| *** TravT_ has joined #openstack-sprint | 02:50 | |
| *** TravT has quit IRC | 02:51 | |
| *** TravT_ has quit IRC | 03:27 | |
| *** TravT has joined #openstack-sprint | 03:28 | |
| *** mrmartin has joined #openstack-sprint | 04:39 | |
| *** mrmartin has quit IRC | 04:54 | |
| *** mrmartin has joined #openstack-sprint | 05:03 | |
| *** mrmartin has quit IRC | 05:04 | |
| *** mrmartin has joined #openstack-sprint | 06:31 | |
| *** mrmartin has quit IRC | 06:45 | |
| *** mrmartin has joined #openstack-sprint | 06:53 | |
| *** mrmartin has quit IRC | 08:50 | |
| *** mrmartin has joined #openstack-sprint | 09:09 | |
| *** jhesketh has quit IRC | 09:39 | |
| *** jhesketh has joined #openstack-sprint | 09:41 | |
| *** pabelanger has joined #openstack-sprint | 10:59 | |
| pabelanger | morning | 10:59 |
|---|---|---|
| *** clayton has joined #openstack-sprint | 11:10 | |
| *** rfolco has joined #openstack-sprint | 11:54 | |
| *** mordred has quit IRC | 12:27 | |
| *** timrc has quit IRC | 12:28 | |
| *** timrc has joined #openstack-sprint | 12:33 | |
| *** ChanServ changes topic to "Puppet Liberty Mid-Cycle Sprint https://etherpad.openstack.org/p/puppet-liberty-mid-cycle" | 12:51 | |
| *** mwhahaha has joined #openstack-sprint | 13:14 | |
| *** TravT has left #openstack-sprint | 13:39 | |
| EmilienM | good morning | 13:42 |
| anteaya | EmilienM: pabelanger morning | 13:42 |
| *** vinsh has joined #openstack-sprint | 13:42 | |
| *** tristanC has joined #openstack-sprint | 13:43 | |
| *** fbo has joined #openstack-sprint | 13:45 | |
| *** richm has joined #openstack-sprint | 13:47 | |
| anteaya | say hello when you arrive so we are sure to include you | 13:54 |
| anteaya | EmilienM pabelanger and I are face to face so we want to be sure to include folks on things | 13:54 |
| vinsh | Hello. | 13:56 |
| EmilienM | we're talking about CI | 13:57 |
| EmilienM | https://etherpad.openstack.org/p/puppet-openstack-CI | 13:57 |
| *** mrmartin has quit IRC | 14:00 | |
| mwhahaha | morning folks | 14:07 |
| *** mrmartin has joined #openstack-sprint | 14:09 | |
| pabelanger | https://wiki.openstack.org/wiki/Infrastructure/Conferencing | 14:12 |
| pabelanger | we are in conference room 6000 | 14:12 |
| anteaya | vinsh mwhahaha and others you are welcome to join us on asterisk | 14:12 |
| anteaya | follow the wiki instructions above | 14:12 |
| *** nibalizer has joined #openstack-sprint | 14:13 | |
| nibalizer | o/ | 14:13 |
| anteaya | nibalizer: hey | 14:13 |
| anteaya | we are on asterisk if you want to join us | 14:13 |
| anteaya | https://wiki.openstack.org/wiki/Infrastructure/Conferencing | 14:13 |
| anteaya | conference room 6000 | 14:13 |
| anteaya | and currently are working on this etherpad: https://etherpad.openstack.org/p/puppet-openstack-CI | 14:14 |
| nibalizer | maybe | 14:14 |
| nibalizer | it is a bit early in the morning for me to phone | 14:14 |
| anteaya | understood | 14:14 |
| anteaya | we are talking about multinode | 14:14 |
| nibalizer | I'm also not sure what my puppet/infra/downstream work split is today | 14:15 |
| anteaya | EmilienM pabelanger and I | 14:15 |
| anteaya | nibalizer: understood | 14:15 |
| EmilienM | nibalizer, we're on https://etherpad.openstack.org/p/puppet-openstack-CI | 14:18 |
| * vinsh dialed in. | 14:20 | |
| anteaya | vinsh: welcome | 14:21 |
| anteaya | how is the audio? | 14:21 |
| vinsh | It sounds good! | 14:21 |
| anteaya | great | 14:21 |
| pabelanger | 609511 | 14:22 |
| pabelanger | oops | 14:22 |
| nibalizer | hh | 14:22 |
| nibalizer | heh | 14:22 |
| *** mfisch has joined #openstack-sprint | 14:23 | |
| EmilienM | mfisch: welcome | 14:23 |
| mfisch | hey | 14:24 |
| mfisch | I'm only half here dealing with a production issue | 14:25 |
| anteaya | ack | 14:27 |
| EmilienM | nibalizer: are you online? | 14:32 |
| EmilienM | on the call I mean | 14:32 |
| nibalizer | I am not on the call | 14:39 |
| anteaya | the zuul3 spec: http://specs.openstack.org/openstack-infra/infra-specs/specs/zuulv3.html | 14:40 |
| anteaya | right now I can find no code in any state regarding the zuul3 implementation | 14:41 |
| nibalizer | also i'd be happy to join the call to sync up as the sprint begins, but 3 days of phone call sounds pretty brutal tbh | 14:42 |
| anteaya | sure that is fine | 14:43 |
| nibalizer | I certainly can't spend all 3 days working on puppet, so if most of the chatter were here I could pop in when needed and read scrollback, if that makes sense | 14:43 |
| anteaya | we are taking notes in the etherpad | 14:43 |
| anteaya | nibalizer: glad to have you for what you can help with | 14:43 |
| anteaya | and if you have any thoughts about current puppet ci | 14:44 |
| anteaya | that is the current topic | 14:44 |
| nibalizer | ya sure | 14:44 |
| nibalizer | so for multinode, I'm guessing here, but a few weeks back someone (paul?) wanted a sepearte node to run tempest from | 14:44 |
| crinkle | I'm a little concerned that this sprint is happening over conference call, if it was happening in this channel as planned then the whole discussion could be logged and folks not in a convenient time zone could participate | 14:45 |
| nibalizer | and my thoughts on that are confused because devstack doesn't use a separate node for tempest, so why should we? | 14:45 |
| pabelanger | nibalizer: Right, that was a simple method of flexing multinode nodepool | 14:46 |
| EmilienM | crinkle: we can stop using the call - though we took notes on https://etherpad.openstack.org/p/puppet-openstack-CI | 14:46 |
| *** _ody has joined #openstack-sprint | 14:47 | |
| *** cdelatte has joined #openstack-sprint | 14:47 | |
| EmilienM | see "Discussion around muti-node" notes | 14:47 |
| mfisch | whats the next topic? | 14:49 |
| EmilienM | crinkle, nibalizer: during the infra sprint, did you use conferencing tool? | 14:49 |
| nibalizer | no we just chat in here | 14:49 |
| pabelanger | are we ready to move on? | 14:49 |
| EmilienM | ok | 14:49 |
| EmilienM | so for "Running OpenStack AIO", we discussed with Paul and everything is written in the etherpad | 14:50 |
| EmilienM | any thoughts ? | 14:51 |
| pabelanger | well, I think virtual people have questions about multinode / single node? | 14:52 |
| EmilienM | single node is for now making progress using scenario001.pp but we might need to split the manifest if running multiple scenarios | 14:53 |
| anteaya | vinsh: http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/projects.yaml#n3257 | 14:54 |
| vinsh | TY | 14:54 |
| anteaya | vinsh: let me get you more | 14:54 |
| anteaya | welcome | 14:54 |
| EmilienM | crinkle: for single node, I would like last reviews on https://review.openstack.org/#/q/status:open+project:openstack/puppet-openstack-integration+owner:%22Emilien+Macchi+%253Cemilien%2540redhat.com%253E%22,n,z | 14:54 |
| EmilienM | so we would enable the jobs for puppet-glance,nova,neutron | 14:55 |
| nibalizer | pabelanger: ya I don't understand the multinode need? | 14:55 |
| nibalizer | infra doesn't have infinite resources so I think we need compelling reasons to use multinode | 14:55 |
| nibalizer | plus it is much more complicated | 14:55 |
| pabelanger | nibalizer: for me, it is the real world example of testing | 14:55 |
| pabelanger | nobody runs a cloud on a single node | 14:55 |
| EmilienM | anteaya: https://review.openstack.org/#/c/219556/ was failing to land, I rebased it, can you +A again? | 14:56 |
| anteaya | vinsh: http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/devstack-gate.yaml#n1889 | 14:56 |
| mfisch | single node is still a good starting point test though | 14:56 |
| nibalizer | so the goal is to have a separate compute node? | 14:56 |
| mfisch | of course anything is better than what we had before you guys started on it | 14:56 |
| crinkle | EmilienM: the integration tests are failing for 3 out of 4 of those | 14:56 |
| EmilienM | crinkle: because of packaging issue that we faced | 14:57 |
| nibalizer | it says zuulv3 right now, which doesn't exist in code form | 14:57 |
| pabelanger | so, my question is, what is the downside of having multinode? | 14:57 |
| EmilienM | crinkle: if can point to you latest working logs | 14:57 |
| * EmilienM look for URL | 14:57 | |
| anteaya | vinsh: http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/openstack_functions.py#n48 | 14:57 |
| mfisch | from a puppet pov whats the upside of a separate node? | 14:57 |
| clayton | I assume the issue is that it chews up x instances instead of 1 instance and there is a hard cap on the total instances nodepool can use at any given time | 14:57 |
| crinkle | EmilienM: okay but it seems silly to merge a patch about tests when the tests its adding are failing | 14:57 |
| mwhahaha | testing cross system dependancies | 14:57 |
| anteaya | vinsh: those parts should at least get you a little bit of context, and be a jumping off place for more questions | 14:57 |
| mwhahaha | ie keystone endpoint creation on a machine that is not the keystone server | 14:58 |
