*** michchap has quit IRC11:06
*** CaptTofu has quit IRC11:16
*** CaptTofu has joined #openstack-infra11:17
*** ruhe has quit IRC11:18
*** seagulls has quit IRC11:28
*** beagles has joined #openstack-infra11:28
*** vogxn has quit IRC11:31
*** ruhe has joined #openstack-infra11:32
*** dims has joined #openstack-infra11:37
*** jhesketh_ has joined #openstack-infra11:46
*** dims has quit IRC11:49
*** anteaya has joined #openstack-infra11:53
*** yaguang has joined #openstack-infra11:54
*** hashar has joined #openstack-infra11:54
*** HenryG has quit IRC11:57
*** dims has joined #openstack-infra12:02
*** weshay has joined #openstack-infra12:06
*** SergeyLukjanov has quit IRC12:09
*** SergeyLukjanov has joined #openstack-infra12:09
*** dkranz has quit IRC12:10
*** dina_belova has quit IRC12:17
*** Mikhail has quit IRC12:18
anteayathe gate is empty12:21
*** dprince has joined #openstack-infra12:22
*** markmcclain has joined #openstack-infra12:23
BobBallmordred: What's the plan for the virtualenv patch?12:24
*** markmcclain has quit IRC12:24
*** sgviking has quit IRC12:25
*** dina_belova has joined #openstack-infra12:31
*** whayutin_ has joined #openstack-infra12:32
*** yaguang has quit IRC12:35
*** weshay has quit IRC12:35
ttxanteaya: this will not last forever.12:36
anteayahey ttx, yes it will not12:36
*** lcestari has joined #openstack-infra12:36
anteayabut zuul and friends have been putting away a lot of patches12:37
anteayathe speed at which they chew through them is impressive12:37
anteayattx, ever the optimist12:38
ttxyeah, looks like the parallel tests move was an improvement overall12:38
anteayais seemed like the pin for so many other changes, nodepool, haproxy for git12:39
anteayaand the zuul decision changes around merging are less evident but they help12:40
*** annegentle_away is now known as annegentle12:46
*** blamar has joined #openstack-infra12:52
*** thomasm has joined #openstack-infra12:57
*** dkehn_away is now known as dkehn13:10
hasharEmilienM: hey there :-]13:11
hasharEmilienM: I noticed a blog post you made on eNovance blogs regarding Jenkins Job Builder http://techs.enovance.com/6006/manage-jenkins-jobs-with-yaml  . Was wondering how you draw the nice pictures :-]13:11
*** lbragstad has quit IRC13:11
EmilienMhashar: hello13:14
EmilienMhashar: http://www.websequencediagrams.com13:14
hasharhence the gnome-screenshot metadata in the .png file13:15
EmilienMhashar: I don't want to pay to download a picture :)13:15
hasharthe 1990's version would be http://www.asciiflow.com/13:15
*** HenryG has joined #openstack-infra13:16
EmilienMhashar: nice one too :)13:16
EmilienMhashar: could you review 2 patchs on JJB ? > https://review.openstack.org/#/c/44032/ and https://review.openstack.org/#/c/43685/13:17
*** ArxCruz has joined #openstack-infra13:17
hasharyou can get them rendered as png with ditaa :D13:17
hasharah I am lagging with jjb development :/13:17
*** whayutin_ has quit IRC13:17
*** adalbas has quit IRC13:17
*** adalbas has joined #openstack-infra13:18
EmilienMok, nw13:19
hasharEmilienM: I will try out the ircbot change13:19
*** michchap has joined #openstack-infra13:20
hasharEmilienM: for sequence diagram, pypi has the 'seqdiag' package : http://blockdiag.com/en/seqdiag/index.html  :-]13:20
EmilienMhashar: thx for review !13:20
*** ruhe has quit IRC13:20
fungialso for more embedded use there's http://bramp.github.io/js-sequence-diagrams/ which is rather neat13:22
anteayahey fungi13:23
anteayawelcome back13:23
anteayahow are you feeling?13:23
fungianteaya: recovering from hives, a respiratory infection and 4 time zones of jet lag all at the same time13:24
fungipretty good, and awake, all things considered13:24
sdagueooof, hope you get better soon13:24
anteayawell I am glad you are back, but only take the priority stuff, leave what you can to me13:25
anteayahey sdague welcome back to you too13:25
fungii'm mostly recovered thanks to the long weekend. just staring at the lingering remains of two weeks of e-mail i spent the past several days digging through13:25
anteayahow was the yurt?13:25
sdagueanteaya: thanks... and very good :)13:25
anteayafungi: ah email, what fun13:25
anteayasdague: yay13:25
anteayabet you have a bit of email too, yeah?13:26
sdagueanteaya: http://dague.net/2013/09/01/vacation/ has some pictures embedded13:26
sdagueanteaya: yeh, I need to get jog0's email cleaner script working before I tackle the review folders13:26
fungiahh, neat. sdague and i were simultaneously venturing up opposite canadian coasts13:27
anteayasdague: ha ha ha13:27
sdaguefungi: awesome :)13:27
fungieastern alaska was also nifty13:27
anteayaoh the yurt looks awesome13:28
anteayafungi: alaska is so beautiful, was that the first time you have been?13:28
fungilots of glaciers there, at least for a little longer13:29
* anteaya raises her hand for being in a future yurt group13:30
anteayafungi were you sick on the boat? or was the boat okay for you?13:30
*** whayutin_ has joined #openstack-infra13:30
* fungi doesn't get seasick, thankfully... even on small craft in rough seas for extended periods13:30
*** ruhe has joined #openstack-infra13:31
fungii could probably live on a boat fairly happily13:31
anteayaI'm glad the boat part was enjoyable13:32
anteayaI too could spend quite a bit of time on a boat13:32
*** ruhe has quit IRC13:32
*** sandywalsh has joined #openstack-infra13:32
*** thomasm has quit IRC13:33
*** mgagne has joined #openstack-infra13:34
*** thomasm has joined #openstack-infra13:36
*** changbl has quit IRC13:43
*** ruhe has joined #openstack-infra13:43
*** dkliban has joined #openstack-infra13:44
*** dina_belova has quit IRC13:44
*** dkliban has quit IRC13:45
*** dkliban has joined #openstack-infra13:45
*** dina_belova has joined #openstack-infra13:46
*** dina_belova has quit IRC13:46
*** lbragstad has joined #openstack-infra13:47
*** markmcclain has joined #openstack-infra13:48
*** boris-42 has joined #openstack-infra13:49
*** prad has joined #openstack-infra13:52
*** burt has joined #openstack-infra13:53
*** ruhe has quit IRC13:54
hasharEmilienM: I have reviewed your IRC publisher for jbb ( https://review.openstack.org/#/c/44032/ ) :-]13:56
*** yaguang has joined #openstack-infra14:02
*** mrodden has joined #openstack-infra14:07
*** vogxn has joined #openstack-infra14:09
*** jjmb has joined #openstack-infra14:09
*** dina_belova has joined #openstack-infra14:12
*** pblaho has quit IRC14:14
*** legend303 has joined #openstack-infra14:19
*** legend303 has left #openstack-infra14:20
BobBallWho can I talk to about what the HP cloud that the CI uses provides?14:27
anteayahey BobBall, fungi is back and might know that one14:27
anteayahe is catching up on email back log and should be here soon14:28
BobBallfungi: Does HP cloud support booting from ISO or importing your own image (i.e. I want something uber-custom), and how does it provide the IP? configdrive?14:28
*** cody-somerville has joined #openstack-infra14:28
*** dkranz has joined #openstack-infra14:28
*** ruhe has joined #openstack-infra14:29
EmilienMhashar: thx14:29
*** pblaho has joined #openstack-infra14:30
EmilienMhashar: could you help me to understand what would you suggest for your first comment ? just a hint in the doc ? or something in the code ?14:32
hasharEmilienM: I should be more explicit probably :]14:33
EmilienMhashar: actually your review is very complete :) but I prefer to be clear14:33
mgagnehashar: looks clear to me =)14:33
hasharEmilienM: so just a hint/warning in the doc.14:33
EmilienMmgagne: o/14:33
EmilienMhashar: ok, thx14:34
hasharI don't think we have any way to fetch the Jenkins  system configuration from JJB14:34
hasharand even if we could, it is not really possible when testing the jobs (jjb test ..)14:34
mgagneEmilienM: \o14:34
hasharso I would just plant a huge warning sign along the way of "WARNING: the IRC plugin requires the channel to be configured in the system wide configuration or the jobs will fail to emit notifications to the channel".14:35
EmilienMhashar: sounds cool14:35
hasharEmilienM: comment 2 and 3 are merely complaining about duplication of dict keys14:35
*** datsun180b has joined #openstack-infra14:37
openstackgerritIlya Shakhat proposed a change to openstack-infra/config: Add doc build process into Stackalytics  https://review.openstack.org/4488714:39
EmilienMhashar: I would say nobody does that AFIK for comment 2 & 314:41
EmilienMbut ok, if you prefer...14:42
openstackgerritAlex Gaynor proposed a change to openstack/requirements: Unpin six now that WSME works with either version  https://review.openstack.org/4488914:42
*** changbl has joined #openstack-infra14:42
*** sdake has joined #openstack-infra14:43
*** sdake has quit IRC14:43
*** sdake has joined #openstack-infra14:43
*** flashgordon is now known as jog014:44
*** dkliban has quit IRC14:47
*** dkliban has joined #openstack-infra14:48
fungiBobBall: production hpcloud does not provide a public glance endpoint (though i gather it is in beta and will eventually be available), and has no other mechanisms i'm aware of to upload your own server images either14:48
fungifor infra/ci we currently use standard images they provide (ubuntu, centos) and then customize and snapshot them as needed in some cases14:49
*** zeus has joined #openstack-infra14:50
BobBallOk... That's a shame... No netbooting then?14:51
BobBallI wonder if I can hack one in by putting a netboot loader on the host14:52
funginot of which i'm aware at any rate14:52
BobBallor ohhhh14:52
BobBallhaha that'd be fun14:52
BobBallputting the XS installer as the kernel/initrd on the image itself14:52
fungiwell, yeah, so you could heavily customize an existing image to turn it into a stub ramdisk loader to bootstrap your network boot image or something, i'm sure14:52
BobBallthat _would_ be fun.14:52
fungior do something like tripleo's takeover-node where you dd the intended image to a second block device and then pivot-root/kexec reboot14:53
BobBallBut once you have an image you can save it and re-use it right?14:54
fungiyep, we do that pretty heavily for devstack-gate jenkins slaves14:54
fungithough i wouldn't call the customization we do particularly heavy. we just update/add some packages, set up accounts and other puppetry, pull down and cache some git repositories, et cetera14:55
*** whayutin_ is now known as weshay14:55
fungiso what we're snapshotting is still fairly close to the base image provided by hpcloud14:57
BobBallI'd want to do some pretty heavy manipulation :)14:57
*** SergeyLukjanov has quit IRC14:57
BobBallOh - the other question - how are IPs provided?  configdrive?14:57
fungivery good question... lemme see if i can figure that out14:58
*** tstevenson has joined #openstack-infra14:58
BobBalldprince: Interesting question on openstack-dev around Review 40396 being hit by torpedo with an empty Neutron DB14:58
*** briancline has joined #openstack-infra15:00
*** mrodden has quit IRC15:01
*** lcestari is now known as lcestari_lunch15:01
BobBallfungi: what instance size do you use as well?15:02
*** vogxn has quit IRC15:02
fungiBobBall: the 8gb flavor15:03
BobBallok, that'd work.15:03
*** pblaho has quit IRC15:04
*** SergeyLukjanov has joined #openstack-infra15:04
fungiBobBall: hpcloud, as i suspected, uses dhcp to set the server's ip address (fairly typical for an openstack provider)15:04
BobBallok - that'd work15:05
BobBalljust that RS cloud doesn't15:05
dprinceBobBall: looking...15:06
*** pentameter has joined #openstack-infra15:06
*** dina_belova has quit IRC15:07
openstackgerritRyan Petrello proposed a change to openstack-infra/config: Provide a more generic run-tox.sh.  https://review.openstack.org/4314515:10
dprinceBobBall: 40396 is a tempest review? (https://review.openstack.org/#/c/40396)15:11
dprinceBobBall: is that the right number?15:11
BobBallsorry, my typo15:11
*** dina_belova has joined #openstack-infra15:12
dprinceBobBall: That looks valid (i.e. it is reproducable) but I'm not sure why. I'll have to have a closer look.15:14
openstackgerritDaniel Izquierdo proposed a change to openstack-infra/config: Addition of the activity-board project to the OpenStack-infra environment  https://review.openstack.org/4405715:15
*** dina_belova has quit IRC15:16
*** vogxn has joined #openstack-infra15:16
*** ruhe has quit IRC15:16
BobBalldprince: the email explains it a little more - firestack is relying on the auto-generated schema which is explicitely removed by that patch15:17
BobBallso I guess the change is incompatible with SS and something needs to fix SS in some way so it has a sensible schema going forward... but I have no idea what that is.15:18