| EmilienM | crinkle: I agree with you, though we are not going to change the patches because of packaging issues, they are not directly related | 14:58 |
| vinsh | anteaya: Very good. I'll study. | 14:58 |
| nibalizer | pabelanger: I think 2 nodes with one controller and one compute is a good idea | 14:58 |
| anteaya | vinsh: great thanks | 14:58 |
| EmilienM | crinkle: http://logs.openstack.org/52/217352/34/check/gate-puppet-openstack-integration-dsvm-centos7/968c4af/ | 14:58 |
| nibalizer | I think a separte node for tempest is not | 14:58 |
| EmilienM | crinkle: from https://review.openstack.org/#/c/217352/ | 14:58 |
| mfisch | yes why would tempest need a node? | 14:58 |
| mfisch | its just making api calls | 14:58 |
| pabelanger | Right, tempest is not an issue right now. Having tempest on a multinode was the simplest way to show multinode working | 14:58 |
| pabelanger | new plan would be to have controller and compute node | 14:59 |
| pabelanger | then, run tempest some place | 14:59 |
| *** krotscheck is now known as kro_afk | 14:59 | |
| mfisch | ok | 14:59 |
| EmilienM | crinkle: I can run the jobs again so we can make sure it pass centos7 at least | 15:00 |
| crinkle | EmilienM: okay | 15:00 |
| crinkle | EmilienM: we could just wait till the packaging issue is fixed | 15:00 |
| pabelanger | AFK, getting drink | 15:00 |
| EmilienM | cool | 15:00 |
| EmilienM | crinkle: lol I hope you don't want to wait for trusty | 15:00 |
| mwhahaha | another point of multinode is that no one runs single nodes in production | 15:01 |
| crinkle | what's the rush? | 15:01 |
| EmilienM | crinkle: see my recent patches in infra, I disabling voting for all jobs that are failing | 15:01 |
| mwhahaha | so we should try and reproduce at least a little separation | 15:01 |
| EmilienM | crinkle: there is no rush I guess, we just want to make progress and iterate our work | 15:01 |
| pabelanger | mwhahaha: exactly | 15:03 |
| crinkle | right but progress in writing tests that don't give us any useful information because of packaging issues | 15:03 |
| *** kro_afk is now known as krotscheck | 15:04 | |
| EmilienM | crinkle: I'm trying to keep centos7 green | 15:05 |
| nibalizer | pabelanger: I think mulitnode is working, and not puppet's job to show that it is or not | 15:05 |
| EmilienM | crinkle: I consider ubuntu unstable now | 15:05 |
| nibalizer | but yea a 2 node multinode test with separte controller I am in favor of | 15:06 |
| pabelanger | nibalizer: working where? | 15:06 |
| EmilienM | nibalizer: controller, compute, network, etc | 15:06 |
| pabelanger | How did packaging for distros get decided to gate on? | 15:07 |
| pabelanger | eg: rdo, ubuntu | 15:07 |
| vinsh | I think it was just pulling latest. right? | 15:07 |
| nibalizer | pabelanger: nodepool multinode testing works? clark has had multinode devstack things going for a while | 15:07 |
| crinkle | we test with packages from distros because those are the packages we support | 15:08 |
| pabelanger | could you consume debian packages, from debian | 15:08 |
| pabelanger | for the gate | 15:08 |
| crinkle | we would still need to test ubuntu | 15:08 |
| EmilienM | right, ubuntu:UCA and centos7:RDO | 15:08 |
| EmilienM | crinkle: right, but during the period where packages are really broken for trusty, I'm not sure we should bump to liberty or keep jobs voting | 15:09 |
| EmilienM | liberty or trunk | 15:09 |
| EmilienM | because it's blocking our CI | 15:09 |
| crinkle | CI is there to be blocked | 15:09 |
| crinkle | if we're going to ignore the tests then why are we writing tests | 15:09 |
| EmilienM | and keeping a backlog of bunch of current patches, is I think a mess to manage | 15:09 |
| pabelanger | It would be worth the effort to get RDO to move packages upstream | 15:09 |
| EmilienM | crinkle: we are not going to ignore | 15:10 |
| EmilienM | crinkle: what is happening now, is, we are ignoring trusty vote for 4 modules | 15:10 |
| EmilienM | we still test them, and still use centos7 to vote | 15:10 |
| pabelanger | Right having downstream packages block our CI, a big issue. Especially if downstream never signed off to support them for us. How we notify them of issues and SLA for fixes? | 15:11 |
| EmilienM | crinkle: this is an issue we will have in the next cycle again | 15:12 |
| EmilienM | crinkle: people want to push features in puppet modules to test trunk - how can we test it if we don't download trunk packages? | 15:13 |
| pabelanger | nibalizer: Yes, starting with a 2 node setup should be fine. I actually have some JJB code up for an experimental job | 15:13 |
| pabelanger | https://review.openstack.org/#/c/215696/ | 15:13 |
| *** richm has quit IRC | 15:14 | |
| crinkle | pabelanger: but these are the packages we support, we write the modules based on decisions these packagers have made, so we need to run CI with them because that is how we're developing | 15:14 |
| EmilienM | crinkle: "run the CI with them" > can you explain? Like keeping synced with their status, etc? | 15:15 |
| crinkle | EmilienM: i mean what we're doing now | 15:15 |
| EmilienM | I'm not sure Ubuntu folks is willing to help | 15:15 |
| crinkle | we install their packages in our ci | 15:15 |
| EmilienM | honestly | 15:15 |
| pabelanger | crinkle: Right, I don't dispute that. How do you guarantee ubuntu will ever get a fix? | 15:15 |
| EmilienM | they don't even reply to my IRC pings | 15:15 |
| pabelanger | Or are you expecting somebody on this team to do the package fix? | 15:15 |
| crinkle | pabelanger: so we should stop supporting ubuntu in our modules? | 15:15 |
| pabelanger | no | 15:16 |
| EmilienM | no | 15:16 |
| pabelanger | I want to understand who at ubuntu is listening to us report problems | 15:16 |
| EmilienM | I just feel we spend too much time on packaging issues | 15:16 |
| mfisch | +1 | 15:16 |
| nibalizer | EmilienM: whats the alternative | 15:16 |
| EmilienM | I have to feeling we are making the job of packaging folks | 15:17 |
| mfisch | you cannot guarantee a fix nor can you guarantee that they wont break again in M | 15:17 |
| EmilienM | because we 1/ install trunk packaging 2/ have a gate - they don't have 2/ | 15:17 |
| EmilienM | nibalizer: well, I see two 3 options, maybe there is more, tell me: | 15:17 |
| EmilienM | 1/ use stable/* until OpenStack is released. So we are not pushing any trunk feature during the cycle | 15:18 |
| EmilienM | this is not the best option, people will fork our modules and push their features | 15:18 |
| clayton | is there any existing tooling for generating reports for gate tests passing? | 15:18 |
| mfisch | have we tried a discussion with the ubuntu packagine guys? | 15:19 |
| EmilienM | mfisch: LOL | 15:19 |
| pabelanger | Right | 15:19 |
| clayton | we could make ubuntu non-voting and consider putting back if it passes regularly | 15:19 |
| EmilienM | that's my proposal in 3/ but I'm not here yet | 15:20 |
| crinkle | it seems slightly backwards to me to expect stability from something that hasn't been released yet | 15:20 |
| EmilienM | well, if they had a CI that would be at least usable | 15:20 |
| EmilienM | we can't even do apt-get install heat today | 15:20 |
| mfisch | I agree not to expect stability but ignoring bugs that are filed is frustratin | 15:20 |
| anteaya | crinkle: that is a good point | 15:21 |
| EmilienM | even if OpenStack is bugged, we should at least being able to do apt-get install heat | 15:21 |
| mwhahaha | imo, we should have the previous openstack package used for voting jobs and the current (dev) packages there but non-voting | 15:22 |
| pabelanger | Ya, I don't have a good solutions. But this should all work under kilo right? | 15:22 |
| pabelanger | otherwise, people have broken production systems | 15:22 |
| EmilienM | so another option is 2/ use trunk and use zuul to vote or not if package is working or not | 15:23 |
| EmilienM | and keep at least one distro voting | 15:23 |
| EmilienM | so we make sure we continue to use beaker jobs part of our voting system | 15:23 |
| pabelanger | Ya, I think we need to atleast reach out to RDO and ubuntu about us wanting to consume there packages for CI. | 15:25 |
| pabelanger | Maybe some point of contact to keep abreast of changes | 15:25 |
| EmilienM | I invited Ubuntu packagers to join here | 15:26 |
| EmilienM | so we can discuss | 15:26 |
| EmilienM | James Page is about to join probably | 15:26 |
| *** jamespage has joined #openstack-sprint | 15:27 | |
| EmilienM | jamespage: welcome here, we are in the first day of Puppet sprint | 15:27 |
| mwhahaha | xarses did have this mail back in july which seems relevant to this conversation again http://lists.openstack.org/pipermail/openstack-dev/2015-July/069123.html | 15:27 |
| EmilienM | jamespage: we are talking about our current issues with Liberty packaging | 15:27 |
| EmilienM | and the alternatives for the next cycle | 15:27 |
| EmilienM | jamespage: we have two options, use stable packaging until openstack is released (this is not the best option, people will fork our modules and push their features that we could not test upstream beacuse of packaging) | 15:28 |