*** danger_fo_away is now known as danger_fo15:18
openstackgerritRyan Petrello proposed a change to openstack-infra/config: Provide a more generic run-tox.sh.  https://review.openstack.org/4314515:20
*** mrodden has joined #openstack-infra15:22
*** yaguang has quit IRC15:23
anteayadhellmann: why do we need a script to run tox, what is wrong with `tox`?15:24
*** vogxn has joined #openstack-infra15:25
*** vogxn has quit IRC15:26
openstackgerritSimon König proposed a change to openstack-infra/jenkins-job-builder: Added support to trigger a build on the same node using the parameterized trigger plugin  https://review.openstack.org/4490015:27
*** ruhe has joined #openstack-infra15:27
*** dina_belova has joined #openstack-infra15:27
*** vogxn has joined #openstack-infra15:28
*** michchap has quit IRC15:28
*** vogxn has quit IRC15:28
*** michchap has joined #openstack-infra15:28
*** thomasbiege has joined #openstack-infra15:33
*** jpmelos has joined #openstack-infra15:35
jpmelosHello. I'm trying to set up a Gerrit/Zuul/Jenkins stack, but I'm having problems. Apparently, Jenkins and Zuul lose connection randomly, and sometimes Zuul misses some patchsets and don't trigger the proper jobs...15:36
jpmelosAny hints on this?15:36
dkehnfungi, question for you: i've got a review https://review.openstack.org/#/c/44564/ that was dependent upon a review that has been merged, hence I can now pull from truck is this a simple pull netron, pull my patch and git review when done and the outdated dependent patch is gone?15:36
*** vogxn has joined #openstack-infra15:37
*** ryanpetrello has joined #openstack-infra15:37
*** vogxn has quit IRC15:37
*** gyee has joined #openstack-infra15:37
*** vogxn has joined #openstack-infra15:38
*** markmcclain has quit IRC15:38
*** UtahDave has joined #openstack-infra15:38
fungidkehn: yeah, that should normally take care of it15:39
fungidkehn: though it's not normally necessary unless you're going to encounter other merge conflicts with changes which landed after the parent change was merged15:39
anteayahi jpmelos there are a few folks doing what you want to be doing and lifeless is documenting his journey15:40
jpmelosanteaya: Thanks, I'll try to talk to him about this! Maybe we can help each other out.15:41
*** thomasbiege has quit IRC15:41
*** Bada has quit IRC15:41
fungijpmelos: since zuul usually watches gerrit's event stream to decide when to run jobs, it could be brief network connectivity interruptions between zuul and gerrit causing changes to be missed15:42
dkehnfungi, shouldnt be the case the ml2 is small, thanks15:42
*** rnirmal has joined #openstack-infra15:44
fungijpmelos: in general, the overall integration design is not completely robust against network interruptions, so while things are generally expected to recover back to an operable state on their own, there is the possibility that they miss events, results, et cetera15:44
anteayajpmelos: I'm sure he would appreciate the support and feedback15:44
jeblairlifeless, mordred: gerrit is rackspaces largest instance size, which is a 30GB vm, and no, that size is not required (though the jvm may be configured by our puppet scripts to use a significant portion of that ram, but that's tunable; review-dev uses a smaller value)15:48
anteayamorning jeblair15:49
jeblairgood morning15:49
anteayadid you have a nice weekend?15:49
anteayajpmelos: lifeless created this document a few days ago: http://ci.openstack.org/running-your-own.html I am sure he would like people to work their way through it and offer feedback in the form of patches if there is anything you think would help.15:53
*** reed has joined #openstack-infra15:54
*** mrmartin has joined #openstack-infra15:55
anteayamorning mrmartin16:02
anteayaclarkb hasn't said hello yet but he should be along shortly16:02
openstackgerritJeremy Stanley proposed a change to openstack-infra/config: Gerrit sysadmin tips for account merge and rename  https://review.openstack.org/4491216:03
*** portante is now known as portante|afk16:03
EmilienMhashar: what would you suggest with your last comment on http://goo.gl/YoLLQQ ? Sorry, maybe I'm not familiar enough16:03
hasharEmilienM: I am not sure whether we could detect the publisher is being used in a matrix job16:04
hasharEmilienM: we should at least support the three parameters available in matrix jobs: ALL ONLY_PARENT and ONLY_CONFIGURATIONS16:04
hasharwill follow up tomorrow, I am in a conf call right now16:04
mrmartinanteaya: thanks, I'll ping him16:04
*** michchap has quit IRC16:05
jeblairmrmartin: i think clarkb was suggesting that you have a script that runs 'tar; cp; chown', and then just exec that one script instead of having the 3 exec statements16:05
anteayahope you get what you need to keep your work moving along16:05
EmilienMhashar: no worry, thx16:05
*** hashar is now known as hasharCall16:05
BobBallfungi: Can you move an image from one acc to another?  I want to set up a XS image so that jenkins can use it - but trying to figure out the best way to do it.  Already proved it works with the RS cloud - although can't inject the IP there because they use their agent to set the networking16:06
fungiBobBall: i haven't seen an option for moving images between accounts or even between regions/zones in hpcloud16:06
fungiwe currently generate our snapshots separately in each zone we use16:07
jeblairBobBall: (you may know this, but just fyi:) we're only running devstack on hpcloud now, but we may run on rax (and others) again in the future.16:07
BobBall*nod* My point was I proved it on rax (except IP addressing doesn't work...) now I want to prove it on hpcloud - where I think it will work and will be useable by jenkins16:08
openstackgerritTobias Stevenson proposed a change to openstack-infra/gerritbot: Add better comment event reporting to Gerritbot  https://review.openstack.org/4491316:08
BobBallIs there a script to generate the snapshots? (i.e. is that what you'd expect us to do?) or can we somehow manually create an image that can then be used16:09
jeblairBobBall: excellent.  :)  http://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/nodepool/scripts16:09
jeblairBobBall: 'prepare_node_devstack.sh' is what gets run to build the snapshot image16:09
jeblairBobBall: so you'd probably want to add a 'prepare_node_xen_devstack' or something that does a superset of prepare_node_devstack16:10
jeblairBobBall: basically what nodepool does is copies those into /opt/nodepool-scripts and runs prepare_node_devstack.sh16:10
BobBallThat might be difficult without the ability to install custom images...16:11
*** lcestari_lunch is now known as lcestari16:11
BobBallI'll have to think16:11
jeblairBobBall: can't you script what you have done?16:11
BobBallwell - the setup I did for RAX depended on pxeboot for XS.  To get it into HP we'd have to boot a CentOS, replace kernel+initrd with XS installer, reboot and then go from there.  Very ugly to script :)16:12
BobBallOnce XS is installed then yes, it's trivial to script16:12
*** jergerber has joined #openstack-infra16:12
*** beagles has quit IRC16:13
jeblairBobBall: we want to replace the simple image-building stuff that nodepool does with diskimage builder eventually anyway...16:13
*** boris-42 has quit IRC16:14
jeblairBobBall: i'm not terribly familiar with it yet, but maybe lifeless or mordred or someone else could suggest whether that would make things any easier16:14
jeblair(both of whom are afk now)16:14
pleia2good morning16:16
pleia2anteaya: thanks for updating the dashboard bug16:17
anteayapleia2: morning, welcome16:17
anteayaany surprises for you in the update?16:17
anteayasorry, just starting to say welcome to everyone16:17
*** beagles has joined #openstack-infra16:17
* anteaya hangs up her walmart greeter vest16:18
*** pcrews has quit IRC16:18
anteayamorning clarkb16:18
mrmartinhi clarkb16:18
clarkbmrmartin: hello.16:19
clarkbmrmartin: I am suggesting either consolidating all of those execs into a single script that can be run so that the creates statement is true and not hard to follow, or documenting the behavior of creates in this particular case because it is slightly different than the typical use case16:19
mrmartinclarkb: could you check the inline comment for this review: https://review.openstack.org/#/c/42608/ ?16:19
mrmartinclarkb: ok16:20
mrmartinI try to do it today, or at least tomorrow, I also think it is cleaner16:20
*** dkehn_ has joined #openstack-infra16:21
*** svarnau has joined #openstack-infra16:21
* anteaya heads out for a walk16:21
*** hasharCall has left #openstack-infra16:22
clarkbmrmartin: I have responded in the gerrit review as well16:22
mrmartinclarkb: thanks16:22
*** dkehn has quit IRC16:22
clarkbmrmartin: typically exec resources will have a 1:1 with the thing they create. Eg exec tar would say creates => foo.tar. It is confusing when 6 exec resources claim to all create the same file16:22
mrmartinclarkb: it was used as a trigger, officially the best and cleanest solution is writing a fact provider, but I like to avoid this now16:24
*** hashar has joined #openstack-infra16:26
*** hashar has left #openstack-infra16:26
openstackgerritJames E. Blair proposed a change to openstack-infra/nodepool: Add a nodepool command  https://review.openstack.org/4455316:26
clarkbmrmartin: ya, avoiding writing a fact provider is probably a good thing. My other suggestion was to document the behavior which I would be fine with as well16:26
mrmartinclarkb: yeap, but this script rewrite will solve this16:27
*** nicedice_ has joined #openstack-infra16:28
openstackgerritA change was merged to openstack-infra/config: Add image logging config for nodepool  https://review.openstack.org/4437216:29
openstackgerritA change was merged to openstack-infra/nodepool: Add image logging  https://review.openstack.org/4437316:32
*** pblaho has joined #openstack-infra16:32
*** lbragstad has left #openstack-infra16:34
*** nati_ueno has joined #openstack-infra16:35
*** nati_ueno has quit IRC16:35
clarkbwow sdague, mordred and fungi are all back on the same day (ish)16:35
*** nati_ueno has joined #openstack-infra16:35
*** markmcclain has joined #openstack-infra16:36
*** Bada has joined #openstack-infra16:36
*** dkehn has joined #openstack-infra16:36
openstackgerritEmilien Macchi proposed a change to openstack-infra/jenkins-job-builder: Add IRCbot plugin support  https://review.openstack.org/4403216:37
*** dkehn_ has quit IRC16:37
*** dizquierdo has left #openstack-infra16:38
*** nati_ueno_2 has joined #openstack-infra16:38
*** zeus has quit IRC16:39
*** Ryan_Lane has quit IRC16:39
fungiclarkb: scary resumption of work, yes16:40
*** nati_ueno has quit IRC16:40
*** zeus has joined #openstack-infra16:40
clarkbmordred: there is new setuptools/pbr madness for when you are ready16:41
openstackgerritTobias Stevenson proposed a change to openstack-infra/gerritbot: Add better comment event reporting to Gerritbot  https://review.openstack.org/4491316:41
*** nati_ueno_2 has quit IRC16:43
*** moted has joined #openstack-infra16:43
fungifor the record (so nobody tries to go in and do it themselves), i've performed the gerrit account changes/merges requested earlier by ichavero, dukhlov and hartsocks16:44
fungithanks anteaya for keeping track of those pending requests16:45
*** dina_belova has quit IRC16:45
fungii e-mailed dukhlov and hartsocks to let them know16:45
fungiichavero is supposedly popping back into irc tomorrow to check on the request status anyway16:46
*** nati_ueno has joined #openstack-infra16:46
clarkbfungi: good to know thanks. In the case of hartsocks that was a merge right?16:47
fungiclarkb: yeah, https://review.openstack.org/44912 is a start at trying to document that sort of thing16:48
fungiif i find out from one of them (or future account merges) that the process is missing something i'll make sure to correct/update it accordingly16:49
*** hashar has joined #openstack-infra16:49
fungiso far it's pretty much just delete everything associated with the old account_id except account_patch_reviews (which should get switched to the new account_id instead)16:51
*** enikanorov-w has quit IRC16:51
*** enikanorov-w has joined #openstack-infra16:52
fungiand expect the user to reapply their preferences, keys, cla signature, contact info, et al to the newer account if needed16:52
fungiabout the only thing they can't change easily is their ssh username, so i added a section for how to change that in the db as well16:52