| EmilienM | and the second one is use trunk packages (liberty staging in your case) and disable voting if packaging is really unstable | 15:29 |
| EmilienM | but keep at least one distro voting | 15:29 |
| EmilienM | so we can actually test our patches | 15:29 |
| *** spredzy has joined #openstack-sprint | 15:29 | |
| jamespage | EmilienM, its sounds like you need some level of stability around the packaging; *staging* is not the right area todo that with | 15:29 |
| pabelanger | so, we are thinking about lunch in 30mins. Does that break work for people? | 15:29 |
| EmilienM | jamespage: you were the one who suggested me staging if I wanted to run liberty | 15:30 |
| *** richm has joined #openstack-sprint | 15:30 | |
| jamespage | EmilienM, that's cause you wanted it before we had done any testing on liberty-1 | 15:30 |
| jamespage | it was the only place we had it right then | 15:30 |
| EmilienM | jamespage: do you run a CI that deploy your packages and test OpenStack ? | 15:30 |
| jamespage | EmilienM, we do and we know that right now things are broken; and we are working to fix that | 15:31 |
| jamespage | EmilienM, in staging | 15:31 |
| mfisch | jamespage: welcome | 15:32 |
| jamespage | once we're happy with the packages in staging, they get promoted to proposed and then updates | 15:32 |
| jamespage | mfisch, hey! | 15:32 |
| EmilienM | jamespage: when is that happen? | 15:33 |
| mfisch | jamespage: I think one of the frustrations is how we can point out packaging issues to you, I'm also happy to submit patches time permitting, but bugs soemtimes end up lost in the shuffle | 15:33 |
| EmilienM | https://etherpad.openstack.org/p/puppet-liberty-blocker | 15:33 |
| EmilienM | they have been for 1 month and a half ^ | 15:33 |
| mfisch | I think LP bugs are easily lost | 15:34 |
| *** med_ has joined #openstack-sprint | 15:35 | |
| EmilienM | crinkle, nibalizer: going back to alternatives, what do you suggest? | 15:35 |
| jamespage | EmilienM, when went offline a few weeks back you only had one outstanding issue | 15:35 |
| jamespage | EmilienM, I see new issues there due to some problems we've had pushing liberty-2 in to the cloud archive | 15:35 |
| EmilienM | that's not true, Heat, Ironic and Glance are broken for a while | 15:36 |
| nibalizer | clayton: I made this which is a hack http://spencerkrum.com/openstack/ci/ | 15:36 |
| EmilienM | pabelanger: ^ | 15:37 |
| nibalizer | clayton: anteaya runs things against elasticsearch | 15:37 |
| mfisch | Guys, if we start from the idea that both of us, puppet & Ubuntu cloud can gain from working together we'll have a better conversation | 15:37 |
| mfisch | We're testing packages all the time and can help find bugs | 15:37 |
| med_ | +1 | 15:38 |
| EmilienM | mfisch: yeah but until now, it's 1/ taking a lot of time who actually dig in beaker failures and report bugs to Launchpad 2/ block our CI | 15:38 |
| mfisch | Yes we have some issues | 15:38 |
| EmilienM | I can't handle 1/ in the next cycle anymore | 15:38 |
| EmilienM | and I don't want 2/ neither | 15:38 |
| mfisch | I think we have 2 main concerns: which repo can we use and how do we better communicate issues and status back & forth | 15:38 |
| jamespage | EmilienM, we need to switch your CI away from staging and to either proposed or updates | 15:39 |
| mfisch | jamespage: sorry if I missed this but is there a liberty proposed/updates? | 15:39 |
| jamespage | EmilienM, right now you're getting stung by some transitory packaging issues due to the way we have to managed the backport to the cloud archive | 15:39 |
| jamespage | mfisch, yes | 15:39 |
| EmilienM | jamespage: we won't test current openstack | 15:39 |
| EmilienM | ie liberty | 15:39 |
| mfisch | we need a balance between "really new" and "working" | 15:40 |
| jamespage | agreed | 15:40 |
| EmilienM | yeah | 15:40 |
| EmilienM | +1 for working | 15:40 |
| EmilienM | we don't mind having last commit from openstack | 15:40 |
| EmilienM | but a package dated from one month could be acceptable | 15:40 |
| mfisch | EmilienM: what about liberty/updates/proposed liek jamespage said? | 15:40 |
| jamespage | EmilienM, to give you an idea, libery-1 was ~90 packaging touchpoints and took about 2 weeks to process | 15:40 |
| jamespage | that's just getting the packaging bits done and through process | 15:40 |
| EmilienM | jamespage: so you're suggesting to use proposed/updates only - 1/ since when could I use this repo to install liberty? 2/ is it tested enough and does it work today? | 15:41 |
| pabelanger | nibalizer: interesting | 15:41 |
| pabelanger | nibalizer: oops | 15:41 |
| pabelanger | EmilienM: interesting | 15:41 |
| jamespage | EmilienM, proposed currently contains the b1 packaging that I think was 'best' | 15:41 |
| jamespage | EmilienM, staging is volatile while we work through the last of b2 - and then we'll run straight into b3 | 15:42 |
| jamespage | unfortunately people like to take holiday | 15:42 |
| EmilienM | ok so the proposal is to use liberty/updates/proposed as soon as it's available | 15:42 |
| EmilienM | jamespage: I like holidays too ;-) | 15:42 |
| mfisch | liberty-b1 is probably okay | 15:43 |
| EmilienM | and you say liberty/updates/proposed should work | 15:43 |
| jamespage | EmilienM, I think that will give you a better balance between current/stable | 15:43 |
| EmilienM | ok | 15:43 |
| crinkle | \o/ | 15:43 |
| EmilienM | so I propose, in the next cycle, we bump our CI as soon liberty/updates/proposed is here | 15:43 |
| mfisch | hate to bring this up but what will we do after L ships | 15:43 |
| jamespage | EmilienM, to the best of my knowledge yes - we're just starting on the next round of testing and updates for the juju charms we use to deploy openstack for our CI | 15:43 |
| jamespage | we will find problems | 15:43 |
| mfisch | will we have m1 by the time we branch liberty to stable? | 15:44 |
| crinkle | mfisch: we can just wait a bit | 15:44 |
| mfisch | makes sense to me | 15:44 |
| EmilienM | yeah | 15:44 |
| EmilienM | liberty/updates/proposed is out, wait a bit (run some tests if we want and see how it behaves) | 15:44 |
| EmilienM | exactly what we did but for staging... | 15:45 |
| mfisch | jamespage: so about communication, I'm not sure LP is working very well for packaging bugs. besides emailing you whats the best way to communicate with ubuntu cloud team? | 15:45 |
| jamespage | mfisch, actually that's something we're working on right now | 15:45 |
| EmilienM | good to know | 15:45 |
| anteaya | jamespage: do you have any expectation of having more people helping with packageing? | 15:46 |
| jamespage | LP bugs can be a fire hose, so we've built out a specific openstack view to make that easier for triage etc... | 15:46 |
| mfisch | jamespage: perfect | 15:46 |
| anteaya | jamespage: as in more that just you and zul? | 15:46 |
| mfisch | that keeps them from getting lost | 15:46 |
| jamespage | anteaya, my packaging team is three - coreycb, zul and me | 15:46 |
| anteaya | jamespage: yay | 15:46 |
| anteaya | I have yet to meet you | 15:46 |
| jamespage | anteaya, we have open headcount to general increase engineering resources in the team | 15:46 |
| anteaya | will you and coreyb be at summit? | 15:46 |
| anteaya | I'm expecting to see zul there | 15:46 |
| jamespage | anteaya, and just to be clear we don't *just* do packaging | 15:46 |
| anteaya | ah sorry | 15:47 |
| EmilienM | jamespage: can you point me an example of apt config file to run liberty/updates/proposed please? | 15:47 |
| jamespage | I'll be their - ditto corey | 15:47 |
| anteaya | my mistake | 15:47 |
| anteaya | thanks for the clarification | 15:47 |
| mfisch | deb http://ubuntu-cloud.archive.canonical.com/ubuntu trusty-updates/liberty main? | 15:47 |
| jamespage | EmilienM, yes one second | 15:47 |
| anteaya | and yay for more folks joining your team! | 15:47 |
| anteaya | \o/ | 15:47 |
| EmilienM | sounds like we have a plan | 15:47 |
| jamespage | mfisch, that sounds right | 15:47 |
| jamespage | mfisch, also need the ubuntu-cloud-keyring package installed | 15:47 |
| mfisch | yep | 15:47 |
| jamespage | EmilienM, stick with trusty-proposed/liberty for noew | 15:48 |
| anteaya | jamespage: I look forward to meeting you and coreyb | 15:48 |
| EmilienM | jamespage: ok, I'm about to patch our CI to run trusty-proposed/liberty today and see how it behaves. If CI works, I'll ask for our team to approve patches | 15:49 |
| EmilienM | crinkle, mfisch: does it work for your? | 15:49 |
| EmilienM | you* | 15:49 |
| jamespage | EmilienM, ack | 15:49 |
| crinkle | EmilienM: yep this seems like the best solution | 15:49 |
| mfisch | EmilienM: I didnt try I was basing it on my kilo entry | 15:49 |
| EmilienM | crinkle: cool, we have a consensus | 15:49 |
| EmilienM | I honestly hope trusty-proposed/liberty is working :) | 15:50 |
| mfisch | jamespage: whats the official status of openstack-client? still universe? | 15:51 |