*** hashar has quit IRC17:00
*** weshay has quit IRC17:02
*** derekh has quit IRC17:02
*** pcrews has joined #openstack-infra17:03
*** dkliban has quit IRC17:04
*** odyssey4me has quit IRC17:05
*** sarob has joined #openstack-infra17:05
*** garyk has quit IRC17:06
*** pblaho has quit IRC17:08
*** anteaya has quit IRC17:08
*** dina_belova has joined #openstack-infra17:09
fungiactually, i've already spotted about half a dozen other tables where update queries are warranted on a merger, and found one where you can encounter collisions too17:09
fungiso i'll have a new rev on that change here in a bit17:09
jeblairfungi: why did you decide to merge?17:09
*** nati_uen_ has joined #openstack-infra17:10
*** nati_ueno has quit IRC17:10
fungii thought it had been requested, and i was already in there handling a couple others anyway17:10
*** nati_ueno has joined #openstack-infra17:11
*** vogxn has quit IRC17:11
jeblairgerrit clearly wasn't designed for it, so i'd rather just not do it.  why not just say, hey, if you get a new openid, you get a new gerrit account.  congrats.17:12
*** SergeyLukjanov has quit IRC17:12
jeblair(we've never done it before for this reason -- it's not supported, documented, tested, etc, and it's not necessary)17:12
*** dina_belova has quit IRC17:13
openstackgerritA change was merged to openstack-infra/config: Rotate gearman log worker logs.  https://review.openstack.org/4430317:13
jeblairwell, that's not entirely true -- the launchpad sync script did do some amount of merging...17:13
fungiwell, it does have the unfortunate side effect that at least ssh username renames are probably warranted, and also the patch owners lose control over open reviews (drafts, et cetera)17:13
jeblairbut we decided to stop doing that17:13
fungiseems like a fine policy going forward, it is finicky and time-consuming17:14
jeblairagreed on the ssh rename bit17:14
fungioh, and also it almost always results in duplicate e-mail addresses, which then breaks gerrit17:15
fungi(keeps you from being able to select the user to add them as reviewers, group members)17:15
*** nati_ueno has quit IRC17:16
fungithe additional/duplicate accounts result in that, i mean, so we still end up needing to clean that up17:16
fungiwhich means figuring out which account is the right one, which is the more time consuming part anyway17:17
jeblairwell, whatever you come up with, having documentation about it will be great.  :)17:17
openstackgerritClark Boylan proposed a change to openstack-infra/config: Rotate Jenkisn log client logs.  https://review.openstack.org/4491917:17
clarkbjeblair: fungi: why do ssh usernames matter?17:17
clarkbit is a configurable gitreview and ssh setting, should be set once and forget17:18
*** weshay has joined #openstack-infra17:18
fungiclarkb: well, because users will continue to upload with their old ssh username and then can't figure out why they don't own their changes... or will get mysterious errors because the account associated with their ssh username lacks cla/contact info and they see it added in the webui17:19
fungibut agreed, it's symptoms which can be recognized, explained in an faq and a workaround recommended17:19
jeblairclarkb: i'd accept that it's a notable inconvenience not to be able to have your ssh username of choice (at least, if you are the user who has it)17:19
clarkbjeblair: fungi: I am going to fix pbr/gear on the logstash workers so that the logrotate (and potentially other puppet changes) can be applied17:20
clarkbworker2 seems happy with the change17:20
jeblairclarkb: ++17:20
openstackgerritJames E. Blair proposed a change to openstack-infra/nodepool: Add image-update command  https://review.openstack.org/4492017:21
jeblairthis ^ gives us "nodepool image-update hpcloud-az1 devstack-precise" which i think is pretty cool.  :)17:22
clarkbjeblair: I haven't looked at the nodepool command change yet (it is starred for review), does that command communicate to the daemon asking it to do work or does it execute a completely distinct nodepool process that uses the same underlying DB?17:25
*** HenryG has quit IRC17:25
jeblairclarkb: distinct17:25
*** portante|afk is now known as portante17:25
*** sarob has quit IRC17:28
*** sarob has joined #openstack-infra17:28
*** dkliban has joined #openstack-infra17:29
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Try additional Rackspace region  https://review.openstack.org/4267717:29
pleia2fungi: our puppet-dashboard server filled up disk space last week and we're brainstorming how to move forward, since you worked with anteaya on some mysql optimizations several months ago I wanted to loop you in for your thoughts when you have time, bug is up to date: https://bugs.launchpad.net/openstack-ci/+bug/1218631 (I added this to meeting agenda too)17:29
uvirtbotLaunchpad bug 1218631 in openstack-ci "Build new puppet-dashboard server" [Undecided,New]17:29
fungipleia2: yeah, the bug is pretty accurate already. i saw mention of it in scrollback from last week and commented in here over the weekend17:30
*** Ryan_Lane has joined #openstack-infra17:30
clarkbjeblair: fungi pleia2 lifeless mordred when fixing the pbr version mismatch then rerunning puppet you may need to double tap puppet if you have mutliple pip packages. The first one that installs a thing that depends on pbr will move pip to /usr/local/bin and subsequent pip runs won't find /usr/bin/pip. Just run puppet again and puppet will refind pip in the new location17:30
pleia2fungi: ah ok, thanks :) should be all set for meeting then17:31
fungipleia2: but in short, the plan sounds sane. last time this happened i had to blow away the db because there wasn't enough room to repair the db (nor was there any good reason to bother). afterward we instituted changes to the mysql configuration so the next time it's rebuilt it should be suitable to support periodic automated maintenance17:32
pleia2fungi: anteaya updated it this morning with a series of other options, including using a fork instead of the unsupported one we're using17:33
*** sarob has quit IRC17:33
fungibut given that we were also starting in earnest to move servers from rs legacy to rs nova, there wasn't a lot of reason to rebuild it in place a second time17:33
* pleia2 nods17:33
fungiand yeah, the concerns over long-term efficacy/support of puppet-dashboard are mostly a new discussion (though it sort of came up the first time they announced they were ceasing to maintain it)17:34
pleia2the main complication of using the fork is ruby version, since we're running this on precise and that's got the ruby version we want17:35
pleia2for regular puppet-dashboard17:35
*** ryanpetrello has quit IRC17:36
*** sarob has joined #openstack-infra17:37
jeblairjhesketh: i just marked your first zuul change as work-in-progress so it doesn't accidentally get merged this week17:38
clarkband for some reason pbr does not want to install on worker4...17:38
jeblairjhesketh: we're trying to avoid disruptive changes until the H3 milestone is released, which is on Friday17:38
jeblairjhesketh: but i think it's ready to be merged first thing next week (yay!)17:39
*** HenryG has joined #openstack-infra17:42
clarkbAttributeError: Distribution instance has no attribute 'include_package_data'17:42
*** ryanpetrello has joined #openstack-infra17:43
*** ryanpetrello has quit IRC17:44
clarkbis that the upgrade setuptools globally error? /me attempts17:44
*** SergeyLukjanov has joined #openstack-infra17:45
clarkbnope that didn't seem to help17:47
*** ryanpetrello has joined #openstack-infra17:48
clarkbjeblair: have you seen that error when installing pbr before?17:48
jeblairclarkb: i don't recall, sorry.17:48
*** garyk has joined #openstack-infra17:50
clarkbI get the same failure with setuptools 0.9.8 and 1.117:52
*** tjones has joined #openstack-infra17:54
clarkbdstufft: ^17:55
openstackgerritA change was merged to openstack-infra/config: Update references to GerritJenkinsGit wiki  https://review.openstack.org/4450717:55
clarkbjeblair: workers1-3 are happy. 4 and 5 are in different degrees of unhappy, 6-8 have been left alone17:55
*** ryanpetrello has quit IRC17:56
*** ryanpetrello has joined #openstack-infra17:56
sdaguejog0 you about?17:56
dstufftclarkb: looks like the same problem as the other day when upstream wasn't patching Distribution somewhere (or something lie that)17:57
*** ryanpetrello has quit IRC17:57
*** yjiang5_away is now known as yjiang517:57
*** ryanpetrello has joined #openstack-infra17:58
*** boris-42 has joined #openstack-infra17:58
clarkbdstufft: interesting, maybe 0.9.8 is too new then?17:58
clarkbI will try with older setuptools17:58
dstufftclarkb: Note: that's just me off the cuffing looking at the error message17:58
*** ryanpetrello has quit IRC17:58
*** ryanpetrello has joined #openstack-infra17:58
jog0sdague: yup17:59
openstackgerritJoe Gordon proposed a change to openstack-infra/devstack-gate: Add support for large_ops tempest test  https://review.openstack.org/4377917:59
jog0sdague: weclome back from vacation17:59
sdaguejog0: never, mind, I just figured it out. Just trying to figure out how to generate the refresh token on your gmail gerrit script17:59
sdaguejog0: thanks :)17:59
jog0sdague: ahh, the code is pretty raw so patches welcome.18:00
clarkbdstufft: that must be it. setuptools 0.7.8 seems to work18:00
*** senk has quit IRC18:00
dstufftclarkb: see if there's a bug with setuptools already and if not file one? it's on bitbucket under pypa/setuptools18:00
dstufftIt looks like it's not patching the Distribution class in all cases18:00
clarkbdstufft: clayg filed one18:00
*** ryanpetrello has quit IRC18:00
*** mrmartin has quit IRC18:01
*** ryanpetrello has joined #openstack-infra18:01
clarkbI think the error clayg ran into was slightly different, but this must be different symptooms for the same thing18:01
clarkbdstufft: I think clayg mentioned that a patch was being worked on18:01
sdaguejog0: yeh, seems like it's gone wonky for me18:01
dstufftclarkb: yea it was a different error18:01
dstuffthis error was a place where setuptools was checking subclass18:01
sdaguejog0: http://paste.openstack.org/show/45688/18:02
dstufftthis is just using the Distribution class from distutils instead of setuptools, andsetutptools added the include_pakage_data bit18:02
sdagueI figured it was worth a shot to help me reduce my review email before I tried to dive in18:02
*** ryanpetrello has quit IRC18:02
*** ryanpetrello has joined #openstack-infra18:03
*** gaelL has quit IRC18:03
jeblairsdague: welcome back!18:03
sdaguejeblair: thanks :)18:03
*** ryanpetrello has quit IRC18:04
*** ryanpetrello has joined #openstack-infra18:04
*** gaelL has joined #openstack-infra18:04
clarkbsdague: if mtreinish hasn't mentioned it yet I think you owe him beer :)18:05
jog0sdague: hmm did you approve gerrit-gmail to access your email in https://code.google.com/apis/console18:05
jeblairi think we all owe him a beer :)18:05
sdaguenice :)18:05
jeblairi think 1200 people owe mtreinish a beer18:05
jog0sdague: under api access18:05
*** ruhe has quit IRC18:06
clarkbjeblair: indeed18:06
*** garyk has quit IRC18:07
sdagueI'll make sure to buy him a few rounds18:07
sdaguehe's actually out this week18:08
jog0sdague: ahh I found the issue18:08
sdaguejog0: I did via the browser....18:08
sdaguejog0: ok, cool18:08
jog0sdague: https://github.com/jogo/gerrit-gmail/blob/master/gerrit-gmail.py#L8218:09
jog0I hard coded in the email tag  Iuse18:09
jog0let me make that a config option, in the mean time you can just swap in your tag there18:09
*** jpmelos has left #openstack-infra18:10
clarkbfungi: when you have a free moment can you look at killing salt salt clients in places that they have gone sidways? the change to disable salt has merged so that we can fix it properly without it interferring with other things, but in the meantime we should probably clean those up18:12
clarkbfungi: I am happy to help but was hoping you could take a look at it since you seem most familiar18:12
fungiclarkb: can do here in a bit18:12
clarkbfungi: just let me know if I need to kill processes on half the slaves or something18:12
clarkbpbr, gear, and puppet are happy on the logstash workers18:13
*** dina_belova has joined #openstack-infra18:13
markmcclainneed update the minimum neutronclient version, but wanted to confirm the proper order for some actions… update requirements first and then the individual projects or the other way around18:14
clarkbmarkmcclain: openstack/requirements first18:15
markmcclainclarkb: will do.. thanks18:15
sdaguejog0: cool, I'll probably turn this into a list, as I actually break them out by projects, but at least that gives me a lead18:16
*** dina_belova has quit IRC18:18
*** jbjohnso has quit IRC18:18
jog0sdague: are those subtags?18:19
sdaguethey are18:19
sdagueworks with subtags?18:19