| jamespage | EmilienM, fingers crossed - like I said we're focussing on updating charms for b2 once we have the packaging complete and to a certain level | 15:51 |
| jamespage | mfisch, nope main | 15:51 |
| mfisch | jamespage: awesome, its a cornerstone of some of the puppet needs for api calls | 15:52 |
| mwhahaha | hey for bug triaging, can I have access to update bug's severity? or should I just update the etherpad and someone else can do i t | 15:54 |
| mfisch | mwhahaha: we're going to discuss them as a group | 15:55 |
| EmilienM | mwhahaha: I would sync with mfisch | 15:55 |
| mfisch | someone here will have access I think I do | 15:55 |
| mwhahaha | K i'll make notes in the etherpad | 15:55 |
| mfisch | mwhahaha: some of those bear discussions | 15:56 |
| mwhahaha | yup | 15:56 |
| mfisch | especially the first 2 | 15:57 |
| mfisch | the keystone ones can be looked at now | 15:57 |
| med_ | is audio still audiating? | 15:57 |
| EmilienM | crinkle: I think I'm about to submit a patch to openstack_extras to support this use case, wdyt? | 15:58 |
| jamespage | EmilienM, the init-system-helpers problems should be resolved for keystone - I'm just looking at other potential impacts | 15:59 |
| jamespage | sahara and manila where impacted as well | 15:59 |
| Hunner | Morning! | 15:59 |
| EmilienM | jamespage: ok - I'm waiting for your apt config to download liberty so I'm sure I'm doing well | 16:00 |
| Hunner | _ody: In the office yet? | 16:00 |
| nibalizer | Hunner: o/ | 16:00 |
| nibalizer | Hunner: _ody said he'd be in the office around nine | 16:00 |
| nibalizer | i'll be heading down there in a few minutes | 16:00 |
| jamespage | EmilienM, "deb http://ubuntu-cloud.archive.canonical.com/ubuntu trusty-proposed/liberty main" | 16:00 |
| jamespage | and install ubuntu-cloud-keyring package | 16:01 |
| Hunner | nibalizer: YAAAAY | 16:01 |
| EmilienM | jamespage: and http://ubuntu-cloud.archive.canonical.com/ubuntu/dists/trusty-updates/liberty/ ? | 16:03 |
| jamespage | EmilienM, you only need updates or proposed, not both | 16:04 |
| EmilienM | ok | 16:04 |
| EmilienM | jamespage: thanks, I'll use proposed then | 16:04 |
| jamespage | EmilienM, yes as updates has nothing liberty in it yet | 16:04 |
| EmilienM | pabelanger, anteaya and I are afk for lunch - happy hacking :-) | 16:05 |
| _ody | Hunner: Yeah. Just came looking for you. | 16:07 |
| _ody | I am in the kitchen now. Need some espressso. | 16:07 |
| mfisch | are we on break? | 16:14 |
| _ody | Hunner: I am in "Big Bird"...I booked the room rest of the week so we do not have to move around if we don't want to. | 16:18 |
| Hunner | _ody: okay, over in a minute | 16:18 |
| _ody | nibalizer: That is the room right behind you if you get to our front desk. | 16:19 |
| mfisch | Hunner: I hope you have your ginormous glass of milk there | 16:23 |
| Hunner | mfisch: I wish :( | 16:24 |
| crinkle | EmilienM: can't we just pass in $source_hash? | 16:25 |
| jamespage | EmilienM, the glanceclient problem should not be impacting the liberty proposed UCA pocket | 16:41 |
| jamespage | as it does not have the newer urllib3 yet | 16:41 |
| Hunner | _ody: https://docs.google.com/spreadsheets/d/1Q8Db0jrv3dZQ8FbKhwpIjJRkI29Q571-FGs6VD_mVSQ/edit#gid=677531679 | 16:59 |
| EmilienM | jamespage: ok perfect | 17:01 |
| EmilienM | crinkle: you're right :) I'll test | 17:01 |
| Hunner | _ody: https://tldrlegal.com/ | 17:17 |
| EmilienM | crinkle: so I still need to patch manifests/repo/debian/ubuntu.pp | 17:28 |
| EmilienM | crinkle: to benefit the "installing ubuntu-cloud-keyring" exec | 17:28 |
| pabelanger | just added info about puppet unit test refactor, if people want to review | 17:34 |
| anteaya | I'm back | 17:34 |
| EmilienM | pabelanger: lgtm | 17:40 |
| EmilienM | it will save nodes consumption | 17:41 |
| EmilienM | and time | 17:41 |
| EmilienM | crinkle: https://review.openstack.org/219800 | 17:44 |
| EmilienM | and https://review.openstack.org/219804 | 17:49 |
| EmilienM | richm: I know you there is a lot of pending patches for keystone v3 work - please go ahead here if you need reviews | 17:52 |
| nibalizer | pabelanger: EmilienM how can I help with the CI progresS? | 17:52 |
| EmilienM | nibalizer: pabelanger is figuring out to consolidate unit test jobs in 2 jobs | 17:52 |
| EmilienM | maybe without a bash script would be cool | 17:53 |
| nibalizer | sweet | 17:53 |
| EmilienM | using JJB | 17:53 |
| EmilienM | one job (voting) to test the versions that we *support* | 17:53 |
| EmilienM | one job (non-voting) to test the latest version that we *don't support yet* but will | 17:53 |
| nibalizer | okay | 17:53 |
| EmilienM | nibalizer: see https://etherpad.openstack.org/p/puppet-openstack-CI | 17:53 |
| nibalizer | these two got created: http://git.openstack.org/cgit/?q=beaker I might build those out | 17:53 |
| EmilienM | pabelanger and nibalizer can work on this together | 17:53 |
| nibalizer | ok | 17:54 |
| nibalizer | these are the rspec unit tests? | 17:54 |
| EmilienM | nibalizer: yes | 17:54 |
| nibalizer | ok | 17:56 |
| pabelanger | what is beaker-nodepool? | 17:56 |
| EmilienM | nibalizer: good question ^ | 17:57 |
| nibalizer | so those are two repos I created | 17:58 |
| nibalizer | with no content, thinking that we would create beaker hypervisors for our use | 17:58 |
| nibalizer | one, beaker-localhost, would standardize and extend the pattern we are using where beaker sshes into itself | 17:58 |
| nibalizer | the other, beaker-nodepool would use either the extra ips in /etc/nodepool for multinode or the (soon to be) nodepool rest api | 17:58 |
| clayton | interesting, you have more details on the nodepool rest api? | 17:59 |
| EmilienM | jamespage, crinkle: it works | 18:03 |
| EmilienM | https://jenkins02.openstack.org/job/gate-puppet-keystone-puppet-beaker-rspec-dsvm-trusty/149/console | 18:03 |
| mfisch | mwhahaha: I have some time to do triage now if you want to dive in we can do it in the main channel | 18:03 |
| EmilienM | for keystone | 18:03 |
| mwhahaha | mfisch: i've got a meeting for 30 right now, you free in 30? | 18:03 |
| EmilienM | mfisch: can you review https://review.openstack.org/#/c/219800 please ? | 18:04 |
| mfisch | mwhahaha: probably free but only until 2pm my time | 18:04 |
| mwhahaha | looks like my meeting is canceled | 18:04 |
| mwhahaha | so let's do this | 18:05 |
| mfisch | EmilienM: looking | 18:05 |
| mfisch | EmilienM: its possible to be something besides updates and proposed too but not currently that way | 18:06 |
| mfisch | since you dont enforce that, +2 | 18:06 |
| mfisch | mwhahaha: I'd like to talk about: https://bugs.launchpad.net/puppet-keystone/+bug/1477093 | 18:07 |
| openstack | Launchpad bug 1477093 in puppet-keystone "Creating Keystone users with a password in the puppet module (Kilo) throws error at second puppetrun" [High,New] | 18:07 |
| mfisch | because its hit me too | 18:07 |
| mfisch | its not necessarily the 2nd puppet run | 18:07 |
| mfisch | but if you source an admin openrc and run puppet it blowsup | 18:07 |
| mwhahaha | So does it happen with a normal puppet run (not manual) | 18:08 |
| mfisch | unsure | 18:08 |
| mfisch | let me see if I can repro it in my env | 18:08 |
| mfisch | I was using a test env | 18:08 |
| mfisch | it blows up trying to get a token | 18:09 |
| mfisch | although the password is 100% right just like he says | 18:09 |
| mwhahaha | and the password is right in the openrc? | 18:09 |
| mwhahaha | os it's a priority thing | 18:09 |
| mfisch | os? | 18:09 |
| mwhahaha | where openrc > parameters | 18:09 |
| mwhahaha | s/os/or | 18:09 |
| mfisch | except the password didnt change | 18:09 |
| mfisch | openrc password matches puppet | 18:09 |
| mfisch | I had them both set to something like "admin_pass" | 18:10 |
| mfisch | at first I was seeing the same "2nd run" until I dropped the shell and came back and it went away | 18:10 |
| mfisch | I asked Van in email about this and this is what he said | 18:10 |
| mfisch | "Found a “sort of” solution. | 18:10 |
| mfisch | If you set "replace_password => false” it will not complain." | 18:10 |
| mfisch | so that implies to me maybe that puppet is trying to change the password even if its the same? | 18:11 |
| mfisch | can you repro? | 18:11 |
| mwhahaha | the provider can't know if the password is the same | 18:12 |
| mwhahaha | so i bet it's that | 18:12 |
| mfisch | right | 18:12 |
| mfisch | I cannot repro using our code but we're on master | 18:12 |
| mfisch | I do see it on stable/kilo | 18:13 |
| mwhahaha | i bet it's the auth url | 18:14 |
| mfisch | oh | 18:15 |
| mwhahaha | https://ask.openstack.org/en/question/68565/error-openstack-the-resource-could-not-be-found-http-404-openstack-user-list/ | 18:15 |
| mfisch | in openrc auth url is port 5000 | 18:15 |
| mfisch | does puppet use 35357 | 18:15 |
| mfisch | let me repro and see | 18:15 |
| mfisch | I can make it happen | 18:15 |
| mwhahaha | it may also be a v3 vs v2 issue | 18:15 |