jog0so just pick the parent tag18:19
clarkbjd__: https://review.openstack.org/#/c/44246/ do those tests pass?18:19
clarkbjd__: Alex_Gaynor mentioned that hacking does but couldn't get pbr to run18:20
Alex_GaynorMy failure to get PBR wasn't related to PyPy (in case that's unclear), tests shouldn't run under CPython either18:20
jog0sdague: for example I have: Openstack/review/nova openstack/review/... so I just specify openstack/review18:20
jog0sdague: just pushed a patch18:21
*** pblaho has joined #openstack-infra18:22
*** dina_belova has joined #openstack-infra18:23
sdaguejog0: parent tags don't seem to work for me18:25
jog0sdague: yeah I am testing that out myself right now.  for now  does a single tag work?18:26
jeblairfungi: do you want to WIP https://review.openstack.org/#/c/44912/1  ?18:26
sdagueit seems to18:26
jog0if so you can just do that for now18:26
*** melwitt has joined #openstack-infra18:27
jog0also note there are some command line options (use -h to seethem)18:27
fungijeblair: yep, thanks--done18:27
*** tjones has left #openstack-infra18:28
*** dina_belova has quit IRC18:28
jog0sdague: mark as read is not enabled by default18:29
sdaguejog0: oh, ok, where do I turn that on?18:29
jog0sdague: run ./gerrit-gmail.py -h18:29
jeblairclarkb: sorry to bother, but can you explain https://review.openstack.org/#/c/44913/2/gerritbot/bot.py to me?18:29
jeblairclarkb: it looks to me like it would send everything that's not patchset-created or change-merged through the comment_added method...18:30
jeblairclarkb: which doesn't sound right18:30
clarkbjeblair: thats correct, but then comment_added() will filter out things that shouldn't actually send messages18:30
clarkbjeblair: it is possibly to move that filtration up higher into the if block that is modified in that change18:31
jeblairclarkb: would it not send, eg, ref-updated events to comment_added?18:31
jeblairclarkb: which would transmit them if a channel was configured for comment-added events?18:31
sdaguejog0: ok, that seems to be doing a thing18:31
sdaguebut only if I narrow it to the final tag18:32
morganfainbergclarkb / jeblair, not sure who to poke at but is there something going on w/ gerrit?  I can ssh to it, very slowly, but git review is just handing indefinitely18:32
jog0sdague: yeah I am looking into why that is.18:32
sdaguejog0: coolio18:32
sdagueotherwise, this is pure awesome18:32
jog0sdague: can you paste the error you get18:32
morganfainbergmight be my conneciton, but just figure i'd ask if something known was going on.18:32
sdagueyeh, one second, letting it complete it's run first18:32
jog0sdague: thanks18:32
clarkbjeblair: oh you are correct, good catch18:32
clarkbjeblair: so we will need to filter in the if block18:33
jeblairmorganfainberg: not that i'm aware of; i'd start looking at your connection (ping/mtr)18:33
jeblairit did, however, just transmit a HUGE amount of data18:34
*** jjmb has quit IRC18:34
jog0sdague: what happens if you set 'tag=[Gmail]/All Mail18:34
morganfainbergjeblair, thanks digging into that now.18:34
jog0in the config file18:34
*** jjmb has joined #openstack-infra18:34
*** ruhe has joined #openstack-infra18:35
jog0that seems to work for me18:35
sdaguejog0: ok, will try, though I liked the idea of limitting it's behavior more so it doesn't accidently do things to the rest of my mail18:37
jog0sdague: yeah, thats why mark as read must explicitly be set as a flag18:37
jog0sdague: I once marked an entire tag as read18:37
sdaguealso, I have a couple GB of email, so I wonder how slow the all mail will be18:38
jog0sdague: it just looks for unread18:38
sdagueok, cool18:38
*** ruhe has quit IRC18:38
sdagueyeh, that seems to be working18:39
*** weshay has quit IRC18:39
jog0sdague: basic flow is: get unread emails, look for X-Gerrit-Change-Id. and compare list to ssh gerrit query18:39
jog0sdague: sweet, I run this tool every morning18:40
sdaguejeblair: I'm trying to read the new devstack node counts in zuul. Should I be concerned about flushing too much stuff to merge, or are we bursting dynamically now?18:42
jeblairsdague: short: don't worry, throw whatevery you want at it.  :)18:42
sdaguejeblair: yes sir :)18:43
*** dprince has quit IRC18:43
sdagueat some point I'll need to debug why the sparklines are flashing on update for me18:43
jeblairsdague: long: it's roughly the same algorithm as before, but it's being run as a daemon now and is very responsive18:43
sdaguebut the ui tweaks are nice18:43
jeblairsdague: ohno!  they should be updating using the same method as the big graphs at the bottom18:44
jeblairsdague: thanks :)18:44
sdagueon chrome I see a momentary flicker on the update18:44
sdaguewhich is not seen on the others18:44
*** dkehn is now known as dkehn_away18:44
openstackgerritJeremy Stanley proposed a change to openstack-infra/config: Gerrit sysadmin tips for account repairs/renaming  https://review.openstack.org/4491218:45
jeblairsdague: i see that too on chromium18:45
*** olaph has joined #openstack-infra18:45
jeblairsdague: the big graphs only update once/min; perhaps they fliker too?18:45
jeblairsdague: the test node graph is still only for devstack nodes at the moment; unit test slaves are still static, though there are a lot more of them18:47
jeblairsdague: (i just realized that may have been what you were asking)18:47
jeblairsdague: hopefully the unit test slaves will be dynamic soon.18:47
jeblairi think we have 40 precise nodes and 14 centos618:49
*** zaro0508 has joined #openstack-infra18:49
openstackgerritmark mcclain proposed a change to openstack/requirements: update min version for neutronclient  https://review.openstack.org/4493618:51
*** weshay has joined #openstack-infra18:53
*** jpeeler has quit IRC18:59
*** jpeeler has joined #openstack-infra18:59
jeblairit's meeting time!  as soon as the keystone folks finish18:59
*** HenryG has quit IRC19:00
*** nati_ueno has joined #openstack-infra19:01
*** syerrapragada has joined #openstack-infra19:13
*** pblaho has quit IRC19:13
*** dripton_ is now known as dripton19:15
*** mrmartin has joined #openstack-infra19:16
*** thomasbiege1 has joined #openstack-infra19:18
*** zul has quit IRC19:21
*** wenlock has joined #openstack-infra19:22
*** anteaya has joined #openstack-infra19:22
*** dina_belova has joined #openstack-infra19:24
anteaya<= has her internet and telephone back now19:24
openstackgerritA change was merged to openstack-infra/config: Backup review and wiki MySQL DBs.  https://review.openstack.org/4412919:25
*** dkliban has quit IRC19:27
*** dina_belova has quit IRC19:29
*** gyee has quit IRC19:30
*** thomasbiege1 has quit IRC19:32
*** zul has joined #openstack-infra19:33
lifelessBobBall: jeblair: hi; I'd be delighted to answer any dib questions (and other folk who are familiar include the rest of tripleo, hub_cap and the trove folk, and [some] heat folk.19:34
ttxmordred: you might want to review https://review.openstack.org/#/c/44489/19:37
ttxit's your idea after all19:37
*** vipul is now known as vipul-away19:41
*** sarob_ has joined #openstack-infra19:43
*** sarob has quit IRC19:45
*** ryanpetrello has quit IRC19:45
*** ryanpetrello has joined #openstack-infra19:46
*** sarob_ has quit IRC19:47
openstackgerritRussell Bryant proposed a change to openstack-infra/reviewstats: Update disagreements calculation.  https://review.openstack.org/4494619:48
*** dkliban has joined #openstack-infra19:49
openstackgerritA change was merged to openstack-infra/reviewstats: Update disagreements calculation.  https://review.openstack.org/4494619:51
*** senk has joined #openstack-infra19:53
* clarkb lunches19:59
*** senk1 has joined #openstack-infra19:59
*** hashar has joined #openstack-infra20:00
mordredttx: looking20:00
mordredjeblair: morning!20:00
*** senk has quit IRC20:00
jeblairmordred: good afternoon!20:00
hashargood timezone everyone20:01
mordredjeblair: indeed! I'm actually in Sacramento right now - so I agree with your view of time20:01
*** vipul-away is now known as vipul20:01
openstackgerritA change was merged to openstack/requirements: Allow use of oslo.messaging 1.2.0a11  https://review.openstack.org/4371220:01
jeblairhashar: good timezone to you too!20:01
jeblairmordred: that's terrible!  i hope you're not there for long!20:01
fungiclarkb: salt cleanup stats so far... of the 91 minions ci-puppetmaster thought it knew about, 21 don't exist any longer, 44 seem to have stopped salt successfully and 26 still have lingering processes. i'm manually killing the stragglers now20:01
mordredjeblair: I'll be taking off in 45 minutes20:01
mordredjeblair: there is no way I can read scrollback for 2 weeks - so as soon as I'm online for more than 5 minutes at a time, I might want to pick your brain for big-ticket things I should know about20:02
jeblairmordred: happy to help!20:03
mordredjeblair: looks like the project is still in one piece though - and I like the new little graphs on the zuul status!20:04
openstackgerritRyan Petrello proposed a change to openstack-infra/config: Provide a more generic run-tox.sh.  https://review.openstack.org/4314520:04
jeblairmordred: they're called 'sparklines'.  edward tufte invented them.20:04
pleia2mordred: we made a whole fleet of git+cgit servers behind haproxy too http://ci.openstack.org/git.html20:05
anteayafungi: yes been chatting with the salt fellers, they want to stop by when we have a bit of space to do some work20:08
mordredpleia2: neat!20:09
mordredanteaya, fungi: are we doing things with salt now?20:09
anteayawe looked at the salt version we are running and it is old and in need of an upgrade, some of the memory leak issues are believed to be attributed to the version of zmq we were running20:09
anteayamordred: being off20:09
anteayawe shut off salt due to its use of resources20:09
mordredah. yes. I believe I remember doing that20:10
anteayaprobably the zmq 2.x memory leak20:10
anteayaI think it got merged last week20:10
mordredso this is part of "fix whatever those problems were"20:10
anteayayeah, jeblair hasn't had time to dig into what the problem was so we just shut it down20:10
anteayaand the I told the salt guys and they want to help us, when we have time to receive the help20:11
anteayaI told them before ff was a bad time20:11
anteayaso once you are up to speed on the priorities and fungi has read enough email to make him happy20:12
anteayaI'll invite them in and we can chat20:12
anteayamake a plan and all20:12
*** sarob has joined #openstack-infra20:13
mordredanteaya: oh right, fungi is just getting back too isn't he?20:16
anteayahe is20:16
* fungi nods20:16
mordredfungi: we should coordinate better ;)20:16
anteayaand getting over hives and a respiratory infection20:16
mordredfungi: how is scrollback/email overload for you?20:16
mordredoh my!20:16
anteayahe had a great vacation20:16
mordredfungi: you should avoid hives and respiratory infections20:16
fungimordred: i declared bankruptcy on irc scrollback besides my nick highlights and am skipping most mailing list stuff unless it looks really important20:17
fungistill taking a while to dig back out20:17
*** zaro0508 has quit IRC20:18
*** reed has quit IRC20:19
sdaguefungi: well I just disconnected my irc proxy when I was out, easier to be preemptively bankrupt on that20:20
fungialso have 6 days to select and work out the schedule of summit tech-deep-dive track talks with the other co-chairs, so this week is going to be fun20:20
* sdague just finished going through openstack-dev list... whew20:20
pleia2anteaya: started an etherpad here to start getting thoughts together re: sodabrew fork https://etherpad.openstack.org/migrate-to-sodabrew20:20
* anteaya looks20:21
*** sarob has quit IRC20:22
openstackgerritEmilien Macchi proposed a change to openstack-infra/jenkins-job-builder: Add IRCbot plugin support  https://review.openstack.org/4403220:23
*** weshay has quit IRC20:23
*** weshay has joined #openstack-infra20:23
anteayapleia2: i don't know what a vscrepo is and the only reference I found that wasn't vbscript is in the puppetlabs comments mentioning vscrepo but not explaining it20:24
*** dina_belova has joined #openstack-infra20:24
pleia2anteaya: in puppet it's how we pull in projects from git directly (rather than with a debian package or similar)20:25
*** russellb is now known as markmc_by_proxy20:25
anteayaah okay20:25
*** markmc_by_proxy is now known as russellb20:25
fungianteaya: in openstack/config try running 'git grep vcsrepo'20:25
pleia2anteaya: https://git.openstack.org/cgit/openstack-infra/config/tree/modules/reviewday/manifests/site.pp#n78 is a simple example20:26
pleia2or what fungi said :)20:26
openstackgerritJoe Gordon proposed a change to openstack-infra/devstack-gate: Add support for large_ops tempest test  https://review.openstack.org/4377920:26
*** rnirmal has quit IRC20:27
anteayahttps://git.openstack.org/cgit/openstack-infra/config/tree/modules/reviewday/manifests/site.pp#n78 so it is a puppet resource/command?20:27
anteayafound it: http://forge.puppetlabs.com/puppetlabs/vcsrepo20:28
pleia2anteaya: it's a puppet module20:29
anteayaI had vsc and it is vcs for version control systems20:29
anteayaokay thanks20:29
*** dina_belova has quit IRC20:29
pleia2anteaya: oops, typing fail, sorry about that20:29
anteayano worries20:29
anteayaI am having swap the two letter beside each other dyslexia all afternoon20:30
anteayahow would our puppet-dashboard module have to be modified to use the sodabrew dashboard?20:31
anteayaany thoughts?20:32
fungihow do they differ? that's probably where the discussion starts20:32
*** mrmartin has quit IRC20:32
pleia2anteaya: I haven't actually looked at either very closely, I just think it's likely it won't work automatically20:33
anteayaa good place to begin, it probably won't I agree20:34
fungiobviously we replace the apt package provider entry for it with the vcsrepo provider and an exec to "install" it from the checkout, but beyond that i have no idea how the codebases diverge20:34
anteayapleia2: do we have the ability to try it and run a stacktrace on how it fails?20:34
fungiso unless there are any already known differences we'll need to cater to, probably best to try installing with a test vm and see what works and what doesn't20:35
pleia2anteaya: I think it's more about the module not syncing up with the exact variables and install procedures of the new version20:35
anteayaI'm all for trying it on a test node and getting some stacktrace20:35
pleia2no need for stacktrace just yet :)20:35
anteayaah okay20:35
pleia2so yeah, what fungi says20:35
pleia2I need to grab lunch and then spend some time on lxc baremetal this afternoon before the nova folks hunt me down ;)20:36
EmilienMmgagne: if you have some time, I miss a last approve on https://review.openstack.org/#/c/43685/20:36
fungiclarkb: all formerly known salt minions are no longer running any remaining salt processes, so should be all set now20:36
anteayapleia2: okay20:36
lifelessok, so who can I pester about hiera content setup for gerrit; like - do you use a different swift user, throwaway keys? what abuot github tokens and do you use a dedicated user or will any old user do?20:39
fungipleia2: also bug 1219076 might be something you can figure out. i don't immediately see any indication we're using puppet to delete it, nor do i recall any reason why we should20:39
uvirtbotLaunchpad bug 1219076 in openstack-ci "puppet-dashboard does not run background services" [Undecided,New] https://launchpad.net/bugs/121907620:39
*** gyee has joined #openstack-infra20:40
lifelessfungi: it may not be puppet to delete it - I couldn't find the deleted file path in a grep... but dpkg -L on the package lists it...20:41
lifelesspleia2: I will hold the net for the nova folk :)20:41
sdagueit's pretty awesome to see all those tempest runs checking in at < 30 minutes20:41
fungilifeless: the swift user is only for openstackwatch (the rss feed generator thing) if memory serves. as for github it needs a dedicated user, i believe, since gerrit needs a privileged one to be able to force push and delete stuff as part of git replication20:42
fungilifeless: though again, if you don't want to replicate to github, then shouldn't be necessary20:42
*** ryanpetrello has quit IRC20:43
lifelessfungi: replicating to github is a good idea for $most people.20:43
lifelessfungi: so I'd rather document it than not.20:43
lifelesshow do you install the rax dns client?20:44
lifelessis it pip installable?20:44
anteayasdague: isn't that lovely?20:44
fungilifeless: maybe, but we have to run it from a virtualenv because its dependencies conflict with other system-installed things20:44
lifelessfungi: so whats the gerrot_github_token for ?20:44
lifelessfungi: so, what is the rax dns clients pip name?20:45
* fungi checks20:45
lifelessfungi: basically, this package exists on puppetmaster and there is no documentation about where to get it20:45
lifelessfungi: so I'm going to ask lots and lots of annoying questions so that I don't install some random *other* package.20:45
lifelesscause there are at least two unofficial rackspace dns clients20:45
openstackgerritSimon König proposed a change to openstack-infra/jenkins-job-builder: Added description of the new option to the doctstring  https://review.openstack.org/4495120:46
fungiyes, it's frustrating. we didn't really puppet any of the launch tool dependencies20:46
clarkbfungi: awesome, thank you (re salt)20:46
fungibecause it started out mostly ad-hoc20:46
lifelessfungi: they are puppetted now, except that one20:46
lifelessfungi: courtesy me asking annoying questions:)20:46
lifelessfungi: ok so what I need to know gerrit wise, is for each hiera key in site.pp for review.$ whether its:20:47
lifeless - a locally created secret20:47
lifeless - a provider allocated secret of some sort, and if so the specific path/docs to follow to make it20:48
clarkbmordred: are you back yet? clayg found a bug in setuptools that is affecting pbr. https://bitbucket.org/pypa/setuptools/issue/73/20:48
lifeless - $other20:48
clarkbmordred: ^ is probably an important thing to at least be apprised of20:48
lifelessfungi: clarkb: ^20:48
fungilifeless: http://paste.openstack.org/show/45693/20:48
lifelessWhat I really need is someone with deep -infra knowledge to give me a dedicated 15m or so to wander through all the keys and answer that stuff20:48
notmynamewhere is the CI config file that shows what tests are running for which project?20:50
*** russellb is now known as markmc_by_proxy20:50
*** markmc_by_proxy is now known as russellb20:50
openstackgerritBoris Pavlovic proposed a change to openstack-infra/config: Add new project Rally  https://review.openstack.org/4495220:50
clarkbnotmyname: https://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/zuul/layout.yaml20:50
funginotmyname: openstack-infra/config modules/openstack_project/files/zuul/layout.yaml20:50
fungid'oh, clarkb beat me to it20:50
notmynamefungi: and clarkb had a clicky link. yay20:51
fungilifeless: i'll make a quick pass through the hiera variables for the review.o.o server and see what i can remember20:51
notmynamethanks :-)20:51
lifelessfungi: thank you!20:51
lifelessfungi: the ssl one I need to dig up doing self signed certs; the contactstore I am ignoring for now.20:51
lifelessfungi: ditto gerritbot20:51
lifelessfungi: email_private_key has a ?, mysql likewise, lp_sync_* likewise.20:52
fungilifeless: github_oauth_token, github_project_username and github_project_password are passed down to jeepyb for manage-projects to use when creating/updating projects mirrored to github20:52
lifelessfungi: how do you get an oauth token out of github? I only found the 'register app' thing20:53
lifelessfungi: which is an API key, not an oauth token20:53
*** sarob has joined #openstack-infra20:54
fungilifeless: i believe if you run manage-projects by hand, the github python module will spawn a (lynx?) browser to interactively request authentication and cache the token20:54
funginot very automation-friendly20:54
lifelessalso a bootstrap problem20:54
*** Adri2000 has quit IRC20:55
lifelessput this thing in hiera to spawn the machine which you can only make on the machine20:55
openstackgerritSimon König proposed a change to openstack-infra/jenkins-job-builder: Added support to trigger a build on the same node using the parameterized trigger plugin  https://review.openstack.org/4490020:55
fungiwe have a lot of gaps where we automated the things which were easy enough and punted on manual intervention where not20:55
fungihopefully there's not a lot of that left at this point though20:55
lifelessfungi: so where does rackdns come from? the rack auth thing includes it?20:55
lifelessRyan_Lane: cause dns.py just prints out the rackdns commands to run20:56
fungilifeless: that -e git+https:// line in the freeze output20:56
lifelessRyan_Lane: sorry20:56
lifelessfungi: so its not just auth? :)20:56
fungilifeless: the -e line is for #egg=rackspace_dnsclient-dev20:57
lifelessoh! I misread. cool.20:57
fungilifeless: rackspace-auth-openstack is a separate module entirely20:57
fungibut yeah, random commit from random person's github repo with nothing to officially tie it to rackspace's dns service. we love that20:58
*** jpmelos has joined #openstack-infra20:58
*** branen__ has quit IRC20:59
fungiat least it's relegated to its own virtualenv, but i got the impression that was done for practical dependency chain reasons as well as just general safety/security20:59
*** svarnau has quit IRC20:59
*** boris-42 has quit IRC20:59
*** sarob_ has joined #openstack-infra21:01
*** danger_fo is now known as danger_fo_away21:01
mordredsdague: you're back from vacation too?21:03
fungilifeless: so proceeding through the list of hiera parameters for review.o.o... mysql_password, mysql_root_password and email_private_key get passed down into ::gerrit to configure the service21:04
lifelessfungi: are manually made? Does puppet make the user and set the password, or do I manually make them and record them in hiera?21:04
lifelessfungi: ^ thats the sort of question I'm really asking.21:04
lifelessfungi: or put another way 'is this where the value is *created* or where the value is *provided* with creation separate?21:05
fungilifeless: ahh, looking at the gerrit module to see21:05
*** sarob has quit IRC21:05
*** jjmb has left #openstack-infra21:05
lifelessfungi: I need to know whether to advise folk to 'make a random $thing of $thing-type' or to 'go look at $other place and grab the $thing from there.21:05
*** sarob_ has quit IRC21:05
fungilifeless: we tend to have a mix of the two, so yeah, i get the lack of clarity there. been a while since i set up a gerrit so i don't remember which bits were automagic but i'm looking21:06
lifelessfungi: thanks!.21:06
Ryan_Lanelifeless: :D21:07
Ryan_Lanepleia2: howdy21:07
Ryan_Lanepleia2: so, I have a devstack image roughly done21:07
*** weshay has quit IRC21:07
Ryan_Laneit's in ovf format and should work with at least vmware and virtualbox21:07
Ryan_Lanefor sure it works with virtualbox, since I tested it. it includes port forwards for ssh and http.21:08
fungilifeless: mysql_password and mysql_root_password get passed to mysql::server to populate/update them in the database21:08
Ryan_LaneI'll need a bit more time to tweak it, but should be done today or tomorrow21:08
fungilifeless: so those are at least something you'd specify in hiera and things should just work21:08
*** senk1 has quit IRC21:08
clarkbmordred: did you catch the setuptools bug I pasted a little earlier?21:10
*** dkehn_away is now known as dkehn21:11
mgagneEmilienM: not getting an approval, there's stuff to improve =)21:11
fungilifeless: and simlarly email_private_key ends up getting inserted as registerEmailPrivateKey in the secure.config template by puppet. trying to refresh my memory on what that's used for21:11
*** svarnau has joined #openstack-infra21:12
*** SergeyLukjanov has quit IRC21:12
*** pentameter has quit IRC21:12
*** sandywalsh has quit IRC21:12
jeblairfungi: https://bugs.launchpad.net/openstack-ci/+bug/98092221:13
uvirtbotLaunchpad bug 980922 in openstack-ci "gerrit unable to register email addresses" [Critical,Fix released]21:13
jeblairfungi: it's used for email confirmations21:13
fungi"Private key used to sign the links emailed to users when they request to register a new email address on their user account."21:13
fungiyep, just found it in the docs21:13
*** rfolco has quit IRC21:15
fungilifeless: gerritbot_password is the nickserv identification for the gerritbot irc bot, and would need to be manually set at the irc network and duplicated in that hiera parameter21:18
fungi(at least i don't think gerritbot has provisions for setting its own identify password the first time it connects to the irc network)21:19
lifelessfungi: is the email private key structured or just arbitrary ascii text?21:19
lifelessjeblair: ^21:19
fungilifeless: checking21:19
jeblairlifeless: i'm not sure; gerrit init created ours for us when we manually initialized the system.  if it's not in the docs (i don't think it is), then i think an examination of the code in gerrit that generates it is in order21:20