| mfisch | let me unset OS_AUTH_ | 18:15 |
| mfisch | yes | 18:15 |
| mfisch | repro | 18:16 |
| mfisch | Error: /Stage[main]/Neutron::Keystone::Auth/Keystone::Resource::Service_identity[neutron]/Keystone_user[neutron]: Could not evaluate: Execution of '/usr/bin/openstack token issue --format value' returned 1: ERROR: openstack The resource could not be found. (HTTP 404) (Request-ID: req-2d82845c-a89e-4d6a-b5ab-fa506b71bf6d) | 18:16 |
| mfisch | I will clear OS_AUTH_ | 18:17 |
| richm | EmilienM: re: pending reviews for v3/federation | 18:17 |
| richm | https://review.openstack.org/#/c/208054 - latest patch hacks the wsgi class to add a custom fragment to the conf - this seems wrong, but I'm not sure how else he can do it | 18:17 |
| richm | https://review.openstack.org/#/c/218044/ - not quite ready | 18:18 |
| richm | https://review.openstack.org/#/c/218059/ - not quite ready | 18:18 |
| EmilienM | crinkle, richm, mfisch: so the bump to trusty-proposed/liberty worked for puppet-keystone: https://review.openstack.org/#/c/219804/ | 18:18 |
| mfisch | mwhahaha: confirmed, no repro with OS_AUTH_URL unset | 18:18 |
| mwhahaha | so did openrc have a v2 url? | 18:19 |
| mfisch | yes I only use V2 | 18:19 |
| mfisch | for everything | 18:19 |
| mwhahaha | and our provider assumes v3 | 18:19 |
| mfisch | but V2 works | 18:19 |
| mfisch | keystone token-get works | 18:19 |
| mfisch | I was thinking it was a port issue | 18:19 |
| mfisch | 5000 vs 35357 | 18:19 |
| mfisch | let me try that | 18:19 |
| richm | https://review.openstack.org/#/c/218281/ - please review the spec/bp | 18:19 |
| EmilienM | richm: ok | 18:20 |
| EmilienM | richm: I dropped a comment this moring already | 18:20 |
| richm | https://review.openstack.org/218688 - not quite ready | 18:20 |
| richm | EmilienM: yes, I saw that - thanks | 18:20 |
| richm | https://review.openstack.org/219084 - please review | 18:20 |
| mfisch | EmilienM: I thought we needed liberty/updates | 18:20 |
| richm | https://review.openstack.org/213957 - please review | 18:20 |
| mfisch | mwhahaha: guess what, it works | 18:20 |
| mfisch | wait nm | 18:21 |
| mfisch | spoke too soon | 18:21 |
| richm | https://review.openstack.org/202409 - not quite ready | 18:21 |
| EmilienM | hum | 18:21 |
| richm | https://review.openstack.org/#/c/219127/ - not quite ready | 18:21 |
| EmilienM | <jamespage> [12:04:39] EmilienM, yes as updates has nothing liberty in it yet | 18:21 |
| EmilienM | <jamespage> [12:00:52] EmilienM, "deb http://ubuntu-cloud.archive.canonical.com/ubuntu trusty-proposed/liberty main" | 18:21 |
| EmilienM | so I think I'm right using proposed | 18:21 |
| mfisch | k | 18:22 |
| EmilienM | jamespage: ^ | 18:22 |
| EmilienM | can you confirm? | 18:22 |
| mfisch | updates is more stable right? | 18:22 |
| richm | https://review.openstack.org/#/c/176150/ - please review | 18:22 |
| mfisch | mwhahaha: no repro with this: export OS_AUTH_URL=http://127.0.0.1:5000/v3/ | 18:22 |
| EmilienM | mfisch: right | 18:23 |
| EmilienM | mfisch: updates will be the one to use when jamespage will finish to upload all liberty | 18:23 |
| EmilienM | at the end | 18:23 |
| EmilienM | iiuc | 18:23 |
| mwhahaha | mfisch: but it repro's if OS_AUTH_URL is v2 right? | 18:23 |
| mfisch | yes | 18:23 |
| mfisch | no matter what port | 18:23 |
| EmilienM | anteaya: https://docs.google.com/presentation/d/1-AS5rYOdq0OVulPHGP3-w5ns3Lz2i3QNBj8pvPQZLkI/edit#slide=id.p | 18:24 |
| mwhahaha | mfisch: https://github.com/openstack/puppet-keystone/blob/master/lib/puppet/provider/keystone_user/openstack.rb#L106 | 18:24 |
| mwhahaha | i think it's that, we need to make sure the auth url is v3 | 18:24 |
| mfisch | does OSC not work with v2? | 18:24 |
| mwhahaha | but we're using v3 creds against v2 | 18:25 |
| mwhahaha | i don't think that works | 18:25 |
| mfisch | v3 creds with a domain you mean | 18:25 |
| mwhahaha | yea | 18:25 |
| mfisch | okay well I consider this triaged | 18:26 |
| mfisch | we could at least toss a useful error | 18:26 |
| mfisch | can you add some comment too | 18:26 |
| mwhahaha | sure | 18:26 |
| mfisch | richm: since you're here, have you seen the 3 v3 related puppet-keystone bugs? https://etherpad.openstack.org/p/puppet-openstack-bugtriage | 18:26 |
| mfisch | mwhahaha: when you are ready I'd like to cover: https://bugs.launchpad.net/puppet-keystone/+bug/1344293 | 18:28 |
| openstack | Launchpad bug 1344293 in puppet-neutron "[library] neutron.rb uses cached call to keystone provider and can end up raising nil exception or otherwise fail to find resources created during this run of puppet" [Undecided,Confirmed] | 18:28 |
| mwhahaha | k let me look at that one | 18:30 |
| richm | mfisch: ok | 18:30 |
| richm | https://bugs.launchpad.net/puppet-keystone/+bug/1488538 | 18:30 |
| openstack | Launchpad bug 1488538 in puppet-keystone "keystone_user_role provider should be `domain` aware" [Undecided,New] - Assigned to Vasyl Saienko (vsaienko) | 18:30 |
| richm | This already works in Kilo - you specify keystone_user_role {'user::domain@project::domain'} | 18:31 |
| mfisch | so thats a wont fix | 18:31 |
| mfisch | can you update? | 18:32 |
| mwhahaha | so i don't think 1344293 is valid anymore | 18:34 |
| mwhahaha | as it appears the affected code is no longer there | 18:34 |
| mwhahaha | at least for keystone | 18:35 |
| mfisch | let me look | 18:35 |
| mfisch | sorry boss came by | 18:35 |
| mwhahaha | bosses always getting in the way :D | 18:36 |
| richm | https://bugs.launchpad.net/puppet-keystone/+bug/1488018 | 18:36 |
| openstack | Launchpad bug 1488018 in puppet-keystone "Does not work if v3 API variables are set in the environment" [Undecided,New] | 18:36 |
| richm | I think Gilles solution is "Don't set OS_URL, OS_AUTH_URL, or OS_IDENTITY_API_VERSION before running puppet" | 18:36 |
| richm | but I've asked him for clarification | 18:37 |
| mfisch | mwhahaha: can you ask bogdando if its still legit? | 18:37 |
| mfisch | richm: thanks | 18:37 |
| richm | so that also would be a Won't Fix | 18:37 |
| mfisch | a comment with "heres the fix is that okay | 18:37 |
| mwhahaha | mfisch: i can but i can tell you we don't have that code in 6.1 or 7.0 anymore | 18:37 |
| mfisch | richm: + marking incomplete | 18:37 |
| mfisch | they will expire off | 18:37 |
| mfisch | mwhahaha: what about upstream in puppet-keystone | 18:38 |
| mwhahaha | it's not in upstream either | 18:38 |
| mfisch | richm: do you have permissions to change prio? | 18:38 |
| mwhahaha | that's what i was refering too | 18:38 |
| mwhahaha | there's no tenant_hash anymore | 18:38 |
| mfisch | mwhahaha: okay I will fix | 18:38 |
| richm | mfisch: For 1488018 ? | 18:38 |
| mfisch | done | 18:38 |
| mfisch | richm: for either of the 2 we've talked about, if you dont have permissions to mark as Incomplete or Wont Fix let me know | 18:38 |
| richm | mfisch: I marked 1488538 as Fix Released since the fix was already in Kilo | 18:39 |
| richm | I would like to wait for feedback from Gilles before doing anything with 1488018 | 18:39 |
| richm | https://bugs.launchpad.net/puppet-keystone/+bug/1470635 | 18:40 |
| openstack | Launchpad bug 1470635 in puppet-keystone "endpoints added with v3 are not visible with v2" [High,Confirmed] | 18:40 |
| richm | This is really a Keystone bug - they would need to change Keystone to allow v2 clients to list endpoints created with v3 | 18:40 |
| richm | we've already worked around it in puppet-keystone | 18:40 |
| richm | mfisch: so I'm not sure how to dispose of this bug | 18:41 |
| mfisch | okay let me catch up | 18:41 |
| mfisch | https://bugs.launchpad.net/puppet-keystone/+bug/1488538 is fixed in kilo and master, yes? | 18:41 |
| openstack | Launchpad bug 1488538 in puppet-keystone "keystone_user_role provider should be `domain` aware" [Undecided,Fix released] - Assigned to Richard Megginson (rmeggins) | 18:41 |
| richm | mfisch: yes | 18:42 |
| mfisch | k | 18:42 |
| mfisch | now let me look at the other bug | 18:42 |
| richm | https://github.com/openstack/puppet-keystone/blob/stable/kilo/lib/puppet/provider/keystone_user_role/openstack.rb#L108 | 18:42 |
| mfisch | okay richm whats the work-around? | 18:42 |
| mfisch | good timing | 18:43 |
| richm | mfisch: the work around for the endpoint bug is that we always use the v2 api for endpoints in keystone_endpoint | 18:43 |
| mfisch | mwhahaha: can you look at 1416701 and 1416638 while we do this? they are related | 18:43 |
| mwhahaha | sure | 18:43 |
| mfisch | richm: can you comment to that affect and I will mark as Wont Fix | 18:44 |
| mfisch | sigh, prod issue, back later | 18:45 |
| richm | mfisch: done | 18:46 |
| pabelanger | should have something here in 20mins for the puppet-unit refactor | 18:47 |
| pabelanger | just testing now | 18:47 |
| mfisch | EmilienM: lets talk about https://bugs.launchpad.net/puppet-keystone/+bug/1423685 | 18:47 |
| openstack | Launchpad bug 1423685 in puppet-keystone "default service should be run under apache, eventlet will deprecated in K, Removed in M" [Wishlist,Triaged] | 18:47 |