*** pcm_ has quit IRC21:20
anteayapleia2: we may need to do some reconfig here: https://git.openstack.org/cgit/openstack-infra/puppet-dashboard/tree/manifests/init.pp#n170 the sodabrew fork uses a Gemfile (which is usually read and processed by bundler): https://github.com/sodabrew/puppet-dashboard/blob/master/Gemfile21:20
anteayaI am trying to find the git repo for the puppet-dashboard that the puppet-dashboard module draws in21:21
fungiyeah, gerrit docs say it'll create one in the db automatically. looks like it's a base64-encoded binary blob of 26 bytes length...21:21
anteayaso as to compare how gems are processed21:21
shadowerhey everyone, a quick question: the Tuskar developers want to hold weekly IRC meetings. Can we use #openstack-meeting(-alt)?21:23
shadowerif we can, what should we do?21:23
*** sarob has joined #openstack-infra21:23
lifelessfungi: ok, so set to nil in hiera?21:23
lifelessfungi: (or does puppet want '' for unset things?)21:23
*** sandywalsh has joined #openstack-infra21:24
jeblairlifeless: no you should set it21:24
mordredlifeless, jeblair it's arbitrary text21:24
jeblairlifeless: https://bugs.launchpad.net/openstack-ci/+bug/98092221:24
uvirtbotLaunchpad bug 980922 in openstack-ci "gerrit unable to register email addresses" [Critical,Fix released]21:24
anteayalooks like this is it: https://github.com/puppetlabs/puppet-dashboard21:24
* mordred generated one for the hp gerrit install21:24
jeblairlifeless: read that bug for what will happen if you don't set it21:24
*** dina_belova has joined #openstack-infra21:25
*** weshay has joined #openstack-infra21:25
lifelessjeblair: so I should run gerrit init after installing gerrit before puppetting it?21:25
clarkbjeblair: https://review.openstack.org/#/c/44553/ please see comment there21:25
jeblairlifeless: i think what mordred is saying is that you should make one up and put it in hiera21:25
jeblairlifeless: like any other password21:25
lifelessok. so the base64 bit is irrelevant?21:26
jeblairlifeless: i have not done the examination of the gerrit code needed to confirm the format, so you are in his hands unless you want to do that yourself21:26
jeblairlifeless: i only assert that you should have _something_ in hiera.21:26
lifelessjeblair: ack, thanks.21:26
mordredlifeless: I believe it is a base64 encoding of a random thing21:26
fungilifeless: i think that's merely what gerrit does to turn random data into text so it can be included in the config21:26
EmilienMmgagne: ack21:26
* lifeless is the great guinea pig21:26
clarkbfor https://review.openstack.org/#/c/44057/ do we need to coordinate the move of activity board into gerrit with reed and bitergia?21:28
jeblairclarkb: responded21:28
clarkbjeblair: fungi ^ if not I can approve that change21:28
*** senk1 has joined #openstack-infra21:29
*** dina_belova has quit IRC21:30
jeblairclarkb: let's ask reed21:30
*** Bada has quit IRC21:30
jeblairit's semi-active and they aren't usually around here21:30
fungiyeah, i don't know the answer21:30
clarkbjeblair: ya I would like to avoid the force pushing that will need to be done with the clouddocs maven repo21:31
* fungi is semi-afk to avoid dripping food on the keyboard before his next meeting21:31
anteayalifeless: you are a great guinea pig21:31
anteayahi shadower, I am not sure who is in charge of the meeting schedule in the meeting and meeting-alt channels21:33
lifelessttx I think21:33
lifelessor even self serve21:33
anteayaah okay21:33
lifelessthere is a calendar somewhere.21:33
fungiself-serve in the wiki i believe21:34
shadowerright, thought about that but I didn't want to step on anyone's toes21:34
fungiand there are definitely stackforge projects (savanna for example) using that channel for scheduled meetings21:34
shadowerttx: can you confirm Tuskar can just pick an empty slot and document it in the wiki?21:34
lifelessshadower: if you can pick a UTC+12 friendly slot, I'd really appreciate that :)21:35
shadowerlifeless: we've already mostly converged on a slot, lemme check it wrt utc+1221:36
shadowermost devs are in UTC+2/-621:36
pleia2Ryan_Lane: awesome! thanks21:37
shadowerlifeless: 14:00UTC is the likely time -> 2PM your time :-(21:37
*** alaski is now known as alaski-vacation21:38
pleia2anteaya: cool, I figured we'd need to adjust the dependencies some21:38
anteayapleia2: k21:38
shadowersorry, 2am21:38
anteayatrying to find out if there is a way to just get the module to load gems from the Gemfile21:39
* shadower confuses left and right, too21:39
openstackgerritA change was merged to openstack-infra/config: Add new savanna repos  https://review.openstack.org/4434821:39
clarkbmrodden: jeblair: fungi: https://review.openstack.org/#/c/39399/ change to add pypi-mirror project we are just going to delete the unneeded files from the copied jeepyb repo?21:41
clarkbI think we discussed this, just double checking before I approve21:41
mordredclarkb: "woohoo! logstash 1.2 is out!" <-- seen on twitter21:42
* fungi vaguely recalls that being the consensus21:42
mroddenclarkb: i think that was the plan21:42
openstackgerritA change was merged to openstack-infra/config: Report cinder patches to #openstack-cinder  https://review.openstack.org/4455421:42
clarkbmordred: "woohoo" they have changed a ton of stuff, it is going to be a bit of work to switch, but the changes are good so just need to power through21:43
clarkbmordred: https://bitbucket.org/pypa/setuptools/issue/73/21:43
mroddeni can like... do a subtree copy of the history of run_mirror.py or something but that wouldn't make it through review very well21:43
clarkbmordred: ^ is causing all osrts of pbr problems21:43
openstackgerritA change was merged to openstack/requirements: update min version for neutronclient  https://review.openstack.org/4493621:43
clarkbmrodden: nevermind I think it is fine as is21:43
clarkbmrodden: it is easy to submit a commit that deletes unneeded things, easy on reviewers too21:43
mordredclarkb: awesome21:44
mroddenthe history will always be in jeepyb if we need it21:44
clarkbmordred: the tl;dr is we need old setuptools but newer than distribute merger21:44
clarkbmrodden: we can keep the history for the non mirror stuff in pypi-mirror too. The repo is small enough21:44
mordredclarkb: you're kidding21:45
clarkbmordred: I wish I was21:45
lifelessshadower: yeah, 0200 is kinda bad :(21:45
* mordred wants to kill someone21:45
lifelessshadower: it just that - well - I kinda think you want close connection to tripleo21:45
clarkbmordred: 0.7.8 works great, 0.9.8 does not. I haven't narrowed down the newest working version21:45
clarkbmordred: clayg may have an idea21:45
lifelessfungi: ok so - github oauth token, just made via the 'register app' thing in github ?21:46
lifelessfungi: project username and password are for the dedicated account ?21:46
shadowerlifeless: yeah, we do. But we'd have to ditch either the US or the EU part of the team for you.21:46
mordredclarkb: why aren't we seeing the problem regularly in our gate then?21:47
lifelessjeblair: should these parameters be documented in the gerrit.rst file or in the gerrit.pp file ?21:47
shadowerlifeless: we'll always send out the notes to openstack-dev and I reckon I'll start attending the tripleo meeting21:47
lifelessjeblair: erm, gerrit.rst /review.pp ?21:47
*** Guest68015 is now known as oubiwann21:47
dstufftmordred: do you upgrade setuptools in the venv?21:47
lifelessshadower: we put the tripleo meeting at a time everyone can get to21:47
clarkbmordred: I'm not sure yet. I believe for unittests tox/virtualenv are protecting us21:47
lifelessshadower: 9/10 at night for our EU folk; midday for US, 0800 for me.21:48
clarkbmordred: then for devstack-gate we install globally and don't explicitly upgrade setuptools21:48
lifelessshadower: adjacent to that, either side, should be tolerable for everyone once a week21:48
*** openstackgerrit has quit IRC21:48
lifelessshadower: or UTC 2000 on a different day ?21:48
*** openstackgerrit has joined #openstack-infra21:48
*** senk1 has quit IRC21:48
shadowerlifeless: most of the team is UTC+1/+2 where it's a bit tough. But I'll bring it up21:48
shadowerbetter than 2am thats for sure21:49
clarkbmordred: jeblair fungi if there is no last minute opposition to using jeepy as the base for pypi-mirror in https://review.openstack.org/#/c/39399/1/modules/openstack_project/templates/review.projects.yaml.erb I will merge that change21:49
lifelessshadower: UTC2000 would be there 9pm/10pm - it's personal time, sure, but it's not [for most folk] actually mid-sleep time.21:50
mordredclarkb: we _do_ upgrade setuptools for devstack-gate21:50
mordredclarkb: so we use latest setuptools everywhere we run tests21:50
fungilifeless: the github parameters (oauth, project_username, project_password) are all needed by the github module and seem to get embedded in one or more of its templates. looking to see what exactly21:50
shadowerlifeless: true. I'd be okay with that (as evidenced by my being online now), but I don't know about the rest of the devs. Will ask & let you know21:50
mordredclarkb: I tihnk I know the problem21:50
mordredgit clone https://github.com/openstack/swift.git21:50
mordredcd ~/swift; sudo python setup.py develop; cd -21:50
fungiclarkb: no opposition from me21:50
mordredare the commands given21:51
mordredthat won't work21:51
mordredsetuptools must be upgraded first21:51
lifelessshadower: thanks21:51
* mordred digging further21:51
clarkbmordred: this was a problem on logstash-worker4 and 5 today21:51
* anteaya is afk for a baiat21:51
clarkbmordred: I upgraded setuptools to 1.1, tried and failed, downgraded to 0.9.8, tried and failed. It wouldn't work until I downgraaded to 0.7.821:52
mordredclarkb: wow. really21:52
*** branen has joined #openstack-infra21:53
openstackgerritMathieu Gagné proposed a change to openstack-infra/jenkins-job-builder: Escape special characters in documentation  https://review.openstack.org/4495921:53
mordredclarkb: and to be clear, it's python-swiftclient you're seeing the problem on, right?21:53
clarkbmordred: the process there was puppet could not upgrade gear because the global pbr 0.5.17 version conflicted with what newer gear needed. So I pip uninstalled gear, rm'd the files in /usr/local/lib/python2.7/dist-packages then pip install gear21:53
clarkbgear 0.3.1 -> 0.4.021:54
mordredclarkb: ok21:54
openstackgerritMathieu Gagné proposed a change to openstack-infra/jenkins-job-builder: Make references to Jenkins plugins uniform  https://review.openstack.org/4496021:54
clarkbbut pip install gear would fail to install pbr, so I had to fiddle with setuptools until I got a version of setuptools that could install pbr without the AttributeError21:54
* mordred wants to cry21:55
mordredok. I'm landing. I'll dig in when I get back on internets21:55
*** branen_ has joined #openstack-infra21:56
fungilifeless: the oauth token seems to correspond to our openstack-gerrit account on github (used for replication) and the "project" username/password correspond to our openstack-project-creator account on github, used by the manage-projects script21:56
EmilienMmgagne: I could not resist to make the patchset now. I just want to build the doc before, but I'm not sure. Could you tell me how you do ?21:56
mgagnecd doc ; make html21:57
EmilienMlike other docs, make sense21:57
mgagnethat should be it, that's what I use21:57
fungilifeless: so precreated with the right privs and then entered into hiera21:57
clarkbmgagne: EmilienM: or `tox -evenv python setup.py build_sphinx`21:57
*** dkranz has quit IRC21:57
EmilienMmgagne: clarkb: thx guys21:57
* fungi needs to afk for a meeting now21:57
clarkbthat deals with any potential dependencies doc building has21:57
mgagneclarkb: thanks, I probably have them (deps) installed already ^^'21:58
lifelessfungi: oh right, and you run jeepyb to get it21:58
lifelessfungi: so bootstrap is enter it as blank21:58
* EmilienM had a lot of dependencies uninstalled 21:58
*** branen has quit IRC21:59
EmilienMmgagne: looks better now indeed, thanks for your review. I push the new patchset now and goes to bed.22:00
mgagneEmilienM: will check asap22:00
EmilienMno worries22:00
openstackgerritEmilien Macchi proposed a change to openstack-infra/jenkins-job-builder: Add Plot plugin support  https://review.openstack.org/4368522:00
mgagneEmilienM: Reference to Jenkins plugin still need to be updated22:01
shadowerlifeless: hey, I've just noticed that the Glance team alternates between 1400 and 2000 UTC. Would you be okay with that as an alternative?22:01
EmilienMmgagne: I did it, but made a mistake.22:01
lifelessshadower: sure22:02
shadowerlifeless: cool22:02
*** ftcjeff has joined #openstack-infra22:02
*** thomasm has quit IRC22:02
openstackgerritA change was merged to openstack-infra/config: Rotate Jenkisn log client logs.  https://review.openstack.org/4491922:03