| mfisch | EmilienM: I thought we'd already agreed to change the default? | 18:47 |
| EmilienM | yeah !!! | 18:47 |
| EmilienM | so | 18:48 |
| mwhahaha | mfisch: those two bugs are dupes, we should mark one the dupe of the other | 18:48 |
| *** iurygregory has joined #openstack-sprint | 18:48 | |
| mfisch | mwhahaha: go for it unless you lack perms | 18:48 |
| mfisch | and note on etherpad plz | 18:48 |
| EmilienM | mfisch: the plan is: deprecate it in liberty and drop it in M | 18:48 |
| mfisch | EmilienM: do we need a BP? | 18:48 |
| EmilienM | no | 18:48 |
| EmilienM | we already did it for liberty | 18:48 |
| EmilienM | mfisch: https://github.com/openstack/puppet-keystone/commit/6c827e49bddbc85875f77c3700b0ecaea58156ea | 18:49 |
| EmilienM | nibalizer: sorry for github url | 18:49 |
| EmilienM | mfisch: what I would love is to drop acceptance code that test eventlet asap | 18:49 |
| EmilienM | probably early M | 18:49 |
| nibalizer | EmilienM: :) | 18:50 |
| mfisch | EmilienM: so we can mark that fixed | 18:50 |
| EmilienM | mfisch: go ahead | 18:50 |
| mwhahaha | mfisch: kicked it back to the reporter because i'm not sure if he's getting an openstack error or a puppet error about duplicated resource definition | 18:51 |
| mfisch | mwhahaha: thanks | 18:51 |
| mfisch | that finishes keystone and now I really need to look at this issue lets take a 15 min break | 18:51 |
| mwhahaha | k, go put out some fires | 18:51 |
| EmilienM | crinkle: I though puppet-unit-latest was not voting, but it does (which is fine but I thought it was not the plan) | 18:53 |
| crinkle | EmilienM: i have no information | 18:53 |
| crinkle | EmilienM: i agree that is surprising | 18:54 |
| nibalizer | pabelanger: so where are we on splitting up those rspec tests? | 18:54 |
| EmilienM | crinkle: well, it works now so... let's keep it voting until it fails and block us... | 18:54 |
| mfisch | mwhahaha: I'm somewhat back | 19:07 |
| mfisch | I wanted to discuss https://bugs.launchpad.net/puppet-glance/+bug/1393490 with mgagne but he's nothere | 19:07 |
| openstack | Launchpad bug 1393490 in puppet-glance "Cannot manage Glance package version" [Low,Triaged] | 19:07 |
| mwhahaha | we may want to make that more general for all modules, not sure all of them support the ability to manage specific versions | 19:08 |
| mwhahaha | some do, some don't | 19:08 |
| mwhahaha | i checked glance and it does allow for the python client version to be specified | 19:08 |
| mwhahaha | but not the api or registry | 19:08 |
| EmilienM | mfisch: I pinged him to join | 19:09 |
| *** mgagne has joined #openstack-sprint | 19:09 | |
| mgagne | hi | 19:09 |
| mwhahaha | as for https://bugs.launchpad.net/puppet-nova/+bug/1350966, i wonder how much effort we want to put into that since we were going to deprecate the rabbitmq config out of nova | 19:09 |
| openstack | Launchpad bug 1350966 in puppet-nova "Incorrect dependency on puppetlabs-rabbitmq" [Undecided,Confirmed] - Assigned to Richard Raseley (richard-raseley) | 19:09 |
| mwhahaha | mgagne: mfisch wanted to talk about https://bugs.launchpad.net/puppet-glance/+bug/1393490 | 19:11 |
| openstack | Launchpad bug 1393490 in puppet-glance "Cannot manage Glance package version" [Low,Triaged] | 19:11 |
| mgagne | AFAIK, it got fixed since: https://github.com/openstack/puppet-glance/blob/master/manifests/api.pp#L250 https://github.com/openstack/puppet-glance/blob/master/manifests/registry.pp#L179 | 19:12 |
| mgagne | mfisch: ^ | 19:12 |
| mfisch | okay can we mark it fixed? | 19:12 |
| *** mrmartin has quit IRC | 19:13 | |
| mgagne | yes | 19:13 |
| mfisch | thanks | 19:15 |
| mfisch | okay | 19:15 |
| mfisch | mwhahaha: this nova one https://bugs.launchpad.net/puppet-nova/+bug/1350966 | 19:15 |
| openstack | Launchpad bug 1350966 in puppet-nova "Incorrect dependency on puppetlabs-rabbitmq" [Undecided,Confirmed] - Assigned to Richard Raseley (richard-raseley) | 19:15 |
| mfisch | is Richard R online? | 19:15 |
| mwhahaha | well based on the irc meeting we are going to deprecate the rabbitmq out of puppet-nova | 19:16 |
| mwhahaha | so we can fix it only to remove it next cycle | 19:16 |
| mwhahaha | the rabbitmq server config that is | 19:16 |
| mfisch | okay can you comment to that and ask Richard to close if he agrees? | 19:16 |
| mwhahaha | sure | 19:16 |
| mfisch | okay I want to skip around some | 19:19 |
| mfisch | this one | 19:19 |
| mfisch | https://bugs.launchpad.net/puppet-neutron/+bug/1420586 | 19:19 |
| openstack | Launchpad bug 1420586 in puppet-neutron "neutron net-show and subnet-show don't scale" [High,Confirmed] | 19:19 |
| mfisch | this is one I filed | 19:19 |
| mfisch | this one is so bad that we gave up managing networks and subnets with neutron | 19:19 |
| mfisch | instead I exec a python script if needed | 19:20 |
| mfisch | its horrible but this was adding 45 minutes per deploy to us | 19:20 |
| mgagne | BTW, I'll switch development focus on LP to liberty instead of kilo | 19:21 |
| mfisch | +1 | 19:21 |
| mfisch | mwhahaha: maybe there's not much to say here | 19:22 |
| mwhahaha | sec still writing something | 19:22 |
| mfisch | k | 19:22 |
| mwhahaha | but if it doesn't scale the question would be how many people at scale are using the functions that require net-show/subnet-show | 19:22 |
| mfisch | and it runs on every control node we have | 19:23 |
| mfisch | You can disable it by removing all the nets and subnets which is what we did | 19:23 |
| mwhahaha | hmm | 19:23 |
| mwhahaha | let me look | 19:23 |
| mfisch | dorman and kris arent here I'd ask them | 19:23 |
| mfisch | Fuel may have the same issue | 19:23 |
| mfisch | if you do a rolling delpoy like us its more painful | 19:23 |
| mwhahaha | we don't :D | 19:24 |
| mfisch | but the catalog runs were taking 15 minutes per control node, over half was this | 19:24 |
| mwhahaha | since we don't run puppet constantly it doesn't really affect fuel | 19:24 |
| clayton | mwhahaha: it's pretty horrible, it does net-list and then net-show one every network and subnet | 19:24 |
| clayton | and no token caching, so it takes a few seconds per network and subnetwork | 19:25 |
| mfisch | when we deploy we do 1 control node, then the next, then the next | 19:25 |
| mgagne | done | 19:25 |
| clayton | once you have a few hundred of those, it's untenable. | 19:25 |
| clayton | my conclusion was that we were the only people using this with more than a few networks | 19:25 |
| mfisch | then lets leave this as-is | 19:26 |
| mfisch | triage wise | 19:26 |
| mfisch | I will note the work around | 19:26 |
| mwhahaha | is there a way to make it optional? | 19:26 |
| mwhahaha | like don't do that unless someone is explictly managing networks | 19:27 |
| clayton | mwhahaha: I believe neutron actually returns all the information it needs from the query in the net-list, it's just not displayed by default | 19:27 |
| mfisch | can the openstack client handle -f? | 19:27 |
| mfisch | for show certain fields? | 19:27 |
| clayton | I suspect a clever person could fix this with neutron net-list -c -c -c or openstack client | 19:28 |
| clayton | I didn't know about telling neutron to display extra fields when we replaced this with a python script and I didn't have the time to wedge openstack client into the neutron module | 19:28 |
| clayton | in fact, you need to do this - neutron net-list -c id -c name -c admin_state_up -c provider:network_type -c provider:physical_network -c provider:segmentation_id -c router:external -c shared -c tenant_id | 19:30 |
| mfisch | does the provider use os-client? | 19:31 |
| mfisch | or neutron | 19:31 |
| mfisch | neutron right? | 19:31 |
| clayton | neutron | 19:31 |
| clayton | that's all the attributes that the neutron_network provider cares about | 19:32 |
| pabelanger | nibalizer: have something local, but ran into an issue with zuul | 19:33 |
| pabelanger | trying to figure out solution | 19:33 |
| mfisch | clayton: sounds like we could fix this since we're the only ones who are hit by it | 19:36 |
| clayton | sure, not sure we have much incentive to at this point though | 19:36 |
| nibalizer | pabelanger: okay | 19:36 |
| nibalizer | pabelanger: let me know how I can help | 19:36 |
| clayton | I'm willing to in theory, but in practice it's not very high up the list of things to do | 19:36 |
| mfisch | agree | 19:36 |
| mfisch | lets move on | 19:36 |
| mfisch | working up | 19:37 |
| mfisch | https://bugs.launchpad.net/puppet-neutron/+bug/1344271 | 19:37 |
| openstack | Launchpad bug 1344271 in puppet-neutron "neutron::server::notifications settings are actually required" [Undecided,Incomplete] | 19:37 |
| mwhahaha | i'm sure that one because we would pull in neutron::server separately from that class | 19:39 |
| mwhahaha | but that's my guess from the description, i'll have to look at the code | 19:39 |
| mwhahaha | may not be valid anymore | 19:40 |
| mfisch | ok | 19:40 |
| mfisch | it was marked incomplete in June without update | 19:41 |