clarkbmrodden: there were merge conflicts on the pypi-mirror change22:03
*** thomasm has joined #openstack-infra22:03
mroddenclarkb: k i'll look at it22:03
clarkbnow I feel bad for approving ^ it has a tyop in the commit message :/22:03
mroddenwas just about to get to that22:03
openstackgerritEmilien Macchi proposed a change to openstack-infra/jenkins-job-builder: Add Plot plugin support  https://review.openstack.org/4368522:03
EmilienMthe good one this time ^ sorry for noise.22:03
clarkbmrodden: sorry I was referring to 4491922:03
hasharEmilienM: will look at your ircbot patch tomorrow :-]22:05
mroddenboth modified:      modules/openstack_project/files/gerrit/acls/openstack-infra/nodepool.config22:05
hasharEmilienM: got to sleep for now sorry22:05
EmilienMhashar: same here22:05
clarkbmrodden: are you doing a rebase?22:06
EmilienMhashar: I did what you asked22:06
mroddendid it detect a bad rename?22:06
hasharEmilienM: wonderful.  then I could get a look at the plot plugin probably22:06
clarkbmrodden: I would check that you aren't rebasing unexpected things. did you rebase -i ?22:06
mroddenonly had one22:06
EmilienMhashar: see you tomorrow, don't hesitate to ping me here :)22:07
mroddeni'm gonna try a cherry-pick instead...22:07
*** burt has quit IRC22:08
*** lcestari has quit IRC22:08
*** flaper87 has joined #openstack-infra22:09
flaper87hey guys, Marconi just got accepted for incubation (w0000t). As for the migration from stackforge to openstack, what's needed? Is it enough to propose a patch to create Marconi (and python-marconiclient) pointing to the stackforge upstream repo?22:10
flaper87What happens with current reviews?22:11
flaper87is it necessary to get them all merged before the migration ?22:11
*** senk has joined #openstack-infra22:11
clarkbflaper87: typically we do a project rename, this requires a short gerrit downtime where we update the gerrit database to capturethe rename in your existing change reviews22:12
*** zeus has quit IRC22:12
flaper87clarkb: ok, how / when can we schedule it ?22:12
clarkbflaper87: coinciding with that is a change to s/stackforge\/*marconi*/openstack\/*marconi*/22:12
clarkbflaper87: first step is to propose the change that does ^22:13
flaper87clarkb: kk, will do!22:13
clarkbflaper87: we won't merge it until after the rename but it gets the conversation going. You might also WIP it to prevent early merging22:13
*** sarob_ has joined #openstack-infra22:13
flaper87clarkb: will do, thanks!22:13
clarkbthen we often try to bundle several renames if possible, but if non are queued we may just do marconi22:13
*** sarob_ has quit IRC22:14
*** sarob_ has joined #openstack-infra22:14
*** syerrapragada has quit IRC22:14
clarkbmordred: https://review.openstack.org/#/c/44496/ is another good one for you22:14
*** melwitt has quit IRC22:14
*** melwitt1 has joined #openstack-infra22:14
*** datsun180b has quit IRC22:15
clarkbflaper87: I wasn't too clear, existing/current reviews remain, we take the short gerrit downtiem so that their details in the DB can be updated22:15
clarkbwe will even update your watch lists in gerrit :)22:15
clarkbAlex_Gaynor: pypy neutronclient tests pass?22:16
Alex_Gaynorclarkb: yuppers!22:16
*** sarob has quit IRC22:17
flaper87clarkb: souuuunds good22:18
*** sarob has joined #openstack-infra22:18
openstackgerritA change was merged to openstack-infra/config: Run the neutronclient tests under PyPy  https://review.openstack.org/4477222:19
clarkbAlex_Gaynor: ^22:19
Alex_Gaynorclarkb: danke!22:19
*** sarob_ has quit IRC22:21
lifelessclarkb: mordred: what is ssh_project_rsa_key_contents= vs ssh_rsa_key_contents ?22:21
clarkblifeless: looking22:23
openstackgerritMathew Odden proposed a change to openstack-infra/config: Create new pypi-mirror project for run_mirror.py  https://review.openstack.org/3939922:23
openstackgerritA change was merged to openstack-infra/jenkins-job-builder: Escape special characters in documentation  https://review.openstack.org/4495922:23
fungilifeless: ssh_rsa_key_contents is the host key content for the gerrit ssh service (~/review_site/etc/ssh_host_rsa_key)22:24
fungiwhereas ssh_project_rsa_key_contents is the key used by manage-projects to authenticate to gerrit's ssh interface22:24
mroddenclarkb: i think i got it... Git was doing something with the rename detection... probably because the ACL files content is so similar22:24
lifelessfungi: thanks22:25
clarkbfungi: any idea why we put the project ssh key in review_site/etc?22:25
lifelesswhat about lp_* ?22:25
lifelesshow are they created ?22:25
*** dina_belova has joined #openstack-infra22:25
clarkblifeless: I believe that is manually following the steps given by the launchpad python lib22:26
openstackgerritMathieu Gagné proposed a change to openstack-infra/jenkins-job-builder: Make references to Jenkins plugins uniform  https://review.openstack.org/4496022:26
clarkbmrodden: thanks22:26
lifelessclarkb: I'm lost; there are 5 values there and no clue about where to start22:27
fungiclarkb: i suspect $no_reason22:27
openstackgerritFlavio Percoco proposed a change to openstack-infra/config: Move marconi under openstack  https://review.openstack.org/4496322:27
flaper87clarkb: ^22:27
*** changbl has quit IRC22:28
*** fbo is now known as fbo_away22:28
clarkbmrodden: any reason for the additional section in pypi-mirror.conf? I don't think we will need to push merge commits22:28
clarkblifeless: actually I think the lp_sync stuff can be ripped out, we aren't syncing with lp anymore22:29
*** dina_belova has quit IRC22:30
mroddenclarkb: ?22:30
mroddenoh whops22:30
clarkbmrodden: https://review.openstack.org/#/c/39399/1..2/modules/openstack_project/files/gerrit/acls/openstack-infra/pypi-mirror.config22:30
mroddenhow did that get there22:30
mroddenstupid rebase...22:30
mroddenlet me fix it22:30
clarkblifeless: http://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/templates/gerrit_lp_creds.erb half of the values go in there22:31
openstackgerritMathew Odden proposed a change to openstack-infra/config: Create new pypi-mirror project for run_mirror.py  https://review.openstack.org/3939922:31
clarkblifeless: that file may be shared with the launchpad bug and blueprint sync scripts22:31
clarkblifeless: I am still trying to figure out what the two rsa keys are for22:31
*** hashar has quit IRC22:32
clarkblifeless: the rsa keys were used by the update_gerrit_users.py script. I believe you can ignore those two parameters (and they should probably be removed). The other three lp items are the ones you want22:34
*** senk has quit IRC22:34
openstackgerritlifeless proposed a change to openstack-infra/config: Document review.pp parameters a bit.  https://review.openstack.org/4496922:36
openstackgerritlifeless proposed a change to openstack-infra/config: Phase 3 infra bootstrap docs: gerrit.  https://review.openstack.org/4497022:36
*** reed has joined #openstack-infra22:36
*** mrodden has quit IRC22:40
jheskethjeblair: yep, that's a good idea (re not merging the refactor). I was wondering if we wanted to wait :-)22:40
*** sarob has quit IRC22:41
*** sarob has joined #openstack-infra22:41
jeblairjhesketh: yeah, if this came up more often, we could set up a branch or something, but it's pretty rare that we get a refactoring zuul change that wants to land right during one of our quiet periods, so i figured it wouldn't really hurt to wait a week.22:41
*** weshay has quit IRC22:43
*** markmcclain has quit IRC22:44
clarkbmrodden left us :( I had one more comment. Oh well22:45
*** prad has quit IRC22:46
*** sarob has quit IRC22:46
*** jergerber has quit IRC22:48
*** alexpilotti has quit IRC22:48
*** ftcjeff has quit IRC22:49
*** mrodden has joined #openstack-infra22:54
*** mrodden1 has joined #openstack-infra22:56
*** mrodden has quit IRC22:58
*** mrodden1 is now known as mrodden23:01
*** tstevenson has quit IRC23:01
russellbso, either we're not merging *that* much code right now, or the vast improvements made to infrastructure and test reliability are making things super smooth23:03
*** sarob has joined #openstack-infra23:03
*** jpich has quit IRC23:03
openstackgerritClark Boylan proposed a change to openstack-infra/publications: Handle lack of tags when running make-index.  https://review.openstack.org/4497423:03
clarkbpleia2: fungi ^23:03
fungihaha! great catch23:04
clarkbrussellb: http://tinyurl.com/k4wa5po we are merging a fair amount of changes but the bulk of it seemed to be before the proposal freeze23:04
pleia2clarkb: great23:05
clarkbrussellb: I wonder if the proposal freezes were misinterpreted as being the feature freeze?23:05
fungirussellb: ttx said the same. i'd like to believe it's thanks to the improvements23:05
jeblairrussellb: http://graphite.openstack.org/render/?_t=0.6905084141797536&from=-8weeks&bgcolor=ffffff&title=Gerrit%20Events%20%28per%20Day%29&width=586&height=308&fgcolor=000000&_salt=1378249530.175&target=alias%28summarize%28stats_counts.gerrit.event.change-merged%2C%20%271d%27%29%2C%20%27Change%20merged%27%2923:05
jeblairrussellb: that's similar to clarkb's graph but expanded to 8 weeks so you can see before the pff23:05
russellbso maybe these freezes kinda worked?23:06
russellba huge push getting code posted23:06
russellband then otherwise, pain was spread over a couple weeks?23:06
russellbbut i think it's the infra improvements too ;-)23:06
*** amotoki_ has joined #openstack-infra23:06
clarkbthey definitely help :)23:06
jeblairrussellb: i think they are helping; it would have been nice if they had been in 2 weeks earlier....23:07
lifelessrdns-create wants a domani parameter23:07
lifelessbut dns.py doesn't spit one out23:07
jeblairrussellb: but at least we know the new system is battle tested.  :)23:08
russellbi bet tomorrow will pick up more than today23:08
jeblairand we have a better idea of what to expect in the future23:08
jeblairlifeless: i wonder if the client changed syntax23:09
clarkblifeless: isn't that what the --name argument is for23:09
lifelessclarkb: apparently not23:09
jeblairlifeless: our checkout hasn't been updated in quite a while23:10
jeblairthat's an absurd parameter to require for rdns-create23:10
lifelessjeblair: I installed the same commit as fungi reported you having23:10
lifeless-e git+https://github.com/kwminnick/rackspace-dns-cli@28df9f5ee79ddae2e24ff91ec34433d90bc41366#egg=rackspace_dnsclient-master23:10
fungilifeless: last nodes i spun up (~3 weeks ago?) i was able to just cut and paste the rackdns commands spat out when running the launch script23:10
fungiso maybe dns.py when run by itself is returning something slightly differenty23:11
lifelessfungi: I didn't run dns.py by itself23:11
lifelessfungi: I ran launch.py23:11
jeblairlifeless: http://paste.openstack.org/show/45700/23:12
fungiahh, okay. yeah that's what i'd been doing... it did work at least recently23:12
jeblairlifeless: that's an actual command i have recently run (with some details removed from the url)23:12
jeblairpulled it right out of my scratch buffer23:12
lifelessjeblair: very odd23:12
fungiahh, right, reverse-dns because it's based on the machine's address doesn't take a domain option23:13
fungibut when specifying a normal record it does need one23:13
mroddenclarkb: so regarding the jenkins_job_builder projects.yaml.... would copying the jeepyb config and changing the name: attribute be sufficient... i'm not sure how this all gets resolved23:13
lifelessjeblair: http://paste.openstack.org/show/45701/23:14
fungilifeless: rdns-create differs from create in that regard23:14
lifelessfungi: indeed; but note both pastebins show rdns-create23:14
fungihow odd23:14
clarkbfungi: pleia2: there is a companion publications change to the overview branch.23:14
lifelessfungi: *I* know, right?23:14
clarkbmrodden: let me look at that section, but yes I think you can copy pasta and rename23:14
mroddenit has gate-{name}-pep8 and -pyflakes23:15
lifelessmrodden: you're working on cloning the infra too ?23:15
mroddenas well as {name}-tarball and pypi-jobs23:15
mroddenuh no... i dont think so23:15
mroddenlifeless: ^23:15
mroddeni'm working on this https://review.openstack.org/#/c/39399/323:16
clarkbmrodden: you can remove the pyflakes job but I would keep tarballs and pypi jobs23:16
clarkbmrodden: that way we will be ready to pypi, and the pyflakes job is superceded by the pep8 job which runs flake823:16
mroddenok, so do these just copy some config.xml template into Jenkins?23:17
jeblairlifeless, fungi: holy unfinished job batman23:17
mroddeni don't need to worry about saying what commands get run etc23:17
jeblairi think i understand23:17
ekarlso--win 3523:17
lifelessmrodden: oh cool23:17
jeblairlifeless, fungi: it appears i made local modifications to that in order for it to work23:17
lifelessjeblair: hah!23:17
fungi'splains it23:18
clarkbmrodden: Jenkins Job Builder sorts it out for you, that is what those yaml files are. They are job templates23:18
clarkbmrodden: tl;dr you get a bunch of stuff for free and it is awesome23:18
jeblairlifeless: want to see if the latest version works?23:18
mroddenclarkb: k... well i guess we'll see what magic happens here in a bit23:18
lifelessjeblair: nto really, I have 50 odd hours to get this all working and then write a presentation about it23:18
*** sarob has quit IRC23:19
clarkbfungi: https://review.openstack.org/#/c/44977/1 is the change that goes along with fixing publications make-index23:19
clarkbjeblair: ^23:19
lifelesswhat creates /etc/init.d/gerrit ?23:19
*** sarob has joined #openstack-infra23:19
*** wenlock has quit IRC23:20
*** jhesketh_ has quit IRC23:21
*** shardy is now known as shardy_afk23:21
jeblairlifeless, fungi: ok; fyi, i am not at the moment interested in contributing my changes to that client upstream.23:21
lifelessjeblair: no worries23:22
jeblair(to be clear, i am pretty much ALWAYS interested in contributing upstream -- that should give you an idea of the degree to which i am morally opposed to its existence)23:22
fungilifeless: modules/gerrit/manifests/init.pp has a symlink23:22
openstackgerritMathew Odden proposed a change to openstack-infra/config: Create new pypi-mirror project for run_mirror.py  https://review.openstack.org/3939923:22
lifelessjeblair: lol23:22
lifelessjeblair: doesn't bother me23:23
fungidoes hpcloud's moniker instance work with the official monikerclient?23:24
*** sarob has quit IRC23:24
fungior whatever it recently renamed to23:24
clarkbfungi: it should23:24
mroddenclarkb: next patchset up... hopefully what i did there makes sense23:25
openstackgerritA change was merged to openstack-infra/publications: Handle lack of tags when running make-index.  https://review.openstack.org/4497423:25
clarkbmrodden: looking23:25
fungilifeless: gimme about 10 minutes and i'll paste.o.o the rackdns diff23:25
*** rfolco has joined #openstack-infra23:26
*** pcrews has quit IRC23:26
clarkbmrodden: LGTM23:26
openstackgerritlifeless proposed a change to openstack-infra/config: Non-openstack-ci support for launch/dns.py.  https://review.openstack.org/4498023:26
openstackgerritlifeless proposed a change to openstack-infra/config: Phase 3 infra bootstrap docs: gerrit.  https://review.openstack.org/4497023:26
openstackgerritlifeless proposed a change to openstack-infra/config: Document review.pp parameters a bit.  https://review.openstack.org/4496923:26
lifelessfungi: no need :) ^23:26
clarkbmrodden: what the stuff in projects.yaml says is instantiate those job templates for this project called openstack-infra/pypi-mirror with these properties23:27
clarkbmrodden: it gives us ~5 jobs for free23:27
lifelessfungi: easier than telling everyone to run a fork of the thing23:27
lifelessfungi: undo your local changes when you apply this, of course.23:27
fungilifeless: so true23:28
jeblairlifeless: that change is not likely to work for us23:28
lifelessjeblair: why not?23:28
lifelessjeblair: (and how can I change it so that it will)23:28
jeblairlifeless: the domain (openstack.org) is not associated with the account most of our servers are in23:28
lifelessjeblair: oh, is the change deeper than just a nonsense-cli-parameter?23:29
mroddenclarkb: jenkins job builder is cool...23:29
jeblairlifeless: i believe it's passed through to the api23:29
jeblairmrodden: thanks :)23:29
lifelessok, looks like I need to eyeball the diff then.23:29
lifelessfungi: if you can get me it that would be great.23:29
lifelessfungi: the replacement of openstack.org with the domain should be ok for create though, right ?23:30
jeblairi'm not comfortable with this system endorsing or requiring rackspace23:31
clarkbmy make-index fix doesn't appear to be sufficient23:31
lifelessjeblair: it does today23:31
jeblairi think the reality of the situation is that openstack does not have a dns api, and so it is incomplete23:31
fungilifeless: http://paste.openstack.org/show/4570323:31
lifelessjeblair: I took a hard coded parameter (openstack.org) and made it inferred on the hostname being written to23:31
fungiand yes, this was/is a stop-gap until there is a good dns story for us, unfortunately23:32
jeblairlifeless: someday this system will be able to support dns23:32
lifelessjeblair: this doesn't seem to make it any more or less endorsed23:32
jeblairfungi: you published my unpublished code23:32
lifelessjeblair: nor make it harder to migrate to moniker23:32
fungid'oh, sorry :(23:32
clarkbjeblair: is ^ a race between zuul and mirroring?23:33
lifelessdelete it from the pastebin?23:33
jeblairlifeless: the launch/ script was never, as you have noticed, intended for consumption by anyone else; i'm happy you are helping to make it so, but i'm really not comfortable with the openstack infra project endorsing, requiring, encouraging, etc something provider-specific23:34
lifelessjeblair: so, I support your goal; I'd like to know how what I've done compromises it, so I can fix it. My goal is to not have to have a separate document that says 'and this is how you bugfix the openstack-ci infra'23:34
jeblairlifeless: if we want launch-node to be part of 'the product', i think it has to punt on dns for now, and say "do whatever you have to do in your environment to set up dns"23:35
clarkbjeblair: running `git remote update && git checkout 54db26ddd088675dcbc114db8c878eb21f72b0ee` by hand in that jenkins workspace works properly. I think there is a race. Now that we load balance across nodes we may need a different way to check for the presence of a ref on the mirrors in zuul23:36
lifelessjeblair: So there are two independent things here: there is 'describe the dns and rdns keys needed'23:36
jeblairlifeless: (and we'll have to move our code somewhere else, and hope no one notices it there until openstack has a dns service, because we have to maintain this thing in the environment we happen to be using)23:36
lifelessjeblair: and there is 'and here is a command to do it on provider $p.23:36
lifelessjeblair: Your code is already extant and public23:37
fungilifeless: jeblair: unpublished (and learned drizzle at the same time)23:37
lifelessjeblair: [launch.py and dns.py]23:37
*** vipul is now known as vipul-away23:37
lifelessjeblair: I *think* you're saying that there is a discontinuity between 'and here is a thing stuck in our puppet tree somewhere' and 'Robert stood up an infrastructure derived from that' ?23:38
jeblairlifeless: yes, and i'm glad that the discontinuity is being removed in general, but i think in this case there is not a good resolution to the discontinuity23:40
jeblairwe have to keep using this script on our systems because it's too painful not to, but i don't want to tell anyone else to use it23:40
lifelessjeblair: so, I don't think there is such a discontinuity, but I can understand your reservations.23:40
lifelessjeblair: so, I propose to add a CLI parameter for the name of the dns tool to use, and document the arguments we use.23:41
lifelessjeblair: that then breaks the visible link to rackspace.23:41
lifelessjeblair: without adding a hard dep on e.g. moniker23:41
lifelessjeblair: would that be sufficient, for now ?23:42
jeblairlifeless: i think so23:42
fungithe rackspace dns interface was never published much less upstreamed, so i second a desire to not use it once we have an option23:42
clarkbjeblair: zuul is checking against gerrit.baseurl which in this case is https://review.openstack.org. Should we change it to be https://git.openstack.org to at least reduce the chance of mirroring being behind zuul?23:42
lifelessanyone got a good off-hand reference for making self-signed ssl certs ?23:42
*** changbl has joined #openstack-infra23:42
jeblairlifeless: 1 sec23:43
*** vipul-away is now known as vipul23:44
jog0fungi: ping23:45
fungijog0: summit abstracts, yes23:46
jeblairlifeless: gah, sorry, i was thinking of csrs; i think i just google it every time23:46
fungijog0: i finally caught up far enough in my e-mail to find the "Rank Your Sessions" e-mail from the events system23:47
fungijog0: (was out of town last week)23:48
mordredclarkb: I just installed gear into a venv that has setuptools 1.1 in it and it worked fine23:48
jog0fungi: haha I just did it myself23:48
jog0I was wondering if you have done it yet, I guess not23:49
fungijog0: yeah, vacation-induced chairing failure on my part :/23:49
fungihoping tomorrow morning, probably23:49
clarkbmordred: did you have preinstalled pbr in that virtualenv?23:50
mordredclarkb: no23:50
mordredclarkb: should I?23:50
jog0fungi: so then I guess nati_ueno may have done some23:50
*** sarob has joined #openstack-infra23:50
clarkbmordred: no you shouldn't but if you did then maybe you wouldn't see the failure23:50
clarkbmordred: you can try with 0.5.17 preinstalled and try to go to 0.5.2123:50
clarkbmordred: as that was what was failing. I wonder if this setuptools bug is not deterministic23:50
openstackgerritlifeless proposed a change to openstack-infra/config: Phase 3 infra bootstrap docs: gerrit.  https://review.openstack.org/4497023:50
openstackgerritlifeless proposed a change to openstack-infra/config: Document review.pp parameters a bit.  https://review.openstack.org/4496923:51
openstackgerritlifeless proposed a change to openstack-infra/config: Non-openstack-ci support for launch/dns.py.  https://review.openstack.org/4498023:51
clarkbmordred: it has to do with lazy loading setuptools on top of distutils.Distribution23:51
*** sarob has quit IRC23:51
mordredclarkb: the bug looks like the 'setuptools got upgraded out of sequence' symptioms23:51
jeblairclarkb: yes, i was starting to look into that issue this morning, actually.  i'm thinking maybe ggp should fetch from zuul in all cases.23:51
lifelessjeblair: tada23:51
clarkbmordred: going through the setuptools install prcess to fix that out of sequence issue does not fix the problem23:51
clarkbmordred: atleast it didn't on the logstash workers23:52
jeblairclarkb: (the race)23:52
jgriffithclarkb: do you guys have any insight on these image/json empty attachment failures?  http://logs.openstack.org/73/44973/1/check/gate-tempest-devstack-vm-postgres-full/f7b9701/console.html23:52
clarkbjeblair: so do a git remote update to get as much as we can, then hit zuul for the particular ref?23:52
clarkbjeblair: wfm23:52
mordredclarkb: although I have found a new way to hate setuiptools just now23:52
jeblairclarkb: yeah, it'd actually be the exact same code for both kinds of checkouts23:52
jog0mordred fungi: I like how 3/4 tech track chairs work for or used to work for hp23:53
*** flaper87 has left #openstack-infra23:53
jgriffithclarkb: or is that cruft because the boot instance failed perhaps23:53
jog0fungi: so as far as I know the web app has no good way of resolving differences so we may have to resort to a email thread23:54
clarkbjgriffith: there are bugs on status.openstack.org/rechecks/ that seem related23:54
clarkbjgriffith: but that is all I know23:54
jgriffithclarkb: yeah, they seem to crop up in various spots23:55
jgriffithclarkb: kk.. thakns23:55
clarkbjgriffith: you might try over in #openstack-wa23:55
jgriffithclarkb: ahh... a new channel to add to my list :)23:55
clarkbI am trying to find this neutron change that keeps sudoing in its unittests and I can't find it23:56
mordredclarkb: ok. so, setuptools has no way to indicate that you want a setup_requires upgraded when you pip install -U -e .23:56
jeblairclarkb: actually, that won't work as-is (because it doesn't generate a zuul_ref); we'd need to switch to review for post jobs or modify zuul to generate zuul refs for all changes23:56
mordreddstufft: ^^ new fuel for the fire of having pip intercept setup_requires23:57
jeblairclarkb: we should perhaps do both of those in that order23:57
jeblairclarkb: (rather, generate zuul refs for ref-updated events)23:57
clarkbmordred: right, so you have to manually upgrade which is where I was having trouble23:57
clarkbclarkb: wfm23:57
clarkber jeblair ^23:58
clarkbjeblair: that way we are consistent regardless of pipeline23:58
clarkbwhich is nice when debugging23:58
jeblairclarkb: yep23:58
jeblairclarkb: i think that fetch should also get tags, right?23:58
jeblairclarkb: if we fetch refs/zuul/Z1234 and that points to a commit that has tags, the tags should be retrieved as well, right?23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!