| mfisch | it should be closed | 19:41 |
| mfisch | mwhahaha: any objections? | 19:42 |
| mwhahaha | Go ahead and close it and i'll look into it later and see if it should be reopened | 19:42 |
| * mfisch is following ubuntu bug control policy here ;) | 19:43 | |
| mfisch | moving up again | 19:43 |
| mfisch | https://bugs.launchpad.net/puppet-neutron/+bug/1474204 | 19:43 |
| openstack | Launchpad bug 1474204 in puppet-neutron "vlan id are invalid" [Undecided,New] | 19:43 |
| mfisch | is this really a puppet bug? | 19:43 |
| mwhahaha | oh going back to that previous bug | 19:44 |
| mfisch | sure | 19:44 |
| mwhahaha | iit's saying that the values have been set to true but unless you include it, neutron won't have the nova creds | 19:44 |
| mwhahaha | i think it's still valid | 19:45 |
| mfisch | ok | 19:45 |
| mwhahaha | because the default is now true but we don't provide the nova creds anywhere | 19:45 |
| mfisch | can you reopen and add that comment? | 19:45 |
| mwhahaha | sure | 19:45 |
| mfisch | the latter bug I moved to neutron | 19:45 |
| mfisch | they will likely close it | 19:45 |
| mwhahaha | k | 19:47 |
| mfisch | ready for https://bugs.launchpad.net/puppet-neutron/+bug/1480277 | 19:47 |
| openstack | Launchpad bug 1480277 in puppet-neutron "linuxbridge agent module in Kilo branch is broken" [Undecided,New] | 19:47 |
| mfisch | ? | 19:47 |
| EmilienM | richm: reviewing https://review.openstack.org/#/c/213957/13/manifests/endpoint.pp,cm - why not using undef? | 19:48 |
| mwhahaha | wonder if that's a packaging issue | 19:48 |
| mwhahaha | similar to the neutron_lbaas.conf missing | 19:48 |
| EmilienM | mwhahaha: no, crinkle did a patch for that | 19:48 |
| mfisch | he's saying the above folder DNE | 19:48 |
| mfisch | the containing folder | 19:48 |
| mfisch | this code is weird | 19:48 |
| EmilienM | https://review.openstack.org/#/c/216523/ | 19:48 |
| EmilienM | oh sorry | 19:49 |
| EmilienM | it's not this bug | 19:49 |
| mwhahaha | that patch might be the fix | 19:50 |
| mfisch | you'd also need to change params.pp | 19:51 |
| mfisch | $linuxbridge_config_file = '/etc/neutron/plugins/linuxbridge/linuxbridge_conf.ini' | 19:51 |
| mfisch | same for ubuntu and rhel | 19:52 |
| mwhahaha | that's not used | 19:52 |
| mwhahaha | probably left over from a previous version | 19:52 |
| mfisch | well it is in 1 place | 19:52 |
| mwhahaha | where | 19:52 |
| mwhahaha | my grep for linuxbridge_config_file returned only params | 19:53 |
| mfisch | in manifests/plugins/linuxbridge.pp | 19:53 |
| mfisch | let me make sure I',m on the right tree | 19:53 |
| EmilienM | richm: err, undef is actually an option, never-mind my comment | 19:53 |
| mfisch | mwhahaha: err I dont see it all anymore on master | 19:53 |
| mfisch | do you? | 19:54 |
| mfisch | I did a clean checkout | 19:54 |
| mfisch | ^%$%$ | 19:54 |
| mfisch | my remote was at stackforge | 19:55 |
| mwhahaha | :D | 19:55 |
| mwhahaha | right so the params.pp needs cleaning up and crinkle's patch should fix it | 19:55 |
| mfisch | which? | 19:55 |
| mwhahaha | https://review.openstack.org/#/c/216523/ | 19:55 |
| mfisch | I dont do much with puppet-neutron | 19:56 |
| mfisch | my contrib of the day | 19:57 |
| mfisch | https://review.openstack.org/219860 | 19:57 |
| mwhahaha | :D | 19:57 |
| mfisch | marking colleen's as a fix | 19:57 |
| mfisch | mwhahaha: I have to go to the DMV now | 19:58 |
| mfisch | I will tether in from there if there is a line | 19:58 |
| mwhahaha | good luck with that | 19:58 |
| mfisch | vinsh: are yo uon? | 19:58 |
| mfisch | you on | 19:58 |
| vinsh | hey | 19:58 |
| EmilienM | mfisch, mwhahaha: fyi, puppet-neutron CI is broken now, I'm unblocking it | 19:59 |
| mfisch | vinsh: can you and mwhahaha do the swift bug triage? | 19:59 |
| mfisch | https://etherpad.openstack.org/p/puppet-openstack-bugtriage | 19:59 |
| mfisch | just dig a bit, see if fixed/fixable update severity etc | 19:59 |
| mfisch | no need to actually fix unles you want | 19:59 |
| mfisch | if not I will catch up after I get my license plates | 19:59 |
| vinsh | Yeah, I could walk through those. mwhahaha you have time? | 20:00 |
| mwhahaha | sure, but let me take a 5 min break. | 20:00 |
| mwhahaha | brb | 20:00 |
| vinsh | kk | 20:00 |
| EmilienM | clarkb: in case you missed mfisch's comment on https://review.openstack.org/#/c/216926/ | 20:04 |
| EmilienM | damn | 20:04 |
| EmilienM | sorry clarkb | 20:04 |
| EmilienM | clayton: https://review.openstack.org/#/c/216926/ | 20:04 |
| clayton | ah, I didn't see that. replied. | 20:06 |
| EmilienM | if anyone can review two puppet-tempest patches: https://review.openstack.org/#/c/218467/ and https://review.openstack.org/#/c/218398/ | 20:06 |
| clayton | I'm fine with adopting the same approach for clients also, but it's less clear what the solution for that would be. For services its' easier, since they access the filesystem in pretty predictable ways | 20:07 |
| clayton | I've been focused on the service side part of this. | 20:07 |
| mwhahaha | vinsh: i'm back | 20:11 |
| vinsh | Right on, first on the list 1466707 | 20:11 |
| vinsh | Sounds like a doc update would help | 20:14 |
| mwhahaha | yea | 20:14 |
| vinsh | Well, I can take that one on | 20:15 |
| vinsh | i'll make note to have mfisch assign it to me when he gets back | 20:15 |
| vinsh | next | 20:16 |
| mwhahaha | https://bugs.launchpad.net/puppet-swift/+bug/1291433 | 20:16 |
| openstack | Launchpad bug 1291433 in puppet-swift "swift storage services can start before databases are rsync'd" [Undecided,Incomplete] - Assigned to David Moreau Simard (dmsimard) | 20:16 |
| vinsh | I feel like that one is no longer an issue... ive never come close to hitting it :) | 20:18 |
| mwhahaha | so i wonder if there's a way to do the check optionally | 20:18 |
| mwhahaha | it is quite old | 20:18 |
| mwhahaha | might be worth at least documenting a work around | 20:19 |
| vinsh | whats up with all these launch pad patches and not a gerrit review? | 20:19 |
| mwhahaha | good question | 20:20 |
| mfisch | which one to you vinsh ? | 20:22 |
| vinsh | first one. I made a note for you | 20:22 |
| vinsh | on etherpad | 20:23 |
| mfisch | lemme reconnect to ep | 20:23 |
| pabelanger | https://review.openstack.org/#/c/219866/ | 20:23 |
| pabelanger | nibalizer: ^ | 20:24 |
| vinsh | mwhahaha: They could probably just add " File['/etc/swift/container.ring.gz'] ~> Class['swift::storage::container']" to their manifest to solve that | 20:25 |
| nibalizer | where does {puppet_unit_version} come from | 20:25 |
| mwhahaha | vinsh: yea which is why i said it should just be documented | 20:25 |
| EmilienM | jamespage: see https://jenkins04.openstack.org/job/gate-puppet-openstack-integration-dsvm-trusty/43/console | 20:26 |
| vinsh | mwhahaha: Ok, so read me update? | 20:26 |
| EmilienM | "WARNING: The following packages cannot be authenticated!" | 20:26 |
| mwhahaha | vinsh: it seems to be something that you can't really solve in the module as it's more of a operator usage thing | 20:26 |
| mwhahaha | vinsh: yea that would be good | 20:26 |
| EmilienM | jamespage: using 'proposed' | 20:26 |
| vinsh | I'll take it on also, seems easy enough to document. | 20:26 |
| nibalizer | pabelanger: where does {puppet_unit_version} come from | 20:27 |
| pabelanger | I'd like to get this one too: https://review.openstack.org/#/c/215696/ | 20:27 |
| pabelanger | nibalizer: the job-group below | 20:27 |
| nibalizer | pabelanger: is there a review containing run_multinode.sh? | 20:28 |
| vinsh | mwhahaha: https://bugs.launchpad.net/puppet-swift/+bug/1289631 | 20:28 |
| openstack | Launchpad bug 1289631 in puppet-swift "Do security hardening for /etc/swift" [High,Triaged] | 20:28 |
| nibalizer | I don't understand why the unit tests have an {ostype} | 20:28 |
| nibalizer | that seems irrelevant to this change but why? | 20:29 |
| mfisch | guys about to take my turn at DMV let me know updates later thanks vinsh and mwhahaha | 20:29 |
| pabelanger | http://paste.openstack.org/show/442385/ JJB output | 20:29 |
| mwhahaha | k | 20:29 |
| pabelanger | ya, ostype could be removed | 20:29 |
| pabelanger | following patch could do that | 20:29 |
| vinsh | mwhahaha: wow, never would have considered this one. | 20:31 |
| mwhahaha | because it doesn't work | 20:31 |
| mwhahaha | and it's a bad idea? :D | 20:31 |
| mwhahaha | recurse doesn't work like that | 20:31 |
| mwhahaha | at at least i don't think it does | 20:31 |
| vinsh | Well, not sure where to go with this one. I don't have time in the next 6 weeks to touch it. | 20:33 |
| pabelanger | nibalizer: nothing yet, this just setups on the nodepool config | 20:33 |
| pabelanger | have something locally I want to push up | 20:33 |
| pabelanger | I've been testing something with ansible-playbook --connection=local | 20:33 |
| mwhahaha | i'd be concerned about the load on the system with recurse = true if it does actually manage all of the items in that folder | 20:33 |
| pabelanger | however, once we get the experimental job merged, we'll have a place to sandbox some of the nodepool functionality | 20:34 |
| mwhahaha | that one seems triaged, it just needs someone to actuallyt est it | 20:34 |
| pabelanger | which, is something I'd like right now | 20:34 |
| vinsh | mwhahaha: the load on /etc/swift? | 20:34 |
| vinsh | not a lot in there. | 20:34 |
| pabelanger | otherwise, I have an internal nodepool I can play with | 20:34 |
| pabelanger | but, would like people to monitor results | 20:34 |
| mwhahaha | vinsh: on the constant checking of perms, the other issue is that really only addresses people who are running puppet constantly | 20:35 |
| vinsh | checking only on each puppet run | 20:35 |
| clayton | I think the advise you see against not using recurse has more to do with people doing it with directories that have a bazillion files | 20:35 |
| vinsh | doesn't seem like a big deal really.. | 20:35 |
| clayton | I've done that before, it's horrible. <100 files has never been an issue | 20:35 |
| mwhahaha | right it should be ok on small directories | 20:36 |
| mwhahaha | but how are we to know what a user puts in /etc/swift | 20:36 |
| vinsh | /etc/swift is >20 files | 20:36 |
| vinsh | I mean < | 20:36 |
| mwhahaha | it should be < 20 files | 20:36 |
| EmilienM | pabelanger: http://logs.openstack.org/18/219818/1/check/gate-puppet-openstack-integration-dsvm-trusty/8d00d68/console.html#_2015-09-02_20_25_04_531 | 20:36 |
| mwhahaha | doesn't mean it is for all cases | 20:36 |
| mwhahaha | anyway back to the triage part, this bug seems fine as is. just needs someone to try it out i guess and make sure it doesn't do anything crazy | 20:37 |
| vinsh | yep, add it to the backlog. | 20:37 |
| *** rfolco has quit IRC | 20:38 | |
| EmilienM | pabelanger: see https://review.openstack.org/#/q/status:open++branch:master+topic:beaker/trusty-proposed,n,z | 20:39 |
| EmilienM | all the same patches | 20:39 |
| mwhahaha | so moving on, https://bugs.launchpad.net/puppet-swift/+bug/1484614 | 20:40 |
| openstack | Launchpad bug 1484614 in puppet-swift "make "log_requests" flag configurable, reduce account/object/container server logs" [Undecided,New] | 20:40 |
| vinsh | Again.. why didn't he put the review in gerrit? :) | 20:40 |
| mwhahaha | was just going to say that | 20:41 |
| vinsh | I would assign the bug to him and let him propose one as he says he can | 20:41 |
| mwhahaha | yes | 20:41 |
| vinsh | I use rysyslog to log.. and I like lots of log data. | 20:41 |
| vinsh | To each his own :) | 20:41 |
| vinsh | i'll comment in the bug | 20:42 |
| mwhahaha | k | 20:42 |
| mwhahaha | next up https://bugs.launchpad.net/puppet-swift/+bug/1484612 | 20:42 |
| openstack | Launchpad bug 1484612 in puppet-swift "enable configuration of [incoming|outgoing]_chmod in ::swift::storage" [Undecided,New] | 20:42 |
| mwhahaha | same guy | 20:43 |
| mwhahaha | same deal | 20:43 |
| vinsh | I think this one is already proposed | 20:43 |
| *** guimaluf has joined #openstack-sprint | 20:43 | |
| vinsh | https://review.openstack.org/#/c/217707/ | 20:44 |
| vinsh | kinda | 20:44 |
| mwhahaha | yup | 20:44 |
| vinsh | suppose we assign it to danp | 20:46 |
| vinsh | since he is working it | 20:46 |
| mwhahaha | sounds ok | 20:46 |
| vinsh | maybe he saw the report. maybe not. :) | 20:46 |
| vinsh | oh the next one is assigned to me :-o | 20:47 |
| EmilienM | mfisch, crinkle: so here are the patches to use proposed repo - you have to know, horizon is broken, ironic too and glanceclient too. that's it - https://review.openstack.org/#/q/status:open++branch:master+topic:beaker/trusty-proposed,n,z | 20:47 |
| mwhahaha | oh it's a dupe | 20:47 |
| vinsh | oh man. i must need coffee | 20:47 |
| vinsh | lol yep. | 20:47 |
| mwhahaha | of the first one | 20:47 |
| mwhahaha | we'll just make that one go away | 20:48 |
| vinsh | Right on. | 20:48 |
| vinsh | swift is done/traiged. mfisch. | 20:48 |
| vinsh | actions listed by each bug for you. | 20:48 |
| EmilienM | mfisch, crinkle: I updated https://etherpad.openstack.org/p/puppet-liberty-blocker so you can see the diff between the 2 repos | 20:50 |
| crinkle | EmilienM: ceilometer is broken too? | 20:58 |
| EmilienM | crinkle: on centos only | 20:58 |
| EmilienM | crinkle: https://bugs.launchpad.net/puppet-ceilometer/+bug/1490986 | 20:58 |
| openstack | Launchpad bug 1490986 in puppet-ceilometer "centos: puppet run is not idempotent (collector & notif agents)" [Critical,In progress] - Assigned to Emilien Macchi (emilienm) | 20:58 |
| EmilienM | https://bugzilla.redhat.com/show_bug.cgi?id=1259075 | 20:59 |
| openstack | bugzilla.redhat.com bug 1259075 in Package Review "Review Request: python-jsonpath-rw-ext - Extensions for JSONPath RW" [Medium,Assigned] - Assigned to rmarko | 20:59 |
| EmilienM | this is a new dep missing | 20:59 |
| EmilienM | RDO guys are doing it right now at this time | 20:59 |
| EmilienM | pabelanger: https://review.openstack.org/#/c/219818 | 21:02 |
| pabelanger | EmilienM: sent again please | 21:02 |
| EmilienM | crinkle: can you also approve https://review.openstack.org/#/c/219393/ ? | 21:03 |
| EmilienM | pabelanger: https://review.openstack.org/#/c/219818 | 21:03 |
| EmilienM | pabelanger: https://review.openstack.org/219804 | 21:03 |
| mfisch | here now | 21:05 |
| mfisch | no new plates just a stupid piece of paper | 21:05 |
| mfisch | mwhahaha: vinsh thanks | 21:05 |
| mfisch | let me read the ep | 21:05 |
| vinsh | :) | 21:06 |
| EmilienM | crinkle: I'm highly concerned by https://review.openstack.org/219818 failures | 21:06 |
| EmilienM | because trusty fails | 21:07 |
| mfisch | mwhahaha: can we pick up the last few tomorrow? | 21:07 |
| mwhahaha | yea sure | 21:07 |
| EmilienM | and it's not failing on puppet-* modules patches because it still adds liberty repo | 21:07 |
| mwhahaha | i'm out of here in a few mins | 21:07 |
| EmilienM | that means, if we merge my patches that use 'proposed' repo, it will break CI | 21:08 |
| EmilienM | since https://review.openstack.org/219818 is failung | 21:08 |
| EmilienM | failing* | 21:08 |
| nibalizer | EmilienM: pabelanger https://review.openstack.org/#/c/219879/ I think will add nonvoting puppet4 beaker tests | 21:09 |
| EmilienM | crinkle: I'm gonna but depends-on on the openstack-integ patch | 21:09 |
| EmilienM | to make sure we don't break our CI | 21:09 |
| mfisch | nice work guys, we only have 6 bugs for tomorrow | 21:10 |
| crinkle | EmilienM: :( my +2s disappeared | 21:12 |
| EmilienM | crinkle: because I added Depends-On for safety | 21:13 |
| EmilienM | crinkle: do you see why I added it? | 21:13 |
| pabelanger | crinkle: nibalizer: EmilienM: lets get https://review.openstack.org/#/c/217346/ merged so we can get a proper depends on for puppet-keystone on puppet-openstack-integration | 21:13 |
| crinkle | EmilienM: yes but :( | 21:13 |
| pabelanger | then you don't have to manually add Depends-On jobs | 21:13 |
| EmilienM | crinkle: you must be angry at me :-P | 21:13 |
| EmilienM | pabelanger: looking | 21:14 |
| nibalizer | pabelanger: I don't understand why we need 217346 | 21:14 |
| EmilienM | crinkle: I think 'proposed' is not installing keystone | 21:15 |
| EmilienM | if I'm right, it really sucks for us | 21:15 |
| pabelanger | nibalizer: so EmilienM can add gate-puppet-keystone-beaker... gate on puppet-openstack-integration | 21:16 |
| pabelanger | which didn't work before because the clone was not properly done | 21:16 |
| EmilienM | crinkle: but looking at http://logs.openstack.org/18/219818/1/check/gate-puppet-openstack-integration-dsvm-trusty/8d00d68/console.html#_2015-09-02_20_25_04_531 from https://review.openstack.org/#/c/219818/ - it looks like it's still broken | 21:16 |
| nibalizer | pabelanger: I don't understand, I'll need a bit more explanation (preferably in the commit message) | 21:17 |
| nibalizer | oh wait maybe I do understand | 21:17 |
| nibalizer | so gate-keystone will get the keystone sourcecode instead of the project sourcecode | 21:18 |
| pabelanger | currently gate-*-beaker uses the jenkins workspace and git-prep | 21:18 |
| nibalizer | yep | 21:19 |
| EmilienM | this is a canari job | 21:19 |
| nibalizer | okay I understand now | 21:19 |
| pabelanger | ok | 21:19 |
| nibalizer | I'm not sure others would though, I would like you to explain it in the commit message please | 21:19 |
| EmilienM | http://goo.gl/tgtiqq | 21:19 |
| EmilienM | nibalizer, crinkle: we are off now | 21:21 |
| EmilienM | crinkle: I'll look status of https://review.openstack.org/#/q/status:open++branch:master+topic:beaker/trusty-proposed,n,z - and see if we should go to 'proposed' repo or not | 21:21 |
| EmilienM | see you later folks | 21:23 |
| nibalizer | okay | 21:23 |
| nibalizer | pabelanger: so f22/23 question | 21:23 |
| nibalizer | the puppet, puppet4 package | 21:24 |
| nibalizer | does it use /etc/puppetlabs? | 21:24 |
| * mwhahaha wanders off | 21:33 | |
| *** krotscheck is now known as kro_paternity | 23:34 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!